Vendor import of clang trunk r321017:vendor/clang/clang-trunk-r321017

https://llvm.org/svn/llvm-project/cfe/trunk@321017

304 files changed, 30619 insertions, 3038 deletions

diff --git a/test/OpenMP/atomic_capture_codegen.cpp b/test/OpenMP/atomic_capture_codegen.cpp
index 0a22e4276b53..306b83f62463 100644
--- a/test/OpenMP/atomic_capture_codegen.cpp
+++ b/test/OpenMP/atomic_capture_codegen.cpp
@@ -1,6 +1,6 @@
-// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -target-cpu core2 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
 // expected-no-diagnostics
 #ifndef HEADER
 #define HEADER
@@ -611,50 +611,6 @@ int main() {
 // CHECK: store i8 [[DESIRED_I8]], i8* @{{.+}},
 #pragma omp atomic capture
   {bx = civ - bx; bv = bx;}
-// CHECK: [[EXPR_RE:%.+]] = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @{{.+}}, i32 0, i32 0)
-// CHECK: [[EXPR_IM:%.+]] = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @{{.+}}, i32 0, i32 1)
-// CHECK: [[X:%.+]] = load atomic i16, i16* [[X_ADDR:@.+]] monotonic
-// CHECK: br label %[[CONT:.+]]
-// CHECK: [[CONT]]
-// CHECK: [[EXPECTED:%.+]] = phi i16 [ [[X]], %{{.+}} ], [ [[OLD_X:%.+]], %[[CONT]] ]
-// CHECK: [[CONV:%.+]] = zext i16 [[EXPECTED]] to i32
-// CHECK: [[X_RVAL:%.+]] = sitofp i32 [[CONV]] to float
-// <Skip checks for complex calculations>
-// CHECK: [[X_RE_ADDR:%.+]] = getelementptr inbounds { float, float }, { float, float }* [[TEMP:%.+]], i32 0, i32 0
-// CHECK: [[X_RE:%.+]] = load float, float* [[X_RE_ADDR]]
-// CHECK: [[X_IM_ADDR:%.+]] = getelementptr inbounds { float, float }, { float, float }* [[TEMP]], i32 0, i32 1
-// CHECK: [[X_IM:%.+]] = load float, float* [[X_IM_ADDR]]
-// CHECK: [[NEW:%.+]] = fptoui float [[X_RE]] to i16
-// CHECK: store i16 [[NEW]], i16* [[TEMP:%.+]],
-// CHECK: [[DESIRED:%.+]] = load i16, i16* [[TEMP]],
-// CHECK: [[RES:%.+]] = cmpxchg i16* [[X_ADDR]], i16 [[EXPECTED]], i16 [[DESIRED]] monotonic monotonic
-// CHECK: [[OLD_X]] = extractvalue { i16, i1 } [[RES]], 0
-// CHECK: [[SUCCESS_FAIL:%.+]] = extractvalue { i16, i1 } [[RES]], 1
-// CHECK: br i1 [[SUCCESS_FAIL]], label %[[EXIT:.+]], label %[[CONT]]
-// CHECK: [[EXIT]]
-// CHECK: store i16 [[NEW]], i16* @{{.+}},
-#pragma omp atomic capture
-  usv = usx /= cfv;
-// CHECK: [[EXPR_RE:%.+]] = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @{{.+}}, i32 0, i32 0)
-// CHECK: [[EXPR_IM:%.+]] = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @{{.+}}, i32 0, i32 1)
-// CHECK: [[X:%.+]] = load atomic i64, i64* [[X_ADDR:@.+]] monotonic
-// CHECK: br label %[[CONT:.+]]
-// CHECK: [[CONT]]
-// CHECK: [[EXPECTED:%.+]] = phi i64 [ [[X]], %{{.+}} ], [ [[OLD_X:%.+]], %[[CONT]] ]
-// CHECK: [[X_RVAL:%.+]] = sitofp i64 [[EXPECTED]] to double
-// CHECK: [[ADD_RE:%.+]] = fadd double [[X_RVAL]], [[EXPR_RE]]
-// CHECK: [[ADD_IM:%.+]] = fadd double 0.000000e+00, [[EXPR_IM]]
-// CHECK: [[DESIRED:%.+]] = fptosi double [[ADD_RE]] to i64
-// CHECK: store i64 [[DESIRED]], i64* [[TEMP:%.+]],
-// CHECK: [[DESIRED:%.+]] = load i64, i64* [[TEMP]],
-// CHECK: [[RES:%.+]] = cmpxchg i64* [[X_ADDR]], i64 [[EXPECTED]], i64 [[DESIRED]] monotonic monotonic
-// CHECK: [[OLD_X]] = extractvalue { i64, i1 } [[RES]], 0
-// CHECK: [[SUCCESS_FAIL:%.+]] = extractvalue { i64, i1 } [[RES]], 1
-// CHECK: br i1 [[SUCCESS_FAIL]], label %[[EXIT:.+]], label %[[CONT]]
-// CHECK: [[EXIT]]
-// CHECK: store i64 [[EXPECTED]], i64* @{{.+}},
-#pragma omp atomic capture
-  {llv = llx; llx += cdv;}
 // CHECK: [[IDX:%.+]] = load i16, i16* @{{.+}}
 // CHECK: load i8, i8*
 // CHECK: [[VEC_ITEM_VAL:%.+]] = zext i1 %{{.+}} to i32
diff --git a/test/OpenMP/atomic_read_codegen.c b/test/OpenMP/atomic_read_codegen.c
index 0cd46e3821fd..0cfb2d26f243 100644
--- a/test/OpenMP/atomic_read_codegen.c
+++ b/test/OpenMP/atomic_read_codegen.c
@@ -1,6 +1,6 @@
-// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -target-cpu core2 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
 // expected-no-diagnostics
 // REQUIRES: x86-registered-target
 #ifndef HEADER
diff --git a/test/OpenMP/atomic_update_codegen.cpp b/test/OpenMP/atomic_update_codegen.cpp
index 8c02a43d64e3..1343cd8ad2cd 100644
--- a/test/OpenMP/atomic_update_codegen.cpp
+++ b/test/OpenMP/atomic_update_codegen.cpp
@@ -1,6 +1,6 @@
-// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -target-cpu core2 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
 // expected-no-diagnostics
 #ifndef HEADER
 #define HEADER
@@ -554,48 +554,6 @@ int main() {
 // CHECK: [[EXIT]]
 #pragma omp atomic
   bx = civ - bx;
-// CHECK: [[EXPR_RE:%.+]] = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @{{.+}}, i32 0, i32 0)
-// CHECK: [[EXPR_IM:%.+]] = load float, float* getelementptr inbounds ({ float, float }, { float, float }* @{{.+}}, i32 0, i32 1)
-// CHECK: [[X:%.+]] = load atomic i16, i16* [[X_ADDR:@.+]] monotonic
-// CHECK: br label %[[CONT:.+]]
-// CHECK: [[CONT]]
-// CHECK: [[EXPECTED:%.+]] = phi i16 [ [[X]], %{{.+}} ], [ [[OLD_X:%.+]], %[[CONT]] ]
-// CHECK: [[CONV:%.+]] = zext i16 [[EXPECTED]] to i32
-// CHECK: [[X_RVAL:%.+]] = sitofp i32 [[CONV]] to float
-// <Skip checks for complex calculations>
-// CHECK: [[X_RE_ADDR:%.+]] = getelementptr inbounds { float, float }, { float, float }* [[TEMP:%.+]], i32 0, i32 0
-// CHECK: [[X_RE:%.+]] = load float, float* [[X_RE_ADDR]]
-// CHECK: [[X_IM_ADDR:%.+]] = getelementptr inbounds { float, float }, { float, float }* [[TEMP]], i32 0, i32 1
-// CHECK: [[X_IM:%.+]] = load float, float* [[X_IM_ADDR]]
-// CHECK: [[DESIRED:%.+]] = fptoui float [[X_RE]] to i16
-// CHECK: store i16 [[DESIRED]], i16* [[TEMP:%.+]]
-// CHECK: [[DESIRED:%.+]] = load i16, i16* [[TEMP]]
-// CHECK: [[RES:%.+]] = cmpxchg i16* [[X_ADDR]], i16 [[EXPECTED]], i16 [[DESIRED]] monotonic monotonic
-// CHECK: [[OLD_X]] = extractvalue { i16, i1 } [[RES]], 0
-// CHECK: [[SUCCESS_FAIL:%.+]] = extractvalue { i16, i1 } [[RES]], 1
-// CHECK: br i1 [[SUCCESS_FAIL]], label %[[EXIT:.+]], label %[[CONT]]
-// CHECK: [[EXIT]]
-#pragma omp atomic update
-  usx /= cfv;
-// CHECK: [[EXPR_RE:%.+]] = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @{{.+}}, i32 0, i32 0)
-// CHECK: [[EXPR_IM:%.+]] = load double, double* getelementptr inbounds ({ double, double }, { double, double }* @{{.+}}, i32 0, i32 1)
-// CHECK: [[X:%.+]] = load atomic i64, i64* [[X_ADDR:@.+]] monotonic
-// CHECK: br label %[[CONT:.+]]
-// CHECK: [[CONT]]
-// CHECK: [[EXPECTED:%.+]] = phi i64 [ [[X]], %{{.+}} ], [ [[OLD_X:%.+]], %[[CONT]] ]
-// CHECK: [[X_RVAL:%.+]] = sitofp i64 [[EXPECTED]] to double
-// CHECK: [[ADD_RE:%.+]] = fadd double [[X_RVAL]], [[EXPR_RE]]
-// CHECK: [[ADD_IM:%.+]] = fadd double 0.000000e+00, [[EXPR_IM]]
-// CHECK: [[DESIRED:%.+]] = fptosi double [[ADD_RE]] to i64
-// CHECK: store i64 [[DESIRED]], i64* [[TEMP:%.+]]
-// CHECK: [[DESIRED:%.+]] = load i64, i64* [[TEMP]]
-// CHECK: [[RES:%.+]] = cmpxchg i64* [[X_ADDR]], i64 [[EXPECTED]], i64 [[DESIRED]] monotonic monotonic
-// CHECK: [[OLD_X]] = extractvalue { i64, i1 } [[RES]], 0
-// CHECK: [[SUCCESS_FAIL:%.+]] = extractvalue { i64, i1 } [[RES]], 1
-// CHECK: br i1 [[SUCCESS_FAIL]], label %[[EXIT:.+]], label %[[CONT]]
-// CHECK: [[EXIT]]
-#pragma omp atomic
-  llx += cdv;
 // CHECK: [[IDX:%.+]] = load i16, i16* @{{.+}}
 // CHECK: load i8, i8*
 // CHECK: [[VEC_ITEM_VAL:%.+]] = zext i1 %{{.+}} to i32
diff --git a/test/OpenMP/atomic_write_codegen.c b/test/OpenMP/atomic_write_codegen.c
index 050d7a510566..0c85b6e88a3f 100644
--- a/test/OpenMP/atomic_write_codegen.c
+++ b/test/OpenMP/atomic_write_codegen.c
@@ -1,6 +1,6 @@
-// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -target-cpu core2 -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c -triple x86_64-apple-darwin10 -target-cpu core2 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
 // expected-no-diagnostics
 // REQUIRES: x86-registered-target
 #ifndef HEADER
diff --git a/test/OpenMP/capturing_in_templates.cpp b/test/OpenMP/capturing_in_templates.cpp
index a4dcbee8fab9..9b2d9d304e0a 100644
--- a/test/OpenMP/capturing_in_templates.cpp
+++ b/test/OpenMP/capturing_in_templates.cpp
@@ -15,7 +15,7 @@ pair<T1, T2> make_pair(T1 &&t1, T2 &&t2) {
 
 // CHECK-LABEL: @main
 int main(int argc, char **argv) {
-// CHECK: call i32 @__tgt_target(i32 -1, i8* @{{.+}}.region_id, i32 0, i8** null, i8** null, i64* null, i32* null)
+// CHECK: call i32 @__tgt_target(i64 -1, i8* @{{.+}}.region_id, i32 0, i8** null, i8** null, i64* null, i64* null)
 #pragma omp target
  {
     for (int i = 0; i < 64; ++i) {
diff --git a/test/OpenMP/declare_reduction_messages.cpp b/test/OpenMP/declare_reduction_messages.cpp
index 198d7756d035..3e5a15ee3dd9 100644
--- a/test/OpenMP/declare_reduction_messages.cpp
+++ b/test/OpenMP/declare_reduction_messages.cpp
@@ -134,3 +134,21 @@ int main() {
   }
   return fun(15) + foo(15); // expected-note {{in instantiation of function template specialization 'foo<int>' requested here}}
 }
+
+#if __cplusplus == 201103L
+struct A {
+  A() {}
+  // expected-note@+1 {{copy constructor is implicitly deleted because 'A' has a user-declared move assignment operator}}
+  A& operator=(A&&) = default;
+};
+
+int A_TEST() {
+  A test;
+// expected-error@+1 {{call to implicitly-deleted copy constructor of 'A'}}
+#pragma omp declare reduction(+ : A : omp_out) initializer(omp_priv = A())
+// expected-error@+1 {{invalid operands to binary expression ('A' and 'A')}}
+#pragma omp parallel reduction(+ : test)
+  {}
+  return 0;
+}
+#endif
diff --git a/test/OpenMP/declare_simd_codegen.cpp b/test/OpenMP/declare_simd_codegen.cpp
index 47e951999b6a..9ae476677c42 100644
--- a/test/OpenMP/declare_simd_codegen.cpp
+++ b/test/OpenMP/declare_simd_codegen.cpp
@@ -8,11 +8,22 @@
 #pragma omp declare simd linear(d : 8)
 #pragma omp declare simd inbranch simdlen(32)
 #pragma omp declare simd notinbranch
+void add_1(float *d);
+
+#pragma omp declare simd linear(d : 8)
+#pragma omp declare simd inbranch simdlen(32)
+#pragma omp declare simd notinbranch
 void add_1(float *d) {}
 
+#pragma omp declare simd linear(d : 8)
+#pragma omp declare simd inbranch simdlen(32)
+#pragma omp declare simd notinbranch
+void add_2(float *d);
+
 #pragma omp declare simd aligned(hp, hp2)
 template <class C>
 void h(C *hp, C *hp2, C *hq, C *lin) {
+  add_2(0);
 }
 
 // Explicit specialization with <C=int>.
@@ -110,6 +121,7 @@ double foo(double x) { return 0; }
 // CHECK-DAG: define {{.+}}@_Z3bax2VVPdi(
 // CHECK-DAG: define {{.+}}@_Z3fooPffi(
 // CHECK-DAG: define {{.+}}@_Z3food(
+// CHECK-DAG: declare {{.+}}@_Z5add_2Pf(
 
 // CHECK-DAG: "_ZGVbM4l8__Z5add_1Pf"
 // CHECK-DAG: "_ZGVbN4l8__Z5add_1Pf"
@@ -277,6 +289,23 @@ double foo(double x) { return 0; }
 // CHECK-DAG: "_ZGVeM16ua16vl1__Z3fooPffi"
 // CHECK-DAG: "_ZGVeN16ua16vl1__Z3fooPffi"
 
+// CHECK-DAG: "_ZGVbM4l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVbN4l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVcM8l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVcN8l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVdM8l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVdN8l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVeM16l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVeN16l8__Z5add_2Pf"
+// CHECK-DAG: "_ZGVbM32v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVcM32v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVdM32v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVeM32v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVbN2v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVcN4v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVdN4v__Z5add_2Pf"
+// CHECK-DAG: "_ZGVeN8v__Z5add_2Pf"
+
 // CHECK-DAG: "_ZGVbN2v__Z3food"
 // CHECK-DAG: "_ZGVcN4v__Z3food"
 // CHECK-DAG: "_ZGVdN4v__Z3food"
diff --git a/test/OpenMP/declare_simd_messages.cpp b/test/OpenMP/declare_simd_messages.cpp
index 15971eb14de5..af46283f9a57 100644
--- a/test/OpenMP/declare_simd_messages.cpp
+++ b/test/OpenMP/declare_simd_messages.cpp
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -triple=x86_64-pc-win32 -verify -fopenmp -x c++ -std=c++11 -fms-extensions %s
+// RUN: %clang_cc1 -triple=x86_64-pc-win32 -verify -fopenmp -x c++ -std=c++11 -fms-extensions -Wno-pragma-pack %s
 
 // expected-error@+1 {{expected an OpenMP directive}}
 #pragma omp declare
diff --git a/test/OpenMP/declare_target_ast_print.cpp b/test/OpenMP/declare_target_ast_print.cpp
index 591844b79498..6c59563d1d1e 100644
--- a/test/OpenMP/declare_target_ast_print.cpp
+++ b/test/OpenMP/declare_target_ast_print.cpp
@@ -108,9 +108,7 @@ void f2() {
 // CHECK: #pragma omp end declare target
 
 int c1, c2, c3;
-void f3() {
-}
-#pragma omp declare target link(c1) link(c2), link(c3, f3)
+#pragma omp declare target link(c1) link(c2), link(c3)
 // CHECK: #pragma omp declare target link
 // CHECK: int c1;
 // CHECK: #pragma omp end declare target
@@ -120,9 +118,33 @@ void f3() {
 // CHECK: #pragma omp declare target link
 // CHECK: int c3;
 // CHECK: #pragma omp end declare target
-// CHECK: #pragma omp declare target link
-// CHECK: void f3()
+
+struct SSSt {
+#pragma omp declare target
+  static int a;
+  int b;
+#pragma omp end declare target
+};
+
+// CHECK: struct SSSt {
+// CHECK: #pragma omp declare target
+// CHECK: static int a;
+// CHECK: #pragma omp end declare target
+// CHECK: int b;
+
+template <class T>
+struct SSSTt {
+#pragma omp declare target
+  static T a;
+  int b;
+#pragma omp end declare target
+};
+
+// CHECK: template <class T> struct SSSTt {
+// CHECK: #pragma omp declare target
+// CHECK: static T a;
 // CHECK: #pragma omp end declare target
+// CHECK: int b;
 
 int main (int argc, char **argv) {
   foo();
diff --git a/test/OpenMP/declare_target_messages.cpp b/test/OpenMP/declare_target_messages.cpp
index b858d53c1ecf..c1c03951a3c1 100644
--- a/test/OpenMP/declare_target_messages.cpp
+++ b/test/OpenMP/declare_target_messages.cpp
@@ -13,8 +13,16 @@ void f();
 
 #pragma omp declare target map(a) // expected-error {{unexpected 'map' clause, only 'to' or 'link' clauses expected}}
 
+#pragma omp declare target to(foo1) // expected-error {{use of undeclared identifier 'foo1'}}
+
+#pragma omp declare target link(foo2) // expected-error {{use of undeclared identifier 'foo2'}}
+
 void c(); // expected-warning {{declaration is not declared in any declare target region}}
 
+void func() {} // expected-note {{'func' defined here}}
+
+#pragma omp declare target link(func) // expected-error {{function name is not allowed in 'link' clause}}
+
 extern int b;
 
 struct NonT {
@@ -28,16 +36,16 @@ typedef int sint;
 extern int b;
 int g;
 
-struct T { // expected-note {{mappable type cannot be polymorphic}}
+struct T {
   int a;
   virtual int method();
 };
 
-class VC { // expected-note {{mappable type cannot be polymorphic}}
+class VC {
   T member;
   NonT member1;
   public:
-    virtual int method() { T a; return 0; } // expected-error {{type 'T' is not mappable to target}}
+    virtual int method() { T a; return 0; }
 };
 
 struct C {
@@ -56,7 +64,7 @@ void foo() {
   b = 0; // expected-note {{used here}}
   t = 1; // expected-error {{threadprivate variables cannot be used in target constructs}}
   C object;
-  VC object1; // expected-error {{type 'VC' is not mappable to target}}
+  VC object1;
   g = object.method();
   g += object.method1();
   g += object1.method();
@@ -73,9 +81,9 @@ int C::method() {
 }
 
 struct S {
-#pragma omp declare target // expected-error {{directive must be at file or namespace scope}}
+#pragma omp declare target
   int v;
-#pragma omp end declare target // expected-error {{unexpected OpenMP directive '#pragma omp end declare target'}}
+#pragma omp end declare target
 };
 
 int main (int argc, char **argv) {
diff --git a/test/OpenMP/distribute_codegen.cpp b/test/OpenMP/distribute_codegen.cpp
index 37f00f09178b..7ea17b116c88 100644
--- a/test/OpenMP/distribute_codegen.cpp
+++ b/test/OpenMP/distribute_codegen.cpp
@@ -23,6 +23,7 @@
 // CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* }
 // CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
 // CHECK-DAG: [[DEF_LOC_0:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+// CHECK-DAG: [[DEF_LOC_DISTRIBUTE_0:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2050, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
 
 // CHECK-LABEL: define {{.*void}} @{{.*}}without_schedule_clause{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}})
 void without_schedule_clause(float *a, float *b, float *c, float *d) {
@@ -34,7 +35,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
   }
 }
 
-// CHECK: define {{.*}}void @.omp_outlined.(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]])
+// CHECK: define {{.*}}void @{{.+}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]])
 // CHECK:  [[TID_ADDR:%.+]] = alloca i32*
 // CHECK:  [[IV:%.+iv]] = alloca i32
 // CHECK:  [[LB:%.+lb]] = alloca i32
@@ -48,7 +49,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
 // CHECK-DAG:  store i32 0, i32* [[LAST]]
 // CHECK-DAG:  [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]]
 // CHECK-DAG:  [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]]
-// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1)
+// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1)
 // CHECK-DAG:  [[UBV0:%.+]] = load i32, i32* [[UB]]
 // CHECK-DAG:  [[USWITCH:%.+]] = icmp sgt i32 [[UBV0]], 4571423
 // CHECK:  br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]]
@@ -82,7 +83,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
 // CHECK:  [[BBINNEND]]:
 // CHECK:  br label %[[LPEXIT:.+]]
 // CHECK:  [[LPEXIT]]:
-// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_0]], i32 [[GBL_TIDV]])
+// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]])
 // CHECK:  ret void
 
 
@@ -110,7 +111,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) {
 // CHECK-DAG:  store i32 0, i32* [[LAST]]
 // CHECK-DAG:  [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]]
 // CHECK-DAG:  [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]]
-// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1)
+// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1)
 // CHECK-DAG:  [[UBV0:%.+]] = load i32, i32* [[UB]]
 // CHECK-DAG:  [[USWITCH:%.+]] = icmp sgt i32 [[UBV0]], 4571423
 // CHECK:  br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]]
@@ -144,7 +145,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) {
 // CHECK:  [[BBINNEND]]:
 // CHECK:  br label %[[LPEXIT:.+]]
 // CHECK:  [[LPEXIT]]:
-// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_0]], i32 [[GBL_TIDV]])
+// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]])
 // CHECK:  ret void
 
 
@@ -172,7 +173,7 @@ void static_chunked(float *a, float *b, float *c, float *d) {
 // CHECK-DAG:  store i32 0, i32* [[LAST]]
 // CHECK-DAG:  [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]]
 // CHECK-DAG:  [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]]
-// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_0]], i32 [[GBL_TIDV]], i32 91, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 5)
+// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 91, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 5)
 // CHECK-DAG:  [[UBV0:%.+]] = load i32, i32* [[UB]]
 // CHECK-DAG:  [[USWITCH:%.+]] = icmp ugt i32 [[UBV0]], 16908288
 // CHECK:  br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]]
@@ -206,7 +207,7 @@ void static_chunked(float *a, float *b, float *c, float *d) {
 // CHECK:  [[BBINNEND]]:
 // CHECK:  br label %[[LPEXIT:.+]]
 // CHECK:  [[LPEXIT]]:
-// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_0]], i32 [[GBL_TIDV]])
+// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]])
 // CHECK:  ret void
 
 // CHECK-LABEL: test_precond
diff --git a/test/OpenMP/distribute_firstprivate_codegen.cpp b/test/OpenMP/distribute_firstprivate_codegen.cpp
index 718fc875254c..98c99d805ed5 100644
--- a/test/OpenMP/distribute_firstprivate_codegen.cpp
+++ b/test/OpenMP/distribute_firstprivate_codegen.cpp
@@ -63,16 +63,13 @@ int main() {
     #pragma omp teams
 #pragma omp distribute firstprivate(g, g1, svar, sfvar)
     for (int i = 0; i < 2; ++i) {
-      // LAMBDA-64: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i{{[0-9]+}} [[G_IN:%.+]], i{{[0-9]+}} [[G1_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]], i{{[0-9]+}} [[SFVAR_IN:%.+]])
-      // LAMBDA-32: define internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.*}} [[G_IN:%.+]], i{{[0-9]+}} [[G1_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]], i{{[0-9]+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: define internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.*}} [[G_IN:%.+]], double*{{.*}} [[G1_IN:%.+]], i32*{{.*}} [[SVAR_IN:%.+]], float*{{.*}} [[SFVAR_IN:%.+]])
       // Private alloca's for conversion
-      // LAMBDA-64: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
-      // LAMBDA-32: [[G_ADDR:%.+]] = alloca double*,
-      // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
-      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
-      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca float*,
       // LAMBDA: [[G1_REF:%.+]] = alloca double*,
-      // LAMBDA: [[TMP:%.+]] = alloca double*,
 
       // Actual private variables to be used in the body (tmp is used for the reference type)
       // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
@@ -82,31 +79,25 @@ int main() {
       // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
 
       // Store input parameter addresses into private alloca's for conversion
-      // LAMBDA-64: store i{{[0-9]+}} [[G_IN]], i{{[0-9]+}}* [[G_ADDR]],
-      // LAMBDA-32: store double* [[G_IN]], double** [[G_ADDR]],
-      // LAMBDA: store i{{[0-9]+}} [[G1_IN]], i{{[0-9]+}}* [[G1_ADDR]],
-      // LAMBDA: store i{{[0-9]+}} [[SVAR_IN]], i{{[0-9]+}}* [[SVAR_ADDR]],
-      // LAMBDA: store i{{[0-9]+}} [[SFVAR_IN]], i{{[0-9]+}}* [[SFVAR_ADDR]],
-
-      // LAMBDA-64-DAG: [[G_CONV:%.+]] = bitcast i{{[0-9]+}}* [[G_ADDR]] to double*
-      // LAMBDA-32-DAG: [[G_ADDR_VAL:%.+]] = load double*, double** [[G_ADDR]],
-      // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast i{{[0-9]+}}* [[G1_ADDR]] to double*
-      // LAMBDA-DAG: store double* [[G1_CONV]], double** [[G1_REF]],
-      // LAMBDA-64-DAG: [[SVAR_CONV:%.+]] = bitcast i{{[0-9]+}}* [[SVAR_ADDR]] to i{{[0-9]+}}*
-      // LAMBDA-DAG: [[SFVAR_CONV:%.+]] = bitcast i{{[0-9]+}}* [[SFVAR_ADDR]] to float*
-      // LAMBDA-DAG: [[G1_REF_VAL:%.+]] = load double*, double** [[G1_REF]],
-      // LAMBDA-DAG: store double* [[G1_REF_VAL]], double** [[TMP]],
-      // LAMBDA-64-DAG: [[G_CONV_VAL:%.+]] = load{{.*}} double, double* [[G_CONV]],
-      // LAMBDA-32-DAG: [[G_CONV_VAL:%.+]] = load{{.*}} double, double* [[G_ADDR_VAL]],
+      // LAMBDA: store double* [[G_IN]], double** [[G_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_ADDR]],
+
+      // LAMBDA-DAG: [[G_ADDR_VAL:%.+]] = load double*, double** [[G_ADDR]],
+      // LAMBDA-DAG: [[SVAR_ADDR_VAL:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+      // LAMBDA-DAG: [[SFVAR_ADDR_VAL:%.+]] = load float*, float** [[SFVAR_ADDR]],
+      // LAMBDA-DAG: [[G1_ADDR_VAL:%.+]] = load double*, double** [[G1_ADDR]],
+      // LAMBDA-DAG: store double* [[G1_ADDR_VAL]], double** [[G1_REF]],
+      // LAMBDA-DAG: [[G_CONV_VAL:%.+]] = load{{.*}} double, double* [[G_ADDR_VAL]],
       // LAMBDA-DAG: store double [[G_CONV_VAL]], double* [[G_PRIVATE]],
-      // LAMBDA-DAG: [[TMP_VAL:%.+]] = load double*, double** [[TMP]],
+      // LAMBDA-DAG: [[TMP_VAL:%.+]] = load double*, double** [[G1_REF]],
       // LAMBDA-DAG: [[TMP_VAL_VAL:%.+]] = load{{.*}} double, double* [[TMP_VAL]],
       // LAMBDA-DAG: store double [[TMP_VAL_VAL]], double* [[G1_PRIVATE]],
       // LAMBDA-DAG: store double* [[G1_PRIVATE]], double** [[TMP_PRIVATE]],
-      // LAMBDA-64-DAG: [[SVAR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_CONV]],
-      // LAMBDA-32-DAG: [[SVAR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_ADDR]],
+      // LAMBDA-DAG: [[SVAR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_ADDR_VAL]],
       // LAMBDA-DAG: store i{{[0-9]+}} [[SVAR_CONV_VAL]], i{{[0-9]+}}* [[SVAR_PRIVATE]],
-      // LAMBDA-DAG: [[SFVAR_CONV_VAL:%.+]] = load float, float* [[SFVAR_CONV]],
+      // LAMBDA-DAG: [[SFVAR_CONV_VAL:%.+]] = load float, float* [[SFVAR_ADDR_VAL]],
       // LAMBDA-DAG: store float [[SFVAR_CONV_VAL]], float* [[SFVAR_PRIVATE]],
       // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
       g += 1;
@@ -204,16 +195,18 @@ int main() {
 // CHECK: ret
 
 // CHECK: define{{.+}} [[OFFLOAD_FUN]](
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}, [2 x i{{[0-9]+}}]*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, i{{[0-9]+}})* [[OMP_OUTLINED:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, i{{[0-9]+}}*)* [[OMP_OUTLINED:@.+]] to void
 // CHECK: ret
 //
-// CHECK: define internal void [[OMP_OUTLINED]](i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.*}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.*}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.*}} [[VAR_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]])
+// CHECK: define internal void [[OMP_OUTLINED]](i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.*}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.*}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.*}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]])
 
-// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
-// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
 // CHECK: [[TMP:%.+]] = alloca [[S_FLOAT_TY]]*,
 
 // discard omp loop variables
@@ -222,6 +215,7 @@ int main() {
 // CHECK: {{.*}} = alloca i{{[0-9]+}},
 // CHECK: {{.*}} = alloca i{{[0-9]+}},
 // CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
 
 // CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
@@ -230,17 +224,16 @@ int main() {
 // CHECK-DAG: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
 // CHECK: [[SVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 
-// CHECK: store i{{[0-9]+}} [[T_VAR_IN]], i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
 // CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
 // CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
 // CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
-// CHECK: store i{{[0-9]+}} [[SVAR_IN]], i{{[0-9]+}}* [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
 
 // init t_var
-// CHECK-64-DAG: [[T_VAR_ADDR_CONV:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR_ADDR]] to i{{[0-9]+}}*
-// CHECK-64-DAG: [[T_VAR_ADDR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_ADDR_CONV]],
-// CHECK-32-DAG: [[T_VAR_ADDR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
-// CHECK-DAG: store i{{[0-9]+}} [[T_VAR_ADDR_CONV_VAL]], i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK-DAG: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_REF]],
+// CHECK-DAG: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_PRIV]],
 
 // init vec
 // CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
@@ -278,10 +271,9 @@ int main() {
 // CHECK-DAG: store [[S_FLOAT_TY]]* [[VAR_PRIV]], [[S_FLOAT_TY]]** [[TMP_PRIV]],
 
 // init svar
-// CHECK-64-DAG: [[SVAR_ADDR_CONV:%.+]] = bitcast{{.+}} [[SVAR_ADDR]] to{{.+}}
-// CHECK-64-DAG: [[SVAR_CONV_VAL:%.+]] = load{{.+}},{{.+}} [[SVAR_ADDR_CONV]],
-// CHECK-32-DAG: [[SVAR_CONV_VAL:%.+]] = load{{.+}},{{.+}} [[SVAR_ADDR]],
-// CHECK-DAG: store{{.+}} [[SVAR_CONV_VAL]],{{.+}} [[SVAR_PRIV]],
+// CHECK-DAG: [[SVAR_REF:%.+]] = load{{.+}},{{.+}} [[SVAR_ADDR]],
+// CHECK-DAG: [[SVAR_VAL:%.+]] = load{{.+}},{{.+}} [[SVAR_REF]],
+// CHECK-DAG: store{{.+}} [[SVAR_VAL]],{{.+}} [[SVAR_PRIV]],
 
 // CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* %.omp{{.+}},
 // CHECK-DAG: store i{{[0-9]+}} 1, i{{[0-9]+}}* %.omp{{.+}},
@@ -301,12 +293,14 @@ int main() {
 // CHECK: ret
 
 // CHECK: define{{.+}} [[OFFLOAD_FUN_1]](
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}, [2 x i{{[0-9]+}}]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[OMP_OUTLINED_1:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[OMP_OUTLINED_1:@.+]] to void
 // CHECK: ret
 //
-// CHECK: define internal void [[OMP_OUTLINED_1]](i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.*}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]*{{.*}} [[S_ARR_IN:%.+]], [[S_INT_TY]]*{{.*}} [[VAR_IN:%.+]])
+// CHECK: define internal void [[OMP_OUTLINED_1]](i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.*}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]*{{.*}} [[S_ARR_IN:%.+]], [[S_INT_TY]]*{{.*}} [[VAR_IN:%.+]])
 
-// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
@@ -318,6 +312,7 @@ int main() {
 // CHECK: {{.*}} = alloca i{{[0-9]+}},
 // CHECK: {{.*}} = alloca i{{[0-9]+}},
 // CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
 
 // CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
@@ -325,16 +320,15 @@ int main() {
 // CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
 // CHECK-DAG: [[TMP_PRIV:%.+]] = alloca [[S_INT_TY]]*,
 
-// CHECK: store i{{[0-9]+}} [[T_VAR_IN]], i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
 // CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
 // CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
 // CHECK: store [[S_INT_TY]]* [[VAR_IN]], [[S_INT_TY]]** [[VAR_ADDR]],
 
 // init t_var
-// CHECK-64-DAG: [[T_VAR_ADDR_CONV:%.+]] = bitcast i{{[0-9]+}}* [[T_VAR_ADDR]] to i{{[0-9]+}}*
-// CHECK-64-DAG: [[T_VAR_ADDR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_ADDR_CONV]],
-// CHECK-32-DAG: [[T_VAR_ADDR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
-// CHECK-DAG: store i{{[0-9]+}} [[T_VAR_ADDR_CONV_VAL]], i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK-DAG: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_REF]],
+// CHECK-DAG: store i{{[0-9]+}} [[T_VAR]], i{{[0-9]+}}* [[T_VAR_PRIV]],
 
 // init vec
 // CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
diff --git a/test/OpenMP/distribute_firstprivate_messages.cpp b/test/OpenMP/distribute_firstprivate_messages.cpp
index 5d371abbfa8e..a59c9528090f 100644
--- a/test/OpenMP/distribute_firstprivate_messages.cpp
+++ b/test/OpenMP/distribute_firstprivate_messages.cpp
@@ -135,11 +135,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo(); // expected-error {{loop iteration variable in the associated loop of 'omp distribute' directive may not be firstprivate, predetermined as private}}
   #pragma omp target
   #pragma omp teams private(argc) // expected-note {{defined as private}}
-  #pragma omp distribute firstprivate(argc) // expected-error {{private variable in '#pragma omp teams' cannot be firstprivate in '#pragma omp distribute'}}
+  #pragma omp distribute firstprivate(argc) // expected-error {{firstprivate variable must be shared}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target
   #pragma omp teams reduction(+:argc) // expected-note {{defined as reduction}}
-  #pragma omp distribute firstprivate(argc) // expected-error {{reduction variable in '#pragma omp teams' cannot be firstprivate in '#pragma omp distribute'}}
+  #pragma omp distribute firstprivate(argc) // expected-error {{firstprivate variable must be shared}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target
   #pragma omp teams
@@ -147,11 +147,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target
   #pragma omp teams
-  #pragma omp distribute lastprivate(argc), firstprivate(argc) // expected-error {{lastprivate variable cannot be firstprivate in '#pragma omp distribute'}} expected-note{{defined as lastprivate}}
+  #pragma omp distribute lastprivate(argc), firstprivate(argc) // expected-error {{lastprivate variable cannot be firstprivate}} expected-note{{defined as lastprivate}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target
   #pragma omp teams
-#pragma omp distribute firstprivate(argc), lastprivate(argc)  // expected-error {{lastprivate variable cannot be firstprivate in '#pragma omp distribute'}} expected-note{{defined as firstprivate}}
+#pragma omp distribute firstprivate(argc), lastprivate(argc)  // expected-error {{firstprivate variable cannot be lastprivate}} expected-note{{defined as firstprivate}}
   for (i = 0; i < argc; ++i) foo();
   return 0;
 }
diff --git a/test/OpenMP/distribute_lastprivate_codegen.cpp b/test/OpenMP/distribute_lastprivate_codegen.cpp
index abcbfd8d69c4..bd9dae956f1c 100644
--- a/test/OpenMP/distribute_lastprivate_codegen.cpp
+++ b/test/OpenMP/distribute_lastprivate_codegen.cpp
@@ -74,6 +74,7 @@ int main() {
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
       // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
       // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
@@ -202,6 +203,7 @@ int main() {
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
@@ -303,6 +305,7 @@ int main() {
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
diff --git a/test/OpenMP/distribute_parallel_for_ast_print.cpp b/test/OpenMP/distribute_parallel_for_ast_print.cpp
index 54a3649e578b..351e15803c09 100644
--- a/test/OpenMP/distribute_parallel_for_ast_print.cpp
+++ b/test/OpenMP/distribute_parallel_for_ast_print.cpp
@@ -82,7 +82,7 @@ T tmain(T argc) {
 // CHECK-NEXT: a = 2;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) schedule(static, N) if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h) dist_schedule(static,N)
+#pragma omp distribute parallel for private(argc, b), firstprivate(c, d), lastprivate(f) collapse(N) schedule(static, N) if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h) dist_schedule(static,N)
   for (int i = 0; i < 2; ++i)
     for (int j = 0; j < 2; ++j)
       for (int j = 0; j < 2; ++j)
@@ -93,8 +93,8 @@ T tmain(T argc) {
       for (int j = 0; j < 2; ++j)
         for (int j = 0; j < 2; ++j)
           for (int j = 0; j < 2; ++j)
-	    a++;
-  // CHECK: #pragma omp distribute parallel for private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) schedule(static, N) if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h) dist_schedule(static, N)
+            a++;
+  // CHECK: #pragma omp distribute parallel for private(argc,b) firstprivate(c,d) lastprivate(f) collapse(N) schedule(static, N) if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h) dist_schedule(static, N)
   // CHECK-NEXT: for (int i = 0; i < 2; ++i)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
diff --git a/test/OpenMP/distribute_parallel_for_codegen.cpp b/test/OpenMP/distribute_parallel_for_codegen.cpp
index 62c38f188851..4fbca7d2f29a 100644
--- a/test/OpenMP/distribute_parallel_for_codegen.cpp
+++ b/test/OpenMP/distribute_parallel_for_codegen.cpp
@@ -28,6 +28,7 @@ T tmain() {
   #pragma omp teams
   #pragma omp distribute parallel for
   for (int i = 0; i < n; ++i) {
+    #pragma omp cancel for
     a[i] = b[i] + c[i];
   }
 
diff --git a/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp b/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp
index d12c0aa11cf2..de348272950c 100644
--- a/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp
+++ b/test/OpenMP/distribute_parallel_for_firstprivate_codegen.cpp
@@ -63,17 +63,14 @@ int main() {
     #pragma omp teams
     #pragma omp distribute parallel for firstprivate(g, g1, svar, sfvar)
     for (int i = 0; i < 2; ++i) {
-      // LAMBDA-64: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i{{[0-9]+}} [[G_IN:%.+]], i{{[0-9]+}} [[G1_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]], i{{[0-9]+}} [[SFVAR_IN:%.+]])
-      // LAMBDA-32: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double* {{.+}} [[G_IN:%.+]], i{{[0-9]+}} [[G1_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]], i{{[0-9]+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double* {{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
 
       // addr alloca's
-      // LAMBDA-64: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
-      // LAMBDA-32: [[G_ADDR:%.+]] = alloca double*,
-      // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
-      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
-      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca float*,
       // LAMBDA: [[G1_REF:%.+]] = alloca double*,
-      // LAMBDA: [[TMP:%.+]] = alloca double*,
 
       // private alloca's
       // LAMBDA: [[G_PRIV:%.+]] = alloca double,
@@ -88,31 +85,27 @@ int main() {
       // LAMBDA-DAG: store {{.+}} [[SVAR_IN]], {{.+}} [[SVAR_ADDR]],
       // LAMBDA-DAG: store {{.+}} [[SFVAR_IN]], {{.+}} [[SFVAR_ADDR]],
 
+      // LAMBDA-DAG: [[G_CONV:%.+]] = load {{.+}}*, {{.+}}** [[G_ADDR]]
+      // LAMBDA-DAG: [[G1_CONV:%.+]] = load {{.+}}*, {{.+}}** [[G1_ADDR]]
+      // LAMBDA-DAG: [[SVAR_CONV:%.+]] = load {{.+}}*, {{.+}}** [[SVAR_ADDR]]
+      // LAMBDA-DAG: [[SFVAR_CONV:%.+]] = load {{.+}}*, {{.+}}** [[SFVAR_ADDR]]
+
       // init private alloca's with addr alloca's
       // g
-      // LAMBDA-64-DAG: [[G_CONV:%.+]] = bitcast {{.+}}* [[G_ADDR]] to
-      // LAMBDA-32-DAG: [[G_CONV:%.+]] = load {{.+}}*, {{.+}}** [[G_ADDR]]
       // LAMBDA-DAG: [[G_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[G_CONV]],
       // LAMBDA-DAG: store {{.+}} [[G_ADDR_VAL]], {{.+}}* [[G_PRIV]],
 
       // g1
-      // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}}* [[G1_ADDR]] to
-      // LAMBDA-DAG: store {{.+}}* [[G1_CONV]], {{.+}}** [[G1_REF]],
-      // LAMBDA-DAG: [[G1_REF_VAL:%.+]] = load {{.+}}*, {{.+}}** [[G1_REF]],
-      // LAMBDA-DAG: store {{.+}}* [[G1_REF_VAL]], {{.+}}** [[TMP]],
-      // LAMBDA-DAG: [[TMP_REF:%.+]] = load {{.+}}*, {{.+}}** [[TMP]],
+      // LAMBDA-DAG: [[TMP_REF:%.+]] = load {{.+}}*, {{.+}}** [[G1_REF]],
       // LAMBDA-DAG: [[TMP_VAL:%.+]] = load {{.+}}, {{.+}}* [[TMP_REF]],
       // LAMBDA-DAG: store {{.+}} [[TMP_VAL]], {{.+}}* [[G1_PRIV]]
       // LAMBDA-DAG: store {{.+}}* [[G1_PRIV]], {{.+}}** [[TMP_PRIV]],
 
       // svar
-      // LAMBDA-64-DAG: [[SVAR_CONV:%.+]] = bitcast {{.+}}* [[SVAR_ADDR]] to
-      // LAMBDA-64-DAG: [[SVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_CONV]],
-      // LAMBDA-32-DAG: [[SVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_ADDR]],
+      // LAMBDA-DAG: [[SVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_CONV]],
       // LAMBDA-DAG: store {{.+}} [[SVAR_VAL]], {{.+}}* [[SVAR_PRIV]],
 
       // sfvar
-      // LAMBDA-DAG: [[SFVAR_CONV:%.+]] = bitcast {{.+}}* [[SFVAR_ADDR]] to
       // LAMBDA-DAG: [[SFVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SFVAR_CONV]],
       // LAMBDA-DAG: store {{.+}} [[SFVAR_VAL]], {{.+}}* [[SFVAR_PRIV]],
 
@@ -271,17 +264,19 @@ int main() {
 // CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
 
 // CHECK: define{{.+}} [[OFFLOAD_FUN_0]](i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]* {{.+}} [[VAR_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]])
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}, [2 x i{{[0-9]+}}]*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, i{{[0-9]+}})* [[OMP_OUTLINED_0:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, i{{[0-9]+}}*)* [[OMP_OUTLINED_0:@.+]] to void
 // CHECK: ret
 
-// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]* {{.+}} [[VAR_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]])
+// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]* {{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]])
 
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
 // addr alloca's
-// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
-// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
 // CHECK: [[TMP:%.+]] = alloca [[S_FLOAT_TY]]*,
 
 // skip loop alloca's
@@ -310,9 +305,8 @@ int main() {
 
 // init private alloca's with addr alloca's
 // t-var
-// CHECK-64-DAG: [[T_VAR_CONV:%.+]] = bitcast {{.+}} [[T_VAR_ADDR]] to
-// CHECK-64-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_CONV]],
-// CHECK-32-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_REF:%.+]] = load {{.+}}, {{.+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_REF]],
 // CHECK-DAG: store {{.+}} [[T_VAR_ADDR_VAL]], {{.+}} [[T_VAR_PRIV]],
 
 // vec
@@ -340,10 +334,9 @@ int main() {
 // CHECK-DAG: store {{.+}}* [[VAR_PRIV]], {{.+}}** [[TMP_PRIV]],
 
 // svar
-// CHECK-64-DAG: [[SVAR_CONV:%.+]] = bitcast {{.+}}* [[SVAR_ADDR]] to
-// CHECK-64-DAG: [[SVAR_CONV_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_CONV]],
-// CHECK-32-DAG: [[SVAR_CONV_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_ADDR]],
-// CHECK-DAG: store {{.+}} [[SVAR_CONV_VAL]], {{.+}}* [[SVAR_PRIV]],
+// CHECK-DAG: [[SVAR_REF:%.+]] = load {{.+}}*, {{.+}}** [[SVAR_ADDR]],
+// CHECK-DAG: [[SVAR:%.+]] = load {{.+}}, {{.+}}* [[SVAR_REF]],
+// CHECK-DAG: store {{.+}} [[SVAR]], {{.+}}* [[SVAR_PRIV]],
 
 // CHECK: call void @__kmpc_for_static_init_4(
 // pass private alloca's to fork
@@ -456,13 +449,15 @@ int main() {
 // CHECK: call {{.*}} [[S_INT_TY_DEF_DESTR:@.+]]([[S_INT_TY]]* [[TEST]])
 
 // CHECK: define{{.+}} [[OFFLOAD_FUN_0]](i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]* {{.+}} [[VAR_IN:%.+]])
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}, [2 x i{{[0-9]+}}]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[OMP_OUTLINED_0:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[OMP_OUTLINED_0:@.+]] to void
 // CHECK: ret
 
-// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]* {{.+}} [[VAR_IN:%.+]])
+// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]* {{.+}} [[VAR_IN:%.+]])
 
 // addr alloca's
-// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
 // CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
 // CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
 // CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
@@ -492,9 +487,8 @@ int main() {
 
 // init private alloca's with addr alloca's
 // t-var
-// CHECK-64-DAG: [[T_VAR_CONV:%.+]] = bitcast {{.+}} [[T_VAR_ADDR]] to
-// CHECK-64-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_CONV]],
-// CHECK-32-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_ADDR_REF:%.+]] = load {{.+}}*, {{.+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_ADDR_REF]],
 // CHECK-DAG: store {{.+}} [[T_VAR_ADDR_VAL]], {{.+}} [[T_VAR_PRIV]],
 
 // vec
diff --git a/test/OpenMP/distribute_parallel_for_firstprivate_messages.cpp b/test/OpenMP/distribute_parallel_for_firstprivate_messages.cpp
index 3e288c37d42c..67075ca12939 100644
--- a/test/OpenMP/distribute_parallel_for_firstprivate_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_firstprivate_messages.cpp
@@ -42,7 +42,7 @@ public:
 };
 class S5 {
   int a;
-  S5(const S5 &s5) : a(s5.a) {} // expected-note 4 {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {} // expected-note 2 {{implicitly declared private here}}
 
 public:
   S5() : a(0) {}
@@ -50,7 +50,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -150,9 +150,10 @@ int foomain(int argc, char **argv) {
 #pragma omp distribute parallel for firstprivate(i)
   for (int k = 0; k < argc; ++k)
     ++k;
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for lastprivate(g) firstprivate(g) // expected-error {{calling a private constructor of class 'S5'}}
+#pragma omp distribute parallel for lastprivate(g) firstprivate(g)
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp parallel private(i)
@@ -314,14 +315,16 @@ int main(int argc, char **argv) {
 #pragma omp distribute parallel for firstprivate(j)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for lastprivate(g) firstprivate(g) // expected-error {{calling a private constructor of class 'S5'}}
+#pragma omp distribute parallel for lastprivate(g) firstprivate(g)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for lastprivate(n) firstprivate(n) // OK
+#pragma omp distribute parallel for lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp parallel
diff --git a/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp b/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp
index 2e8da79c03bd..fccb0cfdb3d8 100644
--- a/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp
+++ b/test/OpenMP/distribute_parallel_for_lastprivate_codegen.cpp
@@ -74,6 +74,7 @@ int main() {
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
       // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
       // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
@@ -141,6 +142,7 @@ int main() {
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
       // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
 
       // private alloca's
       // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
@@ -162,7 +164,7 @@ int main() {
       // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
 
       // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
-      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
 
       // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
 
@@ -273,6 +275,7 @@ int main() {
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
@@ -371,6 +374,7 @@ int main() {
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
@@ -486,6 +490,7 @@ int main() {
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
@@ -566,6 +571,7 @@ int main() {
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
 // CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
diff --git a/test/OpenMP/distribute_parallel_for_lastprivate_messages.cpp b/test/OpenMP/distribute_parallel_for_lastprivate_messages.cpp
index 745007fc48a8..b38d978cca03 100644
--- a/test/OpenMP/distribute_parallel_for_lastprivate_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -313,14 +313,16 @@ int main(int argc, char **argv) {
 #pragma omp distribute parallel for lastprivate(j)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{firstprivate variable cannot be lastprivate}} expected-note@+3 {{defined as firstprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp distribute parallel for firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for lastprivate(n) firstprivate(n) // OK
+#pragma omp distribute parallel for lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i)
     foo();
   static int si;
diff --git a/test/OpenMP/distribute_parallel_for_messages.cpp b/test/OpenMP/distribute_parallel_for_messages.cpp
index 35a61df39e40..b1bb8d088e90 100644
--- a/test/OpenMP/distribute_parallel_for_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_messages.cpp
@@ -38,7 +38,7 @@ int main(int argc, char **argv) {
     foo();
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for
+#pragma omp distribute parallel for linear(argc) // expected-error {{unexpected OpenMP clause 'linear' in directive '#pragma omp distribute parallel for'}}
   for (int i = 0; i < argc; ++i)
     foo();
 #pragma omp target
@@ -116,3 +116,14 @@ void test_ordered() {
     ;
 }
 
+void test_cancel() {
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for
+  for (int i = 0; i < 16; ++i)
+    for (int j = 0; j < 16; ++j) {
+#pragma omp cancel for
+    ;
+    }
+}
+
diff --git a/test/OpenMP/distribute_parallel_for_reduction_messages.cpp b/test/OpenMP/distribute_parallel_for_reduction_messages.cpp
index 95654a9e5018..0bf1bc0472eb 100644
--- a/test/OpenMP/distribute_parallel_for_reduction_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_reduction_messages.cpp
@@ -27,9 +27,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/distribute_parallel_for_simd_ast_print.cpp b/test/OpenMP/distribute_parallel_for_simd_ast_print.cpp
index 6ed3f2192258..21010638be15 100644
--- a/test/OpenMP/distribute_parallel_for_simd_ast_print.cpp
+++ b/test/OpenMP/distribute_parallel_for_simd_ast_print.cpp
@@ -28,17 +28,18 @@ public:
       ++this->a.a;
   }
   S7 &operator=(S7 &s) {
+    int k;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd private(a) private(this->a)
-    for (int k = 0; k < s.a.a; ++k)
+#pragma omp distribute parallel for simd private(a) private(this->a) linear(k)
+    for (k = 0; k < s.a.a; ++k)
       ++s.a.a;
     return *this;
   }
 };
 
 // CHECK: #pragma omp distribute parallel for simd private(this->a) private(this->a) private(T::a)
-// CHECK: #pragma omp distribute parallel for simd private(this->a) private(this->a)
+// CHECK: #pragma omp distribute parallel for simd private(this->a) private(this->a) linear(k)
 // CHECK: #pragma omp distribute parallel for simd private(this->a) private(this->a) private(this->S::a)
 
 class S8 : public S7<S> {
@@ -82,7 +83,7 @@ T tmain(T argc) {
 // CHECK-NEXT: a = 2;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) schedule(static, N) if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h) dist_schedule(static,N)
+#pragma omp distribute parallel for simd private(argc, b), firstprivate(c, d), lastprivate(f) collapse(N) schedule(static, N) if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h) dist_schedule(static,N)
   for (int i = 0; i < 2; ++i)
     for (int j = 0; j < 2; ++j)
       for (int j = 0; j < 2; ++j)
@@ -93,8 +94,8 @@ T tmain(T argc) {
       for (int j = 0; j < 2; ++j)
         for (int j = 0; j < 2; ++j)
           for (int j = 0; j < 2; ++j)
-	    a++;
-  // CHECK: #pragma omp distribute parallel for simd private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) schedule(static, N) if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h) dist_schedule(static, N)
+            a++;
+  // CHECK: #pragma omp distribute parallel for simd private(argc,b) firstprivate(c,d) lastprivate(f) collapse(N) schedule(static, N) if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h) dist_schedule(static, N)
   // CHECK-NEXT: for (int i = 0; i < 2; ++i)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
@@ -135,14 +136,15 @@ int main(int argc, char **argv) {
   // CHECK-NEXT: for (int j = 0; j < 10; ++j)
   // CHECK-NEXT: a++;
 
+  int i;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd aligned(x:8) linear(h:2) safelen(8) simdlen(8)
-  for (int i = 0; i < 100; i++)
+#pragma omp distribute parallel for simd aligned(x:8) linear(i:2) safelen(8) simdlen(8)
+  for (i = 0; i < 100; i++)
     for (int j = 0; j < 200; j++)
       a += h + x[j];
-  // CHECK: #pragma omp distribute parallel for simd aligned(x: 8) linear(h: 2) safelen(8) simdlen(8)
-  // CHECK-NEXT: for (int i = 0; i < 100; i++)
+  // CHECK: #pragma omp distribute parallel for simd aligned(x: 8) linear(i: 2) safelen(8) simdlen(8)
+  // CHECK-NEXT: for (i = 0; i < 100; i++)
   // CHECK-NEXT: for (int j = 0; j < 200; j++)
   // CHECK-NEXT: a += h + x[j];
 
diff --git a/test/OpenMP/distribute_parallel_for_simd_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_codegen.cpp
new file mode 100644
index 000000000000..59e6df90bbcb
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_codegen.cpp
@@ -0,0 +1,2262 @@
+// Test host code gen
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+
+template <typename T>
+T tmain() {
+  T *a, *b, *c;
+  int n = 10000;
+  int ch = 100;
+
+  // no schedule clauses
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  // dist_schedule: static no chunk
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd dist_schedule(static)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  // dist_schedule: static chunk
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd dist_schedule(static, ch)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  // schedule: static no chunk
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd schedule(static)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  // schedule: static chunk
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd schedule(static, ch)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  // schedule: dynamic no chunk
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd schedule(dynamic)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  // schedule: dynamic chunk
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd schedule(dynamic, ch)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+  }
+
+  return T();
+}
+
+int main() {
+  double *a, *b, *c;
+  int n = 10000;
+  int ch = 100;
+
+#ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_1:@.+]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_2:@.+]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_3:@.+]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_4:@.+]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_5:@.+]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_6:@.+]](
+
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN_7:@.+]](
+
+    // no schedule clauses
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_1]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_1]](
+      // LAMBDA-DAG: [[OMP_IV:%.omp.iv]] = alloca
+      // LAMBDA-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+      // LAMBDA-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+      // LAMBDA-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+
+      // check EUB for distribute
+      // LAMBDA-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+      // LAMBDA: [[NUM_IT_1:%.+]] = load{{.+}},
+      // LAMBDA-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+      // LAMBDA: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+      // LAMBDA-DAG: [[EUB_TRUE]]:
+      // LAMBDA: [[NUM_IT_2:%.+]] = load{{.+}},
+      // LAMBDA: br label %[[EUB_END:.+]]
+      // LAMBDA-DAG: [[EUB_FALSE]]:
+      // LAMBDA: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+      // LAMBDA: br label %[[EUB_END]]
+      // LAMBDA-DAG: [[EUB_END]]:
+      // LAMBDA-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+      // LAMBDA: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+      // initialize omp.iv
+      // LAMBDA: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+      // LAMBDA: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_JUMP_BACK:.+]]
+
+      // check exit condition
+      // LAMBDA: [[OMP_JUMP_BACK]]:
+      // LAMBDA-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+      // LAMBDA: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+      // LAMBDA: br {{.+}} [[CMP_IV_UB]], label %[[DIST_BODY:.+]], label %[[DIST_END:.+]]
+
+      // check that PrevLB and PrevUB are passed to the 'for'
+      // LAMBDA: [[DIST_BODY]]:
+      // LAMBDA-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to
+      // LAMBDA-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to
+      // check that distlb and distub are properly passed to fork_call
+      // LAMBDA-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+      // LAMBDA-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+      // LAMBDA: br label %[[DIST_INC:.+]]
+
+      // increment by stride (distInc - 'parallel for' executes the whole chunk) and latch
+      // LAMBDA: [[DIST_INC]]:
+      // LAMBDA-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+      // LAMBDA: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_2]], [[OMP_ST_VAL_1]]
+      // LAMBDA: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_JUMP_BACK]]
+
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+
+      // implementation of 'parallel for'
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED_1]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+      // LAMBDA-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+      // initialize lb and ub to PrevLB and PrevUB
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+      // LAMBDA-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+      // PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+      // In this case we use EUB
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+      // LAMBDA-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+      // LAMBDA: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+      // LAMBDA: [[PF_EUB_TRUE]]:
+      // LAMBDA: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+      // LAMBDA: br label %[[PF_EUB_END:.+]]
+      // LAMBDA-DAG: [[PF_EUB_FALSE]]:
+      // LAMBDA: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA: br label %[[PF_EUB_END]]
+      // LAMBDA-DAG: [[PF_EUB_END]]:
+      // LAMBDA-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+      // LAMBDA: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+      // initialize omp.iv
+      // LAMBDA: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+      // check exit condition
+      // LAMBDA: [[OMP_PF_JUMP_BACK]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+      // LAMBDA: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+      // check that PrevLB and PrevUB are passed to the 'for'
+      // LAMBDA: [[PF_BODY]]:
+      // LAMBDA-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label {{.+}}
+
+      // check stride 1 for 'for' in 'distribute parallel for simd'
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+      // LAMBDA: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_JUMP_BACK]]
+
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+
+    // dist_schedule: static no chunk (same sa default - no dist_schedule)
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_2]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_2:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd dist_schedule(static)
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_2]](
+      // LAMBDA-DAG: [[OMP_IV:%.omp.iv]] = alloca
+      // LAMBDA-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+      // LAMBDA-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+      // LAMBDA-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+
+      // check EUB for distribute
+      // LAMBDA-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+      // LAMBDA: [[NUM_IT_1:%.+]] = load{{.+}},
+      // LAMBDA-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+      // LAMBDA: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+      // LAMBDA-DAG: [[EUB_TRUE]]:
+      // LAMBDA: [[NUM_IT_2:%.+]] = load{{.+}},
+      // LAMBDA: br label %[[EUB_END:.+]]
+      // LAMBDA-DAG: [[EUB_FALSE]]:
+      // LAMBDA: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+      // LAMBDA: br label %[[EUB_END]]
+      // LAMBDA-DAG: [[EUB_END]]:
+      // LAMBDA-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+      // LAMBDA: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+      // initialize omp.iv
+      // LAMBDA: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+      // LAMBDA: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_JUMP_BACK:.+]]
+
+      // check exit condition
+      // LAMBDA: [[OMP_JUMP_BACK]]:
+      // LAMBDA-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+      // LAMBDA: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+      // LAMBDA: br {{.+}} [[CMP_IV_UB]], label %[[DIST_BODY:.+]], label %[[DIST_END:.+]]
+
+      // check that PrevLB and PrevUB are passed to the 'for'
+      // LAMBDA: [[DIST_BODY]]:
+      // LAMBDA-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to
+      // LAMBDA-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to
+      // check that distlb and distub are properly passed to fork_call
+      // LAMBDA-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_2:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+      // LAMBDA-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_2:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+      // LAMBDA: br label %[[DIST_INC:.+]]
+
+      // increment by stride (distInc - 'parallel for' executes the whole chunk) and latch
+      // LAMBDA: [[DIST_INC]]:
+      // LAMBDA-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+      // LAMBDA: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_2]], [[OMP_ST_VAL_1]]
+      // LAMBDA: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_JUMP_BACK]]
+
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+
+      // implementation of 'parallel for'
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED_2]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+      // LAMBDA-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+      // initialize lb and ub to PrevLB and PrevUB
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+      // LAMBDA-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+      // PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+      // In this case we use EUB
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+      // LAMBDA-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+      // LAMBDA: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+      // LAMBDA: [[PF_EUB_TRUE]]:
+      // LAMBDA: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+      // LAMBDA: br label %[[PF_EUB_END:.+]]
+      // LAMBDA-DAG: [[PF_EUB_FALSE]]:
+      // LAMBDA: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA: br label %[[PF_EUB_END]]
+      // LAMBDA-DAG: [[PF_EUB_END]]:
+      // LAMBDA-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+      // LAMBDA: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+      // initialize omp.iv
+      // LAMBDA: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+      // check exit condition
+      // LAMBDA: [[OMP_PF_JUMP_BACK]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+      // LAMBDA: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+      // check that PrevLB and PrevUB are passed to the 'for'
+      // LAMBDA: [[PF_BODY]]:
+      // LAMBDA-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label {{.+}}
+
+      // check stride 1 for 'for' in 'distribute parallel for simd'
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+      // LAMBDA: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_JUMP_BACK]]
+
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+
+    // dist_schedule: static chunk
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_3]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_3:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd dist_schedule(static, ch)
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_3]](
+      // LAMBDA-DAG: [[OMP_IV:%.omp.iv]] = alloca
+      // LAMBDA-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+      // LAMBDA-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+      // LAMBDA-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+      // unlike the previous tests, in this one we have a outer and inner loop for 'distribute'
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91,
+      // LAMBDA: br label %[[DIST_OUTER_LOOP_HEADER:.+]]
+
+      // LAMBDA: [[DIST_OUTER_LOOP_HEADER]]:
+      // check EUB for distribute
+      // LAMBDA-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+      // LAMBDA: [[NUM_IT_1:%.+]] = load{{.+}},
+      // LAMBDA-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+      // LAMBDA: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+      // LAMBDA-DAG: [[EUB_TRUE]]:
+      // LAMBDA: [[NUM_IT_2:%.+]] = load{{.+}},
+      // LAMBDA: br label %[[EUB_END:.+]]
+      // LAMBDA-DAG: [[EUB_FALSE]]:
+      // LAMBDA: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+      // LAMBDA: br label %[[EUB_END]]
+      // LAMBDA-DAG: [[EUB_END]]:
+      // LAMBDA-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+      // LAMBDA: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+      // initialize omp.iv
+      // LAMBDA: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+      // LAMBDA: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+
+      // check exit condition
+      // LAMBDA-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+      // LAMBDA: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+      // LAMBDA: br {{.+}} [[CMP_IV_UB]], label %[[DIST_OUTER_LOOP_BODY:.+]], label %[[DIST_OUTER_LOOP_END:.+]]
+
+      // LAMBDA: [[DIST_OUTER_LOOP_BODY]]:
+      // LAMBDA: br label %[[DIST_INNER_LOOP_HEADER:.+]]
+
+      // LAMBDA: [[DIST_INNER_LOOP_HEADER]]:
+      // LAMBDA-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}} [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_UB_VAL_4:%.+]] = load {{.+}} [[OMP_UB]],
+      // LAMBDA: [[CMP_IV_UB_2:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_2]], [[OMP_UB_VAL_4]]
+      // LAMBDA: br{{.+}} [[CMP_IV_UB_2]], label %[[DIST_INNER_LOOP_BODY:.+]], label %[[DIST_INNER_LOOP_END:.+]]
+
+      // check that PrevLB and PrevUB are passed to the 'for'
+      // LAMBDA: [[DIST_INNER_LOOP_BODY]]:
+      // LAMBDA-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+      // check that distlb and distub are properly passed to fork_call
+      // LAMBDA-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_3:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+      // LAMBDA-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_3:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+      // LAMBDA: br label %[[DIST_INNER_LOOP_INC:.+]]
+
+      // check DistInc
+      // LAMBDA: [[DIST_INNER_LOOP_INC]]:
+      // LAMBDA-DAG: [[OMP_IV_VAL_3:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+      // LAMBDA: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_3]], [[OMP_ST_VAL_1]]
+      // LAMBDA: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[DIST_INNER_LOOP_HEADER]]
+
+      // LAMBDA: [[DIST_INNER_LOOP_END]]:
+      // LAMBDA: br label %[[DIST_OUTER_LOOP_INC:.+]]
+
+      // LAMBDA: [[DIST_OUTER_LOOP_INC]]:
+      // check NextLB and NextUB
+      // LAMBDA-DAG: [[OMP_LB_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_LB]],
+      // LAMBDA-DAG: [[OMP_ST_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_ST]],
+      // LAMBDA-DAG: [[OMP_LB_NEXT:%.+]] = add{{.+}} [[OMP_LB_VAL_2]], [[OMP_ST_VAL_2]]
+      // LAMBDA: store{{.+}} [[OMP_LB_NEXT]], {{.+}}* [[OMP_LB]],
+      // LAMBDA-DAG: [[OMP_UB_VAL_5:%.+]] = load{{.+}}, {{.+}} [[OMP_UB]],
+      // LAMBDA-DAG: [[OMP_ST_VAL_3:%.+]] = load{{.+}}, {{.+}} [[OMP_ST]],
+      // LAMBDA-DAG: [[OMP_UB_NEXT:%.+]] = add{{.+}} [[OMP_UB_VAL_5]], [[OMP_ST_VAL_3]]
+      // LAMBDA: store{{.+}} [[OMP_UB_NEXT]], {{.+}}* [[OMP_UB]],
+      // LAMBDA: br label %[[DIST_OUTER_LOOP_HEADER]]
+
+      // outer loop exit
+      // LAMBDA: [[DIST_OUTER_LOOP_END]]:
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+
+      // skip implementation of 'parallel for': using default scheduling and was tested above
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+
+    // schedule: static no chunk
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_4]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_4:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd schedule(static)
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_4]](
+      // LAMBDA-DAG: [[OMP_IV:%.omp.iv]] = alloca
+      // LAMBDA-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+      // LAMBDA-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+      // LAMBDA-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+      // LAMBDA: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_4:@.+]] to {{.+}},
+      // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+      // LAMBDA: ret
+
+      // 'parallel for' implementation is the same as the case without schedule clase (static no chunk is the default)
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED_4]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+      // LAMBDA-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+      // initialize lb and ub to PrevLB and PrevUB
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+      // LAMBDA-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+      // PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+      // In this case we use EUB
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+      // LAMBDA-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+      // LAMBDA: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+      // LAMBDA: [[PF_EUB_TRUE]]:
+      // LAMBDA: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+      // LAMBDA: br label %[[PF_EUB_END:.+]]
+      // LAMBDA-DAG: [[PF_EUB_FALSE]]:
+      // LAMBDA: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA: br label %[[PF_EUB_END]]
+      // LAMBDA-DAG: [[PF_EUB_END]]:
+      // LAMBDA-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+      // LAMBDA: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+      // initialize omp.iv
+      // LAMBDA: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+      // check exit condition
+      // LAMBDA: [[OMP_PF_JUMP_BACK]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+      // LAMBDA: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+      // check that PrevLB and PrevUB are passed to the 'for'
+      // LAMBDA: [[PF_BODY]]:
+      // LAMBDA-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label {{.+}}
+
+      // check stride 1 for 'for' in 'distribute parallel for simd'
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+      // LAMBDA: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_JUMP_BACK]]
+
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+
+    // schedule: static chunk
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_5]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_5:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd schedule(static, ch)
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_5]](
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+      // LAMBDA: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_5:@.+]] to {{.+}},
+      // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+      // LAMBDA: ret
+
+      // 'parallel for' implementation using outer and inner loops and PrevEUB
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED_5]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}})
+      // LAMBDA-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+      // initialize lb and ub to PrevLB and PrevUB
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+      // LAMBDA-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 33, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+      // check PrevEUB (using PrevUB instead of NumIt as upper bound)
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_HEADER]]:
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA-64-DAG: [[OMP_PF_UB_VAL_CONV:%.+]] = sext{{.+}} [[OMP_PF_UB_VAL_1]] to
+      // LAMBDA: [[PF_PREV_UB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_CONV]], [[PF_PREV_UB_VAL_1]]
+      // LAMBDA-32-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_PREV_UB_VAL_1]]
+      // LAMBDA: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+      // LAMBDA: [[PF_EUB_TRUE]]:
+      // LAMBDA: [[PF_PREV_UB_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA: br label %[[PF_EUB_END:.+]]
+      // LAMBDA-DAG: [[PF_EUB_FALSE]]:
+      // LAMBDA: [[OMP_PF_UB_VAL_2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+      // LAMBDA-64: [[OMP_PF_UB_VAL_2_CONV:%.+]] = sext{{.+}} [[OMP_PF_UB_VAL_2]] to
+      // LAMBDA: br label %[[PF_EUB_END]]
+      // LAMBDA-DAG: [[PF_EUB_END]]:
+      // LAMBDA-64-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_PREV_UB_VAL_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL_2_CONV]], %[[PF_EUB_FALSE]] ]
+      // LAMBDA-32-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_PREV_UB_VAL_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL_2]], %[[PF_EUB_FALSE]] ]
+      // LAMBDA-64-DAG: [[PF_EUB_RES_CONV:%.+]] = trunc{{.+}} [[PF_EUB_RES]] to
+      // LAMBDA-64: store{{.+}} [[PF_EUB_RES_CONV]],{{.+}}  [[OMP_PF_UB]],
+      // LAMBDA-32: store{{.+}} [[PF_EUB_RES]], {{.+}} [[OMP_PF_UB]],
+
+      // initialize omp.iv (IV = LB)
+      // LAMBDA: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+
+      // outer loop: while (IV < UB) {
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB_1:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+      // LAMBDA: br{{.+}} [[PF_CMP_IV_UB_1]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_BODY]]:
+      // LAMBDA: br label %[[OMP_PF_INNER_FOR_HEADER:.+]]
+
+      // LAMBDA: [[OMP_PF_INNER_FOR_HEADER]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+      // LAMBDA: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+      // LAMBDA: [[OMP_PF_INNER_LOOP_BODY]]:
+      // LAMBDA-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // skip body branch
+      // LAMBDA: br{{.+}}
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+      // IV = IV + 1 and inner loop latch
+      // LAMBDA: [[OMP_PF_INNER_LOOP_INC]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+      // LAMBDA-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_PF_INNER_FOR_HEADER]]
+
+      // check NextLB and NextUB
+      // LAMBDA: [[OMP_PF_INNER_LOOP_END]]:
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_INC]]:
+      // LAMBDA-DAG: [[OMP_PF_LB_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA-DAG: [[OMP_PF_ST_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_ST]],
+      // LAMBDA-DAG: [[OMP_PF_LB_NEXT:%.+]] = add{{.+}} [[OMP_PF_LB_VAL_2]], [[OMP_PF_ST_VAL_1]]
+      // LAMBDA: store{{.+}} [[OMP_PF_LB_NEXT]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_5:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+      // LAMBDA-DAG: [[OMP_PF_ST_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_ST]],
+      // LAMBDA-DAG: [[OMP_PF_UB_NEXT:%.+]] = add{{.+}} [[OMP_PF_UB_VAL_5]], [[OMP_PF_ST_VAL_2]]
+      // LAMBDA: store{{.+}} [[OMP_PF_UB_NEXT]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_END]]:
+      // LAMBDA-DAG: call void @__kmpc_for_static_fini(
+      // LAMBDA: ret
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+
+    // schedule: dynamic no chunk
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_6]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_6:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd schedule(dynamic)
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_6]](
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+      // LAMBDA: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_6:@.+]] to {{.+}},
+      // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+      // LAMBDA: ret
+
+      // 'parallel for' implementation using outer and inner loops and PrevEUB
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED_6]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+      // LAMBDA-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+      // initialize lb and ub to PrevLB and PrevUB
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+      // LAMBDA-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-DAG: [[OMP_PF_LB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+      // LAMBDA: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, {{.+}} 35, {{.+}} [[OMP_PF_LB_VAL]], {{.+}} [[OMP_PF_UB_VAL]], {{.+}}, {{.+}})
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_HEADER]]:
+      // LAMBDA: [[IS_FIN:%.+]] = call{{.+}} @__kmpc_dispatch_next_4({{.+}}, {{.+}}, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]], {{.+}}* [[OMP_PF_ST]])
+      // LAMBDA: [[IS_FIN_CMP:%.+]] = icmp{{.+}} [[IS_FIN]], 0
+      // LAMBDA: br{{.+}} [[IS_FIN_CMP]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+      // initialize omp.iv (IV = LB)
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_BODY]]:
+      // LAMBDA-DAG: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA-DAG: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_HEADER:.+]]
+
+      // LAMBDA: [[OMP_PF_INNER_LOOP_HEADER]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+      // LAMBDA: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+      // LAMBDA: [[OMP_PF_INNER_LOOP_BODY]]:
+      // LAMBDA-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // skip body branch
+      // LAMBDA: br{{.+}}
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+      // IV = IV + 1 and inner loop latch
+      // LAMBDA: [[OMP_PF_INNER_LOOP_INC]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+      // LAMBDA-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_HEADER]]
+
+      // check NextLB and NextUB
+      // LAMBDA: [[OMP_PF_INNER_LOOP_END]]:
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_INC]]:
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_END]]:
+      // LAMBDA: ret
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+
+    // schedule: dynamic chunk
+    #pragma omp target
+    #pragma omp teams
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN_7]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_7:@.+]] to {{.+}})
+
+    #pragma omp distribute parallel for simd schedule(dynamic, ch)
+    for (int i = 0; i < n; ++i) {
+      a[i] = b[i] + c[i];
+      // LAMBDA: define{{.+}} void [[OMP_OUTLINED_7]](
+      // LAMBDA: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+      // LAMBDA: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_7:@.+]] to {{.+}},
+      // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+      // LAMBDA: ret
+
+      // 'parallel for' implementation using outer and inner loops and PrevEUB
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED_7]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}})
+      // LAMBDA-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+      // LAMBDA-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+      // initialize lb and ub to PrevLB and PrevUB
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+      // LAMBDA-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+      // LAMBDA-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+      // LAMBDA-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+      // LAMBDA-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+      // LAMBDA-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA-DAG: [[OMP_PF_LB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+      // LAMBDA: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, {{.+}} 35, {{.+}} [[OMP_PF_LB_VAL]], {{.+}} [[OMP_PF_UB_VAL]], {{.+}}, {{.+}})
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_HEADER]]:
+      // LAMBDA: [[IS_FIN:%.+]] = call{{.+}} @__kmpc_dispatch_next_4({{.+}}, {{.+}}, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]], {{.+}}* [[OMP_PF_ST]])
+      // LAMBDA: [[IS_FIN_CMP:%.+]] = icmp{{.+}} [[IS_FIN]], 0
+      // LAMBDA: br{{.+}} [[IS_FIN_CMP]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+      // initialize omp.iv (IV = LB)
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_BODY]]:
+      // LAMBDA-DAG: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+      // LAMBDA-DAG: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_HEADER:.+]]
+
+      // LAMBDA: [[OMP_PF_INNER_LOOP_HEADER]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // LAMBDA-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+      // LAMBDA: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+      // LAMBDA: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+      // LAMBDA: [[OMP_PF_INNER_LOOP_BODY]]:
+      // LAMBDA-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+      // skip body branch
+      // LAMBDA: br{{.+}}
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+      // IV = IV + 1 and inner loop latch
+      // LAMBDA: [[OMP_PF_INNER_LOOP_INC]]:
+      // LAMBDA-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+      // LAMBDA-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+      // LAMBDA-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+      // LAMBDA: br label %[[OMP_PF_INNER_LOOP_HEADER]]
+
+      // check NextLB and NextUB
+      // LAMBDA: [[OMP_PF_INNER_LOOP_END]]:
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_INC]]:
+      // LAMBDA: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+      // LAMBDA: [[OMP_PF_OUTER_LOOP_END]]:
+      // LAMBDA: ret
+      [&]() {
+	a[i] = b[i] + c[i];
+      }();
+    }
+  }();
+  return 0;
+#else
+  // CHECK-LABEL: @main
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_2:@.+]](
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_3:@.+]](
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_4:@.+]](
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_5:@.+]](
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_6:@.+]](
+
+  // CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+  // CHECK: call void [[OFFLOADING_FUN_7:@.+]](
+
+  // CHECK: call{{.+}} [[TMAIN:@.+]]()
+
+  // no schedule clauses
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_1]](
+    // CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+    // CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+    // CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+    // CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+
+    // check EUB for distribute
+    // CHECK-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+    // CHECK: [[NUM_IT_1:%.+]] = load{{.+}},
+    // CHECK-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+    // CHECK: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+    // CHECK-DAG: [[EUB_TRUE]]:
+    // CHECK: [[NUM_IT_2:%.+]] = load{{.+}},
+    // CHECK: br label %[[EUB_END:.+]]
+    // CHECK-DAG: [[EUB_FALSE]]:
+    // CHECK: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+    // CHECK: br label %[[EUB_END]]
+    // CHECK-DAG: [[EUB_END]]:
+    // CHECK-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+    // CHECK: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+    // initialize omp.iv
+    // CHECK: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+    // CHECK: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_JUMP_BACK:.+]]
+
+    // check exit condition
+    // CHECK: [[OMP_JUMP_BACK]]:
+    // CHECK-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+    // CHECK-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+    // CHECK: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+    // CHECK: br {{.+}} [[CMP_IV_UB]], label %[[DIST_BODY:.+]], label %[[DIST_END:.+]]
+
+    // check that PrevLB and PrevUB are passed to the 'for'
+    // CHECK: [[DIST_BODY]]:
+    // CHECK-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+    // check that distlb and distub are properly passed to fork_call
+    // CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+    // CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+    // CHECK: br label %[[DIST_INC:.+]]
+
+    // increment by stride (distInc - 'parallel for' executes the whole chunk) and latch
+    // CHECK: [[DIST_INC]]:
+    // CHECK-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+    // CHECK-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+    // CHECK: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_2]], [[OMP_ST_VAL_1]]
+    // CHECK: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_JUMP_BACK]]
+
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+
+    // implementation of 'parallel for'
+    // CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_1]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+    // CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+    // initialize lb and ub to PrevLB and PrevUB
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+    // CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+    // PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+    // In this case we use EUB
+    // CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+    // CHECK-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+    // CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+    // CHECK: [[PF_EUB_TRUE]]:
+    // CHECK: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+    // CHECK: br label %[[PF_EUB_END:.+]]
+    // CHECK-DAG: [[PF_EUB_FALSE]]:
+    // CHECK: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK: br label %[[PF_EUB_END]]
+    // CHECK-DAG: [[PF_EUB_END]]:
+    // CHECK-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+    // CHECK: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+    // initialize omp.iv
+    // CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+    // check exit condition
+    // CHECK: [[OMP_PF_JUMP_BACK]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+    // CHECK: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+    // check that PrevLB and PrevUB are passed to the 'for'
+    // CHECK: [[PF_BODY]]:
+    // CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label {{.+}}
+
+    // check stride 1 for 'for' in 'distribute parallel for simd'
+    // CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+    // CHECK: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_JUMP_BACK]]
+
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+  }
+
+  // dist_schedule: static no chunk
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define{{.+}} void [[OFFLOADING_FUN_2]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_2:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd dist_schedule(static)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_2]](
+    // CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+    // CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+    // CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+    // CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+
+    // check EUB for distribute
+    // CHECK-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+    // CHECK: [[NUM_IT_1:%.+]] = load{{.+}},
+    // CHECK-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+    // CHECK: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+    // CHECK-DAG: [[EUB_TRUE]]:
+    // CHECK: [[NUM_IT_2:%.+]] = load{{.+}},
+    // CHECK: br label %[[EUB_END:.+]]
+    // CHECK-DAG: [[EUB_FALSE]]:
+    // CHECK: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+    // CHECK: br label %[[EUB_END]]
+    // CHECK-DAG: [[EUB_END]]:
+    // CHECK-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+    // CHECK: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+    // initialize omp.iv
+    // CHECK: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+    // CHECK: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_JUMP_BACK:.+]]
+
+    // check exit condition
+    // CHECK: [[OMP_JUMP_BACK]]:
+    // CHECK-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+    // CHECK-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+    // CHECK: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+    // CHECK: br {{.+}} [[CMP_IV_UB]], label %[[DIST_BODY:.+]], label %[[DIST_END:.+]]
+
+    // check that PrevLB and PrevUB are passed to the 'for'
+    // CHECK: [[DIST_BODY]]:
+    // CHECK-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+    // check that distlb and distub are properly passed to fork_call
+    // CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_2:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+    // CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_2:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+    // CHECK: br label %[[DIST_INC:.+]]
+
+    // increment by stride (distInc - 'parallel for' executes the whole chunk) and latch
+    // CHECK: [[DIST_INC]]:
+    // CHECK-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+    // CHECK-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+    // CHECK: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_2]], [[OMP_ST_VAL_1]]
+    // CHECK: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_JUMP_BACK]]
+
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+
+    // implementation of 'parallel for'
+    // CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_2]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+    // CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+    // initialize lb and ub to PrevLB and PrevUB
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+    // CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+    // PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+    // In this case we use EUB
+    // CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+    // CHECK-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+    // CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+    // CHECK: [[PF_EUB_TRUE]]:
+    // CHECK: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+    // CHECK: br label %[[PF_EUB_END:.+]]
+    // CHECK-DAG: [[PF_EUB_FALSE]]:
+    // CHECK: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK: br label %[[PF_EUB_END]]
+    // CHECK-DAG: [[PF_EUB_END]]:
+    // CHECK-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+    // CHECK: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+    // initialize omp.iv
+    // CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+    // check exit condition
+    // CHECK: [[OMP_PF_JUMP_BACK]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+    // CHECK: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+    // check that PrevLB and PrevUB are passed to the 'for'
+    // CHECK: [[PF_BODY]]:
+    // CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label {{.+}}
+
+    // check stride 1 for 'for' in 'distribute parallel for simd'
+    // CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+    // CHECK: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_JUMP_BACK]]
+
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+  }
+
+  // dist_schedule: static chunk
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define{{.+}} void [[OFFLOADING_FUN_3]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_3:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd dist_schedule(static, ch)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_3]](
+    // CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+    // CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+    // CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+    // CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+    // unlike the previous tests, in this one we have a outer and inner loop for 'distribute'
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91,
+    // CHECK: br label %[[DIST_OUTER_LOOP_HEADER:.+]]
+
+    // CHECK: [[DIST_OUTER_LOOP_HEADER]]:
+    // check EUB for distribute
+    // CHECK-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+    // CHECK: [[NUM_IT_1:%.+]] = load{{.+}},
+    // CHECK-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+    // CHECK: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+    // CHECK-DAG: [[EUB_TRUE]]:
+    // CHECK: [[NUM_IT_2:%.+]] = load{{.+}},
+    // CHECK: br label %[[EUB_END:.+]]
+    // CHECK-DAG: [[EUB_FALSE]]:
+    // CHECK: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+    // CHECK: br label %[[EUB_END]]
+    // CHECK-DAG: [[EUB_END]]:
+    // CHECK-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+    // CHECK: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+    // initialize omp.iv
+    // CHECK: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+    // CHECK: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+
+    // check exit condition
+    // CHECK-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+    // CHECK-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+    // CHECK: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+    // CHECK: br {{.+}} [[CMP_IV_UB]], label %[[DIST_OUTER_LOOP_BODY:.+]], label %[[DIST_OUTER_LOOP_END:.+]]
+
+    // CHECK: [[DIST_OUTER_LOOP_BODY]]:
+    // CHECK: br label %[[DIST_INNER_LOOP_HEADER:.+]]
+
+    // CHECK: [[DIST_INNER_LOOP_HEADER]]:
+    // CHECK-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}} [[OMP_IV]],
+    // CHECK-DAG: [[OMP_UB_VAL_4:%.+]] = load {{.+}} [[OMP_UB]],
+    // CHECK: [[CMP_IV_UB_2:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_2]], [[OMP_UB_VAL_4]]
+    // CHECK: br{{.+}} [[CMP_IV_UB_2]], label %[[DIST_INNER_LOOP_BODY:.+]], label %[[DIST_INNER_LOOP_END:.+]]
+
+    // check that PrevLB and PrevUB are passed to the 'for'
+    // CHECK: [[DIST_INNER_LOOP_BODY]]:
+    // CHECK-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+    // check that distlb and distub are properly passed to fork_call
+    // CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_3:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+    // CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_3:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+    // CHECK: br label %[[DIST_INNER_LOOP_INC:.+]]
+
+    // check DistInc
+    // CHECK: [[DIST_INNER_LOOP_INC]]:
+    // CHECK-DAG: [[OMP_IV_VAL_3:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+    // CHECK-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+    // CHECK: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_3]], [[OMP_ST_VAL_1]]
+    // CHECK: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[DIST_INNER_LOOP_HEADER]]
+
+    // CHECK: [[DIST_INNER_LOOP_END]]:
+    // CHECK: br label %[[DIST_OUTER_LOOP_INC:.+]]
+
+    // CHECK: [[DIST_OUTER_LOOP_INC]]:
+    // check NextLB and NextUB
+    // CHECK-DAG: [[OMP_LB_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_LB]],
+    // CHECK-DAG: [[OMP_ST_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_ST]],
+    // CHECK-DAG: [[OMP_LB_NEXT:%.+]] = add{{.+}} [[OMP_LB_VAL_2]], [[OMP_ST_VAL_2]]
+    // CHECK: store{{.+}} [[OMP_LB_NEXT]], {{.+}}* [[OMP_LB]],
+    // CHECK-DAG: [[OMP_UB_VAL_5:%.+]] = load{{.+}}, {{.+}} [[OMP_UB]],
+    // CHECK-DAG: [[OMP_ST_VAL_3:%.+]] = load{{.+}}, {{.+}} [[OMP_ST]],
+    // CHECK-DAG: [[OMP_UB_NEXT:%.+]] = add{{.+}} [[OMP_UB_VAL_5]], [[OMP_ST_VAL_3]]
+    // CHECK: store{{.+}} [[OMP_UB_NEXT]], {{.+}}* [[OMP_UB]],
+    // CHECK: br label %[[DIST_OUTER_LOOP_HEADER]]
+
+    // outer loop exit
+    // CHECK: [[DIST_OUTER_LOOP_END]]:
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+
+    // skip implementation of 'parallel for': using default scheduling and was tested above
+  }
+
+  // schedule: static no chunk
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define{{.+}} void [[OFFLOADING_FUN_4]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_4:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd schedule(static)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_4]](
+    // CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+    // CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+    // CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+    // CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+    // CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_4:@.+]] to {{.+}},
+    // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+    // CHECK: ret
+
+    // 'parallel for' implementation is the same as the case without schedule clase (static no chunk is the default)
+    // CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_4]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+    // CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+    // initialize lb and ub to PrevLB and PrevUB
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+    // CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+    // PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+    // In this case we use EUB
+    // CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+    // CHECK-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+    // CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+    // CHECK: [[PF_EUB_TRUE]]:
+    // CHECK: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+    // CHECK: br label %[[PF_EUB_END:.+]]
+    // CHECK-DAG: [[PF_EUB_FALSE]]:
+    // CHECK: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK: br label %[[PF_EUB_END]]
+    // CHECK-DAG: [[PF_EUB_END]]:
+    // CHECK-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+    // CHECK: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+    // initialize omp.iv
+    // CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+    // check exit condition
+    // CHECK: [[OMP_PF_JUMP_BACK]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+    // CHECK: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+    // check that PrevLB and PrevUB are passed to the 'for'
+    // CHECK: [[PF_BODY]]:
+    // CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label {{.+}}
+
+    // check stride 1 for 'for' in 'distribute parallel for simd'
+    // CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+    // CHECK: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_JUMP_BACK]]
+
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+  }
+
+  // schedule: static chunk
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define{{.+}} void [[OFFLOADING_FUN_5]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_5:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd schedule(static, ch)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_5]](
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+    // CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_5:@.+]] to {{.+}},
+    // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+    // CHECK: ret
+
+    // 'parallel for' implementation using outer and inner loops and PrevEUB
+    // CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_5]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}})
+    // CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+    // initialize lb and ub to PrevLB and PrevUB
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+    // CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 33, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+    // check PrevEUB (using PrevUB instead of NumIt as upper bound)
+    // CHECK: [[OMP_PF_OUTER_LOOP_HEADER]]:
+    // CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK-64-DAG: [[OMP_PF_UB_VAL_CONV:%.+]] = sext{{.+}} [[OMP_PF_UB_VAL_1]] to
+    // CHECK: [[PF_PREV_UB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_CONV]], [[PF_PREV_UB_VAL_1]]
+    // CHECK-32-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_PREV_UB_VAL_1]]
+    // CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+    // CHECK: [[PF_EUB_TRUE]]:
+    // CHECK: [[PF_PREV_UB_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK: br label %[[PF_EUB_END:.+]]
+    // CHECK-DAG: [[PF_EUB_FALSE]]:
+    // CHECK: [[OMP_PF_UB_VAL_2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+    // CHECK-64: [[OMP_PF_UB_VAL_2_CONV:%.+]] = sext{{.+}} [[OMP_PF_UB_VAL_2]] to
+    // CHECK: br label %[[PF_EUB_END]]
+    // CHECK-DAG: [[PF_EUB_END]]:
+    // CHECK-64-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_PREV_UB_VAL_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL_2_CONV]], %[[PF_EUB_FALSE]] ]
+    // CHECK-32-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_PREV_UB_VAL_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL_2]], %[[PF_EUB_FALSE]] ]
+    // CHECK-64-DAG: [[PF_EUB_RES_CONV:%.+]] = trunc{{.+}} [[PF_EUB_RES]] to
+    // CHECK-64: store{{.+}} [[PF_EUB_RES_CONV]],{{.+}}  [[OMP_PF_UB]],
+    // CHECK-32: store{{.+}} [[PF_EUB_RES]], {{.+}} [[OMP_PF_UB]],
+
+    // initialize omp.iv (IV = LB)
+    // CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+
+    // outer loop: while (IV < UB) {
+    // CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB_1:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+    // CHECK: br{{.+}} [[PF_CMP_IV_UB_1]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_BODY]]:
+    // CHECK: br label %[[OMP_PF_INNER_FOR_HEADER:.+]]
+
+    // CHECK: [[OMP_PF_INNER_FOR_HEADER]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+    // CHECK: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+    // CHECK: [[OMP_PF_INNER_LOOP_BODY]]:
+    // CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // skip body branch
+    // CHECK: br{{.+}}
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+    // IV = IV + 1 and inner loop latch
+    // CHECK: [[OMP_PF_INNER_LOOP_INC]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+    // CHECK-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+    // CHECK-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_PF_INNER_FOR_HEADER]]
+
+    // check NextLB and NextUB
+    // CHECK: [[OMP_PF_INNER_LOOP_END]]:
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_INC]]:
+    // CHECK-DAG: [[OMP_PF_LB_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK-DAG: [[OMP_PF_ST_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_ST]],
+    // CHECK-DAG: [[OMP_PF_LB_NEXT:%.+]] = add{{.+}} [[OMP_PF_LB_VAL_2]], [[OMP_PF_ST_VAL_1]]
+    // CHECK: store{{.+}} [[OMP_PF_LB_NEXT]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_5:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+    // CHECK-DAG: [[OMP_PF_ST_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_ST]],
+    // CHECK-DAG: [[OMP_PF_UB_NEXT:%.+]] = add{{.+}} [[OMP_PF_UB_VAL_5]], [[OMP_PF_ST_VAL_2]]
+    // CHECK: store{{.+}} [[OMP_PF_UB_NEXT]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_END]]:
+    // CHECK-DAG: call void @__kmpc_for_static_fini(
+    // CHECK: ret
+  }
+
+  // schedule: dynamic no chunk
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define{{.+}} void [[OFFLOADING_FUN_6]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_6:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd schedule(dynamic)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_6]](
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+    // CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_6:@.+]] to {{.+}},
+    // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+    // CHECK: ret
+
+    // 'parallel for' implementation using outer and inner loops and PrevEUB
+    // CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_6]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+    // CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+    // initialize lb and ub to PrevLB and PrevUB
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+    // CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-DAG: [[OMP_PF_LB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+    // CHECK: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, {{.+}} 35, {{.+}} [[OMP_PF_LB_VAL]], {{.+}} [[OMP_PF_UB_VAL]], {{.+}}, {{.+}})
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_HEADER]]:
+    // CHECK: [[IS_FIN:%.+]] = call{{.+}} @__kmpc_dispatch_next_4({{.+}}, {{.+}}, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]], {{.+}}* [[OMP_PF_ST]])
+    // CHECK: [[IS_FIN_CMP:%.+]] = icmp{{.+}} [[IS_FIN]], 0
+    // CHECK: br{{.+}} [[IS_FIN_CMP]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+    // initialize omp.iv (IV = LB)
+    // CHECK: [[OMP_PF_OUTER_LOOP_BODY]]:
+    // CHECK-DAG: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK-DAG: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER:.+]]
+
+    // CHECK: [[OMP_PF_INNER_LOOP_HEADER]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+    // CHECK: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+    // CHECK: [[OMP_PF_INNER_LOOP_BODY]]:
+    // CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // skip body branch
+    // CHECK: br{{.+}}
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+    // IV = IV + 1 and inner loop latch
+    // CHECK: [[OMP_PF_INNER_LOOP_INC]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+    // CHECK-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+    // CHECK-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER]]
+
+    // check NextLB and NextUB
+    // CHECK: [[OMP_PF_INNER_LOOP_END]]:
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_INC]]:
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_END]]:
+    // CHECK: ret
+  }
+
+  // schedule: dynamic chunk
+  #pragma omp target
+  #pragma omp teams
+  // CHECK: define{{.+}} void [[OFFLOADING_FUN_7]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_7:@.+]] to {{.+}})
+
+  #pragma omp distribute parallel for simd schedule(dynamic, ch)
+  for (int i = 0; i < n; ++i) {
+    a[i] = b[i] + c[i];
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_7]](
+    // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+    // CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_7:@.+]] to {{.+}},
+    // skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+    // CHECK: ret
+
+    // 'parallel for' implementation using outer and inner loops and PrevEUB
+    // CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_7]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}})
+    // CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+    // CHECK-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+    // initialize lb and ub to PrevLB and PrevUB
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+    // CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+    // CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+    // CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+    // CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+    // CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+    // CHECK-DAG: [[OMP_PF_LB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+    // CHECK: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, {{.+}} 35, {{.+}} [[OMP_PF_LB_VAL]], {{.+}} [[OMP_PF_UB_VAL]], {{.+}}, {{.+}})
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_HEADER]]:
+    // CHECK: [[IS_FIN:%.+]] = call{{.+}} @__kmpc_dispatch_next_4({{.+}}, {{.+}}, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]], {{.+}}* [[OMP_PF_ST]])
+    // CHECK: [[IS_FIN_CMP:%.+]] = icmp{{.+}} [[IS_FIN]], 0
+    // CHECK: br{{.+}} [[IS_FIN_CMP]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+    // initialize omp.iv (IV = LB)
+    // CHECK: [[OMP_PF_OUTER_LOOP_BODY]]:
+    // CHECK-DAG: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+    // CHECK-DAG: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER:.+]]
+
+    // CHECK: [[OMP_PF_INNER_LOOP_HEADER]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // CHECK-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+    // CHECK: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+    // CHECK: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+    // CHECK: [[OMP_PF_INNER_LOOP_BODY]]:
+    // CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+    // skip body branch
+    // CHECK: br{{.+}}
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+    // IV = IV + 1 and inner loop latch
+    // CHECK: [[OMP_PF_INNER_LOOP_INC]]:
+    // CHECK-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+    // CHECK-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+    // CHECK-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+    // CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER]]
+
+    // check NextLB and NextUB
+    // CHECK: [[OMP_PF_INNER_LOOP_END]]:
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_INC]]:
+    // CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+    // CHECK: [[OMP_PF_OUTER_LOOP_END]]:
+    // CHECK: ret
+  }
+
+  return tmain<int>();
+#endif
+}
+
+// check code
+// CHECK: define{{.+}} [[TMAIN]]()
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_2:@.+]](
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_3:@.+]](
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_4:@.+]](
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_5:@.+]](
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_6:@.+]](
+
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_7:@.+]](
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_1]](
+// CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+// CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+// CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+// CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+
+// check EUB for distribute
+// CHECK-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+// CHECK: [[NUM_IT_1:%.+]] = load{{.+}},
+// CHECK-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+// CHECK: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+// CHECK-DAG: [[EUB_TRUE]]:
+// CHECK: [[NUM_IT_2:%.+]] = load{{.+}},
+// CHECK: br label %[[EUB_END:.+]]
+// CHECK-DAG: [[EUB_FALSE]]:
+// CHECK: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+// CHECK: br label %[[EUB_END]]
+// CHECK-DAG: [[EUB_END]]:
+// CHECK-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+// CHECK: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+// initialize omp.iv
+// CHECK: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+// CHECK: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_JUMP_BACK:.+]]
+
+// check exit condition
+// CHECK: [[OMP_JUMP_BACK]]:
+// CHECK-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+// CHECK-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+// CHECK: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+// CHECK: br {{.+}} [[CMP_IV_UB]], label %[[DIST_BODY:.+]], label %[[DIST_END:.+]]
+
+// check that PrevLB and PrevUB are passed to the 'for'
+// CHECK: [[DIST_BODY]]:
+// CHECK-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+// CHECK-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+// CHECK-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+// check that distlb and distub are properly passed to fork_call
+// CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+// CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+// CHECK: br label %[[DIST_INC:.+]]
+
+// increment by stride (distInc - 'parallel for' executes the whole chunk) and latch
+// CHECK: [[DIST_INC]]:
+// CHECK-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+// CHECK-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+// CHECK: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_2]], [[OMP_ST_VAL_1]]
+// CHECK: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_JUMP_BACK]]
+
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// implementation of 'parallel for'
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_1]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+// CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+// initialize lb and ub to PrevLB and PrevUB
+// CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+// CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+// CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+// PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+// In this case we use EUB
+// CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+// CHECK-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+// CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+// CHECK: [[PF_EUB_TRUE]]:
+// CHECK: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+// CHECK: br label %[[PF_EUB_END:.+]]
+// CHECK-DAG: [[PF_EUB_FALSE]]:
+// CHECK: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK: br label %[[PF_EUB_END]]
+// CHECK-DAG: [[PF_EUB_END]]:
+// CHECK-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+// CHECK: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+// initialize omp.iv
+// CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+// check exit condition
+// CHECK: [[OMP_PF_JUMP_BACK]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+// CHECK: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+// check that PrevLB and PrevUB are passed to the 'for'
+// CHECK: [[PF_BODY]]:
+// CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label {{.+}}
+
+// check stride 1 for 'for' in 'distribute parallel for simd'
+// CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+// CHECK: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_JUMP_BACK]]
+
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_2]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_2:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_2]](
+// CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+// CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+// CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+// CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+
+// check EUB for distribute
+// CHECK-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+// CHECK: [[NUM_IT_1:%.+]] = load{{.+}},
+// CHECK-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+// CHECK: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+// CHECK-DAG: [[EUB_TRUE]]:
+// CHECK: [[NUM_IT_2:%.+]] = load{{.+}},
+// CHECK: br label %[[EUB_END:.+]]
+// CHECK-DAG: [[EUB_FALSE]]:
+// CHECK: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+// CHECK: br label %[[EUB_END]]
+// CHECK-DAG: [[EUB_END]]:
+// CHECK-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+// CHECK: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+// initialize omp.iv
+// CHECK: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+// CHECK: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_JUMP_BACK:.+]]
+
+// check exit condition
+// CHECK: [[OMP_JUMP_BACK]]:
+// CHECK-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+// CHECK-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+// CHECK: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+// CHECK: br {{.+}} [[CMP_IV_UB]], label %[[DIST_BODY:.+]], label %[[DIST_END:.+]]
+
+// check that PrevLB and PrevUB are passed to the 'for'
+// CHECK: [[DIST_BODY]]:
+// CHECK-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+// CHECK-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+// CHECK-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+// check that distlb and distub are properly passed to fork_call
+// CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_2:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+// CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_2:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+// CHECK: br label %[[DIST_INC:.+]]
+
+// increment by stride (distInc - 'parallel for' executes the whole chunk) and latch
+// CHECK: [[DIST_INC]]:
+// CHECK-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+// CHECK-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+// CHECK: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_2]], [[OMP_ST_VAL_1]]
+// CHECK: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_JUMP_BACK]]
+
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// implementation of 'parallel for'
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_2]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+// CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+// initialize lb and ub to PrevLB and PrevUB
+// CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+// CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+// CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+// PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+// In this case we use EUB
+// CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+// CHECK-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+// CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+// CHECK: [[PF_EUB_TRUE]]:
+// CHECK: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+// CHECK: br label %[[PF_EUB_END:.+]]
+// CHECK-DAG: [[PF_EUB_FALSE]]:
+// CHECK: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK: br label %[[PF_EUB_END]]
+// CHECK-DAG: [[PF_EUB_END]]:
+// CHECK-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+// CHECK: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+// initialize omp.iv
+// CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+// check exit condition
+// CHECK: [[OMP_PF_JUMP_BACK]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+// CHECK: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+// check that PrevLB and PrevUB are passed to the 'for'
+// CHECK: [[PF_BODY]]:
+// CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label {{.+}}
+
+// check stride 1 for 'for' in 'distribute parallel for simd'
+// CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+// CHECK: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_JUMP_BACK]]
+
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_3]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_3]](
+// CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+// CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+// CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+// CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+// unlike the previous tests, in this one we have a outer and inner loop for 'distribute'
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91,
+// CHECK: br label %[[DIST_OUTER_LOOP_HEADER:.+]]
+
+// CHECK: [[DIST_OUTER_LOOP_HEADER]]:
+// check EUB for distribute
+// CHECK-DAG: [[OMP_UB_VAL_1:%.+]] = load{{.+}} [[OMP_UB]],
+// CHECK: [[NUM_IT_1:%.+]] = load{{.+}},
+// CHECK-DAG: [[CMP_UB_NUM_IT:%.+]] = icmp sgt {{.+}}  [[OMP_UB_VAL_1]], [[NUM_IT_1]]
+// CHECK: br {{.+}} [[CMP_UB_NUM_IT]], label %[[EUB_TRUE:.+]], label %[[EUB_FALSE:.+]]
+// CHECK-DAG: [[EUB_TRUE]]:
+// CHECK: [[NUM_IT_2:%.+]] = load{{.+}},
+// CHECK: br label %[[EUB_END:.+]]
+// CHECK-DAG: [[EUB_FALSE]]:
+// CHECK: [[OMP_UB_VAL2:%.+]] = load{{.+}} [[OMP_UB]],
+// CHECK: br label %[[EUB_END]]
+// CHECK-DAG: [[EUB_END]]:
+// CHECK-DAG: [[EUB_RES:%.+]] = phi{{.+}} [ [[NUM_IT_2]], %[[EUB_TRUE]] ], [ [[OMP_UB_VAL2]], %[[EUB_FALSE]] ]
+// CHECK: store{{.+}} [[EUB_RES]], {{.+}}* [[OMP_UB]],
+
+// initialize omp.iv
+// CHECK: [[OMP_LB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_LB]],
+// CHECK: store {{.+}} [[OMP_LB_VAL_1]], {{.+}}* [[OMP_IV]],
+
+// check exit condition
+// CHECK-DAG: [[OMP_IV_VAL_1:%.+]] = load {{.+}} [[OMP_IV]],
+// CHECK-DAG: [[OMP_UB_VAL_3:%.+]] = load {{.+}} [[OMP_UB]],
+// CHECK: [[CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_1]], [[OMP_UB_VAL_3]]
+// CHECK: br {{.+}} [[CMP_IV_UB]], label %[[DIST_OUTER_LOOP_BODY:.+]], label %[[DIST_OUTER_LOOP_END:.+]]
+
+// CHECK: [[DIST_OUTER_LOOP_BODY]]:
+// CHECK: br label %[[DIST_INNER_LOOP_HEADER:.+]]
+
+// CHECK: [[DIST_INNER_LOOP_HEADER]]:
+// CHECK-DAG: [[OMP_IV_VAL_2:%.+]] = load {{.+}} [[OMP_IV]],
+// CHECK-DAG: [[OMP_UB_VAL_4:%.+]] = load {{.+}} [[OMP_UB]],
+// CHECK: [[CMP_IV_UB_2:%.+]] = icmp sle {{.+}} [[OMP_IV_VAL_2]], [[OMP_UB_VAL_4]]
+// CHECK: br{{.+}} [[CMP_IV_UB_2]], label %[[DIST_INNER_LOOP_BODY:.+]], label %[[DIST_INNER_LOOP_END:.+]]
+
+// check that PrevLB and PrevUB are passed to the 'for'
+// CHECK: [[DIST_INNER_LOOP_BODY]]:
+// CHECK-DAG: [[OMP_PREV_LB:%.+]] = load {{.+}}, {{.+}} [[OMP_LB]],
+// CHECK-64-DAG: [[OMP_PREV_LB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_LB]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB:%.+]] = load {{.+}}, {{.+}} [[OMP_UB]],
+// CHECK-64-DAG: [[OMP_PREV_UB_EXT:%.+]] = zext {{.+}} [[OMP_PREV_UB]] to {{.+}}
+// check that distlb and distub are properly passed to fork_call
+// CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_3:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_EXT]], i{{[0-9]+}} [[OMP_PREV_UB_EXT]], {{.+}})
+// CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_3:@.+]] to {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB]], i{{[0-9]+}} [[OMP_PREV_UB]], {{.+}})
+// CHECK: br label %[[DIST_INNER_LOOP_INC:.+]]
+
+// check DistInc
+// CHECK: [[DIST_INNER_LOOP_INC]]:
+// CHECK-DAG: [[OMP_IV_VAL_3:%.+]] = load {{.+}}, {{.+}}* [[OMP_IV]],
+// CHECK-DAG: [[OMP_ST_VAL_1:%.+]] = load {{.+}}, {{.+}}* [[OMP_ST]],
+// CHECK: [[OMP_IV_INC:%.+]] = add{{.+}} [[OMP_IV_VAL_3]], [[OMP_ST_VAL_1]]
+// CHECK: store{{.+}} [[OMP_IV_INC]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[DIST_INNER_LOOP_HEADER]]
+
+// CHECK: [[DIST_INNER_LOOP_END]]:
+// CHECK: br label %[[DIST_OUTER_LOOP_INC:.+]]
+
+// CHECK: [[DIST_OUTER_LOOP_INC]]:
+// check NextLB and NextUB
+// CHECK-DAG: [[OMP_LB_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_LB]],
+// CHECK-DAG: [[OMP_ST_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_ST]],
+// CHECK-DAG: [[OMP_LB_NEXT:%.+]] = add{{.+}} [[OMP_LB_VAL_2]], [[OMP_ST_VAL_2]]
+// CHECK: store{{.+}} [[OMP_LB_NEXT]], {{.+}}* [[OMP_LB]],
+// CHECK-DAG: [[OMP_UB_VAL_5:%.+]] = load{{.+}}, {{.+}} [[OMP_UB]],
+// CHECK-DAG: [[OMP_ST_VAL_3:%.+]] = load{{.+}}, {{.+}} [[OMP_ST]],
+// CHECK-DAG: [[OMP_UB_NEXT:%.+]] = add{{.+}} [[OMP_UB_VAL_5]], [[OMP_ST_VAL_3]]
+// CHECK: store{{.+}} [[OMP_UB_NEXT]], {{.+}}* [[OMP_UB]],
+// CHECK: br label %[[DIST_OUTER_LOOP_HEADER]]
+
+// outer loop exit
+// CHECK: [[DIST_OUTER_LOOP_END]]:
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// skip implementation of 'parallel for': using default scheduling and was tested above
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_4]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_4:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_4]](
+// CHECK-DAG: [[OMP_IV:%.omp.iv]] = alloca
+// CHECK-DAG: [[OMP_LB:%.omp.comb.lb]] = alloca
+// CHECK-DAG: [[OMP_UB:%.omp.comb.ub]] = alloca
+// CHECK-DAG: [[OMP_ST:%.omp.stride]] = alloca
+
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_4:@.+]] to {{.+}},
+// skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+// CHECK: ret
+
+// 'parallel for' implementation is the same as the case without schedule clase (static no chunk is the default)
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_4]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+
+// CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+
+// initialize lb and ub to PrevLB and PrevUB
+// CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+// CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+// CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 34, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+
+// PrevEUB is only used when 'for' has a chunked schedule, otherwise EUB is used
+// In this case we use EUB
+// CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK: [[PF_NUM_IT_1:%.+]] = load{{.+}},
+// CHECK-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_NUM_IT_1]]
+// CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+// CHECK: [[PF_EUB_TRUE]]:
+// CHECK: [[PF_NUM_IT_2:%.+]] = load{{.+}},
+// CHECK: br label %[[PF_EUB_END:.+]]
+// CHECK-DAG: [[PF_EUB_FALSE]]:
+// CHECK: [[OMP_PF_UB_VAL2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK: br label %[[PF_EUB_END]]
+// CHECK-DAG: [[PF_EUB_END]]:
+// CHECK-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_NUM_IT_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL2]], %[[PF_EUB_FALSE]] ]
+// CHECK: store{{.+}} [[PF_EUB_RES]],{{.+}}  [[OMP_PF_UB]],
+
+// initialize omp.iv
+// CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_JUMP_BACK:.+]]
+
+// check exit condition
+// CHECK: [[OMP_PF_JUMP_BACK]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load {{.+}} [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load {{.+}} [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB:%.+]] = icmp sle {{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+// CHECK: br {{.+}} [[PF_CMP_IV_UB]], label %[[PF_BODY:.+]], label %[[PF_END:.+]]
+
+// check that PrevLB and PrevUB are passed to the 'for'
+// CHECK: [[PF_BODY]]:
+// CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label {{.+}}
+
+// check stride 1 for 'for' in 'distribute parallel for simd'
+// CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load {{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK: [[OMP_PF_IV_INC:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_2]], 1
+// CHECK: store{{.+}} [[OMP_PF_IV_INC]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_JUMP_BACK]]
+
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_5]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_5:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_5]](
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_5:@.+]] to {{.+}},
+// skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+// CHECK: ret
+
+// 'parallel for' implementation using outer and inner loops and PrevEUB
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_5]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}})
+// CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+// initialize lb and ub to PrevLB and PrevUB
+// CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+// CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+// CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 33, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]],{{.+}})
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+// check PrevEUB (using PrevUB instead of NumIt as upper bound)
+// CHECK: [[OMP_PF_OUTER_LOOP_HEADER]]:
+// CHECK-DAG: [[OMP_PF_UB_VAL_1:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK-64-DAG: [[OMP_PF_UB_VAL_CONV:%.+]] = sext{{.+}} [[OMP_PF_UB_VAL_1]] to
+// CHECK: [[PF_PREV_UB_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_CONV]], [[PF_PREV_UB_VAL_1]]
+// CHECK-32-DAG: [[PF_CMP_UB_NUM_IT:%.+]] = icmp{{.+}} [[OMP_PF_UB_VAL_1]], [[PF_PREV_UB_VAL_1]]
+// CHECK: br i1 [[PF_CMP_UB_NUM_IT]], label %[[PF_EUB_TRUE:.+]], label %[[PF_EUB_FALSE:.+]]
+// CHECK: [[PF_EUB_TRUE]]:
+// CHECK: [[PF_PREV_UB_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK: br label %[[PF_EUB_END:.+]]
+// CHECK-DAG: [[PF_EUB_FALSE]]:
+// CHECK: [[OMP_PF_UB_VAL_2:%.+]] = load{{.+}} [[OMP_PF_UB]],
+// CHECK-64: [[OMP_PF_UB_VAL_2_CONV:%.+]] = sext{{.+}} [[OMP_PF_UB_VAL_2]] to
+// CHECK: br label %[[PF_EUB_END]]
+// CHECK-DAG: [[PF_EUB_END]]:
+// CHECK-64-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_PREV_UB_VAL_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL_2_CONV]], %[[PF_EUB_FALSE]] ]
+// CHECK-32-DAG: [[PF_EUB_RES:%.+]] = phi{{.+}} [ [[PF_PREV_UB_VAL_2]], %[[PF_EUB_TRUE]] ], [ [[OMP_PF_UB_VAL_2]], %[[PF_EUB_FALSE]] ]
+// CHECK-64-DAG: [[PF_EUB_RES_CONV:%.+]] = trunc{{.+}} [[PF_EUB_RES]] to
+// CHECK-64: store{{.+}} [[PF_EUB_RES_CONV]],{{.+}}  [[OMP_PF_UB]],
+// CHECK-32: store{{.+}} [[PF_EUB_RES]], {{.+}} [[OMP_PF_UB]],
+
+// initialize omp.iv (IV = LB)
+// CHECK: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+
+// outer loop: while (IV < UB) {
+// CHECK-DAG: [[OMP_PF_IV_VAL_1:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB_1:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_1]], [[OMP_PF_UB_VAL_3]]
+// CHECK: br{{.+}} [[PF_CMP_IV_UB_1]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_BODY]]:
+// CHECK: br label %[[OMP_PF_INNER_FOR_HEADER:.+]]
+
+// CHECK: [[OMP_PF_INNER_FOR_HEADER]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+// CHECK: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+// CHECK: [[OMP_PF_INNER_LOOP_BODY]]:
+// CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// skip body branch
+// CHECK: br{{.+}}
+// CHECK: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+// IV = IV + 1 and inner loop latch
+// CHECK: [[OMP_PF_INNER_LOOP_INC]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+// CHECK-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+// CHECK-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_PF_INNER_FOR_HEADER]]
+
+// check NextLB and NextUB
+// CHECK: [[OMP_PF_INNER_LOOP_END]]:
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_INC]]:
+// CHECK-DAG: [[OMP_PF_LB_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK-DAG: [[OMP_PF_ST_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_ST]],
+// CHECK-DAG: [[OMP_PF_LB_NEXT:%.+]] = add{{.+}} [[OMP_PF_LB_VAL_2]], [[OMP_PF_ST_VAL_1]]
+// CHECK: store{{.+}} [[OMP_PF_LB_NEXT]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_5:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+// CHECK-DAG: [[OMP_PF_ST_VAL_2:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_ST]],
+// CHECK-DAG: [[OMP_PF_UB_NEXT:%.+]] = add{{.+}} [[OMP_PF_UB_VAL_5]], [[OMP_PF_ST_VAL_2]]
+// CHECK: store{{.+}} [[OMP_PF_UB_NEXT]], {{.+}}* [[OMP_PF_UB]],
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_END]]:
+// CHECK-DAG: call void @__kmpc_for_static_fini(
+// CHECK: ret
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_6]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED_6:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_6]](
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_6:@.+]] to {{.+}},
+// skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+// CHECK: ret
+
+// 'parallel for' implementation using outer and inner loops and PrevEUB
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_6]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}})
+// CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+// initialize lb and ub to PrevLB and PrevUB
+// CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+// CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+// CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-DAG: [[OMP_PF_LB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK-DAG: [[OMP_PF_UB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+// CHECK: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, {{.+}} 35, {{.+}} [[OMP_PF_LB_VAL]], {{.+}} [[OMP_PF_UB_VAL]], {{.+}}, {{.+}})
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_HEADER]]:
+// CHECK: [[IS_FIN:%.+]] = call{{.+}} @__kmpc_dispatch_next_4({{.+}}, {{.+}}, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]], {{.+}}* [[OMP_PF_ST]])
+// CHECK: [[IS_FIN_CMP:%.+]] = icmp{{.+}} [[IS_FIN]], 0
+// CHECK: br{{.+}} [[IS_FIN_CMP]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+// initialize omp.iv (IV = LB)
+// CHECK: [[OMP_PF_OUTER_LOOP_BODY]]:
+// CHECK-DAG: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK-DAG: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER:.+]]
+
+// CHECK: [[OMP_PF_INNER_LOOP_HEADER]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+// CHECK: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+// CHECK: [[OMP_PF_INNER_LOOP_BODY]]:
+// CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// skip body branch
+// CHECK: br{{.+}}
+// CHECK: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+// IV = IV + 1 and inner loop latch
+// CHECK: [[OMP_PF_INNER_LOOP_INC]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+// CHECK-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+// CHECK-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER]]
+
+// check NextLB and NextUB
+// CHECK: [[OMP_PF_INNER_LOOP_END]]:
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_INC]]:
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_END]]:
+// CHECK: ret
+
+// CHECK: define{{.+}} void [[OFFLOADING_FUN_7]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}}* [[OMP_OUTLINED_7:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[OMP_OUTLINED_7]](
+// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92,
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_7:@.+]] to {{.+}},
+// skip rest of implementation of 'distribute' as it is tested above for default dist_schedule case
+// CHECK: ret
+
+// 'parallel for' implementation using outer and inner loops and PrevEUB
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_7]]({{.+}}, {{.+}}, i{{[0-9]+}} [[OMP_PREV_LB_IN:%.+]], i{{[0-9]+}} [[OMP_PREV_UB_IN:%.+]], {{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}})
+// CHECK-DAG: [[OMP_PF_LB:%.omp.lb]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_UB:%.omp.ub]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_IV:%.omp.iv]] = alloca{{.+}},
+// CHECK-DAG: [[OMP_PF_ST:%.omp.stride]] = alloca{{.+}},
+
+// initialize lb and ub to PrevLB and PrevUB
+// CHECK-DAG: store{{.+}} [[OMP_PREV_LB_IN]], {{.+}}* [[PREV_LB_ADDR:%.+]],
+// CHECK-DAG: store{{.+}} [[OMP_PREV_UB_IN]], {{.+}}* [[PREV_UB_ADDR:%.+]],
+// CHECK-DAG: [[OMP_PREV_LB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_LB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_LB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_LB_VAL]] to {{.+}}
+// CHECK-DAG: [[OMP_PREV_UB_VAL:%.+]] = load{{.+}}, {{.+}}* [[PREV_UB_ADDR]],
+// CHECK-64-DAG: [[OMP_PREV_UB_TRC:%.+]] = trunc{{.+}} [[OMP_PREV_UB_VAL]] to {{.+}}
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_LB_TRC]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-64-DAG: store{{.+}} [[OMP_PREV_UB_TRC]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_LB_VAL]], {{.+}}* [[OMP_PF_LB]],
+// CHECK-32-DAG: store{{.+}} [[OMP_PREV_UB_VAL]], {{.+}}* [[OMP_PF_UB]],
+// CHECK-DAG: [[OMP_PF_LB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK-DAG: [[OMP_PF_UB_VAL:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_UB]],
+// CHECK: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, {{.+}} 35, {{.+}} [[OMP_PF_LB_VAL]], {{.+}} [[OMP_PF_UB_VAL]], {{.+}}, {{.+}})
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER:.+]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_HEADER]]:
+// CHECK: [[IS_FIN:%.+]] = call{{.+}} @__kmpc_dispatch_next_4({{.+}}, {{.+}}, {{.+}}, {{.+}}* [[OMP_PF_LB]], {{.+}}* [[OMP_PF_UB]], {{.+}}* [[OMP_PF_ST]])
+// CHECK: [[IS_FIN_CMP:%.+]] = icmp{{.+}} [[IS_FIN]], 0
+// CHECK: br{{.+}} [[IS_FIN_CMP]], label %[[OMP_PF_OUTER_LOOP_BODY:.+]], label %[[OMP_PF_OUTER_LOOP_END:.+]]
+
+// initialize omp.iv (IV = LB)
+// CHECK: [[OMP_PF_OUTER_LOOP_BODY]]:
+// CHECK-DAG: [[OMP_PF_LB_VAL_1:%.+]] = load{{.+}}, {{.+}} [[OMP_PF_LB]],
+// CHECK-DAG: store {{.+}} [[OMP_PF_LB_VAL_1]], {{.+}}* [[OMP_PF_IV]],
+// CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER:.+]]
+
+// CHECK: [[OMP_PF_INNER_LOOP_HEADER]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_2:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// CHECK-DAG: [[OMP_PF_UB_VAL_4:%.+]] = load{{.+}}, {{.+}}* [[OMP_PF_UB]],
+// CHECK: [[PF_CMP_IV_UB_2:%.+]] = icmp{{.+}} [[OMP_PF_IV_VAL_2]], [[OMP_PF_UB_VAL_4]]
+// CHECK: br{{.+}} [[PF_CMP_IV_UB_2]], label %[[OMP_PF_INNER_LOOP_BODY:.+]], label %[[OMP_PF_INNER_LOOP_END:.+]]
+
+// CHECK: [[OMP_PF_INNER_LOOP_BODY]]:
+// CHECK-DAG: {{.+}} = load{{.+}}, {{.+}}* [[OMP_PF_IV]],
+// skip body branch
+// CHECK: br{{.+}}
+// CHECK: br label %[[OMP_PF_INNER_LOOP_INC:.+]]
+
+// IV = IV + 1 and inner loop latch
+// CHECK: [[OMP_PF_INNER_LOOP_INC]]:
+// CHECK-DAG: [[OMP_PF_IV_VAL_3:%.+]] = load{{.+}}, {{.+}}* [[OMP_IV]],
+// CHECK-DAG: [[OMP_PF_NEXT_IV:%.+]] = add{{.+}} [[OMP_PF_IV_VAL_3]], 1
+// CHECK-DAG: store{{.+}} [[OMP_PF_NEXT_IV]], {{.+}}* [[OMP_IV]],
+// CHECK: br label %[[OMP_PF_INNER_LOOP_HEADER]]
+
+// check NextLB and NextUB
+// CHECK: [[OMP_PF_INNER_LOOP_END]]:
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_INC:.+]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_INC]]:
+// CHECK: br label %[[OMP_PF_OUTER_LOOP_HEADER]]
+
+// CHECK: [[OMP_PF_OUTER_LOOP_END]]:
+// CHECK: ret
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..b8d0f4ad7540
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_firstprivate_codegen.cpp
@@ -0,0 +1,611 @@
+// RxUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd firstprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams
+    #pragma omp distribute parallel for simd firstprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double* {{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+
+      // addr alloca's
+      // LAMBDA: [[G_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca float*,
+      // LAMBDA: [[G1_REF:%.+]] = alloca double*,
+
+      // private alloca's
+      // LAMBDA: [[G_PRIV:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIV:%.+]] = alloca double,
+      // LAMBDA: [[TMP_PRIV:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIV:%.+]] = alloca float,
+
+      // transfer input parameters into addr alloca's
+      // LAMBDA-DAG: store {{.+}} [[G_IN]], {{.+}} [[G_ADDR]],
+      // LAMBDA-DAG: store {{.+}} [[G1_IN]], {{.+}} [[G1_ADDR]],
+      // LAMBDA-DAG: store {{.+}} [[SVAR_IN]], {{.+}} [[SVAR_ADDR]],
+      // LAMBDA-DAG: store {{.+}} [[SFVAR_IN]], {{.+}} [[SFVAR_ADDR]],
+
+      // init private alloca's with addr alloca's
+      // g
+      // LAMBDA-DAG: [[G_CONV:%.+]] = load {{.+}}*, {{.+}}** [[G_ADDR]]
+      // LAMBDA-DAG: [[G_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[G_CONV]],
+      // LAMBDA-DAG: store {{.+}} [[G_ADDR_VAL]], {{.+}}* [[G_PRIV]],
+
+      // g1
+      // LAMBDA-DAG: [[TMP_REF:%.+]] = load {{.+}}*, {{.+}}** [[G1_REF]],
+      // LAMBDA-DAG: [[TMP_VAL:%.+]] = load {{.+}}, {{.+}}* [[TMP_REF]],
+      // LAMBDA-DAG: store {{.+}} [[TMP_VAL]], {{.+}}* [[G1_PRIV]]
+      // LAMBDA-DAG: store {{.+}}* [[G1_PRIV]], {{.+}}** [[TMP_PRIV]],
+
+      // svar
+      // LAMBDA-DAG: [[SVAR_REF:%.+]] = load {{.+}}*, {{.+}}** [[SVAR_ADDR]],
+      // LAMBDA-DAG: [[SVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_REF]],
+      // LAMBDA-DAG: store {{.+}} [[SVAR_VAL]], {{.+}}* [[SVAR_PRIV]],
+
+      // sfvar
+      // LAMBDA-DAG: [[SFVAR_REF:%.+]] = load {{.+}}*, {{.+}}** [[SFVAR_ADDR]],
+      // LAMBDA-DAG: [[SFVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SFVAR_REF]],
+      // LAMBDA-DAG: store {{.+}} [[SFVAR_VAL]], {{.+}}* [[SFVAR_PRIV]],
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // pass firstprivate parameters to parallel outlined function
+      // g
+      // LAMBDA-64-DAG: [[G_PRIV_VAL:%.+]] = load {{.+}}, {{.+}}* [[G_PRIV]],
+      // LAMBDA-64: [[G_CAST_CONV:%.+]] = bitcast {{.+}}* [[G_CAST:%.+]] to
+      // LAMBDA-64-DAG: store {{.+}} [[G_PRIV_VAL]], {{.+}}* [[G_CAST_CONV]],
+      // LAMBDA-64-DAG: [[G_PAR:%.+]] = load {{.+}}, {{.+}}* [[G_CAST]],
+
+      // g1
+      // LAMBDA-DAG: [[TMP_PRIV_VAL:%.+]] = load {{.+}}, {{.+}}* [[TMP_PRIV]],
+      // LAMBDA-DAG: [[G1_PRIV_VAL:%.+]] = load {{.+}}, {{.+}}* [[TMP_PRIV_VAL]],
+      // LAMBDA: [[G1_CAST_CONV:%.+]] = bitcast {{.+}}* [[G1_CAST:%.+]] to
+      // LAMBDA-DAG: store {{.+}} [[G1_PRIV_VAL]], {{.+}}* [[G1_CAST_CONV]],
+      // LAMBDA-DAG: [[G1_PAR:%.+]] = load {{.+}}, {{.+}}* [[G1_CAST]],
+
+      // svar
+      // LAMBDA: [[SVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_PRIV]],
+      // LAMBDA-64-DAG: [[SVAR_CAST_CONV:%.+]] = bitcast {{.+}}* [[SVAR_CAST:%.+]] to
+      // LAMBDA-64-DAG: store {{.+}} [[SVAR_VAL]], {{.+}}* [[SVAR_CAST_CONV]],
+      // LAMBDA-32-DAG: store {{.+}} [[SVAR_VAL]], {{.+}}* [[SVAR_CAST:%.+]],
+      // LAMBDA-DAG: [[SVAR_PAR:%.+]] = load {{.+}}, {{.+}}* [[SVAR_CAST]],
+
+      // sfvar
+      // LAMBDA: [[SFVAR_VAL:%.+]] = load {{.+}}, {{.+}}* [[SFVAR_PRIV]],
+      // LAMBDA-DAG: [[SFVAR_CAST_CONV:%.+]] = bitcast {{.+}}* [[SFVAR_CAST:%.+]] to
+      // LAMBDA-DAG: store {{.+}} [[SFVAR_VAL]], {{.+}}* [[SFVAR_CAST_CONV]],
+      // LAMBDA-DAG: [[SFVAR_PAR:%.+]] = load {{.+}}, {{.+}}* [[SFVAR_CAST]],
+
+      // LAMBDA-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[G_PAR]], {{.+}} [[G1_PAR]], {{.+}} [[SVAR_PAR]], {{.+}} [[SFVAR_PAR]])
+      // LAMBDA-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[G_PRIV]], {{.+}} [[G1_PAR]], {{.+}} [[SVAR_PAR]], {{.+}} [[SFVAR_PAR]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: ret void
+
+
+      // LAMBDA-64: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, i{{[0-9]+}} [[G_IN:%.+]], i{{[0-9]+}} [[G1_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]], i{{[0-9]+}} [[SFVAR_IN:%.+]])
+      // LAMBDA-32: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, double* {{.+}} [[G_IN:%.+]], i{{[0-9]+}} [[G1_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]], i{{[0-9]+}} [[SFVAR_IN:%.+]])
+      // skip initial params
+      // LAMBDA: {{.+}} = alloca{{.+}},
+      // LAMBDA: {{.+}} = alloca{{.+}},
+      // LAMBDA: {{.+}} = alloca{{.+}},
+      // LAMBDA: {{.+}} = alloca{{.+}},
+
+      // addr alloca's
+      // LAMBDA-64: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA-32: [[G_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G1_REF:%.+]] = alloca double*,
+
+      // private alloca's (only for 32-bit)
+      // LAMBDA-32: [[G_PRIV:%.+]] = alloca double,
+
+      // transfer input parameters into addr alloca's
+      // LAMBDA-DAG: store {{.+}} [[G_IN]], {{.+}} [[G_ADDR]],
+      // LAMBDA-DAG: store {{.+}} [[G1_IN]], {{.+}} [[G1_ADDR]],
+      // LAMBDA-DAG: store {{.+}} [[SVAR_IN]], {{.+}} [[SVAR_ADDR]],
+      // LAMBDA-DAG: store {{.+}} [[SFVAR_IN]], {{.+}} [[SFVAR_ADDR]],
+
+      // prepare parameters for lambda
+      // g
+      // LAMBDA-64-DAG: [[G_CONV:%.+]] = bitcast {{.+}}* [[G_ADDR]] to
+      // LAMBDA-32-DAG: [[G_ADDR_REF:%.+]] = load {{.+}}*, {{.+}}** [[G_ADDR]]
+      // LAMBDA-32-DAG: [[G_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[G_ADDR_REF]],
+      // LAMBDA-32-DAG: store {{.+}} [[G_ADDR_VAL]], {{.+}}* [[G_PRIV]],
+
+      // g1
+      // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}}* [[G1_ADDR]] to
+      // LAMBDA-DAG: store {{.+}}* [[G1_CONV]], {{.+}}* [[G1_REF]],
+
+      // svar
+      // LAMBDA-64-DAG: [[SVAR_CONV:%.+]] = bitcast {{.+}}* [[SVAR_ADDR]] to
+
+      // sfvar
+      // LAMBDA-DAG: [[SFVAR_CONV:%.+]] = bitcast {{.+}}* [[SFVAR_ADDR]] to
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA-64: store double 1.0{{.+}}, double* [[G_CONV]],
+      // LAMBDA-32: store double 1.0{{.+}}, double* [[G_PRIV]],
+      // LAMBDA: [[G1_REF_REF:%.+]] = load {{.+}}*, {{.+}}** [[G1_REF]],
+      // LAMBDA: store {{.+}} 1.0{{.+}}, {{.+}}* [[G1_REF_REF]],
+      // LAMBDA-64: store {{.+}} 3, {{.+}}* [[SVAR_CONV]],
+      // LAMBDA-32: store {{.+}} 3, {{.+}}* [[SVAR_ADDR]],
+      // LAMBDA: store {{.+}} 4.0{{.+}}, {{.+}}* [[SFVAR_CONV]],
+
+      // pass params to inner lambda
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA-64: store double* [[G_CONV]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA-32: store double* [[G_PRIV]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[G1_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF_REF:%.+]] = load double*, double** [[G1_REF]],
+      // LAMBDA: store double* [[G1_REF_REF]], double** [[G1_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA-64: store i{{[0-9]+}}* [[SVAR_CONV]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA-32: store i{{[0-9]+}}* [[SVAR_ADDR]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_CONV]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: ret void
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd firstprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return tmain<int>();
+  #endif
+}
+
+// CHECK-LABEL: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_0:@.+]](
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN_0]](i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]* {{.+}} [[VAR_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]])
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, i{{[0-9]+}}*)* [[OMP_OUTLINED_0:@.+]] to void
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]* {{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]])
+
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// addr alloca's
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[TMP:%.+]] = alloca [[S_FLOAT_TY]]*,
+
+// skip loop alloca's
+// CHECK: [[OMP_IV:.omp.iv+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_LB:.omp.comb.lb+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_UB:.omp.comb.ub+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_ST:.omp.stride+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:.omp.is_last+]] = alloca i{{[0-9]+}},
+
+// private alloca's
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+
+// init addr alloca's with input values
+// CHECK-DAG: store {{.+}} [[T_VAR_IN]], {{.+}}* [[T_VAR_ADDR]],
+// CHECK-DAG: store {{.+}} [[VEC_IN]], {{.+}} [[VEC_ADDR]],
+// CHECK-DAG: store {{.+}} [[S_ARR_IN]], {{.+}} [[S_ARR_ADDR]],
+// CHECK-DAG: store {{.+}} [[VAR_IN]], {{.+}} [[VAR_ADDR]],
+// CHECK-DAG: store {{.+}} [[SVAR_IN]], {{.+}} [[SVAR_ADDR]],
+
+// init private alloca's with addr alloca's
+// t-var
+// CHECK-DAG: [[T_VAR_ADDR_REF:%.+]] = load {{.+}}*, {{.+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_ADDR_REF]],
+// CHECK-DAG: store {{.+}} [[T_VAR]], {{.+}} [[T_VAR_PRIV]],
+
+// vec
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[VEC_ADDR]],
+// CHECK-DAG: [[VEC_PRIV_BCAST:%.+]] = bitcast {{.+}} [[VEC_PRIV]] to
+// CHECK-DAG: [[VEC_ADDR_BCAST:%.+]] = bitcast {{.+}} [[VEC_ADDR_VAL]] to
+// CHECK-DAG: call void @llvm.memcpy{{.+}}({{.+}}* [[VEC_PRIV_BCAST]], {{.+}}* [[VEC_ADDR_BCAST]],
+
+// s_arr
+// CHECK-DAG: [[S_ARR_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[S_ARR_ADDR]],
+// CHECK-DAG: [[S_ARR_BGN:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BCAST:%.+]] = bitcast {{.+}}* [[S_ARR_ADDR_VAL]] to
+// CHECK-DAG: [[S_ARR_BGN_GEP:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_BGN]],
+// CHECK-DAG: [[S_ARR_EMPTY:%.+]] = icmp {{.+}} [[S_ARR_BGN]], [[S_ARR_BGN_GEP]]
+// CHECK-DAG: br {{.+}} [[S_ARR_EMPTY]], label %[[CPY_DONE:.+]], label %[[CPY_BODY:.+]]
+// CHECK-DAG: [[CPY_BODY]]:
+// CHECK-DAG: call void @llvm.memcpy{{.+}}(
+// CHECK-DAG: [[CPY_DONE]]:
+
+// var
+// CHECK-DAG: [[TMP_REF:%.+]] = load {{.+}}*, {{.+}}* [[TMP]],
+// CHECK-DAG: [[VAR_PRIV_BCAST:%.+]] = bitcast {{.+}}* [[VAR_PRIV]] to
+// CHECK-DAG: [[TMP_REF_BCAST:%.+]] = bitcast {{.+}}* [[TMP_REF]] to
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}({{.+}}* [[VAR_PRIV_BCAST]], {{.+}}* [[TMP_REF_BCAST]],
+// CHECK-DAG: store {{.+}}* [[VAR_PRIV]], {{.+}}** [[TMP_PRIV]],
+
+// svar
+// CHECK-DAG: [[SVAR_CONV_REF:%.+]] = load {{.+}}*, {{.+}}** [[SVAR_ADDR]],
+// CHECK-DAG: [[SVAR_CONV_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_CONV_REF]],
+// CHECK-DAG: store {{.+}} [[SVAR_CONV_VAL]], {{.+}}* [[SVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// pass private alloca's to fork
+// CHECK-DAG: [[T_VAR_PRIV_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_PRIV]],
+// not dag to distinguish with S_VAR_CAST
+// CHECK-64: [[T_VAR_CAST_CONV:%.+]] = bitcast {{.+}}* [[T_VAR_CAST:%.+]] to
+// CHECK-64-DAG: store {{.+}} [[T_VAR_PRIV_VAL]], {{.+}} [[T_VAR_CAST_CONV]],
+// CHECK-32: store {{.+}} [[T_VAR_PRIV_VAL]], {{.+}} [[T_VAR_CAST:%.+]],
+// CHECK-DAG: [[T_VAR_CAST_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[TMP_PRIV_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[SVAR_PRIV_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_PRIV]],
+// CHECK-64-DAG: [[SVAR_CAST_CONV:%.+]] = bitcast {{.+}}* [[SVAR_CAST:%.+]] to
+// CHECK-64-DAG: store {{.+}} [[SVAR_PRIV_VAL]], {{.+}}* [[SVAR_CAST_CONV]],
+// CHECK-32-DAG: store {{.+}} [[SVAR_PRIV_VAL]], {{.+}}* [[SVAR_CAST:%.+]],
+// CHECK-DAG: [[SVAR_CAST_VAL:%.+]] = load {{.+}}, {{.+}}* [[SVAR_CAST]],
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_0:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} [[T_VAR_CAST_VAL]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]], [[S_FLOAT_TY]]* [[TMP_PRIV_VAL]], i{{[0-9]+}} [[SVAR_CAST_VAL]])
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+// By OpenMP specifications, 'firstprivate' applies to both distribute and parallel for.
+// However, the support for 'firstprivate' of 'parallel' is only used when 'parallel'
+// is found alone. Therefore we only have one 'firstprivate' support for 'parallel for'
+// in combination
+// CHECK: define internal void [[OMP_PARFOR_OUTLINED_0]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]* {{.+}} [[VAR_IN:%.+]], i{{[0-9]+}} [[SVAR_IN:%.+]])
+
+// addr alloca's
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+
+// skip loop alloca's
+// CHECK: [[OMP_IV:.omp.iv+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_LB:.omp.lb+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_UB:.omp.ub+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_ST:.omp.stride+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:.omp.is_last+]] = alloca i{{[0-9]+}},
+
+// private alloca's
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+
+// init addr alloca's with input values
+// CHECK-DAG: store {{.+}} [[VEC_IN]], {{.+}} [[VEC_ADDR]],
+// CHECK-DAG: store {{.+}} [[T_VAR_IN]], {{.+}}* [[T_VAR_ADDR]],
+// CHECK-DAG: store {{.+}} [[S_ARR_IN]], {{.+}} [[S_ARR_ADDR]],
+// CHECK-DAG: store {{.+}} [[VAR_IN]], {{.+}} [[VAR_ADDR]],
+// CHECK-DAG: store {{.+}} [[SVAR_IN]], {{.+}} [[SVAR_ADDR]],
+
+// init private alloca's with addr alloca's
+// vec
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[VEC_ADDR]],
+// CHECK-DAG: [[VEC_PRIV_BCAST:%.+]] = bitcast {{.+}} [[VEC_PRIV]] to
+// CHECK-DAG: [[VEC_ADDR_BCAST:%.+]] = bitcast {{.+}} [[VEC_ADDR_VAL]] to
+// CHECK-DAG: call void @llvm.memcpy{{.+}}({{.+}}* [[VEC_PRIV_BCAST]], {{.+}}* [[VEC_ADDR_BCAST]],
+
+// s_arr
+// CHECK-DAG: [[S_ARR_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[S_ARR_ADDR]],
+// CHECK-DAG: [[S_ARR_BGN:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BCAST:%.+]] = bitcast {{.+}}* [[S_ARR_ADDR_VAL]] to
+// CHECK-DAG: [[S_ARR_BGN_GEP:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_BGN]],
+// CHECK-DAG: [[S_ARR_EMPTY:%.+]] = icmp {{.+}} [[S_ARR_BGN]], [[S_ARR_BGN_GEP]]
+// CHECK-DAG: br {{.+}} [[S_ARR_EMPTY]], label %[[CPY_DONE:.+]], label %[[CPY_BODY:.+]]
+// CHECK-DAG: [[CPY_BODY]]:
+// CHECK-DAG: call void @llvm.memcpy{{.+}}(
+// CHECK-DAG: [[CPY_DONE]]:
+
+// var
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load {{.+}}*, {{.+}}* [[VAR_ADDR]],
+// CHECK-DAG: [[VAR_PRIV_BCAST:%.+]] = bitcast {{.+}}* [[VAR_PRIV]] to
+// CHECK-DAG: [[VAR_ADDR_BCAST:%.+]] = bitcast {{.+}}* [[VAR_ADDR_REF]] to
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}({{.+}}* [[VAR_PRIV_BCAST]], {{.+}}* [[VAR_ADDR_BCAST]],
+// CHECK-DAG: store {{.+}}* [[VAR_PRIV]], {{.+}}** [[TMP_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+// template tmain with S_INT_TY
+// CHECK-LABEL: define{{.*}} i{{[0-9]+}} @{{.+}}tmain{{.+}}()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_0:@.+]](
+// CHECK: call {{.*}} [[S_INT_TY_DEF_DESTR:@.+]]([[S_INT_TY]]* [[TEST]])
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN_0]](i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]* {{.+}} [[VAR_IN:%.+]])
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[OMP_OUTLINED_0:@.+]] to void
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]* {{.+}} [[VAR_IN:%.+]])
+
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// addr alloca's
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip loop alloca's
+// CHECK: [[OMP_IV:.omp.iv+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_LB:.omp.comb.lb+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_UB:.omp.comb.ub+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_ST:.omp.stride+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:.omp.is_last+]] = alloca i{{[0-9]+}},
+
+// private alloca's
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+
+// init addr alloca's with input values
+// CHECK-DAG: store {{.+}} [[T_VAR_IN]], {{.+}}* [[T_VAR_ADDR]],
+// CHECK-DAG: store {{.+}} [[VEC_IN]], {{.+}} [[VEC_ADDR]],
+// CHECK-DAG: store {{.+}} [[S_ARR_IN]], {{.+}} [[S_ARR_ADDR]],
+// CHECK-DAG: store {{.+}} [[VAR_IN]], {{.+}} [[VAR_ADDR]],
+
+// init private alloca's with addr alloca's
+// t-var
+// CHECK-DAG: [[T_VAR_ADDR_REF:%.+]] = load {{.+}}*, {{.+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_ADDR_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_ADDR_REF]],
+// CHECK-DAG: store {{.+}} [[T_VAR_ADDR_VAL]], {{.+}} [[T_VAR_PRIV]],
+
+// vec
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[VEC_ADDR]],
+// CHECK-DAG: [[VEC_PRIV_BCAST:%.+]] = bitcast {{.+}} [[VEC_PRIV]] to
+// CHECK-DAG: [[VEC_ADDR_BCAST:%.+]] = bitcast {{.+}} [[VEC_ADDR_VAL]] to
+// CHECK-DAG: call void @llvm.memcpy{{.+}}({{.+}}* [[VEC_PRIV_BCAST]], {{.+}}* [[VEC_ADDR_BCAST]],
+
+// s_arr
+// CHECK-DAG: [[S_ARR_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[S_ARR_ADDR]],
+// CHECK-DAG: [[S_ARR_BGN:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BCAST:%.+]] = bitcast {{.+}}* [[S_ARR_ADDR_VAL]] to
+// CHECK-DAG: [[S_ARR_BGN_GEP:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_BGN]],
+// CHECK-DAG: [[S_ARR_EMPTY:%.+]] = icmp {{.+}} [[S_ARR_BGN]], [[S_ARR_BGN_GEP]]
+// CHECK-DAG: br {{.+}} [[S_ARR_EMPTY]], label %[[CPY_DONE:.+]], label %[[CPY_BODY:.+]]
+// CHECK-DAG: [[CPY_BODY]]:
+// CHECK-DAG: call void @llvm.memcpy{{.+}}(
+// CHECK-DAG: [[CPY_DONE]]:
+
+// var
+// CHECK-DAG: [[TMP_REF:%.+]] = load {{.+}}*, {{.+}}* [[TMP]],
+// CHECK-DAG: [[VAR_PRIV_BCAST:%.+]] = bitcast {{.+}}* [[VAR_PRIV]] to
+// CHECK-DAG: [[TMP_REF_BCAST:%.+]] = bitcast {{.+}}* [[TMP_REF]] to
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}({{.+}}* [[VAR_PRIV_BCAST]], {{.+}}* [[TMP_REF_BCAST]],
+// CHECK-DAG: store {{.+}}* [[VAR_PRIV]], {{.+}}** [[TMP_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// pass private alloca's to fork
+// CHECK-DAG: [[T_VAR_PRIV_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_PRIV]],
+// not dag to distinguish with S_VAR_CAST
+// CHECK-64: [[T_VAR_CAST_CONV:%.+]] = bitcast {{.+}}* [[T_VAR_CAST:%.+]] to
+// CHECK-64-DAG: store {{.+}} [[T_VAR_PRIV_VAL]], {{.+}} [[T_VAR_CAST_CONV]],
+// CHECK-32: store {{.+}} [[T_VAR_PRIV_VAL]], {{.+}} [[T_VAR_CAST:%.+]],
+// CHECK-DAG: [[T_VAR_CAST_VAL:%.+]] = load {{.+}}, {{.+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[TMP_PRIV_VAL:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV]],
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_0:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} [[T_VAR_CAST_VAL]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]], [[S_INT_TY]]* [[TMP_PRIV_VAL]])
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_INT_TY_DEF_DESTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_INT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_DESTR]]([[S_INT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+// By OpenMP specifications, 'firstprivate' applies to both distribute and parallel for.
+// However, the support for 'firstprivate' of 'parallel' is only used when 'parallel'
+// is found alone. Therefore we only have one 'firstprivate' support for 'parallel for'
+// in combination
+// CHECK: define internal void [[OMP_PARFOR_OUTLINED_0]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]* {{.+}} [[VEC_IN:%.+]], i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x [[S_INT_TY]]]* {{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]* {{.+}} [[VAR_IN:%.+]])
+
+// addr alloca's
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip loop alloca's
+// CHECK: [[OMP_IV:.omp.iv+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_LB:.omp.lb+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_UB:.omp.ub+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_ST:.omp.stride+]] = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:.omp.is_last+]] = alloca i{{[0-9]+}},
+
+// private alloca's
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+
+// init addr alloca's with input values
+// CHECK-DAG: store {{.+}} [[VEC_IN]], {{.+}} [[VEC_ADDR]],
+// CHECK-DAG: store {{.+}} [[T_VAR_IN]], {{.+}}* [[T_VAR_ADDR]],
+// CHECK-DAG: store {{.+}} [[S_ARR_IN]], {{.+}} [[S_ARR_ADDR]],
+// CHECK-DAG: store {{.+}} [[VAR_IN]], {{.+}} [[VAR_ADDR]],
+
+// init private alloca's with addr alloca's
+// vec
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[VEC_ADDR]],
+// CHECK-DAG: [[VEC_PRIV_BCAST:%.+]] = bitcast {{.+}} [[VEC_PRIV]] to
+// CHECK-DAG: [[VEC_ADDR_BCAST:%.+]] = bitcast {{.+}} [[VEC_ADDR_VAL]] to
+// CHECK-DAG: call void @llvm.memcpy{{.+}}({{.+}}* [[VEC_PRIV_BCAST]], {{.+}}* [[VEC_ADDR_BCAST]],
+
+// s_arr
+// CHECK-DAG: [[S_ARR_ADDR_VAL:%.+]] = load {{.+}}*, {{.+}}** [[S_ARR_ADDR]],
+// CHECK-DAG: [[S_ARR_BGN:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BCAST:%.+]] = bitcast {{.+}}* [[S_ARR_ADDR_VAL]] to
+// CHECK-DAG: [[S_ARR_BGN_GEP:%.+]] = getelementptr {{.+}}, {{.+}}* [[S_ARR_BGN]],
+// CHECK-DAG: [[S_ARR_EMPTY:%.+]] = icmp {{.+}} [[S_ARR_BGN]], [[S_ARR_BGN_GEP]]
+// CHECK-DAG: br {{.+}} [[S_ARR_EMPTY]], label %[[CPY_DONE:.+]], label %[[CPY_BODY:.+]]
+// CHECK-DAG: [[CPY_BODY]]:
+// CHECK-DAG: call void @llvm.memcpy{{.+}}(
+// CHECK-DAG: [[CPY_DONE]]:
+
+// var
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load {{.+}}*, {{.+}}* [[VAR_ADDR]],
+// CHECK-DAG: [[VAR_PRIV_BCAST:%.+]] = bitcast {{.+}}* [[VAR_PRIV]] to
+// CHECK-DAG: [[VAR_ADDR_BCAST:%.+]] = bitcast {{.+}}* [[VAR_ADDR_REF]] to
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}({{.+}}* [[VAR_PRIV_BCAST]], {{.+}}* [[VAR_ADDR_BCAST]],
+// CHECK-DAG: store {{.+}}* [[VAR_PRIV]], {{.+}}** [[TMP_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_INT_TY_DEF_DESTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_INT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_DESTR]]([[S_INT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/distribute_parallel_for_simd_firstprivate_messages.cpp b/test/OpenMP/distribute_parallel_for_simd_firstprivate_messages.cpp
index 07d30e48312b..646407643b25 100644
--- a/test/OpenMP/distribute_parallel_for_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_simd_firstprivate_messages.cpp
@@ -42,7 +42,7 @@ public:
 };
 class S5 {
   int a;
-  S5(const S5 &s5) : a(s5.a) {} // expected-note 4 {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {} // expected-note 2 {{implicitly declared private here}}
 
 public:
   S5() : a(0) {}
@@ -50,7 +50,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -150,9 +150,10 @@ int foomain(int argc, char **argv) {
 #pragma omp distribute parallel for simd firstprivate(i)
   for (int k = 0; k < argc; ++k)
     ++k;
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd lastprivate(g) firstprivate(g) // expected-error {{calling a private constructor of class 'S5'}}
+#pragma omp distribute parallel for simd lastprivate(g) firstprivate(g)
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp parallel private(i)
@@ -314,14 +315,16 @@ int main(int argc, char **argv) {
 #pragma omp distribute parallel for simd firstprivate(j)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd lastprivate(g) firstprivate(g) // expected-error {{calling a private constructor of class 'S5'}}
+#pragma omp distribute parallel for simd lastprivate(g) firstprivate(g)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd lastprivate(n) firstprivate(n) // OK
+#pragma omp distribute parallel for simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp parallel
diff --git a/test/OpenMP/distribute_parallel_for_simd_if_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_if_codegen.cpp
new file mode 100644
index 000000000000..b55038720134
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_if_codegen.cpp
@@ -0,0 +1,192 @@
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple %itanium_abi_triple -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+void fn1();
+void fn2();
+void fn3();
+void fn4();
+void fn5();
+void fn6();
+
+int Arg;
+
+// CHECK-LABEL: define {{.*}}void @{{.+}}gtid_test
+void gtid_test() {
+#pragma omp target
+#pragma omp teams
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+#pragma omp distribute parallel for simd
+  for(int i = 0 ; i < 100; i++) {}
+  // CHECK: define internal void [[OFFLOADING_FUN_0]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+  // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK:  call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_0:@.+]] to void
+  // CHECK: call void @__kmpc_for_static_fini(
+
+  // CHECK: define{{.+}} void [[OMP_OUTLINED_0]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: ret
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (parallel: false)
+  for(int i = 0 ; i < 100; i++) {
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+  // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @__kmpc_serialized_parallel(
+  // CHECK: call void [[OMP_OUTLINED_1:@.+]](
+  // CHECK: call void @__kmpc_end_serialized_parallel(
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: define{{.+}} void [[OMP_OUTLINED_1]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @{{.+}}gtid_test
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: ret
+    gtid_test();
+  }
+}
+
+
+template <typename T>
+int tmain(T Arg) {
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (true)
+  for(int i = 0 ; i < 100; i++) {
+    fn1();
+  }
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (false)
+  for(int i = 0 ; i < 100; i++) {
+    fn2();
+  }
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (parallel: Arg)
+  for(int i = 0 ; i < 100; i++) {
+    fn3();
+  }
+  return 0;
+}
+
+// CHECK-LABEL: define {{.*}}i{{[0-9]+}} @main()
+int main() {
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_2:@.+]](
+// CHECK: = call {{.*}}i{{.+}} @{{.+}}tmain
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (true)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_0]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_2:@.+]] to void
+    // CHECK: call void @__kmpc_for_static_fini(
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_2]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn4
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    fn4();
+  }
+
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (false)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_1]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void @__kmpc_serialized_parallel(
+    // CHECK: call void [[OMP_OUTLINED_3:@.+]](
+    // CHECK: call void @__kmpc_end_serialized_parallel(
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_3]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn5
+    // CHECK: call void @__kmpc_for_static_fini(
+    fn5();
+  }
+
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd if (Arg)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_2]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}}* [[OMP_TEAMS_OUTLINED_2:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_2]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_4:@.+]] to void
+    // CHECK: call void @__kmpc_serialized_parallel(
+    // CHECK: call void [[OMP_OUTLINED_4:@.+]](
+    // CHECK: call void @__kmpc_end_serialized_parallel(
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_4]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn6
+    // CHECK: call void @__kmpc_for_static_fini(
+    fn6();
+  }
+
+  return tmain(Arg);
+}
+
+// CHECK-LABEL: define {{.+}} @{{.+}}tmain
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, void {{.+}}* [[T_OUTLINE_FUN_1:@.+]] to void
+// CHECK: call void @__kmpc_for_static_fini(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_1]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn1
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @__kmpc_serialized_parallel(
+// CHECK: call void [[T_OUTLINE_FUN_2:@.+]](
+// CHECK: call {{.*}}void @__kmpc_end_serialized_parallel(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_2]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn2
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, void {{.+}}* [[T_OUTLINE_FUN_3:@.+]] to void
+// CHECK: call {{.*}}void @__kmpc_serialized_parallel(
+// call void [[T_OUTLINE_FUN_3:@.+]](
+// CHECK: call {{.*}}void @__kmpc_end_serialized_parallel(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_3]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn3
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..89ef33abd64d
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_lastprivate_codegen.cpp
@@ -0,0 +1,671 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute parallel for simd lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams
+#pragma omp distribute parallel for simd lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // LAMBDA: [[TMP_G1:%.+]] = alloca double*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+
+      // init addr alloca's
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN_REF]], double** [[TMP_G1]],
+      // LAMBDA: [[TMP_G1_VAL:%.+]] = load double*, double** [[TMP_G1]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: [[G1_PAR:%.+]] = load{{.+}}, {{.+}} [[TMP_G1_PRIVATE]],
+      // LAMBDA-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[G_PRIVATE]], {{.+}} [[G1_PAR]], {{.+}} [[SVAR_PRIVATE]], {{.+}} [[SFVAR_PRIVATE]])
+      // LAMBDA-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[G_PRIVATE]], {{.+}} [[G1_PAR]], {{.+}} [[SVAR_PRIVATE]], {{.+}} [[SFVAR_PRIVATE]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+
+      // linear counter
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_COUNTER_BLOCK:.+]], label %[[OMP_COUNTER_DONE:.+]]
+      // LAMBDA: [[OMP_COUNTER_BLOCK]]:
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* %
+      // LAMBDA: br label %[[OMP_COUNTER_DONE]]
+      // LAMBDA: [[OMP_COUNTER_DONE]]:
+
+      // lastprivate
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[TMP_G1_VAL]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // outlined function for 'parallel for'
+      // LAMBDA-64: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, {{.+}} [[G_IN:%.+]], {{.+}} [[G1_IN:%.+]], {{.+}} [[SVAR_IN:%.+]], {{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA-32: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, {{.+}} [[G_IN:%.+]], {{.+}} [[G1_IN:%.+]], {{.+}} [[SVAR_IN:%.+]], {{.+}} [[SFVAR_IN:%.+]])
+
+      // addr alloca's
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+
+      // private alloca's
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+
+            // init addr alloca's
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+
+      // loop body
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+
+      // lastprivate
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute parallel for simd lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[TMP:%.*]] = alloca [[S_FLOAT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: [[VAR_ADDR_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_ADDR_REF]], [[S_FLOAT_TY]]** [[TMP]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+
+// call constructor for s_arr
+// CHECK: [[S_ARR_BGN:%.+]] = getelementptr{{.+}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK: [[S_ARR_END:%.+]] = getelementptr {{.+}} [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BGN]],
+// CHECK: br label %[[S_ARR_CST_LOOP:.+]]
+// CHECK: [[S_ARR_CST_LOOP]]:
+// CHECK: [[S_ARR_CTOR:%.+]] = phi {{.+}}
+// CHECK: call void [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[S_ARR_CTOR]])
+// CHECK: [[S_ARR_NEXT:%.+]] = getelementptr {{.+}} [[S_ARR_CTOR]],
+// CHECK: [[S_ARR_DONE:%.+]] = icmp {{.+}} [[S_ARR_NEXT]], [[S_ARR_END]]
+// CHECK: br i1 [[S_ARR_DONE]], label %[[S_ARR_CST_END:.+]], label %[[S_ARR_CST_LOOP]]
+// CHECK: [[S_ARR_CST_END]]:
+// CHECK: [[TMP_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP]],
+// CHECK: call void [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_PRIV]], [[S_FLOAT_TY]]** [[TMP_PRIV]],
+
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: [[TMP_PRIV_VAL:%.+]] = load {{.+}}, {{.+}} [[TMP_PRIV]],
+// CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[VEC_PRIV]], {{.+}} [[T_VAR_PRIV]], {{.+}} [[S_ARR_PRIV]], {{.+}} [[TMP_PRIV_VAL]], {{.+}} [[S_VAR_PRIV]])
+// CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[VEC_PRIV]], {{.+}} [[T_VAR_PRIV]], {{.+}} [[S_ARR_PRIV]], {{.+}} [[TMP_PRIV_VAL]], {{.+}} [[S_VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// outlined function for 'parallel for'
+// CHECK-64: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, {{.+}} [[VEC_IN:%.+]], {{.+}} [[T_VAR_IN:%.+]], {{.+}} [[S_ARR_IN:%.+]], {{.+}} [[VAR_IN:%.+]], {{.+}} [[SVAR_IN:%.+]])
+// CHECK-32: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, {{.+}} [[VEC_IN:%.+]], {{.+}} [[T_VAR_IN:%.+]], {{.+}} [[S_ARR_IN:%.+]], {{.+}} [[VAR_IN:%.+]], {{.+}} [[SVAR_IN:%.+]])
+
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+
+// call constructor for s_arr
+// CHECK: [[S_ARR_BGN:%.+]] = getelementptr{{.+}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK: [[S_ARR_END:%.+]] = getelementptr {{.+}} [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BGN]],
+// CHECK: br label %[[S_ARR_CST_LOOP:.+]]
+// CHECK: [[S_ARR_CST_LOOP]]:
+// CHECK: [[S_ARR_CTOR:%.+]] = phi {{.+}}
+// CHECK: call void [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[S_ARR_CTOR]])
+// CHECK: [[S_ARR_NEXT:%.+]] = getelementptr {{.+}} [[S_ARR_CTOR]],
+// CHECK: [[S_ARR_DONE:%.+]] = icmp {{.+}} [[S_ARR_NEXT]], [[S_ARR_END]]
+// CHECK: br i1 [[S_ARR_DONE]], label %[[S_ARR_CST_END:.+]], label %[[S_ARR_CST_LOOP]]
+// CHECK: [[S_ARR_CST_END]]:
+// CHECK: [[VAR_ADDR_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: call void [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_PRIV]], [[S_FLOAT_TY]]** [[TMP_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+
+// loop body
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]],  [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN]], [[S_INT_TY]]** [[VAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[VAR_ADDR_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_ADDR_REF]], [[S_INT_TY]]** [[TMP]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK-DAG: [[TMP_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: [[TMP_PRIV_VAL:%.+]] = load {{.+}}, {{.+}} [[TMP_PRIV]],
+// CHECK-64: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[VEC_PRIV]], {{.+}} [[T_VAR_PRIV]], {{.+}} [[S_ARR_PRIV]], {{.+}} [[TMP_PRIV_VAL]])
+// CHECK-32: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to void ({{.+}})*), {{.+}}, {{.+}}, {{.+}} [[VEC_PRIV]], {{.+}} [[T_VAR_PRIV]], {{.+}} [[S_ARR_PRIV]], {{.+}} [[TMP_PRIV_VAL]])
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_REF]] to i8*
+// CHECK: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL_BCAST]],{{.+}})
+// CHECK: ret void
+
+// outlined function for 'parallel for'
+// CHECK-64: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, {{.+}} [[VEC_IN:%.+]], {{.+}} [[T_VAR_IN:%.+]], {{.+}} [[S_ARR_IN:%.+]], {{.+}} [[VAR_IN:%.+]])
+// CHECK-32: define{{.+}} void [[OMP_PARFOR_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, {{.+}} [[VEC_IN:%.+]], {{.+}} [[T_VAR_IN:%.+]], {{.+}} [[S_ARR_IN:%.+]], {{.+}} [[VAR_IN:%.+]])
+
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN]], [[S_INT_TY]]** [[VAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+
+// call constructor for s_arr
+// CHECK: [[S_ARR_BGN:%.+]] = getelementptr{{.+}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK: [[S_ARR_END:%.+]] = getelementptr {{.+}} [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BGN]],
+// CHECK: br label %[[S_ARR_CST_LOOP:.+]]
+// CHECK: [[S_ARR_CST_LOOP]]:
+// CHECK: [[S_ARR_CTOR:%.+]] = phi {{.+}}
+// CHECK: call void [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[S_ARR_CTOR]])
+// CHECK: [[S_ARR_NEXT:%.+]] = getelementptr {{.+}} [[S_ARR_CTOR]],
+// CHECK: [[S_ARR_DONE:%.+]] = icmp {{.+}} [[S_ARR_NEXT]], [[S_ARR_END]]
+// CHECK: br i1 [[S_ARR_DONE]], label %[[S_ARR_CST_END:.+]], label %[[S_ARR_CST_LOOP]]
+// CHECK: [[S_ARR_CST_END]]:
+// CHECK: [[VAR_ADDR_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR]],
+// CHECK: call void [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+// CHECK: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
+
diff --git a/test/OpenMP/distribute_parallel_for_simd_lastprivate_messages.cpp b/test/OpenMP/distribute_parallel_for_simd_lastprivate_messages.cpp
index 109fde0577c4..73450d4f1181 100644
--- a/test/OpenMP/distribute_parallel_for_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_simd_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -313,14 +313,16 @@ int main(int argc, char **argv) {
 #pragma omp distribute parallel for simd lastprivate(j)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{firstprivate variable cannot be lastprivate}} expected-note@+3 {{defined as firstprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp distribute parallel for simd firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd lastprivate(n) firstprivate(n) // OK
+#pragma omp distribute parallel for simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i)
     foo();
   static int si;
diff --git a/test/OpenMP/distribute_parallel_for_simd_linear_messages.cpp b/test/OpenMP/distribute_parallel_for_simd_linear_messages.cpp
index 632ef066c0fa..66fd94b9e0f2 100644
--- a/test/OpenMP/distribute_parallel_for_simd_linear_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_simd_linear_messages.cpp
@@ -18,41 +18,49 @@ void test_linear_colons()
 {
   int B = 0;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(B:bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(B:B::bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(X::x : ::z)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(B,::z, X::x)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(::z)
@@ -63,6 +71,7 @@ void test_linear_colons()
 #pragma omp distribute parallel for simd linear(B::bfoo()) // expected-error {{expected variable name}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(B::ib,B:C1+C2)
@@ -148,11 +157,13 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute parallel for simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
@@ -163,6 +174,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute parallel for simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear (argc : 5)
@@ -183,6 +195,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute parallel for simd linear (argv[1]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(e, g)
@@ -193,6 +206,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute parallel for simd linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear(i)
@@ -202,28 +216,13 @@ template<class I, class C> int foomain(I argc, C **argv) {
   {
     int v = 0;
     int i;
+    int k;
     #pragma omp target
     #pragma omp teams
-    #pragma omp distribute parallel for simd linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
+    #pragma omp distribute parallel for simd linear(k:i)
+    for (k = 0; k < argc; ++k) { i = k; v += i; }
   }
 
-#pragma omp target
-#pragma omp teams
-#pragma omp parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute parallel for simd linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute parallel for simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
   return 0;
 }
 
@@ -262,11 +261,13 @@ int main(int argc, char **argv) {
 #pragma omp distribute parallel for simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
@@ -277,6 +278,7 @@ int main(int argc, char **argv) {
 #pragma omp distribute parallel for simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd linear (argc)
@@ -314,24 +316,14 @@ int main(int argc, char **argv) {
     #pragma omp target
     #pragma omp teams
     #pragma omp distribute parallel for simd linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
+      for (i = 0; i < argc; ++i) ++i;
 
     #pragma omp target
     #pragma omp teams
     #pragma omp distribute parallel for simd linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
+      for (i = 0; i < argc; ++i) { ++i; i += 4; }
   }
 
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute parallel for simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
   foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
   return 0;
 }
diff --git a/test/OpenMP/distribute_parallel_for_simd_loop_messages.cpp b/test/OpenMP/distribute_parallel_for_simd_loop_messages.cpp
index 1fedb3c02d79..d137d47ca2a4 100644
--- a/test/OpenMP/distribute_parallel_for_simd_loop_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_simd_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -790,9 +790,10 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute parallel for simd lastprivate(s) firstprivate(s)
+#pragma omp distribute parallel for simd lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/distribute_parallel_for_simd_misc_messages.c b/test/OpenMP/distribute_parallel_for_simd_misc_messages.c
index 01c079e3dac1..4d071621fcb8 100644
--- a/test/OpenMP/distribute_parallel_for_simd_misc_messages.c
+++ b/test/OpenMP/distribute_parallel_for_simd_misc_messages.c
@@ -68,7 +68,7 @@ void test_non_identifiers() {
 #pragma omp target
 #pragma omp teams
 // expected-warning@+1 {{extra tokens at the end of '#pragma omp distribute parallel for simd' are ignored}}
-#pragma omp distribute parallel for simd linear(x);
+#pragma omp distribute parallel for simd firstprivate(x);
   for (i = 0; i < 16; ++i)
     ;
 
@@ -548,89 +548,6 @@ void test_linear() {
 #pragma omp distribute parallel for simd linear(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
-
-  int x, y;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected expression}}
-#pragma omp distribute parallel for simd linear(x :)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected expression}} expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
-#pragma omp distribute parallel for simd linear(x :, )
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute parallel for simd linear(x : 1)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute parallel for simd linear(x : 2 * 2)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
-#pragma omp distribute parallel for simd linear(x : 1, y)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
-#pragma omp distribute parallel for simd linear(x : 1, y, z : 1)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as linear}}
-// expected-error@+1 {{linear variable cannot be linear}}
-#pragma omp distribute parallel for simd linear(x) linear(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as private}}
-// expected-error@+1 {{private variable cannot be linear}}
-#pragma omp distribute parallel for simd private(x) linear(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as linear}}
-// expected-error@+1 {{linear variable cannot be private}}
-#pragma omp distribute parallel for simd linear(x) private(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-warning@+1 {{zero linear step (x and other variables in clause should probably be const)}}
-#pragma omp distribute parallel for simd linear(x, y : 0)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as linear}}
-// expected-error@+1 {{linear variable cannot be lastprivate}}
-#pragma omp distribute parallel for simd linear(x) lastprivate(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as lastprivate}}
-// expected-error@+1 {{lastprivate variable cannot be linear}}
-#pragma omp distribute parallel for simd lastprivate(x) linear(x)
-  for (i = 0; i < 16; ++i)
-    ;
 }
 
 void test_aligned() {
@@ -934,16 +851,19 @@ void test_firstprivate() {
     ;
 
   int x, y, z;
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd lastprivate(x) firstprivate(x)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+3 2 {{lastprivate variable cannot be firstprivate}} expected-note@+3 2 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd lastprivate(x, y) firstprivate(x, y)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+3 3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute parallel for simd lastprivate(x, y, z) firstprivate(x, y, z)
diff --git a/test/OpenMP/distribute_parallel_for_simd_num_threads_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_num_threads_codegen.cpp
new file mode 100644
index 000000000000..5804a2a72b71
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_num_threads_codegen.cpp
@@ -0,0 +1,121 @@
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple %itanium_abi_triple -fexceptions -fcxx-exceptions -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -fexceptions -fcxx-exceptions -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+typedef __INTPTR_TYPE__ intptr_t;
+
+// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[S_TY:%.+]] = type { [[INTPTR_T_TY:i[0-9]+]], [[INTPTR_T_TY]], [[INTPTR_T_TY]] }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_2:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+void foo();
+
+struct S {
+  intptr_t a, b, c;
+  S(intptr_t a) : a(a) {}
+  operator char() { return a; }
+  ~S() {}
+};
+
+template <typename T, int C>
+int tmain() {
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd num_threads(C)
+  for (int i = 0; i < 100; i++)
+    foo();
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd num_threads(T(23))
+  for (int i = 0; i < 100; i++)
+    foo();
+  return 0;
+}
+
+int main() {
+  S s(0);
+  char a = s;
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+// CHECK: invoke{{.+}} [[TMAIN_5:@.+]]()
+// CHECK: invoke{{.+}} [[TMAIN_1:@.+]]()
+#pragma omp target
+#pragma omp teams
+  // CHECK: define internal void [[OFFLOADING_FUN_0]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+#pragma omp distribute parallel for simd num_threads(2)
+  for (int i = 0; i < 100; i++) {
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+    // CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 2)
+    // CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+    foo();
+  }
+#pragma omp target
+#pragma omp teams
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+#pragma omp distribute parallel for simd num_threads(a)
+  for (int i = 0; i < 100; i++) {
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+    // CHECK-DAG: [[A_ADDR:%.+]] = alloca i8*,
+    // CHECK-DAG: [[A_REF:%.+]] = load i8*, i8** [[A_ADDR]],
+    // CHECK-DAG: [[A_VAL:%.+]] = load i8, i8* [[A_REF]],
+    // CHECK-DAG: [[A_EXT:%.+]] = sext i8 [[A_VAL]] to {{.+}}
+    // CHECK: call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 [[A_EXT]])
+    // CHECK: call {{.*}}void {{.*}} @__kmpc_fork_call(
+    foo();
+  }
+  return a + tmain<char, 5>() + tmain<S, 1>();
+}
+
+// tmain 5
+// CHECK-DAG: define {{.*}}i{{[0-9]+}} [[TMAIN_5]]()
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_1:@.+]](
+
+// tmain 1
+// CHECK-DAG: define {{.*}}i{{[0-9]+}} [[TMAIN_1]]()
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_2:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_3:@.+]](
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_0]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_0]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 5)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_1]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 23)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_2]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_2:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_2]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 1)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_3]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_3]](
+// CHECK-DAG:   [[CALL_RES:%.+]] = invoke{{.*}} i8 [[S_TY_CHAR_OP:@.+]]([[S_TY]]* {{.+}})
+// CHECK-DAG:   [[CALL_RES_SEXT:%.+]] = sext i8 [[CALL_RES]] to {{.+}}
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 [[CALL_RES_SEXT]])
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+#endif
diff --git a/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp
new file mode 100644
index 000000000000..8ffe521bfcbc
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_private_codegen.cpp
@@ -0,0 +1,297 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd private(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]]()
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams
+    #pragma omp distribute parallel for simd private(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}})
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double,
+      // LAMBDA: [[TMP_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float,
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR]], double** [[TMP_PRIVATE_ADDR]],
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED:@.+]] to {{.+}},
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: ret void
+
+      // LAMBDA: define{{.+}} void [[OMP_PARFOR_OUTLINED]](
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double,
+      // LAMBDA: [[TMP_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float,
+
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR]], double** [[TMP_PRIVATE_ADDR]],
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE_ADDR]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE_ADDR]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: ret void
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute parallel for simd private(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_0:@.+]](
+
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN_0]]()
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 0, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*)* [[OMP_OUTLINED_0:@.+]] to void
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED_0]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-NOT: alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-NOT: alloca [[S_FLOAT_TY]],
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// this is the ctor loop
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_0:@.+]] to {{.+}},
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+// By OpenMP specifications, private applies to both distribute and parallel for.
+// However, the support for 'private' of 'parallel' is only used when 'parallel'
+// is found alone. Therefore we only have one 'private' support for 'parallel for'
+// in combination
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_0]](
+// CHECK: [[T_VAR_PRIV:%t_var+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%vec+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%s_arr+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-NOT: alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%var+]] = alloca [[S_FLOAT_TY]],
+// CHECK-NOT: alloca [[S_FLOAT_TY]],
+// CHECK: [[S_VAR_PRIV:%svar+]] = alloca i{{[0-9]+}},
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// this is the ctor loop
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_DESTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+// template tmain with S_INT_TY
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](
+// CHECK: call {{.*}} [[S_INT_TY_DEF_DESTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: ret
+
+// CHECK: ret
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]]()
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 0, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*)* [[OMP_OUTLINED_1:@.+]] to void
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED_1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK-NOT: alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK-NOT: alloca [[S_INT_TY]],
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = phi [[S_INT_TY]]*
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[S_ARR_PRIV_ITEM]])
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call{{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}}[[OMP_PARFOR_OUTLINED_1:@.+]] to {{.+}},
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.+}} void [[OMP_PARFOR_OUTLINED_1]](
+// CHECK: [[T_VAR_PRIV:%t_var+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%vec+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%s_arr+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK-NOT: alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%var+]] = alloca [[S_INT_TY]],
+// CHECK-NOT: alloca [[S_INT_TY]],
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// this is the ctor loop
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = phi [[S_INT_TY]]*
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[S_ARR_PRIV_ITEM]])
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// call destructors: var..
+// CHECK-DAG: call {{.+}} [[S_INT_TY_DEF_DESTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+
+// ..and s_arr
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_EL_PAST:%.+]] = phi [[S_INT_TY]]*
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = getelementptr {{.+}}, {{.+}} [[S_ARR_EL_PAST]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_DESTR]]([[S_INT_TY]]* [[S_ARR_PRIV_ITEM]])
+
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/distribute_parallel_for_simd_proc_bind_codegen.cpp b/test/OpenMP/distribute_parallel_for_simd_proc_bind_codegen.cpp
new file mode 100644
index 000000000000..7f0b631aee7f
--- /dev/null
+++ b/test/OpenMP/distribute_parallel_for_simd_proc_bind_codegen.cpp
@@ -0,0 +1,93 @@
+// add -fopenmp-targets
+
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+typedef __INTPTR_TYPE__ intptr_t;
+
+// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_2:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+void foo();
+
+struct S {
+  intptr_t a, b, c;
+  S(intptr_t a) : a(a) {}
+  operator char() { return a; }
+  ~S() {}
+};
+
+template <typename T>
+T tmain() {
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd proc_bind(master)
+  for(int i = 0; i < 1000; i++) {}
+  return T();
+}
+
+int main() {
+  // CHECK-LABEL: @main
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd proc_bind(spread)
+  for(int i = 0; i < 1000; i++) {}
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute parallel for simd proc_bind(close)
+  for(int i = 0; i < 1000; i++) {}
+  return tmain<int>();
+}
+
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL1:@.+]]()
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL2:@.+]]()
+// CHECK: [[CALL_RET:%.+]] = call{{.+}} i32 [[TMAIN:@.+]]()
+// CHECK: ret i32 [[CALL_RET]]
+
+// CHECK: define{{.+}} void [[OFFL1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_1]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 4)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: define{{.+}} [[OFFL2]]()
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_1]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 3)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: define{{.+}} [[TMAIN]]()
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL3:@.+]]()
+
+// CHECK: define{{.+}} [[OFFL3]]()
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_3]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 2)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/distribute_parallel_for_simd_reduction_messages.cpp b/test/OpenMP/distribute_parallel_for_simd_reduction_messages.cpp
index 6ad41d72bf78..94e4541b901b 100644
--- a/test/OpenMP/distribute_parallel_for_simd_reduction_messages.cpp
+++ b/test/OpenMP/distribute_parallel_for_simd_reduction_messages.cpp
@@ -27,9 +27,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/distribute_private_messages.cpp b/test/OpenMP/distribute_private_messages.cpp
index 518b64d9e9ca..0aa1e4796975 100644
--- a/test/OpenMP/distribute_private_messages.cpp
+++ b/test/OpenMP/distribute_private_messages.cpp
@@ -87,7 +87,7 @@ int main(int argc, char **argv) {
   #pragma omp teams
   {
     int i; // expected-note {{predetermined as private}}
-    #pragma omp distribute firstprivate(i), private(i) // expected-error {{private variable in '#pragma omp teams' cannot be firstprivate in '#pragma omp distribute'}}
+    #pragma omp distribute firstprivate(i) // expected-error {{firstprivate variable must be shared}}
     for (int k = 0; k < argc; ++k) ++k;
   }
   #pragma omp target
diff --git a/test/OpenMP/distribute_simd_ast_print.cpp b/test/OpenMP/distribute_simd_ast_print.cpp
index 6d2e916292f5..3246ccb46ae3 100644
--- a/test/OpenMP/distribute_simd_ast_print.cpp
+++ b/test/OpenMP/distribute_simd_ast_print.cpp
@@ -84,14 +84,14 @@ T tmain(T argc) {
 
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) reduction(+ : h) dist_schedule(static,N)
+#pragma omp distribute simd private(argc, b), firstprivate(c, d), lastprivate(f) collapse(N) reduction(+ : h) dist_schedule(static,N)
   for (int i = 0; i < 2; ++i)
     for (int j = 0; j < 2; ++j)
       for (int k = 0; k < 10; ++k)
         for (int m = 0; m < 10; ++m)
           for (int n = 0; n < 10; ++n)
             a++;
-// CHECK: #pragma omp distribute simd private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) reduction(+: h) dist_schedule(static, N)
+// CHECK: #pragma omp distribute simd private(argc,b) firstprivate(c,d) lastprivate(f) collapse(N) reduction(+: h) dist_schedule(static, N)
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
 // CHECK-NEXT: for (int j = 0; j < 2; ++j)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
@@ -129,14 +129,15 @@ int main(int argc, char **argv) {
 // CHECK-NEXT: for (int j = 0; j < 10; ++j)
 // CHECK-NEXT: a++;
 
+  int i;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd aligned(x:8) linear(h:2) safelen(8) simdlen(8)
-  for (int i = 0; i < 100; i++)
+#pragma omp distribute simd aligned(x:8) linear(i:2) safelen(8) simdlen(8)
+  for (i = 0; i < 100; i++)
     for (int j = 0; j < 200; j++)
       a += h + x[j];
-// CHECK: #pragma omp distribute simd aligned(x: 8) linear(h: 2) safelen(8) simdlen(8)
-// CHECK-NEXT: for (int i = 0; i < 100; i++)
+// CHECK: #pragma omp distribute simd aligned(x: 8) linear(i: 2) safelen(8) simdlen(8)
+// CHECK-NEXT: for (i = 0; i < 100; i++)
 // CHECK-NEXT: for (int j = 0; j < 200; j++)
 // CHECK-NEXT: a += h + x[j];
 
diff --git a/test/OpenMP/distribute_simd_codegen.cpp b/test/OpenMP/distribute_simd_codegen.cpp
new file mode 100644
index 000000000000..d8f18c3e431b
--- /dev/null
+++ b/test/OpenMP/distribute_simd_codegen.cpp
@@ -0,0 +1,269 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64  --check-prefix HCHECK
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32  --check-prefix HCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32 --check-prefix HCHECK
+
+// Test target codegen - host bc file has to be created first. (no significant differences with host version of target region)
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_0:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+// CHECK-DAG: [[DEF_LOC_DISTRIBUTE_0:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2050, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+// CHECK-LABEL: define {{.*void}} @{{.*}}without_schedule_clause{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}})
+void without_schedule_clause(float *a, float *b, float *c, float *d) {
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute simd simdlen(8) aligned(a)
+  for (int i = 33; i < 32000000; i += 7) {
+    a[i] = b[i] * c[i] * d[i];
+  }
+}
+
+// CHECK: define {{.*}}void @{{.+}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]])
+// CHECK:  [[TID_ADDR:%.+]] = alloca i32*
+// CHECK:  [[IV:%.+iv]] = alloca i32
+// CHECK:  [[LB:%.+lb]] = alloca i32
+// CHECK:  [[UB:%.+ub]] = alloca i32
+// CHECK:  [[ST:%.+stride]] = alloca i32
+// CHECK:  [[LAST:%.+last]] = alloca i32
+// CHECK-DAG:  store i32* [[GBL_TIDP]], i32** [[TID_ADDR]]
+// CHECK-DAG:  call void @llvm.assume(
+// CHECK-DAG:  store i32 0, i32* [[LB]]
+// CHECK-DAG:  store i32 4571423, i32* [[UB]]
+// CHECK-DAG:  store i32 1, i32* [[ST]]
+// CHECK-DAG:  store i32 0, i32* [[LAST]]
+// CHECK-DAG:  [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]]
+// CHECK-DAG:  [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]]
+// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1)
+// CHECK-DAG:  [[UBV0:%.+]] = load i32, i32* [[UB]]
+// CHECK-DAG:  [[USWITCH:%.+]] = icmp sgt i32 [[UBV0]], 4571423
+// CHECK:  br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]]
+// CHECK-DAG:  [[BBCT]]:
+// CHECK-DAG:  br label %[[BBCE:.+]]
+// CHECK-DAG:  [[BBCF]]:
+// CHECK-DAG:  [[UBV1:%.+]] = load i32, i32* [[UB]]
+// CHECK-DAG:  br label %[[BBCE]]
+// CHECK:  [[BBCE]]:
+// CHECK:  [[SELUB:%.+]] = phi i32 [ 4571423, %[[BBCT]] ], [ [[UBV1]], %[[BBCF]] ]
+// CHECK:  store i32 [[SELUB]], i32* [[UB]]
+// CHECK:  [[LBV0:%.+]] = load i32, i32* [[LB]]
+// CHECK:  store i32 [[LBV0]], i32* [[IV]]
+// CHECK:  br label %[[BBINNFOR:.+]]
+// CHECK:  [[BBINNFOR]]:
+// CHECK:  [[IVVAL0:%.+]] = load i32, i32* [[IV]]
+// CHECK:  [[UBV2:%.+]] = load i32, i32* [[UB]]
+// CHECK:  [[IVLEUB:%.+]] = icmp sle i32 [[IVVAL0]], [[UBV2]]
+// CHECK:  br i1 [[IVLEUB]], label %[[BBINNBODY:.+]], label %[[BBINNEND:.+]]
+// CHECK:  [[BBINNBODY]]:
+// CHECK:  {{.+}} = load i32, i32* [[IV]]
+// ... loop body ...
+// CHECK:  br label %[[BBBODYCONT:.+]]
+// CHECK:  [[BBBODYCONT]]:
+// CHECK:  br label %[[BBINNINC:.+]]
+// CHECK:  [[BBINNINC]]:
+// CHECK:  [[IVVAL1:%.+]] = load i32, i32* [[IV]]
+// CHECK:  [[IVINC:%.+]] = add nsw i32 [[IVVAL1]], 1
+// CHECK:  store i32 [[IVINC]], i32* [[IV]]
+// CHECK:  br label %[[BBINNFOR]]
+// CHECK:  [[BBINNEND]]:
+// CHECK:  br label %[[LPEXIT:.+]]
+// CHECK:  [[LPEXIT]]:
+// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]])
+// CHECK:  ret void
+
+
+// CHECK-LABEL: define {{.*void}} @{{.*}}static_not_chunked{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}})
+void static_not_chunked(float *a, float *b, float *c, float *d) {
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute simd dist_schedule(static) safelen(32)
+  for (int i = 32000000; i > 33; i += -7) {
+        a[i] = b[i] * c[i] * d[i];
+  }
+}
+
+// CHECK: define {{.*}}void @.omp_outlined.{{.*}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]])
+// CHECK:  [[TID_ADDR:%.+]] = alloca i32*
+// CHECK:  [[IV:%.+iv]] = alloca i32
+// CHECK:  [[LB:%.+lb]] = alloca i32
+// CHECK:  [[UB:%.+ub]] = alloca i32
+// CHECK:  [[ST:%.+stride]] = alloca i32
+// CHECK:  [[LAST:%.+last]] = alloca i32
+// CHECK-DAG:  store i32* [[GBL_TIDP]], i32** [[TID_ADDR]]
+// CHECK-DAG:  store i32 0, i32* [[LB]]
+// CHECK-DAG:  store i32 4571423, i32* [[UB]]
+// CHECK-DAG:  store i32 1, i32* [[ST]]
+// CHECK-DAG:  store i32 0, i32* [[LAST]]
+// CHECK-DAG:  [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]]
+// CHECK-DAG:  [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]]
+// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 92, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 1)
+// CHECK-DAG:  [[UBV0:%.+]] = load i32, i32* [[UB]]
+// CHECK-DAG:  [[USWITCH:%.+]] = icmp sgt i32 [[UBV0]], 4571423
+// CHECK:  br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]]
+// CHECK-DAG:  [[BBCT]]:
+// CHECK-DAG:  br label %[[BBCE:.+]]
+// CHECK-DAG:  [[BBCF]]:
+// CHECK-DAG:  [[UBV1:%.+]] = load i32, i32* [[UB]]
+// CHECK-DAG:  br label %[[BBCE]]
+// CHECK:  [[BBCE]]:
+// CHECK:  [[SELUB:%.+]] = phi i32 [ 4571423, %[[BBCT]] ], [ [[UBV1]], %[[BBCF]] ]
+// CHECK:  store i32 [[SELUB]], i32* [[UB]]
+// CHECK:  [[LBV0:%.+]] = load i32, i32* [[LB]]
+// CHECK:  store i32 [[LBV0]], i32* [[IV]]
+// CHECK:  br label %[[BBINNFOR:.+]]
+// CHECK:  [[BBINNFOR]]:
+// CHECK:  [[IVVAL0:%.+]] = load i32, i32* [[IV]]
+// CHECK:  [[UBV2:%.+]] = load i32, i32* [[UB]]
+// CHECK:  [[IVLEUB:%.+]] = icmp sle i32 [[IVVAL0]], [[UBV2]]
+// CHECK:  br i1 [[IVLEUB]], label %[[BBINNBODY:.+]], label %[[BBINNEND:.+]]
+// CHECK:  [[BBINNBODY]]:
+// CHECK:  {{.+}} = load i32, i32* [[IV]]
+// ... loop body ...
+// CHECK:  br label %[[BBBODYCONT:.+]]
+// CHECK:  [[BBBODYCONT]]:
+// CHECK:  br label %[[BBINNINC:.+]]
+// CHECK:  [[BBINNINC]]:
+// CHECK:  [[IVVAL1:%.+]] = load i32, i32* [[IV]]
+// CHECK:  [[IVINC:%.+]] = add nsw i32 [[IVVAL1]], 1
+// CHECK:  store i32 [[IVINC]], i32* [[IV]]
+// CHECK:  br label %[[BBINNFOR]]
+// CHECK:  [[BBINNEND]]:
+// CHECK:  br label %[[LPEXIT:.+]]
+// CHECK:  [[LPEXIT]]:
+// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]])
+// CHECK:  ret void
+
+
+// CHECK-LABEL: define {{.*void}} @{{.*}}static_chunked{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}})
+void static_chunked(float *a, float *b, float *c, float *d) {
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd dist_schedule(static, 5)
+  for (unsigned i = 131071; i <= 2147483647; i += 127) {
+    a[i] = b[i] * c[i] * d[i];
+  }
+}
+
+// CHECK: define {{.*}}void @.omp_outlined.{{.*}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], float** dereferenceable({{[0-9]+}}) [[APTR:%.+]], float** dereferenceable({{[0-9]+}}) [[BPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[CPTR:%.+]], float** dereferenceable({{[0-9]+}}) [[DPTR:%.+]])
+// CHECK:  [[TID_ADDR:%.+]] = alloca i32*
+// CHECK:  [[IV:%.+iv]] = alloca i32
+// CHECK:  [[LB:%.+lb]] = alloca i32
+// CHECK:  [[UB:%.+ub]] = alloca i32
+// CHECK:  [[ST:%.+stride]] = alloca i32
+// CHECK:  [[LAST:%.+last]] = alloca i32
+// CHECK-DAG:  store i32* [[GBL_TIDP]], i32** [[TID_ADDR]]
+// CHECK-DAG:  store i32 0, i32* [[LB]]
+// CHECK-DAG:  store i32 16908288, i32* [[UB]]
+// CHECK-DAG:  store i32 1, i32* [[ST]]
+// CHECK-DAG:  store i32 0, i32* [[LAST]]
+// CHECK-DAG:  [[GBL_TID:%.+]] = load i32*, i32** [[TID_ADDR]]
+// CHECK-DAG:  [[GBL_TIDV:%.+]] = load i32, i32* [[GBL_TID]]
+// CHECK:  call void @__kmpc_for_static_init_{{.+}}(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]], i32 91, i32* %.omp.is_last, i32* %.omp.lb, i32* %.omp.ub, i32* %.omp.stride, i32 1, i32 5)
+// CHECK-DAG:  [[UBV0:%.+]] = load i32, i32* [[UB]]
+// CHECK-DAG:  [[USWITCH:%.+]] = icmp ugt i32 [[UBV0]], 16908288
+// CHECK:  br i1 [[USWITCH]], label %[[BBCT:.+]], label %[[BBCF:.+]]
+// CHECK-DAG:  [[BBCT]]:
+// CHECK-DAG:  br label %[[BBCE:.+]]
+// CHECK-DAG:  [[BBCF]]:
+// CHECK-DAG:  [[UBV1:%.+]] = load i32, i32* [[UB]]
+// CHECK-DAG:  br label %[[BBCE]]
+// CHECK:  [[BBCE]]:
+// CHECK:  [[SELUB:%.+]] = phi i32 [ 16908288, %[[BBCT]] ], [ [[UBV1]], %[[BBCF]] ]
+// CHECK:  store i32 [[SELUB]], i32* [[UB]]
+// CHECK:  [[LBV0:%.+]] = load i32, i32* [[LB]]
+// CHECK:  store i32 [[LBV0]], i32* [[IV]]
+// CHECK:  br label %[[BBINNFOR:.+]]
+// CHECK:  [[BBINNFOR]]:
+// CHECK:  [[IVVAL0:%.+]] = load i32, i32* [[IV]]
+// CHECK:  [[UBV2:%.+]] = load i32, i32* [[UB]]
+// CHECK:  [[IVLEUB:%.+]] = icmp ule i32 [[IVVAL0]], [[UBV2]]
+// CHECK:  br i1 [[IVLEUB]], label %[[BBINNBODY:.+]], label %[[BBINNEND:.+]]
+// CHECK:  [[BBINNBODY]]:
+// CHECK:  {{.+}} = load i32, i32* [[IV]]
+// ... loop body ...
+// CHECK:  br label %[[BBBODYCONT:.+]]
+// CHECK:  [[BBBODYCONT]]:
+// CHECK:  br label %[[BBINNINC:.+]]
+// CHECK:  [[BBINNINC]]:
+// CHECK:  [[IVVAL1:%.+]] = load i32, i32* [[IV]]
+// CHECK:  [[IVINC:%.+]] = add i32 [[IVVAL1]], 1
+// CHECK:  store i32 [[IVINC]], i32* [[IV]]
+// CHECK:  br label %[[BBINNFOR]]
+// CHECK:  [[BBINNEND]]:
+// CHECK:  br label %[[LPEXIT:.+]]
+// CHECK:  [[LPEXIT]]:
+// CHECK:  call void @__kmpc_for_static_fini(%ident_t* [[DEF_LOC_DISTRIBUTE_0]], i32 [[GBL_TIDV]])
+// CHECK:  ret void
+
+// CHECK-LABEL: test_precond
+void test_precond() {
+  char a = 0; char i;
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute simd linear(i)
+  for(i = a; i < 10; ++i);
+}
+
+// a is passed as a parameter to the outlined functions
+// CHECK:  define {{.*}}void @.omp_outlined.{{.*}}(i32* noalias [[GBL_TIDP:%.+]], i32* noalias [[BND_TID:%.+]], i8* dereferenceable({{[0-9]+}}) [[APARM:%.+]])
+// CHECK:  store i8* [[APARM]], i8** [[APTRADDR:%.+]]
+// ..many loads of %0..
+// CHECK:  [[A2:%.+]] = load i8*, i8** [[APTRADDR]]
+// CHECK:  [[AVAL0:%.+]] = load i8, i8* [[A2]]
+// CHECK:  store i8 [[AVAL0]], i8* [[CAP_EXPR:%.+]],
+// CHECK:  [[AVAL1:%.+]] = load i8, i8* [[CAP_EXPR]]
+// CHECK:  load i8, i8* [[CAP_EXPR]]
+// CHECK:  [[AVAL2:%.+]] = load i8, i8* [[CAP_EXPR]]
+// CHECK:  [[ACONV:%.+]] = sext i8 [[AVAL2]] to i32
+// CHECK:  [[ACMP:%.+]] = icmp slt i32 [[ACONV]], 10
+// CHECK:  br i1 [[ACMP]], label %[[PRECOND_THEN:.+]], label %[[PRECOND_END:.+]]
+// CHECK:  [[PRECOND_THEN]]
+// CHECK:  call void @__kmpc_for_static_init_4
+// CHECK:  call void @__kmpc_for_static_fini
+// CHECK:  [[PRECOND_END]]
+
+// no templates for now, as these require special handling in target regions and/or declare target
+
+// HCHECK-LABEL: fint
+// HCHECK: call {{.*}}i32 {{.+}}ftemplate
+// HCHECK: ret i32
+
+// HCHECK: load i16, i16*
+// HCHECK: store i16 %
+// HCHECK: call i32 @__tgt_target_teams(
+// HCHECK: call void @__kmpc_for_static_init_4(
+template <typename T>
+T ftemplate() {
+  short aa = 0;
+
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute simd dist_schedule(static, aa)
+  for (int i = 0; i < 100; i++) {
+  }
+  return T();
+}
+
+int fint(void) { return ftemplate<int>(); }
+
+#endif
+
+// CHECK: !{!"llvm.loop.vectorize.width", i32 8}
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+// CHECK: !{!"llvm.loop.vectorize.width", i32 32}
diff --git a/test/OpenMP/distribute_simd_firstprivate_codegen.cpp b/test/OpenMP/distribute_simd_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..5da9c5dc3b1a
--- /dev/null
+++ b/test/OpenMP/distribute_simd_firstprivate_codegen.cpp
@@ -0,0 +1,380 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd firstprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams
+#pragma omp distribute simd firstprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.*}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // Private alloca's for conversion
+      // LAMBDA: [[G_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_ADDR:%.+]] = alloca float*,
+      // LAMBDA: [[G1_REF:%.+]] = alloca double*,
+
+      // Actual private variables to be used in the body (tmp is used for the reference type)
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+
+      // Store input parameter addresses into private alloca's for conversion
+      // LAMBDA: store double* [[G_IN]], double** [[G_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_ADDR]],
+
+      // LAMBDA-DAG: [[G_ADDR_VAL:%.+]] = load double*, double** [[G_ADDR]],
+      // LAMBDA-DAG: [[G1_ADDR_VAL:%.+]] = load double*, double** [[G1_ADDR]],
+      // LAMBDA-DAG: [[SVAR_ADDR_VAL:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+      // LAMBDA-DAG: [[SFVAR_ADDR_VAL:%.+]] = load float*, float** [[SFVAR_ADDR]],
+      // LAMBDA-DAG: store double* [[G1_ADDR_VAL]], double** [[G1_REF]],
+
+      // LAMBDA-DAG: [[G_VAL:%.+]] = load{{.+}} double, double* [[G_ADDR_VAL]],
+      // LAMBDA-DAG: store double [[G_VAL]], double* [[G_PRIVATE]],
+      // LAMBDA-DAG: [[G1_VAL_REF:%.+]] = load double*, double** [[G1_REF]],
+      // LAMBDA-DAG: [[G1_VAL:%.+]] = load{{.+}} double, double* [[G1_VAL_REF]],
+      // LAMBDA-DAG: store double [[G1_VAL]], double* [[G1_PRIVATE]],
+      // LAMBDA-DAG: store double* [[G1_PRIVATE]], double** [[TMP_PRIVATE]],
+      // LAMBDA-DAG: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_ADDR_VAL]],
+      // LAMBDA-DAG: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA-DAG: [[SFVAR_VAL:%.+]] = load float, float* [[SFVAR_ADDR_VAL]],
+      // LAMBDA-DAG: store float [[SFVAR_VAL]], float* [[SFVAR_PRIVATE]],
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      g += 1;
+      g1 += 1;
+      svar += 3;
+      sfvar += 4.0;
+      // LAMBDA-DAG: [[G_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA-DAG: [[G_NEXT:%.+]] = fadd double [[G_VAL]], 1.{{.+}}
+      // LAMBDA-DAG: store double [[G_NEXT]], double* [[G_PRIVATE]],
+      // LAMBDA-DAG: [[TMP_VAL1:%.+]] = load double*, double** [[TMP_PRIVATE]],
+      // LAMBDA-DAG: [[TMP_VAL_VAL1:%.+]] = load{{.*}} double, double* [[TMP_VAL1]],
+      // LAMBDA-DAG: [[TMP_ADD:%.+]] = fadd double [[TMP_VAL_VAL1]], 1.{{.+}}
+      // LAMBDA-DAG: store{{.*}} double [[TMP_ADD]], double* [[TMP_VAL1]],
+      // LAMBDA-DAG: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA-DAG: [[SVAR_ADD:%.+]] = add{{.*}} i{{[0-9]+}} [[SVAR_VAL]], 3
+      // LAMBDA-DAG: store i{{[0-9]+}} [[SVAR_ADD]], i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA-DAG: [[SFVAR_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA-DAG: [[SFVAR_CONV_VAL1:%.+]] = fpext float [[SFVAR_VAL]] to double
+      // LAMBDA-DAG: [[SFVAR_ADD:%.+]] = fadd double [[SFVAR_CONV_VAL1]], 4.{{.+}}
+      // LAMBDA-DAG: [[SFVAR_CONV_VAL2:%.+]] = fptrunc double [[SFVAR_ADD]] to float
+      // LAMBDA-DAG: store float [[SFVAR_CONV_VAL2:%.+]], float* [[SFVAR_PRIVATE]],
+
+      // call inner lambda (use refs to private alloca's)
+      // LAMBDA: [[GEP_0:%.+]] = getelementptr{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[GEP_0]],
+      // LAMBDA: [[GEP_1:%.+]] = getelementptr{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[TMP_PAR:%.+]] = load double*, double** [[TMP_PRIVATE]],
+      // LAMBDA: store double* [[TMP_PAR]], double** [[GEP_1]],
+      // LAMBDA: [[GEP_2:%.+]] = getelementptr{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[GEP_2]],
+      // LAMBDA: [[GEP_3:%.+]] = getelementptr{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[GEP_3]],
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* {{.+}})
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g += 2;
+	g1 += 2;
+	svar += 4;
+	sfvar += 8.0;
+	// LAMBDA-DAG: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA-DAG: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA-DAG: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]],
+	// LAMBDA-DAG: [[G_REF_VAL:%.+]] = load double, double* [[G_REF]],
+	// LAMBDA-DAG: [[G_REF_ADD:%.+]] = fadd double [[G_REF_VAL]], 2.{{.+}}
+	// LAMBDA-DAG: store double [[G_REF_ADD]], double* [[G_REF]]
+
+	// LAMBDA-DAG: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA-DAG: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA-DAG: [[G1_REF_VAL:%.+]] = load double, double* [[G1_REF]],
+	// LAMBDA-DAG: [[G1_ADD:%.+]] = fadd double [[G1_REF_VAL]], 2.{{.+}}
+	// LAMBDA-DAG: store double [[G1_ADD]], double* [[G1_REF]],
+
+	// LAMBDA-DAG: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA-DAG: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA-DAG: [[SVAR_REF_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA-DAG: [[SVAR_ADD:%.+]] = add{{.*}} i{{[0-9]+}} [[SVAR_REF_VAL]], 4
+	// LAMBDA-DAG: store i{{[0-9]+}} [[SVAR_ADD]], i{{[0-9]+}}* [[SVAR_REF]]
+
+	// LAMBDA-DAG: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA-DAG: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA-DAG: [[SFVAR_REF_VAL:%.+]] = load float, float* [[SFVAR_REF]]
+	// LAMBDA-DAG: [[SFVAR_REF_CONV:%.+]] = fpext float [[SFVAR_REF_VAL]] to double
+	// LAMBDA-DAG: [[SFVAR_ADD:%.+]] = fadd double [[SFVAR_REF_CONV]], 8.{{.+}}
+	// LAMBDA-DAG: [[SFVAR_ADD_CONV:%.+]] = fptrunc double [[SFVAR_ADD]] to float
+	// LAMBDA-DAG: store float [[SFVAR_ADD_CONV]], float* [[SFVAR_REF]],
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams
+  #pragma omp distribute simd firstprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_FLOAT_TY]]]*, [[S_FLOAT_TY]]*, i{{[0-9]+}}*)* [[OMP_OUTLINED:@.+]] to void
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED]](i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.*}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.*}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.*}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]])
+
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[TMP:%.+]] = alloca [[S_FLOAT_TY]]*,
+
+// discard omp loop variables
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// init t_var
+// CHECK-DAG: [[T_VAR_ADDR_CONV_VAL_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_ADDR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_ADDR_CONV_VAL_REF]],
+// CHECK-DAG: store i{{[0-9]+}} [[T_VAR_ADDR_CONV_VAL]], i{{[0-9]+}}* [[T_VAR_PRIV]],
+
+// init vec
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[VEC_ADDR_VAL_BCAST:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i{{[0-9]+}}*
+// CHECK-DAG: [[VEC_PRIV_BCAST:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i{{[0-9]+}}*
+// CHECK-DAG: call void @llvm.memcpy.{{.*}}(i{{[0-9]+}}* [[VEC_PRIV_BCAST]], i{{[0-9]+}}* [[VEC_ADDR_VAL_BCAST]],{{.+}})
+
+// init s_arr
+// CHECK-DAG: [[S_ARR_ADDR_VAL:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_VAL]] to [[S_FLOAT_TY]]*
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.+}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]]{{.+}}
+// CHECK-DAG: [[S_ARR_PRIV_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_PRIV_BGN]]{{.+}}
+// CHECK-DAG: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_PRIV_BGN]], [[S_ARR_PRIV_NEXT]]
+// CHECK-DAG: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_CPY_DONE:.+]], label %[[S_ARR_CPY_BODY:.+]]
+
+// CHECK-DAG: [[S_ARR_CPY_BODY]]:
+// CHECK-DAG: [[S_ARR_SRC_PAST:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BCAST]],{{.+}} ], [ [[S_ARR_SRC:%.+]],{{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_PAST:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]],{{.+}} ], [ [[S_ARR_DST:%.+]],{{.+}} ]
+// CHECK-DAG: [[S_ARR_SRC_BCAST:%.+]] = bitcast{{.+}} [[S_ARR_SRC_PAST]] to{{.+}}
+// CHECK-DAG: [[S_ARR_DST_BCAST:%.+]] = bitcast{{.+}} [[S_ARR_DST_PAST]] to{{.+}}
+// CHECK-DAG: call{{.+}} @llvm.memcpy.{{.+}}({{.+}}* [[S_ARR_DST_BCAST]], {{.+}}* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr{{.+}}
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr{{.+}}
+// CHECK-DAG: [[S_ARR_CPY_FIN:%.+]] = icmp{{.+}} [[S_ARR_DST]], [[S_ARR_PRIV_NEXT]]
+// CHECK-DAG: br i1 [[S_ARR_CPY_FIN]], label %[[S_ARR_CPY_DONE]], label %[[S_ARR_CPY_BODY]]
+// CHECK-DAG: [[S_ARR_CPY_DONE]]:
+
+// init var
+// CHECK-DAG: [[VAR_ADDR_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK-DAG: store{{.+}} [[VAR_ADDR_VAL]],{{.+}} [[TMP]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP]],
+// CHECK-DAG: [[VAR_PRIV_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_PRIV]] to{{.+}}
+// CHECK-DAG: [[TMP_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to{{.+}}
+// CHECK-DAG: call{{.+}} @llvm.memcpy.{{.+}}({{.+}}* [[VAR_PRIV_BCAST]], {{.+}}* [[TMP_BCAST]],{{.+}})
+// CHECK-DAG: store [[S_FLOAT_TY]]* [[VAR_PRIV]], [[S_FLOAT_TY]]** [[TMP_PRIV]],
+
+// init svar
+// CHECK-DAG: [[SVAR_CONV_VAL_REF:%.+]] = load{{.+}},{{.+}} [[SVAR_ADDR]],
+// CHECK-DAG: [[SVAR_CONV_VAL:%.+]] = load{{.+}},{{.+}} [[SVAR_CONV_VAL_REF]],
+// CHECK-DAG: store{{.+}} [[SVAR_CONV_VAL]],{{.+}} [[SVAR_PRIV]],
+
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* %.omp{{.+}},
+// CHECK-DAG: store i{{[0-9]+}} 1, i{{[0-9]+}}* %.omp{{.+}},
+// CHECK-DAG: store i{{[0-9]+}} 1, i{{[0-9]+}}* %.omp{{.+}},
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* %.omp{{.+}},
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// Template
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x i{{[0-9]+}}]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[OMP_OUTLINED_1:@.+]] to void
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED_1]](i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.*}} [[VEC_IN:%.+]], [2 x [[S_INT_TY]]]*{{.*}} [[S_ARR_IN:%.+]], [[S_INT_TY]]*{{.*}} [[VAR_IN:%.+]])
+
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// discard omp loop variables
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+// CHECK: {{.*}} = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK-DAG: [[TMP_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN]], [[S_INT_TY]]** [[VAR_ADDR]],
+
+// init t_var
+// CHECK-DAG: [[T_VAR_ADDR_CONV_VAL_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK-DAG: [[T_VAR_ADDR_CONV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_ADDR_CONV_VAL_REF]],
+// CHECK-DAG: store i{{[0-9]+}} [[T_VAR_ADDR_CONV_VAL]], i{{[0-9]+}}* [[T_VAR_PRIV]],
+
+// init vec
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[VEC_ADDR_VAL_BCAST:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i{{[0-9]+}}*
+// CHECK-DAG: [[VEC_PRIV_BCAST:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i{{[0-9]+}}*
+// CHECK-DAG: call void @llvm.memcpy.{{.*}}(i{{[0-9]+}}* [[VEC_PRIV_BCAST]], i{{[0-9]+}}* [[VEC_ADDR_VAL_BCAST]],{{.+}})
+
+// init s_arr
+// CHECK-DAG: [[S_ARR_ADDR_VAL:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_VAL]] to [[S_INT_TY]]*
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.+}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]]{{.+}}
+// CHECK-DAG: [[S_ARR_PRIV_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_PRIV_BGN]]{{.+}}
+// CHECK-DAG: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_PRIV_BGN]], [[S_ARR_PRIV_NEXT]]
+// CHECK-DAG: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_CPY_DONE:.+]], label %[[S_ARR_CPY_BODY:.+]]
+
+// CHECK-DAG: [[S_ARR_CPY_BODY]]:
+// CHECK-DAG: [[S_ARR_SRC_PAST:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BCAST]],{{.+}} ], [ [[S_ARR_SRC:%.+]],{{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_PAST:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]],{{.+}} ], [ [[S_ARR_DST:%.+]],{{.+}} ]
+// CHECK-DAG: [[S_ARR_SRC_BCAST:%.+]] = bitcast{{.+}} [[S_ARR_SRC_PAST]] to{{.+}}
+// CHECK-DAG: [[S_ARR_DST_BCAST:%.+]] = bitcast{{.+}} [[S_ARR_DST_PAST]] to{{.+}}
+// CHECK-DAG: call{{.+}} @llvm.memcpy.{{.+}}({{.+}}* [[S_ARR_DST_BCAST]], {{.+}}* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr{{.+}}
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr{{.+}}
+// CHECK-DAG: [[S_ARR_CPY_FIN:%.+]] = icmp{{.+}} [[S_ARR_DST]], [[S_ARR_PRIV_NEXT]]
+// CHECK-DAG: br i1 [[S_ARR_CPY_FIN]], label %[[S_ARR_CPY_DONE]], label %[[S_ARR_CPY_BODY]]
+// CHECK-DAG: [[S_ARR_CPY_DONE]]:
+
+// init var
+// CHECK-DAG: [[VAR_ADDR_VAL:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR]],
+// CHECK-DAG: store{{.+}} [[VAR_ADDR_VAL]],{{.+}} [[TMP]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP]],
+// CHECK-DAG: [[VAR_PRIV_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_PRIV]] to{{.+}}
+// CHECK-DAG: [[TMP_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL]] to{{.+}}
+// CHECK-DAG: call{{.+}} @llvm.memcpy.{{.+}}({{.+}}* [[VAR_PRIV_BCAST]], {{.+}}* [[TMP_BCAST]],{{.+}})
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP_PRIV]],
+
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* %.omp{{.+}},
+// CHECK-DAG: store i{{[0-9]+}} 1, i{{[0-9]+}}* %.omp{{.+}},
+// CHECK-DAG: store i{{[0-9]+}} 1, i{{[0-9]+}}* %.omp{{.+}},
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* %.omp{{.+}},
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/distribute_simd_firstprivate_messages.cpp b/test/OpenMP/distribute_simd_firstprivate_messages.cpp
index b9267a3037f0..f0c293b78d80 100644
--- a/test/OpenMP/distribute_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/distribute_simd_firstprivate_messages.cpp
@@ -42,7 +42,7 @@ public:
 };
 class S5 {
   int a;
-  S5(const S5 &s5) : a(s5.a) {} // expected-note 4 {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {} // expected-note 2 {{implicitly declared private here}}
 
 public:
   S5() : a(0) {}
@@ -50,7 +50,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -150,9 +150,10 @@ int foomain(int argc, char **argv) {
 #pragma omp distribute simd firstprivate(i)
   for (int k = 0; k < argc; ++k)
     ++k;
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd lastprivate(g) firstprivate(g) // expected-error {{calling a private constructor of class 'S5'}}
+#pragma omp distribute simd lastprivate(g) firstprivate(g)
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp parallel private(i)
@@ -314,14 +315,16 @@ int main(int argc, char **argv) {
 #pragma omp distribute simd firstprivate(j)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd lastprivate(g) firstprivate(g) // expected-error {{calling a private constructor of class 'S5'}}
+#pragma omp distribute simd lastprivate(g) firstprivate(g)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd lastprivate(n) firstprivate(n) // OK
+#pragma omp distribute simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp parallel
diff --git a/test/OpenMP/distribute_simd_lastprivate_codegen.cpp b/test/OpenMP/distribute_simd_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..df842a009879
--- /dev/null
+++ b/test/OpenMP/distribute_simd_lastprivate_codegen.cpp
@@ -0,0 +1,393 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams
+#pragma omp distribute simd lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // LAMBDA: [[TMP_G1:%.+]] = alloca double*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN_REF]], double** [[TMP_G1]],
+      // LAMBDA: [[TMP_G1_VAL:%.+]] = load double*, double** [[TMP_G1]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+
+      // linear counter
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_COUNTER_BLOCK:.+]], label %[[OMP_COUNTER_DONE:.+]]
+      // LAMBDA: [[OMP_COUNTER_BLOCK]]:
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* %
+      // LAMBDA: br label %[[OMP_COUNTER_DONE]]
+      // LAMBDA: [[OMP_COUNTER_DONE]]:
+
+
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[TMP_G1_VAL]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[TMP:%.*]] = alloca [[S_FLOAT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: [[VAR_ADDR_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_ADDR_REF]], [[S_FLOAT_TY]]** [[TMP]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[TMP_REF:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]],  [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_ADDR1_REF]], [[S_INT_TY]]** [[TMP]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK-DAG: [[TMP_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/distribute_simd_lastprivate_messages.cpp b/test/OpenMP/distribute_simd_lastprivate_messages.cpp
index 0f96cb4c3ec8..82ce15e55048 100644
--- a/test/OpenMP/distribute_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/distribute_simd_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -313,14 +313,16 @@ int main(int argc, char **argv) {
 #pragma omp distribute simd lastprivate(j)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{firstprivate variable cannot be lastprivate}} expected-note@+3 {{defined as firstprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp distribute simd firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i)
     foo();
+// expected-error@+3 {{lastprivate variable cannot be firstprivate}} expected-note@+3 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd lastprivate(n) firstprivate(n) // OK
+#pragma omp distribute simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i)
     foo();
   static int si;
diff --git a/test/OpenMP/distribute_simd_linear_messages.cpp b/test/OpenMP/distribute_simd_linear_messages.cpp
index c60e0a247071..e4d71fee8250 100644
--- a/test/OpenMP/distribute_simd_linear_messages.cpp
+++ b/test/OpenMP/distribute_simd_linear_messages.cpp
@@ -18,41 +18,49 @@ void test_linear_colons()
 {
   int B = 0;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(B:bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(B:B::bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(X::x : ::z)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(B,::z, X::x)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(::z)
@@ -63,6 +71,7 @@ void test_linear_colons()
 #pragma omp distribute simd linear(B::bfoo()) // expected-error {{expected variable name}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+3 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(B::ib,B:C1+C2)
@@ -148,11 +157,13 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
@@ -163,6 +174,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear (argc : 5)
@@ -183,6 +195,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute simd linear (argv[1]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(e, g)
@@ -193,37 +206,12 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp distribute simd linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear(i)
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target
-    #pragma omp teams
-    #pragma omp distribute simd linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target
-#pragma omp teams
-#pragma omp parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute simd linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
   return 0;
 }
 
@@ -262,11 +250,13 @@ int main(int argc, char **argv) {
 #pragma omp distribute simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
@@ -277,6 +267,7 @@ int main(int argc, char **argv) {
 #pragma omp distribute simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams
 #pragma omp distribute simd linear (argc)
@@ -310,28 +301,18 @@ int main(int argc, char **argv) {
 
   #pragma omp parallel
   {
-    int i;
+    int k;
     #pragma omp target
     #pragma omp teams
-    #pragma omp distribute simd linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
+    #pragma omp distribute simd linear(k)
+      for (k = 0; k < argc; ++k) ++k;
 
     #pragma omp target
     #pragma omp teams
-    #pragma omp distribute simd linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
+    #pragma omp distribute simd linear(k : 4)
+      for (k = 0; k < argc; k+=4) { }
   }
 
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
   foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
   return 0;
 }
diff --git a/test/OpenMP/distribute_simd_misc_messages.c b/test/OpenMP/distribute_simd_misc_messages.c
index 5fc2cb64cc62..f643ed1d5b8e 100644
--- a/test/OpenMP/distribute_simd_misc_messages.c
+++ b/test/OpenMP/distribute_simd_misc_messages.c
@@ -561,89 +561,6 @@ void test_linear() {
 #pragma omp distribute simd linear(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
-
-  int x, y;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected expression}}
-#pragma omp distribute simd linear(x :)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected expression}} expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
-#pragma omp distribute simd linear(x :, )
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute simd linear(x : 1)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-#pragma omp distribute simd linear(x : 2 * 2)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
-#pragma omp distribute simd linear(x : 1, y)
-  for (i = 0; i < 16; ++i)
-    ;
-#pragma omp target
-#pragma omp teams
-// expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
-#pragma omp distribute simd linear(x : 1, y, z : 1)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as linear}}
-// expected-error@+1 {{linear variable cannot be linear}}
-#pragma omp distribute simd linear(x) linear(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as private}}
-// expected-error@+1 {{private variable cannot be linear}}
-#pragma omp distribute simd private(x) linear(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as linear}}
-// expected-error@+1 {{linear variable cannot be private}}
-#pragma omp distribute simd linear(x) private(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-warning@+1 {{zero linear step (x and other variables in clause should probably be const)}}
-#pragma omp distribute simd linear(x, y : 0)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as linear}}
-// expected-error@+1 {{linear variable cannot be lastprivate}}
-#pragma omp distribute simd linear(x) lastprivate(x)
-  for (i = 0; i < 16; ++i)
-    ;
-
-#pragma omp target
-#pragma omp teams
-// expected-note@+2 {{defined as lastprivate}}
-// expected-error@+1 {{lastprivate variable cannot be linear}}
-#pragma omp distribute simd lastprivate(x) linear(x)
-  for (i = 0; i < 16; ++i)
-    ;
 }
 
 void test_aligned() {
@@ -1083,26 +1000,26 @@ void test_loop_messages() {
 }
 
 void linear_modifiers(int argc) {
-  int f;
+  int k;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd linear(f)
-  for (int k = 0; k < argc; ++k) ++k;
+#pragma omp distribute simd linear(k)
+  for (k = 0; k < argc; ++k) ++k;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd linear(val(f))
-  for (int k = 0; k < argc; ++k) ++k;
+#pragma omp distribute simd linear(val(k))
+  for (k = 0; k < argc; ++k) ++k;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd linear(uval(f)) // expected-error {{expected 'val' modifier}}
-  for (int k = 0; k < argc; ++k) ++k;
+#pragma omp distribute simd linear(uval(k)) // expected-error {{expected 'val' modifier}}
+  for (k = 0; k < argc; ++k) ++k;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd linear(ref(f)) // expected-error {{expected 'val' modifier}}
-  for (int k = 0; k < argc; ++k) ++k;
+#pragma omp distribute simd linear(ref(k)) // expected-error {{expected 'val' modifier}}
+  for (k = 0; k < argc; ++k) ++k;
 #pragma omp target
 #pragma omp teams
-#pragma omp distribute simd linear(foo(f)) // expected-error {{expected 'val' modifier}}
-  for (int k = 0; k < argc; ++k) ++k;
+#pragma omp distribute simd linear(foo(k)) // expected-error {{expected 'val' modifier}}
+  for (k = 0; k < argc; ++k) ++k;
 }
 
diff --git a/test/OpenMP/distribute_simd_private_codegen.cpp b/test/OpenMP/distribute_simd_private_codegen.cpp
new file mode 100644
index 000000000000..ede807a0fb63
--- /dev/null
+++ b/test/OpenMP/distribute_simd_private_codegen.cpp
@@ -0,0 +1,208 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd private(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]]()
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams
+#pragma omp distribute simd private(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}})
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double,
+      // LAMBDA: [[TMP_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float,
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR]], double** [[TMP_PRIVATE_ADDR]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE_ADDR]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE_ADDR]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd private(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  #pragma omp target
+  #pragma omp teams
+#pragma omp distribute simd
+  for (i = 0; i < 2; ++i) {
+    ;
+  }
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]]()
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 0, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*)* [[OMP_OUTLINED:@.+]] to void
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-NOT: alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-NOT: alloca [[S_FLOAT_TY]],
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = phi [[S_FLOAT_TY]]*
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[S_ARR_PRIV_ITEM]])
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](
+// CHECK: ret
+
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]]()
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 0, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*)* [[OMP_OUTLINED_1:@.+]] to void
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED_1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK-NOT: alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK-NOT: alloca [[S_INT_TY]],
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: {{.+}}:
+// CHECK: [[S_ARR_PRIV_ITEM:%.+]] = phi [[S_INT_TY]]*
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[S_ARR_PRIV_ITEM]])
+// CHECK-NOT: [[T_VAR_PRIV]]
+// CHECK-NOT: [[VEC_PRIV]]
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR]]([[S_INT_TY]]* [[VAR_PRIV]])
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/distribute_simd_reduction_codegen.cpp b/test/OpenMP/distribute_simd_reduction_codegen.cpp
new file mode 100644
index 000000000000..8485d8d27d52
--- /dev/null
+++ b/test/OpenMP/distribute_simd_reduction_codegen.cpp
@@ -0,0 +1,204 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute simd reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: alloca i{{.+}},
+    // LAMBDA: alloca i{{.+}},
+    // LAMBDA: alloca i{{.+}},
+    // LAMBDA: alloca i{{.+}},
+    // LAMBDA: alloca i{{.+}},
+    // LAMBDA: alloca i{{.+}},
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[LAST_ITER:%.+]] = load i32, i32* %
+    // LAMBDA: [[IS_LAST:%.+]] = icmp ne i32 [[LAST_ITER]], 0
+    // LAMBDA: br i1 [[IS_LAST]], label %[[THEN:.+]], label %[[DONE:.+]]
+    // LAMBDA: [[THEN]]
+    // LAMBDA: store i32 2, i32* %
+    // LAMBDA: br label %[[DONE]]
+    // LAMBDA: [[DONE]]
+    // LAMBDA: [[SIVAR_ORIG_VAL:%.+]] = load i32, i32* [[SIVAR_REF]],
+    // LAMBDA: [[SIVAR_PRIV_VAL:%.+]] = load i32, i32* [[SIVAR_PRIV]],
+    // LAMBDA: [[ADD:%.+]] = add nsw i32 [[SIVAR_ORIG_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store i32 [[ADD]], i32* [[SIVAR_REF]],
+    // LAMBDA: ret void
+
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams
+#pragma omp distribute simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK-64: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[LAST_ITER:%.+]] = load i32, i32* %
+// CHECK: [[IS_LAST:%.+]] = icmp ne i32 [[LAST_ITER]], 0
+// CHECK: br i1 [[IS_LAST]], label %[[THEN:.+]], label %[[DONE:.+]]
+// CHECK: [[THEN]]
+// CHECK: store i32 2, i32* %
+// CHECK: br label %[[DONE]]
+// CHECK: [[DONE]]
+// CHECK: [[SIVAR_ORIG_VAL:%.+]] = load i32, i32* [[SIVAR_REF]],
+// CHECK: [[SIVAR_PRIV_VAL:%.+]] = load i32, i32* [[SIVAR_PRIV]],
+// CHECK: [[ADD:%.+]] = add nsw i32 [[SIVAR_ORIG_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store i32 [[ADD]], i32* [[SIVAR_REF]],
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i{{64|32}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK-64: [[TVAR_CONV:%.+]] = bitcast{{.+}} [[TVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: alloca i{{.+}},
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[LAST_ITER:%.+]] = load i32, i32* %
+// CHECK: [[IS_LAST:%.+]] = icmp ne i32 [[LAST_ITER]], 0
+// CHECK: br i1 [[IS_LAST]], label %[[THEN:.+]], label %[[DONE:.+]]
+// CHECK: [[THEN]]
+// CHECK: store i32 2, i32* %
+// CHECK: br label %[[DONE]]
+// CHECK: [[DONE]]
+// CHECK: [[TVAR_ORIG_VAL:%.+]] = load i32, i32* [[TVAR_REF]],
+// CHECK: [[TVAR_PRIV_VAL:%.+]] = load i32, i32* [[TVAR_PRIV]],
+// CHECK: [[ADD:%.+]] = add nsw i32 [[TVAR_ORIG_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store i32 [[ADD]], i32* [[TVAR_REF]],
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/distribute_simd_reduction_messages.cpp b/test/OpenMP/distribute_simd_reduction_messages.cpp
index fca3e85a7104..ddedcb0892ba 100644
--- a/test/OpenMP/distribute_simd_reduction_messages.cpp
+++ b/test/OpenMP/distribute_simd_reduction_messages.cpp
@@ -27,9 +27,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/dump.cpp b/test/OpenMP/dump.cpp
index 378b53ce5bf1..5ec202cc3040 100644
--- a/test/OpenMP/dump.cpp
+++ b/test/OpenMP/dump.cpp
@@ -52,11 +52,11 @@ struct S {
 // CHECK-NEXT: |       |-OMPScheduleClause {{.+}} <col:61, col:79>
 // CHECK-NEXT: |       | `-ImplicitCastExpr {{.+}} <col:78> 'int' <LValueToRValue>
 // CHECK-NEXT: |       |   `-DeclRefExpr {{.+}} <col:78> 'int' lvalue OMPCapturedExpr {{.+}} '.capture_expr.' 'int'
-// CHECK-NEXT: |       |-CapturedStmt {{.+}} <line:40:5, <invalid sloc>>
+// CHECK-NEXT: |       |-CapturedStmt {{.+}} <line:40:5, line:41:9>
 // CHECK-NEXT: |       | |-CapturedDecl {{.+}} <<invalid sloc>> <invalid sloc>
-// CHECK-NEXT: |       | | |-ForStmt {{.+}} <col:5, <invalid sloc>>
-// CHECK:      |       | | | `-UnaryOperator {{.+}} <line:41:7, <invalid sloc>> 'int' lvalue prefix '++'
-// CHECK-NEXT: |       | | |   `-DeclRefExpr {{.+}} <<invalid sloc>> 'int' lvalue OMPCapturedExpr {{.+}} 'a' 'int &'
+// CHECK-NEXT: |       | | |-ForStmt {{.+}} <line:40:5, line:41:9>
+// CHECK:      |       | | | `-UnaryOperator {{.+}} <line:41:7, col:9> 'int' lvalue prefix '++'
+// CHECK-NEXT: |       | | |   `-DeclRefExpr {{.+}} <col:9> 'int' lvalue OMPCapturedExpr {{.+}} 'a' 'int &'
 
 #pragma omp declare simd
 #pragma omp declare simd inbranch
diff --git a/test/OpenMP/for_codegen.cpp b/test/OpenMP/for_codegen.cpp
index 466139cfb2fc..1f5dd7ddd0dd 100644
--- a/test/OpenMP/for_codegen.cpp
+++ b/test/OpenMP/for_codegen.cpp
@@ -11,15 +11,26 @@
 
 // CHECK: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
 // CHECK-DAG: [[IMPLICIT_BARRIER_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 66, i32 0, i32 0, i8*
+// CHECK-DAG: [[LOOP_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 514, i32 0, i32 0, i8*
 // CHECK-DAG: [[I:@.+]] = global i8 1,
 // CHECK-DAG: [[J:@.+]] = global i8 2,
 // CHECK-DAG: [[K:@.+]] = global i8 3,
 
+// CHECK-LABEL: loop_with_counter_collapse
+void loop_with_counter_collapse() {
+  // CHECK: call void @__kmpc_for_static_init_8(%ident_t* @
+  // CHECK: call void @__kmpc_for_static_fini(%ident_t* @
+  #pragma omp for collapse(2)
+  for (int i = 0; i < 4; i++) {
+    for (int j = i; j < 4; j++) {
+    }
+  }
+}
 // CHECK-LABEL: define {{.*void}} @{{.*}}without_schedule_clause{{.*}}(float* {{.+}}, float* {{.+}}, float* {{.+}}, float* {{.+}})
 void without_schedule_clause(float *a, float *b, float *c, float *d) {
 // CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]])
   #pragma omp for nowait
-// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
+// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
 // UB = min(UB, GlobalUB)
 // CHECK-NEXT: [[UB:%.+]] = load i32, i32* [[OMP_UB]]
 // CHECK-NEXT: [[UBCMP:%.+]] = icmp sgt i32 [[UB]], 4571423
@@ -51,7 +62,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: br label %{{.+}}
   }
 // CHECK: [[LOOP1_END]]
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 // CHECK-NOT: __kmpc_barrier
 // CHECK: ret void
 }
@@ -60,7 +71,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
 void static_not_chunked(float *a, float *b, float *c, float *d) {
 // CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]])
   #pragma omp for schedule(static)
-// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
+// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
 // UB = min(UB, GlobalUB)
 // CHECK-NEXT: [[UB:%.+]] = load i32, i32* [[OMP_UB]]
 // CHECK-NEXT: [[UBCMP:%.+]] = icmp sgt i32 [[UB]], 4571423
@@ -92,7 +103,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: br label %{{.+}}
   }
 // CHECK: [[LOOP1_END]]
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 // CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
 // CHECK: ret void
 }
@@ -101,7 +112,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) {
 void static_chunked(float *a, float *b, float *c, float *d) {
 // CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]])
   #pragma omp for schedule(monotonic: static, 5)
-// CHECK: call void @__kmpc_for_static_init_4u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 536870945, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 5)
+// CHECK: call void @__kmpc_for_static_init_4u([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]], i32 536870945, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 5)
 // UB = min(UB, GlobalUB)
 // CHECK: [[UB:%.+]] = load i32, i32* [[OMP_UB]]
 // CHECK-NEXT: [[UBCMP:%.+]] = icmp ugt i32 [[UB]], 16908288
@@ -152,7 +163,7 @@ void static_chunked(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: store i32 [[ADD_UB]], i32* [[OMP_UB]]
 
 // CHECK: [[O_LOOP1_END]]
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 // CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
 // CHECK: ret void
 }
@@ -329,8 +340,8 @@ void runtime(float *a, float *b, float *c, float *d) {
 // CHECK-LABEL: test_precond
 void test_precond() {
   // CHECK: [[A_ADDR:%.+]] = alloca i8,
-  // CHECK: [[CAP:%.+]] = alloca i8,
   // CHECK: [[I_ADDR:%.+]] = alloca i8,
+  // CHECK: [[CAP:%.+]] = alloca i8,
   char a = 0;
   // CHECK: store i8 0,
   // CHECK: store i32
@@ -375,7 +386,9 @@ void parallel_for(float *a) {
 char i = 1, j = 2, k = 3;
 // CHECK-LABEL: for_with_global_lcv
 void for_with_global_lcv() {
+// CHECK: alloca i8,
 // CHECK: [[I_ADDR:%.+]] = alloca i8,
+// CHECK: alloca i8,
 // CHECK: [[J_ADDR:%.+]] = alloca i8,
 
 // CHECK: call void @__kmpc_for_static_init_4(
diff --git a/test/OpenMP/for_firstprivate_codegen.cpp b/test/OpenMP/for_firstprivate_codegen.cpp
index 1e1d404b48db..88f54103ad69 100644
--- a/test/OpenMP/for_firstprivate_codegen.cpp
+++ b/test/OpenMP/for_firstprivate_codegen.cpp
@@ -81,8 +81,10 @@ int main() {
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIVATE_REF:%.+]] = alloca i{{[0-9]+}}*,
     // LAMBDA: [[SIVAR2_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
 
     // LAMBDA:  store i{{[0-9]+}}* [[SIVAR_REF]], i{{[0-9]+}}** %{{.+}},
@@ -91,20 +93,26 @@ int main() {
 
     // LAMBDA: [[G_VAL:%.+]] = load volatile i{{[0-9]+}}, i{{[0-9]+}}* [[G]]
     // LAMBDA: store i{{[0-9]+}} [[G_VAL]], i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
+    // LAMBDA: store i{{[0-9]+}}* [[G1_PRIVATE_ADDR]], i{{[0-9]+}}** [[G1_PRIVATE_REF]],
     // LAMBDA: [[SIVAR2_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR2_PRIVATE_ADDR_REF]]
     // LAMBDA: store i{{[0-9]+}} [[SIVAR2_VAL]], i{{[0-9]+}}* [[SIVAR2_PRIVATE_ADDR]]
 
     // LAMBDA-NOT: call void @__kmpc_barrier(
     g = 1;
-    g1 = 1;
-    sivar = 2;
+    g1 = 2;
+    sivar = 3;
     // LAMBDA: call void @__kmpc_for_static_init_4(
 
     // LAMBDA: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
-    // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[SIVAR2_PRIVATE_ADDR]],
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PRIVATE_REF]],
+    // LAMBDA: store volatile i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_PRIVATE_ADDR]],
+    // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SIVAR2_PRIVATE_ADDR]],
     // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
     // LAMBDA: store i{{[0-9]+}}* [[G_PRIVATE_ADDR]], i{{[0-9]+}}** [[G_PRIVATE_ADDR_REF]]
-    // LAMBDA: [[SIVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+    // LAMBDA: [[G1_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PRIVATE_REF]],
+    // LAMBDA: store i{{[0-9]+}}* [[G1_PRIVATE_ADDR]], i{{[0-9]+}}** [[G1_PRIVATE_ADDR_REF]]
+    // LAMBDA: [[SIVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
     // LAMBDA: store i{{[0-9]+}}* [[SIVAR2_PRIVATE_ADDR]], i{{[0-9]+}}** [[SIVAR_PRIVATE_ADDR_REF]]
     // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
     // LAMBDA: call void @__kmpc_for_static_fini(
@@ -112,17 +120,20 @@ int main() {
     [&]() {
       // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
       // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
-      g = 2;
-      g1 = 2;
-      sivar = 4;
+      g = 4;
+      g1 = 5;
+      sivar = 6;
       // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
 
       // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
       // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
-      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
-      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 5, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
       // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
-      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 6, i{{[0-9]+}}* [[SIVAR_REF]]
     }();
   }
   }();
@@ -144,6 +155,7 @@ int main() {
     // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: alloca i{{[0-9]+}},
+    // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
     // BLOCKS: [[G1_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
     // BLOCKS: [[SIVAR2_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
@@ -208,6 +220,7 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
@@ -262,38 +275,30 @@ int main() {
 // CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]()
 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
 // CHECK: [[TVAR:%.+]] = alloca i32,
-// CHECK: [[TVAR_CAST:%.+]] = alloca i64,
 // CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
-// CHECK: [[TVAR_VAL:%.+]] = load i32, i32* [[TVAR]],
-// CHECK: [[TVAR_CONV:%.+]] = bitcast i64* [[TVAR_CAST]] to i32*
-// CHECK: store i32 [[TVAR_VAL]], i32* [[TVAR_CONV]],
-// CHECK: [[PVT_CASTVAL:%[^,]+]] = load i64, i64* [[TVAR_CAST]],
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, [2 x i32]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void  (i32*, i32*, ...)*), i64 [[PVT_CASTVAL]],
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i32*, [2 x i32]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void  (i32*, i32*, ...)*), i32* [[TVAR]],
 // CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: ret
 //
-// CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 {{.*}}%{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [2 x [[S_INT_TY]]]* dereferenceable(8) %{{.+}}, [[S_INT_TY]]* dereferenceable(4) %{{.+}})
+// CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [2 x [[S_INT_TY]]]* dereferenceable(8) %{{.+}}, [[S_INT_TY]]* dereferenceable(4) %{{.+}})
 // Skip temp vars for loop
-// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
-// CHECK: %{{.+}} = bitcast i64* [[T_VAR_PRIV]] to i32*
 
-// CHECK-NOT: load i{{[0-9]+}}*, i{{[0-9]+}}** %
+// CHECK: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %
 // CHECK: [[VEC_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** %
 // CHECK: [[S_ARR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** %
 // CHECK: [[VAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** %
 
-// firstprivate t_var(t_var)
-// CHECK-NOT: load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_REF]],
-
 // firstprivate vec(vec)
 // CHECK: [[VEC_DEST:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
 // CHECK: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_REF]] to i8*
diff --git a/test/OpenMP/for_lastprivate_codegen.cpp b/test/OpenMP/for_lastprivate_codegen.cpp
index f0fba042c58a..30a17eaf581a 100644
--- a/test/OpenMP/for_lastprivate_codegen.cpp
+++ b/test/OpenMP/for_lastprivate_codegen.cpp
@@ -213,6 +213,7 @@ int main() {
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[B_PRIV:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[C_PRIV:%.+]] = alloca i{{[0-9]+}},
@@ -246,6 +247,7 @@ int main() {
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}, align 128
     // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIVATE_REF:%.+]] = alloca i{{[0-9]+}}*,
     // LAMBDA: [[SIVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[SIVAR_PRIVATE_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %{{.+}},
 
@@ -254,10 +256,15 @@ int main() {
 
     // LAMBDA: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1)
     // LAMBDA: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PRIVATE_REF]],
+    // LAMBDA: store volatile i{{[0-9]+}} 1, i{{[0-9]+}}* [[G1_PRIVATE_ADDR]],
     // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[SIVAR_PRIVATE_ADDR]],
     // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
     // LAMBDA: store i{{[0-9]+}}* [[G_PRIVATE_ADDR]], i{{[0-9]+}}** [[G_PRIVATE_ADDR_REF]]
-    // LAMBDA: [[SIVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+    // LAMBDA: [[G1_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PRIVATE_REF]],
+    // LAMBDA: store i{{[0-9]+}}* [[G1_PRIVATE_ADDR]], i{{[0-9]+}}** [[G1_PRIVATE_ADDR_REF]]
+    // LAMBDA: [[SIVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
     // LAMBDA: store i{{[0-9]+}}* [[SIVAR_PRIVATE_ADDR]], i{{[0-9]+}}** [[SIVAR_PRIVATE_ADDR_REF]]
     // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
     // LAMBDA: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]])
@@ -291,7 +298,10 @@ int main() {
       // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
       // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
       // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
-      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
       // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
       // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
     }();
@@ -398,6 +408,7 @@ int main() {
 // BLOCKS: alloca i{{[0-9]+}},
 // BLOCKS: alloca i{{[0-9]+}},
 // BLOCKS: alloca i{{[0-9]+}},
+// BLOCKS: alloca i{{[0-9]+}},
 // BLOCKS: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
 // BLOCKS: [[B_PRIV:%.+]] = alloca i{{[0-9]+}},
 // BLOCKS: [[C_PRIV:%.+]] = alloca i{{[0-9]+}},
@@ -468,6 +479,7 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
@@ -602,6 +614,7 @@ int main() {
 // CHECK: ret void
 
 // CHECK: define internal void [[MAIN_MICROTASK3]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}})
+// CHECK: alloca i8,
 // CHECK: [[CNT_PRIV:%.+]] = alloca i8,
 
 // CHECK: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[GTID_ADDR_REF]]
@@ -664,6 +677,7 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[B_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[C_PRIV:%.+]] = alloca i{{[0-9]+}},
diff --git a/test/OpenMP/for_lastprivate_messages.cpp b/test/OpenMP/for_lastprivate_messages.cpp
index 79912b6efbca..13ef35e925a7 100644
--- a/test/OpenMP/for_lastprivate_messages.cpp
+++ b/test/OpenMP/for_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/for_linear_codegen.cpp b/test/OpenMP/for_linear_codegen.cpp
index 16a67c0b91d2..54cdf8a4cb33 100644
--- a/test/OpenMP/for_linear_codegen.cpp
+++ b/test/OpenMP/for_linear_codegen.cpp
@@ -157,6 +157,7 @@ int main() {
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[B_PRIV:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[C_PRIV:%.+]] = alloca i{{[0-9]+}},
@@ -181,6 +182,7 @@ int main() {
 
     // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}})
     // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
@@ -233,6 +235,7 @@ int main() {
   for (int i = 0; i < 2; ++i) {
     // BLOCKS: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}})
     // BLOCKS: alloca i{{[0-9]+}},
+    // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}},
     // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: alloca i{{[0-9]+}},
@@ -307,6 +310,7 @@ int main() {
 // BLOCKS: alloca i{{[0-9]+}},
 // BLOCKS: alloca i{{[0-9]+}},
 // BLOCKS: alloca i{{[0-9]+}},
+// BLOCKS: alloca i{{[0-9]+}},
 // BLOCKS: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
 // BLOCKS: [[B_PRIV:%.+]] = alloca i{{[0-9]+}},
 // BLOCKS: [[C_PRIV:%.+]] = alloca i{{[0-9]+}},
@@ -351,13 +355,13 @@ int main() {
 
 // CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, i64* dereferenceable(8) %{{.+}})
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_START:%.+]] = alloca float*,
 // CHECK: [[LVAR_START:%.+]] = alloca i64,
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
-// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_PRIV:%.+]] = alloca float*,
 // CHECK: [[LVAR_PRIV:%.+]] = alloca i64,
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
@@ -419,6 +423,7 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[B_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[C_PRIV:%.+]] = alloca i{{[0-9]+}},
@@ -446,13 +451,13 @@ int main() {
 
 // CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i32** dereferenceable(8) %{{.+}}, i32* dereferenceable(4) %{{.+}})
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_START:%.+]] = alloca i32*,
 // CHECK: [[LVAR_START:%.+]] = alloca i32,
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
-// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_PRIV:%.+]] = alloca i32*,
 // CHECK: [[LVAR_PRIV:%.+]] = alloca i32,
 // CHECK: [[LVAR_PRIV_REF:%.+]] = alloca i32*,
diff --git a/test/OpenMP/for_private_codegen.cpp b/test/OpenMP/for_private_codegen.cpp
index 5bad8d00688d..f1416d75b64b 100644
--- a/test/OpenMP/for_private_codegen.cpp
+++ b/test/OpenMP/for_private_codegen.cpp
@@ -52,6 +52,8 @@ int main() {
   for (int i = 0; i < 2; ++i) {
     // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}})
     // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double,
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double,
+    // LAMBDA: [[G1_PRIVATE_REF:%.+]] = alloca double*,
     // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float,
     g = 1;
@@ -60,13 +62,18 @@ int main() {
     sfvar = 4.0;
     // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
     // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE_ADDR]],
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = load double*, double** [[G1_PRIVATE_REF]],
+    // LAMBDA: store volatile double 1.0{{.+}}, double* [[G1_PRIVATE_ADDR]],
     // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]],
     // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE_ADDR]],
     // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
     // LAMBDA: store double* [[G_PRIVATE_ADDR]], double** [[G_PRIVATE_ADDR_REF]]
-    // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+    // LAMBDA: [[G1_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+    // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = load double*, double** [[G1_PRIVATE_REF]],
+    // LAMBDA: store double* [[G1_PRIVATE_ADDR]], double** [[G1_PRIVATE_ADDR_REF]]
+    // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
     // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
-    // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+    // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
     // LAMBDA: store float* [[SFVAR_PRIVATE_ADDR]], float** [[SFVAR_PRIVATE_ADDR_REF]]
     // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
     // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
@@ -81,10 +88,13 @@ int main() {
       // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
       // LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
       // LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
-      // LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load double*, double** [[G1_PTR_REF]]
+      // LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]]
+      // LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
       // LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
       // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
-      // LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
       // LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
       // LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
     }();
diff --git a/test/OpenMP/for_reduction_codegen.cpp b/test/OpenMP/for_reduction_codegen.cpp
index d5afae990bf4..cef457333062 100644
--- a/test/OpenMP/for_reduction_codegen.cpp
+++ b/test/OpenMP/for_reduction_codegen.cpp
@@ -27,7 +27,7 @@ struct S {
 // CHECK-DAG: [[REDUCTION_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 18, i32 0, i32 0, i8*
 // CHECK-DAG: [[REDUCTION_LOCK:@.+]] = common global [8 x i32] zeroinitializer
 
-template <typename T>
+template <typename T, int length>
 T tmain() {
   T t;
   S<T> test;
@@ -36,6 +36,7 @@ T tmain() {
   S<T> s_arr[] = {1, 2};
   S<T> &var = test;
   S<T> var1;
+  S<T> arr[length];
 #pragma omp parallel
 #pragma omp for reduction(+:t_var) reduction(&:var) reduction(&& : var1) reduction(min: t_var1) nowait
   for (int i = 0; i < 2; ++i) {
@@ -48,6 +49,12 @@ T tmain() {
     vec[i] = t_var;
     s_arr[i] = var;
   }
+#pragma omp parallel
+#pragma omp for reduction(+ : arr[1:length-2])
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
   return T();
 }
 
@@ -180,12 +187,12 @@ int main() {
   S<float> test;
   float t_var = 0, t_var1;
   int vec[] = {1, 2};
-  S<float> s_arr[] = {1, 2};
+  S<float> s_arr[] = {1, 2, 3, 4};
   S<float> &var = test;
   S<float> var1, arrs[10][4];
   S<float> **var2 = foo();
-  S<float> vvar2[2];
-  S<float> (&var3)[2] = s_arr;
+  S<float> vvar2[5];
+  S<float> (&var3)[4] = s_arr;
 #pragma omp parallel
 #pragma omp for reduction(+:t_var) reduction(&:var) reduction(&& : var1) reduction(min: t_var1)
   for (int i = 0; i < 2; ++i) {
@@ -200,11 +207,28 @@ int main() {
 #pragma omp for reduction(+:arr) reduction(&:arrs)
   for (int i = 0; i < 10; ++i)
     ++arr[1][i];
+  // arr is a VLA, but the array section has constant length so we can generate a constant sized array!
+#pragma omp parallel
+#pragma omp for reduction(+:arr[1][0:2])
+  for (int i = 0; i < 10; ++i)
+    ++arr[1][i];
 #pragma omp parallel
 #pragma omp for reduction(& : var2[0 : 5][1 : 6])
   for (int i = 0; i < 10; ++i)
     ;
 #pragma omp parallel
+#pragma omp for reduction(& : var2[1][1 : 6])
+  for (int i = 0; i < 10; ++i)
+    ;
+#pragma omp parallel
+#pragma omp for reduction(& : var2[1 : 1][1 : 6])
+  for (int i = 0; i < 10; ++i)
+    ;
+#pragma omp parallel
+#pragma omp for reduction(& : var2[1 : 1][1])
+  for (int i = 0; i < 10; ++i)
+    ;
+#pragma omp parallel
 #pragma omp for reduction(& : vvar2[0 : 5])
   for (int i = 0; i < 10; ++i)
     ;
@@ -213,28 +237,43 @@ int main() {
   for (int i = 0; i < 10; ++i)
     ;
 #pragma omp parallel
+#pragma omp for reduction(& : var3[ : 2])
+  for (int i = 0; i < 10; ++i)
+    ;
+  // TODO: The compiler should also be able to generate a constant sized array in this case!
+#pragma omp parallel
+#pragma omp for reduction(& : var3[2 : ])
+  for (int i = 0; i < 10; ++i)
+    ;
+#pragma omp parallel
 #pragma omp for reduction(& : var3)
   for (int i = 0; i < 10; ++i)
     ;
-  return tmain<int>();
+  return tmain<int, 42>();
 #endif
 }
 
 // CHECK: define {{.*}}i{{[0-9]+}} @main()
 // CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, float*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float*, [2 x i32]*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, float*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float*, [2 x i32]*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK:@.+]] to void
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, i64, i32*, [2 x i32]*, [10 x [4 x [[S_FLOAT_TY]]]]*)* [[MAIN_MICROTASK1:@.+]] to void
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, i64, i32*, [10 x [4 x [[S_FLOAT_TY]]]]*)* [[MAIN_MICROTASK2:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[S_FLOAT_TY]]***)* [[MAIN_MICROTASK3:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK4:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK5:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK6:@.+]] to void
-// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT:@.+]]()
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 3, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, i64, i32*)* [[MAIN_MICROTASK3:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[S_FLOAT_TY]]***)* [[MAIN_MICROTASK4:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[S_FLOAT_TY]]***)* [[MAIN_MICROTASK5:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[S_FLOAT_TY]]***)* [[MAIN_MICROTASK6:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[S_FLOAT_TY]]***)* [[MAIN_MICROTASK7:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [5 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK8:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK9:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK10:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK11:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK12:@.+]] to void
+// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT_42:@.+]]()
 // CHECK: call {{.*}} [[S_FLOAT_TY_DESTR:@.+]]([[S_FLOAT_TY]]*
 // CHECK: ret
 //
-// CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float* dereferenceable(4) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(4) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(4) %{{.+}}, float* dereferenceable(4) %{{.+}}, [2 x i32]* dereferenceable(8) %vec, [2 x [[S_FLOAT_TY]]]* dereferenceable(8) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float* dereferenceable(4) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(4) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(4) %{{.+}}, float* dereferenceable(4) %{{.+}}, [2 x i32]* dereferenceable(8) %vec, [4 x [[S_FLOAT_TY]]]* dereferenceable(16) %{{.+}})
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca float,
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
@@ -491,7 +530,7 @@ int main() {
 // CHECK: store float [[UP]], float* [[T_VAR1_LHS]],
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* %{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(160) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* {{.+}} %{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(160) %{{.+}})
 
 // Reduction list for runtime.
 // CHECK: [[RED_LIST:%.+]] = alloca [4 x i8*],
@@ -692,7 +731,7 @@ int main() {
 
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK2]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(160) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK2]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* {{.+}} %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(160) %{{.+}})
 
 // CHECK: [[ARRS_PRIV:%.+]] = alloca [10 x [4 x [[S_FLOAT_TY]]]],
 
@@ -882,7 +921,35 @@ int main() {
 
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK3]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [[S_FLOAT_TY]]*** dereferenceable(8) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK3]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i32* {{.*}} %{{.+}})
+
+// CHECK: [[VLA1_ORIG_ADDR:%.+]] = alloca i64
+// CHECK: [[VLA2_ORIG_ADDR:%.+]] = alloca i64
+// CHECK: [[ARR_ORIG_ADDR:%.+]] = alloca i32*,
+// CHECK: [[ARR_PRIV:%.+]] = alloca [1 x [2 x i32]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+// CHECK: [[VLA1:%.+]] = load i64, i64* [[VLA1_ORIG_ADDR]]
+// CHECK: [[VLA2:%.+]] = load i64, i64* [[VLA2_ORIG_ADDR]]
+// CHECK: [[ARR_ORIG:%.+]] = load i32*, i32** [[ARR_ORIG_ADDR]],
+
+// CHECK: [[LOW_OFFSET:%.+]] = mul nsw i64 1, [[VLA2]]
+// CHECK: [[ARRIDX:%.+]] = getelementptr inbounds i32, i32* [[ARR_ORIG]], i64 [[LOW_OFFSET]]
+// CHECK: [[LOW:%.+]] = getelementptr inbounds i32, i32* [[ARRIDX]], i64 0
+
+// CHECK: [[START:%.+]] = ptrtoint i32* [[ARR_ORIG]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint i32* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (i32* getelementptr (i32, i32* null, i32 1) to i64)
+// CHECK: [[ARR_PRIV_PTR:%.+]] = bitcast [1 x [2 x i32]]* [[ARR_PRIV]] to i32*
+// CHECK: [[ARR_PRIV:%.+]] = getelementptr i32, i32* [[ARR_PRIV_PTR]], i64 [[OFFSET]]
+
+// CHECK: ret void
+
+// CHECK: define internal void [[MAIN_MICROTASK4]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [[S_FLOAT_TY]]*** dereferenceable(8) %{{.+}})
 
 // CHECK: [[VAR2_ORIG_ADDR:%.+]] = alloca [[S_FLOAT_TY]]***,
 
@@ -910,44 +977,186 @@ int main() {
 // CHECK: store [[S_FLOAT_TY]]* [[PSEUDO_VAR2_PRIV]], [[S_FLOAT_TY]]** [[REF]]
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK4]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x [[S_FLOAT_TY]]]* dereferenceable(8) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK5]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [[S_FLOAT_TY]]*** dereferenceable(8) %{{.+}})
 
-// CHECK: [[VVAR2_ORIG_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR2_ORIG_ADDR:%.+]] = alloca [[S_FLOAT_TY]]***,
+// CHECK: [[VAR2_PRIV:%.+]] = alloca [1 x [6 x [[S_FLOAT_TY]]]],
 
 // Reduction list for runtime.
-// CHECK: [[RED_LIST:%.+]] = alloca [2 x i8*],
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
-// CHECK: [[VVAR2_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VVAR2_ORIG_ADDR]],
+// CHECK: [[VAR2_ORIG:%.+]] = load [[S_FLOAT_TY]]***, [[S_FLOAT_TY]]**** [[VAR2_ORIG_ADDR]],
 
-// CHECK: [[LOW:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]], i64 0, i64 0
-// CHECK: [[LAST:%.+]] = ptrtoint [[S_FLOAT_TY]]* %{{.+}} to i64
-// CHECK: [[FIRST:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
-// CHECK: [[BYTE_DIF:%.+]] = sub i64 [[LAST]], [[FIRST]]
-// CHECK: [[DIF:%.+]] = sdiv exact i64 [[BYTE_DIF]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
-// CHECK: [[SIZE:%.+]] = add nuw i64 [[DIF]], 1
-// CHECK: call i8* @llvm.stacksave()
-// CHECK: [[VVAR2_PRIV:%.+]] = alloca [[S_FLOAT_TY]], i64 [[SIZE]],
-// CHECK: [[ORIG_START:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+// CHECK: [[ARRIDX:%.+]] = getelementptr inbounds [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[LD]], i64 1
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[ARRIDX]],
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[LD]], i64 1
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+// CHECK: [[ORIG_START:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[LD]],
 // CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
 // CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
 // CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
 // CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
-// CHECK: [[PSEUDO_VVAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VVAR2_PRIV]], i64 [[OFFSET]]
-// CHECK: [[VVAR2_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VVAR2_PRIV]] to [2 x [[S_FLOAT_TY]]]*
+// CHECK: [[VAR2_PRIV_PTR:%.+]] = bitcast [1 x [6 x [[S_FLOAT_TY]]]]* [[VAR2_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[VAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR2_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: store [[S_FLOAT_TY]]** [[REF:.+]], [[S_FLOAT_TY]]*** %
+// CHECK: store [[S_FLOAT_TY]]* [[VAR2_PRIV]], [[S_FLOAT_TY]]** [[REF]]
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK5]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x [[S_FLOAT_TY]]]* dereferenceable(8) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK6]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [[S_FLOAT_TY]]*** dereferenceable(8) %{{.+}})
 
-// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR2_ORIG_ADDR:%.+]] = alloca [[S_FLOAT_TY]]***,
+// CHECK: [[VAR2_PRIV:%.+]] = alloca [1 x [6 x [[S_FLOAT_TY]]]],
 
 // Reduction list for runtime.
-// CHECK: [[RED_LIST:%.+]] = alloca [2 x i8*],
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+// CHECK: [[VAR2_ORIG:%.+]] = load [[S_FLOAT_TY]]***, [[S_FLOAT_TY]]**** [[VAR2_ORIG_ADDR]],
+
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+// CHECK: [[ARRIDX:%.+]] = getelementptr inbounds [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[LD]], i64 1
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[ARRIDX]],
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[LD]], i64 1
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+// CHECK: [[ORIG_START:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[LD]],
+// CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
+// CHECK: [[VAR2_PRIV_PTR:%.+]] = bitcast [1 x [6 x [[S_FLOAT_TY]]]]* [[VAR2_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[VAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR2_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: store [[S_FLOAT_TY]]** [[REF:.+]], [[S_FLOAT_TY]]*** %
+// CHECK: store [[S_FLOAT_TY]]* [[VAR2_PRIV]], [[S_FLOAT_TY]]** [[REF]]
+// CHECK: ret void
+
+// CHECK: define internal void [[MAIN_MICROTASK7]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [[S_FLOAT_TY]]*** dereferenceable(8) %{{.+}})
+
+// CHECK: [[VAR2_ORIG_ADDR:%.+]] = alloca [[S_FLOAT_TY]]***,
+// CHECK: [[VAR2_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+// CHECK: [[VAR2_ORIG:%.+]] = load [[S_FLOAT_TY]]***, [[S_FLOAT_TY]]**** [[VAR2_ORIG_ADDR]],
+
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+// CHECK: [[ARRIDX:%.+]] = getelementptr inbounds [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[LD]], i64 1
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[ARRIDX]],
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[LD]], i64 1
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+
+// CHECK: [[LD:%.+]] = load [[S_FLOAT_TY]]**, [[S_FLOAT_TY]]*** [[VAR2_ORIG]],
+// CHECK: [[ORIG_START:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[LD]],
+// CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
+// CHECK: [[PSEUDO_VAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR2_PRIV]], i64 [[OFFSET]]
+// CHECK: store [[S_FLOAT_TY]]** [[REF:.+]], [[S_FLOAT_TY]]*** %
+// CHECK: store [[S_FLOAT_TY]]* [[PSEUDO_VAR2_PRIV]], [[S_FLOAT_TY]]** [[REF]]
+// CHECK: ret void
+
+// CHECK: define internal void [[MAIN_MICROTASK8]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [5 x [[S_FLOAT_TY]]]* dereferenceable(20) %{{.+}})
+
+// CHECK: [[VVAR2_ORIG_ADDR:%.+]] = alloca [5 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VVAR2_PRIV:%.+]] = alloca [5 x [[S_FLOAT_TY]]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+// CHECK: [[VVAR2_ORIG:%.+]] = load [5 x [[S_FLOAT_TY]]]*, [5 x [[S_FLOAT_TY]]]** [[VVAR2_ORIG_ADDR]],
+
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [5 x [[S_FLOAT_TY]]], [5 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]], i64 0, i64 0
+// CHECK: [[ORIG_START:%.+]] = bitcast [5 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
+// CHECK: [[VVAR2_PRIV_PTR:%.+]] = bitcast [5 x [[S_FLOAT_TY]]]* [[VVAR2_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[VVAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VVAR2_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: ret void
+
+// CHECK: define internal void [[MAIN_MICROTASK9]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [4 x [[S_FLOAT_TY]]]* dereferenceable(16) %{{.+}})
+
+// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [4 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR3_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [4 x [[S_FLOAT_TY]]], [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], i64 0, i64 1
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: [[ORIG_START:%.+]] = bitcast [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
+// CHECK: [[VAR3_PRIV_PTR:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[PSEUDO_VAR3_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR3_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: [[VAR3_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VAR3_PRIV]] to [4 x [[S_FLOAT_TY]]]*
+
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [4 x [[S_FLOAT_TY]]]** %
+
+// CHECK: ret void
+
+// CHECK: define internal void [[MAIN_MICROTASK10]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [4 x [[S_FLOAT_TY]]]* dereferenceable(16) %{{.+}})
+
+// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [4 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR3_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [4 x [[S_FLOAT_TY]]], [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], i64 0, i64 0
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: [[ORIG_START:%.+]] = bitcast [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
+// CHECK: [[VAR3_PRIV_PTR:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[PSEUDO_VAR3_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR3_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: [[VAR3_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VAR3_PRIV]] to [4 x [[S_FLOAT_TY]]]*
+
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [4 x [[S_FLOAT_TY]]]** %
+
+// CHECK: ret void
+
+// CHECK: define internal void [[MAIN_MICROTASK11]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [4 x [[S_FLOAT_TY]]]* dereferenceable(16) %{{.+}})
+
+// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [4 x [[S_FLOAT_TY]]]*,
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [2 x i8*],
+
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
 
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
 // CHECK: [[LAST:%.+]] = ptrtoint [[S_FLOAT_TY]]* %{{.+}} to i64
 // CHECK: [[FIRST:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW:%.+]] to i64
 // CHECK: [[BYTE_DIF:%.+]] = sub i64 [[LAST]], [[FIRST]]
@@ -955,44 +1164,47 @@ int main() {
 // CHECK: [[SIZE:%.+]] = add nuw i64 [[DIF]], 1
 // CHECK: call i8* @llvm.stacksave()
 // CHECK: [[VAR3_PRIV:%.+]] = alloca [[S_FLOAT_TY]], i64 [[SIZE]],
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: [[ORIG_START:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: [[ORIG_START:%.+]] = bitcast [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
 // CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
 // CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
 // CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
 // CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (float* getelementptr (float, float* null, i32 1) to i64)
 // CHECK: [[PSEUDO_VAR3_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR3_PRIV]], i64 [[OFFSET]]
-// CHECK: [[VAR3_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VAR3_PRIV]] to [2 x [[S_FLOAT_TY]]]*
+// CHECK: [[VAR3_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VAR3_PRIV]] to [4 x [[S_FLOAT_TY]]]*
 
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [2 x [[S_FLOAT_TY]]]** %
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [4 x [[S_FLOAT_TY]]]** %
 
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK6]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x [[S_FLOAT_TY]]]* dereferenceable(8) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK12]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [4 x [[S_FLOAT_TY]]]* dereferenceable(16) %{{.+}})
 
-// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
-// CHECK: [[VAR3_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [4 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR3_PRIV:%.+]] = alloca [4 x [[S_FLOAT_TY]]],
 
 // Reduction list for runtime.
 // CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], i32 0, i32 0
-// CHECK: getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* %{{.+}}, i64 2
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: getelementptr inbounds [4 x [[S_FLOAT_TY]]], [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], i32 0, i32 0
+// CHECK: getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* %{{.+}}, i64 4
 
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [2 x [[S_FLOAT_TY]]]** %
-// CHECK: bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [4 x [[S_FLOAT_TY]]]** %
+// CHECK: bitcast [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
 
 // CHECK: ret void
 
-// CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]()
+// CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT_42]]()
 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
 // CHECK: call {{.*}} [[S_INT_TY_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i32*, [[S_INT_TY]]*, [[S_INT_TY]]*, i32*, [2 x i32]*, [2 x [[S_INT_TY]]]*)* [[TMAIN_MICROTASK:@.+]] to void
+// Not interested in this one:
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [42 x [[S_INT_TY]]]*, [2 x i32]*, i32*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK2:@.+]] to void
 // CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: ret
 //
@@ -1002,6 +1214,7 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
 // CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_INT_TY]],
@@ -1226,5 +1439,28 @@ int main() {
 // CHECK: store i{{[0-9]+}} [[UP]], i{{[0-9]+}}* [[T_VAR1_LHS]],
 // CHECK: ret void
 
+// CHECK: define internal void [[TMAIN_MICROTASK2]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [42 x [[S_INT_TY]]]* dereferenceable(168) %{{.*}}, [2 x i32]* dereferenceable(8) %{{.*}}, i32* dereferenceable(4) %{{.*}}, [2 x [[S_INT_TY]]]* dereferenceable(8) %{{.*}}, [[S_INT_TY]]* dereferenceable(4) %{{.*}})
+
+// CHECK: [[ARR_ORIG_ADDR:%.+]] = alloca [42 x [[S_INT_TY]]]*,
+// CHECK: [[ARR_PRIV:%.+]] = alloca [40 x [[S_INT_TY]]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+
+// CHECK: [[ARR_ORIG:%.+]] = load [42 x [[S_INT_TY]]]*, [42 x [[S_INT_TY]]]** [[ARR_ORIG_ADDR]],
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [42 x [[S_INT_TY]]], [42 x [[S_INT_TY]]]* [[ARR_ORIG]], i64 0, i64 1
+// CHECK: [[ORIG_START:%.+]] = bitcast [42 x [[S_INT_TY]]]* [[ARR_ORIG]] to [[S_INT_TY]]*
+// CHECK: [[START:%.+]] = ptrtoint [[S_INT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_INT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint (i32* getelementptr (i32, i32* null, i32 1) to i64)
+// CHECK: [[ARR_PRIV_PTR:%.+]] = bitcast [40 x [[S_INT_TY]]]* [[ARR_PRIV]] to [[S_INT_TY]]*
+// CHECK: [[PSEUDO_ARR_PRIV:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[ARR_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: [[ARR_PRIV:%.+]] = bitcast [[S_INT_TY]]* [[PSEUDO_ARR_PRIV]] to [42 x [[S_INT_TY]]]*
+
+// CHECK: ret void
+
 #endif
 
diff --git a/test/OpenMP/for_reduction_codegen_UDR.cpp b/test/OpenMP/for_reduction_codegen_UDR.cpp
index 95b04f44d468..05a46a6eb7ff 100644
--- a/test/OpenMP/for_reduction_codegen_UDR.cpp
+++ b/test/OpenMP/for_reduction_codegen_UDR.cpp
@@ -40,7 +40,7 @@ void init_plus(BaseS1&, const BaseS1&);
 // CHECK-DAG: [[REDUCTION_LOCK:@.+]] = common global [8 x i32] zeroinitializer
 
 #pragma omp declare reduction(operator&& : int : omp_out = 111 & omp_in)
-template <typename T>
+template <typename T, int length>
 T tmain() {
   T t;
   S<T> test;
@@ -49,6 +49,7 @@ T tmain() {
   S<T> s_arr[] = {1, 2};
   S<T> &var = test;
   S<T> var1;
+  S<T> arr[length];
 #pragma omp declare reduction(operator& : T : omp_out = 15 + omp_in)
 #pragma omp declare reduction(operator+ : T : omp_out = 1513 + omp_in) initializer(omp_priv = 321)
 #pragma omp declare reduction(min : T : omp_out = 47 - omp_in) initializer(omp_priv = 432 / omp_orig)
@@ -66,6 +67,12 @@ T tmain() {
     vec[i] = t_var;
     s_arr[i] = var;
   }
+#pragma omp parallel
+#pragma omp for reduction(+ : arr[1:length-2])
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
   return T();
 }
 
@@ -78,12 +85,12 @@ int main() {
   S<float> test;
   float t_var = 0, t_var1;
   int vec[] = {1, 2};
-  S<float> s_arr[] = {1, 2};
+  S<float> s_arr[] = {1, 2, 3, 4};
   S<float> &var = test;
   S<float> var1, arrs[10][4];
   S<float> **var2 = foo();
-  S<float> vvar2[2];
-  S<float>(&var3)[2] = s_arr;
+  S<float> vvar2[5];
+  S<float>(&var3)[4] = s_arr;
 #pragma omp declare reduction(operator+ : int : omp_out = 555 * omp_in) initializer(omp_priv = 888)
 #pragma omp parallel
 #pragma omp for reduction(+ : t_var) reduction(& : var) reduction(&& : var1) reduction(min : t_var1)
@@ -115,24 +122,24 @@ int main() {
 #pragma omp for reduction(& : var3)
   for (int i = 0; i < 10; ++i)
     ;
-  return tmain<int>();
+  return tmain<int, 42>();
 }
 
 // CHECK: define {{.*}}i{{[0-9]+}} @main()
 // CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: call {{.*}} [[S_FLOAT_TY_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, float*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float*, [2 x i32]*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, float*, [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]*, float*, [2 x i32]*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK:@.+]] to void
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, i64, i32*, [2 x i32]*, [10 x [4 x [[S_FLOAT_TY]]]]*)* [[MAIN_MICROTASK1:@.+]] to void
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, i64, i32*, [10 x [4 x [[S_FLOAT_TY]]]]*)* [[MAIN_MICROTASK2:@.+]] to void
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[S_FLOAT_TY]]***)* [[MAIN_MICROTASK3:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK4:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK5:@.+]] to void
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [2 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK6:@.+]] to void
-// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT:@.+]]()
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [5 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK4:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK5:@.+]] to void
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [4 x [[S_FLOAT_TY]]]*)* [[MAIN_MICROTASK6:@.+]] to void
+// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT_42:@.+]]()
 // CHECK: call {{.*}} [[S_FLOAT_TY_DESTR:@.+]]([[S_FLOAT_TY]]*
 // CHECK: ret
 //
-// CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float* dereferenceable(4) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(12) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(12) %{{.+}}, float* dereferenceable(4) %{{.+}}, [2 x i32]* dereferenceable(8) %vec, [2 x [[S_FLOAT_TY]]]* dereferenceable(24) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float* dereferenceable(4) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(12) %{{.+}}, [[S_FLOAT_TY]]* dereferenceable(12) %{{.+}}, float* dereferenceable(4) %{{.+}}, [2 x i32]* dereferenceable(8) %vec, [4 x [[S_FLOAT_TY]]]* dereferenceable(48) %{{.+}})
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca float,
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
@@ -300,7 +307,7 @@ int main() {
 // CHECK: fadd float 5.550000e+02, %
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* %{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(480) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK1]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* {{.+}} %{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(480) %{{.+}})
 
 // Reduction list for runtime.
 // CHECK: [[RED_LIST:%.+]] = alloca [4 x i8*],
@@ -496,7 +503,7 @@ int main() {
 
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK2]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(480) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK2]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 %{{.+}}, i64 %{{.+}}, i32* {{.+}} %{{.+}}, [10 x [4 x [[S_FLOAT_TY]]]]* dereferenceable(480) %{{.+}})
 
 // CHECK: [[ARRS_PRIV:%.+]] = alloca [10 x [4 x [[S_FLOAT_TY]]]],
 
@@ -711,89 +718,87 @@ int main() {
 // CHECK: store [[S_FLOAT_TY]]* [[PSEUDO_VAR2_PRIV]], [[S_FLOAT_TY]]** [[REF]]
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK4]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x [[S_FLOAT_TY]]]* dereferenceable(24) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK4]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [5 x [[S_FLOAT_TY]]]* dereferenceable(60) %{{.+}})
 
-// CHECK: [[VVAR2_ORIG_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VVAR2_ORIG_ADDR:%.+]] = alloca [5 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VVAR2_PRIV:%.+]] = alloca [5 x [[S_FLOAT_TY]]],
 
 // Reduction list for runtime.
-// CHECK: [[RED_LIST:%.+]] = alloca [2 x i8*],
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
-// CHECK: [[VVAR2_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VVAR2_ORIG_ADDR]],
+// CHECK: [[VVAR2_ORIG:%.+]] = load [5 x [[S_FLOAT_TY]]]*, [5 x [[S_FLOAT_TY]]]** [[VVAR2_ORIG_ADDR]],
 
-// CHECK: [[LOW:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]], i64 0, i64 0
-// CHECK: [[LAST:%.+]] = ptrtoint [[S_FLOAT_TY]]* %{{.+}} to i64
-// CHECK: [[FIRST:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
-// CHECK: [[BYTE_DIF:%.+]] = sub i64 [[LAST]], [[FIRST]]
-// CHECK: [[DIF:%.+]] = sdiv exact i64 [[BYTE_DIF]], ptrtoint ([[S_FLOAT_TY]]* getelementptr ([[S_FLOAT_TY]], [[S_FLOAT_TY]]* null, i32 1) to i64)
-// CHECK: [[SIZE:%.+]] = add nuw i64 [[DIF]], 1
-// CHECK: call i8* @llvm.stacksave()
-// CHECK: [[VVAR2_PRIV:%.+]] = alloca [[S_FLOAT_TY]], i64 [[SIZE]],
-// CHECK: [[ORIG_START:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [5 x [[S_FLOAT_TY]]], [5 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]], i64 0, i64 0
+// CHECK: [[ORIG_START:%.+]] = bitcast [5 x [[S_FLOAT_TY]]]* [[VVAR2_ORIG]] to [[S_FLOAT_TY]]*
 // CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
 // CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
 // CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
 // CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint ([[S_FLOAT_TY]]* getelementptr ([[S_FLOAT_TY]], [[S_FLOAT_TY]]* null, i32 1) to i64)
-// CHECK: [[PSEUDO_VVAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VVAR2_PRIV]], i64 [[OFFSET]]
-// CHECK: [[VVAR2_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VVAR2_PRIV]] to [2 x [[S_FLOAT_TY]]]*
+// CHECK: [[VVAR2_PRIV_PTR:%.+]] = bitcast [5 x [[S_FLOAT_TY]]]* [[VVAR2_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[VVAR2_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VVAR2_PRIV_PTR]], i64 [[OFFSET]]
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK5]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x [[S_FLOAT_TY]]]* dereferenceable(24) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK5]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [4 x [[S_FLOAT_TY]]]* dereferenceable(48) %{{.+}})
 
-// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [4 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR3_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
 
 // Reduction list for runtime.
-// CHECK: [[RED_LIST:%.+]] = alloca [2 x i8*],
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
-// CHECK: [[LAST:%.+]] = ptrtoint [[S_FLOAT_TY]]* %{{.+}} to i64
-// CHECK: [[FIRST:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW:%.+]] to i64
-// CHECK: [[BYTE_DIF:%.+]] = sub i64 [[LAST]], [[FIRST]]
-// CHECK: [[DIF:%.+]] = sdiv exact i64 [[BYTE_DIF]], ptrtoint ([[S_FLOAT_TY]]* getelementptr ([[S_FLOAT_TY]], [[S_FLOAT_TY]]* null, i32 1) to i64)
-// CHECK: [[SIZE:%.+]] = add nuw i64 [[DIF]], 1
-// CHECK: call i8* @llvm.stacksave()
-// CHECK: [[VAR3_PRIV:%.+]] = alloca [[S_FLOAT_TY]], i64 [[SIZE]],
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: [[ORIG_START:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [4 x [[S_FLOAT_TY]]], [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], i64 0, i64 1
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: [[ORIG_START:%.+]] = bitcast [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
 // CHECK: [[START:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[ORIG_START]] to i64
 // CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_FLOAT_TY]]* [[LOW]] to i64
 // CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
 // CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint ([[S_FLOAT_TY]]* getelementptr ([[S_FLOAT_TY]], [[S_FLOAT_TY]]* null, i32 1) to i64)
-// CHECK: [[PSEUDO_VAR3_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR3_PRIV]], i64 [[OFFSET]]
-// CHECK: [[VAR3_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VAR3_PRIV]] to [2 x [[S_FLOAT_TY]]]*
+// CHECK: [[VAR3_PRIV_PTR:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[PSEUDO_VAR3_PRIV:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[VAR3_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: [[VAR3_PRIV:%.+]] = bitcast [[S_FLOAT_TY]]* [[PSEUDO_VAR3_PRIV]] to [4 x [[S_FLOAT_TY]]]*
 
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [2 x [[S_FLOAT_TY]]]** %
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [4 x [[S_FLOAT_TY]]]** %
 
 // CHECK: ret void
 
-// CHECK: define internal void [[MAIN_MICROTASK6]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x [[S_FLOAT_TY]]]* dereferenceable(24) %{{.+}})
+// CHECK: define internal void [[MAIN_MICROTASK6]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [4 x [[S_FLOAT_TY]]]* dereferenceable(48) %{{.+}})
 
-// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
-// CHECK: [[VAR3_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR3_ORIG_ADDR:%.+]] = alloca [4 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR3_PRIV:%.+]] = alloca [4 x [[S_FLOAT_TY]]],
 
 // Reduction list for runtime.
 // CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
 
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
-// CHECK: [[VAR3_ORIG:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
-// CHECK: getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], i32 0, i32 0
-// CHECK: bitcast [2 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
-// CHECK: getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* %{{.+}}, i64 2
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]], [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR:%.+]],
+// CHECK: [[VAR3_ORIG:%.+]] = load [4 x [[S_FLOAT_TY]]]*, [4 x [[S_FLOAT_TY]]]** [[VAR3_ORIG_ADDR]],
+// CHECK: getelementptr inbounds [4 x [[S_FLOAT_TY]]], [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], i32 0, i32 0
+// CHECK: bitcast [4 x [[S_FLOAT_TY]]]* [[VAR3_ORIG]] to [[S_FLOAT_TY]]*
+// CHECK: getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* %{{.+}}, i64 4
 
-// CHECK: store [2 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [2 x [[S_FLOAT_TY]]]** %
+// CHECK: store [4 x [[S_FLOAT_TY]]]* [[VAR3_PRIV]], [4 x [[S_FLOAT_TY]]]** %
 
 // CHECK: ret void
 
-// CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]()
+// CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT_42]]()
 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
 // CHECK: call {{.*}} [[S_INT_TY_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 6, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i32*, [[S_INT_TY]]*, [[S_INT_TY]]*, i32*, [2 x i32]*, [2 x [[S_INT_TY]]]*)* [[TMAIN_MICROTASK:@.+]] to void
+// Not interested in this one:
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 5, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [42 x [[S_INT_TY]]]*, [2 x i32]*, i32*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK2:@.+]] to void
+// CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: ret
 //
@@ -803,6 +808,7 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
 // CHECK: [[VAR1_PRIV:%.+]] = alloca [[S_INT_TY]],
@@ -964,5 +970,28 @@ int main() {
 // CHECK: sub nsw i32 47, %
 // CHECK: ret void
 
+// CHECK: define internal void [[TMAIN_MICROTASK2]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [42 x [[S_INT_TY]]]* dereferenceable(504) %{{.*}}, [2 x i32]* dereferenceable(8) %{{.*}}, i32* dereferenceable(4) %{{.*}}, [2 x [[S_INT_TY]]]* dereferenceable(24) %{{.*}}, [[S_INT_TY]]* dereferenceable(12) %{{.*}})
+
+// CHECK: [[ARR_ORIG_ADDR:%.+]] = alloca [42 x [[S_INT_TY]]]*,
+// CHECK: [[ARR_PRIV:%.+]] = alloca [40 x [[S_INT_TY]]],
+
+// Reduction list for runtime.
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x i8*],
+
+// CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
+
+// CHECK: [[ARR_ORIG:%.+]] = load [42 x [[S_INT_TY]]]*, [42 x [[S_INT_TY]]]** [[ARR_ORIG_ADDR]],
+// CHECK: [[LOW:%.+]] = getelementptr inbounds [42 x [[S_INT_TY]]], [42 x [[S_INT_TY]]]* [[ARR_ORIG]], i64 0, i64 1
+// CHECK: [[ORIG_START:%.+]] = bitcast [42 x [[S_INT_TY]]]* [[ARR_ORIG]] to [[S_INT_TY]]*
+// CHECK: [[START:%.+]] = ptrtoint [[S_INT_TY]]* [[ORIG_START]] to i64
+// CHECK: [[LOW_BOUND:%.+]] = ptrtoint [[S_INT_TY]]* [[LOW]] to i64
+// CHECK: [[OFFSET_BYTES:%.+]] = sub i64 [[START]], [[LOW_BOUND]]
+// CHECK: [[OFFSET:%.+]] = sdiv exact i64 [[OFFSET_BYTES]], ptrtoint ([[S_INT_TY]]* getelementptr ([[S_INT_TY]], [[S_INT_TY]]* null, i32 1) to i64)
+// CHECK: [[ARR_PRIV_PTR:%.+]] = bitcast [40 x [[S_INT_TY]]]* [[ARR_PRIV]] to [[S_INT_TY]]*
+// CHECK: [[PSEUDO_ARR_PRIV:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[ARR_PRIV_PTR]], i64 [[OFFSET]]
+// CHECK: [[ARR_PRIV:%.+]] = bitcast [[S_INT_TY]]* [[PSEUDO_ARR_PRIV]] to [42 x [[S_INT_TY]]]*
+
+// CHECK: ret void
+
 #endif
 
diff --git a/test/OpenMP/for_reduction_messages.cpp b/test/OpenMP/for_reduction_messages.cpp
index bb70ecc2b23e..d4f4a795fe2c 100644
--- a/test/OpenMP/for_reduction_messages.cpp
+++ b/test/OpenMP/for_reduction_messages.cpp
@@ -26,9 +26,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/for_simd_lastprivate_messages.cpp b/test/OpenMP/for_simd_lastprivate_messages.cpp
index d48c07d92b39..a176a2d70eb9 100644
--- a/test/OpenMP/for_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/for_simd_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   S2 &operator =(const S2&);
   const S2 &operator =(const S2&) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/for_simd_reduction_messages.cpp b/test/OpenMP/for_simd_reduction_messages.cpp
index 485070e758ae..a77ab754eff9 100644
--- a/test/OpenMP/for_simd_reduction_messages.cpp
+++ b/test/OpenMP/for_simd_reduction_messages.cpp
@@ -26,9 +26,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/is_initial_device.c b/test/OpenMP/is_initial_device.c
new file mode 100644
index 000000000000..0521f0bd5393
--- /dev/null
+++ b/test/OpenMP/is_initial_device.c
@@ -0,0 +1,36 @@
+// REQUIRES: powerpc-registered-target
+
+// RUN: %clang_cc1 -verify -fopenmp -x c -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-unknown-unknown \
+// RUN:            -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x ir -triple powerpc64le-unknown-unknown -emit-llvm \
+// RUN:             %t-ppc-host.bc -o - | FileCheck %s -check-prefixes HOST,OUTLINED
+// RUN: %clang_cc1 -verify -fopenmp -x c -triple powerpc64le-unknown-unknown -emit-llvm -fopenmp-is-device \
+// RUN:             %s -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefixes DEVICE,OUTLINED
+
+// expected-no-diagnostics
+int check() {
+  int host = omp_is_initial_device();
+  int device;
+#pragma omp target map(tofrom: device)
+  {
+    device = omp_is_initial_device();
+  }
+
+  return host + device;
+}
+
+// The host should get a value of 1:
+// HOST: define{{.*}} @check()
+// HOST: [[HOST:%.*]] = alloca i32
+// HOST: store i32 1, i32* [[HOST]]
+
+// OUTLINED: define{{.*}} @{{.*}}omp_offloading{{.*}}(i32*{{.*}} [[DEVICE_ARGUMENT:%.*]])
+// OUTLINED: [[DEVICE_ADDR_STORAGE:%.*]] = alloca i32*
+// OUTLINED: store i32* [[DEVICE_ARGUMENT]], i32** [[DEVICE_ADDR_STORAGE]]
+// OUTLINED: [[DEVICE_ADDR:%.*]] = load i32*, i32** [[DEVICE_ADDR_STORAGE]]
+
+// The outlined function that is called as fallback also runs on the host:
+// HOST: store i32 1, i32* [[DEVICE_ADDR]]
+
+// The device should get a value of 0:
+// DEVICE: store i32 0, i32* [[DEVICE_ADDR]]
diff --git a/test/OpenMP/nvptx_data_sharing.cpp b/test/OpenMP/nvptx_data_sharing.cpp
new file mode 100644
index 000000000000..65215cd2c4f9
--- /dev/null
+++ b/test/OpenMP/nvptx_data_sharing.cpp
@@ -0,0 +1,57 @@
+// Test device data sharing codegen.
+///==========================================================================///
+
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix CK1
+
+// expected-no-diagnostics
+
+#ifndef HEADER
+#define HEADER
+
+void test_ds(){
+  #pragma omp target
+  {
+    int a = 10;
+    #pragma omp parallel
+    {
+       a = 1000;
+    }
+  }
+}
+
+/// ========= In the worker function ========= ///
+
+// CK1: define internal void @__omp_offloading_{{.*}}test_ds{{.*}}worker() [[ATTR1:#.*]] {
+// CK1: [[SHAREDARGS:%.+]] = alloca i8**
+// CK1: call i1 @__kmpc_kernel_parallel(i8** %work_fn, i8*** [[SHAREDARGS]])
+// CK1: [[SHARGSTMP:%.+]] = load i8**, i8*** [[SHAREDARGS]]
+// CK1: call void @__omp_outlined___wrapper{{.*}}({{.*}}, i8** [[SHARGSTMP]])
+
+/// ========= In the kernel function ========= ///
+
+// CK1: {{.*}}define void @__omp_offloading{{.*}}test_ds{{.*}}() [[ATTR2:#.*]] {
+// CK1: [[SHAREDARGS1:%.+]] = alloca i8**
+// CK1: call void @__kmpc_kernel_prepare_parallel({{.*}}, i8*** [[SHAREDARGS1]], i32 1)
+// CK1: [[SHARGSTMP1:%.+]] = load i8**, i8*** [[SHAREDARGS1]]
+// CK1: [[SHARGSTMP2:%.+]] = getelementptr inbounds i8*, i8** [[SHARGSTMP1]]
+// CK1: [[SHAREDVAR:%.+]] = bitcast i32* {{.*}} to i8*
+// CK1: store i8* [[SHAREDVAR]], i8** [[SHARGSTMP2]]
+
+/// ========= In the data sharing wrapper function ========= ///
+
+// CK1: {{.*}}define internal void @__omp_outlined___wrapper({{.*}}i8**) [[ATTR1]] {
+// CK1: [[SHAREDARGS2:%.+]] = alloca i8**
+// CK1: store i8** %2, i8*** [[SHAREDARGS2]]
+// CK1: [[SHARGSTMP3:%.+]] = load i8**, i8*** [[SHAREDARGS2]]
+// CK1: [[SHARGSTMP4:%.+]] = getelementptr inbounds i8*, i8** [[SHARGSTMP3]]
+// CK1: [[SHARGSTMP5:%.+]] = bitcast i8** [[SHARGSTMP4]] to i32**
+// CK1: [[SHARGSTMP6:%.+]] = load i32*, i32** [[SHARGSTMP5]]
+// CK1: call void @__omp_outlined__({{.*}}, i32* [[SHARGSTMP6]])
+
+/// ========= Attributes ========= ///
+
+// CK1-NOT: attributes [[ATTR1]] = { {{.*}}"has-nvptx-shared-depot"{{.*}} }
+// CK1: attributes [[ATTR2]] = { {{.*}}"has-nvptx-shared-depot"{{.*}} }
+
+#endif
diff --git a/test/OpenMP/nvptx_parallel_codegen.cpp b/test/OpenMP/nvptx_parallel_codegen.cpp
index 224f24569669..ba889bf4975f 100644
--- a/test/OpenMP/nvptx_parallel_codegen.cpp
+++ b/test/OpenMP/nvptx_parallel_codegen.cpp
@@ -78,7 +78,7 @@ int bar(int n){
   //
   // CHECK: [[AWAIT_WORK]]
   // CHECK: call void @llvm.nvvm.barrier0()
-  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]])
+  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]],
   // CHECK: [[KPRB:%.+]] = zext i1 [[KPR]] to i8
   // store i8 [[KPRB]], i8* [[OMP_EXEC_STATUS]], align 1
   // CHECK: [[WORK:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
@@ -92,20 +92,20 @@ int bar(int n){
   //
   // CHECK: [[EXEC_PARALLEL]]
   // CHECK: [[WF1:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
-  // CHECK: [[WM1:%.+]] = icmp eq i8* [[WF1]], bitcast (void (i32*, i32*)* [[PARALLEL_FN1:@.+]] to i8*)
+  // CHECK: [[WM1:%.+]] = icmp eq i8* [[WF1]], bitcast (void (i16, i32, i8**)* [[PARALLEL_FN1:@.+]]_wrapper to i8*)
   // CHECK: br i1 [[WM1]], label {{%?}}[[EXEC_PFN1:.+]], label {{%?}}[[CHECK_NEXT1:.+]]
   //
   // CHECK: [[EXEC_PFN1]]
-  // CHECK: call void [[PARALLEL_FN1]](
+  // CHECK: call void [[PARALLEL_FN1]]_wrapper(
   // CHECK: br label {{%?}}[[TERM_PARALLEL:.+]]
   //
   // CHECK: [[CHECK_NEXT1]]
   // CHECK: [[WF2:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
-  // CHECK: [[WM2:%.+]] = icmp eq i8* [[WF2]], bitcast (void (i32*, i32*)* [[PARALLEL_FN2:@.+]] to i8*)
+  // CHECK: [[WM2:%.+]] = icmp eq i8* [[WF2]], bitcast (void (i16, i32, i8**)* [[PARALLEL_FN2:@.+]]_wrapper to i8*)
   // CHECK: br i1 [[WM2]], label {{%?}}[[EXEC_PFN2:.+]], label {{%?}}[[CHECK_NEXT2:.+]]
   //
   // CHECK: [[EXEC_PFN2]]
-  // CHECK: call void [[PARALLEL_FN2]](
+  // CHECK: call void [[PARALLEL_FN2]]_wrapper(
   // CHECK: br label {{%?}}[[TERM_PARALLEL:.+]]
   //
   // CHECK: [[CHECK_NEXT2]]
@@ -152,13 +152,13 @@ int bar(int n){
   // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize()
   // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]]
   // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]]
-  // CHECK: call void @__kmpc_kernel_prepare_parallel(i8* bitcast (void (i32*, i32*)* [[PARALLEL_FN1]] to i8*))
+  // CHECK: call void @__kmpc_kernel_prepare_parallel(i8* bitcast (void (i16, i32, i8**)* [[PARALLEL_FN1]]_wrapper to i8*),
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: call void @__kmpc_serialized_parallel(
   // CHECK: {{call|invoke}} void [[PARALLEL_FN3:@.+]](
   // CHECK: call void @__kmpc_end_serialized_parallel(
-  // CHECK: call void @__kmpc_kernel_prepare_parallel(i8* bitcast (void (i32*, i32*)* [[PARALLEL_FN2]] to i8*))
+  // CHECK: call void @__kmpc_kernel_prepare_parallel(i8* bitcast (void (i16, i32, i8**)* [[PARALLEL_FN2]]_wrapper to i8*),
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK-64-DAG: load i32, i32* [[REF_A]]
@@ -166,7 +166,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -203,7 +203,7 @@ int bar(int n){
   //
   // CHECK: [[AWAIT_WORK]]
   // CHECK: call void @llvm.nvvm.barrier0()
-  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]])
+  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]],
   // CHECK: [[KPRB:%.+]] = zext i1 [[KPR]] to i8
   // store i8 [[KPRB]], i8* [[OMP_EXEC_STATUS]], align 1
   // CHECK: [[WORK:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
@@ -217,11 +217,11 @@ int bar(int n){
   //
   // CHECK: [[EXEC_PARALLEL]]
   // CHECK: [[WF:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
-  // CHECK: [[WM:%.+]] = icmp eq i8* [[WF]], bitcast (void (i32*, i32*)* [[PARALLEL_FN4:@.+]] to i8*)
+  // CHECK: [[WM:%.+]] = icmp eq i8* [[WF]], bitcast (void (i16, i32, i8**)* [[PARALLEL_FN4:@.+]]_wrapper to i8*)
   // CHECK: br i1 [[WM]], label {{%?}}[[EXEC_PFN:.+]], label {{%?}}[[CHECK_NEXT:.+]]
   //
   // CHECK: [[EXEC_PFN]]
-  // CHECK: call void [[PARALLEL_FN4]](
+  // CHECK: call void [[PARALLEL_FN4]]_wrapper(
   // CHECK: br label {{%?}}[[TERM_PARALLEL:.+]]
   //
   // CHECK: [[CHECK_NEXT]]
@@ -283,7 +283,7 @@ int bar(int n){
   // CHECK: br i1 [[CMP]], label {{%?}}[[IF_THEN:.+]], label {{%?}}[[IF_ELSE:.+]]
   //
   // CHECK: [[IF_THEN]]
-  // CHECK: call void @__kmpc_kernel_prepare_parallel(i8* bitcast (void (i32*, i32*)* [[PARALLEL_FN4]] to i8*))
+  // CHECK: call void @__kmpc_kernel_prepare_parallel(i8* bitcast (void (i16, i32, i8**)* [[PARALLEL_FN4]]_wrapper to i8*),
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[IF_END:.+]]
@@ -303,7 +303,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
diff --git a/test/OpenMP/nvptx_param_translate.c b/test/OpenMP/nvptx_param_translate.c
new file mode 100644
index 000000000000..ec123ce3811b
--- /dev/null
+++ b/test/OpenMP/nvptx_param_translate.c
@@ -0,0 +1,19 @@
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s
+// expected-no-diagnostics
+
+// CHECK: [[MAP_FN:%.+]] = load void (i8*, ...)*, void (i8*, ...)** %
+// CHECK: call void (i8*, ...) [[MAP_FN]](i8* %
+int main() {
+  double a, b;
+
+#pragma omp target map(tofrom      \
+                       : a) map(to \
+                                : b)
+  {
+#pragma omp taskgroup
+#pragma omp task shared(a)
+    a = b;
+  }
+  return 0;
+}
diff --git a/test/OpenMP/nvptx_target_codegen.cpp b/test/OpenMP/nvptx_target_codegen.cpp
index d7ce7dc7d106..23b40e10c4c8 100644
--- a/test/OpenMP/nvptx_target_codegen.cpp
+++ b/test/OpenMP/nvptx_target_codegen.cpp
@@ -91,7 +91,7 @@ int foo(int n) {
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -168,7 +168,7 @@ int foo(int n) {
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -278,7 +278,7 @@ int foo(int n) {
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -441,7 +441,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -531,7 +531,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -616,7 +616,7 @@ int bar(int n){
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
diff --git a/test/OpenMP/nvptx_target_firstprivate_codegen.cpp b/test/OpenMP/nvptx_target_firstprivate_codegen.cpp
index 5dcff8e54842..f750be03e052 100644
--- a/test/OpenMP/nvptx_target_firstprivate_codegen.cpp
+++ b/test/OpenMP/nvptx_target_firstprivate_codegen.cpp
@@ -1,15 +1,14 @@
-
 // Test target codegen - host bc file has to be created first.
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
-// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -debug-info-kind=limited -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc
-// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -debug-info-kind=limited -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
 // expected-no-diagnostics
 #ifndef HEADER
 #define HEADER
 
-template<typename tx, typename ty>
-struct TT{
+template <typename tx, typename ty>
+struct TT {
   tx X;
   ty Y;
 };
@@ -23,29 +22,32 @@ int foo(int n, double *ptr) {
   float b[10];
   double c[5][10];
   TT<long long, char> d;
-  
-  #pragma omp target firstprivate(a)
+
+#pragma omp target firstprivate(a) map(tofrom \
+                                       : b)
   {
+    b[a] = a;
   }
-  
-  // TCHECK:  define void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A_IN:%.+]])
+
+  // TCHECK:  define {{.*}}void @__omp_offloading_{{.+}}([10 x float] addrspace(1)* noalias [[B_IN:%.+]], i{{[0-9]+}} [[A_IN:%.+]])
   // TCHECK:  [[A_ADDR:%.+]] = alloca i{{[0-9]+}},
   // TCHECK-NOT:  alloca i{{[0-9]+}},
+  // TCHECK-64:  call void @llvm.dbg.declare(metadata [10 x float] addrspace(1)** %{{.+}}, metadata !{{[0-9]+}}, metadata !DIExpression())
   // TCHECK:  store i{{[0-9]+}} [[A_IN]], i{{[0-9]+}}* [[A_ADDR]],
-  // TCHECK:  ret void  
+  // TCHECK:  ret void
 
-#pragma omp target firstprivate(aa,b,c,d)
+#pragma omp target firstprivate(aa, b, c, d)
   {
     aa += 1;
     b[2] = 1.0;
     c[1][2] = 1.0;
     d.X = 1;
-    d.Y = 1;    
+    d.Y = 1;
   }
-  
+
   // make sure that firstprivate variables are generated in all cases and that we use those instances for operations inside the
   // target region
-  // TCHECK:  define void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A2_IN:%.+]], [10 x float]* {{.+}} [[B_IN:%.+]], [5 x [10 x double]]* {{.+}} [[C_IN:%.+]], [[TT]]* {{.+}} [[D_IN:%.+]])
+  // TCHECK:  define {{.*}}void @__omp_offloading_{{.+}}(i{{[0-9]+}}{{.*}} [[A2_IN:%.+]], [10 x float]*{{.*}} [[B_IN:%.+]], [5 x [10 x double]]*{{.*}} [[C_IN:%.+]], [[TT]]*{{.*}} [[D_IN:%.+]])
   // TCHECK:  [[A2_ADDR:%.+]] = alloca i{{[0-9]+}},
   // TCHECK:  [[B_ADDR:%.+]] = alloca [10 x float]*,
   // TCHECK:  [[C_ADDR:%.+]] = alloca [5 x [10 x double]]*,
@@ -58,10 +60,12 @@ int foo(int n, double *ptr) {
   // TCHECK:  store [10 x float]* [[B_IN]], [10 x float]** [[B_ADDR]],
   // TCHECK:  store [5 x [10 x double]]* [[C_IN]], [5 x [10 x double]]** [[C_ADDR]],
   // TCHECK:  store [[TT]]* [[D_IN]], [[TT]]** [[D_ADDR]],
-  // TCHECK:  [[CONV_A2ADDR:%.+]] = bitcast i{{[0-9]+}}* [[A2_ADDR]] to i{{[0-9]+}}*
   // TCHECK:  [[B_ADDR_REF:%.+]] = load [10 x float]*, [10 x float]** [[B_ADDR]],
+  // TCHECK:  [[B_ADDR_REF:%.+]] = load [10 x float]*, [10 x float]** %
   // TCHECK:  [[C_ADDR_REF:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[C_ADDR]],
+  // TCHECK:  [[C_ADDR_REF:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** %
   // TCHECK:  [[D_ADDR_REF:%.+]] = load [[TT]]*, [[TT]]** [[D_ADDR]],
+  // TCHECK:  [[D_ADDR_REF:%.+]] = load [[TT]]*, [[TT]]** %
 
   // firstprivate(aa): a_priv = a_in
 
@@ -74,16 +78,15 @@ int foo(int n, double *ptr) {
   // TCHECK:  [[C_PRIV_BCAST:%.+]] = bitcast [5 x [10 x double]]* [[C_PRIV]] to i8*
   // TCHECK:  [[C_IN_BCAST:%.+]] = bitcast [5 x [10 x double]]* [[C_ADDR_REF]] to i8*
   // TCHECK:  call void @llvm.memcpy.{{.+}}(i8* [[C_PRIV_BCAST]], i8* [[C_IN_BCAST]],{{.+}})
-  
+
   // firstprivate(d)
   // TCHECK:  [[D_PRIV_BCAST:%.+]] = bitcast [[TT]]* [[D_PRIV]] to i8*
   // TCHECK:  [[D_IN_BCAST:%.+]] = bitcast [[TT]]* [[D_ADDR_REF]] to i8*
   // TCHECK:  call void @llvm.memcpy.{{.+}}(i8* [[D_PRIV_BCAST]], i8* [[D_IN_BCAST]],{{.+}})
 
-  // TCHECK: load i16, i16* [[CONV_A2ADDR]],
+  // TCHECK: load i16, i16* [[A2_ADDR]],
 
-  
-  #pragma omp target firstprivate(ptr)
+#pragma omp target firstprivate(ptr)
   {
     ptr[0]++;
   }
@@ -98,13 +101,12 @@ int foo(int n, double *ptr) {
   return a;
 }
 
-
-template<typename tx>
+template <typename tx>
 tx ftemplate(int n) {
   tx a = 0;
   tx b[10];
 
-#pragma omp target firstprivate(a,b)
+#pragma omp target firstprivate(a, b)
   {
     a += 1;
     b[2] += 1;
@@ -113,13 +115,12 @@ tx ftemplate(int n) {
   return a;
 }
 
-static
-int fstatic(int n) {
+static int fstatic(int n) {
   int a = 0;
   char aaa = 0;
   int b[10];
 
-#pragma omp target firstprivate(a,aaa,b)
+#pragma omp target firstprivate(a, aaa, b)
   {
     a += 1;
     aaa += 1;
@@ -129,7 +130,7 @@ int fstatic(int n) {
   return a;
 }
 
-// TCHECK: define void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A_IN:%.+]], i{{[0-9]+}} [[A3_IN:%.+]], [10 x i{{[0-9]+}}]*{{.+}} [[B_IN:%.+]])
+// TCHECK: define {{.*}}void @__omp_offloading_{{.+}}(i{{[0-9]+}}{{.*}} [[A_IN:%.+]], i{{[0-9]+}}{{.*}} [[A3_IN:%.+]], [10 x i{{[0-9]+}}]*{{.+}} [[B_IN:%.+]])
 // TCHECK:  [[A_ADDR:%.+]] = alloca i{{[0-9]+}},
 // TCHECK:  [[A3_ADDR:%.+]] = alloca i{{[0-9]+}},
 // TCHECK:  [[B_ADDR:%.+]] = alloca [10 x i{{[0-9]+}}]*,
@@ -138,9 +139,8 @@ int fstatic(int n) {
 // TCHECK:  store i{{[0-9]+}} [[A_IN]], i{{[0-9]+}}* [[A_ADDR]],
 // TCHECK:  store i{{[0-9]+}} [[A3_IN]], i{{[0-9]+}}* [[A3_ADDR]],
 // TCHECK:  store [10 x i{{[0-9]+}}]* [[B_IN]], [10 x i{{[0-9]+}}]** [[B_ADDR]],
-// TCHECK-64:  [[A_CONV:%.+]] = bitcast i{{[0-9]+}}* [[A_ADDR]] to i{{[0-9]+}}*
-// TCHECK:  [[A3_CONV:%.+]] = bitcast i{{[0-9]+}}* [[A3_ADDR]] to i8*
 // TCHECK:  [[B_ADDR_REF:%.+]] = load [10 x i{{[0-9]+}}]*, [10 x i{{[0-9]+}}]** [[B_ADDR]],
+// TCHECK:  [[B_ADDR_REF:%.+]] = load [10 x i{{[0-9]+}}]*, [10 x i{{[0-9]+}}]** %
 
 // firstprivate(a): a_priv = a_in
 
@@ -158,8 +158,8 @@ int fstatic(int n) {
 struct S1 {
   double a;
 
-  int r1(int n){
-    int b = n+1;
+  int r1(int n) {
+    int b = n + 1;
 
 #pragma omp target firstprivate(b)
     {
@@ -169,7 +169,7 @@ struct S1 {
     return (int)b;
   }
 
-  // TCHECK: define void @__omp_offloading_{{.+}}([[S1]]* [[TH:%.+]], i{{[0-9]+}} [[B_IN:%.+]])
+  // TCHECK: define internal void @__omp_offloading_{{.+}}([[S1]]* [[TH:%.+]], i{{[0-9]+}} [[B_IN:%.+]])
   // TCHECK:  [[TH_ADDR:%.+]] = alloca [[S1]]*,
   // TCHECK:  [[B_ADDR:%.+]] = alloca i{{[0-9]+}},
   // TCHECK-NOT: alloca i{{[0-9]+}},
@@ -185,9 +185,7 @@ struct S1 {
   // TCHECK: ret void
 };
 
-
-
-int bar(int n, double *ptr){
+int bar(int n, double *ptr) {
   int a = 0;
   a += foo(n, ptr);
   S1 S;
@@ -200,15 +198,15 @@ int bar(int n, double *ptr){
 
 // template
 
-// TCHECK: define void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A_IN:%.+]], [10 x i{{[0-9]+}}]*{{.+}} [[B_IN:%.+]])
+// TCHECK: define internal void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A_IN:%.+]], [10 x i{{[0-9]+}}]*{{.+}} [[B_IN:%.+]])
 // TCHECK:  [[A_ADDR:%.+]] = alloca i{{[0-9]+}},
 // TCHECK:  [[B_ADDR:%.+]] = alloca [10 x i{{[0-9]+}}]*,
 // TCHECK-NOT: alloca i{{[0-9]+}},
 // TCHECK:  [[B_PRIV:%.+]] = alloca [10 x i{{[0-9]+}}],
 // TCHECK:  store i{{[0-9]+}} [[A_IN]], i{{[0-9]+}}* [[A_ADDR]],
 // TCHECK:  store [10 x i{{[0-9]+}}]* [[B_IN]], [10 x i{{[0-9]+}}]** [[B_ADDR]],
-// TCHECK-64:  [[A_ADDR_CONV:%.+]] = bitcast i{{[0-9]+}}* [[A_ADDR]] to i{{[0-9]+}}*
 // TCHECK:  [[B_ADDR_REF:%.+]] = load [10 x i{{[0-9]+}}]*, [10 x i{{[0-9]+}}]** [[B_ADDR]],
+// TCHECK:  [[B_ADDR_REF:%.+]] = load [10 x i{{[0-9]+}}]*, [10 x i{{[0-9]+}}]** %
 
 // firstprivate(a)
 // TCHECK-NOT:  store i{{[0-9]+}} %{{.+}}, i{{[0-9]+}}*
diff --git a/test/OpenMP/nvptx_target_teams_codegen.cpp b/test/OpenMP/nvptx_target_teams_codegen.cpp
index e823eabbb10f..b79fd185d77a 100644
--- a/test/OpenMP/nvptx_target_teams_codegen.cpp
+++ b/test/OpenMP/nvptx_target_teams_codegen.cpp
@@ -60,7 +60,7 @@ int bar(int n){
   //
   // CHECK: [[AWAIT_WORK]]
   // CHECK: call void @llvm.nvvm.barrier0()
-  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]])
+  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]], i8*** %shared_args)
   // CHECK: [[KPRB:%.+]] = zext i1 [[KPR]] to i8
   // store i8 [[KPRB]], i8* [[OMP_EXEC_STATUS]], align 1
   // CHECK: [[WORK:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
@@ -121,11 +121,13 @@ int bar(int n){
   // CHECK: [[ACV:%.+]] = load i[[SZ]], i[[SZ]]* [[AC]], align
   // CHECK: store i[[SZ]] [[ACV]], i[[SZ]]* [[A_ADDR_T:%.+]], align
   // CHECK: [[CONV2:%.+]] = bitcast i[[SZ]]* [[A_ADDR_T]] to i8*
-  // CHECK: store i8 49, i8* [[CONV2]], align
+  // CHECK: [[LD_CONV2:%.+]] = load i8, i8* [[CONV2]],
+  // CHECK: store i8 [[LD_CONV2]], i8* [[A_PRIV:%[^,]+]],
+  // CHECK: store i8 49, i8* [[A_PRIV]], align
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
@@ -146,7 +148,7 @@ int bar(int n){
   //
   // CHECK: [[AWAIT_WORK]]
   // CHECK: call void @llvm.nvvm.barrier0()
-  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]])
+  // CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]], i8*** %shared_args)
   // CHECK: [[KPRB:%.+]] = zext i1 [[KPR]] to i8
   // store i8 [[KPRB]], i8* [[OMP_EXEC_STATUS]], align 1
   // CHECK: [[WORK:%.+]] = load i8*, i8** [[OMP_WORK_FN]],
@@ -207,11 +209,13 @@ int bar(int n){
   // CHECK: [[ACV:%.+]] = load i[[SZ]], i[[SZ]]* [[AC]], align
   // CHECK: store i[[SZ]] [[ACV]], i[[SZ]]* [[AA_ADDR_T:%.+]], align
   // CHECK: [[CONV2:%.+]] = bitcast i[[SZ]]* [[AA_ADDR_T]] to i16*
-  // CHECK: store i16 1, i16* [[CONV2]], align
+  // CHECK: [[LD_CONV2:%.+]] = load i16, i16* [[CONV2]],
+  // CHECK: store i16 [[LD_CONV2]], i16* [[A_PRIV:%[^,]+]],
+  // CHECK: store i16 1, i16* [[A_PRIV]], align
   // CHECK: br label {{%?}}[[TERMINATE:.+]]
   //
   // CHECK: [[TERMINATE]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
   // CHECK: call void @llvm.nvvm.barrier0()
   // CHECK: br label {{%?}}[[EXIT]]
   //
diff --git a/test/OpenMP/nvptx_teams_reduction_codegen.cpp b/test/OpenMP/nvptx_teams_reduction_codegen.cpp
index ae129ebfae4d..d77231807fb4 100644
--- a/test/OpenMP/nvptx_teams_reduction_codegen.cpp
+++ b/test/OpenMP/nvptx_teams_reduction_codegen.cpp
@@ -84,7 +84,7 @@ int bar(int n){
   // CHECK: br label %[[EXIT]]
   //
   // CHECK: [[EXIT]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
 
   //
   // Reduction function
@@ -360,7 +360,7 @@ int bar(int n){
   // CHECK: br label %[[EXIT]]
   //
   // CHECK: [[EXIT]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
 
   //
   // Reduction function
@@ -776,7 +776,7 @@ int bar(int n){
   // CHECK: br label %[[EXIT]]
   //
   // CHECK: [[EXIT]]
-  // CHECK: call void @__kmpc_kernel_deinit()
+  // CHECK: call void @__kmpc_kernel_deinit(
 
   //
   // Reduction function
diff --git a/test/OpenMP/openmp_offload_codegen.cpp b/test/OpenMP/openmp_offload_codegen.cpp
new file mode 100644
index 000000000000..f89bccf3b552
--- /dev/null
+++ b/test/OpenMP/openmp_offload_codegen.cpp
@@ -0,0 +1,36 @@
+// Test device for mapping codegen.
+///==========================================================================///
+
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -S -emit-llvm %s -o - 2>&1 \
+// RUN:   | FileCheck -check-prefix=CK1 %s
+
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix CK1-DEVICE
+
+// expected-no-diagnostics
+
+#ifdef CK1
+
+void target_maps_parallel_integer(int a){
+  int ParamToKernel = a;
+#pragma omp target map(tofrom: ParamToKernel)
+  {
+    ParamToKernel += 1;
+  }
+}
+
+// CK1-DEVICE: {{.*}}void @__omp_offloading_{{.*}}(i32* dereferenceable(4){{.*}}
+
+// CK1: {{.*}}void {{.*}}target_maps_parallel_integer{{.*}} {
+
+// CK1: [[GEPOBP:%.+]] = getelementptr inbounds {{.*}}
+// CK1: [[GEPOBPBIT:%.+]] = bitcast i8** [[GEPOBP]]
+// CK1: store i32* %ParamToKernel, i32** [[GEPOBPBIT]]
+// CK1: [[GEPOP:%.+]] = getelementptr inbounds {{.*}}
+// CK1: [[GEPOPBIT:%.+]] = bitcast i8** [[GEPOP]]
+// CK1: store i32* %ParamToKernel, i32** [[GEPOPBIT]]
+// CK1: [[GEPOBPARG:%.+]] = getelementptr inbounds {{.*}}
+// CK1: [[GEPOPARG:%.+]] = getelementptr inbounds {{.*}}
+// CK1: call {{.*}}tgt_target({{.*}}i8** [[GEPOBPARG]], i8** [[GEPOPARG]]
+
+#endif
\ No newline at end of file
diff --git a/test/OpenMP/openmp_win_codegen.cpp b/test/OpenMP/openmp_win_codegen.cpp
new file mode 100644
index 000000000000..45b2c185cb8f
--- /dev/null
+++ b/test/OpenMP/openmp_win_codegen.cpp
@@ -0,0 +1,61 @@
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple x86_64-pc-windows-msvc18.0.0 -std=c++11 -fms-compatibility-version=18 -fms-extensions -emit-llvm %s -fexceptions -fcxx-exceptions -o - -O1 | FileCheck %s
+// REQUIRES: x86-registered-target
+// expected-no-diagnostics
+
+void foo();
+void bar();
+
+struct Test {
+  static void main() {
+    int failed = 0;
+    int j = 2;
+
+#pragma omp parallel
+    {
+      int local_j = 3;
+#pragma omp single copyprivate(local_j)
+      {
+        local_j = 4;
+      }
+
+      // Assure reports a data race, but value written to "j"
+      // should always be the same.
+      j = local_j;
+    }
+
+  }
+};
+
+// CHECK-LABEL: @main
+int main() {
+  // CHECK: call void @{{.+}}main
+  Test::main();
+  // CHECK: call void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* {{.*}}@0, i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* [[OUTLINED:@.+]] to void (i32*, i32*, ...)*))
+#pragma omp parallel
+  {
+    try {
+      foo();
+    } catch (int t) {
+#pragma omp critical
+      {
+        bar();
+      };
+    }
+  };
+  // CHECK: ret i32 0
+  return 0;
+}
+
+// CHECK: define internal void [[OUTLINED]](
+// CHECK: [[GID:%.+]] = {{.*}}call i32 @__kmpc_global_thread_num(%ident_t* {{.*}}@0)
+// CHECK: invoke void @{{.+}}foo
+// CHECK: catchswitch within
+// CHECK: catchpad within
+// CHECK: call void @__kmpc_critical(%ident_t* {{.*}}@0, i32 [[GID]],
+// CHECK: invoke void @{{.+}}bar
+// CHECK: call void @__kmpc_end_critical(%ident_t* {{.*}}@0, i32 [[GID]],
+// CHECK: catchret from
+// CHECK: cleanuppad within
+// CHECK: call void @__kmpc_end_critical(%ident_t* {{.*}}@0, i32 [[GID]],
+// CHECK: cleanupret from
+
diff --git a/test/OpenMP/ordered_codegen.cpp b/test/OpenMP/ordered_codegen.cpp
index bd7c07051b04..af6a8722438d 100644
--- a/test/OpenMP/ordered_codegen.cpp
+++ b/test/OpenMP/ordered_codegen.cpp
@@ -216,7 +216,7 @@ float f[10];
 // CHECK-LABEL: foo_simd
 void foo_simd(int low, int up) {
   // CHECK: store float 0.000000e+00, float* %{{.+}}, align {{[0-9]+}}, !llvm.mem.parallel_loop_access !
-  // CHECK-NEXT: call void [[CAP_FUNC:@.+]](i32* %{{.+}}) #{{[0-9]+}}, !llvm.mem.parallel_loop_access !
+  // CHECK-NEXT: call void [[CAP_FUNC:@.+]](i32* %{{.+}}), !llvm.mem.parallel_loop_access !
 #pragma omp simd
   for (int i = low; i < up; ++i) {
     f[i] = 0.0;
@@ -224,7 +224,7 @@ void foo_simd(int low, int up) {
     f[i] = 1.0;
   }
   // CHECK: store float 0.000000e+00, float* %{{.+}}, align {{[0-9]+}}
-  // CHECK-NEXT: call void [[CAP_FUNC:@.+]](i32* %{{.+}}) #{{[0-9]+}}
+  // CHECK-NEXT: call void [[CAP_FUNC:@.+]](i32* %{{.+}})
 #pragma omp for simd ordered
   for (int i = low; i < up; ++i) {
     f[i] = 0.0;
diff --git a/test/OpenMP/parallel_ast_print.cpp b/test/OpenMP/parallel_ast_print.cpp
index 1e24c6e059bd..09141db5841f 100644
--- a/test/OpenMP/parallel_ast_print.cpp
+++ b/test/OpenMP/parallel_ast_print.cpp
@@ -135,6 +135,9 @@ struct S {
 // CHECK-NEXT:   #pragma omp threadprivate(S<long>::TS)
 // CHECK-NEXT: }
 
+int thrp;
+#pragma omp threadprivate(thrp)
+
 template <typename T, int C>
 T tmain(T argc, T *argv) {
   T b = argc, c, d, e, f, g;
@@ -143,7 +146,7 @@ T tmain(T argc, T *argv) {
   T arr[C][10], arr1[C];
 #pragma omp parallel
   a=2;
-#pragma omp parallel default(none), private(argc,b) firstprivate(argv) shared (d) if (parallel:argc > 0) num_threads(C) copyin(S<T>::TS) proc_bind(master) reduction(+:c, arr1[argc]) reduction(max:e, arr[:C][0:10])
+#pragma omp parallel default(none), private(argc,b) firstprivate(argv) shared (d) if (parallel:argc > 0) num_threads(C) copyin(S<T>::TS, thrp) proc_bind(master) reduction(+:c, arr1[argc]) reduction(max:e, arr[:C][0:10])
   foo();
 #pragma omp parallel if (C) num_threads(s) proc_bind(close) reduction(^:e, f, arr[0:C][:argc]) reduction(&& : g)
   foo();
@@ -157,7 +160,7 @@ T tmain(T argc, T *argv) {
 // CHECK-NEXT: T arr[C][10], arr1[C];
 // CHECK-NEXT: #pragma omp parallel
 // CHECK-NEXT: a = 2;
-// CHECK-NEXT: #pragma omp parallel default(none) private(argc,b) firstprivate(argv) shared(d) if(parallel: argc > 0) num_threads(C) copyin(S<T>::TS) proc_bind(master) reduction(+: c,arr1[argc]) reduction(max: e,arr[:C][0:10])
+// CHECK-NEXT: #pragma omp parallel default(none) private(argc,b) firstprivate(argv) shared(d) if(parallel: argc > 0) num_threads(C) copyin(S<T>::TS,thrp) proc_bind(master) reduction(+: c,arr1[argc]) reduction(max: e,arr[:C][0:10])
 // CHECK-NEXT: foo()
 // CHECK-NEXT: #pragma omp parallel if(C) num_threads(s) proc_bind(close) reduction(^: e,f,arr[0:C][:argc]) reduction(&&: g)
 // CHECK-NEXT: foo()
@@ -168,7 +171,7 @@ T tmain(T argc, T *argv) {
 // CHECK-NEXT: int arr[5][10], arr1[5];
 // CHECK-NEXT: #pragma omp parallel
 // CHECK-NEXT: a = 2;
-// CHECK-NEXT: #pragma omp parallel default(none) private(argc,b) firstprivate(argv) shared(d) if(parallel: argc > 0) num_threads(5) copyin(S<int>::TS) proc_bind(master) reduction(+: c,arr1[argc]) reduction(max: e,arr[:5][0:10])
+// CHECK-NEXT: #pragma omp parallel default(none) private(argc,b) firstprivate(argv) shared(d) if(parallel: argc > 0) num_threads(5) copyin(S<int>::TS,thrp) proc_bind(master) reduction(+: c,arr1[argc]) reduction(max: e,arr[:5][0:10])
 // CHECK-NEXT: foo()
 // CHECK-NEXT: #pragma omp parallel if(5) num_threads(s) proc_bind(close) reduction(^: e,f,arr[0:5][:argc]) reduction(&&: g)
 // CHECK-NEXT: foo()
@@ -179,7 +182,7 @@ T tmain(T argc, T *argv) {
 // CHECK-NEXT: long arr[1][10], arr1[1];
 // CHECK-NEXT: #pragma omp parallel
 // CHECK-NEXT: a = 2;
-// CHECK-NEXT: #pragma omp parallel default(none) private(argc,b) firstprivate(argv) shared(d) if(parallel: argc > 0) num_threads(1) copyin(S<long>::TS) proc_bind(master) reduction(+: c,arr1[argc]) reduction(max: e,arr[:1][0:10])
+// CHECK-NEXT: #pragma omp parallel default(none) private(argc,b) firstprivate(argv) shared(d) if(parallel: argc > 0) num_threads(1) copyin(S<long>::TS,thrp) proc_bind(master) reduction(+: c,arr1[argc]) reduction(max: e,arr[:1][0:10])
 // CHECK-NEXT: foo()
 // CHECK-NEXT: #pragma omp parallel if(1) num_threads(s) proc_bind(close) reduction(^: e,f,arr[0:1][:argc]) reduction(&&: g)
 // CHECK-NEXT: foo()
diff --git a/test/OpenMP/parallel_codegen.cpp b/test/OpenMP/parallel_codegen.cpp
index 23b323778b47..77e8efded68b 100644
--- a/test/OpenMP/parallel_codegen.cpp
+++ b/test/OpenMP/parallel_codegen.cpp
@@ -53,7 +53,7 @@ int main (int argc, char **argv) {
 // CHECK-DEBUG:       ret i32
 // CHECK-DEBUG-NEXT:  }
 
-// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i{{[0-9]+}}{{.*}} [[VLA_SIZE:%.+]], i32* [[VLA_ADDR:%[^)]+]])
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i{{[0-9]+}}{{.*}} [[VLA_SIZE:%.+]], i32* {{.+}} [[VLA_ADDR:%[^)]+]])
 // CHECK-SAME:       #[[FN_ATTRS:[0-9]+]]
 // CHECK:       store i32* [[VLA_ADDR]], i32** [[VLA_PTR_ADDR:%.+]],
 // CHECK:       [[VLA_REF:%.+]] = load i32*, i32** [[VLA_PTR_ADDR]]
@@ -64,7 +64,7 @@ int main (int argc, char **argv) {
 // CHECK:       call {{.*}}void @{{.+terminate.*|abort}}(
 // CHECK-NEXT:  unreachable
 // CHECK-NEXT:  }
-// CHECK-DEBUG:       define internal void [[OMP_OUTLINED_DEBUG:@.+]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i64 [[VLA_SIZE:%.+]], i32* [[VLA_ADDR:%[^)]+]])
+// CHECK-DEBUG:       define internal void [[OMP_OUTLINED_DEBUG:@.+]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i64 [[VLA_SIZE:%.+]], i32* {{.+}} [[VLA_ADDR:%[^)]+]])
 // CHECK-DEBUG-SAME:  #[[FN_ATTRS:[0-9]+]]
 // CHECK-DEBUG:       store i32* [[VLA_ADDR]], i32** [[VLA_PTR_ADDR:%.+]],
 // CHECK-DEBUG:       [[VLA_REF:%.+]] = load i32*, i32** [[VLA_PTR_ADDR]]
@@ -80,7 +80,7 @@ int main (int argc, char **argv) {
 // CHECK-DAG: declare {{.*}}void @__kmpc_fork_call(%ident_t*, i32, void (i32*, i32*, ...)*, ...)
 // CHECK-DEBUG-DAG: define linkonce_odr void [[FOO]](i32 %argc)
 // CHECK-DEBUG-DAG: declare void @__kmpc_fork_call(%ident_t*, i32, void (i32*, i32*, ...)*, ...)
-// CHECK-DEBUG-DAG:       define internal void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i64 [[VLA_SIZE:%.+]], i32* [[VLA_ADDR:%[^)]+]])
+// CHECK-DEBUG-DAG:       define internal void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i64 [[VLA_SIZE:%.+]], i32* {{.+}} [[VLA_ADDR:%[^)]+]])
 // CHECK-DEBUG-DAG:       call void [[OMP_OUTLINED_DEBUG]]
 
 // CHECK:       define linkonce_odr {{[a-z\_\b]*[ ]?i32}} [[TMAIN]](i8** %argc)
@@ -109,7 +109,7 @@ int main (int argc, char **argv) {
 // CHECK:       call {{.*}}void @{{.+terminate.*|abort}}(
 // CHECK-NEXT:  unreachable
 // CHECK-NEXT:  }
-// CHECK-DEBUG:       define internal void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i8*** dereferenceable({{4|8}}) %argc)
+// CHECK-DEBUG:       define internal void [[OMP_OUTLINED_DEBUG:@.+]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i8*** dereferenceable({{4|8}}) %argc)
 // CHECK-DEBUG:       store i8*** %argc, i8**** [[ARGC_PTR_ADDR:%.+]],
 // CHECK-DEBUG:       [[ARGC_REF:%.+]] = load i8***, i8**** [[ARGC_PTR_ADDR]]
 // CHECK-DEBUG-NEXT:  [[ARGC:%.+]] = load i8**, i8*** [[ARGC_REF]]
@@ -120,7 +120,9 @@ int main (int argc, char **argv) {
 // CHECK-DEBUG-NEXT:  }
 
 // CHECK: define linkonce_odr {{.*}}void [[FOO1]](i8** %argc)
-// CHECK-DEBUG: define linkonce_odr void [[FOO1]](i8** %argc)
+// CHECK-DEBUG-DAG: define linkonce_odr void [[FOO1]](i8** %argc)
+// CHECK-DEBUG-DAG: define internal void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i8*** dereferenceable({{4|8}}) %argc)
+// CHECK-DEBUG-DAG: call void [[OMP_OUTLINED_DEBUG]]({{[^)]+}}){{[^,]*}}, !dbg
 
 // CHECK: attributes #[[FN_ATTRS]] = {{.+}} nounwind
 // CHECK-DEBUG: attributes #[[FN_ATTRS]] = {{.+}} nounwind
diff --git a/test/OpenMP/parallel_firstprivate_messages.cpp b/test/OpenMP/parallel_firstprivate_messages.cpp
index fc0eb4c7d358..b2cb4ff131ef 100644
--- a/test/OpenMP/parallel_firstprivate_messages.cpp
+++ b/test/OpenMP/parallel_firstprivate_messages.cpp
@@ -56,7 +56,7 @@ using A::x;
 }
 
 int main(int argc, char **argv) {
-  const int d = 5;
+  const int d = 5; // expected-note {{variable 'd' declared const here}}
   const int da[5] = { 0 };
   S4 e(4);
   S5 g(5);
@@ -72,6 +72,8 @@ int main(int argc, char **argv) {
   #pragma omp parallel firstprivate (argc)
   #pragma omp parallel firstprivate (S1) // expected-error {{'S1' does not refer to a value}}
   #pragma omp parallel firstprivate (a, b, c, d, f) // expected-error {{firstprivate variable with incomplete type 'S1'}}
+  #pragma omp parallel firstprivate (d)
+    d = 5; // expected-error {{cannot assign to variable 'd' with const-qualified type}}
   #pragma omp parallel firstprivate (argv[1]) // expected-error {{expected variable name}}
   #pragma omp parallel firstprivate(ba)
   #pragma omp parallel firstprivate(ca)
diff --git a/test/OpenMP/parallel_for_codegen.cpp b/test/OpenMP/parallel_for_codegen.cpp
index 1324ee67b615..1773619bce20 100644
--- a/test/OpenMP/parallel_for_codegen.cpp
+++ b/test/OpenMP/parallel_for_codegen.cpp
@@ -8,6 +8,7 @@
 #define HEADER
 
 // CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[LOOP_LOC:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 514, i32 0, i32 0, i8*
 
 // CHECK-LABEL: with_var_schedule
 void with_var_schedule() {
@@ -19,8 +20,8 @@ void with_var_schedule() {
 // CHECK: [[CHUNK:%.+]] = load i8*, i8** %
 // CHECK: [[CHUNK_VAL:%.+]] = load i8, i8* [[CHUNK]],
 // CHECK: [[CHUNK_SIZE:%.+]] = sext i8 [[CHUNK_VAL]] to i64
-// CHECK: call void @__kmpc_for_static_init_8u([[IDENT_T_TY]]* [[DEFAULT_LOC:@[^,]+]], i32 [[GTID:%[^,]+]], i32 33, i32* [[IS_LAST:%[^,]+]], i64* [[OMP_LB:%[^,]+]], i64* [[OMP_UB:%[^,]+]], i64* [[OMP_ST:%[^,]+]], i64 1, i64 [[CHUNK_SIZE]])
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_init_8u([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID:%[^,]+]], i32 33, i32* [[IS_LAST:%[^,]+]], i64* [[OMP_LB:%[^,]+]], i64* [[OMP_UB:%[^,]+]], i64* [[OMP_ST:%[^,]+]], i64 1, i64 [[CHUNK_SIZE]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 #pragma omp parallel for schedule(static, char(a))
   for (unsigned long long i = 1; i < 2; ++i) {
   }
@@ -32,9 +33,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
+// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
 // UB = min(UB, GlobalUB)
 // CHECK-NEXT: [[UB:%.+]] = load i32, i32* [[OMP_UB]]
 // CHECK-NEXT: [[UBCMP:%.+]] = icmp sgt i32 [[UB]], 4571423
@@ -65,7 +64,7 @@ void without_schedule_clause(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: br label %{{.+}}
   }
 // CHECK: [[LOOP1_END]]
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 // CHECK: ret void
 }
 
@@ -75,9 +74,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
+// CHECK: call void @__kmpc_for_static_init_4([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
 // UB = min(UB, GlobalUB)
 // CHECK-NEXT: [[UB:%.+]] = load i32, i32* [[OMP_UB]]
 // CHECK-NEXT: [[UBCMP:%.+]] = icmp sgt i32 [[UB]], 4571423
@@ -108,7 +105,7 @@ void static_not_chunked(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: br label %{{.+}}
   }
 // CHECK: [[LOOP1_END]]
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 // CHECK: ret void
 }
 
@@ -118,9 +115,7 @@ void static_chunked(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_for_static_init_4u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 33, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 5)
+// CHECK: call void @__kmpc_for_static_init_4u([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID:%.+]], i32 33, i32* [[IS_LAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 5)
 // UB = min(UB, GlobalUB)
 // CHECK: [[UB:%.+]] = load i32, i32* [[OMP_UB]]
 // CHECK-NEXT: [[UBCMP:%.+]] = icmp ugt i32 [[UB]], 16908288
@@ -170,7 +165,7 @@ void static_chunked(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: store i32 [[ADD_UB]], i32* [[OMP_UB]]
 
 // CHECK: [[O_LOOP1_END]]
-// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[LOOP_LOC]], i32 [[GTID]])
 // CHECK: ret void
 }
 
@@ -180,9 +175,7 @@ void dynamic1(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_dispatch_init_8u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 35, i64 0, i64 16908287, i64 1, i64 1)
+// CHECK: call void @__kmpc_dispatch_init_8u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID:%.+]], i32 35, i64 0, i64 16908287, i64 1, i64 1)
 //
 // CHECK: [[HASWORK:%.+]] = call i32 @__kmpc_dispatch_next_8u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32* [[OMP_ISLAST:%[^,]+]], i64* [[OMP_LB:%[^,]+]], i64* [[OMP_UB:%[^,]+]], i64* [[OMP_ST:%[^,]+]])
 // CHECK-NEXT: [[O_CMP:%.+]] = icmp ne i32 [[HASWORK]], 0
@@ -224,9 +217,7 @@ void guided7(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_dispatch_init_8u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 36, i64 0, i64 16908287, i64 1, i64 7)
+// CHECK: call void @__kmpc_dispatch_init_8u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID:%.+]], i32 36, i64 0, i64 16908287, i64 1, i64 7)
 //
 // CHECK: [[HASWORK:%.+]] = call i32 @__kmpc_dispatch_next_8u([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32* [[OMP_ISLAST:%[^,]+]], i64* [[OMP_LB:%[^,]+]], i64* [[OMP_UB:%[^,]+]], i64* [[OMP_ST:%[^,]+]])
 // CHECK-NEXT: [[O_CMP:%.+]] = icmp ne i32 [[HASWORK]], 0
@@ -270,12 +261,8 @@ void test_auto(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 6, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}}, i32* dereferenceable(4) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_dispatch_init_8([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 38, i64 0, i64 [[LAST_ITER:%[^,]+]], i64 1, i64 1)
+// CHECK: call void @__kmpc_dispatch_init_8([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID:%.+]], i32 38, i64 0, i64 [[LAST_ITER:%[^,]+]], i64 1, i64 1)
 //
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
 // CHECK: [[HASWORK:%.+]] = call i32 @__kmpc_dispatch_next_8([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32* [[OMP_ISLAST:%[^,]+]], i64* [[OMP_LB:%[^,]+]], i64* [[OMP_UB:%[^,]+]], i64* [[OMP_ST:%[^,]+]])
 // CHECK-NEXT: [[O_CMP:%.+]] = icmp ne i32 [[HASWORK]], 0
 // CHECK-NEXT: br i1 [[O_CMP]], label %[[O_LOOP1_BODY:[^,]+]], label %[[O_LOOP1_END:[^,]+]]
@@ -318,9 +305,7 @@ void runtime(float *a, float *b, float *c, float *d) {
 // CHECK: call void ([[IDENT_T_TY]]*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call([[IDENT_T_TY]]* [[DEFAULT_LOC:[@%].+]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, float**, float**, float**, float**)* [[OMP_PARALLEL_FUNC:@.+]] to void (i32*, i32*, ...)*),
 // CHECK: define internal void [[OMP_PARALLEL_FUNC]](i32* noalias [[GTID_PARAM_ADDR:%.+]], i32* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}}, float** dereferenceable(8) %{{.+}})
 // CHECK: store i32* [[GTID_PARAM_ADDR]], i32** [[GTID_REF_ADDR:%.+]],
-// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call void @__kmpc_dispatch_init_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32 37, i32 0, i32 199, i32 1, i32 1)
+// CHECK: call void @__kmpc_dispatch_init_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID:%.+]], i32 37, i32 0, i32 199, i32 1, i32 1)
 //
 // CHECK: [[HASWORK:%.+]] = call i32 @__kmpc_dispatch_next_4([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]], i32* [[OMP_ISLAST:%[^,]+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]])
 // CHECK-NEXT: [[O_CMP:%.+]] = icmp ne i32 [[HASWORK]], 0
diff --git a/test/OpenMP/parallel_for_lastprivate_messages.cpp b/test/OpenMP/parallel_for_lastprivate_messages.cpp
index ccfe2ea407b1..134100bc1753 100644
--- a/test/OpenMP/parallel_for_lastprivate_messages.cpp
+++ b/test/OpenMP/parallel_for_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   S2 &operator=(const S2 &);
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/parallel_for_linear_codegen.cpp b/test/OpenMP/parallel_for_linear_codegen.cpp
index d0968d7cd146..2c3cfddbd571 100644
--- a/test/OpenMP/parallel_for_linear_codegen.cpp
+++ b/test/OpenMP/parallel_for_linear_codegen.cpp
@@ -49,6 +49,7 @@ int main() {
   for (int i = 0; i < 2; ++i) {
     // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}})
     // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
     // LAMBDA: alloca i{{[0-9]+}},
@@ -96,6 +97,7 @@ int main() {
   for (int i = 0; i < 2; ++i) {
     // BLOCKS: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}})
     // BLOCKS: alloca i{{[0-9]+}},
+    // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}},
     // BLOCKS: alloca i{{[0-9]+}},
     // BLOCKS: alloca i{{[0-9]+}},
@@ -155,13 +157,13 @@ int main() {
 
 // CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, i64* dereferenceable(8) %{{.+}})
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_START:%.+]] = alloca float*,
 // CHECK: [[LVAR_START:%.+]] = alloca i64,
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
-// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_PRIV:%.+]] = alloca float*,
 // CHECK: [[LVAR_PRIV:%.+]] = alloca i64,
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
@@ -205,13 +207,13 @@ int main() {
 //
 // CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i32** dereferenceable(8) %{{.+}}, i32* dereferenceable(4) %{{.+}})
 // CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_START:%.+]] = alloca i32*,
 // CHECK: [[LVAR_START:%.+]] = alloca i32,
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
-// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[PVAR_PRIV:%.+]] = alloca i32*,
 // CHECK: [[LVAR_PRIV:%.+]] = alloca i32,
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]]
diff --git a/test/OpenMP/parallel_for_reduction_messages.cpp b/test/OpenMP/parallel_for_reduction_messages.cpp
index 57ab1fac9c11..6499ef593a3d 100644
--- a/test/OpenMP/parallel_for_reduction_messages.cpp
+++ b/test/OpenMP/parallel_for_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
@@ -61,6 +61,12 @@ class S5 {
 
 public:
   S5(int v) : a(v) {}
+  void foo() {
+#pragma omp parallel private(a) // expected-note {{defined as private}}
+#pragma omp for reduction(+:a) // expected-error {{reduction variable must be shared}}
+  for (int i = 0; i < 10; ++i)
+    ::foo();
+  }
 };
 class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
 #if __cplusplus >= 201103L // C++11 or later
diff --git a/test/OpenMP/parallel_for_simd_codegen.cpp b/test/OpenMP/parallel_for_simd_codegen.cpp
index 06d96353ee9b..369ea17844e9 100644
--- a/test/OpenMP/parallel_for_simd_codegen.cpp
+++ b/test/OpenMP/parallel_for_simd_codegen.cpp
@@ -114,6 +114,7 @@ void simple(float *a, float *b, float *c, float *d) {
   int lin = 12;
   #pragma omp parallel for simd linear(lin : get_val()), linear(g_ptr)
 
+// CHECK: alloca i32,
 // Init linear private var.
 // CHECK: [[LIN_VAR:%.+]] = load i32*, i32** %
 // CHECK: [[LIN_LOAD:%.+]] = load i32, i32* [[LIN_VAR]]
diff --git a/test/OpenMP/parallel_for_simd_lastprivate_messages.cpp b/test/OpenMP/parallel_for_simd_lastprivate_messages.cpp
index bd1a6d505572..8670e55c6e2b 100644
--- a/test/OpenMP/parallel_for_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/parallel_for_simd_lastprivate_messages.cpp
@@ -17,9 +17,9 @@ public:
   S2(S2 &s2) : a(s2.a) {}
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/parallel_for_simd_reduction_messages.cpp b/test/OpenMP/parallel_for_simd_reduction_messages.cpp
index 60a947dd5f84..8ff246e35f6a 100644
--- a/test/OpenMP/parallel_for_simd_reduction_messages.cpp
+++ b/test/OpenMP/parallel_for_simd_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/parallel_reduction_messages.cpp b/test/OpenMP/parallel_reduction_messages.cpp
index 947353fc2dd4..96f4b58ec8a4 100644
--- a/test/OpenMP/parallel_reduction_messages.cpp
+++ b/test/OpenMP/parallel_reduction_messages.cpp
@@ -24,9 +24,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/parallel_sections_codegen.cpp b/test/OpenMP/parallel_sections_codegen.cpp
index 6fc687d0c503..a26147303611 100644
--- a/test/OpenMP/parallel_sections_codegen.cpp
+++ b/test/OpenMP/parallel_sections_codegen.cpp
@@ -28,9 +28,7 @@ int main() {
   {
 // CHECK:      store i32 0, i32* [[LB_PTR:%.+]],
 // CHECK:      store i32 1, i32* [[UB_PTR:%.+]],
-// CHECK:      [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
-// CHECK:      [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK:      call void @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_PTR:%.+]], i32* [[LB_PTR]], i32* [[UB_PTR]], i32* [[STRIDE_PTR:%.+]], i32 1, i32 1)
+// CHECK:      call void @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST_PTR:%.+]], i32* [[LB_PTR]], i32* [[UB_PTR]], i32* [[STRIDE_PTR:%.+]], i32 1, i32 1)
 // <<UB = min(UB, GlobalUB);>>
 // CHECK:      [[UB:%.+]] = load i32, i32* [[UB_PTR]]
 // CHECK:      [[CMP:%.+]] = icmp slt i32 [[UB]], 1
diff --git a/test/OpenMP/parallel_sections_lastprivate_messages.cpp b/test/OpenMP/parallel_sections_lastprivate_messages.cpp
index af3c5e2a575b..eccbfc5e7e02 100644
--- a/test/OpenMP/parallel_sections_lastprivate_messages.cpp
+++ b/test/OpenMP/parallel_sections_lastprivate_messages.cpp
@@ -17,9 +17,9 @@ public:
   S2(S2 &s2) : a(s2.a) {}
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/parallel_sections_reduction_messages.cpp b/test/OpenMP/parallel_sections_reduction_messages.cpp
index 05cc7f0f6c5f..26b3a5355c19 100644
--- a/test/OpenMP/parallel_sections_reduction_messages.cpp
+++ b/test/OpenMP/parallel_sections_reduction_messages.cpp
@@ -26,9 +26,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/sections_codegen.cpp b/test/OpenMP/sections_codegen.cpp
index 65747e7b912f..94ded37db95d 100644
--- a/test/OpenMP/sections_codegen.cpp
+++ b/test/OpenMP/sections_codegen.cpp
@@ -4,7 +4,8 @@
 // expected-no-diagnostics
 #ifndef HEADER
 #define HEADER
-// CHECK: [[IMPLICIT_BARRIER_SECTIONS_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 194, i32 0, i32 0, i8*
+// CHECK-DAG: [[IMPLICIT_BARRIER_SECTIONS_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 194, i32 0, i32 0, i8*
+// CHECK-DAG: [[SECTIONS_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 1026, i32 0, i32 0, i8*
 // CHECK-LABEL: foo
 void foo() {};
 // CHECK-LABEL: bar
@@ -29,7 +30,7 @@ int main() {
   {
 // CHECK:      store i32 0, i32* [[LB_PTR:%.+]],
 // CHECK:      store i32 1, i32* [[UB_PTR:%.+]],
-// CHECK:      call void @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_PTR:%.+]], i32* [[LB_PTR]], i32* [[UB_PTR]], i32* [[STRIDE_PTR:%.+]], i32 1, i32 1)
+// CHECK:      call void @__kmpc_for_static_init_4(%{{.+}}* [[SECTIONS_LOC]], i32 [[GTID]], i32 34, i32* [[IS_LAST_PTR:%.+]], i32* [[LB_PTR]], i32* [[UB_PTR]], i32* [[STRIDE_PTR:%.+]], i32 1, i32 1)
 // <<UB = min(UB, GlobalUB);>>
 // CHECK:      [[UB:%.+]] = load i32, i32* [[UB_PTR]]
 // CHECK:      [[CMP:%.+]] = icmp slt i32 [[UB]], 1
@@ -69,7 +70,7 @@ int main() {
 // CHECK-NEXT: br label %[[INNER_FOR_COND]]
 // CHECK:      [[INNER_LOOP_END]]
   }
-// CHECK:      call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]])
+// CHECK:      call void @__kmpc_for_static_fini(%{{.+}}* [[SECTIONS_LOC]], i32 [[GTID]])
 // CHECK:      call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_SECTIONS_LOC]],
 #pragma omp sections nowait
   {
diff --git a/test/OpenMP/sections_firstprivate_codegen.cpp b/test/OpenMP/sections_firstprivate_codegen.cpp
index a5426f8bf956..227bb076374f 100644
--- a/test/OpenMP/sections_firstprivate_codegen.cpp
+++ b/test/OpenMP/sections_firstprivate_codegen.cpp
@@ -256,11 +256,7 @@ int main() {
 // CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]()
 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
 // CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
-// CHECK: [[T_VARVAL:%.+]] = load i32, i32* %{{.+}},
-// CHECK: [[T_VARCONV:%.+]] = bitcast i64* [[T_VARCAST:%.+]] to i32*
-// CHECK: store i32  [[T_VARVAL]], i32* [[T_VARCONV]],
-// CHECK: [[T_VARPVT:%.+]] = load i64, i64* [[T_VARCAST]],
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, [2 x i32]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void {{.*}}i64 [[T_VARPVT]],
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i32*, [2 x i32]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void
 // CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: ret
 //
@@ -271,14 +267,13 @@ int main() {
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
 // CHECK: alloca i{{[0-9]+}},
-// CHECK: alloca i{{[0-9]+}},
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK-NOT: load i{{[0-9]+}}*, i{{[0-9]+}}** %
+// CHECK: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %
 // CHECK: [[VEC_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** %
 // CHECK: [[S_ARR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** %
 // CHECK: [[VAR_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** %
diff --git a/test/OpenMP/sections_lastprivate_messages.cpp b/test/OpenMP/sections_lastprivate_messages.cpp
index 5f6b420cb8ea..01bb13de5510 100644
--- a/test/OpenMP/sections_lastprivate_messages.cpp
+++ b/test/OpenMP/sections_lastprivate_messages.cpp
@@ -17,9 +17,9 @@ public:
   S2(S2 &s2) : a(s2.a) {}
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/sections_reduction_messages.cpp b/test/OpenMP/sections_reduction_messages.cpp
index f13c5b3f2862..58b22a5a61d0 100644
--- a/test/OpenMP/sections_reduction_messages.cpp
+++ b/test/OpenMP/sections_reduction_messages.cpp
@@ -27,9 +27,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/simd_lastprivate_messages.cpp b/test/OpenMP/simd_lastprivate_messages.cpp
index 16223db71ad5..349d6a5b904d 100644
--- a/test/OpenMP/simd_lastprivate_messages.cpp
+++ b/test/OpenMP/simd_lastprivate_messages.cpp
@@ -17,9 +17,9 @@ public:
   S2(S2 &s2) : a(s2.a) {}
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/simd_reduction_messages.cpp b/test/OpenMP/simd_reduction_messages.cpp
index 1e1a233ec499..bdf429d1967c 100644
--- a/test/OpenMP/simd_reduction_messages.cpp
+++ b/test/OpenMP/simd_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/single_firstprivate_codegen.cpp b/test/OpenMP/single_firstprivate_codegen.cpp
index 9f059e970470..4a91bc2d5bc5 100644
--- a/test/OpenMP/single_firstprivate_codegen.cpp
+++ b/test/OpenMP/single_firstprivate_codegen.cpp
@@ -222,24 +222,18 @@ int main() {
 // CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]()
 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
 // CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
-// CHECK: [[T_VARVAL:%.+]] = load i32, i32* [[T_VAR:%.+]],
-// CHECK: [[T_VARCONV:%.+]] = bitcast i64* [[T_VARCAST:%.+]] to i32*
-// CHECK: store i32 [[T_VARVAL]], i32* [[T_VARCONV]],
-// CHECK: [[T_VARPVT:%.+]] = load i64, i64* [[T_VARCAST]],
-// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i64, [2 x i32]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void {{.*}}i64 [[T_VARPVT:%.+]],
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i32*, [2 x i32]*, [2 x [[S_INT_TY]]]*, [[S_INT_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void
 // CHECK: call {{.*}} [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]*
 // CHECK: ret
 //
-// CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i64 {{.*}}%{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [2 x [[S_INT_TY]]]* dereferenceable(8) %{{.+}}, [[S_INT_TY]]* dereferenceable(4) %{{.+}})
-// CHECK: [[T_VAR_ARG:%.+]] = alloca i{{[0-9]+}},
+// CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i32* dereferenceable(4) %{{.+}}, [2 x i32]* dereferenceable(8) %{{.+}}, [2 x [[S_INT_TY]]]* dereferenceable(8) %{{.+}}, [[S_INT_TY]]* dereferenceable(4) %{{.+}})
 // CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
 // CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
 // CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
 // CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_ADDR:%.+]],
 
-// CHECK-NOT: load i{{[0-9]+}}*, i{{[0-9]+}}** %
-// CHECK: [[T_VAR_CONV:%.+]] = bitcast i64* [[T_VAR_ARG]] to i32*
+// CHECK: [[T_VAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %
 // CHECK: [[VEC_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** %
 // CHECK: [[S_ARR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** %
 // CHECK: [[VAR_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** %
diff --git a/test/OpenMP/target_codegen.cpp b/test/OpenMP/target_codegen.cpp
index c457045c17c7..20bc96b84cbf 100644
--- a/test/OpenMP/target_codegen.cpp
+++ b/test/OpenMP/target_codegen.cpp
@@ -34,16 +34,18 @@
 // code, only 6 will have mapped arguments, and only 4 have all-constant map
 // sizes.
 
-// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i{{32|64}}] [i[[SZ:32|64]] 2]
-// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: [[SIZET:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 0, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT:@.+]] = private unnamed_addr constant [2 x i64] [i64 32, i64 288]
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i{{32|64}}] [i[[SZ]] 2]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
-// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i32] [i32 288, i32 288]
-// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i32] [i32 288, i32 35, i32 288, i32 35, i32 35, i32 288, i32 288, i32 35, i32 35]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547]
 // CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
-// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i32] [i32 288, i32 288, i32 35]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
 // CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
-// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i32] [i32 288, i32 288, i32 288, i32 35]
-// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i32] [i32 35, i32 288, i32 288, i32 288, i32 35]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
 // CHECK-DAG: @{{.*}} = private constant i8 0
 // CHECK-DAG: @{{.*}} = private constant i8 0
 // CHECK-DAG: @{{.*}} = private constant i8 0
@@ -59,6 +61,7 @@
 // TCHECK: @{{.+}} = constant [[ENTTY]]
 // TCHECK: @{{.+}} = constant [[ENTTY]]
 // TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = {{.*}}constant [[ENTTY]]
 // TCHECK-NOT: @{{.+}} = constant [[ENTTY]]
 
 // Check if offloading descriptor is created.
@@ -79,6 +82,9 @@ struct TT{
   ty Y;
 };
 
+int global;
+extern int global;
+
 // CHECK: define {{.*}}[[FOO:@.+]](
 int foo(int n) {
   int a = 0;
@@ -88,31 +94,60 @@ int foo(int n) {
   double c[5][10];
   double cn[5][n];
   TT<long long, char> d;
+  static long *plocal;
 
-  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i32* null)
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK:       [[ADD:%.+]] = add nsw i32
+  // CHECK:       [[DEVICE:%.+]] = sext i32 [[ADD]] to i64
+  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target(i64 [[DEVICE]], i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i64* null)
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT0:@.+]]()
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
-  #pragma omp target
+  #pragma omp target device(global + a)
   {
   }
 
-  // CHECK:       store i32 0, i32* [[RHV:%.+]], align 4
-  // CHECK:       store i32 -1, i32* [[RHV]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK-DAG:   [[ADD:%.+]] = add nsw i32
+  // CHECK-DAG:   [[DEVICE:%.+]] = sext i32 [[ADD]] to i64
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_nowait(i64 [[DEVICE]], i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT]], i32 0, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]**
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]**
+  // CHECK-DAG:   store i[[SZ]]* [[BP0:%[^,]+]], i[[SZ]]** [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]]* [[BP0]], i[[SZ]]** [[CPADDR0]]
+
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[BP1:%[^,]+]], i[[SZ]]* [[CBPADDR1]]
+  // CHECK-DAG:   store i[[SZ]] [[BP1]], i[[SZ]]* [[CPADDR1]]
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT0_:@.+]](i[[SZ]]* [[BP0]], i[[SZ]] [[BP1]])
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target device(global + a) nowait
+  {
+    static int local1;
+    *plocal = global;
+    local1 = global;
+  }
+
   // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}})
-  #pragma omp target if(0)
+  #pragma omp target if(0) firstprivate(global)
   {
-    a += 1;
+    global += 1;
   }
 
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i32* getelementptr inbounds ([1 x i32], [1 x i32]* [[MAPT2]], i32 0, i32 0))
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT2]], i32 0, i32 0))
   // CHECK-DAG:   [[BP]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[P]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
@@ -122,9 +157,7 @@ int foo(int n) {
   // CHECK-DAG:   store i[[SZ]] [[BP0:%[^,]+]], i[[SZ]]* [[CBPADDR0]]
   // CHECK-DAG:   store i[[SZ]] [[P0:%[^,]+]], i[[SZ]]* [[CPADDR0]]
 
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}})
@@ -138,7 +171,7 @@ int foo(int n) {
   // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
   // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CHECK:       [[IFTHEN]]
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i32* getelementptr inbounds ([2 x i32], [2 x i32]* [[MAPT3]], i32 0, i32 0))
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0))
   // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
 
@@ -155,21 +188,18 @@ int foo(int n) {
   // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
   // CHECK-DAG:   store i[[SZ]] [[BP1:%[^,]+]], i[[SZ]]* [[CBPADDR1]]
   // CHECK-DAG:   store i[[SZ]] [[P1:%[^,]+]], i[[SZ]]* [[CPADDR1]]
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK-NEXT:  br label %[[IFEND:.+]]
-
-  // CHECK:       [[IFELSE]]
-  // CHECK:       store i32 -1, i32* [[RHV]], align 4
-  // CHECK-NEXT:  br label %[[IFEND:.+]]
-
-  // CHECK:       [[IFEND]]
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+
+  // CHECK:       [[IFEND]]
   #pragma omp target if(n>10)
   {
     a += 1;
@@ -191,9 +221,9 @@ int foo(int n) {
   // CHECK:       [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8
 
   // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
-  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[IFELSE:[^,]+]]
   // CHECK:       [[TRY]]
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i32* getelementptr inbounds ([9 x i32], [9 x i32]* [[MAPT4]], i32 0, i32 0))
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0))
   // CHECK-DAG:   [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
@@ -282,15 +312,18 @@ int foo(int n) {
   // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8]]
   // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* [[SADDR8]]
 
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
-  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
-
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT4]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+
+  // CHECK:       [[IFEND]]
   #pragma omp target if(n>20)
   {
     a += 1;
@@ -464,9 +497,9 @@ int bar(int n){
 // CHECK:       [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2
 
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
-// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[TRY]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i32* getelementptr inbounds ([5 x i32], [5 x i32]* [[MAPT7]], i32 0, i32 0))
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0))
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
@@ -507,9 +540,9 @@ int bar(int n){
 // CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* [[SADDR1]]
 
 // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to [[S1]]**
-// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to [[S1]]**
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to double**
 // CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR0]]
-// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR0]]
+// CHECK-DAG:   store double* %{{.+}}, double** [[CPADDR0]]
 // CHECK-DAG:   store i[[SZ]] 8, i[[SZ]]* [[SADDR0]]
 
 // CHECK-DAG:   [[CBPADDR4:%.+]] = bitcast i8** [[BPADDR4]] to i16**
@@ -518,15 +551,18 @@ int bar(int n){
 // CHECK-DAG:   store i16* %{{.+}}, i16** [[CPADDR4]]
 // CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* [[SADDR4]]
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
-// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
-
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT7]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+
+// CHECK:       [[IFEND]]
 
 //
 // CHECK: define {{.*}}[[FSTATIC]]
@@ -534,7 +570,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
 // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[IFTHEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i32* getelementptr inbounds ([4 x i32], [4 x i32]* [[MAPT6]], i32 0, i32 0))
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0))
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0
 
@@ -566,21 +602,18 @@ int bar(int n){
 // CHECK-DAG:   store [10 x i32]* [[VAL3:%[^,]+]], [10 x i32]** [[CBPADDR3]]
 // CHECK-DAG:   store [10 x i32]* [[VAL3]], [10 x i32]** [[CPADDR3]]
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFEND]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+
+// CHECK:       [[IFEND]]
 
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
@@ -588,7 +621,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
 // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[IFTHEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i32 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i32* getelementptr inbounds ([3 x i32], [3 x i32]* [[MAPT5]], i32 0, i32 0))
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0))
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
 
@@ -613,22 +646,18 @@ int bar(int n){
 // CHECK-DAG:   store [10 x i32]* [[VAL2:%[^,]+]], [10 x i32]** [[CBPADDR2]]
 // CHECK-DAG:   store [10 x i32]* [[VAL2]], [10 x i32]** [[CPADDR2]]
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFEND]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT5]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
 
+// CHECK:       [[IFEND]]
 
 
 // Check that the offloading functions are emitted and that the arguments are
diff --git a/test/OpenMP/target_codegen_registration.cpp b/test/OpenMP/target_codegen_registration.cpp
index 064c15b8d8f1..5684b2573b6e 100644
--- a/test/OpenMP/target_codegen_registration.cpp
+++ b/test/OpenMP/target_codegen_registration.cpp
@@ -63,40 +63,40 @@
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // TCHECK-NOT: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 
 // CHECK-NTARGET-NOT: private constant i8 0
 // CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
diff --git a/test/OpenMP/target_data_codegen.cpp b/test/OpenMP/target_data_codegen.cpp
index 7e5e267d13b6..2c9526de73e1 100644
--- a/test/OpenMP/target_data_codegen.cpp
+++ b/test/OpenMP/target_data_codegen.cpp
@@ -22,15 +22,15 @@ ST<int> gb;
 double gc[100];
 
 // CK1: [[SIZE00:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 800]
-// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i32] [i32 34]
+// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i64] [i64 34]
 
 // CK1: [[SIZE02:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] 4]
-// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i64] [i64 33]
 
-// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i32] [i32 37]
+// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i64] [i64 37]
 
 // CK1: [[SIZE04:@.+]] = {{.+}}constant [2 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] 24]
-// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i32] [i32 33, i32 17]
+// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i64] [i64 33, i64 17]
 
 // CK1-LABEL: _Z3fooi
 void foo(int arg) {
@@ -38,8 +38,9 @@ void foo(int arg) {
   float lb[arg];
 
   // Region 00
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
-  // CK1-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK1-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+  // CK1-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -52,8 +53,9 @@ void foo(int arg) {
 
   // CK1: %{{.+}} = add nsw i32 %{{[^,]+}}, 1
 
-  // CK1-DAG: call void @__tgt_target_data_end(i32 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
-  // CK1-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+  // CK1-DAG: call void @__tgt_target_data_end(i64 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK1-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+  // CK1-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P]]
   #pragma omp target data if(1+3-5) device(arg) map(from: gc)
@@ -67,7 +69,7 @@ void foo(int arg) {
   // Region 02
   // CK1: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CK1: [[IFTHEN]]
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -86,7 +88,7 @@ void foo(int arg) {
   // CK1: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 
   // CK1: [[IFTHEN]]
-  // CK1-DAG: call void @__tgt_target_data_end(i32 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_end(i64 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P]]
   // CK1: br label %[[IFEND:[^,]+]]
@@ -97,7 +99,7 @@ void foo(int arg) {
   {++arg;}
 
   // Region 03
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK1-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -113,7 +115,7 @@ void foo(int arg) {
   // CK1-DAG: [[CSVAL0]] = mul nuw i[[sz]] %{{[^,]+}}, 4
   // CK1: %{{.+}} = add nsw i32 %{{[^,]+}}, 1
 
-  // CK1-DAG: call void @__tgt_target_data_end(i32 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_end(i64 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P]]
   // CK1-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S]]
@@ -124,7 +126,7 @@ void foo(int arg) {
   {++arg;}
 
   // Region 04
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -147,7 +149,7 @@ void foo(int arg) {
 
   // CK1: %{{.+}} = add nsw i32 %{{[^,]+}}, 1
 
-  // CK1-DAG: call void @__tgt_target_data_end(i32 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_end(i64 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P]]
   #pragma omp target data map(to: gb.b[:3])
@@ -178,7 +180,7 @@ struct ST {
 };
 
 // CK2: [[SIZE00:@.+]] = {{.+}}constant [2 x i[[sz:64|32]]] [i{{64|32}} {{8|4}}, i{{64|32}} 24]
-// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i32] [i32 37, i32 21]
+// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i64] [i64 37, i64 21]
 
 // CK2-LABEL: _Z3bari
 int bar(int arg){
@@ -189,8 +191,9 @@ int bar(int arg){
 // Region 00
 // CK2: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CK2: [[IFTHEN]]
-// CK2-DAG: call void @__tgt_target_data_begin(i32 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
-// CK2-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+// CK2-DAG: call void @__tgt_target_data_begin(i64 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
+// CK2-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+// CK2-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
 // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -222,8 +225,9 @@ int bar(int arg){
 // CK2: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 
 // CK2: [[IFTHEN]]
-// CK2-DAG: call void @__tgt_target_data_end(i32 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
-// CK2-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+// CK2-DAG: call void @__tgt_target_data_end(i64 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
+// CK2-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+// CK2-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
 // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP]]
 // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P]]
 // CK2: br label %[[IFEND:[^,]+]]
diff --git a/test/OpenMP/target_data_use_device_ptr_codegen.cpp b/test/OpenMP/target_data_use_device_ptr_codegen.cpp
index 5e275565170e..36c42e32ea3f 100644
--- a/test/OpenMP/target_data_use_device_ptr_codegen.cpp
+++ b/test/OpenMP/target_data_use_device_ptr_codegen.cpp
@@ -14,18 +14,18 @@
 double *g;
 
 // CK1: @g = global double*
-// CK1: [[MTYPE00:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE01:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE03:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE04:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE05:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE06:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE07:@.+]] = {{.*}}constant [1 x i32] [i32 99]
-// CK1: [[MTYPE08:@.+]] = {{.*}}constant [2 x i32] [{{i32 35, i32 99|i32 99, i32 35}}]
-// CK1: [[MTYPE09:@.+]] = {{.*}}constant [2 x i32] [i32 99, i32 99]
-// CK1: [[MTYPE10:@.+]] = {{.*}}constant [2 x i32] [i32 99, i32 99]
-// CK1: [[MTYPE11:@.+]] = {{.*}}constant [2 x i32] [i32 96, i32 35]
-// CK1: [[MTYPE12:@.+]] = {{.*}}constant [2 x i32] [i32 96, i32 35]
+// CK1: [[MTYPE00:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE01:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE03:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE04:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE05:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE06:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE07:@.+]] = {{.*}}constant [1 x i64] [i64 99]
+// CK1: [[MTYPE08:@.+]] = {{.*}}constant [2 x i64] [{{i64 35, i64 99|i64 99, i64 35}}]
+// CK1: [[MTYPE09:@.+]] = {{.*}}constant [2 x i64] [i64 99, i64 99]
+// CK1: [[MTYPE10:@.+]] = {{.*}}constant [2 x i64] [i64 99, i64 99]
+// CK1: [[MTYPE11:@.+]] = {{.*}}constant [2 x i64] [i64 96, i64 35]
+// CK1: [[MTYPE12:@.+]] = {{.*}}constant [2 x i64] [i64 96, i64 35]
 
 // CK1-LABEL: @_Z3foo
 template<typename T>
@@ -330,10 +330,10 @@ void bar(float *&a, int *&b) {
 #ifdef CK2
 
 // CK2: [[ST:%.+]] = type { double*, double** }
-// CK2: [[MTYPE00:@.+]] = {{.*}}constant [2 x i32] [i32 35, i32 83]
-// CK2: [[MTYPE01:@.+]] = {{.*}}constant [3 x i32] [i32 32, i32 19, i32 83]
-// CK2: [[MTYPE02:@.+]] = {{.*}}constant [2 x i32] [i32 96, i32 35]
-// CK2: [[MTYPE03:@.+]] = {{.*}}constant [4 x i32] [i32 96, i32 32, i32 19, i32 83]
+// CK2: [[MTYPE00:@.+]] = {{.*}}constant [2 x i64] [i64 35, i64 83]
+// CK2: [[MTYPE01:@.+]] = {{.*}}constant [3 x i64] [i64 32, i64 19, i64 83]
+// CK2: [[MTYPE02:@.+]] = {{.*}}constant [2 x i64] [i64 96, i64 35]
+// CK2: [[MTYPE03:@.+]] = {{.*}}constant [4 x i64] [i64 96, i64 32, i64 19, i64 83]
 
 template <typename T>
 struct ST {
diff --git a/test/OpenMP/target_depend_messages.cpp b/test/OpenMP/target_depend_messages.cpp
index 48bd94180c1e..3eb3af132d8f 100644
--- a/test/OpenMP/target_depend_messages.cpp
+++ b/test/OpenMP/target_depend_messages.cpp
@@ -36,21 +36,21 @@ int main(int argc, char **argv, char *env[]) {
   foo();
   #pragma omp target depend (out: ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
   #pragma omp target depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-  #pragma omp target depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target depend(in : argv[1][1] = '2')
   foo();
-  #pragma omp target depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target depend (in : argv[0])
   foo();
   #pragma omp target depend (in : ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target depend (in : main)
   foo();
-  #pragma omp target depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
diff --git a/test/OpenMP/target_enter_data_codegen.cpp b/test/OpenMP/target_enter_data_codegen.cpp
index 683cd976d395..08b6c7c2702e 100644
--- a/test/OpenMP/target_enter_data_codegen.cpp
+++ b/test/OpenMP/target_enter_data_codegen.cpp
@@ -22,15 +22,15 @@ ST<int> gb;
 double gc[100];
 
 // CK1: [[SIZE00:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 800]
-// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i32] [i32 32]
+// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i64] [i64 32]
 
 // CK1: [[SIZE02:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] 4]
-// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i64] [i64 33]
 
-// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i32] [i32 37]
+// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i64] [i64 37]
 
 // CK1: [[SIZE04:@.+]] = {{.+}}constant [2 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] 24]
-// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i32] [i32 33, i32 17]
+// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i64] [i64 33, i64 17]
 
 // CK1-LABEL: _Z3fooi
 void foo(int arg) {
@@ -38,8 +38,9 @@ void foo(int arg) {
   float lb[arg];
 
   // Region 00
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
-  // CK1-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+  // CK1-DAG: call void @__tgt_target_data_begin_nowait(i64 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK1-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+  // CK1-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -52,7 +53,7 @@ void foo(int arg) {
 
   // CK1: %{{.+}} = add nsw i32 %{{[^,]+}}, 1
   // CK1-NOT: __tgt_target_data_end
-  #pragma omp target enter data if(1+3-5) device(arg) map(alloc: gc)
+  #pragma omp target enter data if(1+3-5) device(arg) map(alloc: gc) nowait
   {++arg;}
 
   // Region 01
@@ -63,7 +64,7 @@ void foo(int arg) {
   // Region 02
   // CK1: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CK1: [[IFTHEN]]
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -87,7 +88,7 @@ void foo(int arg) {
   {++arg;}
 
   // Region 03
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK1-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -110,7 +111,7 @@ void foo(int arg) {
   {++arg;}
 
   // Region 04
-  // CK1-DAG: call void @__tgt_target_data_begin(i32 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_begin(i64 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -161,7 +162,7 @@ struct ST {
 };
 
 // CK2: [[SIZE00:@.+]] = {{.+}}constant [2 x i[[sz:64|32]]] [i{{64|32}} {{8|4}}, i{{64|32}} 24]
-// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i32] [i32 37, i32 21]
+// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i64] [i64 37, i64 21]
 
 // CK2-LABEL: _Z3bari
 int bar(int arg){
@@ -172,8 +173,9 @@ int bar(int arg){
 // Region 00
 // CK2: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CK2: [[IFTHEN]]
-// CK2-DAG: call void @__tgt_target_data_begin(i32 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
-// CK2-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+// CK2-DAG: call void @__tgt_target_data_begin(i64 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
+// CK2-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+// CK2-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
 // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
diff --git a/test/OpenMP/target_enter_data_depend_messages.cpp b/test/OpenMP/target_enter_data_depend_messages.cpp
index 4aa122355ba5..77e7039e8e20 100644
--- a/test/OpenMP/target_enter_data_depend_messages.cpp
+++ b/test/OpenMP/target_enter_data_depend_messages.cpp
@@ -38,21 +38,21 @@ int tmain(T argc, S **argv, R *env[]) {
   foo();
   #pragma omp target enter data map(to: i) depend (out: ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target enter data map(to: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target enter data map(to: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
   #pragma omp target enter data map(to: i) depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-  #pragma omp target enter data map(to: i) depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend(in : argv[1][1] = '2') // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
-  #pragma omp target enter data map(to: i) depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target enter data map(to: i) depend (in : argv[0])
   foo();
   #pragma omp target enter data map(to: i) depend (in : ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target enter data map(to: i) depend (in : tmain) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend (in : tmain)
   foo();
-  #pragma omp target enter data map(to: i) depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target enter data map(to: i) depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
@@ -113,21 +113,21 @@ int main(int argc, char **argv, char *env[]) {
   foo();
   #pragma omp target enter data map(to: i) depend (out: ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target enter data map(to: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target enter data map(to: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
   #pragma omp target enter data map(to: i) depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-  #pragma omp target enter data map(to: i) depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend(in : argv[1][1] = '2')
   foo();
-  #pragma omp target enter data map(to: i) depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target enter data map(to: i) depend (in : argv[0])
   foo();
   #pragma omp target enter data map(to: i) depend (in : ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target enter data map(to: i) depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend (in : main)
   foo();
-  #pragma omp target enter data map(to: i) depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target enter data map(to: i) depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target enter data map(to: i) depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
diff --git a/test/OpenMP/target_exit_data_codegen.cpp b/test/OpenMP/target_exit_data_codegen.cpp
index a82c1c6c4aac..9359e3be451b 100644
--- a/test/OpenMP/target_exit_data_codegen.cpp
+++ b/test/OpenMP/target_exit_data_codegen.cpp
@@ -22,15 +22,15 @@ ST<int> gb;
 double gc[100];
 
 // CK1: [[SIZE00:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 800]
-// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i32] [i32 34]
+// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i64] [i64 34]
 
 // CK1: [[SIZE02:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] 4]
-// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i32] [i32 32]
+// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i64] [i64 32]
 
-// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i32] [i32 38]
+// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i64] [i64 38]
 
 // CK1: [[SIZE04:@.+]] = {{.+}}constant [2 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] 24]
-// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i32] [i32 32, i32 16]
+// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i64] [i64 32, i64 16]
 
 // CK1-LABEL: _Z3fooi
 void foo(int arg) {
@@ -39,8 +39,9 @@ void foo(int arg) {
 
   // Region 00
   // CK1-NOT: __tgt_target_data_begin
-  // CK1-DAG: call void @__tgt_target_data_end(i32 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
-  // CK1-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+  // CK1-DAG: call void @__tgt_target_data_end_nowait(i64 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK1-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+  // CK1-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -52,7 +53,7 @@ void foo(int arg) {
   // CK1-DAG: store [100 x double]* @gc, [100 x double]** [[PC0]]
 
   // CK1: %{{.+}} = add nsw i32 %{{[^,]+}}, 1
-  #pragma omp target exit data if(1+3-5) device(arg) map(from: gc)
+  #pragma omp target exit data if(1+3-5) device(arg) map(from: gc) nowait
   {++arg;}
 
   // Region 01
@@ -64,7 +65,7 @@ void foo(int arg) {
   // CK1-NOT: __tgt_target_data_begin
   // CK1: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CK1: [[IFTHEN]]
-  // CK1-DAG: call void @__tgt_target_data_end(i32 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_end(i64 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -88,7 +89,7 @@ void foo(int arg) {
 
   // Region 03
   // CK1-NOT: __tgt_target_data_begin
-  // CK1-DAG: call void @__tgt_target_data_end(i32 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_end(i64 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK1-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -111,7 +112,7 @@ void foo(int arg) {
 
   // Region 04
   // CK1-NOT: __tgt_target_data_begin
-  // CK1-DAG: call void @__tgt_target_data_end(i32 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_end(i64 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -161,7 +162,7 @@ struct ST {
 };
 
 // CK2: [[SIZE00:@.+]] = {{.+}}constant [2 x i[[sz:64|32]]] [i{{64|32}} {{8|4}}, i{{64|32}} 24]
-// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i32] [i32 36, i32 20]
+// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i64] [i64 36, i64 20]
 
 // CK2-LABEL: _Z3bari
 int bar(int arg){
@@ -173,8 +174,9 @@ int bar(int arg){
 // CK2-NOT: __tgt_target_data_begin
 // CK2: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CK2: [[IFTHEN]]
-// CK2-DAG: call void @__tgt_target_data_end(i32 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
-// CK2-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+// CK2-DAG: call void @__tgt_target_data_end(i64 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
+// CK2-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+// CK2-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
 // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
diff --git a/test/OpenMP/target_exit_data_depend_messages.cpp b/test/OpenMP/target_exit_data_depend_messages.cpp
index cdefbeed08bd..ce8d237f75e9 100644
--- a/test/OpenMP/target_exit_data_depend_messages.cpp
+++ b/test/OpenMP/target_exit_data_depend_messages.cpp
@@ -38,21 +38,21 @@ int tmain(T argc, S **argv, R *env[]) {
   foo();
   #pragma omp target exit data map(from: i) depend (out: ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target exit data map(from: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target exit data map(from: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
   #pragma omp target exit data map(from: i) depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-  #pragma omp target exit data map(from: i) depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend(in : argv[1][1] = '2') // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
-  #pragma omp target exit data map(from: i) depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target exit data map(from: i) depend (in : argv[0])
   foo();
   #pragma omp target exit data map(from: i) depend (in : ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target exit data map(from: i) depend (in : tmain) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend (in : tmain)
   foo();
-  #pragma omp target exit data map(from: i) depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target exit data map(from: i) depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
@@ -113,21 +113,21 @@ int main(int argc, char **argv, char *env[]) {
   foo();
   #pragma omp target exit data map(from: i) depend (out: ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target exit data map(from: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target exit data map(from: i) depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
   #pragma omp target exit data map(from: i) depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-  #pragma omp target exit data map(from: i) depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend(in : argv[1][1] = '2')
   foo();
-  #pragma omp target exit data map(from: i) depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target exit data map(from: i) depend (in : argv[0])
   foo();
   #pragma omp target exit data map(from: i) depend (in : ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target exit data map(from: i) depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend (in : main)
   foo();
-  #pragma omp target exit data map(from: i) depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target exit data map(from: i) depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target exit data map(from: i) depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
diff --git a/test/OpenMP/target_firstprivate_codegen.cpp b/test/OpenMP/target_firstprivate_codegen.cpp
index 1fb0ef9e848c..c2909c6f18b8 100644
--- a/test/OpenMP/target_firstprivate_codegen.cpp
+++ b/test/OpenMP/target_firstprivate_codegen.cpp
@@ -33,15 +33,15 @@ struct TT{
 // TCHECK:  [[S1:%.+]] = type { double }
 
 // CHECK-DAG:  [[SIZET:@.+]] = private unnamed_addr constant [1 x i{{32|64}}] [i[[SZ:32|64]] 4]
-// CHECK:  [[MAPT:@.+]] = private unnamed_addr constant [1 x i32] [i32 288]
-// CHECK-DAG:  [[MAPT2:@.+]] = private unnamed_addr constant [9 x i32] [i32 288, i32 161, i32 288, i32 161, i32 161, i32 288, i32 288, i32 161, i32 161]
+// CHECK:  [[MAPT:@.+]] = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG:  [[MAPT2:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 161, i64 288, i64 161, i64 161, i64 288, i64 288, i64 161, i64 161]
 // CHECK-DAG:  [[SIZET3:@.+]] = private unnamed_addr constant [1 x i{{32|64}}] zeroinitializer
-// CHECK-DAG:  [[MAPT3:@.+]] = private unnamed_addr constant [1 x i32] [i32 32]
-// CHECK-DAG:  [[MAPT4:@.+]] = private unnamed_addr constant [5 x i32] [i32 35, i32 288, i32 288, i32 288, i32 161]
+// CHECK-DAG:  [[MAPT3:@.+]] = private unnamed_addr constant [1 x i64] [i64 32]
+// CHECK-DAG:  [[MAPT4:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 161]
 // CHECK-DAG:  [[SIZET5:@.+]] = private unnamed_addr constant [3 x i{{32|64}}] [i[[SZ]] 4, i[[SZ]] 1, i[[SZ]] 40]
-// CHECK-DAG:  [[MAPT5:@.+]] = private unnamed_addr constant [3 x i32] [i32 288, i32 288, i32 161]
+// CHECK-DAG:  [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 161]
 // CHECK-DAG:  [[SIZET6:@.+]] = private unnamed_addr constant [2 x i{{32|64}}] [i[[SZ]] 4, i[[SZ]] 40]
-// CHECK-DAG:  [[MAPT6:@.+]] = private unnamed_addr constant [2 x i32] [i32 288, i32 161]
+// CHECK-DAG:  [[MAPT6:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 161]
 
 
 // CHECK: define {{.*}}[[FOO:@.+]](
@@ -68,7 +68,6 @@ int foo(int n, double *ptr) {
   // CHECK:  [[C:%.+]] = alloca [5 x [10 x double]],
   // CHECK:  [[D:%.+]] = alloca [[TT]],
   // CHECK:  [[ACAST:%.+]] = alloca i{{[0-9]+}},
-  // CHECK:  {{.+}} = alloca i{{[0-9]+}},
   // CHECK:  [[BASE_PTR_ARR:%.+]] = alloca [1 x i8*],
   // CHECK:  [[PTR_ARR:%.+]] = alloca [1 x i8*],
   // CHECK:  [[A2CAST:%.+]] = alloca i{{[0-9]+}},
@@ -101,7 +100,7 @@ int foo(int n, double *ptr) {
   // CHECK:  store i{{[0-9]+}} [[ACAST_VAL]], i{{[0-9]+}}* [[ACAST_TOPTR2]],
   // CHECK:  [[BASE_PTR_GEP_ARG:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BASE_PTR_ARR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
   // CHECK:  [[PTR_GEP_ARG:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PTR_ARR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
-  // CHECK:  {{.+}} = call i32 @__tgt_target(i32 -1, {{.+}}, i32 1, i8** [[BASE_PTR_GEP_ARG]], i8** [[PTR_GEP_ARG]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i32* getelementptr inbounds ([1 x i32], [1 x i32]* [[MAPT]], i32 0, i32 0))
+  // CHECK:  {{.+}} = call i32 @__tgt_target(i64 -1, {{.+}}, i32 1, i8** [[BASE_PTR_GEP_ARG]], i8** [[PTR_GEP_ARG]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT]], i32 0, i32 0))
   
   // TCHECK:  define void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A_IN:%.+]])
   // TCHECK:  [[A_ADDR:%.+]] = alloca i{{[0-9]+}},
@@ -220,11 +219,11 @@ int foo(int n, double *ptr) {
   // CHECK:  [[BASE_PTR_GEP_ARG2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BASE_PTR_ARR2]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
   // CHECK:  [[PTR_GEP_ARG2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[PTR_ARR2]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
   // CHECK:  [[SIZES_ARG2:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[SIZET2]],  i{{[0-9]+}} 0, i{{[0-9]+}} 0
-  // CHECK: {{.+}} = call i32 @__tgt_target(i32 -1, {{.+}}, i32 9, i8** [[BASE_PTR_GEP_ARG2]], i8** [[PTR_GEP_ARG2]], i[[SZ]]* [[SIZES_ARG2]], i32* getelementptr inbounds ([9 x i32], [9 x i32]* [[MAPT2]], i32 0, i32 0))
+  // CHECK: {{.+}} = call i32 @__tgt_target(i64 -1, {{.+}}, i32 9, i8** [[BASE_PTR_GEP_ARG2]], i8** [[PTR_GEP_ARG2]], i[[SZ]]* [[SIZES_ARG2]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT2]], i32 0, i32 0))
   
   // make sure that firstprivate variables are generated in all cases and that we use those instances for operations inside the
   // target region
-  // TCHECK:  define {{.*}}void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A2_IN:%.+]], [10 x float]* {{.+}} [[B_IN:%.+]], i{{[0-9]+}} [[BN_SZ:%.+]], float* [[BN_IN:%.+]], [5 x [10 x double]]* {{.+}} [[C_IN:%.+]], i{{[0-9]+}} [[CN_SZ1:%.+]], i{{[0-9]+}} [[CN_SZ2:%.+]], double* [[CN_IN:%.+]], [[TT]]* {{.+}} [[D_IN:%.+]])
+  // TCHECK:  define {{.*}}void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A2_IN:%.+]], [10 x float]* {{.+}} [[B_IN:%.+]], i{{[0-9]+}} [[BN_SZ:%.+]], float* {{.+}} [[BN_IN:%.+]], [5 x [10 x double]]* {{.+}} [[C_IN:%.+]], i{{[0-9]+}} [[CN_SZ1:%.+]], i{{[0-9]+}} [[CN_SZ2:%.+]], double* {{.+}} [[CN_IN:%.+]], [[TT]]* {{.+}} [[D_IN:%.+]])
   // TCHECK:  [[A2_ADDR:%.+]] = alloca i{{[0-9]+}},
   // TCHECK:  [[B_ADDR:%.+]] = alloca [10 x float]*,
   // TCHECK:  [[VLA_ADDR:%.+]] = alloca i{{[0-9]+}},
@@ -311,7 +310,7 @@ int foo(int n, double *ptr) {
 
   // CHECK:  [[BASE_PTR_GEP_ARG3:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BASE_PTR_ARR3]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
   // CHECK:  [[PTR_GEP_ARG3:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PTR_ARR3]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
-  // CHECK: {{.+}} = call i32 @__tgt_target(i32 -1, {{.+}}, i32 1, i8** [[BASE_PTR_GEP_ARG3]], i8** [[PTR_GEP_ARG3]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i32* getelementptr inbounds ([1 x i32], [1 x i32]* [[MAPT3]], i32 0, i32 0))
+  // CHECK: {{.+}} = call i32 @__tgt_target(i64 -1, {{.+}}, i32 1, i8** [[BASE_PTR_GEP_ARG3]], i8** [[PTR_GEP_ARG3]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT3]], i32 0, i32 0))
 
   // TCHECK:  define void @__omp_offloading_{{.+}}(double* [[PTR_IN:%.+]])
   // TCHECK:  [[PTR_ADDR:%.+]] = alloca double*,
@@ -447,7 +446,7 @@ struct S1 {
   // CHECK:  store i{{[0-9]+}} [[B_SIZE:%.+]], i{{[0-9]+}}* [[SIZES_GEP4_4]],
 
   // only check that we use the map types stored in the global variable
-  // CHECK:  call i32 @__tgt_target(i32 -1, {{.+}}, i32 5, i8** {{.+}}, i8** {{.+}}, i{{[0-9]+}}* {{.+}}, i32* getelementptr inbounds ([5 x i32], [5 x i32]* [[MAPT4]], i32 0, i32 0))
+  // CHECK:  call i32 @__tgt_target(i64 -1, {{.+}}, i32 5, i8** {{.+}}, i8** {{.+}}, i{{[0-9]+}}* {{.+}}, i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT4]], i32 0, i32 0))
   
   // TCHECK: define void @__omp_offloading_{{.+}}([[S1]]* [[TH:%.+]], i{{[0-9]+}} [[B_IN:%.+]], i{{[0-9]+}} [[VLA:%.+]], i{{[0-9]+}} [[VLA1:%.+]], i{{[0-9]+}}{{.+}} [[C_IN:%.+]])
   // TCHECK:  [[TH_ADDR:%.+]] = alloca [[S1]]*,
@@ -520,7 +519,7 @@ struct S1 {
   // CHECK:  store [10 x i{{[0-9]+}}]* [[B]], [10 x i{{[0-9]+}}]** [[BCAST_TOPTR]],
 
   // only check that the right sizes and map types are used
-  // CHECK:  call i32 @__tgt_target(i32 -1, {{.+}}, i32 3, i8** {{.+}}, i8** {{.+}}, i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i32* getelementptr inbounds ([3 x i32], [3 x i32]* [[MAPT5]], i32 0, i32 0))
+  // CHECK:  call i32 @__tgt_target(i64 -1, {{.+}}, i32 3, i8** {{.+}}, i8** {{.+}}, i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0))
 };
 
 
@@ -558,7 +557,7 @@ int bar(int n, double *ptr){
 // CHECK:  [[BCAST_TOPTR:%.+]] = bitcast i8** [[PTRS_GEP6_1]] to [10 x i{{[0-9]+}}]**
 // CHECK:  store [10 x i{{[0-9]+}}]* [[B]], [10 x i{{[0-9]+}}]** [[BCAST_TOPTR]],
 
-// CHECK:  call i32 @__tgt_target(i32 -1, {{.+}}, i32 2, i8** {{.+}}, i8** {{.+}}, i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i32* getelementptr inbounds ([2 x i32], [2 x i32]* [[MAPT6]], i32 0, i32 0))
+// CHECK:  call i32 @__tgt_target(i64 -1, {{.+}}, i32 2, i8** {{.+}}, i8** {{.+}}, i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT6]], i32 0, i32 0))
 
 
 // TCHECK: define void @__omp_offloading_{{.+}}(i{{[0-9]+}} [[A_IN:%.+]], [10 x i{{[0-9]+}}]*{{.+}} [[B_IN:%.+]])
diff --git a/test/OpenMP/target_is_device_ptr_codegen.cpp b/test/OpenMP/target_is_device_ptr_codegen.cpp
index 1a54aa18c00b..1491f41a787f 100644
--- a/test/OpenMP/target_is_device_ptr_codegen.cpp
+++ b/test/OpenMP/target_is_device_ptr_codegen.cpp
@@ -15,25 +15,25 @@ double *g;
 
 // CK1: @g = global double*
 // CK1: [[SIZES00:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} {{8|4}}]
-// CK1: [[TYPES00:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1: [[TYPES00:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1: [[SIZES01:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] {{8|4}}]
-// CK1: [[TYPES01:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1: [[TYPES01:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1: [[SIZES02:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] {{8|4}}]
-// CK1: [[TYPES02:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1: [[TYPES02:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1: [[SIZES03:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] {{8|4}}]
-// CK1: [[TYPES03:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1: [[TYPES03:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1: [[SIZES04:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] {{8|4}}]
-// CK1: [[TYPES04:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1: [[TYPES04:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1: [[SIZES05:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] {{8|4}}]
-// CK1: [[TYPES05:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1: [[TYPES05:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1: [[SIZES06:@.+]] = {{.+}}constant [2 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] {{8|4}}]
-// CK1: [[TYPES06:@.+]] = {{.+}}constant [2 x i32] [i32 288, i32 288]
+// CK1: [[TYPES06:@.+]] = {{.+}}constant [2 x i64] [i64 288, i64 288]
 
 // CK1-LABEL: @_Z3foo
 template<typename T>
@@ -41,7 +41,7 @@ void foo(float *&lr, T *&tr) {
   float *l;
   T *t;
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES00]]{{.+}}, {{.+}}[[TYPES00]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES00]]{{.+}}, {{.+}}[[TYPES00]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -58,7 +58,7 @@ void foo(float *&lr, T *&tr) {
     ++g;
   }
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES01]]{{.+}}, {{.+}}[[TYPES01]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES01]]{{.+}}, {{.+}}[[TYPES01]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -75,7 +75,7 @@ void foo(float *&lr, T *&tr) {
     ++l;
   }
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES02]]{{.+}}, {{.+}}[[TYPES02]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES02]]{{.+}}, {{.+}}[[TYPES02]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -92,7 +92,7 @@ void foo(float *&lr, T *&tr) {
     ++t;
   }
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES03]]{{.+}}, {{.+}}[[TYPES03]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES03]]{{.+}}, {{.+}}[[TYPES03]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -110,7 +110,7 @@ void foo(float *&lr, T *&tr) {
     ++lr;
   }
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES04]]{{.+}}, {{.+}}[[TYPES04]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES04]]{{.+}}, {{.+}}[[TYPES04]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -128,7 +128,7 @@ void foo(float *&lr, T *&tr) {
     ++tr;
   }
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES05]]{{.+}}, {{.+}}[[TYPES05]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES05]]{{.+}}, {{.+}}[[TYPES05]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -146,7 +146,7 @@ void foo(float *&lr, T *&tr) {
     ++tr;
   }
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 2, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES06]]{{.+}}, {{.+}}[[TYPES06]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 2, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES06]]{{.+}}, {{.+}}[[TYPES06]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -191,13 +191,13 @@ void bar(float *&a, int *&b) {
 // CK2: [[ST:%.+]] = type { double*, double** }
 
 // CK2: [[SIZE00:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} {{8|4}}]
-// CK2: [[MTYPE00:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// CK2: [[MTYPE00:@.+]] = {{.+}}constant [1 x i64] [i64 33]
 
 // CK2: [[SIZE01:@.+]] = {{.+}}constant [2 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] {{8|4}}]
-// CK2: [[MTYPE01:@.+]] = {{.+}}constant [2 x i32] [i32 32, i32 17]
+// CK2: [[MTYPE01:@.+]] = {{.+}}constant [2 x i64] [i64 32, i64 17]
 
 // CK2: [[SIZE02:@.+]] = {{.+}}constant [3 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] {{8|4}}, i[[sz]] {{8|4}}]
-// CK2: [[MTYPE02:@.+]] = {{.+}}constant [3 x i32] [i32 33, i32 0, i32 17]
+// CK2: [[MTYPE02:@.+]] = {{.+}}constant [3 x i64] [i64 33, i64 0, i64 17]
 
 template <typename T>
 struct ST {
@@ -209,7 +209,7 @@ struct ST {
   void foo(double *&arg) {
     int *la = 0;
 
-    // CK2-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+    // CK2-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
     // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -225,7 +225,7 @@ struct ST {
       a++;
     }
 
-    // CK2-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE01]]{{.+}})
+    // CK2-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE01]]{{.+}})
     // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -249,7 +249,7 @@ struct ST {
       b++;
     }
 
-    // CK2-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE02]]{{.+}})
+    // CK2-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE02]]{{.+}})
     // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
diff --git a/test/OpenMP/target_map_codegen.cpp b/test/OpenMP/target_map_codegen.cpp
index 69e80bb9505b..34b239504462 100644
--- a/test/OpenMP/target_map_codegen.cpp
+++ b/test/OpenMP/target_map_codegen.cpp
@@ -15,15 +15,33 @@
 // RUN: %clang_cc1 -fopenmp -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s  --check-prefix CK1 --check-prefix CK1-32
 #ifdef CK1
 
+class B {
+public:
+  static double VAR;
+  B() {
+  }
+
+  static void modify(int &res) {
+#pragma omp target map(tofrom \
+                       : res)
+    {
+      res = B::VAR;
+    }
+  }
+};
+double B::VAR = 1.0;
+
 // CK1-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK1-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK1-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK1-LABEL: implicit_maps_integer
 void implicit_maps_integer (int a){
+  // CK1: call void{{.*}}modify
+  B::modify(a);
   int i = a;
 
-  // CK1-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK1-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK1-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK1-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -62,15 +80,15 @@ void implicit_maps_integer (int a){
 
 // CK2: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK2: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK2: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 // CK2: [[SIZES2:@.+]] = {{.+}}constant [1 x i[[sz]]] zeroinitializer
 // Map types: OMP_MAP_IS_PTR = 32
-// CK2: [[TYPES2:@.+]] = {{.+}}constant [1 x i32] [i32 32]
+// CK2: [[TYPES2:@.+]] = {{.+}}constant [1 x i64] [i64 32]
 
 // CK2-LABEL: implicit_maps_reference
 void implicit_maps_reference (int a, int *b){
   int &i = a;
-  // CK2-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK2-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK2-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK2-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK2-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -90,7 +108,7 @@ void implicit_maps_reference (int a, int *b){
   }
 
   int *&p = b;
-  // CK2-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES2]]{{.+}}, {{.+}}[[TYPES2]]{{.+}})
+  // CK2-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES2]]{{.+}}, {{.+}}[[TYPES2]]{{.+}})
   // CK2-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK2-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK2-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -141,12 +159,12 @@ void implicit_maps_reference (int a, int *b){
 
 // CK3-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK3-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK3-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK3-LABEL: implicit_maps_parameter
 void implicit_maps_parameter (int a){
 
-  // CK3-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK3-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK3-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK3-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK3-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -185,7 +203,7 @@ void implicit_maps_parameter (int a){
 
 // CK4-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK4-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK4-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK4-LABEL: implicit_maps_nested_integer
 void implicit_maps_nested_integer (int a){
@@ -198,7 +216,7 @@ void implicit_maps_nested_integer (int a){
   // CK4: define internal void [[KERNELP1]](i32* {{[^,]+}}, i32* {{[^,]+}}, i32* {{[^,]+}})
   #pragma omp parallel
   {
-    // CK4-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+    // CK4-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
     // CK4-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
     // CK4-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
     // CK4-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -241,7 +259,7 @@ void implicit_maps_nested_integer (int a){
 
 // CK5-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK5-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK5-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK5-LABEL: implicit_maps_nested_integer_and_enum
 void implicit_maps_nested_integer_and_enum (int a){
@@ -252,7 +270,7 @@ void implicit_maps_nested_integer_and_enum (int a){
   // Using an enum should not change the mapping information.
   int  i = a;
 
-  // CK5-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK5-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK5-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK5-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK5-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -292,12 +310,12 @@ void implicit_maps_nested_integer_and_enum (int a){
 // CK6-DAG: [[GBL:@Gi]] = global i32 0
 // CK6-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK6-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK6-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK6-LABEL: implicit_maps_host_global
 int Gi;
 void implicit_maps_host_global (int a){
-  // CK6-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK6-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK6-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK6-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK6-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -341,15 +359,15 @@ void implicit_maps_host_global (int a){
 
 // CK7-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 8]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK7-64-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
-// Map types: OMP_MAP_TO  | OMP_MAP_IS_FIRST = 33
-// CK7-32-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// CK7-64-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
+// Map types: OMP_MAP_TO  | OMP_MAP_PRIVATE_PTR | OMP_MAP_FIRST_REF = 161
+// CK7-32-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 161]
 
 // CK7-LABEL: implicit_maps_double
 void implicit_maps_double (int a){
   double d = (double)a;
 
-  // CK7-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK7-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK7-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK7-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK7-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -400,13 +418,13 @@ void implicit_maps_double (int a){
 
 // CK8-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_IS_FIRST = 288
-// CK8-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK8-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 // CK8-LABEL: implicit_maps_float
 void implicit_maps_float (int a){
   float f = (float)a;
 
-  // CK8-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK8-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK8-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK8-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK8-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -443,14 +461,14 @@ void implicit_maps_float (int a){
 #ifdef CK9
 
 // CK9-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 16]
-// Map types: OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST = 35
-// CK9-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 35]
+// Map types: OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST + OMP_MAP_IMPLICIT = 547
+// CK9-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 547]
 
 // CK9-LABEL: implicit_maps_array
 void implicit_maps_array (int a){
   double darr[2] = {(double)a, (double)a};
 
-  // CK9-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK9-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK9-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK9-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK9-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -485,13 +503,13 @@ void implicit_maps_array (int a){
 
 // CK10-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] zeroinitializer
 // Map types: OMP_MAP_IS_FIRST = 32
-// CK10-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 32]
+// CK10-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 32]
 
 // CK10-LABEL: implicit_maps_pointer
 void implicit_maps_pointer (){
   double *ddyn;
 
-  // CK10-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK10-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK10-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK10-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK10-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -526,14 +544,14 @@ void implicit_maps_pointer (){
 #ifdef CK11
 
 // CK11-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 16]
-// Map types: OMP_MAP_TO + OMP_MAP_IS_FIRST = 33
-// CK11-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// Map types: OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST + OMP_MAP_IMPLICIT = 547
+// CK11-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 547]
 
 // CK11-LABEL: implicit_maps_double_complex
 void implicit_maps_double_complex (int a){
   double _Complex dc = (double)a;
 
-  // CK11-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK11-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK11-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK11-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK11-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -544,7 +562,7 @@ void implicit_maps_double_complex (int a){
   // CK11-DAG: store { double, double }* [[PTR]], { double, double }** [[CP1]]
 
   // CK11: call void [[KERNEL:@.+]]({ double, double }* [[PTR]])
-  #pragma omp target
+  #pragma omp target defaultmap(tofrom:scalar)
   {
    dc *= dc;
   }
@@ -570,15 +588,15 @@ void implicit_maps_double_complex (int a){
 
 // CK12-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 8]
 // Map types: OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288
-// CK12-64-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
-// Map types: OMP_MAP_TO + OMP_MAP_IS_FIRST = 33
-// CK12-32-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// CK12-64-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
+// Map types: OMP_MAP_TO  | OMP_MAP_PRIVATE_PTR | OMP_MAP_FIRST_REF = 161
+// CK12-32-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 161]
 
 // CK12-LABEL: implicit_maps_float_complex
 void implicit_maps_float_complex (int a){
   float _Complex fc = (float)a;
 
-  // CK12-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK12-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK12-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK12-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK12-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -630,14 +648,14 @@ void implicit_maps_float_complex (int a){
 // Map types:
 //  - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288 (vla size)
 //  - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288 (vla size)
-//  - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST = 35
-// CK13-DAG: [[TYPES:@.+]] = {{.+}}constant [3 x i32] [i32 288, i32 288, i32 35]
+//  - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST + OMP_IMPICIT_MAP = 547
+// CK13-DAG: [[TYPES:@.+]] = {{.+}}constant [3 x i64] [i64 288, i64 288, i64 547]
 
 // CK13-LABEL: implicit_maps_variable_length_array
 void implicit_maps_variable_length_array (int a){
   double vla[2][a];
 
-  // CK13-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 3, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], i[[sz:64|32]]* [[SGEP:%[^,]+]], {{.+}}[[TYPES]]{{.+}})
+  // CK13-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 3, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], i[[sz:64|32]]* [[SGEP:%[^,]+]], {{.+}}[[TYPES]]{{.+}})
   // CK13-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK13-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK13-DAG: [[SGEP]] = getelementptr inbounds {{.+}}[[SS:%[^,]+]], i32 0, i32 0
@@ -699,11 +717,12 @@ void implicit_maps_variable_length_array (int a){
 #ifdef CK14
 
 // CK14-DAG: [[ST:%.+]] = type { i32, double }
-// CK14-DAG: [[SIZES:@.+]] = {{.+}}constant [2 x i[[sz:64|32]]] [i{{64|32}} {{16|12}}, i{{64|32}} 4]
+// CK14-DAG: [[SIZES:@.+]] = {{.+}}constant [3 x i[[sz:64|32]]] [i{{64|32}} 4, i{{64|32}} 8, i{{64|32}} 4]
 // Map types:
-// - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST = 35
+// - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST + OMP_IMPLICIT_MAP = 547
+// - OMP_MAP_TO + OMP_MAP_FROM + OMP_IMPLICIT_MAP = 515
 // - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288
-// CK14-DAG: [[TYPES:@.+]] = {{.+}}constant [2 x i32] [i32 35, i32 288]
+// CK14-DAG: [[TYPES:@.+]] = {{.+}}constant [3 x i64] [i64 547, i64 515, i64 288]
 
 class SSS {
 public:
@@ -726,19 +745,26 @@ void implicit_maps_class (int a){
   SSS sss(a, (double)a);
 
   // CK14: define {{.*}}void @{{.+}}foo{{.+}}([[ST]]* {{[^,]+}}, i32 {{[^,]+}})
-  // CK14-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 2, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK14-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 3, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK14-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK14-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
 
   // CK14-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
   // CK14-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 0
   // CK14-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[ST]]**
-  // CK14-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[ST]]**
+  // CK14-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
   // CK14-DAG: store [[ST]]* [[DECL:%.+]], [[ST]]** [[CBP0]]
-  // CK14-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CP0]]
+  // CK14-DAG: store i32* %{{.+}}, i32** [[CP0]]
 
   // CK14-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 1
   // CK14-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 1
+  // CK14-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[ST]]**
+  // CK14-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to double**
+  // CK14-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CBP1]]
+  // CK14-DAG: store double* %{{.+}}, double** [[CP1]]
+
+  // CK14-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 2
+  // CK14-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 2
   // CK14-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to i[[sz]]*
   // CK14-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i[[sz]]*
   // CK14-DAG: store i[[sz]] [[VAL:%.+]], i[[sz]]* [[CBP1]]
@@ -773,17 +799,17 @@ void implicit_maps_class (int a){
 #ifdef CK15
 
 // CK15: [[ST:%.+]] = type { i32, double, i32* }
-// CK15: [[SIZES:@.+]] = {{.+}}constant [2 x i[[sz:64|32]]] [i{{64|32}} {{24|16}}, i{{64|32}} 4]
+// CK15: [[SIZES:@.+]] = {{.+}}constant [5 x i[[sz:64|32]]] [i{{64|32}} 4, i{{64|32}} 8, i{{64|32}} {{8|4}}, i{{64|32}} 4, i{{64|32}} 4]
 // Map types:
-// - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST = 35
+// - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST + OMP_MAP_IMPLICIT = 547
 // - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288
-// CK15: [[TYPES:@.+]] = {{.+}}constant [2 x i32] [i32 35, i32 288]
+// CK15: [[TYPES:@.+]] = {{.+}}constant [5 x i64] [i64 547, i64 515, i64 512, i64 531, i64 288]
 
-// CK15: [[SIZES2:@.+]] = {{.+}}constant [2 x i[[sz]]] [i{{64|32}} {{24|16}}, i{{64|32}} 4]
+// CK15: [[SIZES2:@.+]] = {{.+}}constant [5 x i[[sz]]] [i{{64|32}} 4, i{{64|32}} 8, i{{64|32}} {{8|4}}, i{{64|32}} 4, i{{64|32}} 4]
 // Map types:
 // - OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST = 35
 // - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288
-// CK15: [[TYPES2:@.+]] = {{.+}}constant [2 x i32] [i32 35, i32 288]
+// CK15: [[TYPES2:@.+]] = {{.+}}constant [5 x i64] [i64 547, i64 515, i64 512, i64 531, i64 288]
 
 template<int x>
 class SSST {
@@ -818,23 +844,44 @@ void implicit_maps_templated_class (int a){
   SSST<123> ssst(a, (double)a, a);
 
   // CK15: define {{.*}}void @{{.+}}foo{{.+}}([[ST]]* {{[^,]+}}, i32 {{[^,]+}})
-  // CK15-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 2, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK15-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 5, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK15-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK15-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
 
   // CK15-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
   // CK15-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 0
   // CK15-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[ST]]**
-  // CK15-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[ST]]**
+  // CK15-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
   // CK15-DAG: store [[ST]]* [[DECL:%.+]], [[ST]]** [[CBP0]]
-  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CP0]]
+  // CK15-DAG: store i32* %{{.+}}, i32** [[CP0]]
 
   // CK15-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 1
   // CK15-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 1
-  // CK15-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to i[[sz]]*
-  // CK15-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i[[sz]]*
-  // CK15-DAG: store i[[sz]] [[VAL:%.+]], i[[sz]]* [[CBP1]]
-  // CK15-DAG: store i[[sz]] [[VAL]], i[[sz]]* [[CP1]]
+  // CK15-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[ST]]**
+  // CK15-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to double**
+  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CBP1]]
+  // CK15-DAG: store double* %{{.+}}, double** [[CP1]]
+
+  // CK15-DAG: [[BP2:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 2
+  // CK15-DAG: [[P2:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 2
+  // CK15-DAG: [[CBP2:%.+]] = bitcast i8** [[BP2]] to [[ST]]**
+  // CK15-DAG: [[CP2:%.+]] = bitcast i8** [[P2]] to i32***
+  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CBP2]]
+  // CK15-DAG: store i32** %{{.+}}, i32*** [[CP2]]
+
+  // CK15-DAG: [[BP3:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 3
+  // CK15-DAG: [[P3:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 3
+  // CK15-DAG: [[CBP3:%.+]] = bitcast i8** [[BP3]] to i32***
+  // CK15-DAG: [[CP3:%.+]] = bitcast i8** [[P3]] to i32**
+  // CK15-DAG: store i32** %{{.+}}, i32*** [[CBP3]]
+  // CK15-DAG: store i32* %{{.+}}, i32** [[CP3]]
+
+  // CK15-DAG: [[BP4:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 4
+  // CK15-DAG: [[P4:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 4
+  // CK15-DAG: [[CBP4:%.+]] = bitcast i8** [[BP4]] to i[[sz]]*
+  // CK15-DAG: [[CP4:%.+]] = bitcast i8** [[P4]] to i[[sz]]*
+  // CK15-DAG: store i[[sz]] [[VAL:%.+]], i[[sz]]* [[CBP4]]
+  // CK15-DAG: store i[[sz]] [[VAL]], i[[sz]]* [[CP4]]
   // CK15-DAG: [[VAL]] = load i[[sz]], i[[sz]]* [[ADDR:%.+]],
   // CK15-64-DAG: [[CADDR:%.+]] = bitcast i[[sz]]* [[ADDR]] to i32*
   // CK15-64-DAG: store i32 {{.+}}, i32* [[CADDR]],
@@ -843,23 +890,44 @@ void implicit_maps_templated_class (int a){
   ssst.foo(456);
 
   // CK15: define {{.*}}void @{{.+}}bar{{.+}}([[ST]]* {{[^,]+}}, i32 {{[^,]+}})
-  // CK15-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 2, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES2]]{{.+}}, {{.+}}[[TYPES2]]{{.+}})
+  // CK15-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 5, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES2]]{{.+}}, {{.+}}[[TYPES2]]{{.+}})
   // CK15-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK15-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
 
   // CK15-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
   // CK15-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 0
   // CK15-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[ST]]**
-  // CK15-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[ST]]**
-  // CK15-DAG: store [[ST]]* [[DECL:%.+]], [[ST]]** [[CBP0]]
-  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CP0]]
+  // CK15-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
+  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CBP0]]
+  // CK15-DAG: store i32* %{{.+}}, i32** [[CP0]]
 
   // CK15-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 1
   // CK15-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 1
-  // CK15-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to i[[sz]]*
-  // CK15-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i[[sz]]*
-  // CK15-DAG: store i[[sz]] [[VAL:%.+]], i[[sz]]* [[CBP1]]
-  // CK15-DAG: store i[[sz]] [[VAL]], i[[sz]]* [[CP1]]
+  // CK15-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[ST]]**
+  // CK15-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to double**
+  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CBP1]]
+  // CK15-DAG: store double* %{{.+}}, double** [[CP1]]
+
+  // CK15-DAG: [[BP2:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 2
+  // CK15-DAG: [[P2:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 2
+  // CK15-DAG: [[CBP2:%.+]] = bitcast i8** [[BP2]] to [[ST]]**
+  // CK15-DAG: [[CP2:%.+]] = bitcast i8** [[P2]] to i32***
+  // CK15-DAG: store [[ST]]* [[DECL]], [[ST]]** [[CBP2]]
+  // CK15-DAG: store i32** %{{.+}}, i32*** [[CP2]]
+
+  // CK15-DAG: [[BP3:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 3
+  // CK15-DAG: [[P3:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 3
+  // CK15-DAG: [[CBP3:%.+]] = bitcast i8** [[BP3]] to i32***
+  // CK15-DAG: [[CP3:%.+]] = bitcast i8** [[P3]] to i32**
+  // CK15-DAG: store i32** %{{.+}}, i32*** [[CBP3]]
+  // CK15-DAG: store i32* %{{.+}}, i32** [[CP3]]
+
+  // CK15-DAG: [[BP4:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 4
+  // CK15-DAG: [[P4:%.+]] = getelementptr inbounds {{.+}}[[PS]], i32 0, i32 4
+  // CK15-DAG: [[CBP4:%.+]] = bitcast i8** [[BP4]] to i[[sz]]*
+  // CK15-DAG: [[CP4:%.+]] = bitcast i8** [[P4]] to i[[sz]]*
+  // CK15-DAG: store i[[sz]] [[VAL:%.+]], i[[sz]]* [[CBP4]]
+  // CK15-DAG: store i[[sz]] [[VAL]], i[[sz]]* [[CP4]]
   // CK15-DAG: [[VAL]] = load i[[sz]], i[[sz]]* [[ADDR:%.+]],
   // CK15-64-DAG: [[CADDR:%.+]] = bitcast i[[sz]]* [[ADDR]] to i32*
   // CK15-64-DAG: store i32 {{.+}}, i32* [[CADDR]],
@@ -903,7 +971,7 @@ void implicit_maps_templated_class (int a){
 // CK16-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types:
 // - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288
-// CK16-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK16-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 template<int y>
 int foo(int d) {
@@ -919,7 +987,7 @@ void implicit_maps_templated_function (int a){
   int i = a;
 
   // CK16: define {{.*}}i32 @{{.+}}foo{{.+}}(i32 {{[^,]+}})
-  // CK16-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK16-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK16-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK16-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
 
@@ -955,8 +1023,8 @@ void implicit_maps_templated_function (int a){
 
 // CK17-DAG: [[ST:%.+]] = type { i32, double }
 // CK17-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} {{16|12}}]
-// Map types: OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST = 35
-// CK17-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 35]
+// Map types: OMP_MAP_TO + OMP_MAP_FROM + OMP_MAP_IS_FIRST + OMP_MAP_IMPLICIT = 547
+// CK17-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 547]
 
 class SSS {
 public:
@@ -968,7 +1036,7 @@ public:
 void implicit_maps_struct (int a){
   SSS s = {a, (double)a};
 
-  // CK17-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK17-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK17-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK17-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK17-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -1004,7 +1072,7 @@ void implicit_maps_struct (int a){
 // CK18-DAG: [[SIZES:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 4]
 // Map types:
 // - OMP_MAP_PRIVATE_VAL + OMP_MAP_IS_FIRST = 288
-// CK18-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i32] [i32 288]
+// CK18-DAG: [[TYPES:@.+]] = {{.+}}constant [1 x i64] [i64 288]
 
 template<typename T>
 int foo(T d) {
@@ -1019,7 +1087,7 @@ void implicit_maps_template_type_capture (int a){
   int i = a;
 
   // CK18: define {{.*}}i32 @{{.+}}foo{{.+}}(i32 {{[^,]+}})
-  // CK18-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
+  // CK18-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZES]]{{.+}}, {{.+}}[[TYPES]]{{.+}})
   // CK18-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK18-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
 
@@ -1054,125 +1122,125 @@ void implicit_maps_template_type_capture (int a){
 #ifdef CK19
 
 // CK19: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK19: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK19: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK19: [[SIZE00n:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK19: [[MTYPE00n:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK19: [[MTYPE00n:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK19: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK19: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK19: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
 // CK19: [[SIZE02:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 240]
-// CK19: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK19: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
 // CK19: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 240]
-// CK19: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE04:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK19: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK19: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK19: [[SIZE05:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK19: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK19: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
-// CK19: [[MTYPE06:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE06:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
-// CK19: [[MTYPE07:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK19: [[MTYPE07:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK19: [[SIZE08:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK19: [[MTYPE08:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE08:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE09:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{8|4}}]
-// CK19: [[MTYPE09:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK19: [[MTYPE09:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
 // CK19: [[SIZE10:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 240]
-// CK19: [[MTYPE10:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE10:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE11:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 240]
-// CK19: [[MTYPE11:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK19: [[MTYPE11:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK19: [[SIZE12:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK19: [[MTYPE12:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK19: [[MTYPE12:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
-// CK19: [[MTYPE13:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK19: [[MTYPE13:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
-// CK19: [[MTYPE14:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK19: [[MTYPE14:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
 // CK19: [[SIZE15:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK19: [[MTYPE15:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK19: [[MTYPE15:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
-// CK19: [[MTYPE16:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 33]
+// CK19: [[MTYPE16:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 33]
 
 // CK19: [[SIZE17:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 240]
-// CK19: [[MTYPE17:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 34]
+// CK19: [[MTYPE17:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 34]
 
 // CK19: [[SIZE18:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 240]
-// CK19: [[MTYPE18:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 35]
+// CK19: [[MTYPE18:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 35]
 
-// CK19: [[MTYPE19:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 32]
+// CK19: [[MTYPE19:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 32]
 
 // CK19: [[SIZE20:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK19: [[MTYPE20:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 33]
+// CK19: [[MTYPE20:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 33]
 
-// CK19: [[MTYPE21:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 35]
+// CK19: [[MTYPE21:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 35]
 
 // CK19: [[SIZE22:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK19: [[MTYPE22:@.+]] = private {{.*}}constant [2 x i32] [i32 288, i32 35]
+// CK19: [[MTYPE22:@.+]] = private {{.*}}constant [2 x i64] [i64 288, i64 35]
 
 // CK19: [[SIZE23:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK19: [[MTYPE23:@.+]] = private {{.*}}constant [1 x i32] [i32 39]
+// CK19: [[MTYPE23:@.+]] = private {{.*}}constant [1 x i64] [i64 39]
 
 // CK19: [[SIZE24:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 480]
-// CK19: [[MTYPE24:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE24:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE25:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 16]
-// CK19: [[MTYPE25:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE25:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE26:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 24]
-// CK19: [[MTYPE26:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE26:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE27:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK19: [[MTYPE27:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE27:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE28:@.+]] = private {{.*}}constant [3 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 16]
-// CK19: [[MTYPE28:@.+]] = private {{.*}}constant [3 x i32] [i32 35, i32 19, i32 19]
+// CK19: [[MTYPE28:@.+]] = private {{.*}}constant [3 x i64] [i64 35, i64 19, i64 19]
 
 // CK19: [[SIZE29:@.+]] = private {{.*}}constant [3 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK19: [[MTYPE29:@.+]] = private {{.*}}constant [3 x i32] [i32 35, i32 19, i32 19]
+// CK19: [[MTYPE29:@.+]] = private {{.*}}constant [3 x i64] [i64 35, i64 19, i64 19]
 
-// CK19: [[MTYPE30:@.+]] = private {{.*}}constant [4 x i32] [i32 288, i32 288, i32 288, i32 35]
+// CK19: [[MTYPE30:@.+]] = private {{.*}}constant [4 x i64] [i64 288, i64 288, i64 288, i64 35]
 
 // CK19: [[SIZE31:@.+]] = private {{.*}}constant [4 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 40]
-// CK19: [[MTYPE31:@.+]] = private {{.*}}constant [4 x i32] [i32 288, i32 288, i32 288, i32 35]
+// CK19: [[MTYPE31:@.+]] = private {{.*}}constant [4 x i64] [i64 288, i64 288, i64 288, i64 35]
 
 // CK19: [[SIZE32:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 13728]
-// CK19: [[MTYPE32:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE32:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE33:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 13728]
-// CK19: [[MTYPE33:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE33:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE34:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 13728]
-// CK19: [[MTYPE34:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE34:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
-// CK19: [[MTYPE35:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE35:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19: [[SIZE36:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 208]
-// CK19: [[MTYPE36:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE36:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
-// CK19: [[MTYPE37:@.+]] = private {{.*}}constant [3 x i32] [i32 288, i32 288, i32 35]
+// CK19: [[MTYPE37:@.+]] = private {{.*}}constant [3 x i64] [i64 288, i64 288, i64 35]
 
-// CK19: [[MTYPE38:@.+]] = private {{.*}}constant [3 x i32] [i32 288, i32 288, i32 35]
+// CK19: [[MTYPE38:@.+]] = private {{.*}}constant [3 x i64] [i64 288, i64 288, i64 35]
 
-// CK19: [[MTYPE39:@.+]] = private {{.*}}constant [3 x i32] [i32 288, i32 288, i32 35]
+// CK19: [[MTYPE39:@.+]] = private {{.*}}constant [3 x i64] [i64 288, i64 288, i64 35]
 
-// CK19: [[MTYPE40:@.+]] = private {{.*}}constant [3 x i32] [i32 288, i32 288, i32 35]
+// CK19: [[MTYPE40:@.+]] = private {{.*}}constant [3 x i64] [i64 288, i64 288, i64 35]
 
 // CK19: [[SIZE41:@.+]] = private {{.*}}constant [3 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 208]
-// CK19: [[MTYPE41:@.+]] = private {{.*}}constant [3 x i32] [i32 288, i32 288, i32 35]
+// CK19: [[MTYPE41:@.+]] = private {{.*}}constant [3 x i64] [i64 288, i64 288, i64 35]
 
 // CK19: [[SIZE42:@.+]] = private {{.*}}constant [3 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 104]
-// CK19: [[MTYPE42:@.+]] = private {{.*}}constant [3 x i32] [i32 35, i32 19, i32 19]
+// CK19: [[MTYPE42:@.+]] = private {{.*}}constant [3 x i64] [i64 35, i64 19, i64 19]
 
-// CK19: [[MTYPE43:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK19: [[MTYPE43:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK19-LABEL: explicit_maps_single
 void explicit_maps_single (int ii){
@@ -1180,7 +1248,7 @@ void explicit_maps_single (int ii){
   int a = ii;
 
   // Region 00
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1201,7 +1269,7 @@ void explicit_maps_single (int ii){
   int b = a;
 
   // Region 00n
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00n]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00n]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00n]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00n]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1223,7 +1291,7 @@ void explicit_maps_single (int ii){
   int arra[100];
 
   // Region 01
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1241,7 +1309,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 02
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1260,7 +1328,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 03
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1279,7 +1347,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 04
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1298,7 +1366,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 05
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1317,7 +1385,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 06
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE06]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE06]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1340,7 +1408,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 07
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE07]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE07]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1363,7 +1431,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 08
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE08]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE08]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE08]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE08]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1385,7 +1453,7 @@ void explicit_maps_single (int ii){
   int *pa;
 
   // Region 09
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE09]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE09]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1403,7 +1471,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 10
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE10]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE10]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE10]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE10]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1424,7 +1492,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 11
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE11]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE11]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE11]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE11]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1445,7 +1513,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 12
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE12]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE12]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE12]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE12]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1466,7 +1534,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 13
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE13]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE13]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1491,7 +1559,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 14
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE14]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE14]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1516,7 +1584,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 15
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE15]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE15]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE15]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE15]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1540,7 +1608,7 @@ void explicit_maps_single (int ii){
   int va[ii];
 
   // Region 16
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE16]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE16]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1571,7 +1639,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 17
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE17]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE17]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE17]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE17]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1597,7 +1665,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 18
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE18]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE18]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE18]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE18]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1623,7 +1691,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 19
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE19]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE19]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1655,7 +1723,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 20
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE20]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE20]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE20]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE20]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1681,7 +1749,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 21
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE21]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE21]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1713,7 +1781,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 22
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE22]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE22]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE22]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE22]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1740,7 +1808,7 @@ void explicit_maps_single (int ii){
 
   // Always.
   // Region 23
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE23]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE23]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE23]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE23]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1762,7 +1830,7 @@ void explicit_maps_single (int ii){
   int ***mptr;
 
   // Region 24
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE24]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE24]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE24]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE24]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1780,7 +1848,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 25
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE25]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE25]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE25]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE25]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1801,7 +1869,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 26
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE26]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE26]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE26]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE26]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1822,7 +1890,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 27
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE27]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE27]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE27]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE27]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1843,7 +1911,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 28
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE28]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE28]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE28]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE28]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1888,7 +1956,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 29
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE29]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE29]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE29]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE29]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -1936,7 +2004,7 @@ void explicit_maps_single (int ii){
   double mva[23][ii][ii+5];
 
   // Region 30
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE30]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE30]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -1989,7 +2057,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 31
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE31]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE31]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE31]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE31]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   //
@@ -2038,7 +2106,7 @@ void explicit_maps_single (int ii){
   double ***mptras;
 
   // Region 32
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE32]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE32]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE32]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE32]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2056,7 +2124,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 33
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE33]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE33]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE33]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE33]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2075,7 +2143,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 34
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE34]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE34]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE34]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE34]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2094,7 +2162,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 35
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE35]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE35]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -2119,7 +2187,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 36
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE36]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE36]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE36]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE36]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2140,7 +2208,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 37
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE37]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE37]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -2180,7 +2248,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 38
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE38]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE38]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -2222,7 +2290,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 39
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE39]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE39]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -2264,7 +2332,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 40
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE40]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE40]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -2306,7 +2374,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 41
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE41]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE41]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE41]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE41]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   //
@@ -2341,7 +2409,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 42
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE42]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE42]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 3, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[SIZE42]], {{.+}}getelementptr {{.+}}[3 x i{{.+}}]* [[MTYPE42]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2386,7 +2454,7 @@ void explicit_maps_single (int ii){
   }
 
   // Region 43 - the memory is not contiguous for this map - will map the whole last dimension.
-  // CK19-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE43]]{{.+}})
+  // CK19-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[Z]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE43]]{{.+}})
   // CK19-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK19-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK19-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -2469,16 +2537,16 @@ void explicit_maps_single (int ii){
 #ifdef CK20
 
 // CK20: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK20: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK20: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
 // CK20: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 20]
-// CK20: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK20: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
 // CK20: [[SIZE02:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK20: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK20: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
 // CK20: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 12]
-// CK20: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK20: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
 // CK20-LABEL: explicit_maps_references_and_function_args
 void explicit_maps_references_and_function_args (int a, float b, int (&c)[10], float *d){
@@ -2489,7 +2557,7 @@ void explicit_maps_references_and_function_args (int a, float b, int (&c)[10], f
   float *&dd = d;
 
   // Region 00
-  // CK20-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK20-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
   // CK20-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK20-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2509,7 +2577,7 @@ void explicit_maps_references_and_function_args (int a, float b, int (&c)[10], f
   }
 
   // Region 01
-  // CK20-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+  // CK20-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
   // CK20-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK20-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2530,7 +2598,7 @@ void explicit_maps_references_and_function_args (int a, float b, int (&c)[10], f
   }
 
   // Region 02
-  // CK20-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK20-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK20-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK20-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2548,7 +2616,7 @@ void explicit_maps_references_and_function_args (int a, float b, int (&c)[10], f
   }
 
   // Region 03
-  // CK20-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK20-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK20-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK20-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2586,22 +2654,22 @@ void explicit_maps_references_and_function_args (int a, float b, int (&c)[10], f
 // CK21: [[ST:%.+]] = type { i32, i32, float* }
 
 // CK21: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK21: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK21: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK21: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 492]
-// CK21: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK21: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK21: [[SIZE02:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 500]
-// CK21: [[MTYPE02:@.+]] = private {{.*}}constant [2 x i32] [i32 34, i32 18]
+// CK21: [[MTYPE02:@.+]] = private {{.*}}constant [2 x i64] [i64 34, i64 18]
 
 // CK21: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 492]
-// CK21: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK21: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
 // CK21: [[SIZE04:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK21: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i32] [i32 34]
+// CK21: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i64] [i64 34]
 
 // CK21: [[SIZE05:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] 4, i[[Z]] 4]
-// CK21: [[MTYPE05:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 3]
+// CK21: [[MTYPE05:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 3]
 
 // CK21-LABEL: explicit_maps_template_args_and_members
 
@@ -2616,7 +2684,7 @@ struct CC {
     T *lb;
 
     // Region 00
-    // CK21-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+    // CK21-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
     // CK21-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK21-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2635,7 +2703,7 @@ struct CC {
     }
     
     // Region 01
-    // CK21-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+    // CK21-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
     // CK21-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK21-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2656,7 +2724,7 @@ struct CC {
     }
     
     // Region 02
-    // CK21-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE02]]{{.+}})
+    // CK21-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE02]]{{.+}})
     // CK21-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK21-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2685,7 +2753,7 @@ struct CC {
     }
     
     // Region 03
-    // CK21-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+    // CK21-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
     // CK21-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK21-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2703,7 +2771,7 @@ struct CC {
     }
     
     // Region 04
-    // CK21-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
+    // CK21-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
     // CK21-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK21-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2722,7 +2790,7 @@ struct CC {
     
     // Make sure the extra flag is passed to the second map.
     // Region 05
-    // CK21-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE05]]{{.+}})
+    // CK21-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE05]]{{.+}})
     // CK21-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK21-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2777,49 +2845,49 @@ int explicit_maps_template_args_and_members(int a){
 // CK22-DAG: [[STT:%.+]] = type { i32 }
 
 // CK22: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK22: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK22: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE02:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{8|4}}]
-// CK22: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 16]
-// CK22: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE04:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 20]
-// CK22: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE05:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK22: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE06:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK22: [[MTYPE06:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE06:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE07:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{8|4}}]
-// CK22: [[MTYPE07:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE07:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE08:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 16]
-// CK22: [[MTYPE08:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE08:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE09:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 20]
-// CK22: [[MTYPE09:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE09:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE10:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK22: [[MTYPE10:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE10:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE11:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK22: [[MTYPE11:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE11:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE12:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{8|4}}]
-// CK22: [[MTYPE12:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE12:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE13:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 16]
-// CK22: [[MTYPE13:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE13:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK22: [[SIZE14:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 20]
-// CK22: [[MTYPE14:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK22: [[MTYPE14:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 int a;
 int c[100];
@@ -2845,7 +2913,7 @@ STT<int> *std;
 // CK22-LABEL: explicit_maps_globals
 int explicit_maps_globals(void){
   // Region 00
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2861,7 +2929,7 @@ int explicit_maps_globals(void){
   { a+=1; }
   
   // Region 01
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2877,7 +2945,7 @@ int explicit_maps_globals(void){
   { c[3]+=1; }
   
   // Region 02
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2893,7 +2961,7 @@ int explicit_maps_globals(void){
   { d[3]+=1; }
     
   // Region 03
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2909,7 +2977,7 @@ int explicit_maps_globals(void){
   { c[3]+=1; }
   
   // Region 04
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2928,7 +2996,7 @@ int explicit_maps_globals(void){
   { d[3]+=1; }
   
   // Region 05
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2944,7 +3012,7 @@ int explicit_maps_globals(void){
   { sa.fa+=1; }
   
   // Region 06
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE06]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE06]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE06]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE06]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2960,7 +3028,7 @@ int explicit_maps_globals(void){
   { sc[3].fa+=1; }
   
   // Region 07
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE07]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE07]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE07]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE07]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2976,7 +3044,7 @@ int explicit_maps_globals(void){
   { sd[3].fa+=1; }
   
   // Region 08
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE08]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE08]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE08]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE08]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -2992,7 +3060,7 @@ int explicit_maps_globals(void){
   { sc[3].fa+=1; }
   
   // Region 09
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE09]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE09]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3011,7 +3079,7 @@ int explicit_maps_globals(void){
   { sd[3].fa+=1; }
   
   // Region 10
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE10]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE10]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE10]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE10]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3027,7 +3095,7 @@ int explicit_maps_globals(void){
   { sta.fa+=1; }
   
   // Region 11
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE11]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE11]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE11]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE11]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3043,7 +3111,7 @@ int explicit_maps_globals(void){
   { stc[3].fa+=1; }
   
   // Region 12
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE12]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE12]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE12]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE12]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3059,7 +3127,7 @@ int explicit_maps_globals(void){
   { std[3].fa+=1; }
   
   // Region 13
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE13]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE13]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE13]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE13]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3075,7 +3143,7 @@ int explicit_maps_globals(void){
   { stc[3].fa+=1; }
   
   // Region 14
-  // CK22-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE14]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE14]]{{.+}})
+  // CK22-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE14]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE14]]{{.+}})
   // CK22-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK22-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3121,22 +3189,22 @@ int explicit_maps_globals(void){
 #ifdef CK23
 
 // CK23: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK23: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK23: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK23: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK23: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK23: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK23: [[SIZE02:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK23: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK23: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK23: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{8|4}}]
-// CK23: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK23: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK23: [[SIZE04:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 16]
-// CK23: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK23: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK23: [[SIZE05:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 16]
-// CK23: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK23: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK23-LABEL: explicit_maps_inside_captured
 int explicit_maps_inside_captured(int a){
@@ -3148,7 +3216,7 @@ int explicit_maps_inside_captured(int a){
   // CK23: define {{.*}}explicit_maps_inside_captured{{.*}}
   [&](void){
     // Region 00
-    // CK23-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+    // CK23-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
     // CK23-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK23-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3167,7 +3235,7 @@ int explicit_maps_inside_captured(int a){
     #pragma omp target map(a)
       { a+=1; }
     // Region 01
-    // CK23-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+    // CK23-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
     // CK23-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK23-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3186,7 +3254,7 @@ int explicit_maps_inside_captured(int a){
     #pragma omp target map(b)
       { b+=1; }
     // Region 02
-    // CK23-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+    // CK23-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
     // CK23-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK23-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3206,7 +3274,7 @@ int explicit_maps_inside_captured(int a){
       { c[3]+=1; }
       
     // Region 03
-    // CK23-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+    // CK23-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
     // CK23-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK23-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3225,7 +3293,7 @@ int explicit_maps_inside_captured(int a){
     #pragma omp target map(d)
       { d[3]+=1; }
     // Region 04
-    // CK23-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
+    // CK23-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
     // CK23-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK23-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3246,7 +3314,7 @@ int explicit_maps_inside_captured(int a){
       { c[3]+=1; }
       
     // Region 05
-    // CK23-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
+    // CK23-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
     // CK23-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK23-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3311,76 +3379,43 @@ struct SC{
 };
 
 // CK24: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK24: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
-
-// CK24: [[SIZE02:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{56|48}}]
-// CK24: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
-
-// CK24: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK24: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
-
-// CK24: [[SIZE04:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 20]
-// CK24: [[MTYPE04:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
-
-// CK24: [[SIZE05:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{3560|2880}}]
-// CK24: [[MTYPE05:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
-
-// CK24: [[SIZE06:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK24: [[MTYPE06:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
-
-// CK24: [[SIZE07:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE07:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
-
-// CK24: [[SIZE08:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE08:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
-
-// CK24: [[SIZE09:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE09:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
-
-// CK24: [[SIZE10:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 8]
-// CK24: [[MTYPE10:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
-
-// CK24: [[SIZE11:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}]
-// CK24: [[MTYPE11:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
-
-// CK24: [[SIZE12:@.+]] = private {{.*}}constant [4 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE12:@.+]] = private {{.*}}constant [4 x i32] [i32 35, i32 19, i32 19, i32 19]
+// CK24: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE13:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK24: [[MTYPE13:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK24: [[MTYPE13:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE14:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{56|48}}]
-// CK24: [[MTYPE14:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK24: [[MTYPE14:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE15:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK24: [[MTYPE15:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK24: [[MTYPE15:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE16:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 20]
-// CK24: [[MTYPE16:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK24: [[MTYPE16:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE17:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{3560|2880}}]
-// CK24: [[MTYPE17:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
+// CK24: [[MTYPE17:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 19]
 
 // CK24: [[SIZE18:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK24: [[MTYPE18:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK24: [[MTYPE18:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE19:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE19:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
+// CK24: [[MTYPE19:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 19]
 
 // CK24: [[SIZE20:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE20:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
+// CK24: [[MTYPE20:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 19]
 
 // CK24: [[SIZE21:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE21:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
+// CK24: [[MTYPE21:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 19]
 
 // CK24: [[SIZE22:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] {{8|4}}]
-// CK24: [[MTYPE22:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK24: [[MTYPE22:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK24: [[SIZE23:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}]
-// CK24: [[MTYPE23:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 19]
+// CK24: [[MTYPE23:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 19]
 
 // CK24: [[SIZE24:@.+]] = private {{.*}}constant [4 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 4]
-// CK24: [[MTYPE24:@.+]] = private {{.*}}constant [4 x i32] [i32 35, i32 19, i32 19, i32 19]
+// CK24: [[MTYPE24:@.+]] = private {{.*}}constant [4 x i64] [i64 35, i64 19, i64 19, i64 19]
 
 // CK24-LABEL: explicit_maps_struct_fields
 int explicit_maps_struct_fields(int a){
@@ -3388,7 +3423,7 @@ int explicit_maps_struct_fields(int a){
   SC *p;
 
 // Region 01
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3404,303 +3439,11 @@ int explicit_maps_struct_fields(int a){
 #pragma omp target map(s.a)
   { s.a++; }
   
-// Region 02
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SA]]**
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SA]]* [[SEC0:%.+]], [[SA]]** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SB]]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL02:@.+]]([[SC]]* {{[^,]+}})
-#pragma omp target map(s.s.s)
-  { s.a++; }
-  
-// Region 03
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store i32* [[SEC0:%.+]], i32** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SA]]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SB]]* [[SEC000:%[^,]+]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[SEC000]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL03:@.+]]([[SC]]* {{[^,]+}})
-#pragma omp target map(s.s.s.a)
-  { s.a++; }
-
-// Region 04
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE04]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store i32* [[SEC0:%.+]], i32** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[10 x i32]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 3
-
-// CK24: call void [[CALL04:@.+]]([[SC]]* {{[^,]+}})
-#pragma omp target map(s.b[:5])
-  { s.a++; }
-  
-// Region 05
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE05]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SB]]***
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SB]]** [[SEC0:%.+]], [[SB]]*** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[SB]]***
-// CK24-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to [[SB]]**
-// CK24-DAG: store [[SB]]** [[SEC0]], [[SB]]*** [[CBP1]]
-// CK24-DAG: store [[SB]]* [[SEC1:%.+]], [[SB]]** [[CP1]]
-// CK24-DAG: [[SEC1]] = getelementptr {{.*}}[[SB]]* [[SEC11:%[^,]+]], i{{.+}} 0
-// CK24-DAG: [[SEC11]] = load [[SB]]*, [[SB]]** [[SEC111:%[^,]+]],
-// CK24-DAG: [[SEC111]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24: call void [[CALL05:@.+]]([[SC]]* {{[^,]+}})  
-#pragma omp target map(s.p[:5])
-  { s.a++; }
-
-// Region 06
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE06]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE06]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store i32* [[SEC0:%.+]], i32** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SA]]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[10 x [[SA]]]* [[SEC000:%[^,]+]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[SEC000]] = getelementptr {{.*}}[[SB]]* [[SEC0000:%[^,]+]], i{{.+}} 0, i{{.+}} 2
-// CK24-DAG: [[SEC0000]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL06:@.+]]([[SC]]* {{[^,]+}})
-#pragma omp target map(s.s.sa[3].a)
-  { s.a++; }
-  
-// Region 07
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE07]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE07]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SA]]***
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SA]]** [[SEC0:%.+]], [[SA]]*** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[10 x [[SA]]*]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SB]]* [[SEC000:%[^,]+]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[SEC000]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[SA]]***
-// CK24-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i32**
-// CK24-DAG: store [[SA]]** [[SEC0]], [[SA]]*** [[CBP1]]
-// CK24-DAG: store i32* [[SEC1:%.+]], i32** [[CP1]]
-// CK24-DAG: [[SEC1]] = getelementptr {{.*}}[[SA]]* [[SEC11:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC11]] = load [[SA]]*, [[SA]]** [[SEC111:%[^,]+]],
-// CK24-DAG: [[SEC111]] = getelementptr {{.*}}[10 x [[SA]]*]* [[SEC1111:%[^,]+]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[SEC1111]] = getelementptr {{.*}}[[SB]]* [[SEC11111:%[^,]+]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[SEC11111]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL07:@.+]]([[SC]]* {{[^,]+}}) 
-#pragma omp target map(s.s.sp[3]->a)
-  { s.a++; }
-
-// Region 08
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE08]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE08]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SB]]***
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SB]]** [[SEC0:%.+]], [[SB]]*** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[SB]]***
-// CK24-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i32**
-// CK24-DAG: store [[SB]]** [[SEC0]], [[SB]]*** [[CBP1]]
-// CK24-DAG: store i32* [[SEC1:%.+]], i32** [[CP1]]
-// CK24-DAG: [[SEC1]] = getelementptr {{.*}}[[SB]]* [[SEC11:%[^,]+]], i{{.+}} 0
-// CK24-DAG: [[SEC11]] = load [[SB]]*, [[SB]]** [[SEC111:%[^,]+]],
-// CK24-DAG: [[SEC111]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24: call void [[CALL08:@.+]]([[SC]]* {{[^,]+}}) 
-#pragma omp target map(s.p->a)
-  { s.a++; }
-  
-// Region 09
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE09]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SA]]***
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SA]]** [[SEC0:%.+]], [[SA]]*** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SB]]* [[SEC00:[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[SA]]***
-// CK24-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i32**
-// CK24-DAG: store [[SA]]** [[SEC0]], [[SA]]*** [[CBP1]]
-// CK24-DAG: store i32* [[SEC1:%.+]], i32** [[CP1]]
-// CK24-DAG: [[SEC1]] = getelementptr {{.*}}[[SA]]* [[SEC11:%[^,]+]], i{{.+}} 0
-// CK24-DAG: [[SEC11]] = load [[SA]]*, [[SA]]** [[SEC111:%[^,]+]],
-// CK24-DAG: [[SEC111]] = getelementptr {{.*}}[[SB]]* [[SEC1111:[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC1111]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL09:@.+]]([[SC]]* {{[^,]+}}) 
-#pragma omp target map(s.s.p->a)
-  { s.a++; }
-
-// Region 10
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE10]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE10]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to i32**
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store i32* [[SEC0:%.+]], i32** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[10 x i32]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SA]]* [[SEC000:%[^,]+]], i{{.+}} 0, i{{.+}} 2
-// CK24-DAG: [[SEC000]] = getelementptr {{.*}}[[SB]]* [[SEC0000:%[^,]+]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[SEC0000]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL10:@.+]]([[SC]]* {{[^,]+}})
-#pragma omp target map(s.s.s.b[:2])
-  { s.a++; }
-  
-// Region 11
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE11]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE11]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SA]]***
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SA]]** [[SEC0:%.+]], [[SA]]*** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SB]]* [[SEC00:%[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC00]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[SA]]***
-// CK24-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to i32**
-// CK24-DAG: store [[SA]]** [[SEC0]], [[SA]]*** [[CBP1]]
-// CK24-DAG: store i32* [[SEC1:%.+]], i32** [[CP1]]
-// CK24-DAG: [[SEC1]] = getelementptr {{.*}}[10 x i32]* [[SEC11:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC11]] = getelementptr {{.*}}[[SA]]* [[SEC111:%[^,]+]], i{{.+}} 0, i{{.+}} 2
-// CK24-DAG: [[SEC111]] = load [[SA]]*, [[SA]]** [[SEC1111:%[^,]+]],
-// CK24-DAG: [[SEC1111]] = getelementptr {{.*}}[[SB]]* [[SEC11111:%[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC11111]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 1
-
-// CK24: call void [[CALL11:@.+]]([[SC]]* {{[^,]+}}) 
-#pragma omp target map(s.s.p->b[:2])
-  { s.a++; }
-
-// Region 12
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE12]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE12]]{{.+}})
-// CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
-// CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
-
-// CK24-DAG: [[BP0:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[P0:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[CBP0:%.+]] = bitcast i8** [[BP0]] to [[SC]]**
-// CK24-DAG: [[CP0:%.+]] = bitcast i8** [[P0]] to [[SB]]***
-// CK24-DAG: store [[SC]]* [[VAR0:%.+]], [[SC]]** [[CBP0]]
-// CK24-DAG: store [[SB]]** [[SEC0:%.+]], [[SB]]*** [[CP0]]
-// CK24-DAG: [[SEC0]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[P1:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[CBP1:%.+]] = bitcast i8** [[BP1]] to [[SB]]***
-// CK24-DAG: [[CP1:%.+]] = bitcast i8** [[P1]] to [[SA]]***
-// CK24-DAG: store [[SB]]** [[SEC0:%.+]], [[SB]]*** [[CBP1]]
-// CK24-DAG: store [[SA]]** [[SEC1:%.+]], [[SA]]*** [[CP1]]
-// CK24-DAG: [[SEC1]] = getelementptr {{.*}}[[SB]]* [[SEC11:%[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC11]] = load [[SB]]*, [[SB]]** [[SEC111:%[^,]+]],
-// CK24-DAG: [[SEC111]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24-DAG: [[BP2:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 2
-// CK24-DAG: [[P2:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 2
-// CK24-DAG: [[CBP2:%.+]] = bitcast i8** [[BP2]] to [[SA]]***
-// CK24-DAG: [[CP2:%.+]] = bitcast i8** [[P2]] to [[SA]]***
-// CK24-DAG: store [[SA]]** [[SEC1:%.+]], [[SA]]*** [[CBP2]]
-// CK24-DAG: store [[SA]]** [[SEC2:%.+]], [[SA]]*** [[CP2]]
-// CK24-DAG: [[SEC2]] = getelementptr {{.*}}[[SA]]* [[SEC22:%[^,]+]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[SEC22]] = load [[SA]]*, [[SA]]** [[SEC222:%[^,]+]],
-// CK24-DAG: [[SEC222]] = getelementptr {{.*}}[[SB]]* [[SEC2222:%[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC2222]] = load [[SB]]*, [[SB]]** [[SEC22222:%[^,]+]],
-// CK24-DAG: [[SEC22222]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24-DAG: [[BP3:%.+]] = getelementptr inbounds {{.+}}[[BP]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[P3:%.+]] = getelementptr inbounds {{.+}}[[P]], i{{.+}} 0, i{{.+}} 3
-// CK24-DAG: [[CBP3:%.+]] = bitcast i8** [[BP3]] to [[SA]]***
-// CK24-DAG: [[CP3:%.+]] = bitcast i8** [[P3]] to i32**
-// CK24-DAG: store [[SA]]** [[SEC2]], [[SA]]*** [[CBP3]]
-// CK24-DAG: store i32* [[SEC3:%.+]], i32** [[CP3]]
-// CK24-DAG: [[SEC3]] = getelementptr {{.*}}[[SA]]* [[SEC33:%[^,]+]], i{{.+}} 0, i{{.+}} 0
-// CK24-DAG: [[SEC33]] = load [[SA]]*, [[SA]]** [[SEC333:%[^,]+]],
-// CK24-DAG: [[SEC333]] = getelementptr {{.*}}[[SA]]* [[SEC3333:%[^,]+]], i{{.+}} 0, i{{.+}} 1
-// CK24-DAG: [[SEC3333]] = load [[SA]]*, [[SA]]** [[SEC33333:%[^,]+]],
-// CK24-DAG: [[SEC33333]] = getelementptr {{.*}}[[SB]]* [[SEC333333:%[^,]+]], i{{.+}} 0, i{{.+}} 4
-// CK24-DAG: [[SEC333333]] = load [[SB]]*, [[SB]]** [[SEC3333333:%[^,]+]],
-// CK24-DAG: [[SEC3333333]] = getelementptr {{.*}}[[SC]]* [[VAR0]], i{{.+}} 0, i{{.+}} 2
-
-// CK24: call void [[CALL12:@.+]]([[SC]]* {{[^,]+}})
-#pragma omp target map(s.p->p->p->a)
-  { s.a++; }
-
 //
 // Same thing but starting from a pointer.
 //
 // Region 13
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE13]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE13]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE13]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE13]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3720,7 +3463,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
   
 // Region 14
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE14]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE14]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE14]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE14]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3741,7 +3484,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
   
 // Region 15
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE15]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE15]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE15]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE15]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3763,7 +3506,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
 
 // Region 16
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE16]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE16]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE16]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE16]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3784,7 +3527,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
   
 // Region 17
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE17]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE17]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE17]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE17]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3815,7 +3558,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
 
 // Region 18
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE18]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE18]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE18]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE18]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3838,7 +3581,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
   
 // Region 19
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE19]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE19]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE19]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE19]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3873,7 +3616,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
 
 // Region 20
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE20]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE20]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE20]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE20]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3904,7 +3647,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
   
 // Region 21
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE21]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE21]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE21]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE21]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3937,7 +3680,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
 
 // Region 22
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE22]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE22]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE22]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE22]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3960,7 +3703,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
   
 // Region 23
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE23]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE23]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE23]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE23]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -3994,7 +3737,7 @@ int explicit_maps_struct_fields(int a){
   { p->a++; }
 
 // Region 24
-// CK24-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE24]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE24]]{{.+}})
+// CK24-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE24]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE24]]{{.+}})
 // CK24-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK24-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4056,17 +3799,6 @@ int explicit_maps_struct_fields(int a){
 }
 
 // CK24: define {{.+}}[[CALL01]]
-// CK24: define {{.+}}[[CALL02]]
-// CK24: define {{.+}}[[CALL03]]
-// CK24: define {{.+}}[[CALL04]]
-// CK24: define {{.+}}[[CALL05]]
-// CK24: define {{.+}}[[CALL06]]
-// CK24: define {{.+}}[[CALL07]]
-// CK24: define {{.+}}[[CALL08]]
-// CK24: define {{.+}}[[CALL09]]
-// CK24: define {{.+}}[[CALL10]]
-// CK24: define {{.+}}[[CALL11]]
-// CK24: define {{.+}}[[CALL12]]
 // CK24: define {{.+}}[[CALL13]]
 // CK24: define {{.+}}[[CALL14]]
 // CK24: define {{.+}}[[CALL15]]
@@ -4093,10 +3825,10 @@ int explicit_maps_struct_fields(int a){
 // CK25: [[CA01:%.+]] = type { i32* }
 
 // CK25: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] 4]
-// CK25: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK25: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
 // CK25: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK25: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 33]
+// CK25: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 33]
 
 // CK25-LABEL: explicit_maps_with_inner_lambda
 
@@ -4107,7 +3839,7 @@ struct CC {
 
   int foo(T arg) {
     // Region 00
-    // CK25-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+    // CK25-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
     // CK25-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK25-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4128,7 +3860,7 @@ struct CC {
     }
 
     // Region 01
-    // CK25-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+    // CK25-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
     // CK25-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK25-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4185,16 +3917,16 @@ int explicit_maps_with_inner_lambda(int a){
 // CK26: [[ST:%.+]] = type { i32, float*, i32, float* }
 
 // CK26: [[SIZE00:@.+]] = private {{.*}}constant [2 x i[[Z:64|32]]] [i[[Z:64|32]] {{32|16}}, i[[Z:64|32]] 4]
-// CK26: [[MTYPE00:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 35]
+// CK26: [[MTYPE00:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 35]
 
 // CK26: [[SIZE01:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{32|16}}, i[[Z]] 4]
-// CK26: [[MTYPE01:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 35]
+// CK26: [[MTYPE01:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 35]
 
 // CK26: [[SIZE02:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{32|16}}, i[[Z]] 4]
-// CK26: [[MTYPE02:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 35]
+// CK26: [[MTYPE02:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 35]
 
 // CK26: [[SIZE03:@.+]] = private {{.*}}constant [2 x i[[Z]]] [i[[Z]] {{32|16}}, i[[Z]] 4]
-// CK26: [[MTYPE03:@.+]] = private {{.*}}constant [2 x i32] [i32 35, i32 35]
+// CK26: [[MTYPE03:@.+]] = private {{.*}}constant [2 x i64] [i64 35, i64 35]
 
 // CK26-LABEL: explicit_maps_with_private_class_members
 
@@ -4211,7 +3943,7 @@ struct CC {
     #pragma omp parallel firstprivate(fA,fB) private(pA,pB)
     {
       // Region 00
-      // CK26-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
+      // CK26-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
       // CK26-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
       // CK26-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4238,7 +3970,7 @@ struct CC {
       }
 
       // Region 01
-      // CK26-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE01]]{{.+}})
+      // CK26-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE01]]{{.+}})
       // CK26-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
       // CK26-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4265,7 +3997,7 @@ struct CC {
       }
 
       // Region 02
-      // CK26-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE02]]{{.+}})
+      // CK26-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE02]]{{.+}})
       // CK26-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
       // CK26-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4292,7 +4024,7 @@ struct CC {
       }
 
       // Region 01
-      // CK26-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE03]]{{.+}})
+      // CK26-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE03]]{{.+}})
       // CK26-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
       // CK26-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4368,25 +4100,25 @@ int explicit_maps_with_private_class_members(){
 #ifdef CK27
 
 // CK27: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] zeroinitializer
-// CK27: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK27: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK27: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] zeroinitializer
-// CK27: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK27: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK27: [[SIZE02:@.+]] = private {{.*}}constant [1 x i[[Z]]] zeroinitializer
-// CK27: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK27: [[MTYPE02:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK27: [[SIZE03:@.+]] = private {{.*}}constant [1 x i[[Z]]] zeroinitializer
-// CK27: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK27: [[MTYPE03:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK27: [[SIZE05:@.+]] = private {{.*}}constant [1 x i[[Z]]] zeroinitializer
-// CK27: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i32] [i32 32]
+// CK27: [[MTYPE05:@.+]] = private {{.*}}constant [1 x i64] [i64 32]
 
 // CK27: [[SIZE07:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 4]
-// CK27: [[MTYPE07:@.+]] = private {{.*}}constant [1 x i32] [i32 288]
+// CK27: [[MTYPE07:@.+]] = private {{.*}}constant [1 x i64] [i64 288]
 
 // CK27: [[SIZE09:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 40]
-// CK27: [[MTYPE09:@.+]] = private {{.*}}constant [1 x i32] [i32 161]
+// CK27: [[MTYPE09:@.+]] = private {{.*}}constant [1 x i64] [i64 161]
 
 // CK27-LABEL: zero_size_section_and_private_maps
 void zero_size_section_and_private_maps (int ii){
@@ -4395,7 +4127,7 @@ void zero_size_section_and_private_maps (int ii){
   int *pa;
 
   // Region 00
-  // CK27-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
   // CK27-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK27-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4413,7 +4145,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 01
-  // CK27-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
   // CK27-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK27-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4434,7 +4166,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 02
-  // CK27-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK27-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK27-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4455,7 +4187,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 03
-  // CK27-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE03]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK27-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK27-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4480,7 +4212,7 @@ void zero_size_section_and_private_maps (int ii){
   int pvtArr[10];
 
   // Region 04
-  // CK27: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i32* null)
+  // CK27: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null)
   // CK27: call void [[CALL04:@.+]]()
   #pragma omp target private(pvtPtr)
   {
@@ -4488,7 +4220,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 05
-  // CK27-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE05]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE05]]{{.+}})
   // CK27-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK27-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4506,7 +4238,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 06
-  // CK27: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i32* null)
+  // CK27: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null)
   // CK27: call void [[CALL06:@.+]]()
   #pragma omp target private(pvtScl)
   {
@@ -4514,7 +4246,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 07
-  // CK27-DAG: call i32 @__tgt_target(i32 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZE07]]{{.+}}, {{.+}}[[MTYPE07]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{.+}}, i8* {{.+}}, i32 1, i8** [[BPGEP:%[0-9]+]], i8** [[PGEP:%[0-9]+]], {{.+}}[[SIZE07]]{{.+}}, {{.+}}[[MTYPE07]]{{.+}})
   // CK27-DAG: [[BPGEP]] = getelementptr inbounds {{.+}}[[BPS:%[^,]+]], i32 0, i32 0
   // CK27-DAG: [[PGEP]] = getelementptr inbounds {{.+}}[[PS:%[^,]+]], i32 0, i32 0
   // CK27-DAG: [[BP1:%.+]] = getelementptr inbounds {{.+}}[[BPS]], i32 0, i32 0
@@ -4534,7 +4266,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 08
-  // CK27: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i32* null)
+  // CK27: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null)
   // CK27: call void [[CALL08:@.+]]()
   #pragma omp target private(pvtArr)
   {
@@ -4542,7 +4274,7 @@ void zero_size_section_and_private_maps (int ii){
   }
 
   // Region 09
-  // CK27-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE09]]{{.+}})
+  // CK27-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE09]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE09]]{{.+}})
   // CK27-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK27-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4579,17 +4311,17 @@ void zero_size_section_and_private_maps (int ii){
 #ifdef CK28
 
 // CK28: [[SIZE00:@.+]] = private {{.*}}constant [1 x i[[Z:64|32]]] [i[[Z:64|32]] {{8|4}}]
-// CK28: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK28: [[MTYPE00:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK28: [[SIZE01:@.+]] = private {{.*}}constant [1 x i[[Z]]] [i[[Z]] 400]
-// CK28: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i32] [i32 35]
+// CK28: [[MTYPE01:@.+]] = private {{.*}}constant [1 x i64] [i64 35]
 
 // CK28-LABEL: explicit_maps_pointer_references
 void explicit_maps_pointer_references (int *p){
   int *&a = p;
 
   // Region 00
-  // CK28-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK28-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
   // CK28-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK28-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4609,7 +4341,7 @@ void explicit_maps_pointer_references (int *p){
   }
 
   // Region 01
-  // CK28-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
+  // CK28-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE01]]{{.+}})
   // CK28-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK28-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -4645,13 +4377,13 @@ void explicit_maps_pointer_references (int *p){
 // CK29: [[SSB:%.+]]  = type { [[SSA]]*, [[SSA]]** }
 
 // CK29: [[SIZE00:@.+]] = private {{.*}}constant [4 x i[[Z:64|32]]] [i[[Z:64|32]] {{8|4}}, i[[Z:64|32]] {{8|4}}, i[[Z:64|32]] {{8|4}}, i[[Z:64|32]] 80]
-// CK29: [[MTYPE00:@.+]] = private {{.*}}constant [4 x i32] [i32 35, i32 16, i32 19, i32 19]
+// CK29: [[MTYPE00:@.+]] = private {{.*}}constant [4 x i64] [i64 35, i64 16, i64 19, i64 19]
 
 // CK29: [[SIZE01:@.+]] = private {{.*}}constant [4 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 80]
-// CK29: [[MTYPE01:@.+]] = private {{.*}}constant [4 x i32] [i32 32, i32 19, i32 19, i32 19]
+// CK29: [[MTYPE01:@.+]] = private {{.*}}constant [4 x i64] [i64 32, i64 19, i64 19, i64 19]
 
 // CK29: [[SIZE02:@.+]] = private {{.*}}constant [5 x i[[Z]]] [i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] {{8|4}}, i[[Z]] 80]
-// CK29: [[MTYPE02:@.+]] = private {{.*}}constant [5 x i32] [i32 32, i32 19, i32 16, i32 19, i32 19]
+// CK29: [[MTYPE02:@.+]] = private {{.*}}constant [5 x i64] [i64 32, i64 19, i64 16, i64 19, i64 19]
 
 struct SSA{
   double *p;
@@ -4668,7 +4400,7 @@ struct SSB{
   void foo() {
 
     // Region 00
-    // CK29-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE00]]{{.+}})
+    // CK29-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE00]]{{.+}})
 
     // CK29-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK29-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
@@ -4715,7 +4447,7 @@ struct SSB{
     }
 
     // Region 01
-    // CK29-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE01]]{{.+}})
+    // CK29-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 4, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[SIZE01]], {{.+}}getelementptr {{.+}}[4 x i{{.+}}]* [[MTYPE01]]{{.+}})
 
     // CK29-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK29-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
@@ -4760,7 +4492,7 @@ struct SSB{
     }
 
     // Region 02
-    // CK29-DAG: call i32 @__tgt_target(i32 {{[^,]+}}, i8* {{[^,]+}}, i32 5, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[5 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[5 x i{{.+}}]* [[MTYPE02]]{{.+}})
+    // CK29-DAG: call i32 @__tgt_target(i64 {{[^,]+}}, i8* {{[^,]+}}, i32 5, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[5 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[5 x i{{.+}}]* [[MTYPE02]]{{.+}})
 
     // CK29-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
     // CK29-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
diff --git a/test/OpenMP/target_map_messages.cpp b/test/OpenMP/target_map_messages.cpp
index b9640b965d50..9ffbec472572 100644
--- a/test/OpenMP/target_map_messages.cpp
+++ b/test/OpenMP/target_map_messages.cpp
@@ -26,7 +26,14 @@ struct SA {
   T d;
   float e[I];
   T *f;
+  int bf : 20;
   void func(int arg) {
+    #pragma omp target
+    {
+      a = 0.0;
+      func(arg);
+      bf = 20;
+    }
     #pragma omp target map(arg,a,d)
     {}
     #pragma omp target map(arg[2:2],a,d) // expected-error {{subscripted value is not an array or pointer}}
@@ -267,8 +274,13 @@ void SAclient(int arg) {
   {}
   #pragma omp target map((p+1)->A)  // expected-error {{expected expression containing only member accesses and/or array sections based on named variables}}
   {}
-  #pragma omp target map(u.B)  // expected-error {{mapped storage cannot be derived from a union}}
+  #pragma omp target map(u.B)  // expected-error {{mapping of union members is not allowed}}
   {}
+  #pragma omp target
+  {
+    u.B = 0;
+    r.S.foo();
+  }
 
   #pragma omp target data map(to: r.C) //expected-note {{used here}}
   {
@@ -319,8 +331,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -353,7 +365,7 @@ template <class T>
 struct S6;
 
 template<>
-struct S6<int>  // expected-note {{mappable type cannot be polymorphic}}
+struct S6<int>
 {
    virtual void foo();
 };
@@ -420,8 +432,8 @@ T tmain(T argc) {
 #pragma omp target data map(tofrom: argc > 0 ? x : y) // expected-error 2 {{expected expression containing only member accesses and/or array sections based on named variables}}
 #pragma omp target data map(argc)
 #pragma omp target data map(S1) // expected-error {{'S1' does not refer to a value}}
-#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
-#pragma omp target data map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
+#pragma omp target data map(ba)
 #pragma omp target data map(ca)
 #pragma omp target data map(da)
 #pragma omp target data map(S2::S2s)
@@ -454,6 +466,25 @@ T tmain(T argc) {
   return 0;
 }
 
+struct SA1{
+  int a;
+  struct SA1 *p;
+  int b[10];
+};
+struct SB1{
+  int a;
+  struct SA1 s;
+  struct SA1 sa[10];
+  struct SA1 *sp[10];
+  struct SA1 *p;
+};
+struct SC1{
+  int a;
+  struct SB1 s;
+  struct SB1 *p;
+  int b[10];
+};
+
 int main(int argc, char **argv) {
   const int d = 5;
   const int da[5] = { 0 };
@@ -467,6 +498,8 @@ int main(int argc, char **argv) {
   int y;
   int to, tofrom, always;
   const int (&l)[5] = da;
+  SC1 s;
+  SC1 *p;
 #pragma omp target data map // expected-error {{expected '(' after 'map'}} expected-error {{expected at least one 'map' or 'use_device_ptr' clause for '#pragma omp target data'}}
 #pragma omp target data map( // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{expected expression}}
 #pragma omp target data map() // expected-error {{expected expression}}
@@ -489,9 +522,9 @@ int main(int argc, char **argv) {
 #pragma omp target data map(tofrom: argc > 0 ? argv[1] : argv[2]) // expected-error {{xpected expression containing only member accesses and/or array sections based on named variables}}
 #pragma omp target data map(argc)
 #pragma omp target data map(S1) // expected-error {{'S1' does not refer to a value}}
-#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
 #pragma omp target data map(argv[1])
-#pragma omp target data map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target data map(ba)
 #pragma omp target data map(ca)
 #pragma omp target data map(da)
 #pragma omp target data map(S2::S2s)
@@ -525,8 +558,55 @@ int main(int argc, char **argv) {
   {}
 #pragma omp target firstprivate(j) map(j)  // expected-error {{firstprivate variable cannot be in a map clause in '#pragma omp target' directive}} expected-note {{defined as firstprivate}}
   {}
-#pragma omp target map(m) // expected-error {{type 'S6<int>' is not mappable to target}}
-  {}
+#pragma omp target map(m)
+  {}
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.s)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.s.a)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.b[:5])
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.p[:5])
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.sa[3].a)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.sp[3]->a)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.p->a)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.p->a)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.s.b[:2])
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.s.p->b[:2])
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+// expected-note@+1 {{used here}}
+#pragma omp target map(s.p->p->p->a)
+// expected-error@+1 {{variable already marked as mapped in current construct}}
+  { s.a++; }
+#pragma omp target map(s.s.s.b[:2])
+  { s.s.s.b[0]++; }
+
   return tmain<int, 3>(argc)+tmain<from, 4>(argc); // expected-note {{in instantiation of function template specialization 'tmain<int, 3>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<int, 4>' requested here}}
 }
 #endif
diff --git a/test/OpenMP/target_messages.cpp b/test/OpenMP/target_messages.cpp
index 6f79f44627fa..77d808493789 100644
--- a/test/OpenMP/target_messages.cpp
+++ b/test/OpenMP/target_messages.cpp
@@ -4,6 +4,8 @@
 // RUN: not %clang_cc1 -fopenmp -std=c++11 -triple nvptx64-nvidia-cuda -o - %s 2>&1 | FileCheck --check-prefix CHECK-UNSUPPORTED-HOST-TARGET %s
 // RUN: not %clang_cc1 -fopenmp -std=c++11 -triple nvptx-nvidia-cuda -o - %s 2>&1 | FileCheck --check-prefix CHECK-UNSUPPORTED-HOST-TARGET %s
 // CHECK-UNSUPPORTED-HOST-TARGET: error: The target '{{nvptx64-nvidia-cuda|nvptx-nvidia-cuda}}' is not a supported OpenMP host target.
+// RUN: not %clang_cc1 -fopenmp -std=c++11 -fopenmp-targets=hexagon-linux-gnu -o - %s 2>&1 | FileCheck --check-prefix CHECK-UNSUPPORTED-DEVICE-TARGET %s
+// CHECK-UNSUPPORTED-DEVICE-TARGET: OpenMP target is invalid: 'hexagon-linux-gnu'
 
 void foo() {
 }
diff --git a/test/OpenMP/target_parallel_codegen.cpp b/test/OpenMP/target_parallel_codegen.cpp
index 3063ab16abc8..b9d00da02f45 100644
--- a/test/OpenMP/target_parallel_codegen.cpp
+++ b/test/OpenMP/target_parallel_codegen.cpp
@@ -39,15 +39,15 @@
 // sizes.
 
 // CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i{{32|64}}] [i[[SZ:32|64]] 2]
-// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
-// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i32] [i32 288, i32 288]
-// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i32] [i32 288, i32 35, i32 288, i32 35, i32 35, i32 288, i32 288, i32 35, i32 35]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547]
 // CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
-// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i32] [i32 288, i32 288, i32 35]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
 // CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
-// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i32] [i32 288, i32 288, i32 288, i32 35]
-// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i32] [i32 35, i32 288, i32 288, i32 288, i32 35]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
 // CHECK-DAG: @{{.*}} = private constant i8 0
 // CHECK-DAG: @{{.*}} = private constant i8 0
 // CHECK-DAG: @{{.*}} = private constant i8 0
@@ -93,30 +93,24 @@ int foo(int n) {
   double cn[5][n];
   TT<long long, char> d;
 
-  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i32* null, i32 1, i32 0)
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target_teams_nowait(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i64* null, i32 1, i32 0)
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT0:@.+]]()
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
-  #pragma omp target parallel
+  #pragma omp target parallel nowait
   {
   }
 
-  // CHECK:       store i32 0, i32* [[RHV:%.+]], align 4
-  // CHECK:       store i32 -1, i32* [[RHV]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
   // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}})
   #pragma omp target parallel if(target: 0)
   {
     a += 1;
   }
 
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i32* getelementptr inbounds ([1 x i32], [1 x i32]* [[MAPT2]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT2]], i32 0, i32 0), i32 1, i32 0)
   // CHECK-DAG:   [[BP]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[P]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
@@ -126,9 +120,7 @@ int foo(int n) {
   // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
   // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
 
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}})
@@ -142,7 +134,7 @@ int foo(int n) {
   // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
   // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CHECK:       [[IFTHEN]]
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i32* getelementptr inbounds ([2 x i32], [2 x i32]* [[MAPT3]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 1, i32 0)
   // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
 
@@ -159,21 +151,18 @@ int foo(int n) {
   // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
   // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
   // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK-NEXT:  br label %[[IFEND:.+]]
-
-  // CHECK:       [[IFELSE]]
-  // CHECK:       store i32 -1, i32* [[RHV]], align 4
-  // CHECK-NEXT:  br label %[[IFEND:.+]]
-
-  // CHECK:       [[IFEND]]
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
+
   #pragma omp target parallel if(target: n>10)
   {
     a += 1;
@@ -197,7 +186,7 @@ int foo(int n) {
   // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
   // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
   // CHECK:       [[TRY]]
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i32* getelementptr inbounds ([9 x i32], [9 x i32]* [[MAPT4]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0), i32 1, i32 0)
   // CHECK-DAG:   [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
@@ -286,9 +275,7 @@ int foo(int n) {
   // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
   // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
 
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
 
   // CHECK:       [[FAIL]]
@@ -441,7 +428,7 @@ int foo(int n) {
 // CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 9, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*)* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]])
 //
 //
-// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* %{{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* %{{.+}}, [[TT]]* {{.+}})
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}})
 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
 
 template<typename tx>
@@ -533,7 +520,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
 // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
 // CHECK:       [[TRY]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i32* getelementptr inbounds ([5 x i32], [5 x i32]* [[MAPT7]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0), i32 1, i32 0)
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
@@ -582,9 +569,7 @@ int bar(int n){
 // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
 // CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
 
 // CHECK:       [[FAIL]]
@@ -598,7 +583,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
 // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[IFTHEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i32* getelementptr inbounds ([4 x i32], [4 x i32]* [[MAPT6]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0), i32 1, i32 0)
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0
 
@@ -630,21 +615,17 @@ int bar(int n){
 // CHECK-DAG:   store [10 x i32]* [[VAL3:%.+]], [10 x i32]** [[CBPADDR3]],
 // CHECK-DAG:   store [10 x i32]* [[VAL3]], [10 x i32]** [[CPADDR3]],
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFEND]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
@@ -652,7 +633,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
 // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[IFTHEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i32* getelementptr inbounds ([3 x i32], [3 x i32]* [[MAPT5]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0), i32 1, i32 0)
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
 
@@ -677,23 +658,17 @@ int bar(int n){
 // CHECK-DAG:   store [10 x i32]* [[VAL2:%.+]], [10 x i32]** [[CBPADDR2]],
 // CHECK-DAG:   store [10 x i32]* [[VAL2]], [10 x i32]** [[CPADDR2]],
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFEND]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
-
-
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 
 // Check that the offloading functions are emitted and that the arguments are
 // correct and loaded correctly for the target regions of the callees of bar().
@@ -728,7 +703,7 @@ int bar(int n){
 // CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]])
 //
 //
-// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* %{{.+}})
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}})
 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
 
 
diff --git a/test/OpenMP/target_parallel_codegen_registration.cpp b/test/OpenMP/target_parallel_codegen_registration.cpp
index 2511bc4c9d3c..d2cb1ec6f778 100644
--- a/test/OpenMP/target_parallel_codegen_registration.cpp
+++ b/test/OpenMP/target_parallel_codegen_registration.cpp
@@ -63,40 +63,40 @@
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // TCHECK-NOT: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 
 // CHECK-NTARGET-NOT: private constant i8 0
 // CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
diff --git a/test/OpenMP/target_parallel_debug_codegen.cpp b/test/OpenMP/target_parallel_debug_codegen.cpp
new file mode 100644
index 000000000000..385f7a29cb8f
--- /dev/null
+++ b/test/OpenMP/target_parallel_debug_codegen.cpp
@@ -0,0 +1,123 @@
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -debug-info-kind=limited | FileCheck %s
+// expected-no-diagnostics
+
+int main() {
+  /* int(*b)[a]; */
+  /* int *(**c)[a]; */
+  bool bb;
+  int a;
+  int b[10][10];
+  int c[10][10][10];
+#pragma omp target parallel firstprivate(a, b) map(tofrom          \
+                                                   : c) map(tofrom \
+                                                            : bb) if (a)
+  {
+    int &f = c[1][1][1];
+    int &g = a;
+    int &h = b[1][1];
+    int d = 15;
+    a = 5;
+    b[0][a] = 10;
+    c[0][0][a] = 11;
+    b[0][a] = c[0][0][a];
+    bb |= b[0][a];
+  }
+#pragma omp target parallel firstprivate(a) map(tofrom         \
+                                                : c, b) map(to \
+                                                            : bb)
+  {
+    int &f = c[1][1][1];
+    int &g = a;
+    int &h = b[1][1];
+    int d = 15;
+    a = 5;
+    b[0][a] = 10;
+    c[0][0][a] = 11;
+    b[0][a] = c[0][0][a];
+    d = bb;
+  }
+#pragma omp target parallel map(tofrom              \
+                                : a, c, b) map(from \
+                                               : bb)
+  {
+    int &f = c[1][1][1];
+    int &g = a;
+    int &h = b[1][1];
+    int d = 15;
+    a = 5;
+    b[0][a] = 10;
+    c[0][0][a] = 11;
+    b[0][a] = c[0][0][a];
+    bb = b[0][a];
+  }
+  return 0;
+}
+
+// CHECK: define internal void @__omp_offloading{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8 addrspace(1)* noalias{{[^,]+}}, i1 {{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: call void [[NONDEBUG_WRAPPER:.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* {{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8* {{[^)]+}})
+
+// CHECK: define internal void [[DEBUG_PARALLEL:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+
+// CHECK: define internal void [[NONDEBUG_WRAPPER]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: call void [[DEBUG_PARALLEL]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: call void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define internal void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+// CHECK: call void [[NONDEBUG_WRAPPER:.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* {{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8* {{[^)]+}})
+
+// CHECK: define internal void [[DEBUG_PARALLEL:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+
+// CHECK: define internal void [[NONDEBUG_WRAPPER]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void [[DEBUG_PARALLEL]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define internal void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 addrspace(1)* noalias{{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast i32 addrspace(1)* %{{.+}} to i32*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+// CHECK: call void [[NONDEBUG_WRAPPER:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8* {{[^)]+}})
+
+// CHECK: define internal void [[DEBUG_PARALLEL:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 addrspace(1)* noalias{{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast i32 addrspace(1)* %{{.+}} to i32*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+
+// CHECK: define internal void [[NONDEBUG_WRAPPER]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i32* dereferenceable{{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast i32* %{{.+}} to i32 addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void [[DEBUG_PARALLEL]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 addrspace(1)* {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i32* dereferenceable{{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast i32* %{{.+}} to i32 addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 addrspace(1)* {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: !DILocalVariable(name: ".global_tid.",
+// CHECK-SAME: DIFlagArtificial
+// CHECK: !DILocalVariable(name: ".bound_tid.",
+// CHECK-SAME: DIFlagArtificial
+// CHECK: !DILocalVariable(name: "c",
+// CHECK-SAMEi-NOT: DIFlagArtificial
+// CHECK: !DILocalVariable(name: "a",
+// CHECK-SAMEi-NOT: DIFlagArtificial
+// CHECK: !DILocalVariable(name: "b",
+// CHECK-SAMEi-NOT: DIFlagArtificial
diff --git a/test/OpenMP/target_parallel_depend_messages.cpp b/test/OpenMP/target_parallel_depend_messages.cpp
index fde940be1aa4..24c69e95cc0a 100644
--- a/test/OpenMP/target_parallel_depend_messages.cpp
+++ b/test/OpenMP/target_parallel_depend_messages.cpp
@@ -36,21 +36,21 @@ int main(int argc, char **argv, char *env[]) {
   foo();
   #pragma omp target parallel depend (out: ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target parallel depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target parallel depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
   #pragma omp target parallel depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-  #pragma omp target parallel depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel depend(in : argv[1][1] = '2')
   foo();
-  #pragma omp target parallel depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target parallel depend (in : argv[0])
   foo();
   #pragma omp target parallel depend (in : ) // expected-error {{expected expression}}
   foo();
-  #pragma omp target parallel depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel depend (in : main)
   foo();
-  #pragma omp target parallel depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target parallel depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
   #pragma omp target parallel depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
diff --git a/test/OpenMP/target_parallel_for_codegen.cpp b/test/OpenMP/target_parallel_for_codegen.cpp
new file mode 100644
index 000000000000..58ac10bf9b2c
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_codegen.cpp
@@ -0,0 +1,812 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// Test target codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+// CHECK-DAG: [[TT:%.+]] = type { i64, i8 }
+// CHECK-DAG: [[S1:%.+]] = type { double }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK: [[ENTTY:%.+]] = type { i8*, i8*, i{{32|64}}, i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// We have 8 target regions, but only 7 that actually will generate offloading
+// code, only 6 will have mapped arguments, and only 4 have all-constant map
+// sizes.
+
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288]
+// CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [10 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547, i64 288]
+// CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK-NOT: @{{.+}} = constant [[ENTTY]]
+
+// Check if offloading descriptor is created.
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// Check target registration is registered as a Ctor.
+// CHECK: appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+
+template<typename tx, typename ty>
+struct TT{
+  tx X;
+  ty Y;
+};
+
+// CHECK-LABEL: get_val
+long long get_val() { return 0; }
+
+// CHECK: define {{.*}}[[FOO:@.+]](
+int foo(int n) {
+  int a = 0;
+  short aa = 0;
+  float b[10];
+  float bn[n];
+  double c[5][10];
+  double cn[5][n];
+  TT<long long, char> d;
+
+  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i64* null, i32 1, i32 0)
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT0:@.+]]()
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target parallel for
+  for (int i = 3; i < 32; i += 5) {
+#pragma omp cancel for
+#pragma omp cancellation point for
+  }
+
+  // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}}, i{{32|64}}{{[*]*}} {{[^)]+}})
+  long long k = get_val();
+  #pragma omp target parallel for if(target: 0) linear(k : 3) schedule(dynamic)
+  for (int i = 10; i > 1; i--) {
+    a += 1;
+  }
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams_nowait(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT2]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2]],
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  int lin = 12;
+  #pragma omp target parallel for if(target: 1) linear(lin, a : get_val()) nowait
+  for (unsigned long long it = 2000; it >= 600; it-=400) {
+    aa += 1;
+  }
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
+  // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+  // CHECK:       [[IFTHEN]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
+
+  #pragma omp target parallel for if(target: n>10)
+  for (short it = 6; it <= 20; it-=-4) {
+    a += 1;
+    aa += 1;
+  }
+
+  // We capture 3 VLA sizes in this target region
+  // CHECK:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-64:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-64:       [[A_ADDR:%.+]] = bitcast i[[SZ]]* [[A_CADDR:%.+]] to i32*
+  // CHECK-64:       store i32 [[A_VAL]], i32* [[A_ADDR]],
+  // CHECK-64:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK-32:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-32:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-32:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK:       [[BNSIZE:%.+]] = mul nuw i[[SZ]] [[VLA0:%.+]], 4
+  // CHECK:       [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]]
+  // CHECK:       [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
+  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+  // CHECK:       [[TRY]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 10, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([10 x i64], [10 x i64]* [[MAPT4]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[SR]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX2:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX3:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[SADDR4:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX4:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[SADDR5:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX5:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR5:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[PADDR5:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[SADDR6:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX6:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR6:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[PADDR6:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[SADDR7:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX7:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR7:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[PADDR7:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[SADDR8:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX8:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR8:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX8]]
+  // CHECK-DAG:   [[PADDR8:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX8]]
+
+  // The names below are not necessarily consistent with the names used for the
+  // addresses above as some are repeated.
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+  // CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CBPADDR1:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CPADDR1:%.+]],
+  // CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CBPADDR2:%.+]],
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CPADDR2:%.+]],
+  // CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CBPADDR3:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CPADDR3:%.+]],
+  // CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CBPADDR4:%.+]],
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CPADDR4:%.+]],
+  // CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   store i[[SZ]] 40, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CBPADDR5:%.+]],
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CPADDR5:%.+]],
+  // CHECK-DAG:   [[CBPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   [[CPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   store i[[SZ]] [[BNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CBPADDR6:%.+]],
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CPADDR6:%.+]],
+  // CHECK-DAG:   [[CBPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   [[CPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   store i[[SZ]] 400, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CBPADDR7:%.+]],
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CPADDR7:%.+]],
+  // CHECK-DAG:   [[CBPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   [[CPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   store i[[SZ]] [[CNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CBPADDR8:%.+]],
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8:%.+]],
+  // CHECK-DAG:   [[CBPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target parallel for if(target: n>20) schedule(static, a)
+  for (unsigned char it = 'z'; it >= 'a'; it+=-1) {
+    a += 1;
+    b[2] += 1.0;
+    bn[3] += 1.0;
+    c[1][2] += 1.0;
+    cn[1][3] += 1.0;
+    d.X += 1;
+    d.Y += 1;
+  }
+
+  return a;
+}
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions in foo().
+
+// CHECK:       define internal void [[HVT0]]()
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*))
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid.)
+// CHECK:       call i32 @__kmpc_cancel(%ident_t* @
+// CHECK:       call i32 @__kmpc_cancellationpoint(%ident_t* @
+// CHECK:       ret void
+// CHECK:       }
+
+
+// CHECK:       define internal void [[HVT1]](i[[SZ]] %{{.+}}, i{{32|64}}{{[*]*.*}} %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       alloca i{{32|64}}{{[*]*}}, align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK-64:    [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i32*
+// CHECK-64:    store i32 [[AA]], i32* [[AA_C]], align
+// CHECK-32:    store i32 [[AA]], i32* [[AA_CASTED]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i{{32|64}}{{[*]*}})* [[OMP_OUTLINED1:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]], i{{32|64}}{{[*]*}} %{{.+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED1]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT2]](i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       alloca i[[SZ]], align
+// CHECK:       alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK:       store i16 [[AA]], i16* [[AA_C]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]])* [[OMP_OUTLINED2:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]], i[[SZ]] {{.+}}, i[[SZ]] {{.+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED2]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT3]]
+// Create stack storage and store argument in there.
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[A_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK-64-DAG:[[A:%.+]] = load i32, i32* [[A_CADDR]], align
+// CHECK-32-DAG:[[A:%.+]] = load i32, i32* [[A_ADDR]], align
+// CHECK-64-DAG:[[A_C:%.+]] = bitcast i[[SZ]]* [[A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[A]], i32* [[A_C]], align
+// CHECK-32-DAG:store i32 [[A]], i32* [[A_CASTED]], align
+// CHECK-DAG:   [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK-DAG:   [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[AA]], i16* [[AA_C]], align
+// CHECK-DAG:   [[PARAM1:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CASTED]], align
+// CHECK-DAG:   [[PARAM2:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK-DAG:   call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]])* [[OMP_OUTLINED3:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM1]], i[[SZ]] [[PARAM2]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED3]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT4]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x float]*
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_BN:%.+]] = alloca float*
+// CHECK:       [[LOCAL_C:%.+]] = alloca [5 x [10 x double]]*
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA3:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_CN:%.+]] = alloca double*
+// CHECK:       [[LOCAL_D:%.+]] = alloca [[TT]]*
+// CHECK:       alloca i[[SZ]]
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store [10 x float]* [[ARG_B:%.+]], [10 x float]** [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store float* [[ARG_BN:%.+]], float** [[LOCAL_BN]]
+// CHECK-DAG:   store [5 x [10 x double]]* [[ARG_C:%.+]], [5 x [10 x double]]** [[LOCAL_C]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA3:%.+]], i[[SZ]]* [[LOCAL_VLA3]]
+// CHECK-DAG:   store double* [[ARG_CN:%.+]], double** [[LOCAL_CN]]
+// CHECK-DAG:   store [[TT]]* [[ARG_D:%.+]], [[TT]]** [[LOCAL_D]]
+
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x float]*, [10 x float]** [[LOCAL_B]],
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[REF_BN:%.+]] = load float*, float** [[LOCAL_BN]],
+// CHECK-DAG:   [[REF_C:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[LOCAL_C]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[VAL_VLA3:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA3]],
+// CHECK-DAG:   [[REF_CN:%.+]] = load double*, double** [[LOCAL_CN]],
+// CHECK-DAG:   [[REF_D:%.+]] = load [[TT]]*, [[TT]]** [[LOCAL_D]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 10, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*, i[[SZ]])* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]], i[[SZ]] %{{.+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+template<typename tx>
+tx ftemplate(int n) {
+  tx a = 0;
+  short aa = 0;
+  tx b[10];
+
+  #pragma omp target parallel for if(target: n>40)
+  for (long long i = -10; i < 10; i += 3) {
+    a += 1;
+    aa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+static
+int fstatic(int n) {
+  int a = 0;
+  short aa = 0;
+  char aaa = 0;
+  int b[10];
+
+  #pragma omp target parallel for if(target: n>50)
+  for (unsigned i=100; i<10; i+=10) {
+    a += 1;
+    aa += 1;
+    aaa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+struct S1 {
+  double a;
+
+  int r1(int n){
+    int b = n+1;
+    short int c[2][n];
+
+    #pragma omp target parallel for if(target: n>60)
+    for (unsigned long long it = 2000; it >= 600; it -= 400) {
+      this->a = (double)b + 1.5;
+      c[1][1] = ++a;
+    }
+
+    return c[1][1] + (int)b;
+  }
+};
+
+// CHECK: define {{.*}}@{{.*}}bar{{.*}}
+int bar(int n){
+  int a = 0;
+
+  // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}})
+  a += foo(n);
+
+  S1 S;
+  // CHECK: call {{.*}}i32 [[FS1:@.+]]([[S1]]* {{.*}}, i32 {{.*}})
+  a += S.r1(n);
+
+  // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}})
+  a += fstatic(n);
+
+  // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}})
+  a += ftemplate<int>(n);
+
+  return a;
+}
+
+//
+// CHECK: define {{.*}}[[FS1]]
+//
+// CHECK:          i8* @llvm.stacksave()
+// CHECK-64:       [[B_ADDR:%.+]] = bitcast i[[SZ]]* [[B_CADDR:%.+]] to i32*
+// CHECK-64:       store i32 %{{.+}}, i32* [[B_ADDR]],
+// CHECK-64:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_CADDR]],
+
+// CHECK-32:       store i32 %{{.+}}, i32* [[B_ADDR:%.+]],
+// CHECK-32:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_ADDR]],
+
+// We capture 2 VLA sizes in this target region
+// CHECK:       [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]]
+// CHECK:       [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2
+
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
+// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+// CHECK:       [[TRY]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX0:[0-9]+]]
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX0]]
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX0]]
+// CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX1:[0-9]+]]
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX1]]
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX1]]
+// CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX2:[0-9]+]]
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX2]]
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX2]]
+// CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX3:[0-9]+]]
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX3]]
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX3]]
+
+// The names below are not necessarily consistent with the names used for the
+// addresses above as some are repeated.
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR3:%.+]],
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   store i[[SZ]] 8, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
+
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+
+//
+// CHECK: define {{.*}}[[FSTATIC]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2]],
+
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 3
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 3
+// CHECK-DAG:   [[CBPADDR3:%.+]] = bitcast i8** [[BPADDR3]] to [10 x i32]**
+// CHECK-DAG:   [[CPADDR3:%.+]] = bitcast i8** [[PADDR3]] to [10 x i32]**
+// CHECK-DAG:   store [10 x i32]* [[VAL3:%.+]], [10 x i32]** [[CBPADDR3]],
+// CHECK-DAG:   store [10 x i32]* [[VAL3]], [10 x i32]** [[CPADDR3]],
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+//
+// CHECK: define {{.*}}[[FTEMPLATE]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to [10 x i32]**
+// CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to [10 x i32]**
+// CHECK-DAG:   store [10 x i32]* [[VAL2:%.+]], [10 x i32]** [[CBPADDR2]],
+// CHECK-DAG:   store [10 x i32]* [[VAL2]], [10 x i32]** [[CPADDR2]],
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions of the callees of bar().
+
+// CHECK:       define internal void [[HVT7]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_THIS:%.+]] = alloca [[S1]]*
+// CHECK:       [[LOCAL_B:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_C:%.+]] = alloca i16*
+// CHECK:       [[LOCAL_B_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store [[S1]]* [[ARG_THIS:%.+]], [[S1]]** [[LOCAL_THIS]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_B:%.+]], i[[SZ]]* [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i16* [[ARG_C:%.+]], i16** [[LOCAL_C]]
+// Store captures in the context.
+// CHECK-DAG:   [[REF_THIS:%.+]] = load [[S1]]*, [[S1]]** [[LOCAL_THIS]],
+// CHECK-64-DAG:[[CONV_BP:%.+]] = bitcast i[[SZ]]* [[LOCAL_B]] to i32*
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[REF_C:%.+]] = load i16*, i16** [[LOCAL_C]],
+
+// CHECK-64-DAG:[[CONV_B:%.+]] = load i32, i32* [[CONV_BP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_B_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_B]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_BV:%.+]] = load i32, i32* [[LOCAL_B]]
+// CHECK-32-DAG:store i32 [[LOCAL_BV]], i32* [[LOCAL_B_CASTED]], align
+// CHECK-DAG:   [[REF_B:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_B_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+
+// CHECK:       define internal void [[HVT6]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AAA:%.+]], i[[SZ]]* [[LOCAL_AAA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[CONV_AAAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA]] to i8*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK-DAG:   [[CONV_AAA:%.+]] = load i8, i8* [[CONV_AAAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA_CASTED]] to i8*
+// CHECK-DAG:   store i8 [[CONV_AAA]], i8* [[CONV]], align
+// CHECK-DAG:   [[REF_AAA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AAA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED6:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], i[[SZ]] [[REF_AAA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED6]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+// CHECK:       define internal void [[HVT5]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED7:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED7]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+#endif
diff --git a/test/OpenMP/target_parallel_for_codegen_registration.cpp b/test/OpenMP/target_parallel_for_codegen_registration.cpp
new file mode 100644
index 000000000000..24f0c06771ec
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_codegen_registration.cpp
@@ -0,0 +1,451 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target parallel for codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// Check that no target code is emmitted if no omptests flag was provided.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK-NTARGET
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK:    [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// CHECK-DAG: [[A1:@.+]] = internal global [[SA]]
+// CHECK-DAG: [[A2:@.+]] = global [[SA]]
+// CHECK-DAG: [[B1:@.+]] = global [[SB]]
+// CHECK-DAG: [[B2:@.+]] = global [[SB]]
+// CHECK-DAG: [[C1:@.+]] = internal global [[SC]]
+// CHECK-DAG: [[D1:@.+]] = global [[SD]]
+// CHECK-DAG: [[E1:@.+]] = global [[SE]]
+// CHECK-DAG: [[T1:@.+]] = global [[ST1]]
+// CHECK-DAG: [[T2:@.+]] = global [[ST2]]
+
+// CHECK-NTARGET-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-NTARGET-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-NTARGET-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-NTARGET-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-NTARGET-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-NTARGET-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-NTARGET-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-NTARGET-NOT: type { i8*, i8*, %
+// CHECK-NTARGET-NOT: type { i32, %
+
+// We have 7 target regions
+
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// TCHECK-NOT: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+
+// CHECK-NTARGET-NOT: private constant i8 0
+// CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
+
+// CHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// CHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// CHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// CHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// CHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// CHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// CHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// CHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// CHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// CHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// CHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// CHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// CHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// TCHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// TCHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// TCHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// TCHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// TCHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// TCHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// TCHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// TCHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// TCHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// TCHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// TCHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// TCHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// TCHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// We have 4 initializers, one for the 500 priority, another one for 501, or more for the default priority, and the last one for the offloading registration function.
+// CHECK: @llvm.global_ctors = appending global [4 x { i32, void ()*, i8* }] [
+// CHECK-SAME: { i32, void ()*, i8* } { i32 500, void ()* [[P500:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 501, void ()* [[P501:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 65535, void ()* [[PMAX:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+// CHECK-NTARGET: @llvm.global_ctors = appending global [3   x { i32, void ()*, i8* }] [
+
+extern int *R;
+
+struct SA {
+  int arr[4];
+  void foo() {
+    int a = *R;
+    a += 1;
+    *R = a;
+  }
+  SA() {
+    int a = *R;
+    a += 2;
+    *R = a;
+  }
+  ~SA() {
+    int a = *R;
+    a += 3;
+    *R = a;
+  }
+};
+
+struct SB {
+  int arr[8];
+  void foo() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 4;
+    *R = a;
+  }
+  SB() {
+    int a = *R;
+    a += 5;
+    *R = a;
+  }
+  ~SB() {
+    int a = *R;
+    a += 6;
+    *R = a;
+  }
+};
+
+struct SC {
+  int arr[16];
+  void foo() {
+    int a = *R;
+    a += 7;
+    *R = a;
+  }
+  SC() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 8;
+    *R = a;
+  }
+  ~SC() {
+    int a = *R;
+    a += 9;
+    *R = a;
+  }
+};
+
+struct SD {
+  int arr[32];
+  void foo() {
+    int a = *R;
+    a += 10;
+    *R = a;
+  }
+  SD() {
+    int a = *R;
+    a += 11;
+    *R = a;
+  }
+  ~SD() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 12;
+    *R = a;
+  }
+};
+
+struct SE {
+  int arr[64];
+  void foo() {
+    int a = *R;
+    #pragma omp target parallel for if(target: 0)
+    for (int i = 0; i < 10; ++i)
+      a += 13;
+    *R = a;
+  }
+  SE() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 14;
+    *R = a;
+  }
+  ~SE() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 15;
+    *R = a;
+  }
+};
+
+template <int x>
+struct ST {
+  int arr[128 + x];
+  void foo() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 16 + x;
+    *R = a;
+  }
+  ST() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 17 + x;
+    *R = a;
+  }
+  ~ST() {
+    int a = *R;
+    #pragma omp target parallel for
+    for (int i = 0; i < 10; ++i)
+      a += 18 + x;
+    *R = a;
+  }
+};
+
+// We have to make sure we us all the target regions:
+//CHECK-DAG: define internal void @[[NAME1]](
+//CHECK-DAG: call void @[[NAME1]](
+//CHECK-DAG: define internal void @[[NAME2]](
+//CHECK-DAG: call void @[[NAME2]](
+//CHECK-DAG: define internal void @[[NAME3]](
+//CHECK-DAG: call void @[[NAME3]](
+//CHECK-DAG: define internal void @[[NAME4]](
+//CHECK-DAG: call void @[[NAME4]](
+//CHECK-DAG: define internal void @[[NAME5]](
+//CHECK-DAG: call void @[[NAME5]](
+//CHECK-DAG: define internal void @[[NAME6]](
+//CHECK-DAG: call void @[[NAME6]](
+//CHECK-DAG: define internal void @[[NAME7]](
+//CHECK-DAG: call void @[[NAME7]](
+//CHECK-DAG: define internal void @[[NAME8]](
+//CHECK-DAG: call void @[[NAME8]](
+//CHECK-DAG: define internal void @[[NAME9]](
+//CHECK-DAG: call void @[[NAME9]](
+//CHECK-DAG: define internal void @[[NAME10]](
+//CHECK-DAG: call void @[[NAME10]](
+//CHECK-DAG: define internal void @[[NAME11]](
+//CHECK-DAG: call void @[[NAME11]](
+//CHECK-DAG: define internal void @[[NAME12]](
+//CHECK-DAG: call void @[[NAME12]](
+
+//TCHECK-DAG: define void @[[NAME1]](
+//TCHECK-DAG: define void @[[NAME2]](
+//TCHECK-DAG: define void @[[NAME3]](
+//TCHECK-DAG: define void @[[NAME4]](
+//TCHECK-DAG: define void @[[NAME5]](
+//TCHECK-DAG: define void @[[NAME6]](
+//TCHECK-DAG: define void @[[NAME7]](
+//TCHECK-DAG: define void @[[NAME8]](
+//TCHECK-DAG: define void @[[NAME9]](
+//TCHECK-DAG: define void @[[NAME10]](
+//TCHECK-DAG: define void @[[NAME11]](
+//TCHECK-DAG: define void @[[NAME12]](
+
+// CHECK-NTARGET-NOT: __tgt_target
+// CHECK-NTARGET-NOT: __tgt_register_lib
+// CHECK-NTARGET-NOT: __tgt_unregister_lib
+
+// TCHECK-NOT: __tgt_target
+// TCHECK-NOT: __tgt_register_lib
+// TCHECK-NOT: __tgt_unregister_lib
+
+// We have 2 initializers with priority 500
+//CHECK: define internal void [[P500]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 1 initializers with priority 501
+//CHECK: define internal void [[P501]](
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 6 initializers with default priority
+//CHECK: define internal void [[PMAX]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// Check registration and unregistration
+
+//CHECK:     define internal void @[[UNREGFN:.+]](i8*)
+//CHECK-SAME: comdat($[[REGFN]]) {
+//CHECK:     call i32 @__tgt_unregister_lib([[DSCTY]]* [[DESC]])
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_unregister_lib([[DSCTY]]*)
+
+//CHECK:     define linkonce hidden void @[[REGFN]](i8*)
+//CHECK-SAME: comdat {
+//CHECK:     call i32 @__tgt_register_lib([[DSCTY]]* [[DESC]])
+//CHECK:     call i32 @__cxa_atexit(void (i8*)* @[[UNREGFN]], i8* bitcast ([[DSCTY]]* [[DESC]] to i8*),
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_register_lib([[DSCTY]]*)
+
+static __attribute__((init_priority(500))) SA a1;
+SA a2;
+SB __attribute__((init_priority(500))) b1;
+SB __attribute__((init_priority(501))) b2;
+static SC c1;
+SD d1;
+SE e1;
+ST<100> t1;
+ST<1000> t2;
+
+
+int bar(int a){
+  int r = a;
+
+  a1.foo();
+  a2.foo();
+  b1.foo();
+  b2.foo();
+  c1.foo();
+  d1.foo();
+  e1.foo();
+  t1.foo();
+  t2.foo();
+
+  #pragma omp target parallel for
+  for (int i = 0; i < 10; ++i)
+    ++r;
+
+  return r + *R;
+}
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+#endif
diff --git a/test/OpenMP/target_parallel_for_codegen_registration_naming.cpp b/test/OpenMP/target_parallel_for_codegen_registration_naming.cpp
new file mode 100644
index 000000000000..b42b5b55a505
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_codegen_registration_naming.cpp
@@ -0,0 +1,68 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target parallel for codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[CA:%.+]] = type { i32* }
+
+// CHECK: define {{.*}}i32 @[[NNAME:.+]](i32 {{.*}}%{{.+}})
+int nested(int a){
+  // CHECK: call void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME]]_l[[T1L:[0-9]+]](
+  #pragma omp target parallel for
+  for (int i = 0; i < 10; ++i)
+    ++a;
+
+  // CHECK: call void @"[[LNAME:.+]]"([[CA]]*
+  auto F = [&](){
+    #pragma omp parallel
+    {
+      #pragma omp target parallel for
+      for (int i = 0; i < 10; ++i)
+        ++a;
+    }
+  };
+
+  F();
+
+  return a;
+}
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T1L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME:.+]]_l[[T1L:[0-9]+]](
+
+// CHECK: define {{.*}}void @"[[LNAME]]"(
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.+}}[[PNAME:@.+]] to
+
+// CHECK: define {{.*}}void [[PNAME]](
+// CHECK: call void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L:[0-9]+]](
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME:.+]]_l[[T2L:[0-9]+]](
+
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+#endif
diff --git a/test/OpenMP/target_parallel_for_debug_codegen.cpp b/test/OpenMP/target_parallel_for_debug_codegen.cpp
new file mode 100644
index 000000000000..0e0bb7e85306
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_debug_codegen.cpp
@@ -0,0 +1,113 @@
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - -debug-info-kind=limited | FileCheck %s
+// expected-no-diagnostics
+
+int main() {
+  /* int(*b)[a]; */
+  /* int *(**c)[a]; */
+  bool bb;
+  int a;
+  int b[10][10];
+  int c[10][10][10];
+#pragma omp target parallel for firstprivate(a, b) map(tofrom          \
+                                                       : c) map(tofrom \
+                                                                : bb) if (a)
+  for (int i = 0; i < 10; ++i) {
+    int &f = c[1][1][1];
+    int &g = a;
+    int &h = b[1][1];
+    int d = 15;
+    a = 5;
+    b[0][a] = 10;
+    c[0][0][a] = 11;
+    b[0][a] = c[0][0][a];
+    bb |= b[0][a];
+  }
+#pragma omp target parallel for firstprivate(a) map(tofrom         \
+                                                    : c, b) map(to \
+                                                                : bb)
+  for (int i = 0; i < 10; ++i) {
+    int &f = c[1][1][1];
+    int &g = a;
+    int &h = b[1][1];
+    int d = 15;
+    a = 5;
+    b[0][a] = 10;
+    c[0][0][a] = 11;
+    b[0][a] = c[0][0][a];
+    d = bb;
+  }
+#pragma omp target parallel for map(tofrom              \
+                                    : a, c, b) map(from \
+                                                   : bb)
+  for (int i = 0; i < 10; ++i) {
+    int &f = c[1][1][1];
+    int &g = a;
+    int &h = b[1][1];
+    int d = 15;
+    a = 5;
+    b[0][a] = 10;
+    c[0][0][a] = 11;
+    b[0][a] = c[0][0][a];
+    bb = b[0][a];
+  }
+  return 0;
+}
+
+// CHECK: define internal void @__omp_offloading{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8 addrspace(1)* noalias{{[^,]+}}, i1 {{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: call void [[NONDEBUG_WRAPPER:.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* {{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8* {{[^)]+}})
+
+// CHECK: define internal void [[DEBUG_PARALLEL:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+
+// CHECK: define internal void [[NONDEBUG_WRAPPER]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: call void [[DEBUG_PARALLEL]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: call void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define internal void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+// CHECK: call void [[NONDEBUG_WRAPPER:.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* {{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8* {{[^)]+}})
+
+// CHECK: define internal void [[DEBUG_PARALLEL:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+
+// CHECK: define internal void [[NONDEBUG_WRAPPER]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void [[DEBUG_PARALLEL]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i64 {{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define internal void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 addrspace(1)* noalias{{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast i32 addrspace(1)* %{{.+}} to i32*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+// CHECK: call void [[NONDEBUG_WRAPPER:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x i32]]* {{[^,]+}}, i8* {{[^)]+}})
+
+// CHECK: define internal void [[DEBUG_PARALLEL:@.+]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* noalias{{[^,]+}}, i32 addrspace(1)* noalias{{[^,]+}}, [10 x [10 x i32]] addrspace(1)* noalias{{[^,]+}}, i8 addrspace(1)* noalias{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]] addrspace(1)* %{{.+}} to [10 x [10 x [10 x i32]]]*
+// CHECK: addrspacecast i32 addrspace(1)* %{{.+}} to i32*
+// CHECK: addrspacecast [10 x [10 x i32]] addrspace(1)* %{{.+}} to [10 x [10 x i32]]*
+
+// CHECK: define internal void [[NONDEBUG_WRAPPER]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i32* dereferenceable{{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast i32* %{{.+}} to i32 addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void [[DEBUG_PARALLEL]](i32* {{[^,]+}}, i32* {{[^,]+}}, [10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 addrspace(1)* {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
+// CHECK: define void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]]* dereferenceable{{[^,]+}}, i32* dereferenceable{{[^,]+}}, [10 x [10 x i32]]* dereferenceable{{[^,]+}}, i8* dereferenceable{{[^)]+}})
+// CHECK: addrspacecast [10 x [10 x [10 x i32]]]* %{{.+}} to [10 x [10 x [10 x i32]]] addrspace(1)*
+// CHECK: addrspacecast i32* %{{.+}} to i32 addrspace(1)*
+// CHECK: addrspacecast [10 x [10 x i32]]* %{{.+}} to [10 x [10 x i32]] addrspace(1)*
+// CHECK: call void @__omp_offloading_{{[^(]+}}([10 x [10 x [10 x i32]]] addrspace(1)* {{[^,]+}}, i32 addrspace(1)* {{[^,]+}}, [10 x [10 x i32]] addrspace(1)* {{[^,]+}}, i8 addrspace(1)* {{[^)]+}})
+
diff --git a/test/OpenMP/target_parallel_for_depend_messages.cpp b/test/OpenMP/target_parallel_for_depend_messages.cpp
index cf17d7a5def2..19872566161a 100644
--- a/test/OpenMP/target_parallel_for_depend_messages.cpp
+++ b/test/OpenMP/target_parallel_for_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target parallel for depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel for depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel for depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel for depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target parallel for depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_parallel_for_lastprivate_messages.cpp b/test/OpenMP/target_parallel_for_lastprivate_messages.cpp
index c001b7f0d168..ee703a1ce32e 100644
--- a/test/OpenMP/target_parallel_for_lastprivate_messages.cpp
+++ b/test/OpenMP/target_parallel_for_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   S2 &operator=(const S2 &);
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/target_parallel_for_map_messages.cpp b/test/OpenMP/target_parallel_for_map_messages.cpp
index f0019f94f493..f3e1c616b406 100644
--- a/test/OpenMP/target_parallel_for_map_messages.cpp
+++ b/test/OpenMP/target_parallel_for_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -222,11 +222,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_parallel_for_reduction_messages.cpp b/test/OpenMP/target_parallel_for_reduction_messages.cpp
index 234b71aec218..9f1f68f25f83 100644
--- a/test/OpenMP/target_parallel_for_reduction_messages.cpp
+++ b/test/OpenMP/target_parallel_for_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_parallel_for_simd_ast_print.cpp b/test/OpenMP/target_parallel_for_simd_ast_print.cpp
index 82d72c3b1a4b..8f0ed97d5cd9 100644
--- a/test/OpenMP/target_parallel_for_simd_ast_print.cpp
+++ b/test/OpenMP/target_parallel_for_simd_ast_print.cpp
@@ -77,14 +77,14 @@ T tmain(T argc, T *argv) {
     a = 2;
 // CHECK-NEXT: for (T i = 0; i < 2; ++i)
 // CHECK-NEXT: a = 2;
-#pragma omp target parallel for simd private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) schedule(static, N) ordered(N) if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h)
+#pragma omp target parallel for simd private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) schedule(static, N) ordered if (parallel :argc) num_threads(N) default(shared) shared(e) reduction(+ : h)
   for (int i = 0; i < 2; ++i)
     for (int j = 0; j < 2; ++j)
       for (int j = 0; j < 2; ++j)
         for (int j = 0; j < 2; ++j)
           for (int j = 0; j < 2; ++j)
             foo();
-  // CHECK-NEXT: #pragma omp target parallel for simd private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) schedule(static, N) ordered(N) if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h)
+  // CHECK-NEXT: #pragma omp target parallel for simd private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) schedule(static, N) ordered if(parallel: argc) num_threads(N) default(shared) shared(e) reduction(+: h)
   // CHECK-NEXT: for (int i = 0; i < 2; ++i)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
diff --git a/test/OpenMP/target_parallel_for_simd_codegen.cpp b/test/OpenMP/target_parallel_for_simd_codegen.cpp
new file mode 100644
index 000000000000..d168214df35f
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_simd_codegen.cpp
@@ -0,0 +1,812 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// Test target codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+// CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+// CHECK-DAG: [[TT:%.+]] = type { i64, i8 }
+// CHECK-DAG: [[S1:%.+]] = type { double }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK: [[ENTTY:%.+]] = type { i8*, i8*, i{{32|64}}, i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// We have 8 target regions, but only 7 that actually will generate offloading
+// code, only 6 will have mapped arguments, and only 4 have all-constant map
+// sizes.
+
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288]
+// CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [10 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547, i64 288]
+// CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK-NOT: @{{.+}} = constant [[ENTTY]]
+
+// Check if offloading descriptor is created.
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// Check target registration is registered as a Ctor.
+// CHECK: appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+
+template<typename tx, typename ty>
+struct TT{
+  tx X;
+  ty Y;
+};
+
+// CHECK-LABEL: get_val
+long long get_val() { return 0; }
+
+// CHECK: define {{.*}}[[FOO:@.+]](
+int foo(int n) {
+  int a = 0;
+  short aa = 0;
+  float b[10];
+  float bn[n];
+  double c[5][10];
+  double cn[5][n];
+  TT<long long, char> d;
+
+  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target_teams_nowait(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i64* null, i32 1, i32 0)
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT0:@.+]]()
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target parallel for simd nowait
+  for (int i = 3; i < 32; i += 5) {
+  }
+
+  // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}}, i{{32|64}}{{[*]*}} {{[^)]+}})
+  long long k = get_val();
+  #pragma omp target parallel for simd if(target: 0) linear(k : 3) schedule(dynamic)
+  for (int i = 10; i > 1; i--) {
+    a += 1;
+  }
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT2]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2]],
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  int lin = 12;
+  #pragma omp target parallel for simd if(target: 1) linear(lin, a : get_val())
+  for (unsigned long long it = 2000; it >= 600; it-=400) {
+    aa += 1;
+  }
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
+  // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+  // CHECK:       [[IFTHEN]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
+
+  #pragma omp target parallel for simd if(target: n>10)
+  for (short it = 6; it <= 20; it-=-4) {
+    a += 1;
+    aa += 1;
+  }
+
+  // We capture 3 VLA sizes in this target region
+  // CHECK:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-64:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-64:       [[A_ADDR:%.+]] = bitcast i[[SZ]]* [[A_CADDR:%.+]] to i32*
+  // CHECK-64:       store i32 [[A_VAL]], i32* [[A_ADDR]],
+  // CHECK-64:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK-32:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-32:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-32:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK:       [[BNSIZE:%.+]] = mul nuw i[[SZ]] [[VLA0:%.+]], 4
+  // CHECK:       [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]]
+  // CHECK:       [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
+  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+  // CHECK:       [[TRY]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 10, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([10 x i64], [10 x i64]* [[MAPT4]], i32 0, i32 0), i32 1, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[SR]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX2:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX3:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[SADDR4:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX4:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[SADDR5:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX5:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR5:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[PADDR5:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[SADDR6:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX6:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR6:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[PADDR6:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[SADDR7:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX7:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR7:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[PADDR7:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[SADDR8:%.+]] = getelementptr inbounds [10 x i[[SZ]]], [10 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX8:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR8:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[BP]], i32 0, i32 [[IDX8]]
+  // CHECK-DAG:   [[PADDR8:%.+]] = getelementptr inbounds [10 x i8*], [10 x i8*]* [[P]], i32 0, i32 [[IDX8]]
+
+  // The names below are not necessarily consistent with the names used for the
+  // addresses above as some are repeated.
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+  // CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CBPADDR1:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CPADDR1:%.+]],
+  // CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CBPADDR2:%.+]],
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CPADDR2:%.+]],
+  // CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CBPADDR3:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CPADDR3:%.+]],
+  // CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CBPADDR4:%.+]],
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CPADDR4:%.+]],
+  // CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   store i[[SZ]] 40, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CBPADDR5:%.+]],
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CPADDR5:%.+]],
+  // CHECK-DAG:   [[CBPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   [[CPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   store i[[SZ]] [[BNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CBPADDR6:%.+]],
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CPADDR6:%.+]],
+  // CHECK-DAG:   [[CBPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   [[CPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   store i[[SZ]] 400, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CBPADDR7:%.+]],
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CPADDR7:%.+]],
+  // CHECK-DAG:   [[CBPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   [[CPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   store i[[SZ]] [[CNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CBPADDR8:%.+]],
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8:%.+]],
+  // CHECK-DAG:   [[CBPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target parallel for simd if(target: n>20) schedule(static, a)
+  for (unsigned char it = 'z'; it >= 'a'; it+=-1) {
+    a += 1;
+    b[2] += 1.0;
+    bn[3] += 1.0;
+    c[1][2] += 1.0;
+    cn[1][3] += 1.0;
+    d.X += 1;
+    d.Y += 1;
+  }
+
+  return a;
+}
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions in foo().
+
+// CHECK:       define internal void [[HVT0]]()
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*))
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid.)
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+
+// CHECK:       define internal void [[HVT1]](i[[SZ]] %{{.+}}, i{{32|64}}{{[*]*.*}} %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       alloca i{{32|64}}{{[*]*}}, align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK-64:    [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i32*
+// CHECK-64:    store i32 [[AA]], i32* [[AA_C]], align
+// CHECK-32:    store i32 [[AA]], i32* [[AA_CASTED]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i{{32|64}}{{[*]*}})* [[OMP_OUTLINED1:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]], i{{32|64}}{{[*]*}} %{{.+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED1]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK:       !llvm.mem.parallel_loop_access
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT2]](i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       alloca i[[SZ]], align
+// CHECK:       alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK:       store i16 [[AA]], i16* [[AA_C]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]])* [[OMP_OUTLINED2:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]], i[[SZ]] {{.+}}, i[[SZ]] {{.+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED2]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT3]]
+// Create stack storage and store argument in there.
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[A_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK-64-DAG:[[A:%.+]] = load i32, i32* [[A_CADDR]], align
+// CHECK-32-DAG:[[A:%.+]] = load i32, i32* [[A_ADDR]], align
+// CHECK-64-DAG:[[A_C:%.+]] = bitcast i[[SZ]]* [[A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[A]], i32* [[A_C]], align
+// CHECK-32-DAG:store i32 [[A]], i32* [[A_CASTED]], align
+// CHECK-DAG:   [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK-DAG:   [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[AA]], i16* [[AA_C]], align
+// CHECK-DAG:   [[PARAM1:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CASTED]], align
+// CHECK-DAG:   [[PARAM2:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK-DAG:   call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]])* [[OMP_OUTLINED3:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM1]], i[[SZ]] [[PARAM2]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED3]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT4]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x float]*
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_BN:%.+]] = alloca float*
+// CHECK:       [[LOCAL_C:%.+]] = alloca [5 x [10 x double]]*
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA3:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_CN:%.+]] = alloca double*
+// CHECK:       [[LOCAL_D:%.+]] = alloca [[TT]]*
+// CHECK:       alloca i[[SZ]]
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store [10 x float]* [[ARG_B:%.+]], [10 x float]** [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store float* [[ARG_BN:%.+]], float** [[LOCAL_BN]]
+// CHECK-DAG:   store [5 x [10 x double]]* [[ARG_C:%.+]], [5 x [10 x double]]** [[LOCAL_C]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA3:%.+]], i[[SZ]]* [[LOCAL_VLA3]]
+// CHECK-DAG:   store double* [[ARG_CN:%.+]], double** [[LOCAL_CN]]
+// CHECK-DAG:   store [[TT]]* [[ARG_D:%.+]], [[TT]]** [[LOCAL_D]]
+
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x float]*, [10 x float]** [[LOCAL_B]],
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[REF_BN:%.+]] = load float*, float** [[LOCAL_BN]],
+// CHECK-DAG:   [[REF_C:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[LOCAL_C]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[VAL_VLA3:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA3]],
+// CHECK-DAG:   [[REF_CN:%.+]] = load double*, double** [[LOCAL_CN]],
+// CHECK-DAG:   [[REF_D:%.+]] = load [[TT]]*, [[TT]]** [[LOCAL_D]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 10, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*, i[[SZ]])* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]], i[[SZ]] %{{.+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+template<typename tx>
+tx ftemplate(int n) {
+  tx a = 0;
+  short aa = 0;
+  tx b[10];
+
+  #pragma omp target parallel for simd if(target: n>40)
+  for (long long i = -10; i < 10; i += 3) {
+    a += 1;
+    aa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+static
+int fstatic(int n) {
+  int a = 0;
+  short aa = 0;
+  char aaa = 0;
+  int b[10];
+
+  #pragma omp target parallel for simd if(target: n>50)
+  for (unsigned i=100; i<10; i+=10) {
+    a += 1;
+    aa += 1;
+    aaa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+struct S1 {
+  double a;
+
+  int r1(int n){
+    int b = n+1;
+    short int c[2][n];
+
+    #pragma omp target parallel for simd if(target: n>60)
+    for (unsigned long long it = 2000; it >= 600; it -= 400) {
+      this->a = (double)b + 1.5;
+      c[1][1] = ++a;
+    }
+
+    return c[1][1] + (int)b;
+  }
+};
+
+// CHECK: define {{.*}}@{{.*}}bar{{.*}}
+int bar(int n){
+  int a = 0;
+
+  // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}})
+  a += foo(n);
+
+  S1 S;
+  // CHECK: call {{.*}}i32 [[FS1:@.+]]([[S1]]* {{.*}}, i32 {{.*}})
+  a += S.r1(n);
+
+  // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}})
+  a += fstatic(n);
+
+  // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}})
+  a += ftemplate<int>(n);
+
+  return a;
+}
+
+//
+// CHECK: define {{.*}}[[FS1]]
+//
+// CHECK:          i8* @llvm.stacksave()
+// CHECK-64:       [[B_ADDR:%.+]] = bitcast i[[SZ]]* [[B_CADDR:%.+]] to i32*
+// CHECK-64:       store i32 %{{.+}}, i32* [[B_ADDR]],
+// CHECK-64:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_CADDR]],
+
+// CHECK-32:       store i32 %{{.+}}, i32* [[B_ADDR:%.+]],
+// CHECK-32:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_ADDR]],
+
+// We capture 2 VLA sizes in this target region
+// CHECK:       [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]]
+// CHECK:       [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2
+
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
+// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+// CHECK:       [[TRY]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX0:[0-9]+]]
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX0]]
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX0]]
+// CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX1:[0-9]+]]
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX1]]
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX1]]
+// CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX2:[0-9]+]]
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX2]]
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX2]]
+// CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX3:[0-9]+]]
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX3]]
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX3]]
+
+// The names below are not necessarily consistent with the names used for the
+// addresses above as some are repeated.
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR3:%.+]],
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   store i[[SZ]] 8, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
+
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+
+//
+// CHECK: define {{.*}}[[FSTATIC]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2]],
+
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 3
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 3
+// CHECK-DAG:   [[CBPADDR3:%.+]] = bitcast i8** [[BPADDR3]] to [10 x i32]**
+// CHECK-DAG:   [[CPADDR3:%.+]] = bitcast i8** [[PADDR3]] to [10 x i32]**
+// CHECK-DAG:   store [10 x i32]* [[VAL3:%.+]], [10 x i32]** [[CBPADDR3]],
+// CHECK-DAG:   store [10 x i32]* [[VAL3]], [10 x i32]** [[CPADDR3]],
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+//
+// CHECK: define {{.*}}[[FTEMPLATE]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0), i32 1, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to [10 x i32]**
+// CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to [10 x i32]**
+// CHECK-DAG:   store [10 x i32]* [[VAL2:%.+]], [10 x i32]** [[CBPADDR2]],
+// CHECK-DAG:   store [10 x i32]* [[VAL2]], [10 x i32]** [[CPADDR2]],
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions of the callees of bar().
+
+// CHECK:       define internal void [[HVT7]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_THIS:%.+]] = alloca [[S1]]*
+// CHECK:       [[LOCAL_B:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_C:%.+]] = alloca i16*
+// CHECK:       [[LOCAL_B_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store [[S1]]* [[ARG_THIS:%.+]], [[S1]]** [[LOCAL_THIS]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_B:%.+]], i[[SZ]]* [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i16* [[ARG_C:%.+]], i16** [[LOCAL_C]]
+// Store captures in the context.
+// CHECK-DAG:   [[REF_THIS:%.+]] = load [[S1]]*, [[S1]]** [[LOCAL_THIS]],
+// CHECK-64-DAG:[[CONV_BP:%.+]] = bitcast i[[SZ]]* [[LOCAL_B]] to i32*
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[REF_C:%.+]] = load i16*, i16** [[LOCAL_C]],
+
+// CHECK-64-DAG:[[CONV_B:%.+]] = load i32, i32* [[CONV_BP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_B_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_B]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_BV:%.+]] = load i32, i32* [[LOCAL_B]]
+// CHECK-32-DAG:store i32 [[LOCAL_BV]], i32* [[LOCAL_B_CASTED]], align
+// CHECK-DAG:   [[REF_B:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_B_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+
+// CHECK:       define internal void [[HVT6]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AAA:%.+]], i[[SZ]]* [[LOCAL_AAA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[CONV_AAAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA]] to i8*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK-DAG:   [[CONV_AAA:%.+]] = load i8, i8* [[CONV_AAAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA_CASTED]] to i8*
+// CHECK-DAG:   store i8 [[CONV_AAA]], i8* [[CONV]], align
+// CHECK-DAG:   [[REF_AAA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AAA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 4, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED6:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], i[[SZ]] [[REF_AAA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED6]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+// CHECK:       define internal void [[HVT5]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED7:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED7]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+#endif
diff --git a/test/OpenMP/target_parallel_for_simd_codegen_registration.cpp b/test/OpenMP/target_parallel_for_simd_codegen_registration.cpp
new file mode 100644
index 000000000000..5cc9d7015cbd
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_simd_codegen_registration.cpp
@@ -0,0 +1,451 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target parallel for simd codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// Check that no target code is emmitted if no omptests flag was provided.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK-NTARGET
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK:    [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// CHECK-DAG: [[A1:@.+]] = internal global [[SA]]
+// CHECK-DAG: [[A2:@.+]] = global [[SA]]
+// CHECK-DAG: [[B1:@.+]] = global [[SB]]
+// CHECK-DAG: [[B2:@.+]] = global [[SB]]
+// CHECK-DAG: [[C1:@.+]] = internal global [[SC]]
+// CHECK-DAG: [[D1:@.+]] = global [[SD]]
+// CHECK-DAG: [[E1:@.+]] = global [[SE]]
+// CHECK-DAG: [[T1:@.+]] = global [[ST1]]
+// CHECK-DAG: [[T2:@.+]] = global [[ST2]]
+
+// CHECK-NTARGET-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-NTARGET-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-NTARGET-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-NTARGET-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-NTARGET-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-NTARGET-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-NTARGET-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-NTARGET-NOT: type { i8*, i8*, %
+// CHECK-NTARGET-NOT: type { i32, %
+
+// We have 7 target regions
+
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// TCHECK-NOT: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+
+// CHECK-NTARGET-NOT: private constant i8 0
+// CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
+
+// CHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// CHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// CHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// CHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// CHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// CHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// CHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// CHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// CHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// CHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// CHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// CHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// CHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// TCHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// TCHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// TCHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// TCHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// TCHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// TCHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// TCHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// TCHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// TCHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// TCHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// TCHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// TCHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// TCHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// We have 4 initializers, one for the 500 priority, another one for 501, or more for the default priority, and the last one for the offloading registration function.
+// CHECK: @llvm.global_ctors = appending global [4 x { i32, void ()*, i8* }] [
+// CHECK-SAME: { i32, void ()*, i8* } { i32 500, void ()* [[P500:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 501, void ()* [[P501:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 65535, void ()* [[PMAX:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+// CHECK-NTARGET: @llvm.global_ctors = appending global [3   x { i32, void ()*, i8* }] [
+
+extern int *R;
+
+struct SA {
+  int arr[4];
+  void foo() {
+    int a = *R;
+    a += 1;
+    *R = a;
+  }
+  SA() {
+    int a = *R;
+    a += 2;
+    *R = a;
+  }
+  ~SA() {
+    int a = *R;
+    a += 3;
+    *R = a;
+  }
+};
+
+struct SB {
+  int arr[8];
+  void foo() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 4;
+    *R = a;
+  }
+  SB() {
+    int a = *R;
+    a += 5;
+    *R = a;
+  }
+  ~SB() {
+    int a = *R;
+    a += 6;
+    *R = a;
+  }
+};
+
+struct SC {
+  int arr[16];
+  void foo() {
+    int a = *R;
+    a += 7;
+    *R = a;
+  }
+  SC() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 8;
+    *R = a;
+  }
+  ~SC() {
+    int a = *R;
+    a += 9;
+    *R = a;
+  }
+};
+
+struct SD {
+  int arr[32];
+  void foo() {
+    int a = *R;
+    a += 10;
+    *R = a;
+  }
+  SD() {
+    int a = *R;
+    a += 11;
+    *R = a;
+  }
+  ~SD() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 12;
+    *R = a;
+  }
+};
+
+struct SE {
+  int arr[64];
+  void foo() {
+    int a = *R;
+    #pragma omp target parallel for simd if(target: 0)
+    for (int i = 0; i < 10; ++i)
+      a += 13;
+    *R = a;
+  }
+  SE() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 14;
+    *R = a;
+  }
+  ~SE() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 15;
+    *R = a;
+  }
+};
+
+template <int x>
+struct ST {
+  int arr[128 + x];
+  void foo() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 16 + x;
+    *R = a;
+  }
+  ST() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 17 + x;
+    *R = a;
+  }
+  ~ST() {
+    int a = *R;
+    #pragma omp target parallel for simd
+    for (int i = 0; i < 10; ++i)
+      a += 18 + x;
+    *R = a;
+  }
+};
+
+// We have to make sure we us all the target regions:
+//CHECK-DAG: define internal void @[[NAME1]](
+//CHECK-DAG: call void @[[NAME1]](
+//CHECK-DAG: define internal void @[[NAME2]](
+//CHECK-DAG: call void @[[NAME2]](
+//CHECK-DAG: define internal void @[[NAME3]](
+//CHECK-DAG: call void @[[NAME3]](
+//CHECK-DAG: define internal void @[[NAME4]](
+//CHECK-DAG: call void @[[NAME4]](
+//CHECK-DAG: define internal void @[[NAME5]](
+//CHECK-DAG: call void @[[NAME5]](
+//CHECK-DAG: define internal void @[[NAME6]](
+//CHECK-DAG: call void @[[NAME6]](
+//CHECK-DAG: define internal void @[[NAME7]](
+//CHECK-DAG: call void @[[NAME7]](
+//CHECK-DAG: define internal void @[[NAME8]](
+//CHECK-DAG: call void @[[NAME8]](
+//CHECK-DAG: define internal void @[[NAME9]](
+//CHECK-DAG: call void @[[NAME9]](
+//CHECK-DAG: define internal void @[[NAME10]](
+//CHECK-DAG: call void @[[NAME10]](
+//CHECK-DAG: define internal void @[[NAME11]](
+//CHECK-DAG: call void @[[NAME11]](
+//CHECK-DAG: define internal void @[[NAME12]](
+//CHECK-DAG: call void @[[NAME12]](
+
+//TCHECK-DAG: define void @[[NAME1]](
+//TCHECK-DAG: define void @[[NAME2]](
+//TCHECK-DAG: define void @[[NAME3]](
+//TCHECK-DAG: define void @[[NAME4]](
+//TCHECK-DAG: define void @[[NAME5]](
+//TCHECK-DAG: define void @[[NAME6]](
+//TCHECK-DAG: define void @[[NAME7]](
+//TCHECK-DAG: define void @[[NAME8]](
+//TCHECK-DAG: define void @[[NAME9]](
+//TCHECK-DAG: define void @[[NAME10]](
+//TCHECK-DAG: define void @[[NAME11]](
+//TCHECK-DAG: define void @[[NAME12]](
+
+// CHECK-NTARGET-NOT: __tgt_target
+// CHECK-NTARGET-NOT: __tgt_register_lib
+// CHECK-NTARGET-NOT: __tgt_unregister_lib
+
+// TCHECK-NOT: __tgt_target
+// TCHECK-NOT: __tgt_register_lib
+// TCHECK-NOT: __tgt_unregister_lib
+
+// We have 2 initializers with priority 500
+//CHECK: define internal void [[P500]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 1 initializers with priority 501
+//CHECK: define internal void [[P501]](
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 6 initializers with default priority
+//CHECK: define internal void [[PMAX]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// Check registration and unregistration
+
+//CHECK:     define internal void @[[UNREGFN:.+]](i8*)
+//CHECK-SAME: comdat($[[REGFN]]) {
+//CHECK:     call i32 @__tgt_unregister_lib([[DSCTY]]* [[DESC]])
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_unregister_lib([[DSCTY]]*)
+
+//CHECK:     define linkonce hidden void @[[REGFN]](i8*)
+//CHECK-SAME: comdat {
+//CHECK:     call i32 @__tgt_register_lib([[DSCTY]]* [[DESC]])
+//CHECK:     call i32 @__cxa_atexit(void (i8*)* @[[UNREGFN]], i8* bitcast ([[DSCTY]]* [[DESC]] to i8*),
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_register_lib([[DSCTY]]*)
+
+static __attribute__((init_priority(500))) SA a1;
+SA a2;
+SB __attribute__((init_priority(500))) b1;
+SB __attribute__((init_priority(501))) b2;
+static SC c1;
+SD d1;
+SE e1;
+ST<100> t1;
+ST<1000> t2;
+
+
+int bar(int a){
+  int r = a;
+
+  a1.foo();
+  a2.foo();
+  b1.foo();
+  b2.foo();
+  c1.foo();
+  d1.foo();
+  e1.foo();
+  t1.foo();
+  t2.foo();
+
+  #pragma omp target parallel for simd
+  for (int i = 0; i < 10; ++i)
+    ++r;
+
+  return r + *R;
+}
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+#endif
diff --git a/test/OpenMP/target_parallel_for_simd_codegen_registration_naming.cpp b/test/OpenMP/target_parallel_for_simd_codegen_registration_naming.cpp
new file mode 100644
index 000000000000..bdfa38669859
--- /dev/null
+++ b/test/OpenMP/target_parallel_for_simd_codegen_registration_naming.cpp
@@ -0,0 +1,68 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target parallel for simd codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[CA:%.+]] = type { i32* }
+
+// CHECK: define {{.*}}i32 @[[NNAME:.+]](i32 {{.*}}%{{.+}})
+int nested(int a){
+  // CHECK: call void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME]]_l[[T1L:[0-9]+]](
+  #pragma omp target parallel for simd
+  for (int i = 0; i < 10; ++i)
+    ++a;
+
+  // CHECK: call void @"[[LNAME:.+]]"([[CA]]*
+  auto F = [&](){
+    #pragma omp parallel
+    {
+      #pragma omp target parallel for simd
+      for (int i = 0; i < 10; ++i)
+        ++a;
+    }
+  };
+
+  F();
+
+  return a;
+}
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T1L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME:.+]]_l[[T1L:[0-9]+]](
+
+// CHECK: define {{.*}}void @"[[LNAME]]"(
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.+}}[[PNAME:@.+]] to
+
+// CHECK: define {{.*}}void [[PNAME]](
+// CHECK: call void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L:[0-9]+]](
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME:.+]]_l[[T2L:[0-9]+]](
+
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+#endif
diff --git a/test/OpenMP/target_parallel_for_simd_depend_messages.cpp b/test/OpenMP/target_parallel_for_simd_depend_messages.cpp
index a8b4de7ff755..4375941036cf 100644
--- a/test/OpenMP/target_parallel_for_simd_depend_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for simd depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target parallel for simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for simd depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for simd depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel for simd depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for simd depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel for simd depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for simd depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for simd depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for simd depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target parallel for simd depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target parallel for simd depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target parallel for simd depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target parallel for simd depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_parallel_for_simd_firstprivate_messages.cpp b/test/OpenMP/target_parallel_for_simd_firstprivate_messages.cpp
index 9397314d87cc..9ad255f033e4 100644
--- a/test/OpenMP/target_parallel_for_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_firstprivate_messages.cpp
@@ -125,11 +125,11 @@ int foomain(int argc, char **argv) {
     foo();
 #pragma omp parallel private(i)
 #pragma omp target parallel for simd firstprivate(i) // expected-note {{defined as firstprivate}}
-  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as private}}
+  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as linear}}
     foo();
 #pragma omp parallel reduction(+ : i)
 #pragma omp target parallel for simd firstprivate(i) // expected-note {{defined as firstprivate}}
-  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as private}}
+  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as linear}}
     foo();
   return 0;
 }
@@ -219,7 +219,7 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i)
     foo();
 #pragma omp target parallel for simd firstprivate(i) // expected-note {{defined as firstprivate}}
-  for (i = 0; i < argc; ++i)    // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as private}}
+  for (i = 0; i < argc; ++i)    // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as linear}}
     foo();
 #pragma omp parallel shared(xa)
 #pragma omp target parallel for simd firstprivate(xa) // OK: may be firstprivate
@@ -246,11 +246,11 @@ int main(int argc, char **argv) {
   }
 #pragma omp parallel private(i)
 #pragma omp target parallel for simd firstprivate(i) // expected-note {{defined as firstprivate}}
-  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as private}}
+  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as linear}}
     foo();
 #pragma omp parallel reduction(+ : i)
 #pragma omp target parallel for simd firstprivate(i) // expected-note {{defined as firstprivate}}
-  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as private}}
+  for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as linear}}
     foo();
   static int si;
 #pragma omp target parallel for simd firstprivate(si) // OK
diff --git a/test/OpenMP/target_parallel_for_simd_lastprivate_messages.cpp b/test/OpenMP/target_parallel_for_simd_lastprivate_messages.cpp
index 51fc72464c9a..313192dbb5f6 100644
--- a/test/OpenMP/target_parallel_for_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   S2 &operator=(const S2 &);
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
@@ -209,6 +209,8 @@ int main(int argc, char **argv) {
 #pragma omp target parallel for simd private(xa), lastprivate(xa) // expected-error {{private variable cannot be lastprivate}} expected-note {{defined as private}}
   for (i = 0; i < argc; ++i)
     foo();
+// expected-note@+2 {{defined as lastprivate}}
+// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be lastprivate, predetermined as linear}}
 #pragma omp target parallel for simd lastprivate(i)
   for (i = 0; i < argc; ++i)
     foo();
diff --git a/test/OpenMP/target_parallel_for_simd_loop_messages.cpp b/test/OpenMP/target_parallel_for_simd_loop_messages.cpp
index ed184a5e4124..f07f19e2a690 100644
--- a/test/OpenMP/target_parallel_for_simd_loop_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_loop_messages.cpp
@@ -238,27 +238,29 @@ int test_iteration_spaces() {
     c[ii] = a[ii];
 
 // expected-note@+2  {{defined as firstprivate}}
-// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as private}}
+// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be firstprivate, predetermined as linear}}
 #pragma omp target parallel for simd firstprivate(ii)
   for (ii = 0; ii < 10; ii++)
     c[ii] = a[ii];
 
-// expected-note@+2  {{defined as linear}}
-// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be linear, predetermined as private}}
 #pragma omp target parallel for simd linear(ii)
   for (ii = 0; ii < 10; ii++)
     c[ii] = a[ii];
 
+// expected-note@+2  {{defined as private}}
+// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be private, predetermined as linear}}
 #pragma omp target parallel for simd private(ii)
   for (ii = 0; ii < 10; ii++)
     c[ii] = a[ii];
 
+// expected-note@+2  {{defined as lastprivate}}
+// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be lastprivate, predetermined as linear}}
 #pragma omp target parallel for simd lastprivate(ii)
   for (ii = 0; ii < 10; ii++)
     c[ii] = a[ii];
 
   {
-// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be threadprivate or thread local, predetermined as private}}
+// expected-error@+2 {{loop iteration variable in the associated loop of 'omp target parallel for simd' directive may not be threadprivate or thread local, predetermined as linear}}
 #pragma omp target parallel for simd
     for (sii = 0; sii < 10; sii += 1)
       c[sii] = a[sii];
@@ -585,15 +587,15 @@ void test_loop_eh() {
 #pragma omp target parallel for simd
   for (int i = 0; i < 10; i++) {
     c[i] = a[i] + b[i];
-    try {
+    try { // expected-error {{'try' statement cannot be used in OpenMP simd region}}
       for (int j = 0; j < 10; ++j) {
         if (a[i] > b[j])
-          throw a[i];
+          throw a[i]; // expected-error {{'throw' statement cannot be used in OpenMP simd region}}
       }
-      throw a[i];
+      throw a[i]; // expected-error {{'throw' statement cannot be used in OpenMP simd region}}
     } catch (float f) {
       if (f > 0.1)
-        throw a[i];
+        throw a[i]; // expected-error {{'throw' statement cannot be used in OpenMP simd region}}
       return; // expected-error {{cannot return from OpenMP region}}
     }
     switch (i) {
@@ -605,7 +607,7 @@ void test_loop_eh() {
     }
     for (int j = 0; j < 10; j++) {
       if (c[i] > 10)
-        throw c[i];
+        throw c[i]; // expected-error {{'throw' statement cannot be used in OpenMP simd region}}
     }
   }
   if (c[9] > 10)
diff --git a/test/OpenMP/target_parallel_for_simd_map_messages.cpp b/test/OpenMP/target_parallel_for_simd_map_messages.cpp
index 195b39c595c3..7e95e10ae5ca 100644
--- a/test/OpenMP/target_parallel_for_simd_map_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for simd map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -222,11 +222,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for simd map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target parallel for simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel for simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target parallel for simd map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_parallel_for_simd_misc_messages.c b/test/OpenMP/target_parallel_for_simd_misc_messages.c
index 2adc6f85116e..f2574a35472d 100644
--- a/test/OpenMP/target_parallel_for_simd_misc_messages.c
+++ b/test/OpenMP/target_parallel_for_simd_misc_messages.c
@@ -166,7 +166,7 @@ void test_collapse() {
 // expected-note@+1 {{variable with automatic storage duration is predetermined as private; perhaps you forget to enclose 'omp for' directive into a parallel or another task region?}}
     for (int j = 0; j < 16; ++j)
 // expected-error@+2 2 {{reduction variable must be shared}}
-// expected-error@+1 {{region cannot be closely nested inside 'target parallel for simd' region; perhaps you forget to enclose 'omp for' directive into a parallel region?}}
+// expected-error@+1 {{OpenMP constructs may not be nested inside a simd region}}
 #pragma omp for reduction(+ : i, j)
       for (int k = 0; k < 16; ++k)
         i += j;
diff --git a/test/OpenMP/target_parallel_for_simd_ordered_messages.cpp b/test/OpenMP/target_parallel_for_simd_ordered_messages.cpp
index 70a3b4ef6245..065d81517f83 100644
--- a/test/OpenMP/target_parallel_for_simd_ordered_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_ordered_messages.cpp
@@ -6,7 +6,7 @@ void foo() {
 }
 
 #if __cplusplus >= 201103L
- // expected-note@+2 4 {{declared here}}
+ // expected-note@+2 2 {{declared here}}
 #endif
 bool foobool(int argc) {
   return argc;
@@ -15,7 +15,7 @@ bool foobool(int argc) {
 struct S1; // expected-note {{declared here}}
 
 template <class T, typename S, int N, int ST> // expected-note {{declared here}}
-T tmain(T argc, S **argv) {                   //expected-note 2 {{declared here}}
+T tmain(T argc, S **argv) {
   int j; // expected-note {{declared here}}
 #pragma omp target parallel for simd ordered
   for (int i = ST; i < N; i++)
@@ -26,28 +26,26 @@ T tmain(T argc, S **argv) {                   //expected-note 2 {{declared here}
 #pragma omp target parallel for simd ordered() // expected-error {{expected expression}}
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
-// expected-error@+3 {{expected ')'}} expected-note@+3 {{to match this '('}}
-// expected-error@+2 2 {{expression is not an integral constant expression}}
-// expected-note@+1 2 {{read of non-const variable 'argc' is not allowed in a constant expression}}
+// expected-error@+2 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+// expected-error@+1 {{expected ')'}} expected-note@+1 {{to match this '('}}
 #pragma omp target parallel for simd ordered(argc
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
-// expected-error@+1 2 {{argument to 'ordered' clause must be a strictly positive integer value}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
 #pragma omp target parallel for simd ordered(ST // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
 #pragma omp target parallel for simd ordered(1)) // expected-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}}
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
-#pragma omp target parallel for simd ordered((ST > 0) ? 1 + ST : 2) // expected-note 2 {{as specified in 'ordered' clause}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+#pragma omp target parallel for simd ordered((ST > 0) ? 1 + ST : 2)
   for (int i = ST; i < N; i++)
-    argv[0][i] = argv[0][i] - argv[0][i - ST]; // expected-error 2 {{expected 2 for loops after '#pragma omp target parallel for simd', but found only 1}}
-#if __cplusplus >= 201103L
-  // expected-note@+5 2 {{non-constexpr function 'foobool' cannot be used in a constant expression}}
-#endif
-// expected-error@+3 2 {{directive '#pragma omp target parallel for simd' cannot contain more than one 'ordered' clause}}
-// expected-error@+2 2 {{argument to 'ordered' clause must be a strictly positive integer value}}
-// expected-error@+1 2 {{expression is not an integral constant expression}}
+    argv[0][i] = argv[0][i] - argv[0][i - ST];
+// expected-error@+3 2 {{argument to 'ordered' clause must be a strictly positive integer value}}
+// expected-error@+2 2 {{directive '#pragma omp target parallel for simd' cannot contain more than one 'ordered' clause}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
 #pragma omp target parallel for simd ordered(foobool(argc)), ordered(true), ordered(-5)
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
@@ -59,14 +57,17 @@ T tmain(T argc, S **argv) {                   //expected-note 2 {{declared here}
 #pragma omp target parallel for simd ordered(j = 2) // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
 #pragma omp target parallel for simd ordered(1)
   for (int i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
-#pragma omp target parallel for simd ordered(N) // expected-error {{argument to 'ordered' clause must be a strictly positive integer value}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+#pragma omp target parallel for simd ordered(N)
   for (T i = ST; i < N; i++)
     argv[0][i] = argv[0][i] - argv[0][i - ST];
-#pragma omp target parallel for simd ordered(2) // expected-note {{as specified in 'ordered' clause}}
-  foo();                            // expected-error {{expected 2 for loops after '#pragma omp target parallel for simd'}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+#pragma omp target parallel for simd ordered(2)
+  foo();
   return argc;
 }
 
@@ -81,12 +82,14 @@ int main(int argc, char **argv) {
 #pragma omp target parallel for simd ordered() // expected-error {{expected expression}}
   for (int i = 4; i < 12; i++)
     argv[0][i] = argv[0][i] - argv[0][i - 4];
-#pragma omp target parallel for simd ordered(4 // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-note {{as specified in 'ordered' clause}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+#pragma omp target parallel for simd ordered(4 // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int i = 4; i < 12; i++)
-    argv[0][i] = argv[0][i] - argv[0][i - 4]; // expected-error {{expected 4 for loops after '#pragma omp target parallel for simd', but found only 1}}
-#pragma omp target parallel for simd ordered(2 + 2))      // expected-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}}  expected-note {{as specified in 'ordered' clause}}
+    argv[0][i] = argv[0][i] - argv[0][i - 4];
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+#pragma omp target parallel for simd ordered(2 + 2))      // expected-warning {{extra tokens at the end of '#pragma omp target parallel for simd' are ignored}}
   for (int i = 4; i < 12; i++)
-    argv[0][i] = argv[0][i] - argv[0][i - 4];            // expected-error {{expected 4 for loops after '#pragma omp target parallel for simd', but found only 1}}
+    argv[0][i] = argv[0][i] - argv[0][i - 4];
 #if __cplusplus >= 201103L
   // expected-note@+2 {{non-constexpr function 'foobool' cannot be used in a constant expression}}
 #endif
@@ -110,13 +113,12 @@ int main(int argc, char **argv) {
 #pragma omp target parallel for simd ordered(j = 2) // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int i = 4; i < 12; i++)
     argv[0][i] = argv[0][i] - argv[0][i - 4];
-// expected-error@+3 {{statement after '#pragma omp target parallel for simd' must be a for loop}}
-// expected-note@+1 {{in instantiation of function template specialization 'tmain<int, char, -1, -2>' requested here}}
+// expected-error@+2 {{statement after '#pragma omp target parallel for simd' must be a for loop}}
 #pragma omp target parallel for simd ordered(ordered(tmain < int, char, -1, -2 > (argc, argv) // expected-error 2 {{expected ')'}} expected-note 2 {{to match this '('}}
   foo();
-#pragma omp target parallel for simd ordered(2) // expected-note {{as specified in 'ordered' clause}}
-  foo();                            // expected-error {{expected 2 for loops after '#pragma omp target parallel for simd'}}
-  // expected-note@+1 {{in instantiation of function template specialization 'tmain<int, char, 1, 0>' requested here}}
+// expected-error@+1 {{'ordered' clause with a parameter can not be specified in '#pragma omp target parallel for simd' directive}}
+#pragma omp target parallel for simd ordered(2)
+  foo();
   return tmain<int, char, 1, 0>(argc, argv);
 }
 
diff --git a/test/OpenMP/target_parallel_for_simd_reduction_messages.cpp b/test/OpenMP/target_parallel_for_simd_reduction_messages.cpp
index 289b5b2641b6..72e2130f3df4 100644
--- a/test/OpenMP/target_parallel_for_simd_reduction_messages.cpp
+++ b/test/OpenMP/target_parallel_for_simd_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_parallel_if_codegen.cpp b/test/OpenMP/target_parallel_if_codegen.cpp
index 0aebcc218fbb..c155bf7c051a 100644
--- a/test/OpenMP/target_parallel_if_codegen.cpp
+++ b/test/OpenMP/target_parallel_if_codegen.cpp
@@ -130,8 +130,6 @@ int bar(int n){
   return a;
 }
 
-
-
 //
 // CHECK: define {{.*}}[[FS1]]([[S1]]* {{%.+}}, i32 {{[^%]*}}[[PARM:%.+]])
 //
@@ -147,10 +145,8 @@ int bar(int n){
 // CHECK:       store i8 [[FB]], i8* [[CONV]], align
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 1, i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 1, i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -175,28 +171,18 @@ int bar(int n){
 // CHECK:       br i1 [[CMP]], label {{%?}}[[IF_THEN:.+]], label {{%?}}[[IF_ELSE:.+]]
 //
 // CHECK:       [[IF_THEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, {{.*}}, i32 1, i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       br label {{%?}}[[END:.+]]
-//
-// CHECK:       [[IF_ELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align
-// CHECK:       br label {{%?}}[[END]]
-//
-// CHECK:       [[END]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
-// CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
-//
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, {{.*}}, i32 1, i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT2:@.+]]([[S1]]* {{%.+}}, i[[SZ]] [[ARG]])
-// CHECK:       br label {{%?}}[[END]]
+// CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
-
-
-
-
-
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IF_ELSE]]
+// CHECK:       call void [[HVT2]]([[S1]]* {{%.+}}, i[[SZ]] [[ARG]])
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 
 //
 // CHECK: define {{.*}}[[FSTATIC]](i32 {{[^%]*}}[[PARM:%.+]])
@@ -217,23 +203,18 @@ int bar(int n){
 // CHECK:       br i1 [[TB]], label {{%?}}[[IF_THEN:.+]], label {{%?}}[[IF_ELSE:.+]]
 //
 // CHECK:       [[IF_THEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       br label {{%?}}[[END:.+]]
-//
-// CHECK:       [[IF_ELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align
-// CHECK:       br label {{%?}}[[END]]
-//
-// CHECK:       [[END]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
-// CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
-//
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT3:@.+]](i[[SZ]] [[ARG]])
-// CHECK:       br label {{%?}}[[END]]
+// CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IF_ELSE]]
+// CHECK:       call void [[HVT3]](i[[SZ]] [[ARG]])
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 //
 //
 //
@@ -243,23 +224,18 @@ int bar(int n){
 // CHECK:       br i1 [[CMP]], label {{%?}}[[IF_THEN:.+]], label {{%?}}[[IF_ELSE:.+]]
 //
 // CHECK:       [[IF_THEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 1, i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       br label {{%?}}[[END:.+]]
-//
-// CHECK:       [[IF_ELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align
-// CHECK:       br label {{%?}}[[END]]
-//
-// CHECK:       [[END]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
-// CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
-//
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 1, i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT4:@.+]]()
-// CHECK:       br label {{%?}}[[END]]
+// CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IF_ELSE]]
+// CHECK:       call void [[HVT4]]()
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 
 
 
@@ -269,10 +245,8 @@ int bar(int n){
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 0)
-// CHECK-NEXT:  store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK-NEXT:  [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 0)
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -283,10 +257,8 @@ int bar(int n){
 //
 //
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, {{.*}}, i32 1, i32 0)
-// CHECK-NEXT:  store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK-NEXT:  [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, {{.*}}, i32 1, i32 0)
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
diff --git a/test/OpenMP/target_parallel_map_messages.cpp b/test/OpenMP/target_parallel_map_messages.cpp
index ca794cecf4c5..d3cc742decd5 100644
--- a/test/OpenMP/target_parallel_map_messages.cpp
+++ b/test/OpenMP/target_parallel_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   foo();
 #pragma omp target parallel map(S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-#pragma omp target parallel map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   foo();
-#pragma omp target parallel map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel map(ba)
   foo();
 #pragma omp target parallel map(ca)
   foo();
@@ -221,11 +221,11 @@ int main(int argc, char **argv) {
   foo();
 #pragma omp target parallel map(S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-#pragma omp target parallel map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   foo();
 #pragma omp target parallel map(argv[1])
   foo();
-#pragma omp target parallel map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target parallel map(ba)
   foo();
 #pragma omp target parallel map(ca)
   foo();
diff --git a/test/OpenMP/target_parallel_no_exceptions.cpp b/test/OpenMP/target_parallel_no_exceptions.cpp
new file mode 100644
index 000000000000..95189a3ade2b
--- /dev/null
+++ b/test/OpenMP/target_parallel_no_exceptions.cpp
@@ -0,0 +1,18 @@
+/// Make sure no exception messages are inclided in the llvm output.
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix CHK-EXCEPTION
+
+void test_increment() {
+#pragma omp target
+  {
+    []() { return; }();
+  }
+}
+
+int main() {
+  test_increment();
+  return 0;
+}
+
+//CHK-EXCEPTION-NOT: invoke
+
diff --git a/test/OpenMP/target_parallel_num_threads_codegen.cpp b/test/OpenMP/target_parallel_num_threads_codegen.cpp
index 225c2dd1d06a..836a92d9a5db 100644
--- a/test/OpenMP/target_parallel_num_threads_codegen.cpp
+++ b/test/OpenMP/target_parallel_num_threads_codegen.cpp
@@ -147,10 +147,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[THREADS:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 1, i32 [[THREADS]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 1, i32 [[THREADS]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -160,10 +158,8 @@ int bar(int n){
 //
 //
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.+}}, i32 1, i32 1024)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.+}}, i32 1, i32 1024)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -190,10 +186,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[THREADS:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 [[THREADS]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 [[THREADS]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -213,10 +207,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[THREADS:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 [[THREADS]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 1, i32 [[THREADS]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -233,10 +225,8 @@ int bar(int n){
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 1, i32 20)
-// CHECK-NEXT:  store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK-NEXT:  [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 1, i32 20)
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -257,10 +247,8 @@ int bar(int n){
 // CHECK:       [[T:%.+]] = load i16, i16* [[CAPE_ADDR]], align
 // CHECK:       [[THREADS:%.+]] = sext i16 [[T]] to i32
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 1, i32 [[THREADS]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 1, i32 [[THREADS]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
diff --git a/test/OpenMP/target_parallel_reduction_messages.cpp b/test/OpenMP/target_parallel_reduction_messages.cpp
index 52338ee71cbe..038f099c5a6c 100644
--- a/test/OpenMP/target_parallel_reduction_messages.cpp
+++ b/test/OpenMP/target_parallel_reduction_messages.cpp
@@ -24,9 +24,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_reduction_codegen.cpp b/test/OpenMP/target_reduction_codegen.cpp
new file mode 100644
index 000000000000..f86719b55fa1
--- /dev/null
+++ b/test/OpenMP/target_reduction_codegen.cpp
@@ -0,0 +1,215 @@
+// Only test codegen on target side, as private clause does not require any action on the host side
+// Test target codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template<typename tx, typename ty>
+struct TT{
+  tx X;
+  ty Y;
+  TT<tx, ty> operator*(const TT<tx, ty> &) { return *this; }
+};
+
+// TCHECK: [[S1:%.+]] = type { double }
+
+int foo(int n) {
+  int a = 0;
+  short aa = 0;
+  float b[10];
+  float bn[n];
+  double c[5][10];
+  double cn[5][n];
+  TT<long long, char> d;
+
+  #pragma omp target reduction(*:a)
+  {
+  }
+
+  // TCHECK: define void @__omp_offloading_{{.+}}(i32*{{.+}} %{{.+}})
+  // TCHECK: [[A:%.+]] = alloca i{{[0-9]+}}*,
+  // TCHECK: store {{.+}}, {{.+}} [[A]],
+  // TCHECK: load i32*, i32** [[A]],
+  // TCHECK: ret void
+
+#pragma omp target reduction(+:a)
+  {
+    a = 1;
+  }
+
+  // TCHECK:  define void @__omp_offloading_{{.+}}(i32*{{.+}} %{{.+}})
+  // TCHECK: [[A:%.+]] = alloca i{{[0-9]+}}*,
+  // TCHECK: store {{.+}}, {{.+}} [[A]],
+  // TCHECK: [[REF:%.+]] = load i32*, i32** [[A]],
+  // TCHECK: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[REF]],
+  // TCHECK: ret void
+
+  #pragma omp target reduction(-:a, aa)
+  {
+    a = 1;
+    aa = 1;
+  }
+
+  // TCHECK:  define void @__omp_offloading_{{.+}}(i32*{{.+}} [[A:%.+]], i16*{{.+}} [[AA:%.+]])
+  // TCHECK:  [[A:%.+]] = alloca i{{[0-9]+}}*,
+  // TCHECK:  [[AA:%.+]] = alloca i{{[0-9]+}}*,
+  // TCHECK: store {{.+}}, {{.+}} [[A]],
+  // TCHECK: store {{.+}}, {{.+}} [[AA]],
+  // TCHECK: [[A_REF:%.+]] = load i32*, i32** [[A]],
+  // TCHECK: [[AA_REF:%.+]] = load i16*, i16** [[AA]],
+  // TCHECK:  store i{{[0-9]+}} 1, i{{[0-9]+}}* [[A_REF]],
+  // TCHECK:  store i{{[0-9]+}} 1, i{{[0-9]+}}* [[AA_REF]],
+  // TCHECK:  ret void
+
+  return a;
+}
+
+
+template<typename tx>
+tx ftemplate(int n) {
+  tx a = 0;
+  short aa = 0;
+  tx b[10];
+
+#pragma omp target reduction(+:a,aa,b)
+  {
+    a = 1;
+    aa = 1;
+    b[2] = 1;
+  }
+
+  return a;
+}
+
+static
+int fstatic(int n) {
+  int a = 0;
+  short aa = 0;
+  char aaa = 0;
+  int b[10];
+
+#pragma omp target reduction(-:a,aa,aaa,b)
+  {
+    a = 1;
+    aa = 1;
+    aaa = 1;
+    b[2] = 1;
+  }
+
+  return a;
+}
+
+// TCHECK: define void @__omp_offloading_{{.+}}(i32*{{.+}}, i16*{{.+}}, i8*{{.+}}, [10 x i32]*{{.+}})
+// TCHECK:  [[A:%.+]] = alloca i{{[0-9]+}}*,
+// TCHECK:  [[A2:%.+]] = alloca i{{[0-9]+}}*,
+// TCHECK:  [[A3:%.+]] = alloca i{{[0-9]+}}*,
+// TCHECK:  [[B:%.+]] = alloca [10 x i{{[0-9]+}}]*,
+// TCHECK: store {{.+}}, {{.+}} [[A]],
+// TCHECK: store {{.+}}, {{.+}} [[A2]],
+// TCHECK: store {{.+}}, {{.+}} [[A3]],
+// TCHECK: store {{.+}}, {{.+}} [[B]],
+// TCHECK: [[A_REF:%.+]] = load i32*, i32** [[A]],
+// TCHECK: [[AA_REF:%.+]] = load i16*, i16** [[AA]],
+// TCHECK: [[A3_REF:%.+]] = load i8*, i8** [[A3]],
+// TCHECK: [[B_REF:%.+]] = load {{.+}}*, {{.+}}** [[B]],
+// TCHECK:  store i{{[0-9]+}} 1, i{{[0-9]+}}* [[A_REF]],
+// TCHECK:  store i{{[0-9]+}} 1, i{{[0-9]+}}* [[AA_REF]],
+// TCHECK:  store i{{[0-9]+}} 1, i{{[0-9]+}}* [[A3_REF]],
+// TCHECK:  [[B_GEP:%.+]] = getelementptr inbounds [10 x i{{[0-9]+}}], [10 x i{{[0-9]+}}]* [[B_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+// TCHECK:  store i{{[0-9]+}} 1, i{{[0-9]+}}* [[B_GEP]],
+// TCHECK:  ret void
+
+struct S1 {
+  double a;
+
+  int r1(int n){
+    int b = n+1;
+    short int c[2][n];
+
+#pragma omp target reduction(max:b,c)
+    {
+      this->a = (double)b + 1.5;
+      c[1][1] = ++a;
+    }
+
+    return c[1][1] + (int)b;
+  }
+
+  // TCHECK: define void @__omp_offloading_{{.+}}([[S1]]* [[TH:%.+]], i32*{{.+}}, i{{[0-9]+}} [[VLA:%.+]], i{{[0-9]+}} [[VLA1:%.+]], i16*{{.+}})
+  // TCHECK: [[TH_ADDR:%.+]] = alloca [[S1]]*,
+  // TCHECK: [[B_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+  // TCHECK: [[VLA_ADDR:%.+]] = alloca i{{[0-9]+}},
+  // TCHECK: [[VLA_ADDR2:%.+]] = alloca i{{[0-9]+}},
+  // TCHECK: [[C_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+  // TCHECK: store [[S1]]* [[TH]], [[S1]]** [[TH_ADDR]],
+  // TCHECK: store i{{[0-9]+}}* {{.+}}, i{{[0-9]+}}** [[B_ADDR]],
+  // TCHECK: store i{{[0-9]+}} [[VLA]], i{{[0-9]+}}* [[VLA_ADDR]],
+  // TCHECK: store i{{[0-9]+}} [[VLA1]], i{{[0-9]+}}* [[VLA_ADDR2]],
+  // TCHECK: store i{{[0-9]+}}* {{.+}}, i{{[0-9]+}}** [[C_ADDR]],
+  // TCHECK: [[TH_ADDR_REF:%.+]] = load [[S1]]*, [[S1]]** [[TH_ADDR]],
+  // TCHECK: [[B_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[B_ADDR]],
+  // TCHECK: [[VLA_ADDR_REF:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[VLA_ADDR]],
+  // TCHECK: [[VLA_ADDR_REF2:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[VLA_ADDR2]],
+  // TCHECK: [[C_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[C_ADDR]],
+
+  // this->a = (double)b + 1.5;
+  // TCHECK: [[B_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[B_REF]],
+  // TCHECK: [[B_CONV:%.+]] = sitofp i{{[0-9]+}} [[B_VAL]] to double
+  // TCHECK: [[NEW_A_VAL:%.+]] = fadd double [[B_CONV]], 1.5{{.+}}+00
+  // TCHECK: [[A_FIELD:%.+]] = getelementptr inbounds [[S1]], [[S1]]* [[TH_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+  // TCHECK: store double [[NEW_A_VAL]], double* [[A_FIELD]],
+
+  // c[1][1] = ++a;
+  // TCHECK: [[A_FIELD4:%.+]] = getelementptr inbounds [[S1]], [[S1]]* [[TH_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+  // TCHECK: [[A_FIELD4_VAL:%.+]] = load double, double* [[A_FIELD4]],
+  // TCHECK: [[A_FIELD_INC:%.+]] = fadd double [[A_FIELD4_VAL]], 1.0{{.+}}+00
+  // TCHECK: store double [[A_FIELD_INC]], double* [[A_FIELD4]],
+  // TCHECK: [[A_FIELD_INC_CONV:%.+]] = fptosi double [[A_FIELD_INC]] to i{{[0-9]+}}
+  // TCHECK: [[C_IND:%.+]] = mul{{.+}} i{{[0-9]+}} 1, [[VLA_ADDR_REF2]]
+  // TCHECK: [[C_1_REF:%.+]] = getelementptr inbounds i{{[0-9]+}}, i{{[0-9]+}}* [[C_REF]], i{{[0-9]+}} [[C_IND]]
+  // TCHECK: [[C_1_1_REF:%.+]] = getelementptr inbounds i{{[0-9]+}}, i{{[0-9]+}}* [[C_1_REF]], i{{[0-9]+}} 1
+  // TCHECK: store i{{[0-9]+}} [[A_FIELD_INC_CONV]], i{{[0-9]+}}* [[C_1_1_REF]],
+
+  // finish
+  // TCHECK: ret void
+};
+
+
+int bar(int n){
+  int a = 0;
+  a += foo(n);
+  S1 S;
+  a += S.r1(n);
+  a += fstatic(n);
+  a += ftemplate<int>(n);
+
+  return a;
+}
+
+// template
+// TCHECK: define void @__omp_offloading_{{.+}}(i{{[0-9]+}}*{{.+}}, i{{[0-9]+}}*{{.+}}, [10 x i32]*{{.+}})
+// TCHECK: [[A:%.+]] = alloca i{{[0-9]+}}*,
+// TCHECK: [[A2:%.+]] = alloca i{{[0-9]+}}*,
+// TCHECK: [[B:%.+]] = alloca [10 x i{{[0-9]+}}]*,
+// TCHECK: store {{.+}}, {{.+}} [[A]],
+// TCHECK: store {{.+}}, {{.+}} [[A2]],
+// TCHECK: store {{.+}}, {{.+}} [[B]],
+// TCHECK: [[A_REF:%.+]] = load i32*, i32** [[A]],
+// TCHECK: [[AA_REF:%.+]] = load i16*, i16** [[AA]],
+// TCHECK: [[B_REF:%.+]] = load {{.+}}*, {{.+}}** [[B]],
+// TCHECK: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[A_REF]],
+// TCHECK: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[AA_REF]],
+// TCHECK: [[B_GEP:%.+]] = getelementptr inbounds [10 x i{{[0-9]+}}], [10 x i{{[0-9]+}}]* [[B_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+// TCHECK: store i{{[0-9]+}} 1, i{{[0-9]+}}* [[B_GEP]],
+// TCHECK: ret void
+
+#endif
diff --git a/test/OpenMP/target_reduction_messages.cpp b/test/OpenMP/target_reduction_messages.cpp
new file mode 100644
index 000000000000..8cf79f9a5d99
--- /dev/null
+++ b/test/OpenMP/target_reduction_messages.cpp
@@ -0,0 +1,262 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++98 -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++11 -ferror-limit 150 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+void foobar(int &ref) {
+#pragma omp target reduction(+:ref)
+  foo();
+}
+
+struct S1; // expected-note {{declared here}} expected-note 4 {{forward declaration of 'S1'}}
+extern S1 a;
+class S2 {
+  mutable int a;
+  S2 &operator+(const S2 &arg) { return (*this); } // expected-note 3 {{implicitly declared private here}}
+
+public:
+  S2() : a(0) {}
+  S2(S2 &s2) : a(s2.a) {}
+  static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
+};
+const float S2::S2sc = 0;
+S2 b;                     // expected-note 3 {{'b' defined here}}
+const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
+class S3 {
+  int a;
+
+public:
+  int b;
+  S3() : a(0) {}
+  S3(const S3 &s3) : a(s3.a) {}
+  S3 operator+(const S3 &arg1) { return arg1; }
+};
+int operator+(const S3 &arg1, const S3 &arg2) { return 5; }
+S3 c;               // expected-note 3 {{'c' defined here}}
+const S3 ca[5];     // expected-note 2 {{'ca' defined here}}
+extern const int f; // expected-note 4 {{'f' declared here}}
+class S4 {
+  int a;
+  S4(); // expected-note {{implicitly declared private here}}
+  S4(const S4 &s4);
+  S4 &operator+(const S4 &arg) { return (*this); }
+
+public:
+  S4(int v) : a(v) {}
+};
+S4 &operator&=(S4 &arg1, S4 &arg2) { return arg1; }
+class S5 {
+  int a;
+  S5() : a(0) {} // expected-note {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {}
+  S5 &operator+(const S5 &arg);
+
+public:
+  S5(int v) : a(v) {}
+};
+class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
+#if __cplusplus >= 201103L // C++11 or later
+// expected-note@-2 3 {{candidate function (the implicit move assignment operator) not viable}}
+#endif
+  int a;
+
+public:
+  S6() : a(6) {}
+  operator int() { return 6; }
+} o;
+
+S3 h, k;
+#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
+
+template <class T>       // expected-note {{declared here}}
+T tmain(T argc) {
+  const T d = T();       // expected-note 4 {{'d' defined here}}
+  const T da[5] = {T()}; // expected-note 2 {{'da' defined here}}
+  T qa[5] = {T()};
+  T i;
+  T &j = i;                    // expected-note 4 {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const T &r = da[(int)i];     // expected-note 2 {{'r' defined here}}
+  T &q = qa[(int)i];           // expected-note 2 {{'q' defined here}}
+  T fl;
+#pragma omp target reduction // expected-error {{expected '(' after 'reduction'}}
+  foo();
+#pragma omp target reduction + // expected-error {{expected '(' after 'reduction'}} expected-warning {{extra tokens at the end of '#pragma omp target' are ignored}}
+  foo();
+#pragma omp target reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp target reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp target reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp target reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}}
+  foo();
+#pragma omp target reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp target reduction(& : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp target reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp target reduction(|| : argc ? i : argc) // expected-error 2 {{expected variable name, array element or array section}}
+  foo();
+#pragma omp target reduction(foo : argc) //expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'float'}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'int'}}
+  foo();
+#pragma omp target reduction(&& : argc)
+  foo();
+#pragma omp target reduction(^ : T) // expected-error {{'T' does not refer to a value}}
+  foo();
+#pragma omp target reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 3 {{const-qualified list item cannot be reduction}} expected-error 2 {{'operator+' is a private member of 'S2'}}
+  foo();
+#pragma omp target reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 4 {{arguments of OpenMP clause 'reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 3 {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  foo();
+#pragma omp target reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}} expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp target reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  foo();
+#pragma omp target reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(+ : h, k) // expected-error {{threadprivate or thread local variable cannot be reduction}}
+  foo();
+#pragma omp target reduction(+ : o) // expected-error 2 {{no viable overloaded '='}}
+  foo();
+#pragma omp parallel private(i), reduction(+ : j), reduction(+ : q) // expected-error 4 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp parallel private(k)
+#pragma omp target reduction(+ : p), reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp target reduction(+ : p), reduction(+ : p) // expected-error 2 {{variable can appear only once in OpenMP 'reduction' clause}} expected-note 2 {{previously referenced here}}
+  foo();
+#pragma omp target reduction(+ : r) // expected-error 2 {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp target reduction(min : i)
+#pragma omp parallel reduction(max : j) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp parallel
+#pragma omp for private(fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp target reduction(+ : fl)
+    foo();
+#pragma omp parallel
+#pragma omp for reduction(- : fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp target reduction(+ : fl)
+    foo();
+
+  return T();
+}
+
+namespace A {
+double x;
+#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  const int d = 5;       // expected-note 2 {{'d' defined here}}
+  const int da[5] = {0}; // expected-note {{'da' defined here}}
+  int qa[5] = {0};
+  S4 e(4);
+  S5 g(5);
+  int i;
+  int &j = i;                  // expected-note 2 {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const int &r = da[i];        // expected-note {{'r' defined here}}
+  int &q = qa[i];              // expected-note {{'q' defined here}}
+  float fl;
+#pragma omp target reduction // expected-error {{expected '(' after 'reduction'}}
+  foo();
+#pragma omp target reduction + // expected-error {{expected '(' after 'reduction'}} expected-warning {{extra tokens at the end of '#pragma omp target' are ignored}}
+  foo();
+#pragma omp target reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp target reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp target reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp target reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}}
+  foo();
+#pragma omp target reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp target reduction(foo : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max'}}
+  foo();
+#pragma omp target reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp target reduction(|| : argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name, array element or array section}}
+  foo();
+#pragma omp target reduction(~ : argc) // expected-error {{expected unqualified-id}}
+  foo();
+#pragma omp target reduction(&& : argc)
+  foo();
+#pragma omp target reduction(^ : S1) // expected-error {{'S1' does not refer to a value}}
+  foo();
+#pragma omp target reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{const-qualified list item cannot be reduction}} expected-error {{'operator+' is a private member of 'S2'}}
+  foo();
+#pragma omp target reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{arguments of OpenMP clause 'reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 2 {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  foo();
+#pragma omp target reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp target reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  foo();
+#pragma omp target reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp target reduction(& : e, g) // expected-error {{calling a private constructor of class 'S4'}} expected-error {{nvalid operands to binary expression ('S4' and 'S4')}} expected-error {{calling a private constructor of class 'S5'}} expected-error {{invalid operands to binary expression ('S5' and 'S5')}}
+  foo();
+#pragma omp target reduction(+ : h, k, B::x) // expected-error 2 {{threadprivate or thread local variable cannot be reduction}}
+  foo();
+#pragma omp target reduction(+ : o) // expected-error {{no viable overloaded '='}}
+  foo();
+#pragma omp parallel private(i), reduction(+ : j), reduction(+ : q) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp parallel private(k)
+#pragma omp target reduction(+ : p), reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp target reduction(+ : p), reduction(+ : p) // expected-error {{variable can appear only once in OpenMP 'reduction' clause}} expected-note {{previously referenced here}}
+  foo();
+#pragma omp target reduction(+ : r) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp target reduction(min : i)
+#pragma omp parallel reduction(max : j) // expected-error {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp parallel
+#pragma omp for private(fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp target reduction(+ : fl)
+    foo();
+#pragma omp parallel
+#pragma omp for reduction(- : fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp target reduction(+ : fl)
+    foo();
+  static int m;
+#pragma omp target reduction(+ : m) // OK
+  m++;
+
+  return tmain(argc) + tmain(fl); // expected-note {{in instantiation of function template specialization 'tmain<int>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<float>' requested here}}
+}
diff --git a/test/OpenMP/target_simd_codegen.cpp b/test/OpenMP/target_simd_codegen.cpp
new file mode 100644
index 000000000000..d57e3818559a
--- /dev/null
+++ b/test/OpenMP/target_simd_codegen.cpp
@@ -0,0 +1,675 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// Test target codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: [[TT:%.+]] = type { i64, i8 }
+// CHECK-DAG: [[S1:%.+]] = type { double }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK: [[ENTTY:%.+]] = type { i8*, i8*, i{{32|64}}, i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// We have 8 target regions, but only 7 that actually will generate offloading
+// code, only 6 will have mapped arguments, and only 4 have all-constant map
+// sizes.
+
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288]
+// CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547]
+// CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK-NOT: @{{.+}} = constant [[ENTTY]]
+
+// Check if offloading descriptor is created.
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// Check target registration is registered as a Ctor.
+// CHECK: appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+
+template<typename tx, typename ty>
+struct TT{
+  tx X;
+  ty Y;
+};
+
+// CHECK-LABEL: get_val
+long long get_val() { return 0; }
+
+// CHECK: define {{.*}}[[FOO:@.+]](
+int foo(int n) {
+  int a = 0;
+  short aa = 0;
+  float b[10];
+  float bn[n];
+  double c[5][10];
+  double cn[5][n];
+  TT<long long, char> d;
+
+  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target_nowait(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i64* null)
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT0:@.+]]()
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target simd nowait
+  for (int i = 3; i < 32; i += 5) {
+  }
+
+  // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}}, i{{32|64}}{{[*]*}} {{[^)]+}})
+  long long k = get_val();
+  #pragma omp target simd if(target: 0) linear(k : 3)
+  for (int i = 10; i > 1; i--) {
+    a += 1;
+  }
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT2]], i32 0, i32 0))
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2]],
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  int lin = 12;
+  #pragma omp target simd if(target: 1) linear(lin, a : get_val())
+  for (unsigned long long it = 2000; it >= 600; it-=400) {
+    aa += 1;
+  }
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
+  // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+  // CHECK:       [[IFTHEN]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0))
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+  // CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
+
+  #pragma omp target simd if(target: n>10)
+  for (short it = 6; it <= 20; it-=-4) {
+    a += 1;
+    aa += 1;
+  }
+
+  // We capture 3 VLA sizes in this target region
+  // CHECK-64:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-64:       [[A_ADDR:%.+]] = bitcast i[[SZ]]* [[A_CADDR:%.+]] to i32*
+  // CHECK-64:       store i32 [[A_VAL]], i32* [[A_ADDR]],
+  // CHECK-64:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK-32:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-32:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-32:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK:       [[BNSIZE:%.+]] = mul nuw i[[SZ]] [[VLA0:%.+]], 4
+  // CHECK:       [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]]
+  // CHECK:       [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
+  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+  // CHECK:       [[TRY]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0))
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX2:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX3:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[SADDR4:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX4:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[SADDR5:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX5:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[PADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[SADDR6:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX6:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[PADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[SADDR7:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX7:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[PADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[SADDR8:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX8:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX8]]
+  // CHECK-DAG:   [[PADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX8]]
+
+  // The names below are not necessarily consistent with the names used for the
+  // addresses above as some are repeated.
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+  // CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CBPADDR1:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CPADDR1:%.+]],
+  // CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CBPADDR2:%.+]],
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CPADDR2:%.+]],
+  // CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CBPADDR3:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CPADDR3:%.+]],
+  // CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CBPADDR4:%.+]],
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CPADDR4:%.+]],
+  // CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   store i[[SZ]] 40, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CBPADDR5:%.+]],
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CPADDR5:%.+]],
+  // CHECK-DAG:   [[CBPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   [[CPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   store i[[SZ]] [[BNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CBPADDR6:%.+]],
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CPADDR6:%.+]],
+  // CHECK-DAG:   [[CBPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   [[CPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   store i[[SZ]] 400, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CBPADDR7:%.+]],
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CPADDR7:%.+]],
+  // CHECK-DAG:   [[CBPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   [[CPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   store i[[SZ]] [[CNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CBPADDR8:%.+]],
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8:%.+]],
+  // CHECK-DAG:   [[CBPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target simd if(target: n>20)
+  for (unsigned char it = 'z'; it >= 'a'; it+=-1) {
+    a += 1;
+    b[2] += 1.0;
+    bn[3] += 1.0;
+    c[1][2] += 1.0;
+    cn[1][3] += 1.0;
+    d.X += 1;
+    d.Y += 1;
+  }
+
+  return a;
+}
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions in foo().
+
+// CHECK:       define internal void [[HVT0]]()
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+
+// CHECK:       define internal void [[HVT1]](i[[SZ]] %{{.+}}, i{{32|64}}{{[*]*.*}} %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK:       !llvm.mem.parallel_loop_access
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT2]](i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT3]]
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       !llvm.loop
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT4]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x float]*
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_BN:%.+]] = alloca float*
+// CHECK:       [[LOCAL_C:%.+]] = alloca [5 x [10 x double]]*
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA3:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_CN:%.+]] = alloca double*
+// CHECK:       [[LOCAL_D:%.+]] = alloca [[TT]]*
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store [10 x float]* [[ARG_B:%.+]], [10 x float]** [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store float* [[ARG_BN:%.+]], float** [[LOCAL_BN]]
+// CHECK-DAG:   store [5 x [10 x double]]* [[ARG_C:%.+]], [5 x [10 x double]]** [[LOCAL_C]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA3:%.+]], i[[SZ]]* [[LOCAL_VLA3]]
+// CHECK-DAG:   store double* [[ARG_CN:%.+]], double** [[LOCAL_CN]]
+// CHECK-DAG:   store [[TT]]* [[ARG_D:%.+]], [[TT]]** [[LOCAL_D]]
+
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x float]*, [10 x float]** [[LOCAL_B]],
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[REF_BN:%.+]] = load float*, float** [[LOCAL_BN]],
+// CHECK-DAG:   [[REF_C:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[LOCAL_C]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[VAL_VLA3:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA3]],
+// CHECK-DAG:   [[REF_CN:%.+]] = load double*, double** [[LOCAL_CN]],
+// CHECK-DAG:   [[REF_D:%.+]] = load [[TT]]*, [[TT]]** [[LOCAL_D]],
+
+
+template<typename tx>
+tx ftemplate(int n) {
+  tx a = 0;
+  short aa = 0;
+  tx b[10];
+
+  #pragma omp target simd if(target: n>40)
+  for (long long i = -10; i < 10; i += 3) {
+    a += 1;
+    aa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+static
+int fstatic(int n) {
+  int a = 0;
+  short aa = 0;
+  char aaa = 0;
+  int b[10];
+
+  #pragma omp target simd if(target: n>50)
+  for (unsigned i=100; i<10; i+=10) {
+    a += 1;
+    aa += 1;
+    aaa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+struct S1 {
+  double a;
+
+  int r1(int n){
+    int b = n+1;
+    short int c[2][n];
+
+    #pragma omp target simd if(target: n>60)
+    for (unsigned long long it = 2000; it >= 600; it -= 400) {
+      this->a = (double)b + 1.5;
+      c[1][1] = ++a;
+    }
+
+    return c[1][1] + (int)b;
+  }
+};
+
+// CHECK: define {{.*}}@{{.*}}bar{{.*}}
+int bar(int n){
+  int a = 0;
+
+  // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}})
+  a += foo(n);
+
+  S1 S;
+  // CHECK: call {{.*}}i32 [[FS1:@.+]]([[S1]]* {{.*}}, i32 {{.*}})
+  a += S.r1(n);
+
+  // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}})
+  a += fstatic(n);
+
+  // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}})
+  a += ftemplate<int>(n);
+
+  return a;
+}
+
+//
+// CHECK: define {{.*}}[[FS1]]
+//
+// CHECK:          i8* @llvm.stacksave()
+// CHECK-64:       [[B_ADDR:%.+]] = bitcast i[[SZ]]* [[B_CADDR:%.+]] to i32*
+// CHECK-64:       store i32 %{{.+}}, i32* [[B_ADDR]],
+// CHECK-64:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_CADDR]],
+
+// CHECK-32:       store i32 %{{.+}}, i32* [[B_ADDR:%.+]],
+// CHECK-32:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_ADDR]],
+
+// We capture 2 VLA sizes in this target region
+// CHECK:       [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]]
+// CHECK:       [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2
+
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
+// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+// CHECK:       [[TRY]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0))
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX0:[0-9]+]]
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX0]]
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX0]]
+// CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX1:[0-9]+]]
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX1]]
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX1]]
+// CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX2:[0-9]+]]
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX2]]
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX2]]
+// CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX3:[0-9]+]]
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX3]]
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX3]]
+
+// The names below are not necessarily consistent with the names used for the
+// addresses above as some are repeated.
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR3:%.+]],
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   store i[[SZ]] 8, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
+
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+
+//
+// CHECK: define {{.*}}[[FSTATIC]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0))
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2]],
+
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP]], i32 0, i32 3
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P]], i32 0, i32 3
+// CHECK-DAG:   [[CBPADDR3:%.+]] = bitcast i8** [[BPADDR3]] to [10 x i32]**
+// CHECK-DAG:   [[CPADDR3:%.+]] = bitcast i8** [[PADDR3]] to [10 x i32]**
+// CHECK-DAG:   store [10 x i32]* [[VAL3:%.+]], [10 x i32]** [[CBPADDR3]],
+// CHECK-DAG:   store [10 x i32]* [[VAL3]], [10 x i32]** [[CPADDR3]],
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+//
+// CHECK: define {{.*}}[[FTEMPLATE]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0))
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0]],
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1]],
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to [10 x i32]**
+// CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to [10 x i32]**
+// CHECK-DAG:   store [10 x i32]* [[VAL2:%.+]], [10 x i32]** [[CBPADDR2]],
+// CHECK-DAG:   store [10 x i32]* [[VAL2]], [10 x i32]** [[CPADDR2]],
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions of the callees of bar().
+
+// CHECK:       define internal void [[HVT7]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_THIS:%.+]] = alloca [[S1]]*
+// CHECK:       [[LOCAL_B:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_C:%.+]] = alloca i16*
+// CHECK-DAG:   store [[S1]]* [[ARG_THIS:%.+]], [[S1]]** [[LOCAL_THIS]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_B:%.+]], i[[SZ]]* [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i16* [[ARG_C:%.+]], i16** [[LOCAL_C]]
+// Store captures in the context.
+// CHECK-DAG:   [[REF_THIS:%.+]] = load [[S1]]*, [[S1]]** [[LOCAL_THIS]],
+// CHECK-64-DAG:[[CONV_BP:%.+]] = bitcast i[[SZ]]* [[LOCAL_B]] to i32*
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[REF_C:%.+]] = load i16*, i16** [[LOCAL_C]],
+
+
+// CHECK:       define internal void [[HVT6]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AAA:%.+]], i[[SZ]]* [[LOCAL_AAA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[CONV_AAAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA]] to i8*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK:       define internal void [[HVT5]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+#endif
diff --git a/test/OpenMP/target_simd_codegen_registration.cpp b/test/OpenMP/target_simd_codegen_registration.cpp
new file mode 100644
index 000000000000..a6e6e9b38565
--- /dev/null
+++ b/test/OpenMP/target_simd_codegen_registration.cpp
@@ -0,0 +1,451 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target simd codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// Check that no target code is emmitted if no omptests flag was provided.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK-NTARGET
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK:    [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// CHECK-DAG: [[A1:@.+]] = internal global [[SA]]
+// CHECK-DAG: [[A2:@.+]] = global [[SA]]
+// CHECK-DAG: [[B1:@.+]] = global [[SB]]
+// CHECK-DAG: [[B2:@.+]] = global [[SB]]
+// CHECK-DAG: [[C1:@.+]] = internal global [[SC]]
+// CHECK-DAG: [[D1:@.+]] = global [[SD]]
+// CHECK-DAG: [[E1:@.+]] = global [[SE]]
+// CHECK-DAG: [[T1:@.+]] = global [[ST1]]
+// CHECK-DAG: [[T2:@.+]] = global [[ST2]]
+
+// CHECK-NTARGET-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-NTARGET-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-NTARGET-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-NTARGET-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-NTARGET-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-NTARGET-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-NTARGET-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-NTARGET-NOT: type { i8*, i8*, %
+// CHECK-NTARGET-NOT: type { i32, %
+
+// We have 7 target regions
+
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// TCHECK-NOT: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+
+// CHECK-NTARGET-NOT: private constant i8 0
+// CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
+
+// CHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// CHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// CHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// CHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// CHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// CHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// CHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// CHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// CHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// CHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// CHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// CHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// CHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// TCHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// TCHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// TCHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// TCHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// TCHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// TCHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// TCHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// TCHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// TCHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// TCHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// TCHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// TCHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// TCHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// We have 4 initializers, one for the 500 priority, another one for 501, or more for the default priority, and the last one for the offloading registration function.
+// CHECK: @llvm.global_ctors = appending global [4 x { i32, void ()*, i8* }] [
+// CHECK-SAME: { i32, void ()*, i8* } { i32 500, void ()* [[P500:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 501, void ()* [[P501:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 65535, void ()* [[PMAX:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+// CHECK-NTARGET: @llvm.global_ctors = appending global [3   x { i32, void ()*, i8* }] [
+
+extern int *R;
+
+struct SA {
+  int arr[4];
+  void foo() {
+    int a = *R;
+    a += 1;
+    *R = a;
+  }
+  SA() {
+    int a = *R;
+    a += 2;
+    *R = a;
+  }
+  ~SA() {
+    int a = *R;
+    a += 3;
+    *R = a;
+  }
+};
+
+struct SB {
+  int arr[8];
+  void foo() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 4;
+    *R = a;
+  }
+  SB() {
+    int a = *R;
+    a += 5;
+    *R = a;
+  }
+  ~SB() {
+    int a = *R;
+    a += 6;
+    *R = a;
+  }
+};
+
+struct SC {
+  int arr[16];
+  void foo() {
+    int a = *R;
+    a += 7;
+    *R = a;
+  }
+  SC() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 8;
+    *R = a;
+  }
+  ~SC() {
+    int a = *R;
+    a += 9;
+    *R = a;
+  }
+};
+
+struct SD {
+  int arr[32];
+  void foo() {
+    int a = *R;
+    a += 10;
+    *R = a;
+  }
+  SD() {
+    int a = *R;
+    a += 11;
+    *R = a;
+  }
+  ~SD() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 12;
+    *R = a;
+  }
+};
+
+struct SE {
+  int arr[64];
+  void foo() {
+    int a = *R;
+    #pragma omp target simd if(target: 0)
+    for (int i = 0; i < 10; ++i)
+      a += 13;
+    *R = a;
+  }
+  SE() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 14;
+    *R = a;
+  }
+  ~SE() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 15;
+    *R = a;
+  }
+};
+
+template <int x>
+struct ST {
+  int arr[128 + x];
+  void foo() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 16 + x;
+    *R = a;
+  }
+  ST() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 17 + x;
+    *R = a;
+  }
+  ~ST() {
+    int a = *R;
+    #pragma omp target simd
+    for (int i = 0; i < 10; ++i)
+      a += 18 + x;
+    *R = a;
+  }
+};
+
+// We have to make sure we us all the target regions:
+//CHECK-DAG: define internal void @[[NAME1]](
+//CHECK-DAG: call void @[[NAME1]](
+//CHECK-DAG: define internal void @[[NAME2]](
+//CHECK-DAG: call void @[[NAME2]](
+//CHECK-DAG: define internal void @[[NAME3]](
+//CHECK-DAG: call void @[[NAME3]](
+//CHECK-DAG: define internal void @[[NAME4]](
+//CHECK-DAG: call void @[[NAME4]](
+//CHECK-DAG: define internal void @[[NAME5]](
+//CHECK-DAG: call void @[[NAME5]](
+//CHECK-DAG: define internal void @[[NAME6]](
+//CHECK-DAG: call void @[[NAME6]](
+//CHECK-DAG: define internal void @[[NAME7]](
+//CHECK-DAG: call void @[[NAME7]](
+//CHECK-DAG: define internal void @[[NAME8]](
+//CHECK-DAG: call void @[[NAME8]](
+//CHECK-DAG: define internal void @[[NAME9]](
+//CHECK-DAG: call void @[[NAME9]](
+//CHECK-DAG: define internal void @[[NAME10]](
+//CHECK-DAG: call void @[[NAME10]](
+//CHECK-DAG: define internal void @[[NAME11]](
+//CHECK-DAG: call void @[[NAME11]](
+//CHECK-DAG: define internal void @[[NAME12]](
+//CHECK-DAG: call void @[[NAME12]](
+
+//TCHECK-DAG: define void @[[NAME1]](
+//TCHECK-DAG: define void @[[NAME2]](
+//TCHECK-DAG: define void @[[NAME3]](
+//TCHECK-DAG: define void @[[NAME4]](
+//TCHECK-DAG: define void @[[NAME5]](
+//TCHECK-DAG: define void @[[NAME6]](
+//TCHECK-DAG: define void @[[NAME7]](
+//TCHECK-DAG: define void @[[NAME8]](
+//TCHECK-DAG: define void @[[NAME9]](
+//TCHECK-DAG: define void @[[NAME10]](
+//TCHECK-DAG: define void @[[NAME11]](
+//TCHECK-DAG: define void @[[NAME12]](
+
+// CHECK-NTARGET-NOT: __tgt_target
+// CHECK-NTARGET-NOT: __tgt_register_lib
+// CHECK-NTARGET-NOT: __tgt_unregister_lib
+
+// TCHECK-NOT: __tgt_target
+// TCHECK-NOT: __tgt_register_lib
+// TCHECK-NOT: __tgt_unregister_lib
+
+// We have 2 initializers with priority 500
+//CHECK: define internal void [[P500]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 1 initializers with priority 501
+//CHECK: define internal void [[P501]](
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 6 initializers with default priority
+//CHECK: define internal void [[PMAX]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// Check registration and unregistration
+
+//CHECK:     define internal void @[[UNREGFN:.+]](i8*)
+//CHECK-SAME: comdat($[[REGFN]]) {
+//CHECK:     call i32 @__tgt_unregister_lib([[DSCTY]]* [[DESC]])
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_unregister_lib([[DSCTY]]*)
+
+//CHECK:     define linkonce hidden void @[[REGFN]](i8*)
+//CHECK-SAME: comdat {
+//CHECK:     call i32 @__tgt_register_lib([[DSCTY]]* [[DESC]])
+//CHECK:     call i32 @__cxa_atexit(void (i8*)* @[[UNREGFN]], i8* bitcast ([[DSCTY]]* [[DESC]] to i8*),
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_register_lib([[DSCTY]]*)
+
+static __attribute__((init_priority(500))) SA a1;
+SA a2;
+SB __attribute__((init_priority(500))) b1;
+SB __attribute__((init_priority(501))) b2;
+static SC c1;
+SD d1;
+SE e1;
+ST<100> t1;
+ST<1000> t2;
+
+
+int bar(int a){
+  int r = a;
+
+  a1.foo();
+  a2.foo();
+  b1.foo();
+  b2.foo();
+  c1.foo();
+  d1.foo();
+  e1.foo();
+  t1.foo();
+  t2.foo();
+
+  #pragma omp target simd
+  for (int i = 0; i < 10; ++i)
+    ++r;
+
+  return r + *R;
+}
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+#endif
diff --git a/test/OpenMP/target_simd_codegen_registration_naming.cpp b/test/OpenMP/target_simd_codegen_registration_naming.cpp
new file mode 100644
index 000000000000..2ed74da4f05b
--- /dev/null
+++ b/test/OpenMP/target_simd_codegen_registration_naming.cpp
@@ -0,0 +1,68 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target simd codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[CA:%.+]] = type { i32* }
+
+// CHECK: define {{.*}}i32 @[[NNAME:.+]](i32 {{.*}}%{{.+}})
+int nested(int a){
+  // CHECK: call void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME]]_l[[T1L:[0-9]+]](
+  #pragma omp target simd
+  for (int i = 0; i < 10; ++i)
+    ++a;
+
+  // CHECK: call void @"[[LNAME:.+]]"([[CA]]*
+  auto F = [&](){
+    #pragma omp parallel
+    {
+      #pragma omp target simd
+      for (int i = 0; i < 10; ++i)
+        ++a;
+    }
+  };
+
+  F();
+
+  return a;
+}
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T1L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME:.+]]_l[[T1L:[0-9]+]](
+
+// CHECK: define {{.*}}void @"[[LNAME]]"(
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.+}}[[PNAME:@.+]] to
+
+// CHECK: define {{.*}}void [[PNAME]](
+// CHECK: call void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L:[0-9]+]](
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME:.+]]_l[[T2L:[0-9]+]](
+
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+#endif
diff --git a/test/OpenMP/target_simd_depend_messages.cpp b/test/OpenMP/target_simd_depend_messages.cpp
index 3fc46f4c0848..89184f25d7a1 100644
--- a/test/OpenMP/target_simd_depend_messages.cpp
+++ b/test/OpenMP/target_simd_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target simd depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target simd depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target simd depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target simd depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target simd depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target simd depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target simd depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target simd depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target simd depend (in : main) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target simd depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-  #pragma omp target simd depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target simd depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
   #pragma omp target simd depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_simd_lastprivate_messages.cpp b/test/OpenMP/target_simd_lastprivate_messages.cpp
index afb38d9c9222..9a6eb241728b 100644
--- a/test/OpenMP/target_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/target_simd_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   S2 &operator=(const S2 &);
   const S2 &operator=(const S2 &) const;
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/target_simd_map_messages.cpp b/test/OpenMP/target_simd_map_messages.cpp
index 63bbdde77255..2473bbb49bea 100644
--- a/test/OpenMP/target_simd_map_messages.cpp
+++ b/test/OpenMP/target_simd_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -112,9 +112,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target simd map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -214,11 +214,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target simd map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target simd map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_simd_reduction_messages.cpp b/test/OpenMP/target_simd_reduction_messages.cpp
index ca14acb2293f..9e1bb0521d71 100644
--- a/test/OpenMP/target_simd_reduction_messages.cpp
+++ b/test/OpenMP/target_simd_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_teams_codegen.cpp b/test/OpenMP/target_teams_codegen.cpp
index 208cda6bb350..e6c250178d06 100644
--- a/test/OpenMP/target_teams_codegen.cpp
+++ b/test/OpenMP/target_teams_codegen.cpp
@@ -38,16 +38,18 @@
 // code, only 6 will have mapped arguments, and only 4 have all-constant map
 // sizes.
 
-// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i{{32|64}}] [i[[SZ:32|64]] 2]
-// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: [[SIZET:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288]
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 2]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
-// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i32] [i32 288, i32 288]
-// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i32] [i32 288, i32 35, i32 288, i32 35, i32 35, i32 288, i32 288, i32 35, i32 35]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547]
 // CHECK-DAG: [[SIZET5:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
-// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i32] [i32 288, i32 288, i32 35]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
 // CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [4 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
-// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i32] [i32 288, i32 288, i32 288, i32 35]
-// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i32] [i32 35, i32 288, i32 288, i32 288, i32 35]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [4 x i64] [i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
 // CHECK-DAG: @{{.*}} = private constant i8 0
 // CHECK-DAG: @{{.*}} = private constant i8 0
 // CHECK-DAG: @{{.*}} = private constant i8 0
@@ -83,6 +85,8 @@ struct TT{
   ty Y;
 };
 
+int global;
+
 // CHECK: define {{.*}}[[FOO:@.+]](
 int foo(int n) {
   int a = 0;
@@ -93,30 +97,44 @@ int foo(int n) {
   double cn[5][n];
   TT<long long, char> d;
 
-  // CHECK:       [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i[[SZ]]* null, i32* null, i32 0, i32 0)
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams_nowait(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT]], i32 0, i32 0), i32 {{[^,]+}}, i32 {{[^)]+}})
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR2]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR2]]
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
-  // CHECK:       call void [[HVT0:@.+]]()
+  // CHECK:       call void [[HVT0:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
-  #pragma omp target teams
+  #pragma omp target teams num_teams(a) thread_limit(a) firstprivate(aa) nowait
   {
   }
 
-  // CHECK:       store i32 0, i32* [[RHV:%.+]], align 4
-  // CHECK:       store i32 -1, i32* [[RHV]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
   // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}})
   #pragma omp target teams if(target: 0)
   {
     a += 1;
   }
 
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i32* getelementptr inbounds ([1 x i32], [1 x i32]* [[MAPT2]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT2]], i32 0, i32 0), i32 0, i32 0)
   // CHECK-DAG:   [[BP]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[P]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
@@ -126,9 +144,7 @@ int foo(int n) {
   // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
   // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
 
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}})
@@ -142,7 +158,7 @@ int foo(int n) {
   // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
   // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CHECK:       [[IFTHEN]]
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i32* getelementptr inbounds ([2 x i32], [2 x i32]* [[MAPT3]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 0, i32 0)
   // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
 
@@ -159,21 +175,17 @@ int foo(int n) {
   // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
   // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]]
   // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]]
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK-NEXT:  br label %[[IFEND:.+]]
-
-  // CHECK:       [[IFELSE]]
-  // CHECK:       store i32 -1, i32* [[RHV]], align 4
-  // CHECK-NEXT:  br label %[[IFEND:.+]]
-
-  // CHECK:       [[IFEND]]
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
   // CHECK:       [[FAIL]]
   // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
   // CHECK-NEXT:  br label %[[END]]
   // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
   #pragma omp target teams if(target: n>10)
   {
     a += 1;
@@ -197,7 +209,7 @@ int foo(int n) {
   // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
   // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
   // CHECK:       [[TRY]]
-  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i32* getelementptr inbounds ([9 x i32], [9 x i32]* [[MAPT4]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0), i32 0, i32 0)
   // CHECK-DAG:   [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0
   // CHECK-DAG:   [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
@@ -286,9 +298,7 @@ int foo(int n) {
   // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
   // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
 
-  // CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-  // CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
   // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
 
   // CHECK:       [[FAIL]]
@@ -312,11 +322,12 @@ int foo(int n) {
 // Check that the offloading functions are emitted and that the arguments are
 // correct and loaded correctly for the target regions in foo().
 
-// CHECK:       define internal void [[HVT0]]()
-// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 0, void (i32*, i32*, ...)* bitcast (void (i32*, i32*)* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*))
+// CHECK:       define internal void [[HVT0]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] {{[^)]+}})
 //
 //
-// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid.)
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] {{[^)]+}})
+// CHECK:       alloca i16,
 // CHECK:       ret void
 // CHECK-NEXT:  }
 
@@ -441,7 +452,7 @@ int foo(int n) {
 // CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 9, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*)* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]])
 //
 //
-// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* %{{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* %{{.+}}, [[TT]]* {{.+}})
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}})
 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
 
 template<typename tx>
@@ -533,7 +544,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
 // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
 // CHECK:       [[TRY]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i32* getelementptr inbounds ([5 x i32], [5 x i32]* [[MAPT7]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT7]], i32 0, i32 0), i32 0, i32 0)
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
@@ -582,9 +593,7 @@ int bar(int n){
 // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
 // CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
 
 // CHECK:       [[FAIL]]
@@ -598,7 +607,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
 // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[IFTHEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i32* getelementptr inbounds ([4 x i32], [4 x i32]* [[MAPT6]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([4 x i[[SZ]]], [4 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([4 x i64], [4 x i64]* [[MAPT6]], i32 0, i32 0), i32 0, i32 0)
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [4 x i8*], [4 x i8*]* [[P:%.+]], i32 0, i32 0
 
@@ -630,21 +639,17 @@ int bar(int n){
 // CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
 // CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFEND]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
@@ -652,7 +657,7 @@ int bar(int n){
 // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
 // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CHECK:       [[IFTHEN]]
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i32* getelementptr inbounds ([3 x i32], [3 x i32]* [[MAPT5]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET5]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT5]], i32 0, i32 0), i32 0, i32 0)
 // CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
 // CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
 
@@ -677,21 +682,17 @@ int bar(int n){
 // CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
 // CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
 
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFELSE]]
-// CHECK:       store i32 -1, i32* [[RHV]], align 4
-// CHECK-NEXT:  br label %[[IFEND:.+]]
-
-// CHECK:       [[IFEND]]
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align 4
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 // CHECK:       [[FAIL]]
 // CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
 // CHECK-NEXT:  br label %[[END]]
 // CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT5]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
 
 
 
@@ -728,7 +729,7 @@ int bar(int n){
 // CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]])
 //
 //
-// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* %{{.+}})
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}})
 // To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
 
 
diff --git a/test/OpenMP/target_teams_codegen_registration.cpp b/test/OpenMP/target_teams_codegen_registration.cpp
index b8154538df04..38668a46b161 100644
--- a/test/OpenMP/target_teams_codegen_registration.cpp
+++ b/test/OpenMP/target_teams_codegen_registration.cpp
@@ -63,40 +63,40 @@
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // TCHECK-NOT: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 // CHECK-DAG: {{@.+}} = private constant i8 0
 // CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
-// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i32] [i32 288]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
 
 // CHECK-NTARGET-NOT: private constant i8 0
 // CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
diff --git a/test/OpenMP/target_teams_depend_messages.cpp b/test/OpenMP/target_teams_depend_messages.cpp
index 9308e8972872..57983eec8882 100644
--- a/test/OpenMP/target_teams_depend_messages.cpp
+++ b/test/OpenMP/target_teams_depend_messages.cpp
@@ -36,21 +36,21 @@ int main(int argc, char **argv, char *env[]) {
   foo();
 #pragma omp target teams depend (out: ) // expected-error {{expected expression}}
   foo();
-#pragma omp target teams depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+#pragma omp target teams depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   foo();
 #pragma omp target teams depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   foo();
-#pragma omp target teams depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams depend(in : argv[1][1] = '2')
   foo();
-#pragma omp target teams depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   foo();
 #pragma omp target teams depend (in : argv[0])
   foo();
 #pragma omp target teams depend (in : ) // expected-error {{expected expression}}
   foo();
-#pragma omp target teams depend (in : main) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams depend (in : main)
   foo();
-#pragma omp target teams depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+#pragma omp target teams depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   foo();
 #pragma omp target teams depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   foo();
diff --git a/test/OpenMP/target_teams_distribute_codegen.cpp b/test/OpenMP/target_teams_distribute_codegen.cpp
new file mode 100644
index 000000000000..17baa603d12b
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_codegen.cpp
@@ -0,0 +1,820 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// Test target codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+// CHECK-DAG: [[TT:%.+]] = type { i64, i8 }
+// CHECK-DAG: [[S1:%.+]] = type { double }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK: [[ENTTY:%.+]] = type { i8*, i8*, i{{32|64}}, i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// We have 8 target regions, but only 7 that actually will generate offloading
+// code, only 6 will have mapped arguments, and only 4 have all-constant map
+// sizes.
+
+// CHECK-DAG: [[SIZET:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288]
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 2]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [5 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [5 x i64] [i64 288, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET7:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK-NOT: @{{.+}} = constant [[ENTTY]]
+
+// Check if offloading descriptor is created.
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// Check target registration is registered as a Ctor.
+// CHECK: appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+
+template<typename tx, typename ty>
+struct TT{
+  tx X;
+  ty Y;
+};
+
+int global;
+
+// CHECK: define {{.*}}[[FOO:@.+]](
+int foo(int n) {
+  int a = 0;
+  short aa = 0;
+  float b[10];
+  float bn[n];
+  double c[5][10];
+  double cn[5][n];
+  TT<long long, char> d;
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams_nowait(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT]], i32 0, i32 0), i32 {{[^,]+}}, i32 {{[^)]+}})
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR2]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR2]]
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT0:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target teams distribute num_teams(a) thread_limit(a) firstprivate(aa) nowait
+  for (int i = 0; i < 10; ++i) {
+  }
+
+  // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}})
+  #pragma omp target teams distribute if(target: 0)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+  }
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT2]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target teams distribute if(target: 1)
+  for (int i = 0; i < 10; ++i) {
+    aa += 1;
+  }
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
+  // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+  // CHECK:       [[IFTHEN]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]]
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
+  #pragma omp target teams distribute if(target: n>10)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+    aa += 1;
+  }
+
+  // We capture 3 VLA sizes in this target region
+  // CHECK-64:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-64:       [[A_ADDR:%.+]] = bitcast i[[SZ]]* [[A_CADDR:%.+]] to i32*
+  // CHECK-64:       store i32 [[A_VAL]], i32* [[A_ADDR]],
+  // CHECK-64:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK-32:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-32:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-32:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK:       [[BNSIZE:%.+]] = mul nuw i[[SZ]] [[VLA0:%.+]], 4
+  // CHECK:       [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]]
+  // CHECK:       [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
+  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+  // CHECK:       [[TRY]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX2:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX3:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[SADDR4:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX4:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[SADDR5:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX5:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[PADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[SADDR6:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX6:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[PADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[SADDR7:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX7:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[PADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[SADDR8:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX8:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX8]]
+  // CHECK-DAG:   [[PADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX8]]
+
+  // The names below are not necessarily consistent with the names used for the
+  // addresses above as some are repeated.
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+  // CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CBPADDR1:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CPADDR1:%.+]],
+  // CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CBPADDR2:%.+]],
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CPADDR2:%.+]],
+  // CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CBPADDR3:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CPADDR3:%.+]],
+  // CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CBPADDR4:%.+]],
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CPADDR4:%.+]],
+  // CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   store i[[SZ]] 40, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CBPADDR5:%.+]],
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CPADDR5:%.+]],
+  // CHECK-DAG:   [[CBPADDR5]] = bitcast i8** {{%[^,]+}} to  float**
+  // CHECK-DAG:   [[CPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   store i[[SZ]] [[BNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CBPADDR6:%.+]],
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CPADDR6:%.+]],
+  // CHECK-DAG:   [[CBPADDR6]] = bitcast i8** {{%[^,]+}} to  [5 x [10 x double]]**
+  // CHECK-DAG:   [[CPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   store i[[SZ]] 400, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CBPADDR7:%.+]],
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CPADDR7:%.+]],
+  // CHECK-DAG:   [[CBPADDR7]] = bitcast i8** {{%[^,]+}} to  double**
+  // CHECK-DAG:   [[CPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   store i[[SZ]] [[CNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CBPADDR8:%.+]],
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8:%.+]],
+  // CHECK-DAG:   [[CBPADDR8]] = bitcast i8** {{%[^,]+}} to  [[TT]]**
+  // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target teams distribute if(target: n>20)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+    b[2] += 1.0;
+    bn[3] += 1.0;
+    c[1][2] += 1.0;
+    cn[1][3] += 1.0;
+    d.X += 1;
+    d.Y += 1;
+  }
+
+  return a;
+}
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions in foo().
+
+// CHECK:       define internal void [[HVT0]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] {{[^)]+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] {{[^)]+}})
+// CHECK:       alloca i16,
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+
+// CHECK:       define internal void [[HVT1]](i[[SZ]] %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK-64:    [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i32*
+// CHECK-64:    store i32 [[AA]], i32* [[AA_C]], align
+// CHECK-32:    store i32 [[AA]], i32* [[AA_CASTED]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED1:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED1]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT2]](i[[SZ]] %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK:       store i16 [[AA]], i16* [[AA_C]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED2:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED2]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT3]]
+// Create stack storage and store argument in there.
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[A_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK-64-DAG:[[A:%.+]] = load i32, i32* [[A_CADDR]], align
+// CHECK-32-DAG:[[A:%.+]] = load i32, i32* [[A_ADDR]], align
+// CHECK-64-DAG:[[A_C:%.+]] = bitcast i[[SZ]]* [[A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[A]], i32* [[A_C]], align
+// CHECK-32-DAG:store i32 [[A]], i32* [[A_CASTED]], align
+// CHECK-DAG:   [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK-DAG:   [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[AA]], i16* [[AA_C]], align
+// CHECK-DAG:   [[PARAM1:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CASTED]], align
+// CHECK-DAG:   [[PARAM2:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK-DAG:   call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]])* [[OMP_OUTLINED3:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM1]], i[[SZ]] [[PARAM2]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED3]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT4]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x float]*
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_BN:%.+]] = alloca float*
+// CHECK:       [[LOCAL_C:%.+]] = alloca [5 x [10 x double]]*
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA3:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_CN:%.+]] = alloca double*
+// CHECK:       [[LOCAL_D:%.+]] = alloca [[TT]]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store [10 x float]* [[ARG_B:%.+]], [10 x float]** [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store float* [[ARG_BN:%.+]], float** [[LOCAL_BN]]
+// CHECK-DAG:   store [5 x [10 x double]]* [[ARG_C:%.+]], [5 x [10 x double]]** [[LOCAL_C]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA3:%.+]], i[[SZ]]* [[LOCAL_VLA3]]
+// CHECK-DAG:   store double* [[ARG_CN:%.+]], double** [[LOCAL_CN]]
+// CHECK-DAG:   store [[TT]]* [[ARG_D:%.+]], [[TT]]** [[LOCAL_D]]
+
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x float]*, [10 x float]** [[LOCAL_B]],
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[REF_BN:%.+]] = load float*, float** [[LOCAL_BN]],
+// CHECK-DAG:   [[REF_C:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[LOCAL_C]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[VAL_VLA3:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA3]],
+// CHECK-DAG:   [[REF_CN:%.+]] = load double*, double** [[LOCAL_CN]],
+// CHECK-DAG:   [[REF_D:%.+]] = load [[TT]]*, [[TT]]** [[LOCAL_D]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 9, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*)* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+template<typename tx>
+tx ftemplate(int n) {
+  tx a = 0;
+  short aa = 0;
+  tx b[10];
+
+  #pragma omp target teams distribute if(target: n>40)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+    aa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+static
+int fstatic(int n) {
+  int a = 0;
+  short aa = 0;
+  char aaa = 0;
+  int b[10];
+
+  #pragma omp target teams distribute if(target: n>50)
+  for (int i = a; i < n; ++i) {
+    a += 1;
+    aa += 1;
+    aaa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+struct S1 {
+  double a;
+
+  int r1(int n){
+    int b = n+1;
+    short int c[2][n];
+
+    #pragma omp target teams distribute if(target: n>60)
+    for (int i = 0; i < 10; ++i) {
+      this->a = (double)b + 1.5;
+      c[1][1] = ++a;
+    }
+
+    return c[1][1] + (int)b;
+  }
+};
+
+// CHECK: define {{.*}}@{{.*}}bar{{.*}}
+int bar(int n){
+  int a = 0;
+
+  // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}})
+  a += foo(n);
+
+  S1 S;
+  // CHECK: call {{.*}}i32 [[FS1:@.+]]([[S1]]* {{.*}}, i32 {{.*}})
+  a += S.r1(n);
+
+  // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}})
+  a += fstatic(n);
+
+  // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}})
+  a += ftemplate<int>(n);
+
+  return a;
+}
+
+//
+// CHECK: define {{.*}}[[FS1]]
+//
+// CHECK:          i8* @llvm.stacksave()
+// CHECK-64:       [[B_ADDR:%.+]] = bitcast i[[SZ]]* [[B_CADDR:%.+]] to i32*
+// CHECK-64:       store i32 %{{.+}}, i32* [[B_ADDR]],
+// CHECK-64:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_CADDR]],
+
+// CHECK-32:       store i32 %{{.+}}, i32* [[B_ADDR:%.+]],
+// CHECK-32:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_ADDR]],
+
+// We capture 2 VLA sizes in this target region
+// CHECK:       [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]]
+// CHECK:       [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2
+
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
+// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+// CHECK:       [[TRY]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT5]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX0:[0-9]+]]
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX0]]
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX0]]
+// CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX1:[0-9]+]]
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX1]]
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX1]]
+// CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX2:[0-9]+]]
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX2]]
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX2]]
+// CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX3:[0-9]+]]
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX3]]
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX3]]
+
+// The names below are not necessarily consistent with the names used for the
+// addresses above as some are repeated.
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR3:%.+]],
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   store i[[SZ]] 8, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
+
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+
+//
+// CHECK: define {{.*}}[[FSTATIC]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([5 x i[[SZ]]], [5 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT6]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 3
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 3
+// CHECK-DAG:   store i[[SZ]] [[VAL3:%.+]], i[[SZ]]* [[CBPADDR3:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL3]], i[[SZ]]* [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 4
+// CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 4
+// CHECK-DAG:   store [10 x i32]* %{{.+}}, [10 x i32]** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store [10 x i32]* %{{.+}}, [10 x i32]** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+//
+// CHECK: define {{.*}}[[FTEMPLATE]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET7]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT7]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT5]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions of the callees of bar().
+
+// CHECK:       define internal void [[HVT7]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_THIS:%.+]] = alloca [[S1]]*
+// CHECK:       [[LOCAL_B:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_C:%.+]] = alloca i16*
+// CHECK:       [[LOCAL_B_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store [[S1]]* [[ARG_THIS:%.+]], [[S1]]** [[LOCAL_THIS]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_B:%.+]], i[[SZ]]* [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i16* [[ARG_C:%.+]], i16** [[LOCAL_C]]
+// Store captures in the context.
+// CHECK-DAG:   [[REF_THIS:%.+]] = load [[S1]]*, [[S1]]** [[LOCAL_THIS]],
+// CHECK-64-DAG:[[CONV_BP:%.+]] = bitcast i[[SZ]]* [[LOCAL_B]] to i32*
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[REF_C:%.+]] = load i16*, i16** [[LOCAL_C]],
+
+// CHECK-64-DAG:[[CONV_B:%.+]] = load i32, i32* [[CONV_BP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_B_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_B]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_BV:%.+]] = load i32, i32* [[LOCAL_B]]
+// CHECK-32-DAG:store i32 [[LOCAL_BV]], i32* [[LOCAL_B_CASTED]], align
+// CHECK-DAG:   [[REF_B:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_B_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+
+// CHECK:       define internal void [[HVT6]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       alloca i[[SZ]],
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       alloca i[[SZ]],
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AAA:%.+]], i[[SZ]]* [[LOCAL_AAA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[CONV_AAAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA]] to i8*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK-DAG:   [[CONV_AAA:%.+]] = load i8, i8* [[CONV_AAAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA_CASTED]] to i8*
+// CHECK-DAG:   store i8 [[CONV_AAA]], i8* [[CONV]], align
+// CHECK-DAG:   [[REF_AAA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AAA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED6:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] {{.+}}, i[[SZ]] [[REF_AA]], i[[SZ]] [[REF_AAA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED6]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+// CHECK:       define internal void [[HVT5]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED7:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED7]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_codegen_registration.cpp b/test/OpenMP/target_teams_distribute_codegen_registration.cpp
new file mode 100644
index 000000000000..c031731c8065
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_codegen_registration.cpp
@@ -0,0 +1,451 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target teams distribute codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// Check that no target code is emmitted if no omptests flag was provided.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK-NTARGET
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK:    [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// CHECK-DAG: [[A1:@.+]] = internal global [[SA]]
+// CHECK-DAG: [[A2:@.+]] = global [[SA]]
+// CHECK-DAG: [[B1:@.+]] = global [[SB]]
+// CHECK-DAG: [[B2:@.+]] = global [[SB]]
+// CHECK-DAG: [[C1:@.+]] = internal global [[SC]]
+// CHECK-DAG: [[D1:@.+]] = global [[SD]]
+// CHECK-DAG: [[E1:@.+]] = global [[SE]]
+// CHECK-DAG: [[T1:@.+]] = global [[ST1]]
+// CHECK-DAG: [[T2:@.+]] = global [[ST2]]
+
+// CHECK-NTARGET-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-NTARGET-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-NTARGET-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-NTARGET-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-NTARGET-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-NTARGET-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-NTARGET-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-NTARGET-NOT: type { i8*, i8*, %
+// CHECK-NTARGET-NOT: type { i32, %
+
+// We have 7 target regions
+
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// TCHECK-NOT: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+
+// CHECK-NTARGET-NOT: private constant i8 0
+// CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
+
+// CHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// CHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// CHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// CHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// CHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// CHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// CHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// CHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// CHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// CHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// CHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// CHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// CHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// TCHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// TCHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// TCHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// TCHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// TCHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// TCHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// TCHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// TCHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// TCHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// TCHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// TCHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// TCHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// TCHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// We have 4 initializers, one for the 500 priority, another one for 501, or more for the default priority, and the last one for the offloading registration function.
+// CHECK: @llvm.global_ctors = appending global [4 x { i32, void ()*, i8* }] [
+// CHECK-SAME: { i32, void ()*, i8* } { i32 500, void ()* [[P500:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 501, void ()* [[P501:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 65535, void ()* [[PMAX:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+// CHECK-NTARGET: @llvm.global_ctors = appending global [3   x { i32, void ()*, i8* }] [
+
+extern int *R;
+
+struct SA {
+  int arr[4];
+  void foo() {
+    int a = *R;
+    a += 1;
+    *R = a;
+  }
+  SA() {
+    int a = *R;
+    a += 2;
+    *R = a;
+  }
+  ~SA() {
+    int a = *R;
+    a += 3;
+    *R = a;
+  }
+};
+
+struct SB {
+  int arr[8];
+  void foo() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 4;
+    *R = a;
+  }
+  SB() {
+    int a = *R;
+    a += 5;
+    *R = a;
+  }
+  ~SB() {
+    int a = *R;
+    a += 6;
+    *R = a;
+  }
+};
+
+struct SC {
+  int arr[16];
+  void foo() {
+    int a = *R;
+    a += 7;
+    *R = a;
+  }
+  SC() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 8;
+    *R = a;
+  }
+  ~SC() {
+    int a = *R;
+    a += 9;
+    *R = a;
+  }
+};
+
+struct SD {
+  int arr[32];
+  void foo() {
+    int a = *R;
+    a += 10;
+    *R = a;
+  }
+  SD() {
+    int a = *R;
+    a += 11;
+    *R = a;
+  }
+  ~SD() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 12;
+    *R = a;
+  }
+};
+
+struct SE {
+  int arr[64];
+  void foo() {
+    int a = *R;
+    #pragma omp target teams distribute if(target: 0)
+    for (int i = 0; i < 10; ++i)
+      a += 13;
+    *R = a;
+  }
+  SE() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 14;
+    *R = a;
+  }
+  ~SE() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 15;
+    *R = a;
+  }
+};
+
+template <int x>
+struct ST {
+  int arr[128 + x];
+  void foo() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 16 + x;
+    *R = a;
+  }
+  ST() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 17 + x;
+    *R = a;
+  }
+  ~ST() {
+    int a = *R;
+    #pragma omp target teams distribute
+    for (int i = 0; i < 10; ++i)
+      a += 18 + x;
+    *R = a;
+  }
+};
+
+// We have to make sure we us all the target regions:
+//CHECK-DAG: define internal void @[[NAME1]](
+//CHECK-DAG: call void @[[NAME1]](
+//CHECK-DAG: define internal void @[[NAME2]](
+//CHECK-DAG: call void @[[NAME2]](
+//CHECK-DAG: define internal void @[[NAME3]](
+//CHECK-DAG: call void @[[NAME3]](
+//CHECK-DAG: define internal void @[[NAME4]](
+//CHECK-DAG: call void @[[NAME4]](
+//CHECK-DAG: define internal void @[[NAME5]](
+//CHECK-DAG: call void @[[NAME5]](
+//CHECK-DAG: define internal void @[[NAME6]](
+//CHECK-DAG: call void @[[NAME6]](
+//CHECK-DAG: define internal void @[[NAME7]](
+//CHECK-DAG: call void @[[NAME7]](
+//CHECK-DAG: define internal void @[[NAME8]](
+//CHECK-DAG: call void @[[NAME8]](
+//CHECK-DAG: define internal void @[[NAME9]](
+//CHECK-DAG: call void @[[NAME9]](
+//CHECK-DAG: define internal void @[[NAME10]](
+//CHECK-DAG: call void @[[NAME10]](
+//CHECK-DAG: define internal void @[[NAME11]](
+//CHECK-DAG: call void @[[NAME11]](
+//CHECK-DAG: define internal void @[[NAME12]](
+//CHECK-DAG: call void @[[NAME12]](
+
+//TCHECK-DAG: define void @[[NAME1]](
+//TCHECK-DAG: define void @[[NAME2]](
+//TCHECK-DAG: define void @[[NAME3]](
+//TCHECK-DAG: define void @[[NAME4]](
+//TCHECK-DAG: define void @[[NAME5]](
+//TCHECK-DAG: define void @[[NAME6]](
+//TCHECK-DAG: define void @[[NAME7]](
+//TCHECK-DAG: define void @[[NAME8]](
+//TCHECK-DAG: define void @[[NAME9]](
+//TCHECK-DAG: define void @[[NAME10]](
+//TCHECK-DAG: define void @[[NAME11]](
+//TCHECK-DAG: define void @[[NAME12]](
+
+// CHECK-NTARGET-NOT: __tgt_target
+// CHECK-NTARGET-NOT: __tgt_register_lib
+// CHECK-NTARGET-NOT: __tgt_unregister_lib
+
+// TCHECK-NOT: __tgt_target
+// TCHECK-NOT: __tgt_register_lib
+// TCHECK-NOT: __tgt_unregister_lib
+
+// We have 2 initializers with priority 500
+//CHECK: define internal void [[P500]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 1 initializers with priority 501
+//CHECK: define internal void [[P501]](
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 6 initializers with default priority
+//CHECK: define internal void [[PMAX]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// Check registration and unregistration
+
+//CHECK:     define internal void @[[UNREGFN:.+]](i8*)
+//CHECK-SAME: comdat($[[REGFN]]) {
+//CHECK:     call i32 @__tgt_unregister_lib([[DSCTY]]* [[DESC]])
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_unregister_lib([[DSCTY]]*)
+
+//CHECK:     define linkonce hidden void @[[REGFN]](i8*)
+//CHECK-SAME: comdat {
+//CHECK:     call i32 @__tgt_register_lib([[DSCTY]]* [[DESC]])
+//CHECK:     call i32 @__cxa_atexit(void (i8*)* @[[UNREGFN]], i8* bitcast ([[DSCTY]]* [[DESC]] to i8*),
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_register_lib([[DSCTY]]*)
+
+static __attribute__((init_priority(500))) SA a1;
+SA a2;
+SB __attribute__((init_priority(500))) b1;
+SB __attribute__((init_priority(501))) b2;
+static SC c1;
+SD d1;
+SE e1;
+ST<100> t1;
+ST<1000> t2;
+
+
+int bar(int a){
+  int r = a;
+
+  a1.foo();
+  a2.foo();
+  b1.foo();
+  b2.foo();
+  c1.foo();
+  d1.foo();
+  e1.foo();
+  t1.foo();
+  t2.foo();
+
+  #pragma omp target teams distribute
+  for (int i = 0; i < 10; ++i)
+    ++r;
+
+  return r + *R;
+}
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_codegen_registration_naming.cpp b/test/OpenMP/target_teams_distribute_codegen_registration_naming.cpp
new file mode 100644
index 000000000000..c03466a52646
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_codegen_registration_naming.cpp
@@ -0,0 +1,68 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target teams distribute codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[CA:%.+]] = type { i32* }
+
+// CHECK: define {{.*}}i32 @[[NNAME:.+]](i32 {{.*}}%{{.+}})
+int nested(int a){
+  // CHECK: call void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME]]_l[[T1L:[0-9]+]](
+  #pragma omp target teams distribute
+  for (int i = 0; i < 10; ++i)
+    ++a;
+
+  // CHECK: call void @"[[LNAME:.+]]"([[CA]]*
+  auto F = [&](){
+    #pragma omp parallel
+    {
+      #pragma omp target teams distribute
+      for (int i = 0; i < 10; ++i)
+        ++a;
+    }
+  };
+
+  F();
+
+  return a;
+}
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T1L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME:.+]]_l[[T1L:[0-9]+]](
+
+// CHECK: define {{.*}}void @"[[LNAME]]"(
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.+}}[[PNAME:@.+]] to
+
+// CHECK: define {{.*}}void [[PNAME]](
+// CHECK: call void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L:[0-9]+]](
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME:.+]]_l[[T2L:[0-9]+]](
+
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+#endif
diff --git a/test/OpenMP/target_teams_distribute_collapse_codegen.cpp b/test/OpenMP/target_teams_distribute_collapse_codegen.cpp
new file mode 100644
index 000000000000..be99186e4681
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_collapse_codegen.cpp
@@ -0,0 +1,125 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X][Y];
+
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+    // CK1: call i32 @__tgt_target_teams(
+    // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target teams distribute collapse(2)
+    for(int i = 0; i < X; i++) {
+      for(int j = 0; j < Y; j++) {
+        a[i][j] = (T)0;
+      }
+    }
+    // CK1: define internal void @[[OFFL1]](
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+    // CK1: ret void
+
+    // CK1: define internal void @[[OUTL1]]({{.+}})
+    // discard loop variables not needed here
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: [[OMP_UB:%.+]] = alloca i32,
+    // CK1: store i32 56087, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void
+
+    return a[0][0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n, int m>
+int tmain(T argc) {
+  T a[n][m];
+  #pragma omp target teams distribute collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = (T)0;
+    }
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int m = 2;
+  int a[n][m];
+  #pragma omp target teams distribute collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = 0;
+    }
+  }
+  return tmain<int, 10, 2>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i64,
+// CK2: store i64 {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_8({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// discard loop variables not needed here
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i32,
+// CK2: store i32 {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/target_teams_distribute_depend_messages.cpp b/test/OpenMP/target_teams_distribute_depend_messages.cpp
index b3e733141070..fbdd49590f87 100644
--- a/test/OpenMP/target_teams_distribute_depend_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+#pragma omp target teams distribute depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute depend (in : main) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+#pragma omp target teams distribute depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_dist_schedule_codegen.cpp b/test/OpenMP/target_teams_distribute_dist_schedule_codegen.cpp
new file mode 100644
index 000000000000..0c9b2387debf
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_dist_schedule_codegen.cpp
@@ -0,0 +1,197 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target teams distribute
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target teams distribute dist_schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target teams distribute dist_schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+#pragma omp target teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target teams distribute dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target teams distribute dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target teams distribute dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target teams distribute dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp b/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..35abffa92600
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_firstprivate_codegen.cpp
@@ -0,0 +1,342 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+// CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target teams distribute firstprivate(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}})
+    // LAMBDA:  ret
+#pragma omp target teams distribute firstprivate(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}})
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]],
+    // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]],
+    // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_PRIV_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // LAMBDA: [[SIVAR_PRIV_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // skip loop vars
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_PRIV_ADDR]], {{.+}} [[G1_TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV_ADDR]],
+    // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP]]
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV_ADDR]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target teams distribute firstprivate(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]],
+// CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]],
+// CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]],
+// CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[TOFFL1:.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]({{.+}})
+// CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]],
+// CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]],
+// CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]],
+// CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_firstprivate_messages.cpp b/test/OpenMP/target_teams_distribute_firstprivate_messages.cpp
index 602e039802d1..934149e7c139 100644
--- a/test/OpenMP/target_teams_distribute_firstprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_firstprivate_messages.cpp
@@ -125,7 +125,8 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute lastprivate(argc), firstprivate(argc) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
   return 0;
diff --git a/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp b/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..b41a6016079b
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_lastprivate_codegen.cpp
@@ -0,0 +1,384 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target teams distribute lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target teams distribute lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA-64: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i64 [[G_IN:%.+]], i64 [[G1_IN:%.+]], i64 [[SVAR_IN:%.+]], i64 [[SFVAR_IN:%.+]])
+      // LAMBDA-32: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i32 [[SVAR_IN:%.+]], i32 [[SFVAR_IN:%.+]])
+      // LAMBDA-64: [[G_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA-64: [[G1_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA-64: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i32,
+      // LAMBDA-64: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca i32,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA-64: store i64 [[G_IN]], i64* [[G_PRIVATE_ADDR]],
+      // LAMBDA-32: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA-64: store i64 [[G1_IN]], i64* [[G1_PRIVATE_ADDR]],
+      // LAMBDA-32: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA-64: store i64 [[SVAR_IN]], i64* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA-32: store i32 [[SVAR_IN]], i32* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA-64: store i64 [[SFVAR_IN]], i64* [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA-32: store i32 [[SFVAR_IN]], i32* [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA-64: [[G_IN_REF:%.+]] = bitcast i64* [[G_PRIVATE_ADDR]] to double*
+      // LAMBDA-32: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA-64: [[G1_IN_REF:%.+]] = bitcast i64* [[G1_PRIVATE_ADDR]] to double*
+      // LAMBDA-64: store double* [[G1_IN_REF]], double** [[G1_IN_ADDR_REF:%.+]],
+      // LAMBDA-64: [[SVAR_IN_REF:%.+]] = bitcast i64* [[SVAR_PRIVATE_ADDR]] to i32*
+      // LAMBDA-64: [[SFVAR_IN_REF:%.+]] = bitcast i64* [[SFVAR_PRIVATE_ADDR]] to float*
+      // LAMBDA-32: [[SFVAR_IN_REF:%.+]] = bitcast i32* [[SFVAR_PRIVATE_ADDR]] to float*
+      // LAMBDA-64: [[G1_IN_REF:%.+]] = load double*, double** [[G1_IN_ADDR_REF]],
+      // LAMBDA-32: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA-64: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA-32: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+      [&]() {
+        // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+        // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+        g = 2;
+        g1 = 2;
+        svar = 4;
+        sfvar = 8.0;
+        // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+        // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+        // LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+        // LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+        // LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+        // LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+        // LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+        // LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+        // LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+        // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+        // LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+        // LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+        // LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target teams distribute lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]]([2 x i{{[0-9]+}}]* {{.+}}, i{{[0-9]+}} {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]]([2 x i{{[0-9]+}}]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_IN]], i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} [[S_VAR_IN]], i{{[0-9]+}}* [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[T_VAR_ADDR_REF:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-64: [[SVAR_ADDR_REF:%.+]] = bitcast i64* [[SVAR_ADDR]] to i32*
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK-64: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK-32: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK-64: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK-32: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]]([2 x i{{[0-9]+}}]* {{.+}}, i{{[0-9]+}} {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_IN1]], i{{[0-9]+}}* [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK-64: [[T_VAR_ADDR_REF1:%.+]] = bitcast i64* [[T_VAR_ADDR1]] to i32*
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK-64: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK-32: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/target_teams_distribute_lastprivate_messages.cpp b/test/OpenMP/target_teams_distribute_lastprivate_messages.cpp
index 8d594e94b09f..898291833d20 100644
--- a/test/OpenMP/target_teams_distribute_lastprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -215,10 +215,12 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+// expected-error@+1 {{firstprivate variable cannot be lastprivate}} expected-note@+1 {{defined as firstprivate}}
+#pragma omp target teams distribute firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute lastprivate(n) firstprivate(n) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/target_teams_distribute_loop_messages.cpp b/test/OpenMP/target_teams_distribute_loop_messages.cpp
index 441abcba0049..0a825ab1e609 100644
--- a/test/OpenMP/target_teams_distribute_loop_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -607,7 +607,8 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
-#pragma omp target teams distribute lastprivate(s) firstprivate(s)
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/target_teams_distribute_map_messages.cpp b/test/OpenMP/target_teams_distribute_map_messages.cpp
index 5d815de85b04..eafdccb66bc9 100644
--- a/test/OpenMP/target_teams_distribute_map_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -222,11 +222,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_misc_messages.c b/test/OpenMP/target_teams_distribute_misc_messages.c
index bf4df0894ad3..33496a70d140 100644
--- a/test/OpenMP/target_teams_distribute_misc_messages.c
+++ b/test/OpenMP/target_teams_distribute_misc_messages.c
@@ -285,12 +285,15 @@ void test_firstprivate() {
     ;
 
   int x, y, z;
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
 #pragma omp target teams distribute lastprivate(x) firstprivate(x)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 2 {{lastprivate variable cannot be firstprivate}} expected-note@+1 2 {{defined as lastprivate}}
 #pragma omp target teams distribute lastprivate(x, y) firstprivate(x, y)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 3 {{lastprivate variable cannot be firstprivate}} expected-note@+1 3 {{defined as lastprivate}}
 #pragma omp target teams distribute lastprivate(x, y, z) firstprivate(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_ast_print.cpp b/test/OpenMP/target_teams_distribute_parallel_for_ast_print.cpp
index 8df3c972dce0..f51c4fe60d7f 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_ast_print.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_ast_print.cpp
@@ -24,8 +24,10 @@ protected:
 public:
   S7(typename T::type v) : a(v) {
 #pragma omp target teams distribute parallel for private(a) private(this->a) private(T::a)
-    for (int k = 0; k < a.a; ++k)
+    for (int k = 0; k < a.a; ++k) {
       ++this->a.a;
+#pragma omp cancel for
+    }
   }
   S7 &operator=(S7 &s) {
 #pragma omp target teams distribute parallel for private(a) private(this->a)
@@ -43,6 +45,7 @@ public:
   }
 };
 // CHECK: #pragma omp target teams distribute parallel for private(this->a) private(this->a) private(T::a)
+// CHECK: #pragma omp cancel for
 // CHECK: #pragma omp target teams distribute parallel for private(this->a) private(this->a)
 // CHECK: #pragma omp target teams distribute parallel for default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 
@@ -113,10 +116,10 @@ T tmain(T argc) {
 // CHECK: #pragma omp target teams distribute parallel for default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
-#pragma omp target teams distribute parallel for linear(d)
+#pragma omp target teams distribute parallel for
   for (int k = 0; k < 10; ++k)
     e += d + argc;
-// CHECK: #pragma omp target teams distribute parallel for linear(d)
+// CHECK: #pragma omp target teams distribute parallel for
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
   return T();
@@ -158,10 +161,10 @@ int main (int argc, char **argv) {
 // CHECK: #pragma omp target teams distribute parallel for default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
-#pragma omp target teams distribute parallel for linear(d)
+#pragma omp target teams distribute parallel for
   for (int k = 0; k < 10; ++k)
     e += d + argc;
-// CHECK: #pragma omp target teams distribute parallel for linear(d)
+// CHECK: #pragma omp target teams distribute parallel for
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
   return (0);
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_depend_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_depend_messages.cpp
index ca7241a16c0a..c4688958e554 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_depend_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+#pragma omp target teams distribute parallel for depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for depend (in : main) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_messages.cpp
index d4e708308a02..0d8355047830 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_firstprivate_messages.cpp
@@ -124,7 +124,8 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute parallel for lastprivate(argc), firstprivate(argc) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute parallel for lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
   return 0;
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_messages.cpp
index d31acef0039d..84a6df28174d 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -216,10 +216,12 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute parallel for firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+// expected-error@+1 {{firstprivate variable cannot be lastprivate}} expected-note@+1 {{defined as firstprivate}}
+#pragma omp target teams distribute parallel for firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute parallel for lastprivate(n) firstprivate(n) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute parallel for lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_linear_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_linear_messages.cpp
deleted file mode 100644
index 82220dcf45c6..000000000000
--- a/test/OpenMP/target_teams_distribute_parallel_for_linear_messages.cpp
+++ /dev/null
@@ -1,248 +0,0 @@
-// RUN: %clang_cc1 -verify -fopenmp %s
-
-namespace X {
-  int x;
-};
-
-struct B {
-  static int ib; // expected-note {{'B::ib' declared here}}
-  static int bfoo() { return 8; }
-};
-
-int bfoo() { return 4; }
-
-int z;
-const int C1 = 1;
-const int C2 = 2;
-void test_linear_colons()
-{
-  int B = 0;
-
-#pragma omp target teams distribute parallel for linear(B:bfoo())
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(B:B::bfoo())
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(X::x : ::z)
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(B,::z, X::x)
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(::z)
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(B::bfoo()) // expected-error {{expected variable name}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target teams distribute parallel for linear(B::ib,B:C1+C2)
-  for (int i = 0; i < 10; ++i) ;
-}
-
-template<int L, class T, class N> T test_template(T* arr, N num) {
-  N i;
-  T sum = (T)0;
-  T ind2 = - num * L; // expected-note {{'ind2' defined here}}
-
-#pragma omp target teams distribute parallel for linear(ind2:L) // expected-error {{argument of a linear clause should be of integral or pointer type}}
-  for (i = 0; i < num; ++i) {
-    T cur = arr[(int)ind2];
-    ind2 += L;
-    sum += cur;
-  }
-  return T();
-}
-
-template<int LEN> int test_warn() {
-  int ind2 = 0;
-#pragma omp target teams distribute parallel for linear(ind2:LEN) // expected-warning {{zero linear step (ind2 should probably be const)}}
-  for (int i = 0; i < 100; i++) {
-    ind2 += LEN;
-  }
-  return ind2;
-}
-
-struct S1; // expected-note 2 {{declared here}} expected-note 2 {{forward declaration of 'S1'}}
-extern S1 a;
-class S2 {
-  mutable int a;
-public:
-  S2():a(0) { }
-};
-const S2 b; // expected-note 2 {{'b' defined here}}
-const S2 ba[5];
-class S3 {
-  int a;
-public:
-  S3():a(0) { }
-};
-const S3 ca[5];
-class S4 {
-  int a;
-  S4();
-public:
-  S4(int v):a(v) { }
-};
-class S5 {
-  int a;
-  S5():a(0) {}
-public:
-  S5(int v):a(v) { }
-};
-
-S3 h;
-#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
-
-template<class I, class C> int foomain(I argc, C **argv) {
-  I e(4);
-  I g(5);
-  int i;
-  int &j = i;
-
-#pragma omp target teams distribute parallel for linear // expected-error {{expected '(' after 'linear'}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear ( // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear () // expected-error {{expected expression}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc : 5)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (S1) // expected-error {{'S1' does not refer to a value}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (a, b:B::ib) // expected-error {{linear variable with incomplete type 'S1'}} expected-error {{const-qualified variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argv[1]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(e, g)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target teams distribute parallel for linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target teams distribute parallel for linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target teams distribute parallel for linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-  return 0;
-}
-
-namespace A {
-double x;
-#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
-}
-namespace C {
-using A::x;
-}
-
-int main(int argc, char **argv) {
-  double darr[100];
-  // expected-note@+1 {{in instantiation of function template specialization 'test_template<-4, double, int>' requested here}}
-  test_template<-4>(darr, 4);
-  // expected-note@+1 {{in instantiation of function template specialization 'test_warn<0>' requested here}}
-  test_warn<0>();
-
-  S4 e(4); // expected-note {{'e' defined here}}
-  S5 g(5); // expected-note {{'g' defined here}}
-  int i;
-  int &j = i;
-
-#pragma omp target teams distribute parallel for linear // expected-error {{expected '(' after 'linear'}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear ( // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear () // expected-error {{expected expression}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argc)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (S1) // expected-error {{'S1' does not refer to a value}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-
-#pragma omp target teams distribute parallel for linear (a, b) // expected-error {{linear variable with incomplete type 'S1'}} expected-error {{const-qualified variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear (argv[1]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(e, g) // expected-error {{argument of a linear clause should be of integral or pointer type, not 'S4'}} expected-error {{argument of a linear clause should be of integral or pointer type, not 'S5'}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(h, C::x) // expected-error 2 {{threadprivate or thread local variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp parallel
-{
-  int i;
-#pragma omp target teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(i : 4)
-  for (int k = 0; k < argc; ++k) { ++k; i += 4; }
-}
-
-#pragma omp target teams distribute parallel for linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
-  return 0;
-}
-
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_loop_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_loop_messages.cpp
index d912737a2146..57ab273c4601 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_loop_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -605,7 +605,8 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
-#pragma omp target teams distribute parallel for lastprivate(s) firstprivate(s)
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute parallel for lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_map_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_map_messages.cpp
index 326db95f90f8..1fb466ee5518 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_map_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -222,11 +222,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_messages.cpp
index cb76ef2e0124..15135838e981 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_messages.cpp
@@ -27,7 +27,7 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for } // expected-warning {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}}
   for (int i = 0; i < argc; ++i)
     foo();
-#pragma omp target teams distribute parallel for
+#pragma omp target teams distribute parallel for linear(argc) // expected-error {{unexpected OpenMP clause 'linear' in directive '#pragma omp target teams distribute parallel for'}}
   for (int i = 0; i < argc; ++i)
     foo();
 // expected-warning@+1 {{extra tokens at the end of '#pragma omp target teams distribute parallel for' are ignored}}
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_misc_messages.c b/test/OpenMP/target_teams_distribute_parallel_for_misc_messages.c
index 769be6c10c3a..c48a2dfa9e68 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_misc_messages.c
+++ b/test/OpenMP/target_teams_distribute_parallel_for_misc_messages.c
@@ -285,12 +285,15 @@ void test_firstprivate() {
     ;
 
   int x, y, z;
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
 #pragma omp target teams distribute parallel for lastprivate(x) firstprivate(x)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 2 {{lastprivate variable cannot be firstprivate}} expected-note@+1 2 {{defined as lastprivate}}
 #pragma omp target teams distribute parallel for lastprivate(x, y) firstprivate(x, y)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 3 {{lastprivate variable cannot be firstprivate}} expected-note@+1 3 {{defined as lastprivate}}
 #pragma omp target teams distribute parallel for lastprivate(x, y, z) firstprivate(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_reduction_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_reduction_messages.cpp
index ba9586a1cef8..59b35417d401 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_reduction_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_reduction_messages.cpp
@@ -19,9 +19,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_ast_print.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_ast_print.cpp
index 29a0c6d36333..af588f1b81d9 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_ast_print.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_ast_print.cpp
@@ -28,8 +28,9 @@ public:
       ++this->a.a;
   }
   S7 &operator=(S7 &s) {
-#pragma omp target teams distribute parallel for simd private(a) private(this->a)
-    for (int k = 0; k < s.a.a; ++k)
+    int k;
+#pragma omp target teams distribute parallel for simd private(a) private(this->a) linear(k)
+    for (k = 0; k < s.a.a; ++k)
       ++s.a.a;
 
     foo();
@@ -53,7 +54,7 @@ public:
   }
 };
 // CHECK: #pragma omp target teams distribute parallel for simd private(this->a) private(this->a) private(T::a)
-// CHECK: #pragma omp target teams distribute parallel for simd private(this->a) private(this->a)
+// CHECK: #pragma omp target teams distribute parallel for simd private(this->a) private(this->a) linear(k)
 // CHECK: #pragma omp target teams distribute parallel for simd default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK: #pragma omp target teams distribute parallel for simd simdlen(slen1) safelen(slen2) aligned(arr: alen)
 
@@ -135,10 +136,10 @@ T tmain(T argc) {
 // CHECK: #pragma omp target teams distribute parallel for simd default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
-#pragma omp target teams distribute parallel for simd simdlen(clen-1) linear(d)
+#pragma omp target teams distribute parallel for simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
-// CHECK: #pragma omp target teams distribute parallel for simd simdlen(clen - 1) linear(d)
+// CHECK: #pragma omp target teams distribute parallel for simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target teams distribute parallel for simd safelen(clen-1) aligned(arr:alen)
@@ -186,10 +187,10 @@ int main (int argc, char **argv) {
 // CHECK: #pragma omp target teams distribute parallel for simd default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
-#pragma omp target teams distribute parallel for simd simdlen(clen-1) linear(d)
+#pragma omp target teams distribute parallel for simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
-// CHECK: #pragma omp target teams distribute parallel for simd simdlen(clen - 1) linear(d)
+// CHECK: #pragma omp target teams distribute parallel for simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target teams distribute parallel for simd safelen(clen-1) aligned(arr:N+6)
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_depend_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_depend_messages.cpp
index 2dc9979cdc1c..4a9bf01e40bf 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_depend_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+#pragma omp target teams distribute parallel for simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for simd depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for simd depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd depend (in : main) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for simd depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+#pragma omp target teams distribute parallel for simd depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_messages.cpp
index d26044abc81b..4790b4bffdaf 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_firstprivate_messages.cpp
@@ -124,7 +124,8 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for simd firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute parallel for simd lastprivate(argc), firstprivate(argc) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute parallel for simd lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
   return 0;
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_messages.cpp
index 954830eb19f5..156f1a66c191 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -216,10 +216,12 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for simd lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute parallel for simd firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+// expected-error@+1 {{firstprivate variable cannot be lastprivate}} expected-note@+1 {{defined as firstprivate}}
+#pragma omp target teams distribute parallel for simd firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute parallel for simd lastprivate(n) firstprivate(n) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute parallel for simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_linear_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_linear_messages.cpp
index 24c432ade9e4..257244ede444 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_linear_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_linear_messages.cpp
@@ -18,33 +18,42 @@ void test_linear_colons()
 {
   int B = 0;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(B:bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(B:B::bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(X::x : ::z)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(B,::z, X::x)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(::z)
   for (int i = 0; i < 10; ++i) ;
 
 #pragma omp target teams distribute parallel for simd linear(B::bfoo()) // expected-error {{expected variable name}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(B::ib,B:C1+C2)
   for (int i = 0; i < 10; ++i) ;
 }
@@ -65,6 +74,7 @@ template<int L, class T, class N> T test_template(T* arr, N num) {
 
 template<int LEN> int test_warn() {
   int ind2 = 0;
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
   #pragma omp target teams distribute parallel for simd linear(ind2:LEN) // expected-warning {{zero linear step (ind2 should probably be const)}}
   for (int i = 0; i < 100; i++) {
     ind2 += LEN;
@@ -118,15 +128,18 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp target teams distribute parallel for simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
 #pragma omp target teams distribute parallel for simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear (argc : 5)
   for (int k = 0; k < argc; ++k) ++k;
 
@@ -139,33 +152,17 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp target teams distribute parallel for simd linear (argv[1]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(e, g)
   for (int k = 0; k < argc; ++k) ++k;
 
 #pragma omp target teams distribute parallel for simd linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear(i)
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target teams distribute parallel for simd linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target teams distribute parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target teams distribute parallel for simd linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target teams distribute parallel for simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
   return 0;
 }
 
@@ -198,15 +195,18 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
 #pragma omp target teams distribute parallel for simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute parallel for simd linear (argc)
   for (int k = 0; k < argc; ++k) ++k;
 
@@ -226,22 +226,6 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute parallel for simd linear(h, C::x) // expected-error 2 {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int i;
-    #pragma omp target teams distribute parallel for simd linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
-
-    #pragma omp target teams distribute parallel for simd linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
-  }
-
-#pragma omp target teams distribute parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute parallel for simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
   foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
   return 0;
 }
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_loop_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_loop_messages.cpp
index 0ed57baa5c3b..a16dec190dd2 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_loop_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -608,7 +608,8 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
-#pragma omp target teams distribute parallel for simd lastprivate(s) firstprivate(s)
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute parallel for simd lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_map_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_map_messages.cpp
index 8f9c3e7e073c..b230d9ab24fe 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_map_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -222,11 +222,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute parallel for simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute parallel for simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute parallel for simd map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_misc_messages.c b/test/OpenMP/target_teams_distribute_parallel_for_simd_misc_messages.c
index d444844be5b4..1ea02f6d2b57 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_misc_messages.c
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_misc_messages.c
@@ -285,15 +285,22 @@ void test_firstprivate() {
     ;
 
   int x, y, z;
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
 #pragma omp target teams distribute parallel for simd lastprivate(x) firstprivate(x)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 2 {{lastprivate variable cannot be firstprivate}} expected-note@+1 2 {{defined as lastprivate}}
 #pragma omp target teams distribute parallel for simd lastprivate(x, y) firstprivate(x, y)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 3 {{lastprivate variable cannot be firstprivate}} expected-note@+1 3 {{defined as lastprivate}}
 #pragma omp target teams distribute parallel for simd lastprivate(x, y, z) firstprivate(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 {{the value of 'simdlen' parameter must be less than or equal to the value of the 'safelen' parameter}}
+#pragma omp target teams distribute parallel for simd simdlen(64) safelen(8)
+  for (i = 0; i < 16; ++i)
+    ;
 }
 
 void test_loop_messages() {
diff --git a/test/OpenMP/target_teams_distribute_parallel_for_simd_reduction_messages.cpp b/test/OpenMP/target_teams_distribute_parallel_for_simd_reduction_messages.cpp
index cadccbf5cda3..63744c77fb3a 100644
--- a/test/OpenMP/target_teams_distribute_parallel_for_simd_reduction_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_parallel_for_simd_reduction_messages.cpp
@@ -19,9 +19,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_teams_distribute_private_codegen.cpp b/test/OpenMP/target_teams_distribute_private_codegen.cpp
new file mode 100644
index 000000000000..7c88c01c4ab3
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_private_codegen.cpp
@@ -0,0 +1,233 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target teams distribute private(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]]()
+    // LAMBDA:  ret
+#pragma omp target teams distribute private(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]]()
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target teams distribute private(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]()
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[SIVAR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0,
+// CHECK: call void @[[TOFFL1:.+]]()
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_reduction_codegen.cpp b/test/OpenMP/target_teams_distribute_reduction_codegen.cpp
new file mode 100644
index 000000000000..188b75306453
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_reduction_codegen.cpp
@@ -0,0 +1,211 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target teams distribute reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target teams distribute reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i32*{{.+}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR:%.+]] = load i32*, i32** [[SIVAR_ADDR]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}}*{{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target teams distribute reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i32* {{.+}})
+// CHECK: [[RES:%.+]] = call{{.*}} i32 @[[TMAIN_INT:[^(]+]]()
+// CHECK: ret i32 [[RES]]
+
+// CHECK: define{{.*}} void @[[OFFL1]](i32*{{.+}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}}** [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_LOAD:%.+]] = load i32*, i32** [[SIVAR_ADDR]],
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_LOAD]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, i32*{{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i32,
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load i32*, i32** [[SIVAR_ADDR]],
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}}* {{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i32*{{.+}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR:%.+]] = load i32*, i32** [[TVAR_ADDR]],
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}}*{{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load i32*, i32** [[TVAR_ADDR]],
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_reduction_messages.cpp b/test/OpenMP/target_teams_distribute_reduction_messages.cpp
index d52ee076130b..3cfb08651a61 100644
--- a/test/OpenMP/target_teams_distribute_reduction_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_reduction_messages.cpp
@@ -24,9 +24,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_teams_distribute_simd_ast_print.cpp b/test/OpenMP/target_teams_distribute_simd_ast_print.cpp
index 77732ef56011..31b9ad0c5fc1 100644
--- a/test/OpenMP/target_teams_distribute_simd_ast_print.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_ast_print.cpp
@@ -28,8 +28,9 @@ public:
       ++this->a.a;
   }
   S7 &operator=(S7 &s) {
-#pragma omp target teams distribute simd private(a) private(this->a)
-    for (int k = 0; k < s.a.a; ++k)
+    int k;
+#pragma omp target teams distribute simd private(a) private(this->a) linear(k)
+    for (k = 0; k < s.a.a; ++k)
       ++s.a.a;
     return *this;
   }
@@ -50,7 +51,7 @@ public:
   }
 };
 // CHECK: #pragma omp target teams distribute simd private(this->a) private(this->a) private(T::a)
-// CHECK: #pragma omp target teams distribute simd private(this->a) private(this->a)
+// CHECK: #pragma omp target teams distribute simd private(this->a) private(this->a) linear(k)
 // CHECK: #pragma omp target teams distribute simd private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK: #pragma omp target teams distribute simd simdlen(slen1) safelen(slen2) aligned(arr: alen)
 // CHECK: #pragma omp target teams distribute simd private(this->a) private(this->a) private(this->S::a)
@@ -129,10 +130,10 @@ T tmain(T argc) {
 // CHECK: #pragma omp target teams distribute simd private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
-#pragma omp target teams distribute simd simdlen(clen-1) linear(d)
+#pragma omp target teams distribute simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
-// CHECK: #pragma omp target teams distribute simd simdlen(clen - 1) linear(d)
+// CHECK: #pragma omp target teams distribute simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target teams distribute simd safelen(clen-1) aligned(arr:alen)
@@ -180,10 +181,10 @@ int main (int argc, char **argv) {
 // CHECK: #pragma omp target teams distribute simd private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
-#pragma omp target teams distribute simd simdlen(clen-1) linear(d)
+#pragma omp target teams distribute simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
-// CHECK: #pragma omp target teams distribute simd simdlen(clen - 1) linear(d)
+// CHECK: #pragma omp target teams distribute simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target teams distribute simd safelen(clen-1) aligned(arr:N+6)
diff --git a/test/OpenMP/target_teams_distribute_simd_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_codegen.cpp
new file mode 100644
index 000000000000..7fb76d9198db
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_codegen.cpp
@@ -0,0 +1,824 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// Test target codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-64
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix TCHECK --check-prefix TCHECK-32
+// expected-no-diagnostics
+
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: %ident_t = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC:@.+]] = private unnamed_addr constant %ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+// CHECK-DAG: [[TT:%.+]] = type { i64, i8 }
+// CHECK-DAG: [[S1:%.+]] = type { double }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK: [[ENTTY:%.+]] = type { i8*, i8*, i{{32|64}}, i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// We have 8 target regions, but only 7 that actually will generate offloading
+// code, only 6 will have mapped arguments, and only 4 have all-constant map
+// sizes.
+
+// CHECK-DAG: [[SIZET:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 2, i[[SZ]] 4, i[[SZ]] 4]
+// CHECK-DAG: [[MAPT:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 288]
+// CHECK-DAG: [[SIZET2:@.+]] = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 2]
+// CHECK-DAG: [[MAPT2:@.+]] = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: [[SIZET3:@.+]] = private unnamed_addr constant [2 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2]
+// CHECK-DAG: [[MAPT3:@.+]] = private unnamed_addr constant [2 x i64] [i64 288, i64 288]
+// CHECK-DAG: [[MAPT4:@.+]] = private unnamed_addr constant [9 x i64] [i64 288, i64 547, i64 288, i64 547, i64 547, i64 288, i64 288, i64 547, i64 547]
+// CHECK-DAG: [[MAPT5:@.+]] = private unnamed_addr constant [5 x i64] [i64 547, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET6:@.+]] = private unnamed_addr constant [5 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 1, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT6:@.+]] = private unnamed_addr constant [5 x i64] [i64 288, i64 288, i64 288, i64 288, i64 547]
+// CHECK-DAG: [[SIZET7:@.+]] = private unnamed_addr constant [3 x i[[SZ]]] [i[[SZ]] 4, i[[SZ]] 2, i[[SZ]] 40]
+// CHECK-DAG: [[MAPT7:@.+]] = private unnamed_addr constant [3 x i64] [i64 288, i64 288, i64 547]
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+// CHECK-DAG: @{{.*}} = private constant i8 0
+
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK: @{{.+}} = constant [[ENTTY]]
+// TCHECK-NOT: @{{.+}} = constant [[ENTTY]]
+
+// Check if offloading descriptor is created.
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// Check target registration is registered as a Ctor.
+// CHECK: appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+
+template<typename tx, typename ty>
+struct TT{
+  tx X;
+  ty Y;
+};
+
+int global;
+
+// CHECK: define {{.*}}[[FOO:@.+]](
+int foo(int n) {
+  int a = 0;
+  short aa = 0;
+  float b[10];
+  float bn[n];
+  double c[5][10];
+  double cn[5][n];
+  TT<long long, char> d;
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams_nowait(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT]], i32 0, i32 0), i32 {{[^,]+}}, i32 {{[^)]+}})
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BPR]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PR]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[CBPADDR2:%.+]] = bitcast i8** [[BPADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2:%.+]] = bitcast i8** [[PADDR2]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR2]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR2]]
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT0:@.+]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target teams distribute simd num_teams(a) thread_limit(a) firstprivate(aa) simdlen(16) nowait
+  for (int i = 0; i < 10; ++i) {
+  }
+
+  // CHECK:       call void [[HVT1:@.+]](i[[SZ]] {{[^,]+}})
+  #pragma omp target teams distribute simd if(target: 0) safelen(32) linear(a)
+  for (a = 0; a < 10; ++a) {
+    a += 1;
+  }
+
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** [[BP:%[^,]+]], i8** [[P:%[^,]+]], i[[SZ]]* getelementptr inbounds ([1 x i[[SZ]]], [1 x i[[SZ]]]* [[SIZET2]], i32 0, i32 0), i64* getelementptr inbounds ([1 x i64], [1 x i64]* [[MAPT2]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[BP]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[P]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[BPR]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[PR]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT2:@.+]](i[[SZ]] {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target teams distribute simd if(target: 1)
+  for (int i = 0; i < 10; ++i) {
+    aa += 1;
+  }
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 10
+  // CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+  // CHECK:       [[IFTHEN]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([2 x i[[SZ]]], [2 x i[[SZ]]]* [[SIZET3]], i32 0, i32 0), i64* getelementptr inbounds ([2 x i64], [2 x i64]* [[MAPT3]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 0
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 0
+  // CHECK-DAG:   [[CBPADDR0:%.+]] = bitcast i8** [[BPADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0:%.+]] = bitcast i8** [[PADDR0]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR0]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR0]]
+
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[BP]], i32 0, i32 1
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [2 x i8*], [2 x i8*]* [[P]], i32 0, i32 1
+  // CHECK-DAG:   [[CBPADDR1:%.+]] = bitcast i8** [[BPADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1:%.+]] = bitcast i8** [[PADDR1]] to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CBPADDR1]]
+  // CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[CPADDR1]]
+  // CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT3:@.+]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  // CHECK-NEXT:  br label %[[IFEND:.+]]
+  // CHECK:       [[IFELSE]]
+  // CHECK:       call void [[HVT3]]({{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[IFEND]]
+  // CHECK:       [[IFEND]]
+  #pragma omp target teams distribute simd if(target: n>10)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+    aa += 1;
+  }
+
+  // We capture 3 VLA sizes in this target region
+  // CHECK-64:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-64:       [[A_ADDR:%.+]] = bitcast i[[SZ]]* [[A_CADDR:%.+]] to i32*
+  // CHECK-64:       store i32 [[A_VAL]], i32* [[A_ADDR]],
+  // CHECK-64:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK-32:       [[A_VAL:%.+]] = load i32, i32* %{{.+}},
+  // CHECK-32:       store i32 [[A_VAL]], i32* [[A_CADDR:%.+]],
+  // CHECK-32:       [[A_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CADDR]],
+
+  // CHECK:       [[BNSIZE:%.+]] = mul nuw i[[SZ]] [[VLA0:%.+]], 4
+  // CHECK:       [[CNELEMSIZE2:%.+]] = mul nuw i[[SZ]] 5, [[VLA1:%.+]]
+  // CHECK:       [[CNSIZE:%.+]] = mul nuw i[[SZ]] [[CNELEMSIZE2]], 8
+
+  // CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 20
+  // CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+  // CHECK:       [[TRY]]
+  // CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 9, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([9 x i64], [9 x i64]* [[MAPT4]], i32 0, i32 0), i32 0, i32 0)
+  // CHECK-DAG:   [[BPR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[PR]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P:%[^,]+]], i32 0, i32 0
+  // CHECK-DAG:   [[SR]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S:%[^,]+]], i32 0, i32 0
+
+  // CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX0:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX0]]
+  // CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX1:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX1]]
+  // CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX2:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX2]]
+  // CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX3:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX3]]
+  // CHECK-DAG:   [[SADDR4:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX4:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX4]]
+  // CHECK-DAG:   [[SADDR5:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX5:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[PADDR5:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX5]]
+  // CHECK-DAG:   [[SADDR6:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX6:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[PADDR6:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX6]]
+  // CHECK-DAG:   [[SADDR7:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX7:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[PADDR7:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX7]]
+  // CHECK-DAG:   [[SADDR8:%.+]] = getelementptr inbounds [9 x i[[SZ]]], [9 x i[[SZ]]]* [[S]], i32 0, i32 [[IDX8:[0-9]+]]
+  // CHECK-DAG:   [[BPADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[BP]], i32 0, i32 [[IDX8]]
+  // CHECK-DAG:   [[PADDR8:%.+]] = getelementptr inbounds [9 x i8*], [9 x i8*]* [[P]], i32 0, i32 [[IDX8]]
+
+  // The names below are not necessarily consistent with the names used for the
+  // addresses above as some are repeated.
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+  // CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CBPADDR1:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[VLA1]], i[[SZ]]* [[CPADDR1:%.+]],
+  // CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CBPADDR2:%.+]],
+  // CHECK-DAG:   store i[[SZ]] 5, i[[SZ]]* [[CPADDR2:%.+]],
+  // CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CBPADDR3:%.+]],
+  // CHECK-DAG:   store i[[SZ]] [[A_CVAL]], i[[SZ]]* [[CPADDR3:%.+]],
+  // CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+  // CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CBPADDR4:%.+]],
+  // CHECK-DAG:   store [10 x float]* %{{.+}}, [10 x float]** [[CPADDR4:%.+]],
+  // CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x float]**
+  // CHECK-DAG:   store i[[SZ]] 40, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CBPADDR5:%.+]],
+  // CHECK-DAG:   store float* %{{.+}}, float** [[CPADDR5:%.+]],
+  // CHECK-DAG:   [[CBPADDR5]] = bitcast i8** {{%[^,]+}} to  float**
+  // CHECK-DAG:   [[CPADDR5]] = bitcast i8** {{%[^,]+}} to float**
+  // CHECK-DAG:   store i[[SZ]] [[BNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CBPADDR6:%.+]],
+  // CHECK-DAG:   store [5 x [10 x double]]* %{{.+}}, [5 x [10 x double]]** [[CPADDR6:%.+]],
+  // CHECK-DAG:   [[CBPADDR6]] = bitcast i8** {{%[^,]+}} to  [5 x [10 x double]]**
+  // CHECK-DAG:   [[CPADDR6]] = bitcast i8** {{%[^,]+}} to [5 x [10 x double]]**
+  // CHECK-DAG:   store i[[SZ]] 400, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CBPADDR7:%.+]],
+  // CHECK-DAG:   store double* %{{.+}}, double** [[CPADDR7:%.+]],
+  // CHECK-DAG:   [[CBPADDR7]] = bitcast i8** {{%[^,]+}} to  double**
+  // CHECK-DAG:   [[CPADDR7]] = bitcast i8** {{%[^,]+}} to double**
+  // CHECK-DAG:   store i[[SZ]] [[CNSIZE]], i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CBPADDR8:%.+]],
+  // CHECK-DAG:   store [[TT]]* %{{.+}}, [[TT]]** [[CPADDR8:%.+]],
+  // CHECK-DAG:   [[CBPADDR8]] = bitcast i8** {{%[^,]+}} to  [[TT]]**
+  // CHECK-DAG:   [[CPADDR8]] = bitcast i8** {{%[^,]+}} to [[TT]]**
+  // CHECK-DAG:   store i[[SZ]] {{12|16}}, i[[SZ]]* {{%[^,]+}}
+
+  // CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+  // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+  // CHECK:       [[FAIL]]
+  // CHECK:       call void [[HVT4:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+  // CHECK-NEXT:  br label %[[END]]
+  // CHECK:       [[END]]
+  #pragma omp target teams distribute simd if(target: n>20) aligned(b)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+    b[2] += 1.0;
+    bn[3] += 1.0;
+    c[1][2] += 1.0;
+    cn[1][3] += 1.0;
+    d.X += 1;
+    d.Y += 1;
+  }
+
+  return a;
+}
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions in foo().
+
+// CHECK:       define internal void [[HVT0]](i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^,]+}}, i[[SZ]] {{[^)]+}})
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] {{[^)]+}})
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] {{[^)]+}})
+// CHECK:       alloca i16,
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+
+// CHECK:       define internal void [[HVT1]](i[[SZ]] %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    [[AA:%.+]] = load i32, i32* [[AA_CADDR]], align
+// CHECK-32:    [[AA:%.+]] = load i32, i32* [[AA_ADDR]], align
+// CHECK-64:    [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i32*
+// CHECK-64:    store i32 [[AA]], i32* [[AA_C]], align
+// CHECK-32:    store i32 [[AA]], i32* [[AA_CASTED]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED1:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED1]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64:    [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i32*
+// CHECK-64:    store i32 10, i32* [[AA_CADDR]], align
+// CHECK-32:    store i32 10, i32* [[AA_ADDR]], align
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT2]](i[[SZ]] %{{.+}})
+// Create stack storage and store argument in there.
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK:       store i16 [[AA]], i16* [[AA_C]], align
+// CHECK:       [[PARAM:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]])* [[OMP_OUTLINED2:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED2]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}})
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK:       [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT3]]
+// Create stack storage and store argument in there.
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[A_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_CASTED:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK-64-DAG:[[A:%.+]] = load i32, i32* [[A_CADDR]], align
+// CHECK-32-DAG:[[A:%.+]] = load i32, i32* [[A_ADDR]], align
+// CHECK-64-DAG:[[A_C:%.+]] = bitcast i[[SZ]]* [[A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[A]], i32* [[A_C]], align
+// CHECK-32-DAG:store i32 [[A]], i32* [[A_CASTED]], align
+// CHECK-DAG:   [[AA:%.+]] = load i16, i16* [[AA_CADDR]], align
+// CHECK-DAG:   [[AA_C:%.+]] = bitcast i[[SZ]]* [[AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[AA]], i16* [[AA_C]], align
+// CHECK-DAG:   [[PARAM1:%.+]] = load i[[SZ]], i[[SZ]]* [[A_CASTED]], align
+// CHECK-DAG:   [[PARAM2:%.+]] = load i[[SZ]], i[[SZ]]* [[AA_CASTED]], align
+// CHECK-DAG:   call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 2, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]])* [[OMP_OUTLINED3:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[PARAM1]], i[[SZ]] [[PARAM2]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED3]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}})
+// CHECK:       [[A_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK:       [[AA_ADDR:%.+]] = alloca i[[SZ]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[A_ADDR]], align
+// CHECK-DAG:   store i[[SZ]] %{{.+}}, i[[SZ]]* [[AA_ADDR]], align
+// CHECK-64-DAG:[[A_CADDR:%.+]] = bitcast i[[SZ]]* [[A_ADDR]] to i32*
+// CHECK-DAG:   [[AA_CADDR:%.+]] = bitcast i[[SZ]]* [[AA_ADDR]] to i16*
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal void [[HVT4]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x float]*
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_BN:%.+]] = alloca float*
+// CHECK:       [[LOCAL_C:%.+]] = alloca [5 x [10 x double]]*
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA3:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_CN:%.+]] = alloca double*
+// CHECK:       [[LOCAL_D:%.+]] = alloca [[TT]]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store [10 x float]* [[ARG_B:%.+]], [10 x float]** [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store float* [[ARG_BN:%.+]], float** [[LOCAL_BN]]
+// CHECK-DAG:   store [5 x [10 x double]]* [[ARG_C:%.+]], [5 x [10 x double]]** [[LOCAL_C]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA3:%.+]], i[[SZ]]* [[LOCAL_VLA3]]
+// CHECK-DAG:   store double* [[ARG_CN:%.+]], double** [[LOCAL_CN]]
+// CHECK-DAG:   store [[TT]]* [[ARG_D:%.+]], [[TT]]** [[LOCAL_D]]
+
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x float]*, [10 x float]** [[LOCAL_B]],
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[REF_BN:%.+]] = load float*, float** [[LOCAL_BN]],
+// CHECK-DAG:   [[REF_C:%.+]] = load [5 x [10 x double]]*, [5 x [10 x double]]** [[LOCAL_C]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[VAL_VLA3:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA3]],
+// CHECK-DAG:   [[REF_CN:%.+]] = load double*, double** [[LOCAL_CN]],
+// CHECK-DAG:   [[REF_D:%.+]] = load [[TT]]*, [[TT]]** [[LOCAL_D]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 9, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], [10 x float]*, i[[SZ]], float*, [5 x [10 x double]]*, i[[SZ]], i[[SZ]], double*, [[TT]]*)* [[OMP_OUTLINED4:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], [10 x float]* [[REF_B]], i[[SZ]] [[VAL_VLA1]], float* [[REF_BN]], [5 x [10 x double]]* [[REF_C]], i[[SZ]] [[VAL_VLA2]], i[[SZ]] [[VAL_VLA3]], double* [[REF_CN]], [[TT]]* [[REF_D]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED4]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, [10 x float]* {{.+}}, i[[SZ]] %{{.+}}, float* {{.+}}, [5 x [10 x double]]* {{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, double* {{.+}}, [[TT]]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+template<typename tx>
+tx ftemplate(int n) {
+  tx a = 0;
+  short aa = 0;
+  tx b[10];
+
+  #pragma omp target teams distribute simd if(target: n>40)
+  for (int i = 0; i < 10; ++i) {
+    a += 1;
+    aa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+static
+int fstatic(int n) {
+  int a = 0;
+  short aa = 0;
+  char aaa = 0;
+  int b[10];
+
+  #pragma omp target teams distribute simd if(target: n>50)
+  for (int i = a; i < n; ++i) {
+    a += 1;
+    aa += 1;
+    aaa += 1;
+    b[2] += 1;
+  }
+
+  return a;
+}
+
+struct S1 {
+  double a;
+
+  int r1(int n){
+    int b = n+1;
+    short int c[2][n];
+
+    #pragma omp target teams distribute simd if(target: n>60)
+    for (int i = 0; i < 10; ++i) {
+      this->a = (double)b + 1.5;
+      c[1][1] = ++a;
+    }
+
+    return c[1][1] + (int)b;
+  }
+};
+
+// CHECK: define {{.*}}@{{.*}}bar{{.*}}
+int bar(int n){
+  int a = 0;
+
+  // CHECK: call {{.*}}i32 [[FOO]](i32 {{.*}})
+  a += foo(n);
+
+  S1 S;
+  // CHECK: call {{.*}}i32 [[FS1:@.+]]([[S1]]* {{.*}}, i32 {{.*}})
+  a += S.r1(n);
+
+  // CHECK: call {{.*}}i32 [[FSTATIC:@.+]](i32 {{.*}})
+  a += fstatic(n);
+
+  // CHECK: call {{.*}}i32 [[FTEMPLATE:@.+]](i32 {{.*}})
+  a += ftemplate<int>(n);
+
+  return a;
+}
+
+//
+// CHECK: define {{.*}}[[FS1]]
+//
+// CHECK:          i8* @llvm.stacksave()
+// CHECK-64:       [[B_ADDR:%.+]] = bitcast i[[SZ]]* [[B_CADDR:%.+]] to i32*
+// CHECK-64:       store i32 %{{.+}}, i32* [[B_ADDR]],
+// CHECK-64:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_CADDR]],
+
+// CHECK-32:       store i32 %{{.+}}, i32* [[B_ADDR:%.+]],
+// CHECK-32:       [[B_CVAL:%.+]] = load i[[SZ]], i[[SZ]]* [[B_ADDR]],
+
+// We capture 2 VLA sizes in this target region
+// CHECK:       [[CELEMSIZE2:%.+]] = mul nuw i[[SZ]] 2, [[VLA0:%.+]]
+// CHECK:       [[CSIZE:%.+]] = mul nuw i[[SZ]] [[CELEMSIZE2]], 2
+
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 60
+// CHECK:       br i1 [[IF]], label %[[TRY:[^,]+]], label %[[FAIL:[^,]+]]
+// CHECK:       [[TRY]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* [[SR:%[^,]+]], i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT5]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SR]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[SADDR0:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX0:[0-9]+]]
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX0]]
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX0]]
+// CHECK-DAG:   [[SADDR1:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX1:[0-9]+]]
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX1]]
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX1]]
+// CHECK-DAG:   [[SADDR2:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX2:[0-9]+]]
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX2]]
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX2]]
+// CHECK-DAG:   [[SADDR3:%.+]] = getelementptr inbounds [5 x i[[SZ]]], [5 x i[[SZ]]]* [[S]], i32 [[IDX3:[0-9]+]]
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 [[IDX3]]
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 [[IDX3]]
+
+// The names below are not necessarily consistent with the names used for the
+// addresses above as some are repeated.
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VLA0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] 2, i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] {{4|8}}, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[B_CVAL]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   store i[[SZ]] 4, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CBPADDR3:%.+]],
+// CHECK-DAG:   store [[S1]]* %{{.+}}, [[S1]]** [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to [[S1]]**
+// CHECK-DAG:   store i[[SZ]] 8, i[[SZ]]* {{%[^,]+}}
+
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store i16* %{{.+}}, i16** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to i16**
+// CHECK-DAG:   store i[[SZ]] [[CSIZE]], i[[SZ]]* {{%[^,]+}}
+
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:[^,]+]], label %[[END:[^,]+]]
+
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT7:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+
+//
+// CHECK: define {{.*}}[[FSTATIC]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 50
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([5 x i[[SZ]]], [5 x i[[SZ]]]* [[SIZET6]], i32 0, i32 0), i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[MAPT6]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 3
+// CHECK-DAG:   [[PADDR3:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 3
+// CHECK-DAG:   store i[[SZ]] [[VAL3:%.+]], i[[SZ]]* [[CBPADDR3:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL3]], i[[SZ]]* [[CPADDR3:%.+]],
+// CHECK-DAG:   [[CBPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR3]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR4:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BP]], i32 0, i32 4
+// CHECK-DAG:   [[PADDR4:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[P]], i32 0, i32 4
+// CHECK-DAG:   store [10 x i32]* %{{.+}}, [10 x i32]** [[CBPADDR4:%.+]],
+// CHECK-DAG:   store [10 x i32]* %{{.+}}, [10 x i32]** [[CPADDR4:%.+]],
+// CHECK-DAG:   [[CBPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
+// CHECK-DAG:   [[CPADDR4]] = bitcast i8** {{%[^,]+}} to [10 x i32]**
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT6:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT6]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+//
+// CHECK: define {{.*}}[[FTEMPLATE]]
+//
+// CHECK:       [[IF:%.+]] = icmp sgt i32 {{[^,]+}}, 40
+// CHECK:       br i1 [[IF]], label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
+// CHECK:       [[IFTHEN]]
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** [[BPR:%[^,]+]], i8** [[PR:%[^,]+]], i[[SZ]]* getelementptr inbounds ([3 x i[[SZ]]], [3 x i[[SZ]]]* [[SIZET7]], i32 0, i32 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[MAPT7]], i32 0, i32 0), i32 0, i32 0)
+// CHECK-DAG:   [[BPR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP:%.+]], i32 0, i32 0
+// CHECK-DAG:   [[PR]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P:%.+]], i32 0, i32 0
+
+// CHECK-DAG:   [[BPADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 0
+// CHECK-DAG:   [[PADDR0:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 0
+// CHECK-DAG:   store i[[SZ]] [[VAL0:%.+]], i[[SZ]]* [[CBPADDR0:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL0]], i[[SZ]]* [[CPADDR0:%.+]],
+// CHECK-DAG:   [[CBPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR0]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 1
+// CHECK-DAG:   [[PADDR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 1
+// CHECK-DAG:   store i[[SZ]] [[VAL1:%.+]], i[[SZ]]* [[CBPADDR1:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL1]], i[[SZ]]* [[CPADDR1:%.+]],
+// CHECK-DAG:   [[CBPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR1]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK-DAG:   [[BPADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BP]], i32 0, i32 2
+// CHECK-DAG:   [[PADDR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[P]], i32 0, i32 2
+// CHECK-DAG:   store i[[SZ]] [[VAL2:%.+]], i[[SZ]]* [[CBPADDR2:%.+]],
+// CHECK-DAG:   store i[[SZ]] [[VAL2]], i[[SZ]]* [[CPADDR2:%.+]],
+// CHECK-DAG:   [[CBPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+// CHECK-DAG:   [[CPADDR2]] = bitcast i8** {{%[^,]+}} to i[[SZ]]*
+
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
+// CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
+// CHECK:       [[FAIL]]
+// CHECK:       call void [[HVT5:@.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[END]]
+// CHECK:       [[END]]
+// CHECK-NEXT:  br label %[[IFEND:.+]]
+// CHECK:       [[IFELSE]]
+// CHECK:       call void [[HVT5]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK-NEXT:  br label %[[IFEND]]
+// CHECK:       [[IFEND]]
+
+
+
+// Check that the offloading functions are emitted and that the arguments are
+// correct and loaded correctly for the target regions of the callees of bar().
+
+// CHECK:       define internal void [[HVT7]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_THIS:%.+]] = alloca [[S1]]*
+// CHECK:       [[LOCAL_B:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA1:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_VLA2:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_C:%.+]] = alloca i16*
+// CHECK:       [[LOCAL_B_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store [[S1]]* [[ARG_THIS:%.+]], [[S1]]** [[LOCAL_THIS]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_B:%.+]], i[[SZ]]* [[LOCAL_B]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA1:%.+]], i[[SZ]]* [[LOCAL_VLA1]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_VLA2:%.+]], i[[SZ]]* [[LOCAL_VLA2]]
+// CHECK-DAG:   store i16* [[ARG_C:%.+]], i16** [[LOCAL_C]]
+// Store captures in the context.
+// CHECK-DAG:   [[REF_THIS:%.+]] = load [[S1]]*, [[S1]]** [[LOCAL_THIS]],
+// CHECK-64-DAG:[[CONV_BP:%.+]] = bitcast i[[SZ]]* [[LOCAL_B]] to i32*
+// CHECK-DAG:   [[VAL_VLA1:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA1]],
+// CHECK-DAG:   [[VAL_VLA2:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_VLA2]],
+// CHECK-DAG:   [[REF_C:%.+]] = load i16*, i16** [[LOCAL_C]],
+
+// CHECK-64-DAG:[[CONV_B:%.+]] = load i32, i32* [[CONV_BP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_B_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_B]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_BV:%.+]] = load i32, i32* [[LOCAL_B]]
+// CHECK-32-DAG:store i32 [[LOCAL_BV]], i32* [[LOCAL_B_CASTED]], align
+// CHECK-DAG:   [[REF_B:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_B_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, [[S1]]*, i[[SZ]], i[[SZ]], i[[SZ]], i16*)* [[OMP_OUTLINED5:@.+]] to void (i32*, i32*, ...)*), [[S1]]* [[REF_THIS]], i[[SZ]] [[REF_B]], i[[SZ]] [[VAL_VLA1]], i[[SZ]] [[VAL_VLA2]], i16* [[REF_C]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED5]](i32* noalias %.global_tid., i32* noalias %.bound_tid., [[S1]]* %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i16* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+
+// CHECK:       define internal void [[HVT6]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       alloca i[[SZ]],
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       alloca i[[SZ]],
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AAA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AAA:%.+]], i[[SZ]]* [[LOCAL_AAA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[CONV_AAAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA]] to i8*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK-DAG:   [[CONV_AAA:%.+]] = load i8, i8* [[CONV_AAAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AAA_CASTED]] to i8*
+// CHECK-DAG:   store i8 [[CONV_AAA]], i8* [[CONV]], align
+// CHECK-DAG:   [[REF_AAA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AAA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED6:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] {{.+}}, i[[SZ]] [[REF_AA]], i[[SZ]] [[REF_AAA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED6]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+// CHECK:       define internal void [[HVT5]]
+// Create local storage for each capture.
+// CHECK:       [[LOCAL_A:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_B:%.+]] = alloca [10 x i32]*
+// CHECK:       [[LOCAL_A_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK:       [[LOCAL_AA_CASTED:%.+]] = alloca i[[SZ]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_A:%.+]], i[[SZ]]* [[LOCAL_A]]
+// CHECK-DAG:   store i[[SZ]] [[ARG_AA:%.+]], i[[SZ]]* [[LOCAL_AA]]
+// CHECK-DAG:   store [10 x i32]* [[ARG_B:%.+]], [10 x i32]** [[LOCAL_B]]
+// Store captures in the context.
+// CHECK-64-DAG:[[CONV_AP:%.+]] = bitcast i[[SZ]]* [[LOCAL_A]] to i32*
+// CHECK-DAG:   [[CONV_AAP:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA]] to i16*
+// CHECK-DAG:   [[REF_B:%.+]] = load [10 x i32]*, [10 x i32]** [[LOCAL_B]],
+
+// CHECK-64-DAG:[[CONV_A:%.+]] = load i32, i32* [[CONV_AP]]
+// CHECK-64-DAG:[[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_A_CASTED]] to i32*
+// CHECK-64-DAG:store i32 [[CONV_A]], i32* [[CONV]], align
+// CHECK-32-DAG:[[LOCAL_AV:%.+]] = load i32, i32* [[LOCAL_A]]
+// CHECK-32-DAG:store i32 [[LOCAL_AV]], i32* [[LOCAL_A_CASTED]], align
+// CHECK-DAG:   [[REF_A:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_A_CASTED]],
+
+// CHECK-DAG:   [[CONV_AA:%.+]] = load i16, i16* [[CONV_AAP]]
+// CHECK-DAG:   [[CONV:%.+]] = bitcast i[[SZ]]* [[LOCAL_AA_CASTED]] to i16*
+// CHECK-DAG:   store i16 [[CONV_AA]], i16* [[CONV]], align
+// CHECK-DAG:   [[REF_AA:%.+]] = load i[[SZ]], i[[SZ]]* [[LOCAL_AA_CASTED]],
+
+// CHECK:       call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_teams(%ident_t* [[DEF_LOC]], i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i[[SZ]], i[[SZ]], [10 x i32]*)* [[OMP_OUTLINED7:@.+]] to void (i32*, i32*, ...)*), i[[SZ]] [[REF_A]], i[[SZ]] [[REF_AA]], [10 x i32]* [[REF_B]])
+//
+//
+// CHECK:       define internal {{.*}}void [[OMP_OUTLINED7]](i32* noalias %.global_tid., i32* noalias %.bound_tid., i[[SZ]] %{{.+}}, i[[SZ]] %{{.+}}, [10 x i32]* {{.+}})
+// To reduce complexity, we're only going as far as validating the signature of the outlined parallel function.
+
+// CHECK-DAG: !{!"llvm.loop.vectorize.width", i32 16}
+// CHECK-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CHECK-DAG: !{!"llvm.loop.vectorize.width", i32 32}
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_codegen_registration.cpp b/test/OpenMP/target_teams_distribute_simd_codegen_registration.cpp
new file mode 100644
index 000000000000..d138114633ea
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_codegen_registration.cpp
@@ -0,0 +1,454 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target teams distribute simd codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// Check that no target code is emmitted if no omptests flag was provided.
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK-NTARGET
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-DAG: [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+// CHECK-DAG: [[DEVTY:%.+]] = type { i8*, i8*, [[ENTTY]]*, [[ENTTY]]* }
+// CHECK-DAG: [[DSCTY:%.+]] = type { i32, [[DEVTY]]*, [[ENTTY]]*, [[ENTTY]]* }
+
+// TCHECK:    [[ENTTY:%.+]] = type { i8*, i8*, i[[SZ:32|64]], i32, i32 }
+
+// CHECK-DAG: $[[REGFN:\.omp_offloading\..+]] = comdat
+
+// CHECK-DAG: [[A1:@.+]] = internal global [[SA]]
+// CHECK-DAG: [[A2:@.+]] = global [[SA]]
+// CHECK-DAG: [[B1:@.+]] = global [[SB]]
+// CHECK-DAG: [[B2:@.+]] = global [[SB]]
+// CHECK-DAG: [[C1:@.+]] = internal global [[SC]]
+// CHECK-DAG: [[D1:@.+]] = global [[SD]]
+// CHECK-DAG: [[E1:@.+]] = global [[SE]]
+// CHECK-DAG: [[T1:@.+]] = global [[ST1]]
+// CHECK-DAG: [[T2:@.+]] = global [[ST2]]
+
+// CHECK-NTARGET-DAG: [[SA:%.+]] = type { [4 x i32] }
+// CHECK-NTARGET-DAG: [[SB:%.+]] = type { [8 x i32] }
+// CHECK-NTARGET-DAG: [[SC:%.+]] = type { [16 x i32] }
+// CHECK-NTARGET-DAG: [[SD:%.+]] = type { [32 x i32] }
+// CHECK-NTARGET-DAG: [[SE:%.+]] = type { [64 x i32] }
+// CHECK-NTARGET-DAG: [[ST1:%.+]] = type { [228 x i32] }
+// CHECK-NTARGET-DAG: [[ST2:%.+]] = type { [1128 x i32] }
+// CHECK-NTARGET-NOT: type { i8*, i8*, %
+// CHECK-NTARGET-NOT: type { i32, %
+
+// We have 7 target regions
+
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// TCHECK-NOT: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+// CHECK-DAG: {{@.+}} = private constant i8 0
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i[[SZ]]] [i[[SZ]] 4]
+// CHECK-DAG: {{@.+}} = private unnamed_addr constant [1 x i64] [i64 288]
+
+// CHECK-NTARGET-NOT: private constant i8 0
+// CHECK-NTARGET-NOT: private unnamed_addr constant [1 x i
+
+// CHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// CHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// CHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// CHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// CHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// CHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// CHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// CHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// CHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// CHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// CHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// CHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// CHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// CHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* @{{.*}}, i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// TCHECK-DAG: [[NAMEPTR1:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME1:__omp_offloading_[0-9a-f]+_[0-9a-f]+__Z.+_l[0-9]+]]\00"
+// TCHECK-DAG: [[ENTRY1:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR1]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR2:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME2:.+]]\00"
+// TCHECK-DAG: [[ENTRY2:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR2]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR3:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME3:.+]]\00"
+// TCHECK-DAG: [[ENTRY3:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR3]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR4:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME4:.+]]\00"
+// TCHECK-DAG: [[ENTRY4:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR4]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR5:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME5:.+]]\00"
+// TCHECK-DAG: [[ENTRY5:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR5]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR6:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME6:.+]]\00"
+// TCHECK-DAG: [[ENTRY6:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR6]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR7:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME7:.+]]\00"
+// TCHECK-DAG: [[ENTRY7:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR7]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR8:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME8:.+]]\00"
+// TCHECK-DAG: [[ENTRY8:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR8]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR9:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME9:.+]]\00"
+// TCHECK-DAG: [[ENTRY9:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR9]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR10:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME10:.+]]\00"
+// TCHECK-DAG: [[ENTRY10:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR10]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR11:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME11:.+]]\00"
+// TCHECK-DAG: [[ENTRY11:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR11]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+// TCHECK-DAG: [[NAMEPTR12:@.+]] = internal unnamed_addr constant [{{.*}} x i8] c"[[NAME12:.+]]\00"
+// TCHECK-DAG: [[ENTRY12:@.+]] = constant [[ENTTY]] { i8* bitcast (void (i[[SZ]])* @{{.*}} to i8*), i8* getelementptr inbounds ([{{.*}} x i8], [{{.*}} x i8]* [[NAMEPTR12]], i32 0, i32 0), i[[SZ]] 0, i32 0, i32 0 }, section ".omp_offloading.entries", align 1
+
+// CHECK: [[ENTBEGIN:@.+]] = external constant [[ENTTY]]
+// CHECK: [[ENTEND:@.+]] = external constant [[ENTTY]]
+// CHECK: [[DEVBEGIN:@.+]] = external constant i8
+// CHECK: [[DEVEND:@.+]] = external constant i8
+// CHECK: [[IMAGES:@.+]] = internal unnamed_addr constant [1 x [[DEVTY]]] [{{.+}} { i8* [[DEVBEGIN]], i8* [[DEVEND]], [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }], comdat($[[REGFN]])
+// CHECK: [[DESC:@.+]] = internal constant [[DSCTY]] { i32 1, [[DEVTY]]* getelementptr inbounds ([1 x [[DEVTY]]], [1 x [[DEVTY]]]* [[IMAGES]], i32 0, i32 0), [[ENTTY]]* [[ENTBEGIN]], [[ENTTY]]* [[ENTEND]] }, comdat($[[REGFN]])
+
+// We have 4 initializers, one for the 500 priority, another one for 501, or more for the default priority, and the last one for the offloading registration function.
+// CHECK: @llvm.global_ctors = appending global [4 x { i32, void ()*, i8* }] [
+// CHECK-SAME: { i32, void ()*, i8* } { i32 500, void ()* [[P500:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 501, void ()* [[P501:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 65535, void ()* [[PMAX:@[^,]+]], i8* null },
+// CHECK-SAME: { i32, void ()*, i8* } { i32 0, void ()* bitcast (void (i8*)* @[[REGFN]] to void ()*), i8* bitcast (void (i8*)* @[[REGFN]] to i8*) }]
+
+// CHECK-NTARGET: @llvm.global_ctors = appending global [3   x { i32, void ()*, i8* }] [
+
+extern int *R;
+
+struct SA {
+  int arr[4];
+  void foo() {
+    int a = *R;
+    a += 1;
+    *R = a;
+  }
+  SA() {
+    int a = *R;
+    a += 2;
+    *R = a;
+  }
+  ~SA() {
+    int a = *R;
+    a += 3;
+    *R = a;
+  }
+};
+
+struct SB {
+  int arr[8];
+  void foo() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 4;
+    *R = a;
+  }
+  SB() {
+    int a = *R;
+    a += 5;
+    *R = a;
+  }
+  ~SB() {
+    int a = *R;
+    a += 6;
+    *R = a;
+  }
+};
+
+struct SC {
+  int arr[16];
+  void foo() {
+    int a = *R;
+    a += 7;
+    *R = a;
+  }
+  SC() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 8;
+    *R = a;
+  }
+  ~SC() {
+    int a = *R;
+    a += 9;
+    *R = a;
+  }
+};
+
+struct SD {
+  int arr[32];
+  void foo() {
+    int a = *R;
+    a += 10;
+    *R = a;
+  }
+  SD() {
+    int a = *R;
+    a += 11;
+    *R = a;
+  }
+  ~SD() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 12;
+    *R = a;
+  }
+};
+
+struct SE {
+  int arr[64];
+  void foo() {
+    int a = *R;
+    #pragma omp target teams distribute simd if(target: 0)
+    for (int i = 0; i < 10; ++i)
+      a += 13;
+    *R = a;
+  }
+  SE() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 14;
+    *R = a;
+  }
+  ~SE() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 15;
+    *R = a;
+  }
+};
+
+template <int x>
+struct ST {
+  int arr[128 + x];
+  void foo() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 16 + x;
+    *R = a;
+  }
+  ST() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 17 + x;
+    *R = a;
+  }
+  ~ST() {
+    int a = *R;
+    #pragma omp target teams distribute simd
+    for (int i = 0; i < 10; ++i)
+      a += 18 + x;
+    *R = a;
+  }
+};
+
+// We have to make sure we us all the target regions:
+//CHECK-DAG: define internal void @[[NAME1]](
+//CHECK-DAG: call void @[[NAME1]](
+//CHECK-DAG: define internal void @[[NAME2]](
+//CHECK-DAG: call void @[[NAME2]](
+//CHECK-DAG: define internal void @[[NAME3]](
+//CHECK-DAG: call void @[[NAME3]](
+//CHECK-DAG: define internal void @[[NAME4]](
+//CHECK-DAG: call void @[[NAME4]](
+//CHECK-DAG: define internal void @[[NAME5]](
+//CHECK-DAG: call void @[[NAME5]](
+//CHECK-DAG: define internal void @[[NAME6]](
+//CHECK-DAG: call void @[[NAME6]](
+//CHECK-DAG: define internal void @[[NAME7]](
+//CHECK-DAG: call void @[[NAME7]](
+//CHECK-DAG: define internal void @[[NAME8]](
+//CHECK-DAG: call void @[[NAME8]](
+//CHECK-DAG: define internal void @[[NAME9]](
+//CHECK-DAG: call void @[[NAME9]](
+//CHECK-DAG: define internal void @[[NAME10]](
+//CHECK-DAG: call void @[[NAME10]](
+//CHECK-DAG: define internal void @[[NAME11]](
+//CHECK-DAG: call void @[[NAME11]](
+//CHECK-DAG: define internal void @[[NAME12]](
+//CHECK-DAG: call void @[[NAME12]](
+
+//TCHECK-DAG: define void @[[NAME1]](
+//TCHECK-DAG: define void @[[NAME2]](
+//TCHECK-DAG: define void @[[NAME3]](
+//TCHECK-DAG: define void @[[NAME4]](
+//TCHECK-DAG: define void @[[NAME5]](
+//TCHECK-DAG: define void @[[NAME6]](
+//TCHECK-DAG: define void @[[NAME7]](
+//TCHECK-DAG: define void @[[NAME8]](
+//TCHECK-DAG: define void @[[NAME9]](
+//TCHECK-DAG: define void @[[NAME10]](
+//TCHECK-DAG: define void @[[NAME11]](
+//TCHECK-DAG: define void @[[NAME12]](
+
+// CHECK-NTARGET-NOT: __tgt_target
+// CHECK-NTARGET-NOT: __tgt_register_lib
+// CHECK-NTARGET-NOT: __tgt_unregister_lib
+
+// TCHECK-NOT: __tgt_target
+// TCHECK-NOT: __tgt_register_lib
+// TCHECK-NOT: __tgt_unregister_lib
+
+// We have 2 initializers with priority 500
+//CHECK: define internal void [[P500]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 1 initializers with priority 501
+//CHECK: define internal void [[P501]](
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// We have 6 initializers with default priority
+//CHECK: define internal void [[PMAX]](
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK:     call void @{{.+}}()
+//CHECK-NOT: call void @{{.+}}()
+//CHECK:     ret void
+
+// Check registration and unregistration
+
+//CHECK:     define internal void @[[UNREGFN:.+]](i8*)
+//CHECK-SAME: comdat($[[REGFN]]) {
+//CHECK:     call i32 @__tgt_unregister_lib([[DSCTY]]* [[DESC]])
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_unregister_lib([[DSCTY]]*)
+
+//CHECK:     define linkonce hidden void @[[REGFN]](i8*)
+//CHECK-SAME: comdat {
+//CHECK:     call i32 @__tgt_register_lib([[DSCTY]]* [[DESC]])
+//CHECK:     call i32 @__cxa_atexit(void (i8*)* @[[UNREGFN]], i8* bitcast ([[DSCTY]]* [[DESC]] to i8*),
+//CHECK:     ret void
+//CHECK:     declare i32 @__tgt_register_lib([[DSCTY]]*)
+
+static __attribute__((init_priority(500))) SA a1;
+SA a2;
+SB __attribute__((init_priority(500))) b1;
+SB __attribute__((init_priority(501))) b2;
+static SC c1;
+SD d1;
+SE e1;
+ST<100> t1;
+ST<1000> t2;
+
+
+int bar(int a){
+  int r = a;
+
+  a1.foo();
+  a2.foo();
+  b1.foo();
+  b2.foo();
+  c1.foo();
+  d1.foo();
+  e1.foo();
+  t1.foo();
+  t2.foo();
+
+  #pragma omp target teams distribute simd
+  for (int i = 0; i < 10; ++i)
+    ++r;
+
+  return r + *R;
+}
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID:-?[0-9]+]], i32 [[FILEID:-?[0-9]+]], !"_ZN2SB3fooEv", i32 195, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SDD1Ev", i32 247, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SEC1Ev", i32 265, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SED1Ev", i32 272, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_Z3bari", i32 415, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EEC1Ev", i32 291, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi1000EED1Ev", i32 298, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2STILi100EE3fooEv", i32 284, i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 [[DEVID]], i32 [[FILEID]], !"_ZN2SCC1Ev", i32 221, i32 {{[0-9]+}}}
+
+// TCHECK-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CHECK-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_codegen_registration_naming.cpp b/test/OpenMP/target_teams_distribute_simd_codegen_registration_naming.cpp
new file mode 100644
index 000000000000..76618accbd68
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_codegen_registration_naming.cpp
@@ -0,0 +1,71 @@
+// Test host codegen.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+
+// Test target teams distribute simd codegen - host bc file has to be created first.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm-bc %s -o %t-ppc-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm-bc %s -o %t-x86-host.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s -check-prefix=TCHECK
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=TCHECK
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[CA:%.+]] = type { i32* }
+
+// CHECK: define {{.*}}i32 @[[NNAME:.+]](i32 {{.*}}%{{.+}})
+int nested(int a){
+  // CHECK: call void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME]]_l[[T1L:[0-9]+]](
+  #pragma omp target teams distribute simd
+  for (int i = 0; i < 10; ++i)
+    ++a;
+
+  // CHECK: call void @"[[LNAME:.+]]"([[CA]]*
+  auto F = [&](){
+    #pragma omp parallel
+    {
+      #pragma omp target teams distribute simd
+      for (int i = 0; i < 10; ++i)
+        ++a;
+    }
+  };
+
+  F();
+
+  return a;
+}
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T1L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID:[0-9a-f]+_[0-9a-f]+]]_[[NNAME:.+]]_l[[T1L:[0-9]+]](
+
+// CHECK: define {{.*}}void @"[[LNAME]]"(
+// CHECK: call void {{.*}}@__kmpc_fork_call{{.+}}[[PNAME:@.+]] to
+
+// CHECK: define {{.*}}void [[PNAME]](
+// CHECK: call void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L:[0-9]+]](
+
+// CHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME]]_l[[T2L]](
+// TCHECK: define {{.*}}void @__omp_offloading_[[FILEID]]_[[NNAME:.+]]_l[[T2L:[0-9]+]](
+
+
+// Check metadata is properly generated:
+// CHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// CHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+
+// TCHECK:     !omp_offload.info = !{!{{[0-9]+}}, !{{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T1L]], i32 {{[0-9]+}}}
+// TCHECK-DAG: = !{i32 0, i32 {{-?[0-9]+}}, i32 {{-?[0-9]+}}, !"[[NNAME]]", i32 [[T2L]], i32 {{[0-9]+}}}
+
+// CHECK-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// TCHECK-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_collapse_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_collapse_codegen.cpp
new file mode 100644
index 000000000000..703136f0652e
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_collapse_codegen.cpp
@@ -0,0 +1,125 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X][Y];
+
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+    // CK1: call i32 @__tgt_target_teams(
+    // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target teams distribute simd collapse(2)
+    for(int i = 0; i < X; i++) {
+      for(int j = 0; j < Y; j++) {
+        a[i][j] = (T)0;
+      }
+    }
+    // CK1: define internal void @[[OFFL1]](
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+    // CK1: ret void
+
+    // CK1: define internal void @[[OUTL1]]({{.+}})
+    // discard loop variables not needed here
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: [[OMP_UB:%.+]] = alloca i32,
+    // CK1: store i32 56087, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void
+
+    return a[0][0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n, int m>
+int tmain(T argc) {
+  T a[n][m];
+  #pragma omp target teams distribute simd collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = (T)0;
+    }
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int m = 2;
+  int a[n][m];
+  #pragma omp target teams distribute simd collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = 0;
+    }
+  }
+  return tmain<int, 10, 2>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i64,
+// CK2: store i64 {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_8({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// discard loop variables not needed here
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i32,
+// CK2: store i32 {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/target_teams_distribute_simd_depend_messages.cpp b/test/OpenMP/target_teams_distribute_simd_depend_messages.cpp
index 39d12d8d4d0e..c24c7438b0b5 100644
--- a/test/OpenMP/target_teams_distribute_simd_depend_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_depend_messages.cpp
@@ -37,21 +37,21 @@ int main(int argc, char **argv, char *env[]) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd depend (out: ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+#pragma omp target teams distribute simd depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd depend (out :S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute simd depend(in : argv[1][1] = '2')
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute simd depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd depend (in : argv[0])
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd depend (in : ) // expected-error {{expected expression}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd depend (in : main) // expected-error {{expected variable name, array element or array section}}
+#pragma omp target teams distribute simd depend (in : main)
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+#pragma omp target teams distribute simd depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_simd_dist_schedule_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_dist_schedule_codegen.cpp
new file mode 100644
index 000000000000..d096da329abc
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_dist_schedule_codegen.cpp
@@ -0,0 +1,197 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target teams distribute simd
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target teams distribute simd dist_schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target teams distribute simd dist_schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+#pragma omp target teams distribute simd
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target teams distribute simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target teams distribute simd dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target teams distribute simd
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target teams distribute simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target teams distribute simd dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..f2871f811833
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_firstprivate_codegen.cpp
@@ -0,0 +1,342 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+// CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target teams distribute simd firstprivate(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}})
+    // LAMBDA:  ret
+#pragma omp target teams distribute simd firstprivate(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}})
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]],
+    // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]],
+    // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_PRIV_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // LAMBDA: [[SIVAR_PRIV_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // skip loop vars
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_PRIV_ADDR]], {{.+}} [[G1_TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV_ADDR]],
+    // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP]]
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV_ADDR]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target teams distribute simd firstprivate(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]],
+// CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]],
+// CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]],
+// CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[TOFFL1:.+]]({{[^,]+}}, {{[^,]+}}, {{[^,]+}}, {{[^,]+}})
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]({{.+}})
+// CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]],
+// CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]],
+// CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]],
+// CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_firstprivate_messages.cpp b/test/OpenMP/target_teams_distribute_simd_firstprivate_messages.cpp
index b9b039efd929..235533b49d16 100644
--- a/test/OpenMP/target_teams_distribute_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_firstprivate_messages.cpp
@@ -124,7 +124,8 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute simd firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute simd lastprivate(argc), firstprivate(argc) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute simd lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
 #pragma omp target teams distribute simd firstprivate(argc) map(argc) // expected-error {{firstprivate variable cannot be in a map clause in '#pragma omp target teams distribute simd' directive}} expected-note {{defined as firstprivate}}
diff --git a/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..8a978a221954
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_lastprivate_codegen.cpp
@@ -0,0 +1,385 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target teams distribute simd lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target teams distribute simd lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA-64: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, i64 [[G_IN:%.+]], i64 [[G1_IN:%.+]], i64 [[SVAR_IN:%.+]], i64 [[SFVAR_IN:%.+]])
+      // LAMBDA-32: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i32 [[SVAR_IN:%.+]], i32 [[SFVAR_IN:%.+]])
+      // LAMBDA-64: [[G_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA-64: [[G1_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA-64: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i32,
+      // LAMBDA-64: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca i64,
+      // LAMBDA-32: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca i32,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA-64: store i64 [[G_IN]], i64* [[G_PRIVATE_ADDR]],
+      // LAMBDA-32: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA-64: store i64 [[G1_IN]], i64* [[G1_PRIVATE_ADDR]],
+      // LAMBDA-32: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA-64: store i64 [[SVAR_IN]], i64* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA-32: store i32 [[SVAR_IN]], i32* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA-64: store i64 [[SFVAR_IN]], i64* [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA-32: store i32 [[SFVAR_IN]], i32* [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA-64: [[G_IN_REF:%.+]] = bitcast i64* [[G_PRIVATE_ADDR]] to double*
+      // LAMBDA-32: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA-64: [[G1_IN_REF:%.+]] = bitcast i64* [[G1_PRIVATE_ADDR]] to double*
+      // LAMBDA-64: store double* [[G1_IN_REF]], double** [[G1_IN_ADDR_REF:%.+]],
+      // LAMBDA-64: [[SVAR_IN_REF:%.+]] = bitcast i64* [[SVAR_PRIVATE_ADDR]] to i32*
+      // LAMBDA-64: [[SFVAR_IN_REF:%.+]] = bitcast i64* [[SFVAR_PRIVATE_ADDR]] to float*
+      // LAMBDA-32: [[SFVAR_IN_REF:%.+]] = bitcast i32* [[SFVAR_PRIVATE_ADDR]] to float*
+      // LAMBDA-64: [[G1_IN_REF:%.+]] = load double*, double** [[G1_IN_ADDR_REF]],
+      // LAMBDA-32: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: store i32 2, i32* %
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA-64: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA-32: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+      [&]() {
+        // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+        // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+        g = 2;
+        g1 = 2;
+        svar = 4;
+        sfvar = 8.0;
+        // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+        // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+        // LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+        // LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+        // LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+        // LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+        // LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+        // LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+        // LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+        // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+        // LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+        // LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+        // LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target teams distribute simd lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]]([2 x i{{[0-9]+}}]* {{.+}}, i{{[0-9]+}} {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]]([2 x i{{[0-9]+}}]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_IN]], i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} [[S_VAR_IN]], i{{[0-9]+}}* [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[T_VAR_ADDR_REF:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-64: [[SVAR_ADDR_REF:%.+]] = bitcast i64* [[SVAR_ADDR]] to i32*
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK-64: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK-32: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK-64: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK-32: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]]([2 x i{{[0-9]+}}]* {{.+}}, i{{[0-9]+}} {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_IN1]], i{{[0-9]+}}* [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK-64: [[T_VAR_ADDR_REF1:%.+]] = bitcast i64* [[T_VAR_ADDR1]] to i32*
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK-64: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK-32: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_lastprivate_messages.cpp b/test/OpenMP/target_teams_distribute_simd_lastprivate_messages.cpp
index 3b533e2e5484..a143dc2fd883 100644
--- a/test/OpenMP/target_teams_distribute_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -216,10 +216,12 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute simd lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute simd firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+// expected-error@+1 {{firstprivate variable cannot be lastprivate}} expected-note@+1 {{defined as firstprivate}}
+#pragma omp target teams distribute simd firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
-#pragma omp target teams distribute simd lastprivate(n) firstprivate(n) // OK
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/target_teams_distribute_simd_linear_messages.cpp b/test/OpenMP/target_teams_distribute_simd_linear_messages.cpp
index 8db57a0f3665..2be792bbda74 100644
--- a/test/OpenMP/target_teams_distribute_simd_linear_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_linear_messages.cpp
@@ -18,33 +18,42 @@ void test_linear_colons()
 {
   int B = 0;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(B:bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(B:B::bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(X::x : ::z)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(B,::z, X::x)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(::z)
   for (int i = 0; i < 10; ++i) ;
 
 #pragma omp target teams distribute simd linear(B::bfoo()) // expected-error {{expected variable name}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+1 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(B::ib,B:C1+C2)
   for (int i = 0; i < 10; ++i) ;
 }
@@ -65,6 +74,7 @@ template<int L, class T, class N> T test_template(T* arr, N num) {
 
 template<int LEN> int test_warn() {
   int ind2 = 0;
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
   #pragma omp target teams distribute simd linear(ind2:LEN) // expected-warning {{zero linear step (ind2 should probably be const)}}
   for (int i = 0; i < 100; i++) {
     ind2 += LEN;
@@ -118,15 +128,18 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp target teams distribute simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
 #pragma omp target teams distribute simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear (argc : 5)
   for (int k = 0; k < argc; ++k) ++k;
 
@@ -139,33 +152,17 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp target teams distribute simd linear (argv[1]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(e, g)
   for (int k = 0; k < argc; ++k) ++k;
 
 #pragma omp target teams distribute simd linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear(i)
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target teams distribute simd linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target teams distribute simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target teams distribute simd linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target teams distribute simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
   return 0;
 }
 
@@ -198,15 +195,18 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
 #pragma omp target teams distribute simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+1 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target teams distribute simd linear (argc)
   for (int k = 0; k < argc; ++k) ++k;
 
@@ -226,22 +226,6 @@ int main(int argc, char **argv) {
 #pragma omp target teams distribute simd linear(h, C::x) // expected-error 2 {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int i;
-    #pragma omp target teams distribute simd linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
-
-    #pragma omp target teams distribute simd linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
-  }
-
-#pragma omp target teams distribute simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target teams distribute simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
   foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
   return 0;
 }
diff --git a/test/OpenMP/target_teams_distribute_simd_loop_messages.cpp b/test/OpenMP/target_teams_distribute_simd_loop_messages.cpp
index aae00a4433bc..09aa548bd5c6 100644
--- a/test/OpenMP/target_teams_distribute_simd_loop_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -607,7 +607,8 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
-#pragma omp target teams distribute simd lastprivate(s) firstprivate(s)
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
+#pragma omp target teams distribute simd lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/target_teams_distribute_simd_map_messages.cpp b/test/OpenMP/target_teams_distribute_simd_map_messages.cpp
index 414f6f83debb..29d48d958a3d 100644
--- a/test/OpenMP/target_teams_distribute_simd_map_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_map_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -116,9 +116,9 @@ T tmain(T argc) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd map(ca)
   for (i = 0; i < argc; ++i) foo();
@@ -222,11 +222,11 @@ int main(int argc, char **argv) {
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd map(S1) // expected-error {{'S1' does not refer to a value}}
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute simd map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd map(argv[1])
   for (i = 0; i < argc; ++i) foo();
-#pragma omp target teams distribute simd map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target teams distribute simd map(ba)
   for (i = 0; i < argc; ++i) foo();
 #pragma omp target teams distribute simd map(ca)
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/target_teams_distribute_simd_misc_messages.c b/test/OpenMP/target_teams_distribute_simd_misc_messages.c
index c76fdea4fcf1..70ec508f6798 100644
--- a/test/OpenMP/target_teams_distribute_simd_misc_messages.c
+++ b/test/OpenMP/target_teams_distribute_simd_misc_messages.c
@@ -285,15 +285,22 @@ void test_firstprivate() {
     ;
 
   int x, y, z;
+// expected-error@+1 {{lastprivate variable cannot be firstprivate}} expected-note@+1 {{defined as lastprivate}}
 #pragma omp target teams distribute simd lastprivate(x) firstprivate(x)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 2 {{lastprivate variable cannot be firstprivate}} expected-note@+1 2 {{defined as lastprivate}}
 #pragma omp target teams distribute simd lastprivate(x, y) firstprivate(x, y)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 3 {{lastprivate variable cannot be firstprivate}} expected-note@+1 3 {{defined as lastprivate}}
 #pragma omp target teams distribute simd lastprivate(x, y, z) firstprivate(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 {{the value of 'simdlen' parameter must be less than or equal to the value of the 'safelen' parameter}}
+#pragma omp target teams distribute simd simdlen(64) safelen(8)
+  for (i = 0; i < 16; ++i)
+    ;
 }
 
 void test_loop_messages() {
diff --git a/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp
new file mode 100644
index 000000000000..c8883e18d7d9
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_private_codegen.cpp
@@ -0,0 +1,233 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target teams distribute simd private(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]]()
+    // LAMBDA:  ret
+#pragma omp target teams distribute simd private(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]]()
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target teams distribute simd private(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]()
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[SIVAR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0,
+// CHECK: call void @[[TOFFL1:.+]]()
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_reduction_codegen.cpp b/test/OpenMP/target_teams_distribute_simd_reduction_codegen.cpp
new file mode 100644
index 000000000000..76a1299981cb
--- /dev/null
+++ b/test/OpenMP/target_teams_distribute_simd_reduction_codegen.cpp
@@ -0,0 +1,211 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target teams distribute simd reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target teams distribute simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i32*{{.+}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR:%.+]] = load i32*, i32** [[SIVAR_ADDR]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}}*{{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target teams distribute simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i32* {{.+}})
+// CHECK: [[RES:%.+]] = call{{.*}} i32 @[[TMAIN_INT:[^(]+]]()
+// CHECK: ret i32 [[RES]]
+
+// CHECK: define{{.*}} void @[[OFFL1]](i32*{{.+}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}}** [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_LOAD:%.+]] = load i32*, i32** [[SIVAR_ADDR]],
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_LOAD]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, i32*{{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i32,
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load i32*, i32** [[SIVAR_ADDR]],
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}}* {{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i32*{{.+}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR:%.+]] = load i32*, i32** [[TVAR_ADDR]],
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}}*{{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load i32*, i32** [[TVAR_ADDR]],
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+#endif
diff --git a/test/OpenMP/target_teams_distribute_simd_reduction_messages.cpp b/test/OpenMP/target_teams_distribute_simd_reduction_messages.cpp
index 1c0283285aa5..35d4e12dc339 100644
--- a/test/OpenMP/target_teams_distribute_simd_reduction_messages.cpp
+++ b/test/OpenMP/target_teams_distribute_simd_reduction_messages.cpp
@@ -19,9 +19,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_teams_map_messages.cpp b/test/OpenMP/target_teams_map_messages.cpp
index 89425d622c10..7874c2e105cf 100644
--- a/test/OpenMP/target_teams_map_messages.cpp
+++ b/test/OpenMP/target_teams_map_messages.cpp
@@ -265,7 +265,7 @@ void SAclient(int arg) {
   {}
   #pragma omp target teams map((p+1)->A)  // expected-error {{expected expression containing only member accesses and/or array sections based on named variables}}
   {}
-  #pragma omp target teams map(u.B)  // expected-error {{mapped storage cannot be derived from a union}}
+  #pragma omp target teams map(u.B)  // expected-error {{mapping of union members is not allowed}}
   {}
 
   #pragma omp target data map(to: r.C) //expected-note {{used here}}
@@ -312,8 +312,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -346,7 +346,7 @@ template <class T>
 struct S6;
 
 template<>
-struct S6<int>  // expected-note {{mappable type cannot be polymorphic}}
+struct S6<int>
 {
    virtual void foo();
 };
@@ -414,8 +414,8 @@ T tmain(T argc) {
 #pragma omp target data map(tofrom: argc > 0 ? x : y) // expected-error 2 {{expected expression containing only member accesses and/or array sections based on named variables}}
 #pragma omp target data map(argc)
 #pragma omp target data map(S1) // expected-error {{'S1' does not refer to a value}}
-#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
-#pragma omp target data map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
+#pragma omp target data map(ba)
 #pragma omp target data map(ca)
 #pragma omp target data map(da)
 #pragma omp target data map(S2::S2s)
@@ -488,9 +488,9 @@ int main(int argc, char **argv) {
 #pragma omp target data map(tofrom: argc > 0 ? argv[1] : argv[2]) // expected-error {{xpected expression containing only member accesses and/or array sections based on named variables}}
 #pragma omp target data map(argc)
 #pragma omp target data map(S1) // expected-error {{'S1' does not refer to a value}}
-#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target data map(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
 #pragma omp target data map(argv[1])
-#pragma omp target data map(ba) // expected-error 2 {{type 'S2' is not mappable to target}}
+#pragma omp target data map(ba)
 #pragma omp target data map(ca)
 #pragma omp target data map(da)
 #pragma omp target data map(S2::S2s)
@@ -530,7 +530,7 @@ int main(int argc, char **argv) {
 #pragma omp target teams firstprivate(j) map(j)  // expected-error {{firstprivate variable cannot be in a map clause in '#pragma omp target teams' directive}} expected-note {{defined as firstprivate}}
   {}
 
-#pragma omp target teams map(m) // expected-error {{type 'S6<int>' is not mappable to target}}
+#pragma omp target teams map(m)
   {}
 
   return tmain<int, 3>(argc)+tmain<from, 4>(argc); // expected-note {{in instantiation of function template specialization 'tmain<int, 3>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<int, 4>' requested here}}
diff --git a/test/OpenMP/target_teams_num_teams_codegen.cpp b/test/OpenMP/target_teams_num_teams_codegen.cpp
index ad3626515f50..91288458c5aa 100644
--- a/test/OpenMP/target_teams_num_teams_codegen.cpp
+++ b/test/OpenMP/target_teams_num_teams_codegen.cpp
@@ -147,10 +147,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[TEAMS:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 [[TEAMS]], i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 [[TEAMS]], i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -160,10 +158,8 @@ int bar(int n){
 //
 //
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.+}}, i32 1024, i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.+}}, i32 1024, i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -190,10 +186,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[TEAMS:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 [[TEAMS]], i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 [[TEAMS]], i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -213,10 +207,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[TEAMS:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 [[TEAMS]], i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 [[TEAMS]], i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -233,10 +225,8 @@ int bar(int n){
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 20, i32 0)
-// CHECK-NEXT:  store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK-NEXT:  [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 20, i32 0)
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -257,10 +247,8 @@ int bar(int n){
 // CHECK:       [[T:%.+]] = load i16, i16* [[CAPE_ADDR]], align
 // CHECK:       [[TEAMS:%.+]] = sext i16 [[T]] to i32
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 [[TEAMS]], i32 0)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 [[TEAMS]], i32 0)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
diff --git a/test/OpenMP/target_teams_reduction_messages.cpp b/test/OpenMP/target_teams_reduction_messages.cpp
index 48fa310253c8..b5876b108508 100644
--- a/test/OpenMP/target_teams_reduction_messages.cpp
+++ b/test/OpenMP/target_teams_reduction_messages.cpp
@@ -24,9 +24,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/target_teams_thread_limit_codegen.cpp b/test/OpenMP/target_teams_thread_limit_codegen.cpp
index d89f51561712..5d93c1de77fe 100644
--- a/test/OpenMP/target_teams_thread_limit_codegen.cpp
+++ b/test/OpenMP/target_teams_thread_limit_codegen.cpp
@@ -147,10 +147,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[TL:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 0, i32 [[TL]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 0, i32 [[TL]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -160,10 +158,8 @@ int bar(int n){
 //
 //
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.+}}, i32 0, i32 1024)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.+}}, i32 0, i32 1024)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -199,10 +195,8 @@ int bar(int n){
 // CHECK:       [[TEAMS:%.+]] = load i32, i32* [[CAPE_ADDR1]], align
 // CHECK:       [[TL:%.+]] = load i32, i32* [[CAPE_ADDR2]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, {{.*}}, i32 [[TEAMS]], i32 [[TL]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, {{.*}}, i32 [[TEAMS]], i32 [[TL]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -222,10 +216,8 @@ int bar(int n){
 // CHECK:       [[ARG:%.+]] = load i[[SZ]], i[[SZ]]* [[CAPEC_ADDR]], align
 // CHECK:       [[TL:%.+]] = load i32, i32* [[CAPE_ADDR]], align
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 0, i32 [[TL]])
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, {{.*}}, i32 0, i32 [[TL]])
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -242,10 +234,8 @@ int bar(int n){
 //
 // CHECK: define {{.*}}[[FTEMPLATE]]
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 0, i32 20)
-// CHECK-NEXT:  store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK-NEXT:  [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, {{.*}}, i32 0, i32 20)
+// CHECK-NEXT:  [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK-NEXT:  br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
@@ -266,10 +256,8 @@ int bar(int n){
 // CHECK:       [[T:%.+]] = load i16, i16* [[CAPE_ADDR]], align
 // CHECK:       [[TEAMS:%.+]] = sext i16 [[T]] to i32
 //
-// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 [[TEAMS]], i32 1024)
-// CHECK:       store i32 [[RET]], i32* [[RHV:%.+]], align
-// CHECK:       [[RET2:%.+]] = load i32, i32* [[RHV]], align
-// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET2]], 0
+// CHECK-DAG:   [[RET:%.+]] = call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, {{.*}}, i32 [[TEAMS]], i32 1024)
+// CHECK:       [[ERROR:%.+]] = icmp ne i32 [[RET]], 0
 // CHECK:       br i1 [[ERROR]], label %[[FAIL:.+]], label %[[END:[^,]+]]
 //
 // CHECK:       [[FAIL]]
diff --git a/test/OpenMP/target_update_codegen.cpp b/test/OpenMP/target_update_codegen.cpp
index f1e61a54c6e2..7f45c313a8a1 100644
--- a/test/OpenMP/target_update_codegen.cpp
+++ b/test/OpenMP/target_update_codegen.cpp
@@ -22,15 +22,15 @@ ST<int> gb;
 double gc[100];
 
 // CK1: [[SIZE00:@.+]] = {{.+}}constant [1 x i[[sz:64|32]]] [i{{64|32}} 800]
-// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i32] [i32 34]
+// CK1: [[MTYPE00:@.+]] = {{.+}}constant [1 x i64] [i64 34]
 
 // CK1: [[SIZE02:@.+]] = {{.+}}constant [1 x i[[sz]]] [i[[sz]] 4]
-// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i32] [i32 33]
+// CK1: [[MTYPE02:@.+]] = {{.+}}constant [1 x i64] [i64 33]
 
-// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i32] [i32 34]
+// CK1: [[MTYPE03:@.+]] = {{.+}}constant [1 x i64] [i64 34]
 
 // CK1: [[SIZE04:@.+]] = {{.+}}constant [2 x i[[sz]]] [i[[sz]] {{8|4}}, i[[sz]] 24]
-// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i32] [i32 33, i32 17]
+// CK1: [[MTYPE04:@.+]] = {{.+}}constant [2 x i64] [i64 33, i64 17]
 
 // CK1-LABEL: _Z3fooi
 void foo(int arg) {
@@ -38,8 +38,9 @@ void foo(int arg) {
   float lb[arg];
 
   // Region 00
-  // CK1-DAG: call void @__tgt_target_data_update(i32 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
-  // CK1-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+  // CK1-DAG: call void @__tgt_target_data_update_nowait(i64 [[DEV:%[^,]+]], i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE00]]{{.+}})
+  // CK1-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+  // CK1-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -51,7 +52,7 @@ void foo(int arg) {
   // CK1-DAG: store [100 x double]* @gc, [100 x double]** [[PC0]]
 
   // CK1: %{{.+}} = add nsw i32 %{{[^,]+}}, 1
-  #pragma omp target update if(1+3-5) device(arg) from(gc)
+  #pragma omp target update if(1+3-5) device(arg) from(gc) nowait
   {++arg;}
 
   // Region 01
@@ -62,7 +63,7 @@ void foo(int arg) {
   // Region 02
   // CK1: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
   // CK1: [[IFTHEN]]
-  // CK1-DAG: call void @__tgt_target_data_update(i32 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_update(i64 4, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[SIZE02]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE02]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -85,7 +86,7 @@ void foo(int arg) {
   {++arg;}
 
   // Region 03
-  // CK1-DAG: call void @__tgt_target_data_update(i32 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_update(i64 -1, i32 1, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], i[[sz]]* [[GEPS:%.+]], {{.+}}getelementptr {{.+}}[1 x i{{.+}}]* [[MTYPE03]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
   // CK1-DAG: [[GEPS]] = getelementptr inbounds {{.+}}[[S:%[^,]+]]
@@ -108,7 +109,7 @@ void foo(int arg) {
   {++arg;}
 
   // Region 04
-  // CK1-DAG: call void @__tgt_target_data_update(i32 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
+  // CK1-DAG: call void @__tgt_target_data_update(i64 -1, i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE04]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE04]]{{.+}})
   // CK1-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
   // CK1-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
@@ -159,7 +160,7 @@ struct ST {
 };
 
 // CK2: [[SIZE00:@.+]] = {{.+}}constant [2 x i[[sz:64|32]]] [i{{64|32}} {{8|4}}, i{{64|32}} 24]
-// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i32] [i32 34, i32 18]
+// CK2: [[MTYPE00:@.+]] = {{.+}}constant [2 x i64] [i64 34, i64 18]
 
 // CK2-LABEL: _Z3bari
 int bar(int arg){
@@ -170,8 +171,9 @@ int bar(int arg){
 // Region 00
 // CK2: br i1 %{{[^,]+}}, label %[[IFTHEN:[^,]+]], label %[[IFELSE:[^,]+]]
 // CK2: [[IFTHEN]]
-// CK2-DAG: call void @__tgt_target_data_update(i32 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
-// CK2-DAG: [[DEV]] = load i32, i32* %{{[^,]+}},
+// CK2-DAG: call void @__tgt_target_data_update(i64 [[DEV:%[^,]+]], i32 2, i8** [[GEPBP:%.+]], i8** [[GEPP:%.+]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[SIZE00]], {{.+}}getelementptr {{.+}}[2 x i{{.+}}]* [[MTYPE00]]{{.+}})
+// CK2-DAG: [[DEV]] = sext i32 [[DEVi32:%[^,]+]] to i64
+// CK2-DAG: [[DEVi32]] = load i32, i32* %{{[^,]+}},
 // CK2-DAG: [[GEPBP]] = getelementptr inbounds {{.+}}[[BP:%[^,]+]]
 // CK2-DAG: [[GEPP]] = getelementptr inbounds {{.+}}[[P:%[^,]+]]
 
diff --git a/test/OpenMP/target_update_depend_messages.cpp b/test/OpenMP/target_update_depend_messages.cpp
index 64383a049104..59e159e99fdb 100644
--- a/test/OpenMP/target_update_depend_messages.cpp
+++ b/test/OpenMP/target_update_depend_messages.cpp
@@ -35,14 +35,14 @@ int tmain(T argc, S **argv, R *env[]) {
   #pragma omp target update to(z) depend(source) // expected-error {{expected expression}} expected-warning {{missing ':' after dependency type - ignoring}}
   #pragma omp target update to(z) depend(in : argc)) // expected-warning {{extra tokens at the end of '#pragma omp target update' are ignored}}
   #pragma omp target update to(z) depend(out: ) // expected-error {{expected expression}}
-  #pragma omp target update to(z) depend(inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target update to(z) depend(inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   #pragma omp target update to(z) depend(out :S1) // expected-error {{'S1' does not refer to a value}}
-  #pragma omp target update to(z) depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
-  #pragma omp target update to(z) depend(in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target update to(z) depend(in : argv[1][1] = '2') // expected-error {{expected addressable lvalue expression, array element or array section}}
+  #pragma omp target update to(z) depend(in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   #pragma omp target update to(z) depend(in : argv[0])
   #pragma omp target update to(z) depend(in : ) // expected-error {{expected expression}}
-  #pragma omp target update to(z) depend(in : tmain) // expected-error {{expected variable name, array element or array section}}
-  #pragma omp target update to(z) depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target update to(z) depend(in : tmain)
+  #pragma omp target update to(z) depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   #pragma omp target update to(z) depend(in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   #pragma omp target update to(z) depend(in : argv[ // expected-error {{expected expression}} expected-error {{expected ']'}} expected-error {{expected ')'}} expected-note {{to match this '['}} expected-note {{to match this '('}}
   #pragma omp target update to(z) depend(in : argv[: // expected-error {{expected expression}} expected-error {{expected ']'}} expected-error {{expected ')'}} expected-note {{to match this '['}} expected-note {{to match this '('}}
@@ -83,14 +83,14 @@ int main(int argc, char **argv, char *env[]) {
   #pragma omp target update to(z) depend(source) // expected-error {{expected expression}} expected-warning {{missing ':' after dependency type - ignoring}}
   #pragma omp target update to(z) depend(in : argc)) // expected-warning {{extra tokens at the end of '#pragma omp target update' are ignored}}
   #pragma omp target update to(z) depend(out: ) // expected-error {{expected expression}}
-  #pragma omp target update to(z) depend(inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp target update to(z) depend(inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   #pragma omp target update to(z) depend(out :S1) // expected-error {{'S1' does not refer to a value}}
-  #pragma omp target update to(z) depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
-  #pragma omp target update to(z) depend(in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp target update to(z) depend(in : argv[1][1] = '2')
+  #pragma omp target update to(z) depend(in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   #pragma omp target update to(z) depend(in : argv[0])
   #pragma omp target update to(z) depend(in : ) // expected-error {{expected expression}}
-  #pragma omp target update to(z) depend(in : main) // expected-error {{expected variable name, array element or array section}}
-  #pragma omp target update to(z) depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp target update to(z) depend(in : main)
+  #pragma omp target update to(z) depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   #pragma omp target update to(z) depend(in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   #pragma omp target update to(z) depend(in : argv[ // expected-error {{expected expression}} expected-error {{expected ']'}} expected-error {{expected ')'}} expected-note {{to match this '['}} expected-note {{to match this '('}}
   #pragma omp target update to(z) depend(in : argv[: // expected-error {{expected expression}} expected-error {{expected ']'}} expected-error {{expected ')'}} expected-note {{to match this '['}} expected-note {{to match this '('}}
diff --git a/test/OpenMP/target_update_from_messages.cpp b/test/OpenMP/target_update_from_messages.cpp
index 6aff083b6a48..c42bd12f75b8 100644
--- a/test/OpenMP/target_update_from_messages.cpp
+++ b/test/OpenMP/target_update_from_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -94,8 +94,8 @@ T tmain(T argc) {
 #pragma omp target update from(y x) // expected-error {{expected ',' or ')' in 'from' clause}}
 #pragma omp target update from(argc > 0 ? x : y) // expected-error 2 {{expected expression containing only member accesses and/or array sections based on named variables}} 
 #pragma omp target update from(S1) // expected-error {{'S1' does not refer to a value}}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
-#pragma omp target update from(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
-#pragma omp target update from(ba) // expected-error 2 {{type 'S2' is not mappable to target}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
+#pragma omp target update from(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
+#pragma omp target update from(ba)
 #pragma omp target update from(h) // expected-error {{threadprivate variables are not allowed in 'from' clause}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
 #pragma omp target update from(k), to(k) // expected-error 2 {{variable can appear only once in OpenMP 'target update' construct}} expected-note 2 {{used here}}
 #pragma omp target update from(t), from(t[:5]) // expected-error 2 {{variable can appear only once in OpenMP 'target update' construct}} expected-note 2 {{used here}}
@@ -148,8 +148,8 @@ int main(int argc, char **argv) {
 #pragma omp target update from(y x) // expected-error {{expected ',' or ')' in 'from' clause}}
 #pragma omp target update from(argc > 0 ? x : y) // expected-error {{expected expression containing only member accesses and/or array sections based on named variables}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
 #pragma omp target update from(S1) // expected-error {{'S1' does not refer to a value}}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
-#pragma omp target update from(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
-#pragma omp target update from(ba) // expected-error 2 {{type 'S2' is not mappable to target}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
+#pragma omp target update from(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
+#pragma omp target update from(ba)
 #pragma omp target update from(h) // expected-error {{threadprivate variables are not allowed in 'from' clause}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
 #pragma omp target update from(k), to(k) // expected-error {{variable can appear only once in OpenMP 'target update' construct}} expected-note {{used here}}
 #pragma omp target update from(t), from(t[:5]) // expected-error {{variable can appear only once in OpenMP 'target update' construct}} expected-note {{used here}}
diff --git a/test/OpenMP/target_update_to_messages.cpp b/test/OpenMP/target_update_to_messages.cpp
index 641d0bd949a2..9bde51e9839b 100644
--- a/test/OpenMP/target_update_to_messages.cpp
+++ b/test/OpenMP/target_update_to_messages.cpp
@@ -14,8 +14,8 @@ class S2 {
 public:
   S2():a(0) { }
   S2(S2 &s2):a(s2.a) { }
-  static float S2s; // expected-note 4 {{mappable type cannot contain static members}}
-  static const float S2sc; // expected-note 4 {{mappable type cannot contain static members}}
+  static float S2s;
+  static const float S2sc;
 };
 const float S2::S2sc = 0;
 const S2 b;
@@ -94,8 +94,8 @@ T tmain(T argc) {
 #pragma omp target update to(y x) // expected-error {{expected ',' or ')' in 'to' clause}}
 #pragma omp target update to(argc > 0 ? x : y) // expected-error 2 {{expected expression containing only member accesses and/or array sections based on named variables}} 
 #pragma omp target update to(S1) // expected-error {{'S1' does not refer to a value}}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
-#pragma omp target update to(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
-#pragma omp target update to(ba) // expected-error 2 {{type 'S2' is not mappable to target}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
+#pragma omp target update to(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
+#pragma omp target update to(ba)
 #pragma omp target update to(h) // expected-error {{threadprivate variables are not allowed in 'to' clause}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
 #pragma omp target update to(k), from(k) // expected-error 2 {{variable can appear only once in OpenMP 'target update' construct}} expected-note 2 {{used here}}
 #pragma omp target update to(t), to(t[:5]) // expected-error 2 {{variable can appear only once in OpenMP 'target update' construct}} expected-note 2 {{used here}}
@@ -147,8 +147,8 @@ int main(int argc, char **argv) {
 #pragma omp target update to(y x) // expected-error {{expected ',' or ')' in 'to' clause}}
 #pragma omp target update to(argc > 0 ? x : y) // expected-error {{expected expression containing only member accesses and/or array sections based on named variables}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
 #pragma omp target update to(S1) // expected-error {{'S1' does not refer to a value}}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
-#pragma omp target update to(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}} expected-error 2 {{type 'S2' is not mappable to target}}
-#pragma omp target update to(ba) // expected-error 2 {{type 'S2' is not mappable to target}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
+#pragma omp target update to(a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
+#pragma omp target update to(ba)
 #pragma omp target update to(h) // expected-error {{threadprivate variables are not allowed in 'to' clause}} expected-error {{expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'}}
 #pragma omp target update to(k), from(k) // expected-error {{variable can appear only once in OpenMP 'target update' construct}} expected-note {{used here}}
 #pragma omp target update to(t), to(t[:5]) // expected-error {{variable can appear only once in OpenMP 'target update' construct}} expected-note {{used here}}
diff --git a/test/OpenMP/target_vla_messages.cpp b/test/OpenMP/target_vla_messages.cpp
new file mode 100644
index 000000000000..c970d0def661
--- /dev/null
+++ b/test/OpenMP/target_vla_messages.cpp
@@ -0,0 +1,201 @@
+// PowerPC supports VLAs.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-unknown-unknown -emit-llvm-bc %s -o %t-ppc-host-ppc.bc
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-unknown-unknown -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host-ppc.bc -o %t-ppc-device.ll
+
+// Nvidia GPUs don't support VLAs.
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host-nvptx.bc
+// RUN: %clang_cc1 -verify -DNO_VLA -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host-nvptx.bc -o %t-nvptx-device.ll
+
+#ifndef NO_VLA
+// expected-no-diagnostics
+#endif
+
+#pragma omp declare target
+void declare(int arg) {
+  int a[2];
+#ifdef NO_VLA
+  // expected-error@+2 {{variable length arrays are not supported for the current target}}
+#endif
+  int vla[arg];
+}
+
+void declare_parallel_reduction(int arg) {
+  int a[2];
+
+#pragma omp parallel reduction(+: a)
+  { }
+
+#pragma omp parallel reduction(+: a[0:2])
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+3 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+2 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp parallel reduction(+: a[0:arg])
+  { }
+}
+#pragma omp end declare target
+
+template <typename T>
+void target_template(int arg) {  
+#pragma omp target
+  {
+#ifdef NO_VLA
+    // expected-error@+2 {{variable length arrays are not supported for the current target}}
+#endif
+    T vla[arg];
+  }
+}
+
+void target(int arg) {
+#pragma omp target
+  {
+#ifdef NO_VLA
+    // expected-error@+2 {{variable length arrays are not supported for the current target}}
+#endif
+    int vla[arg];
+  }
+
+#pragma omp target
+  {
+#pragma omp parallel
+    {
+#ifdef NO_VLA
+    // expected-error@+2 {{variable length arrays are not supported for the current target}}
+#endif
+      int vla[arg];
+    }
+  }
+
+  target_template<long>(arg);
+}
+
+void teams_reduction(int arg) {
+  int a[2];
+  int vla[arg];
+
+#pragma omp target map(a)
+#pragma omp teams reduction(+: a)
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+4 {{cannot generate code for reduction on variable length array}}
+  // expected-note@+3 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(vla)
+#pragma omp teams reduction(+: vla)
+  { }
+  
+#pragma omp target map(a[0:2])
+#pragma omp teams reduction(+: a[0:2])
+  { }
+
+#pragma omp target map(vla[0:2])
+#pragma omp teams reduction(+: vla[0:2])
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+4 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+3 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(a[0:arg])
+#pragma omp teams reduction(+: a[0:arg])
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+4 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+3 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(vla[0:arg])
+#pragma omp teams reduction(+: vla[0:arg])
+  { }
+}
+
+void parallel_reduction(int arg) {
+  int a[2];
+  int vla[arg];
+
+#pragma omp target map(a)
+#pragma omp parallel reduction(+: a)
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+4 {{cannot generate code for reduction on variable length array}}
+  // expected-note@+3 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(vla)
+#pragma omp parallel reduction(+: vla)
+  { }
+
+#pragma omp target map(a[0:2])
+#pragma omp parallel reduction(+: a[0:2])
+  { }
+
+#pragma omp target map(vla[0:2])
+#pragma omp parallel reduction(+: vla[0:2])
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+4 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+3 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(a[0:arg])
+#pragma omp parallel reduction(+: a[0:arg])
+  { }
+
+#ifdef NO_VLA
+  // expected-error@+4 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+3 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(vla[0:arg])
+#pragma omp parallel reduction(+: vla[0:arg])
+  { }
+}
+
+void for_reduction(int arg) {
+  int a[2];
+  int vla[arg];
+
+#pragma omp target map(a)
+#pragma omp parallel
+#pragma omp for reduction(+: a)
+  for (int i = 0; i < arg; i++) ;
+
+#ifdef NO_VLA
+  // expected-error@+5 {{cannot generate code for reduction on variable length array}}
+  // expected-note@+4 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(vla)
+#pragma omp parallel
+#pragma omp for reduction(+: vla)
+  for (int i = 0; i < arg; i++) ;
+
+#pragma omp target map(a[0:2])
+#pragma omp parallel
+#pragma omp for reduction(+: a[0:2])
+  for (int i = 0; i < arg; i++) ;
+
+#pragma omp target map(vla[0:2])
+#pragma omp parallel
+#pragma omp for reduction(+: vla[0:2])
+  for (int i = 0; i < arg; i++) ;
+
+#ifdef NO_VLA
+  // expected-error@+5 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+4 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(a[0:arg])
+#pragma omp parallel
+#pragma omp for reduction(+: a[0:arg])
+  for (int i = 0; i < arg; i++) ;
+
+#ifdef NO_VLA
+  // expected-error@+5 {{cannot generate code for reduction on array section, which requires a variable length array}}
+  // expected-note@+4 {{variable length arrays are not supported for the current target}}
+#endif
+#pragma omp target map(vla[0:arg])
+#pragma omp parallel
+#pragma omp for reduction(+: vla[0:arg])
+  for (int i = 0; i < arg; i++) ;
+}
diff --git a/test/OpenMP/task_ast_print.cpp b/test/OpenMP/task_ast_print.cpp
index 9a7d64ee5462..54c05d8ed410 100644
--- a/test/OpenMP/task_ast_print.cpp
+++ b/test/OpenMP/task_ast_print.cpp
@@ -13,17 +13,19 @@ struct S1 {
   S1(int v) : a(v) {}
   int a;
   typedef int type;
+  S1 operator +(const S1&);
 };
 
 template <typename T>
 class S7 : public T {
 protected:
-  T a;
+  T a, b;
   S7() : a(0) {}
 
 public:
   S7(typename T::type v) : a(v) {
-#pragma omp task private(a) private(this->a) private(T::a)
+#pragma omp taskgroup task_reduction(+:b)
+#pragma omp task private(a) private(this->a) private(T::a) in_reduction(+:this->b)
     for (int k = 0; k < a.a; ++k)
       ++this->a.a;
   }
@@ -35,7 +37,8 @@ public:
   }
 };
 
-// CHECK: #pragma omp task private(this->a) private(this->a) private(T::a)
+// CHECK: #pragma omp taskgroup task_reduction(+: this->b)
+// CHECK: #pragma omp task private(this->a) private(this->a) private(T::a) in_reduction(+: this->b)
 // CHECK: #pragma omp task private(this->a) private(this->a)
 // CHECK: #pragma omp task private(this->a) private(this->a) private(this->S1::a)
 
@@ -89,7 +92,8 @@ T tmain(T argc, T *argv) {
   a = 2;
 #pragma omp task default(none), private(argc, b) firstprivate(argv) shared(d) if (argc > 0) final(S<T>::TS > 0) priority(argc)
   foo();
-#pragma omp task if (C) mergeable priority(C)
+#pragma omp taskgroup task_reduction(-: argc)
+#pragma omp task if (C) mergeable priority(C) in_reduction(-: argc)
   foo();
   return 0;
 }
@@ -103,7 +107,8 @@ T tmain(T argc, T *argv) {
 // CHECK-NEXT: a = 2;
 // CHECK-NEXT: #pragma omp task default(none) private(argc,b) firstprivate(argv) shared(d) if(argc > 0) final(S<T>::TS > 0) priority(argc)
 // CHECK-NEXT: foo()
-// CHECK-NEXT: #pragma omp task if(C) mergeable priority(C)
+// CHECK-NEXT: #pragma omp taskgroup task_reduction(-: argc)
+// CHECK-NEXT: #pragma omp task if(C) mergeable priority(C) in_reduction(-: argc)
 // CHECK-NEXT: foo()
 // CHECK: template<> int tmain<int, 5>(int argc, int *argv) {
 // CHECK-NEXT: int b = argc, c, d, e, f, g;
@@ -114,7 +119,8 @@ T tmain(T argc, T *argv) {
 // CHECK-NEXT: a = 2;
 // CHECK-NEXT: #pragma omp task default(none) private(argc,b) firstprivate(argv) shared(d) if(argc > 0) final(S<int>::TS > 0) priority(argc)
 // CHECK-NEXT: foo()
-// CHECK-NEXT: #pragma omp task if(5) mergeable priority(5)
+// CHECK-NEXT: #pragma omp taskgroup task_reduction(-: argc)
+// CHECK-NEXT: #pragma omp task if(5) mergeable priority(5) in_reduction(-: argc)
 // CHECK-NEXT: foo()
 // CHECK: template<> long tmain<long, 1>(long argc, long *argv) {
 // CHECK-NEXT: long b = argc, c, d, e, f, g;
@@ -125,7 +131,8 @@ T tmain(T argc, T *argv) {
 // CHECK-NEXT: a = 2;
 // CHECK-NEXT: #pragma omp task default(none) private(argc,b) firstprivate(argv) shared(d) if(argc > 0) final(S<long>::TS > 0) priority(argc)
 // CHECK-NEXT: foo()
-// CHECK-NEXT: #pragma omp task if(1) mergeable priority(1)
+// CHECK-NEXT: #pragma omp taskgroup task_reduction(-: argc)
+// CHECK-NEXT: #pragma omp task if(1) mergeable priority(1) in_reduction(-: argc)
 // CHECK-NEXT: foo()
 
 enum Enum {};
@@ -134,7 +141,7 @@ int main(int argc, char **argv) {
   long x;
   int b = argc, c, d, e, f, g;
   static int a;
-  int arr[10];
+  int arr[10], arr1[argc];
 #pragma omp threadprivate(a)
   Enum ee;
 // CHECK: Enum ee;
@@ -142,8 +149,18 @@ int main(int argc, char **argv) {
   // CHECK-NEXT: #pragma omp task untied mergeable depend(out : argv[:a][1],(arr)[0:]) if(task: argc > 0) priority(f)
   a = 2;
 // CHECK-NEXT: a = 2;
-#pragma omp task default(none), private(argc, b) firstprivate(argv) if (argc > 0) final(a > 0) depend(inout : a, argv[:argc],arr[:a]) priority(23)
-  // CHECK-NEXT: #pragma omp task default(none) private(argc,b) firstprivate(argv) if(argc > 0) final(a > 0) depend(inout : a,argv[:argc],arr[:a]) priority(23)
+#pragma omp taskgroup task_reduction(min: arr1)
+#pragma omp task default(none), private(argc, b) firstprivate(argv) if (argc > 0) final(a > 0) depend(inout : a, argv[:argc],arr[:a]) priority(23) in_reduction(min: arr1)
+  // CHECK-NEXT: #pragma omp taskgroup task_reduction(min: arr1)
+  // CHECK-NEXT: #pragma omp task default(none) private(argc,b) firstprivate(argv) if(argc > 0) final(a > 0) depend(inout : a,argv[:argc],arr[:a]) priority(23) in_reduction(min: arr1)
+  foo();
+  // CHECK-NEXT: foo();
+#pragma omp taskgroup task_reduction(min: arr1)
+#pragma omp parallel reduction(+:arr1)
+#pragma omp task in_reduction(min: arr1)
+  // CHECK-NEXT: #pragma omp taskgroup task_reduction(min: arr1)
+  // CHECK-NEXT: #pragma omp parallel reduction(+: arr1)
+  // CHECK-NEXT: #pragma omp task in_reduction(min: arr1)
   foo();
   // CHECK-NEXT: foo();
   return tmain<int, 5>(b, &b) + tmain<long, 1>(x, &x);
diff --git a/test/OpenMP/task_depend_messages.cpp b/test/OpenMP/task_depend_messages.cpp
index 576738ceac46..1e51b479dd96 100644
--- a/test/OpenMP/task_depend_messages.cpp
+++ b/test/OpenMP/task_depend_messages.cpp
@@ -28,14 +28,14 @@ int main(int argc, char **argv, char *env[]) {
   #pragma omp task depend (source) // expected-error {{expected expression}} expected-warning {{missing ':' after dependency type - ignoring}}
   #pragma omp task depend (in : argc)) // expected-warning {{extra tokens at the end of '#pragma omp task' are ignored}}
   #pragma omp task depend (out: ) // expected-error {{expected expression}}
-  #pragma omp task depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected variable name, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
+  #pragma omp task depend (inout : foobool(argc)), depend (in, argc) // expected-error {{expected addressable lvalue expression, array element or array section}} expected-warning {{missing ':' after dependency type - ignoring}} expected-error {{expected expression}}
   #pragma omp task depend (out :S1) // expected-error {{'S1' does not refer to a value}}
-  #pragma omp task depend(in : argv[1][1] = '2') // expected-error {{expected variable name, array element or array section}}
-  #pragma omp task depend (in : vec[1]) // expected-error {{expected variable name, array element or array section}}
+  #pragma omp task depend(in : argv[1][1] = '2')
+  #pragma omp task depend (in : vec[1]) // expected-error {{expected addressable lvalue expression, array element or array section}}
   #pragma omp task depend (in : argv[0])
   #pragma omp task depend (in : ) // expected-error {{expected expression}}
-  #pragma omp task depend (in : main) // expected-error {{expected variable name, array element or array section}}
-  #pragma omp task depend(in : a[0]) // expected-error{{expected variable name, array element or array section}}
+  #pragma omp task depend (in : main)
+  #pragma omp task depend(in : a[0]) // expected-error{{expected addressable lvalue expression, array element or array section}}
   #pragma omp task depend (in : vec[1:2]) // expected-error {{ value is not an array or pointer}}
   #pragma omp task depend (in : argv[ // expected-error {{expected expression}} expected-error {{expected ']'}} expected-error {{expected ')'}} expected-note {{to match this '['}} expected-note {{to match this '('}}
   #pragma omp task depend (in : argv[: // expected-error {{expected expression}} expected-error {{expected ']'}} expected-error {{expected ')'}} expected-note {{to match this '['}} expected-note {{to match this '('}}
diff --git a/test/OpenMP/task_in_reduction_codegen.cpp b/test/OpenMP/task_in_reduction_codegen.cpp
new file mode 100644
index 000000000000..2ad013a14a57
--- /dev/null
+++ b/test/OpenMP/task_in_reduction_codegen.cpp
@@ -0,0 +1,81 @@
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c++ -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[PRIVATES:%.+]] = type { i8*, i8* }
+
+struct S {
+  int a;
+  S() : a(0) {}
+  S(const S&) {}
+  S& operator=(const S&) {return *this;}
+  ~S() {}
+  friend S operator+(const S&a, const S&b) {return a;}
+};
+
+
+int main(int argc, char **argv) {
+  int a;
+  float b;
+  S c[5];
+  short d[argc];
+#pragma omp taskgroup task_reduction(+: a, b, argc)
+  {
+#pragma omp taskgroup task_reduction(-:c, d)
+#pragma omp parallel
+#pragma omp task in_reduction(+:a) in_reduction(-:d)
+    a += d[a];
+  }
+  return 0;
+}
+
+// CHECK-LABEL: @main
+// CHECK:       void @__kmpc_taskgroup(%ident_t* @0, i32 [[GTID:%.+]])
+// CHECK:       [[TD1:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 3, i8* %
+// CHECK-NEXT:  store i8* [[TD1]], i8** [[TD1_ADDR:%[^,]+]],
+// CHECK-NEXT:  call void @__kmpc_taskgroup(%ident_t* @0, i32 [[GTID]])
+// CHECK:       [[TD2:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 2, i8* %
+// CHECK-NEXT:  store i8* [[TD2]], i8** [[TD2_ADDR:%[^,]+]],
+// CHECK-NEXT:  call void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* @0, i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i16*, i8**, i8**)* [[OMP_PARALLEL:@.+]] to void (i32*, i32*, ...)*), i32* %{{.+}}, i64 %{{.+}}, i16* %{{.+}}, i8** [[TD1_ADDR]], i8** [[TD2_ADDR]])
+// CHECK-NEXT:  call void @__kmpc_end_taskgroup(%ident_t* @0, i32 [[GTID]])
+// CHECK-NEXT:  call void @__kmpc_end_taskgroup(%ident_t* @0, i32 [[GTID]])
+
+// CHECK:       define internal void [[OMP_PARALLEL]](
+// CHECK:       [[TASK_T:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* @0, i32 [[GTID:%.+]], i32 1, i64 56, i64 40, i32 (i32, i8*)* bitcast (i32 (i32, [[T:%.+]]*)* [[OMP_TASK:@.+]] to i32 (i32, i8*)*))
+// CHECK-NEXT:  [[TASK_T_WITH_PRIVS:%.+]] = bitcast i8* [[TASK_T]] to [[T]]*
+// CHECK:       [[PRIVS:%.+]] = getelementptr inbounds [[T]], [[T]]* [[TASK_T_WITH_PRIVS]], i32 0, i32 1
+// CHECK:       [[TD1_REF:%.+]] = getelementptr inbounds [[PRIVATES]], [[PRIVATES]]* [[PRIVS]], i32 0, i32 0
+// CHECK-NEXT:  [[TD1_SHAR:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[TD1_ADDR:%.+]] = load i8**, i8*** [[TD1_SHAR]],
+// CHECK-NEXT:  [[TD1:%.+]] = load i8*, i8** [[TD1_ADDR]],
+// CHECK-NEXT:  store i8* [[TD1]], i8** [[TD1_REF]],
+// CHECK-NEXT:  [[TD2_REF:%.+]] = getelementptr inbounds [[PRIVATES]], [[PRIVATES]]* [[PRIVS]], i32 0, i32 1
+// CHECK-NEXT:  [[TD2_SHAR:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[TD2_ADDR:%.+]] = load i8**, i8*** [[TD2_SHAR]],
+// CHECK-NEXT:  [[TD2:%.+]] = load i8*, i8** [[TD2_ADDR]],
+// CHECK-NEXT:  store i8* [[TD2]], i8** [[TD2_REF]],
+// CHECK-NEXT:  call i32 @__kmpc_omp_task(%ident_t* @0, i32 [[GTID]], i8* [[TASK_T]])
+// CHECK-NEXT:  ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal {{.*}} [[OMP_TASK]](
+// CHECK:       call void (i8*, ...) %{{[^(]+}}(i8* %{{.+}}, i8*** [[TD1_REF:%[^,]+]], i8*** [[TD2_REF:%[^,]+]])
+// CHECK-NEXT:  [[TD1_ADDR:%.+]] = load i8**, i8*** [[TD1_REF]],
+// CHECK-NEXT:  [[TD2_ADDR:%.+]] = load i8**, i8*** [[TD2_REF]],
+// CHECK-NEXT:  [[A_REF:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[A_ADDR:%.+]] = load i32*, i32** [[A_REF]],
+// CHECK-NEXT:  [[TD1:%.+]] = load i8*, i8** [[TD1_ADDR]],
+// CHECK-NEXT:  [[GTID:%.+]] = load i32, i32* %
+// CHECK-NEXT:  [[A_PTR:%.+]] = bitcast i32* [[A_ADDR]] to i8*
+// CHECK-NEXT:  call i8* @__kmpc_task_reduction_get_th_data(i32 [[GTID]], i8* [[TD1]], i8* [[A_PTR]])
+// CHECK:       [[D_REF:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[D_ADDR:%.+]] = load i16*, i16** [[D_REF]],
+// CHECK:       [[TD2:%.+]] = load i8*, i8** [[TD2_ADDR]],
+// CHECK-NEXT:  [[D_PTR:%.+]] = bitcast i16* [[D_ADDR]] to i8*
+// CHECK-NEXT:  call i8* @__kmpc_task_reduction_get_th_data(i32 [[GTID]], i8* [[TD2]], i8* [[D_PTR]])
+// CHECK:       add nsw i32
+// CHECK:       store i32 %
+#endif
diff --git a/test/OpenMP/task_in_reduction_message.cpp b/test/OpenMP/task_in_reduction_message.cpp
new file mode 100644
index 000000000000..e176dc9305b3
--- /dev/null
+++ b/test/OpenMP/task_in_reduction_message.cpp
@@ -0,0 +1,308 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++98 -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++11 -ferror-limit 150 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+void foobar(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref)
+#pragma omp task in_reduction(+:ref)
+  foo();
+}
+
+void foobar1(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref)
+#pragma omp task in_reduction(-:ref)
+  foo();
+}
+
+#pragma omp declare reduction (red:int:omp_out += omp_in)
+
+void foobar2(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref) // expected-note {{previously marked as task_reduction with different reduction operation}}
+#pragma omp task in_reduction(red:ref) // expected-error{{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  foo();
+}
+
+void foobar3(int &ref) {
+#pragma omp taskgroup task_reduction(red:ref) // expected-note {{previously marked as task_reduction with different reduction operation}}
+#pragma omp task in_reduction(min:ref)  // expected-error{{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  foo();
+}
+
+void foobar4(int &ref) {
+#pragma omp task in_reduction(min:ref) // expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  foo();
+}
+
+struct S1; // expected-note {{declared here}} expected-note 4 {{forward declaration of 'S1'}}
+extern S1 a;
+class S2 {
+  mutable int a;
+  S2 &operator+(const S2 &arg) { return (*this); } // expected-note 3 {{implicitly declared private here}}
+
+public:
+  S2() : a(0) {}
+  S2(S2 &s2) : a(s2.a) {}
+  static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
+};
+const float S2::S2sc = 0;
+S2 b;                     // expected-note 3 {{'b' defined here}}
+const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
+class S3 {
+  int a;
+
+public:
+  int b;
+  S3() : a(0) {}
+  S3(const S3 &s3) : a(s3.a) {}
+  S3 operator+(const S3 &arg1) { return arg1; }
+};
+int operator+(const S3 &arg1, const S3 &arg2) { return 5; }
+S3 c;               // expected-note 3 {{'c' defined here}}
+const S3 ca[5];     // expected-note 2 {{'ca' defined here}}
+extern const int f; // expected-note 4 {{'f' declared here}}
+class S4 {
+  int a;
+  S4(); // expected-note {{implicitly declared private here}}
+  S4(const S4 &s4);
+  S4 &operator+(const S4 &arg) { return (*this); }
+
+public:
+  S4(int v) : a(v) {}
+};
+S4 &operator&=(S4 &arg1, S4 &arg2) { return arg1; }
+class S5 {
+  int a;
+  S5() : a(0) {} // expected-note {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {}
+  S5 &operator+(const S5 &arg);
+
+public:
+  S5(int v) : a(v) {}
+};
+class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
+#if __cplusplus >= 201103L // C++11 or later
+// expected-note@-2 3 {{candidate function (the implicit move assignment operator) not viable}}
+#endif
+  int a;
+
+public:
+  S6() : a(6) {}
+  operator int() { return 6; }
+} o;
+
+S3 h, k;
+#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
+
+template <class T>       // expected-note {{declared here}}
+T tmain(T argc) {
+  const T d = T();       // expected-note 4 {{'d' defined here}}
+  const T da[5] = {T()}; // expected-note 2 {{'da' defined here}}
+  T qa[5] = {T()};
+  T i;
+  T &j = i;                    // expected-note 2 {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const T &r = da[(int)i];     // expected-note 2 {{'r' defined here}}
+  T &q = qa[(int)i];
+  T fl;
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction // expected-error {{expected '(' after 'in_reduction'}}
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction + // expected-error {{expected '(' after 'in_reduction'}} expected-warning {{extra tokens at the end of '#pragma omp task' are ignored}}
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp task in_reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp taskgroup task_reduction(&:argc) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+#pragma omp task in_reduction(& : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp taskgroup task_reduction(|:argc) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+#pragma omp task in_reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp task in_reduction(|| : argc ? i : argc) // expected-error 2 {{expected variable name, array element or array section}}
+  foo();
+#pragma omp task in_reduction(foo : argc) //expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'float'}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'int'}}
+  foo();
+#pragma omp taskgroup task_reduction(&&:argc)
+#pragma omp task in_reduction(&& : argc)
+  foo();
+#pragma omp task in_reduction(^ : T) // expected-error {{'T' does not refer to a value}}
+  foo();
+#pragma omp taskgroup task_reduction(+:c)
+#pragma omp task in_reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 3 {{const-qualified list item cannot be reduction}} expected-error 2 {{'operator+' is a private member of 'S2'}} expected-error 2 {{in_reduction variable must appear in a task_reduction clause}}
+  foo();
+#pragma omp task in_reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 4 {{arguments of OpenMP clause 'in_reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 3 {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  foo();
+#pragma omp task in_reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}} expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}} expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  foo();
+#pragma omp task in_reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp taskgroup task_reduction(+:k)
+#pragma omp task in_reduction(+ : h, k) // expected-error {{threadprivate or thread local variable cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(+ : o) // expected-error 2 {{no viable overloaded '='}}
+  foo();
+#pragma omp parallel private(k)
+#pragma omp task in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp taskgroup task_reduction(+:p)
+#pragma omp task in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{variable can appear only once in OpenMP 'in_reduction' clause}} expected-note 2 {{previously referenced here}}
+  foo();
+#pragma omp task in_reduction(+ : r) // expected-error 2 {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp task in_reduction(max : j) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp task in_reduction(+ : fl)
+    foo();
+#pragma omp parallel
+#pragma omp for reduction(- : fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp task in_reduction(+ : fl)
+    foo();
+
+  return T();
+}
+
+namespace A {
+double x;
+#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  const int d = 5;       // expected-note 2 {{'d' defined here}}
+  const int da[5] = {0}; // expected-note {{'da' defined here}}
+  int qa[5] = {0};
+  S4 e(4);
+  S5 g(5);
+  int i;
+  int &j = i;                  // expected-note {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const int &r = da[i];        // expected-note {{'r' defined here}}
+  int &q = qa[i];
+  float fl;
+#pragma omp task in_reduction // expected-error {{expected '(' after 'in_reduction'}}
+  foo();
+#pragma omp task in_reduction + // expected-error {{expected '(' after 'in_reduction'}} expected-warning {{extra tokens at the end of '#pragma omp task' are ignored}}
+  foo();
+#pragma omp task in_reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp task in_reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  foo();
+#pragma omp task in_reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp task in_reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp task in_reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  foo();
+#pragma omp task in_reduction(foo : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max'}}
+  foo();
+#pragma omp taskgroup task_reduction(|:argc)
+{
+#pragma omp taskgroup task_reduction(+:argc) // expected-note {{previously marked as task_reduction with different reduction operation}}
+{
+#pragma omp task in_reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  foo();
+}
+#pragma omp task in_reduction(| : argc)
+  foo();
+}
+#pragma omp task in_reduction(|| : argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name, array element or array section}}
+  foo();
+#pragma omp task in_reduction(~ : argc) // expected-error {{expected unqualified-id}}
+  foo();
+#pragma omp taskgroup task_reduction(&&:argc)
+#pragma omp task in_reduction(&& : argc)
+  foo();
+#pragma omp task in_reduction(^ : S1) // expected-error {{'S1' does not refer to a value}}
+  foo();
+#pragma omp taskgroup task_reduction(+:c)
+#pragma omp task in_reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{const-qualified list item cannot be reduction}} expected-error {{'operator+' is a private member of 'S2'}} expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  foo();
+#pragma omp task in_reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{arguments of OpenMP clause 'in_reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 2 {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  foo();
+#pragma omp task in_reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  foo();
+#pragma omp task in_reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(& : e, g) // expected-error {{calling a private constructor of class 'S4'}} expected-error {{nvalid operands to binary expression ('S4' and 'S4')}} expected-error {{calling a private constructor of class 'S5'}} expected-error {{invalid operands to binary expression ('S5' and 'S5')}}
+  foo();
+#pragma omp taskgroup task_reduction(+:k)
+#pragma omp task in_reduction(+ : h, k, B::x) // expected-error 2 {{threadprivate or thread local variable cannot be reduction}}
+  foo();
+#pragma omp task in_reduction(+ : o) // expected-error {{no viable overloaded '='}}
+  foo();
+#pragma omp parallel private(k)
+#pragma omp task in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp taskgroup task_reduction(+:p)
+#pragma omp task in_reduction(+ : p), in_reduction(+ : p) // expected-error {{variable can appear only once in OpenMP 'in_reduction' clause}} expected-note {{previously referenced here}}
+  foo();
+#pragma omp task in_reduction(+ : r) // expected-error {{const-qualified list item cannot be reduction}}
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp task in_reduction(max : j) // expected-error {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  foo();
+#pragma omp parallel
+#pragma omp for private(fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp task in_reduction(+ : fl)
+    foo();
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp task in_reduction(+ : fl)
+    foo();
+  static int m;
+#pragma omp taskgroup task_reduction(+:m)
+#pragma omp task in_reduction(+ : m) // OK
+  m++;
+
+  return tmain(argc) + tmain(fl); // expected-note {{in instantiation of function template specialization 'tmain<int>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<float>' requested here}}
+}
diff --git a/test/OpenMP/taskgroup_task_reduction_codegen.cpp b/test/OpenMP/taskgroup_task_reduction_codegen.cpp
new file mode 100644
index 000000000000..63e015182420
--- /dev/null
+++ b/test/OpenMP/taskgroup_task_reduction_codegen.cpp
@@ -0,0 +1,212 @@
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c++ -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ %s -verify -debug-info-kind=limited -emit-llvm -o - -triple x86_64-apple-darwin10 | FileCheck %s --check-prefix=CHECK --check-prefix=DEBUG
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct S {
+  int a;
+  S() : a(0) {}
+  S(const S&) {}
+  S& operator=(const S&) {return *this;}
+  ~S() {}
+  friend S operator+(const S&a, const S&b) {return a;}
+};
+
+
+int main(int argc, char **argv) {
+  int a;
+  float b;
+  S c[5];
+  short d[argc];
+#pragma omp taskgroup task_reduction(+: a, b, argc)
+  {
+#pragma omp taskgroup task_reduction(-:c, d)
+    ;
+  }
+  return 0;
+}
+// CHECK-LABEL: @main
+// CHECK:       alloca i32,
+// CHECK:       [[ARGC_ADDR:%.+]] = alloca i32,
+// CHECK:       [[ARGV_ADDR:%.+]] = alloca i8**,
+// CHECK:       [[A:%.+]] = alloca i32,
+// CHECK:       [[B:%.+]] = alloca float,
+// CHECK:       [[C:%.+]] = alloca [5 x %struct.S],
+// CHECK:       [[GTID:%.+]] = call i32 @__kmpc_global_thread_num(%ident_t*
+// CHECK:       [[RD_IN1:%.+]] = alloca [3 x [[T1:%[^,]+]]],
+// CHECK:       [[TD1:%.+]] = alloca i8*,
+// CHECK:       [[RD_IN2:%.+]] = alloca [2 x [[T2:%[^,]+]]],
+// CHECK:       [[TD2:%.+]] = alloca i8*,
+
+// CHECK:       [[VLA:%.+]] = alloca i16, i64 [[VLA_SIZE:%[^,]+]],
+
+// CHECK:       call void @__kmpc_taskgroup(%ident_t* {{[^,]+}}, i32 [[GTID]])
+// CHECK-DAG:   [[BC_A:%.+]] = bitcast i32* [[A]] to i8*
+// CHECK-DAG:   store i8* [[BC_A]], i8** [[A_REF:[^,]+]],
+// CHECK-DAG:   [[A_REF]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPA:%[^,]+]], i32 0, i32 0
+// CHECK-DAG:   [[GEPA]] = getelementptr inbounds [3 x [[T1]]], [3 x [[T1]]]* [[RD_IN1]], i64 0, i64
+// CHECK-DAG:   [[TMP6:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPA]], i32 0, i32 1
+// CHECK-DAG:   store i64 4, i64* [[TMP6]],
+// CHECK-DAG:   [[TMP7:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPA]], i32 0, i32 2
+// CHECK-DAG:   store i8* bitcast (void (i8*)* [[AINIT:@.+]] to i8*), i8** [[TMP7]],
+// CHECK-DAG:   [[TMP8:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPA]], i32 0, i32 3
+// CHECK-DAG:   store i8* null, i8** [[TMP8]],
+// CHECK-DAG:   [[TMP9:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPA]], i32 0, i32 4
+// CHECK-DAG:   store i8* bitcast (void (i8*, i8*)* [[ACOMB:@.+]] to i8*), i8** [[TMP9]],
+// CHECK-DAG:   [[TMP10:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPA]], i32 0, i32 5
+// CHECK-DAG:   [[TMP11:%.+]] = bitcast i32* [[TMP10]] to i8*
+// CHECK-DAG:   call void @llvm.memset.p0i8.i64(i8* [[TMP11]], i8 0, i64 4, i32 8, i1 false)
+// CHECK-DAG:   [[TMP13:%.+]] = bitcast float* [[B]] to i8*
+// CHECK-DAG:   store i8* [[TMP13]], i8** [[TMP12:%[^,]+]],
+// CHECK-DAG:   [[TMP12]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPB:%[^,]+]], i32 0, i32 0
+// CHECK-DAG:   [[GEPB]] = getelementptr inbounds [3 x [[T1]]], [3 x [[T1]]]* [[RD_IN1]], i64 0, i64
+// CHECK-DAG:   [[TMP14:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPB]], i32 0, i32 1
+// CHECK-DAG:   store i64 4, i64* [[TMP14]],
+// CHECK-DAG:   [[TMP15:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPB]], i32 0, i32 2
+// CHECK-DAG:   store i8* bitcast (void (i8*)* [[BINIT:@.+]] to i8*), i8** [[TMP15]],
+// CHECK-DAG:   [[TMP16:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPB]], i32 0, i32 3
+// CHECK-DAG:   store i8* null, i8** [[TMP16]],
+// CHECK-DAG:   [[TMP17:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPB]], i32 0, i32 4
+// CHECK-DAG:   store i8* bitcast (void (i8*, i8*)* [[BCOMB:@.+]] to i8*), i8** [[TMP17]],
+// CHECK-DAG:   [[TMP18:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPB]], i32 0, i32 5
+// CHECK-DAG:   [[TMP19:%.+]] = bitcast i32* [[TMP18]] to i8*
+// CHECK-DAG:   call void @llvm.memset.p0i8.i64(i8* [[TMP19]], i8 0, i64 4, i32 8, i1 false)
+// CHECK-DAG:   [[TMP21:%.+]] = bitcast i32* [[ARGC_ADDR]] to i8*
+// CHECK-DAG:   store i8* [[TMP21]], i8** [[TMP20:%[^,]+]],
+// CHECK-DAG:   [[TMP20]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPARGC:%[^,]+]], i32 0, i32 0
+// CHECK-DAG:   [[GEPARGC]] = getelementptr inbounds [3 x [[T1]]], [3 x [[T1]]]* [[RD_IN1]], i64 0, i64
+// CHECK-DAG:   [[TMP22:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPARGC]], i32 0, i32 1
+// CHECK-DAG:   store i64 4, i64* [[TMP22]],
+// CHECK-DAG:   [[TMP23:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPARGC]], i32 0, i32 2
+// CHECK-DAG:   store i8* bitcast (void (i8*)* [[ARGCINIT:@.+]] to i8*), i8** [[TMP23]],
+// CHECK-DAG:   [[TMP24:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPARGC]], i32 0, i32 3
+// CHECK-DAG:   store i8* null, i8** [[TMP24]],
+// CHECK-DAG:   [[TMP25:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPARGC]], i32 0, i32 4
+// CHECK-DAG:   store i8* bitcast (void (i8*, i8*)* [[ARGCCOMB:@.+]] to i8*), i8** [[TMP25]],
+// CHECK-DAG:   [[TMP26:%.+]] = getelementptr inbounds [[T1]], [[T1]]* [[GEPARGC]], i32 0, i32 5
+// CHECK-DAG:   [[TMP27:%.+]] = bitcast i32* [[TMP26]] to i8*
+// CHECK-DAG:   call void @llvm.memset.p0i8.i64(i8* [[TMP27]], i8 0, i64 4, i32 8, i1 false)
+// CHECK-DAG:   [[TMP28:%.+]] = bitcast [3 x [[T1]]]* [[RD_IN1]] to i8*
+// CHECK-DAG:   [[TMP29:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 3, i8* [[TMP28]])
+// DEBUG-DAG:   call void @llvm.dbg.declare(metadata i8** [[TD1]],
+// CHECK-DAG:   store i8* [[TMP29]], i8** [[TD1]],
+// CHECK-DAG:   call void @__kmpc_taskgroup(%ident_t* {{[^,]+}}, i32 [[GTID]])
+// CHECK-DAG:   [[TMP31:%.+]] = bitcast [5 x %struct.S]* [[C]] to i8*
+// CHECK-DAG:   store i8* [[TMP31]], i8** [[TMP30:%[^,]+]],
+// CHECK-DAG:   [[TMP30]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPC:%[^,]+]], i32 0, i32 0
+// CHECK-DAG:   [[GEPC]] = getelementptr inbounds [2 x [[T2]]], [2 x [[T2]]]* [[RD_IN2]], i64 0, i64
+// CHECK-DAG:   [[TMP32:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPC]], i32 0, i32 1
+// CHECK-DAG:   store i64 20, i64* [[TMP32]],
+// CHECK-DAG:   [[TMP33:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPC]], i32 0, i32 2
+// CHECK-DAG:   store i8* bitcast (void (i8*)* [[CINIT:@.+]] to i8*), i8** [[TMP33]],
+// CHECK-DAG:   [[TMP34:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPC]], i32 0, i32 3
+// CHECK-DAG:   store i8* bitcast (void (i8*)* [[CFINI:@.+]] to i8*), i8** [[TMP34]],
+// CHECK-DAG:   [[TMP35:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPC]], i32 0, i32 4
+// CHECK-DAG:   store i8* bitcast (void (i8*, i8*)* [[CCOMB:@.+]] to i8*), i8** [[TMP35]],
+// CHECK-DAG:   [[TMP36:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPC]], i32 0, i32 5
+// CHECK-DAG:   [[TMP37:%.+]] = bitcast i32* [[TMP36]] to i8*
+// CHECK-DAG:   call void @llvm.memset.p0i8.i64(i8* [[TMP37]], i8 0, i64 4, i32 8, i1 false)
+// CHECK-DAG:   [[TMP39:%.+]] = bitcast i16* [[VLA]] to i8*
+// CHECK-DAG:   store i8* [[TMP39]], i8** [[TMP38:%[^,]+]],
+// CHECK-DAG:   [[TMP38]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPVLA:%[^,]+]], i32 0, i32 0
+// CHECK-DAG:   [[GEPVLA]] = getelementptr inbounds [2 x [[T2]]], [2 x [[T2]]]* [[RD_IN2]], i64 0, i64
+// CHECK-DAG:   [[TMP40:%.+]] = mul nuw i64 [[VLA_SIZE]], 2
+// CHECK-DAG:   [[TMP41:%.+]] = udiv exact i64 [[TMP40]], ptrtoint (i16* getelementptr (i16, i16* null, i32 1) to i64)
+// CHECK-DAG:   [[TMP42:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPVLA]], i32 0, i32 1
+// CHECK-DAG:   store i64 [[TMP40]], i64* [[TMP42]],
+// CHECK-DAG:   [[TMP43:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPVLA]], i32 0, i32 2
+// CHECK-DAG:   store i8* bitcast (void (i8*)* [[VLAINIT:@.+]] to i8*), i8** [[TMP43]],
+// CHECK-DAG:   [[TMP44:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPVLA]], i32 0, i32 3
+// CHECK-DAG:   store i8* null, i8** [[TMP44]],
+// CHECK-DAG:   [[TMP45:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPVLA]], i32 0, i32 4
+// CHECK-DAG:   store i8* bitcast (void (i8*, i8*)* [[VLACOMB:@.+]] to i8*), i8** [[TMP45]],
+// CHECK-DAG:   [[TMP46:%.+]] = getelementptr inbounds [[T2]], [[T2]]* [[GEPVLA]], i32 0, i32 5
+// CHECK-DAG:   store i32 1, i32* [[TMP46]],
+// CHECK:       [[TMP47:%.+]] = bitcast [2 x [[T2]]]* [[RD_IN2]] to i8*
+// CHECK:       [[TMP48:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 2, i8* [[TMP47]])
+// CHECK:       store i8* [[TMP48]], i8** [[TD2]],
+// CHECK:       call void @__kmpc_end_taskgroup(%ident_t* {{[^,]+}}, i32 [[GTID]])
+// CHECK:       call void @__kmpc_end_taskgroup(%ident_t* {{[^,]+}}, i32 [[GTID]])
+
+// CHECK-DAG: define internal void [[AINIT]](i8*)
+// CHECK-DAG: store i32 0, i32* %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[ACOMB]](i8*, i8*)
+// CHECK-DAG: add nsw i32 %
+// CHECK-DAG: store i32 %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[BINIT]](i8*)
+// CHECK-DAG: store float 0.000000e+00, float* %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[BCOMB]](i8*, i8*)
+// CHECK-DAG: fadd float %
+// CHECK-DAG: store float %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[ARGCINIT]](i8*)
+// CHECK-DAG: store i32 0, i32* %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[ARGCCOMB]](i8*, i8*)
+// CHECK-DAG: add nsw i32 %
+// CHECK-DAG: store i32 %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[CINIT]](i8*)
+// CHECK-DAG: phi %struct.S* [
+// CHECK-DAG: call {{.+}}(%struct.S* {{.+}})
+// CHECK-DAG: br i1 %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[CFINI]](i8*)
+// CHECK-DAG: phi %struct.S* [
+// CHECK-DAG: call {{.+}}(%struct.S* {{.+}})
+// CHECK-DAG: br i1 %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[CCOMB]](i8*, i8*)
+// CHECK-DAG: phi %struct.S* [
+// CHECK-DAG: phi %struct.S* [
+// CHECK-DAG: call {{.+}}(%struct.S* {{.+}}, %struct.S* {{.+}}, %struct.S* {{.+}})
+// CHECK-DAG: call {{.+}}(%struct.S* {{.+}}, %struct.S* {{.+}})
+// CHECK-DAG: call {{.+}}(%struct.S* {{.+}})
+// CHECK-DAG: br i1 %
+// CHECK-DAG: ret void
+// CHECK_DAG: }
+
+// CHECK-DAG: define internal void [[VLAINIT]](i8*)
+// CHECK-DAG: call i32 @__kmpc_global_thread_num(%ident_t* {{[^,]+}})
+// CHECK-DAG: call i8* @__kmpc_threadprivate_cached(%ident_t*
+// CHECK-DAG: phi i16* [
+// CHECK-DAG: store i16 0, i16* %
+// CHECK-DAG: br i1 %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+
+// CHECK-DAG: define internal void [[VLACOMB]](i8*, i8*)
+// CHECK-DAG: call i32 @__kmpc_global_thread_num(%ident_t* {{[^,]+}})
+// CHECK-DAG: call i8* @__kmpc_threadprivate_cached(%ident_t*
+// CHECK-DAG: phi i16* [
+// CHECK-DAG: phi i16* [
+// CHECK-DAG: sext i16 %{{.+}} to i32
+// CHECK-DAG: add nsw i32 %
+// CHECK-DAG: trunc i32 %{{.+}} to i16
+// CHECK-DAG: store i16 %
+// CHECK_DAG: br i1 %
+// CHECK-DAG: ret void
+// CHECK-DAG: }
+#endif
diff --git a/test/OpenMP/taskgroup_task_reduction_messages.cpp b/test/OpenMP/taskgroup_task_reduction_messages.cpp
index 9c8177b0bd43..819dc5bdaf68 100644
--- a/test/OpenMP/taskgroup_task_reduction_messages.cpp
+++ b/test/OpenMP/taskgroup_task_reduction_messages.cpp
@@ -24,9 +24,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/taskloop_ast_print.cpp b/test/OpenMP/taskloop_ast_print.cpp
index dd0af85916cf..891b31dce4db 100644
--- a/test/OpenMP/taskloop_ast_print.cpp
+++ b/test/OpenMP/taskloop_ast_print.cpp
@@ -13,8 +13,10 @@ T tmain(T argc) {
   T b = argc, c, d, e, f, g;
   static T a;
 // CHECK: static T a;
-#pragma omp taskloop if(taskloop: argc > N) default(shared) untied priority(N) grainsize(N) reduction(+:g)
-  // CHECK-NEXT: #pragma omp taskloop if(taskloop: argc > N) default(shared) untied priority(N) grainsize(N) reduction(+: g)
+#pragma omp taskgroup task_reduction(+: d)
+#pragma omp taskloop if(taskloop: argc > N) default(shared) untied priority(N) grainsize(N) reduction(+:g) in_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskgroup task_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskloop if(taskloop: argc > N) default(shared) untied priority(N) grainsize(N) reduction(+: g) in_reduction(+: d)
   for (int i = 0; i < 2; ++i)
     a = 2;
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
@@ -53,8 +55,10 @@ int main(int argc, char **argv) {
   int b = argc, c, d, e, f, g;
   static int a;
 // CHECK: static int a;
-#pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) num_tasks(argc) reduction(*: g)
-  // CHECK-NEXT: #pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) num_tasks(argc) reduction(*: g)
+#pragma omp taskgroup task_reduction(+: d)
+#pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) num_tasks(argc) reduction(*: g) in_reduction(+:d)
+  // CHECK-NEXT: #pragma omp taskgroup task_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) num_tasks(argc) reduction(*: g) in_reduction(+: d)
   for (int i = 0; i < 2; ++i)
     a = 2;
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
diff --git a/test/OpenMP/taskloop_firstprivate_messages.cpp b/test/OpenMP/taskloop_firstprivate_messages.cpp
index fe22311f650c..e63d6a678b8d 100644
--- a/test/OpenMP/taskloop_firstprivate_messages.cpp
+++ b/test/OpenMP/taskloop_firstprivate_messages.cpp
@@ -295,9 +295,9 @@ int main(int argc, char **argv) {
 #pragma omp taskloop firstprivate(i) // expected-note {{defined as firstprivate}}
   for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp taskloop' directive may not be firstprivate, predetermined as private}}
     foo();
-#pragma omp parallel reduction(+ : i) // expected-note 4 {{defined as reduction}}
+#pragma omp parallel reduction(+ : i) // expected-note 2 {{defined as reduction}}
 #pragma omp taskloop firstprivate(i) //expected-error {{argument of a reduction clause of a parallel construct must not appear in a firstprivate clause on a task construct}}
-  for (i = 0; i < argc; ++i) // expected-error 3 {{reduction variables may not be accessed in an explicit task}}
+  for (i = 0; i < argc; ++i) // expected-error {{reduction variables may not be accessed in an explicit task}}
     foo();
 #pragma omp parallel
 #pragma omp taskloop firstprivate(B::x) // expected-error {{threadprivate or thread local variable cannot be firstprivate}}
diff --git a/test/OpenMP/taskloop_in_reduction_codegen.cpp b/test/OpenMP/taskloop_in_reduction_codegen.cpp
new file mode 100644
index 000000000000..e9ee9a32b97b
--- /dev/null
+++ b/test/OpenMP/taskloop_in_reduction_codegen.cpp
@@ -0,0 +1,82 @@
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c++ -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[PRIVATES:%.+]] = type { i8*, i8* }
+
+struct S {
+  int a;
+  S() : a(0) {}
+  S(const S&) {}
+  S& operator=(const S&) {return *this;}
+  ~S() {}
+  friend S operator+(const S&a, const S&b) {return a;}
+};
+
+
+int main(int argc, char **argv) {
+  int a;
+  float b;
+  S c[5];
+  short d[argc];
+#pragma omp taskgroup task_reduction(+: a, b, argc)
+  {
+#pragma omp taskgroup task_reduction(-:c, d)
+#pragma omp parallel
+#pragma omp taskloop in_reduction(+:a) in_reduction(-:d)
+    for (int i = 0; i < 5; ++i)
+      a += d[a];
+  }
+  return 0;
+}
+
+// CHECK-LABEL: @main
+// CHECK:       void @__kmpc_taskgroup(%ident_t* @0, i32 [[GTID:%.+]])
+// CHECK:       [[TD1:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 3, i8* %
+// CHECK-NEXT:  store i8* [[TD1]], i8** [[TD1_ADDR:%[^,]+]],
+// CHECK-NEXT:  call void @__kmpc_taskgroup(%ident_t* @0, i32 [[GTID]])
+// CHECK:       [[TD2:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 2, i8* %
+// CHECK-NEXT:  store i8* [[TD2]], i8** [[TD2_ADDR:%[^,]+]],
+// CHECK-NEXT:  call void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* @0, i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i16*, i8**, i8**)* [[OMP_PARALLEL:@.+]] to void (i32*, i32*, ...)*), i32* %{{.+}}, i64 %{{.+}}, i16* %{{.+}}, i8** [[TD1_ADDR]], i8** [[TD2_ADDR]])
+// CHECK-NEXT:  call void @__kmpc_end_taskgroup(%ident_t* @0, i32 [[GTID]])
+// CHECK-NEXT:  call void @__kmpc_end_taskgroup(%ident_t* @0, i32 [[GTID]])
+
+// CHECK:       define internal void [[OMP_PARALLEL]](
+// CHECK:       [[TASK_T:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* @0, i32 [[GTID:%.+]], i32 1, i64 96, i64 40, i32 (i32, i8*)* bitcast (i32 (i32, [[T:%.+]]*)* [[OMP_TASK:@.+]] to i32 (i32, i8*)*))
+// CHECK-NEXT:  [[TASK_T_WITH_PRIVS:%.+]] = bitcast i8* [[TASK_T]] to [[T]]*
+// CHECK:       [[PRIVS:%.+]] = getelementptr inbounds [[T]], [[T]]* [[TASK_T_WITH_PRIVS]], i32 0, i32 1
+// CHECK:       [[TD1_REF:%.+]] = getelementptr inbounds [[PRIVATES]], [[PRIVATES]]* [[PRIVS]], i32 0, i32 0
+// CHECK-NEXT:  [[TD1_SHAR:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[TD1_ADDR:%.+]] = load i8**, i8*** [[TD1_SHAR]],
+// CHECK-NEXT:  [[TD1:%.+]] = load i8*, i8** [[TD1_ADDR]],
+// CHECK-NEXT:  store i8* [[TD1]], i8** [[TD1_REF]],
+// CHECK-NEXT:  [[TD2_REF:%.+]] = getelementptr inbounds [[PRIVATES]], [[PRIVATES]]* [[PRIVS]], i32 0, i32 1
+// CHECK-NEXT:  [[TD2_SHAR:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[TD2_ADDR:%.+]] = load i8**, i8*** [[TD2_SHAR]],
+// CHECK-NEXT:  [[TD2:%.+]] = load i8*, i8** [[TD2_ADDR]],
+// CHECK-NEXT:  store i8* [[TD2]], i8** [[TD2_REF]],
+// CHECK:       call void @__kmpc_taskloop(%ident_t* @0, i32 [[GTID]], i8* [[TASK_T]], i32 1,
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal {{.*}} [[OMP_TASK]](
+// CHECK:       call void (i8*, ...) %{{[^(]+}}(i8* %{{.+}}, i8*** [[TD1_REF:%[^,]+]], i8*** [[TD2_REF:%[^,]+]])
+// CHECK-NEXT:  [[TD1_ADDR:%.+]] = load i8**, i8*** [[TD1_REF]],
+// CHECK-NEXT:  [[TD2_ADDR:%.+]] = load i8**, i8*** [[TD2_REF]],
+// CHECK-NEXT:  [[A_REF:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[A_ADDR:%.+]] = load i32*, i32** [[A_REF]],
+// CHECK-NEXT:  [[TD1:%.+]] = load i8*, i8** [[TD1_ADDR]],
+// CHECK-NEXT:  [[GTID:%.+]] = load i32, i32* %
+// CHECK-NEXT:  [[A_PTR:%.+]] = bitcast i32* [[A_ADDR]] to i8*
+// CHECK-NEXT:  call i8* @__kmpc_task_reduction_get_th_data(i32 [[GTID]], i8* [[TD1]], i8* [[A_PTR]])
+// CHECK:       [[D_REF:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[D_ADDR:%.+]] = load i16*, i16** [[D_REF]],
+// CHECK:       [[TD2:%.+]] = load i8*, i8** [[TD2_ADDR]],
+// CHECK-NEXT:  [[D_PTR:%.+]] = bitcast i16* [[D_ADDR]] to i8*
+// CHECK-NEXT:  call i8* @__kmpc_task_reduction_get_th_data(i32 [[GTID]], i8* [[TD2]], i8* [[D_PTR]])
+// CHECK:       add nsw i32
+// CHECK:       store i32 %
+#endif
diff --git a/test/OpenMP/taskloop_in_reduction_messages.cpp b/test/OpenMP/taskloop_in_reduction_messages.cpp
new file mode 100644
index 000000000000..409975565a91
--- /dev/null
+++ b/test/OpenMP/taskloop_in_reduction_messages.cpp
@@ -0,0 +1,376 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++98 -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++11 -ferror-limit 150 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+void foobar(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref)
+#pragma omp taskloop in_reduction(+:ref)
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+void foobar1(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref)
+#pragma omp taskloop in_reduction(-:ref)
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+#pragma omp declare reduction (red:int:omp_out += omp_in)
+
+void foobar2(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref) // expected-note {{previously marked as task_reduction with different reduction operation}}
+#pragma omp taskloop in_reduction(red:ref) // expected-error{{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+void foobar3(int &ref) {
+#pragma omp taskgroup task_reduction(red:ref) // expected-note {{previously marked as task_reduction with different reduction operation}}
+#pragma omp taskloop in_reduction(min:ref)  // expected-error{{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+void foobar4(int &ref) {
+#pragma omp taskloop in_reduction(min:ref) // expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+struct S1; // expected-note {{declared here}} expected-note 4 {{forward declaration of 'S1'}}
+extern S1 a;
+class S2 {
+  mutable int a;
+  S2 &operator+(const S2 &arg) { return (*this); } // expected-note 3 {{implicitly declared private here}}
+
+public:
+  S2() : a(0) {}
+  S2(S2 &s2) : a(s2.a) {}
+  static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
+};
+const float S2::S2sc = 0;
+S2 b;                     // expected-note 3 {{'b' defined here}}
+const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
+class S3 {
+  int a;
+
+public:
+  int b;
+  S3() : a(0) {}
+  S3(const S3 &s3) : a(s3.a) {}
+  S3 operator+(const S3 &arg1) { return arg1; }
+};
+int operator+(const S3 &arg1, const S3 &arg2) { return 5; }
+S3 c;               // expected-note 3 {{'c' defined here}}
+const S3 ca[5];     // expected-note 2 {{'ca' defined here}}
+extern const int f; // expected-note 4 {{'f' declared here}}
+class S4 {
+  int a;
+  S4(); // expected-note {{implicitly declared private here}}
+  S4(const S4 &s4);
+  S4 &operator+(const S4 &arg) { return (*this); }
+
+public:
+  S4(int v) : a(v) {}
+};
+S4 &operator&=(S4 &arg1, S4 &arg2) { return arg1; }
+class S5 {
+  int a;
+  S5() : a(0) {} // expected-note {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {}
+  S5 &operator+(const S5 &arg);
+
+public:
+  S5(int v) : a(v) {}
+};
+class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
+#if __cplusplus >= 201103L // C++11 or later
+// expected-note@-2 3 {{candidate function (the implicit move assignment operator) not viable}}
+#endif
+  int a;
+
+public:
+  S6() : a(6) {}
+  operator int() { return 6; }
+} o;
+
+S3 h, k;
+#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
+
+template <class T>       // expected-note {{declared here}}
+T tmain(T argc) {
+  const T d = T();       // expected-note 4 {{'d' defined here}}
+  const T da[5] = {T()}; // expected-note 2 {{'da' defined here}}
+  T qa[5] = {T()};
+  T i;
+  T &j = i;                    // expected-note 2 {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const T &r = da[(int)i];     // expected-note 2 {{'r' defined here}}
+  T &q = qa[(int)i];
+  T fl;
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction // expected-error {{expected '(' after 'in_reduction'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction + // expected-error {{expected '(' after 'in_reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop' are ignored}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop in_reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(&:argc) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+#pragma omp taskloop in_reduction(& : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(|:argc) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+#pragma omp taskloop in_reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(|| : argc ? i : argc) // expected-error 2 {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(foo : argc) //expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'float'}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'int'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(&&:argc)
+#pragma omp taskloop in_reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(^ : T) // expected-error {{'T' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:c)
+#pragma omp taskloop in_reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 3 {{const-qualified list item cannot be reduction}} expected-error 2 {{'operator+' is a private member of 'S2'}} expected-error 2 {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 4 {{arguments of OpenMP clause 'in_reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 3 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}} expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}} expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:k)
+#pragma omp taskloop in_reduction(+ : h, k) // expected-error {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(+ : o) // expected-error 2 {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:p)
+#pragma omp taskloop in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{variable can appear only once in OpenMP 'in_reduction' clause}} expected-note 2 {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(+ : r) // expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop in_reduction(max : j) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:fl)
+{
+#pragma omp taskloop in_reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskgroup task_reduction(*:fl) // expected-note 2 {{previously marked as task_reduction with different reduction operation}}
+{
+#pragma omp taskloop in_reduction(+ : fl) // expected-error 2 {{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+}
+}
+#pragma omp parallel
+#pragma omp for reduction(- : fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp taskloop in_reduction(+ : fl)
+  for (int j = 0; j < 10; ++j)
+    foo();
+
+  return T();
+}
+
+namespace A {
+double x;
+#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  const int d = 5;       // expected-note 2 {{'d' defined here}}
+  const int da[5] = {0}; // expected-note {{'da' defined here}}
+  int qa[5] = {0};
+  S4 e(4);
+  S5 g(5);
+  int i;
+  int &j = i;                  // expected-note {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const int &r = da[i];        // expected-note {{'r' defined here}}
+  int &q = qa[i];
+  float fl;
+#pragma omp taskloop in_reduction // expected-error {{expected '(' after 'in_reduction'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction + // expected-error {{expected '(' after 'in_reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop' are ignored}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(foo : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(|:argc)
+#pragma omp taskloop in_reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(|| : argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(~ : argc) // expected-error {{expected unqualified-id}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(&&:argc)
+#pragma omp taskloop in_reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(^ : S1) // expected-error {{'S1' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:c)
+#pragma omp taskloop in_reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{const-qualified list item cannot be reduction}} expected-error {{'operator+' is a private member of 'S2'}} expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{arguments of OpenMP clause 'in_reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(& : e, g) // expected-error {{calling a private constructor of class 'S4'}} expected-error {{nvalid operands to binary expression ('S4' and 'S4')}} expected-error {{calling a private constructor of class 'S5'}} expected-error {{invalid operands to binary expression ('S5' and 'S5')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:k)
+#pragma omp taskloop in_reduction(+ : h, k, B::x) // expected-error 2 {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(+ : o) // expected-error {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:p)
+#pragma omp taskloop in_reduction(+ : p), in_reduction(+ : p) // expected-error {{variable can appear only once in OpenMP 'in_reduction' clause}} expected-note {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop in_reduction(+ : r) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop in_reduction(max : j) // expected-error {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel
+#pragma omp for private(fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp taskloop in_reduction(+ : fl)
+  for (int j = 0; j < 10; ++j)
+    foo();
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp taskloop in_reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+  static int m;
+#pragma omp taskgroup task_reduction(+:m)
+#pragma omp taskloop in_reduction(+ : m) // OK
+  for (int i = 0; i < 10; ++i)
+  m++;
+
+  return tmain(argc) + tmain(fl); // expected-note {{in instantiation of function template specialization 'tmain<int>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<float>' requested here}}
+}
diff --git a/test/OpenMP/taskloop_lastprivate_messages.cpp b/test/OpenMP/taskloop_lastprivate_messages.cpp
index e4364448159c..1d238ea083e6 100644
--- a/test/OpenMP/taskloop_lastprivate_messages.cpp
+++ b/test/OpenMP/taskloop_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/taskloop_reduction_codegen.cpp b/test/OpenMP/taskloop_reduction_codegen.cpp
index af071ae8bd4e..6af067642fea 100644
--- a/test/OpenMP/taskloop_reduction_codegen.cpp
+++ b/test/OpenMP/taskloop_reduction_codegen.cpp
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -fopenmp -x c++ %s -verify -debug-info-kind=limited -emit-llvm -o - -triple powerpc64le-unknown-linux-gnu | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ %s -verify -debug-info-kind=limited -emit-llvm -o - -triple powerpc64le-unknown-linux-gnu -std=c++98 | FileCheck %s
 // expected-no-diagnostics
 
 struct S {
@@ -51,6 +51,7 @@ sum = 0.0;
 // CHECK:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]],
 // CHECK:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%ident_t*
 // CHECK:    [[DOTRD_INPUT_:%.*]] = alloca [4 x %struct.kmp_task_red_input_t],
+// CHECK:    alloca i32,
 // CHECK:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32,
 // CHECK:    [[DOTCAPTURE_EXPR_9:%.*]] = alloca i32,
 // CHECK:    store i32 0, i32* [[RETVAL]],
@@ -143,7 +144,7 @@ sum = 0.0;
 // CHECK:    [[DIV:%.*]] = sdiv i32 [[ADD11]], 1
 // CHECK:    [[SUB12:%.*]] = sub nsw i32 [[DIV]], 1
 // CHECK:    store i32 [[SUB12]], i32* [[DOTCAPTURE_EXPR_9]],
-// CHECK:    [[TMP65:%.*]] = call i8* @__kmpc_omp_task_alloc(%ident_t* %{{.+}}, i32 [[TMP0]], i32 1, i64 888, i64 72, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates*)* @{{.+}} to i32 (i32, i8*)*))
+// CHECK:    [[TMP65:%.*]] = call i8* @__kmpc_omp_task_alloc(%ident_t* %{{.+}}, i32 [[TMP0]], i32 1, i64 888, i64 72, i32 (i32, i8*)* bitcast (i32 (i32, %struct.kmp_task_t_with_privates*)* @[[TASK:.+]] to i32 (i32, i8*)*))
 // CHECK:    call void @__kmpc_taskloop(%ident_t* %{{.+}}, i32 [[TMP0]], i8* [[TMP65]], i32 1, i64* %{{.+}}, i64* %{{.+}}, i64 %{{.+}}, i32 0, i32 0, i64 0, i8* null)
 // CHECK:    call void @__kmpc_end_taskgroup(%ident_t*
 
@@ -195,3 +196,4 @@ sum = 0.0;
 // CHECK: store float %{{.+}}, float* %
 // CHECK: ret void
 
+// CHECK: distinct !DISubprogram(linkageName: "[[TASK]]", scope: !
diff --git a/test/OpenMP/taskloop_reduction_messages.cpp b/test/OpenMP/taskloop_reduction_messages.cpp
index 7295075c3cb8..23f70b568525 100644
--- a/test/OpenMP/taskloop_reduction_messages.cpp
+++ b/test/OpenMP/taskloop_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/taskloop_simd_ast_print.cpp b/test/OpenMP/taskloop_simd_ast_print.cpp
index 8678b9122e29..6f13c7e0baa5 100644
--- a/test/OpenMP/taskloop_simd_ast_print.cpp
+++ b/test/OpenMP/taskloop_simd_ast_print.cpp
@@ -14,8 +14,10 @@ T tmain(T argc) {
   T *ptr;
   static T a;
 // CHECK: static T a;
-#pragma omp taskloop simd if(taskloop: argc > N) default(shared) untied priority(N) safelen(N) linear(c) aligned(ptr) grainsize(N) reduction(+:g)
-  // CHECK-NEXT: #pragma omp taskloop simd if(taskloop: argc > N) default(shared) untied priority(N) safelen(N) linear(c) aligned(ptr) grainsize(N) reduction(+: g)
+#pragma omp taskgroup task_reduction(+: d)
+#pragma omp taskloop simd if(taskloop: argc > N) default(shared) untied priority(N) safelen(N) linear(c) aligned(ptr) grainsize(N) reduction(+:g) in_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskgroup task_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskloop simd if(taskloop: argc > N) default(shared) untied priority(N) safelen(N) linear(c) aligned(ptr) grainsize(N) reduction(+: g) in_reduction(+: d)
   for (int i = 0; i < 2; ++i)
     a = 2;
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
@@ -54,8 +56,10 @@ int main(int argc, char **argv) {
   int b = argc, c, d, e, f, g;
   static int a;
 // CHECK: static int a;
-#pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b), aligned(argv) num_tasks(argc) reduction(*: g)
-  // CHECK-NEXT: #pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b) aligned(argv) num_tasks(argc) reduction(*: g)
+#pragma omp taskgroup task_reduction(+: d)
+#pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b), aligned(argv) num_tasks(argc) reduction(*: g) in_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskgroup task_reduction(+: d)
+  // CHECK-NEXT: #pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b) aligned(argv) num_tasks(argc) reduction(*: g) in_reduction(+: d)
   for (int i = 0; i < 2; ++i)
     a = 2;
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
diff --git a/test/OpenMP/taskloop_simd_codegen.cpp b/test/OpenMP/taskloop_simd_codegen.cpp
index f787689688e3..006fd59af3fa 100644
--- a/test/OpenMP/taskloop_simd_codegen.cpp
+++ b/test/OpenMP/taskloop_simd_codegen.cpp
@@ -158,7 +158,7 @@ struct S {
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: [[NUM_TASKS:%.+]] = zext i32 %{{.+}} to i64
 // CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 2, i64 [[NUM_TASKS]], i8* null)
-#pragma omp taskloop simd shared(c) num_tasks(a) simdlen(64) safelen(8)
+#pragma omp taskloop simd shared(c) num_tasks(a) simdlen(8) safelen(64)
     for (a = 0; a < c; ++a)
       ;
   }
@@ -201,6 +201,6 @@ struct S {
 // CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
 // CHECK: !{!"llvm.loop.vectorize.width", i32 4}
 // CHECK: !{!"llvm.loop.vectorize.width", i32 32}
-// CHECK: !{!"llvm.loop.vectorize.width", i32 64}
+// CHECK: !{!"llvm.loop.vectorize.width", i32 8}
 
 #endif
diff --git a/test/OpenMP/taskloop_simd_firstprivate_messages.cpp b/test/OpenMP/taskloop_simd_firstprivate_messages.cpp
index 18cefc1beeb2..176dcd7f8f9c 100644
--- a/test/OpenMP/taskloop_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/taskloop_simd_firstprivate_messages.cpp
@@ -295,9 +295,9 @@ int main(int argc, char **argv) {
 #pragma omp taskloop simd firstprivate(i) // expected-note {{defined as firstprivate}}
   for (i = 0; i < argc; ++i) // expected-error {{loop iteration variable in the associated loop of 'omp taskloop simd' directive may not be firstprivate, predetermined as linear}}
     foo();
-#pragma omp parallel reduction(+ : i) // expected-note 4 {{defined as reduction}}
+#pragma omp parallel reduction(+ : i) // expected-note 2 {{defined as reduction}}
 #pragma omp taskloop simd firstprivate(i) // expected-error {{argument of a reduction clause of a parallel construct must not appear in a firstprivate clause on a task construct}}
-  for (i = 0; i < argc; ++i) // expected-error 3 {{reduction variables may not be accessed in an explicit task}}
+  for (i = 0; i < argc; ++i) // expected-error {{reduction variables may not be accessed in an explicit task}}
     foo();
 #pragma omp parallel
 #pragma omp taskloop simd firstprivate(B::x) // expected-error {{threadprivate or thread local variable cannot be firstprivate}}
diff --git a/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp b/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp
new file mode 100644
index 000000000000..d894943cc359
--- /dev/null
+++ b/test/OpenMP/taskloop_simd_in_reduction_codegen.cpp
@@ -0,0 +1,82 @@
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -x c++ -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// CHECK: [[PRIVATES:%.+]] = type { i8*, i8* }
+
+struct S {
+  int a;
+  S() : a(0) {}
+  S(const S&) {}
+  S& operator=(const S&) {return *this;}
+  ~S() {}
+  friend S operator+(const S&a, const S&b) {return a;}
+};
+
+
+int main(int argc, char **argv) {
+  int a;
+  float b;
+  S c[5];
+  short d[argc];
+#pragma omp taskgroup task_reduction(+: a, b, argc)
+  {
+#pragma omp taskgroup task_reduction(-:c, d)
+#pragma omp parallel
+#pragma omp taskloop simd in_reduction(+:a) in_reduction(-:d)
+    for (int i = 0; i < 5; ++i)
+      a += d[a];
+  }
+  return 0;
+}
+
+// CHECK-LABEL: @main
+// CHECK:       void @__kmpc_taskgroup(%ident_t* @0, i32 [[GTID:%.+]])
+// CHECK:       [[TD1:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 3, i8* %
+// CHECK-NEXT:  store i8* [[TD1]], i8** [[TD1_ADDR:%[^,]+]],
+// CHECK-NEXT:  call void @__kmpc_taskgroup(%ident_t* @0, i32 [[GTID]])
+// CHECK:       [[TD2:%.+]] = call i8* @__kmpc_task_reduction_init(i32 [[GTID]], i32 2, i8* %
+// CHECK-NEXT:  store i8* [[TD2]], i8** [[TD2_ADDR:%[^,]+]],
+// CHECK-NEXT:  call void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%ident_t* @0, i32 5, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, i64, i16*, i8**, i8**)* [[OMP_PARALLEL:@.+]] to void (i32*, i32*, ...)*), i32* %{{.+}}, i64 %{{.+}}, i16* %{{.+}}, i8** [[TD1_ADDR]], i8** [[TD2_ADDR]])
+// CHECK-NEXT:  call void @__kmpc_end_taskgroup(%ident_t* @0, i32 [[GTID]])
+// CHECK-NEXT:  call void @__kmpc_end_taskgroup(%ident_t* @0, i32 [[GTID]])
+
+// CHECK:       define internal void [[OMP_PARALLEL]](
+// CHECK:       [[TASK_T:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* @0, i32 [[GTID:%.+]], i32 1, i64 96, i64 40, i32 (i32, i8*)* bitcast (i32 (i32, [[T:%.+]]*)* [[OMP_TASK:@.+]] to i32 (i32, i8*)*))
+// CHECK-NEXT:  [[TASK_T_WITH_PRIVS:%.+]] = bitcast i8* [[TASK_T]] to [[T]]*
+// CHECK:       [[PRIVS:%.+]] = getelementptr inbounds [[T]], [[T]]* [[TASK_T_WITH_PRIVS]], i32 0, i32 1
+// CHECK:       [[TD1_REF:%.+]] = getelementptr inbounds [[PRIVATES]], [[PRIVATES]]* [[PRIVS]], i32 0, i32 0
+// CHECK-NEXT:  [[TD1_SHAR:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[TD1_ADDR:%.+]] = load i8**, i8*** [[TD1_SHAR]],
+// CHECK-NEXT:  [[TD1:%.+]] = load i8*, i8** [[TD1_ADDR]],
+// CHECK-NEXT:  store i8* [[TD1]], i8** [[TD1_REF]],
+// CHECK-NEXT:  [[TD2_REF:%.+]] = getelementptr inbounds [[PRIVATES]], [[PRIVATES]]* [[PRIVS]], i32 0, i32 1
+// CHECK-NEXT:  [[TD2_SHAR:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[TD2_ADDR:%.+]] = load i8**, i8*** [[TD2_SHAR]],
+// CHECK-NEXT:  [[TD2:%.+]] = load i8*, i8** [[TD2_ADDR]],
+// CHECK-NEXT:  store i8* [[TD2]], i8** [[TD2_REF]],
+// CHECK:       call void @__kmpc_taskloop(%ident_t* @0, i32 [[GTID]], i8* [[TASK_T]], i32 1,
+// CHECK:       ret void
+// CHECK-NEXT:  }
+
+// CHECK:       define internal {{.*}} [[OMP_TASK]](
+// CHECK:       call void (i8*, ...) %{{[^(]+}}(i8* %{{.+}}, i8*** [[TD1_REF:%[^,]+]], i8*** [[TD2_REF:%[^,]+]])
+// CHECK-NEXT:  [[TD1_ADDR:%.+]] = load i8**, i8*** [[TD1_REF]],
+// CHECK-NEXT:  [[TD2_ADDR:%.+]] = load i8**, i8*** [[TD2_REF]],
+// CHECK-NEXT:  [[A_REF:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[A_ADDR:%.+]] = load i32*, i32** [[A_REF]],
+// CHECK-NEXT:  [[TD1:%.+]] = load i8*, i8** [[TD1_ADDR]],
+// CHECK-NEXT:  [[GTID:%.+]] = load i32, i32* %
+// CHECK-NEXT:  [[A_PTR:%.+]] = bitcast i32* [[A_ADDR]] to i8*
+// CHECK-NEXT:  call i8* @__kmpc_task_reduction_get_th_data(i32 [[GTID]], i8* [[TD1]], i8* [[A_PTR]])
+// CHECK:       [[D_REF:%.+]] = getelementptr inbounds %
+// CHECK-NEXT:  [[D_ADDR:%.+]] = load i16*, i16** [[D_REF]],
+// CHECK:       [[TD2:%.+]] = load i8*, i8** [[TD2_ADDR]],
+// CHECK-NEXT:  [[D_PTR:%.+]] = bitcast i16* [[D_ADDR]] to i8*
+// CHECK-NEXT:  call i8* @__kmpc_task_reduction_get_th_data(i32 [[GTID]], i8* [[TD2]], i8* [[D_PTR]])
+// CHECK:       add nsw i32
+// CHECK:       store i32 %
+#endif
diff --git a/test/OpenMP/taskloop_simd_in_reduction_messages.cpp b/test/OpenMP/taskloop_simd_in_reduction_messages.cpp
new file mode 100644
index 000000000000..f3b1a855e320
--- /dev/null
+++ b/test/OpenMP/taskloop_simd_in_reduction_messages.cpp
@@ -0,0 +1,376 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++98 -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++11 -ferror-limit 150 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+void foobar(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref)
+#pragma omp taskloop simd in_reduction(+:ref)
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+void foobar1(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref)
+#pragma omp taskloop simd in_reduction(-:ref)
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+#pragma omp declare reduction (red:int:omp_out += omp_in)
+
+void foobar2(int &ref) {
+#pragma omp taskgroup task_reduction(+:ref) // expected-note {{previously marked as task_reduction with different reduction operation}}
+#pragma omp taskloop simd in_reduction(red:ref) // expected-error{{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+void foobar3(int &ref) {
+#pragma omp taskgroup task_reduction(red:ref) // expected-note {{previously marked as task_reduction with different reduction operation}}
+#pragma omp taskloop simd in_reduction(min:ref)  // expected-error{{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+void foobar4(int &ref) {
+#pragma omp taskloop simd in_reduction(min:ref) // expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+}
+
+struct S1; // expected-note {{declared here}} expected-note 4 {{forward declaration of 'S1'}}
+extern S1 a;
+class S2 {
+  mutable int a;
+  S2 &operator+(const S2 &arg) { return (*this); } // expected-note 3 {{implicitly declared private here}}
+
+public:
+  S2() : a(0) {}
+  S2(S2 &s2) : a(s2.a) {}
+  static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
+};
+const float S2::S2sc = 0;
+S2 b;                     // expected-note 3 {{'b' defined here}}
+const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
+class S3 {
+  int a;
+
+public:
+  int b;
+  S3() : a(0) {}
+  S3(const S3 &s3) : a(s3.a) {}
+  S3 operator+(const S3 &arg1) { return arg1; }
+};
+int operator+(const S3 &arg1, const S3 &arg2) { return 5; }
+S3 c;               // expected-note 3 {{'c' defined here}}
+const S3 ca[5];     // expected-note 2 {{'ca' defined here}}
+extern const int f; // expected-note 4 {{'f' declared here}}
+class S4 {
+  int a;
+  S4(); // expected-note {{implicitly declared private here}}
+  S4(const S4 &s4);
+  S4 &operator+(const S4 &arg) { return (*this); }
+
+public:
+  S4(int v) : a(v) {}
+};
+S4 &operator&=(S4 &arg1, S4 &arg2) { return arg1; }
+class S5 {
+  int a;
+  S5() : a(0) {} // expected-note {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {}
+  S5 &operator+(const S5 &arg);
+
+public:
+  S5(int v) : a(v) {}
+};
+class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
+#if __cplusplus >= 201103L // C++11 or later
+// expected-note@-2 3 {{candidate function (the implicit move assignment operator) not viable}}
+#endif
+  int a;
+
+public:
+  S6() : a(6) {}
+  operator int() { return 6; }
+} o;
+
+S3 h, k;
+#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
+
+template <class T>       // expected-note {{declared here}}
+T tmain(T argc) {
+  const T d = T();       // expected-note 4 {{'d' defined here}}
+  const T da[5] = {T()}; // expected-note 2 {{'da' defined here}}
+  T qa[5] = {T()};
+  T i;
+  T &j = i;                    // expected-note 2 {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const T &r = da[(int)i];     // expected-note 2 {{'r' defined here}}
+  T &q = qa[(int)i];
+  T fl;
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction // expected-error {{expected '(' after 'in_reduction'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction + // expected-error {{expected '(' after 'in_reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop simd' are ignored}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:argc)
+#pragma omp taskloop simd in_reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(&:argc) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+#pragma omp taskloop simd in_reduction(& : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(|:argc) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+#pragma omp taskloop simd in_reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(|| : argc ? i : argc) // expected-error 2 {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(foo : argc) //expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'float'}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'int'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(&&:argc)
+#pragma omp taskloop simd in_reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(^ : T) // expected-error {{'T' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:c)
+#pragma omp taskloop simd in_reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 3 {{const-qualified list item cannot be reduction}} expected-error 2 {{'operator+' is a private member of 'S2'}} expected-error 2 {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 4 {{arguments of OpenMP clause 'in_reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 3 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}} expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}} expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:k)
+#pragma omp taskloop simd in_reduction(+ : h, k) // expected-error {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(+ : o) // expected-error 2 {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop simd in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:p)
+#pragma omp taskloop simd in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{variable can appear only once in OpenMP 'in_reduction' clause}} expected-note 2 {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(+ : r) // expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop simd in_reduction(max : j) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:fl)
+{
+#pragma omp taskloop simd in_reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskgroup task_reduction(*:fl) // expected-note 2 {{previously marked as task_reduction with different reduction operation}}
+{
+#pragma omp taskloop simd in_reduction(+ : fl) // expected-error 2 {{in_reduction variable must have the same reduction operation as in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+}
+}
+#pragma omp parallel
+#pragma omp for reduction(- : fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp taskloop simd in_reduction(+ : fl)
+  for (int j = 0; j < 10; ++j)
+    foo();
+
+  return T();
+}
+
+namespace A {
+double x;
+#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  const int d = 5;       // expected-note 2 {{'d' defined here}}
+  const int da[5] = {0}; // expected-note {{'da' defined here}}
+  int qa[5] = {0};
+  S4 e(4);
+  S5 g(5);
+  int i;
+  int &j = i;                  // expected-note {{'j' defined here}}
+  S3 &p = k;                   // expected-note 2 {{'p' defined here}}
+  const int &r = da[i];        // expected-note {{'r' defined here}}
+  int &q = qa[i];
+  float fl;
+#pragma omp taskloop simd in_reduction // expected-error {{expected '(' after 'in_reduction'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction + // expected-error {{expected '(' after 'in_reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop simd' are ignored}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(foo : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max'}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(|:argc)
+#pragma omp taskloop simd in_reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(|| : argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(~ : argc) // expected-error {{expected unqualified-id}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(&&:argc)
+#pragma omp taskloop simd in_reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(^ : S1) // expected-error {{'S1' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:c)
+#pragma omp taskloop simd in_reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{const-qualified list item cannot be reduction}} expected-error {{'operator+' is a private member of 'S2'}} expected-error {{in_reduction variable must appear in a task_reduction clause}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{arguments of OpenMP clause 'in_reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(& : e, g) // expected-error {{calling a private constructor of class 'S4'}} expected-error {{nvalid operands to binary expression ('S4' and 'S4')}} expected-error {{calling a private constructor of class 'S5'}} expected-error {{invalid operands to binary expression ('S5' and 'S5')}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:k)
+#pragma omp taskloop simd in_reduction(+ : h, k, B::x) // expected-error 2 {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(+ : o) // expected-error {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop simd in_reduction(+ : p), in_reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskgroup task_reduction(+:p)
+#pragma omp taskloop simd in_reduction(+ : p), in_reduction(+ : p) // expected-error {{variable can appear only once in OpenMP 'in_reduction' clause}} expected-note {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp taskloop simd in_reduction(+ : r) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop simd in_reduction(max : j) // expected-error {{argument of OpenMP clause 'in_reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+  foo();
+#pragma omp parallel
+#pragma omp for private(fl)
+  for (int i = 0; i < 10; ++i)
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp taskloop simd in_reduction(+ : fl)
+  for (int j = 0; j < 10; ++j)
+    foo();
+#pragma omp taskgroup task_reduction(+:fl)
+#pragma omp taskloop simd in_reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+  static int m;
+#pragma omp taskgroup task_reduction(+:m)
+#pragma omp taskloop simd in_reduction(+ : m) // OK
+  for (int i = 0; i < 10; ++i)
+  m++;
+
+  return tmain(argc) + tmain(fl); // expected-note {{in instantiation of function template specialization 'tmain<int>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<float>' requested here}}
+}
diff --git a/test/OpenMP/taskloop_simd_lastprivate_messages.cpp b/test/OpenMP/taskloop_simd_lastprivate_messages.cpp
index ed1bdf5a6e3e..e5cba6c64b6b 100644
--- a/test/OpenMP/taskloop_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/taskloop_simd_lastprivate_messages.cpp
@@ -18,9 +18,9 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
diff --git a/test/OpenMP/taskloop_simd_misc_messages.c b/test/OpenMP/taskloop_simd_misc_messages.c
index 61c7b107d496..35cc62cad585 100644
--- a/test/OpenMP/taskloop_simd_misc_messages.c
+++ b/test/OpenMP/taskloop_simd_misc_messages.c
@@ -346,6 +346,10 @@ void test_firstprivate() {
 #pragma omp taskloop simd lastprivate(x, y, z) firstprivate(x, y, z)
   for (i = 0; i < 16; ++i)
     ;
+// expected-error@+1 {{the value of 'simdlen' parameter must be less than or equal to the value of the 'safelen' parameter}}
+#pragma omp taskloop simd simdlen(64) safelen(8)
+  for (i = 0; i < 16; ++i)
+    ;
 }
 
 void test_loop_messages() {
diff --git a/test/OpenMP/taskloop_simd_reduction_codegen.cpp b/test/OpenMP/taskloop_simd_reduction_codegen.cpp
index b96c8d50a3ff..7e8d54a2a1da 100644
--- a/test/OpenMP/taskloop_simd_reduction_codegen.cpp
+++ b/test/OpenMP/taskloop_simd_reduction_codegen.cpp
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -fopenmp -x c++ %s -verify -debug-info-kind=limited -emit-llvm -o - -triple powerpc64le-unknown-linux-gnu | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ %s -verify -debug-info-kind=limited -emit-llvm -o - -triple powerpc64le-unknown-linux-gnu -std=c++98 | FileCheck %s
 // expected-no-diagnostics
 
 struct S {
@@ -51,6 +51,7 @@ sum = 0.0;
 // CHECK:    [[AGG_CAPTURED:%.*]] = alloca [[STRUCT_ANON:%.*]],
 // CHECK:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%ident_t*
 // CHECK:    [[DOTRD_INPUT_:%.*]] = alloca [4 x %struct.kmp_task_red_input_t],
+// CHECK:    alloca i32,
 // CHECK:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32,
 // CHECK:    [[DOTCAPTURE_EXPR_9:%.*]] = alloca i32,
 // CHECK:    store i32 0, i32* [[RETVAL]],
diff --git a/test/OpenMP/taskloop_simd_reduction_messages.cpp b/test/OpenMP/taskloop_simd_reduction_messages.cpp
index f9739f708a28..886b685f6325 100644
--- a/test/OpenMP/taskloop_simd_reduction_messages.cpp
+++ b/test/OpenMP/taskloop_simd_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/teams_codegen.cpp b/test/OpenMP/teams_codegen.cpp
index 8aaa206abc3b..28f6ca0714d4 100644
--- a/test/OpenMP/teams_codegen.cpp
+++ b/test/OpenMP/teams_codegen.cpp
@@ -21,7 +21,7 @@ int teams_argument_global_local(int a){
   int la = 23;
   float lc = 25.0;
 
-  // CK1: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
   // CK1: call void @{{.+}}(i{{64|32}} %{{.+}})
   #pragma omp target
   #pragma omp teams
@@ -29,7 +29,7 @@ int teams_argument_global_local(int a){
     ++comp;
   }
 
-  // CK1: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
   // CK1: call void @{{.+}}(i{{64|32}} %{{.+}})
   #pragma omp target
   {{{
@@ -39,7 +39,7 @@ int teams_argument_global_local(int a){
     }
   }}}
 
-  // CK1-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 0)
+  // CK1-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 0)
   // CK1-DAG: [[NT]] = load i32, i32* [[NTA:%[^,]+]],
 
   // CK1: call void @{{.+}}(i{{64|32}} %{{.+}})
@@ -49,7 +49,7 @@ int teams_argument_global_local(int a){
     ++comp;
   }
 
-  // CK1-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 [[NT:%[^,]+]])
+  // CK1-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 [[NT:%[^,]+]])
   // CK1-DAG: [[NT]] = load i32, i32* [[NTA:%[^,]+]],
 
   // CK1: call void @{{.+}}(i{{64|32}} %{{.+}})
@@ -59,7 +59,7 @@ int teams_argument_global_local(int a){
     ++comp;
   }
 
-  // CK1-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
+  // CK1-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
 
   // CK1-DAG: [[NT]] = add nsw i32 [[NTA:%[^,]+]], [[NTB:%[^,]+]]
   // CK1-DAG: [[NTA]] = load i32, i32* @Gbla,
@@ -78,7 +78,7 @@ int teams_argument_global_local(int a){
     ++comp;
   }
 
-  // CK1-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 {{.+}}, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
+  // CK1-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 {{.+}}, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
 
   // CK1-DAG: [[NT]] = add nsw i32 [[NTA:%[^,]+]], 1
   // CK1-DAG: [[NTA]] = load i32, i32* @Gbla,
@@ -125,7 +125,7 @@ int teams_template_arg(void) {
   SS<int> la;
   SS<long long> lb;
 
-  // CK2-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
+  // CK2-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
 
   // CK2-DAG: [[NT]] = load i32, i32* getelementptr inbounds ([[SSI]], [[SSI]]* @Gbla, i32 0, i32 0)
 
@@ -141,7 +141,7 @@ int teams_template_arg(void) {
     ++comp;
   }
 
-  // CK2-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
+  // CK2-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 [[TL:%[^,]+]])
 
   // CK2-DAG: [[TL]] = trunc i64 [[TLD:%[^,]+]] to i32
   // CK2-DAG: [[TLD]] = load i64, i64* getelementptr inbounds ([[SSL]], [[SSL]]* @Gblb, i32 0, i32 0),
@@ -181,7 +181,7 @@ struct SS{
   int foo(void) {
     int comp = 1;
 
-    // CK3-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 123)
+    // CK3-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 [[NT:%[^,]+]], i32 123)
 
     // CK3-DAG: [[NT]] = load i32, i32* [[NTA:%[^,]+]],
     // CK3-DAG: [[NTA]] = getelementptr inbounds [[SSI]], [[SSI]]* [[NTB:%[^,]+]], i32 0, i32 0
@@ -194,7 +194,7 @@ struct SS{
       ++comp;
     }
 
-    // CK3-DAG: call i32 @__tgt_target_teams(i32 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 456, i32 [[TL:%[^,]+]])
+    // CK3-DAG: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 456, i32 [[TL:%[^,]+]])
 
     // CK3-DAG: [[TL]] = add nsw i32 [[TLA:%[^,]+]], 123
     // CK3-DAG: [[TLA]] = fptosi float [[TLB:%[^,]+]] to i32
diff --git a/test/OpenMP/teams_distribute_codegen.cpp b/test/OpenMP/teams_distribute_codegen.cpp
new file mode 100644
index 000000000000..8a2afbd4341d
--- /dev/null
+++ b/test/OpenMP/teams_distribute_codegen.cpp
@@ -0,0 +1,241 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+int a[100];
+
+// CK1: define {{.*}}i32 @{{.+}}teams_argument_globali(
+int teams_argument_global(int n){  
+  int te = n / 128;
+  int th = 128;
+  // discard n_addr
+  // CK1: alloca i32,
+  // CK1: [[TE:%.+]] = alloca i32,
+  // CK1: [[TH:%.+]] = alloca i32,
+  // CK1: [[TE_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TE_PAR:%.+]] = load{{.+}}, {{.+}} [[TE_CAST]],
+  // CK1: [[TH_PAR:%.+]] = load{{.+}}, {{.+}} [[TH_CAST]],
+
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 {{.+}}, i32 {{.+}})
+
+  // CK1: call void @[[OFFL1:.+]](i{{32|64}} [[TE_PAR]], i{{32|64}} [[TH_PAR]],
+  #pragma omp target
+  #pragma omp teams distribute num_teams(te), thread_limit(th)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK1: call void @[[OFFL2:.+]](i{{64|32}} %{{.+}})
+  #pragma omp target
+  {{{
+  #pragma omp teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  }}}
+
+  // outlined target regions
+  // CK1: define internal void @[[OFFL1]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], i{{32|64}} {{.+}}, {{.+}})
+  // CK1: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+  // CK1: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+  // CK1-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+  // CK1-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+  // CK1-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+  // CK1-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+  // CK1-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+  // CK1-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+  // CK1: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]]({{.+}}, {{.+}})
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL2:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  return a[0];
+}
+
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+// CK2: define {{.*}}i32 @{{.+}}teams_local_argv(
+int teams_local_arg(void) {
+  int n = 100;
+  int a[n];
+
+  // CK2: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK2: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+  #pragma omp target
+  #pragma omp teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  // outlined target region
+  // CK2: define internal void @[[OFFL1]]({{.+}}, {{.+}})
+  // CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK2: ret void
+
+  // CK2: define internal void @[[OUTL1]]({{.+}})
+  // CK2: call void @__kmpc_for_static_init_4(
+  // CK2: call void @__kmpc_for_static_fini(
+  // CK2: ret void  
+
+  return a[0];
+}
+#endif // CK2
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+#ifdef CK3
+
+// CK3: [[SSI:%.+]] = type { [{{.+}} x i32], float }
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK3: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK3: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK3: call void @[[OFFL1:.+]]([[SSI]]* %{{.+}})
+    #pragma omp target
+    #pragma omp teams distribute
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+      // outlined target region
+  // CK3: define internal void @[[OFFL1]]([[SSI]]* {{.+}})
+  // CK3: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK3: ret void
+
+  // CK3: define internal void @[[OUTL1]]({{.+}})
+  // CK3: call void @__kmpc_for_static_init_4(
+  // CK3: call void @__kmpc_for_static_fini(
+  // CK3: ret void  
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK3
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+
+#ifdef CK4
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int te = n/128;
+  int th = 128;
+#pragma omp target
+#pragma omp teams distribute num_teams(te) thread_limit(th)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target
+#pragma omp teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK4:  define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CK4: call void @[[OFFL1:.+]]({{.+}})
+// CK4: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK4:  ret
+
+// CK4:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTL1]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+// CK4:  define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 {{.+}}, i32 {{.+}})
+// CK4: call void @[[OFFLT:.+]]({{.+}})
+// CK4:  ret
+// CK4-NEXT: }
+
+// CK4: define {{.*}}void @[[OFFLT]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], {{.+}})
+// CK4: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+// CK4: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+// CK4-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+// CK4-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+// CK4-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+// CK4-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+// CK4-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+// CK4-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+// CK4: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTLT]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+#endif // CK4
+#endif
diff --git a/test/OpenMP/teams_distribute_collapse_codegen.cpp b/test/OpenMP/teams_distribute_collapse_codegen.cpp
new file mode 100644
index 000000000000..a65c0a2e63f7
--- /dev/null
+++ b/test/OpenMP/teams_distribute_collapse_codegen.cpp
@@ -0,0 +1,128 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X][Y];
+
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+    
+    // CK1: call i32 @__tgt_target_teams(
+    // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute collapse(2)
+    for(int i = 0; i < X; i++) {
+      for(int j = 0; j < Y; j++) {
+	a[i][j] = (T)0;
+      }
+    }
+    // CK1: define internal void @[[OFFL1]](
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+    // CK1: ret void
+    
+    // CK1: define internal void @[[OUTL1]]({{.+}})
+    // discard loop variables not needed here
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: [[OMP_UB:%.+]] = alloca i32,
+    // CK1: store i32 56087, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void  
+
+    return a[0][0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n, int m>
+int tmain(T argc) {
+  T a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = (T)0;
+    }
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int m = 2;
+  int a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = 0;
+    }
+  }
+  return tmain<int, 10, 2>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i64,
+// CK2: store i64 {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_8({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// discard loop variables not needed here
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i32,
+// CK2: store i32 {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_dist_schedule_codegen.cpp b/test/OpenMP/teams_distribute_dist_schedule_codegen.cpp
new file mode 100644
index 000000000000..8bd8825e00c4
--- /dev/null
+++ b/test/OpenMP/teams_distribute_dist_schedule_codegen.cpp
@@ -0,0 +1,206 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target
+    #pragma omp teams distribute dist_schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target
+    #pragma omp teams distribute dist_schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void  
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void  
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void  
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+#pragma omp target
+#pragma omp teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target
+#pragma omp teams distribute
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_firstprivate_codegen.cpp b/test/OpenMP/teams_distribute_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..a126b1d16b77
--- /dev/null
+++ b/test/OpenMP/teams_distribute_firstprivate_codegen.cpp
@@ -0,0 +1,336 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+// CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute firstprivate(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}})
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute firstprivate(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}})
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]],
+    // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]],
+    // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // skip loop vars
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_CONV]],
+    // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP]]
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_CONV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute firstprivate(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]],
+// CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]],
+// CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]],
+// CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[TOFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]({{.+}})
+// CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]],
+// CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]],
+// CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]],
+// CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/teams_distribute_firstprivate_messages.cpp b/test/OpenMP/teams_distribute_firstprivate_messages.cpp
index a71080774331..1a2b5f3782f9 100644
--- a/test/OpenMP/teams_distribute_firstprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_firstprivate_messages.cpp
@@ -144,8 +144,9 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute lastprivate(argc), firstprivate(argc) // OK
+#pragma omp teams distribute lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
   return 0;
diff --git a/test/OpenMP/teams_distribute_lastprivate_codegen.cpp b/test/OpenMP/teams_distribute_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..2eab60b9dd3c
--- /dev/null
+++ b/test/OpenMP/teams_distribute_lastprivate_codegen.cpp
@@ -0,0 +1,369 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams distribute lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams distribute lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams distribute lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/teams_distribute_lastprivate_messages.cpp b/test/OpenMP/teams_distribute_lastprivate_messages.cpp
index b892141d5014..392607ea14ba 100644
--- a/test/OpenMP/teams_distribute_lastprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -255,12 +255,14 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{firstprivate variable cannot be lastprivate}} expected-note@+2 {{defined as firstprivate}}
 #pragma omp target
-#pragma omp teams distribute firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp teams distribute firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute lastprivate(n) firstprivate(n) // OK
+#pragma omp teams distribute lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/teams_distribute_loop_messages.cpp b/test/OpenMP/teams_distribute_loop_messages.cpp
index ea2604eebf0f..acab1f76a2d4 100644
--- a/test/OpenMP/teams_distribute_loop_messages.cpp
+++ b/test/OpenMP/teams_distribute_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -689,12 +689,16 @@ void test_loop_eh() {
       void g() { throw 0; }
     };
   }
+  #pragma omp target
+  #pragma omp teams distribute
+  f; // expected-error {{use of undeclared identifier 'f'}}
 }
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute lastprivate(s) firstprivate(s)
+#pragma omp teams distribute lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/teams_distribute_parallel_for_ast_print.cpp b/test/OpenMP/teams_distribute_parallel_for_ast_print.cpp
index ad14794116ed..68bd424ade01 100644
--- a/test/OpenMP/teams_distribute_parallel_for_ast_print.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_ast_print.cpp
@@ -8,6 +8,9 @@
 
 void foo() {}
 
+int x;
+#pragma omp threadprivate(x)
+
 struct S {
   S(): a(0) {}
   S(int v) : a(v) {}
@@ -40,7 +43,7 @@ public:
   void foo() {
     int b, argv, d, c, e, f;
 #pragma omp target
-#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argv) shared(d) reduction(+:c) reduction(max:e) num_teams(f) thread_limit(d)
+#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argv) shared(d) reduction(+:c) reduction(max:e) num_teams(f) thread_limit(d) copyin(x)
     for (int k = 0; k < a.a; ++k)
       ++a.a;
   }
@@ -50,7 +53,7 @@ public:
 // CHECK: #pragma omp target
 // CHECK-NEXT: #pragma omp teams distribute parallel for private(this->a) private(this->a)
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d) copyin(x)
 
 class S8 : public S7<S> {
   S8() {}
@@ -74,7 +77,7 @@ public:
   void bar() {
     int b, argv, d, c, e, f8;
 #pragma omp target
-#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argv) shared(d) reduction(+:c) reduction(max:e) num_teams(f8) thread_limit(d)
+#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argv) shared(d) reduction(+:c) reduction(max:e) num_teams(f8) thread_limit(d) copyin(x)
     for (int k = 0; k < a.a; ++k)
       ++a.a;
   }
@@ -86,7 +89,7 @@ public:
 // CHECK: #pragma omp target
 // CHECK-NEXT: #pragma omp teams distribute parallel for private(this->a) private(this->a)
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f8) thread_limit(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f8) thread_limit(d) copyin(x)
 
 template <class T, int N>
 T tmain(T argc) {
@@ -127,19 +130,19 @@ T tmain(T argc) {
 // CHECK-NEXT: for (int i = 0; i < 10; ++i)
 // CHECK-NEXT: foo();  
 #pragma omp target
-#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argc) shared(d) reduction(+:c) reduction(max:e) num_teams(f) thread_limit(d)
+#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argc) shared(d) reduction(+:c) reduction(max:e) num_teams(f) thread_limit(d) copyin(x)
     for (int k = 0; k < 10; ++k)
       e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d) copyin(x)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
-#pragma omp teams distribute parallel for linear(d)
+#pragma omp teams distribute parallel for
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for linear(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
   return T();
@@ -182,19 +185,19 @@ int main (int argc, char **argv) {
 // CHECK-NEXT: for (int i = 0; i < 10; ++i)
 // CHECK-NEXT: foo();
 #pragma omp target
-#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argc) shared(d) reduction(+:c) reduction(max:e) num_teams(f) thread_limit(d)
+#pragma omp teams distribute parallel for default(none), private(b) firstprivate(argc) shared(d) reduction(+:c) reduction(max:e) num_teams(f) thread_limit(d) copyin(x)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for default(none) private(b) firstprivate(argc) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d) copyin(x)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
-#pragma omp teams distribute parallel for linear(d)
+#pragma omp teams distribute parallel for
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for linear(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
   return (0);
diff --git a/test/OpenMP/teams_distribute_parallel_for_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_codegen.cpp
new file mode 100644
index 000000000000..5711ede87c86
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_codegen.cpp
@@ -0,0 +1,247 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+int a[100];
+
+// CK1: define {{.*}}i32 @{{.+}}teams_argument_globali(
+int teams_argument_global(int n){
+  int te = n / 128;
+  int th = 128;
+  // discard n_addr
+  // CK1: alloca i32,
+  // CK1: [[TE:%.+]] = alloca i32,
+  // CK1: [[TH:%.+]] = alloca i32,
+  // CK1: [[TE_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TE_PAR:%.+]] = load{{.+}}, {{.+}} [[TE_CAST]],
+  // CK1: [[TH_PAR:%.+]] = load{{.+}}, {{.+}} [[TH_CAST]],
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 {{.+}}, i32 {{.+}})
+
+  // CK1: call void @[[OFFL1:.+]](i{{32|64}} [[TE_PAR]], i{{32|64}} [[TH_PAR]],
+  #pragma omp target
+  #pragma omp teams distribute parallel for num_teams(te), thread_limit(th)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+    #pragma omp cancel for
+  }
+
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK1: call void @[[OFFL2:.+]](i{{64|32}} %{{.+}})
+  #pragma omp target
+  {{{
+  #pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  }}}
+
+  // outlined target regions
+  // CK1: define internal void @[[OFFL1]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], i{{32|64}} {{.+}}, {{.+}})
+  // CK1: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+  // CK1: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+  // CK1-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+  // CK1-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+  // CK1-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+  // CK1-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+  // CK1-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+  // CK1-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+  // CK1: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.+}} @__kmpc_fork_call(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]]({{.+}}, {{.+}})
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL2:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.+}} @__kmpc_fork_call(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  return a[0];
+}
+
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+// CK2: define {{.*}}i32 @{{.+}}teams_local_argv(
+int teams_local_arg(void) {
+  int n = 100;
+  int a[n];
+
+  // CK2: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK2: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+  #pragma omp target
+  #pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  // outlined target region
+  // CK2: define internal void @[[OFFL1]]({{.+}}, {{.+}})
+  // CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK2: ret void
+
+  // CK2: define internal void @[[OUTL1]]({{.+}})
+  // CK2: call void @__kmpc_for_static_init_4(
+  // CK2: call void {{.+}} @__kmpc_fork_call(
+  // CK2: call void @__kmpc_for_static_fini(
+  // CK2: ret void
+
+  return a[0];
+}
+#endif // CK2
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+#ifdef CK3
+
+// CK3: [[SSI:%.+]] = type { [{{.+}} x i32], float }
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK3: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK3: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK3: call void @[[OFFL1:.+]]([[SSI]]* %{{.+}})
+    #pragma omp target
+    #pragma omp teams distribute parallel for
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+      // outlined target region
+  // CK3: define internal void @[[OFFL1]]([[SSI]]* {{.+}})
+  // CK3: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK3: ret void
+
+  // CK3: define internal void @[[OUTL1]]({{.+}})
+  // CK3: call void @__kmpc_for_static_init_4(
+  // CK3: call void {{.+}} @__kmpc_fork_call(
+  // CK3: call void @__kmpc_for_static_fini(
+  // CK3: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK3
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+
+#ifdef CK4
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int te = n/128;
+  int th = 128;
+#pragma omp target
+#pragma omp teams distribute parallel for num_teams(te) thread_limit(th)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target
+#pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK4:  define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CK4: call void @[[OFFL1:.+]]({{.+}})
+// CK4: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK4:  ret
+
+// CK4:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTL1]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void {{.+}} @__kmpc_fork_call(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+// CK4:  define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 {{.+}}, i32 {{.+}})
+// CK4: call void @[[OFFLT:.+]]({{.+}})
+// CK4:  ret
+// CK4-NEXT: }
+
+// CK4: define {{.*}}void @[[OFFLT]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], {{.+}})
+// CK4: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+// CK4: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+// CK4-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+// CK4-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+// CK4-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+// CK4-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+// CK4-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+// CK4-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+// CK4: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTLT]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void {{.+}} @__kmpc_fork_call(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+#endif // CK4
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_collapse_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_collapse_codegen.cpp
new file mode 100644
index 000000000000..7522288f75e2
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_collapse_codegen.cpp
@@ -0,0 +1,144 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X][Y];
+
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+    // CK1: call i32 @__tgt_target_teams(
+    // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for collapse(2)
+    for(int i = 0; i < X; i++) {
+      for(int j = 0; j < Y; j++) {
+        a[i][j] = (T)0;
+      }
+    }
+    // CK1: define internal void @[[OFFL1]](
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+    // CK1: ret void
+
+    // CK1: define internal void @[[OUTL1]]({{.+}})
+    // discard loop variables not needed here
+    // CK1: [[OMP_UB:%.omp.comb.ub]] = alloca i32,
+    // CK1: store i32 56087, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+    // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void
+
+    // CK1: define internal void @[[PAR_OUTL1]]({{.+}})
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34, {{.+}}, {{.+}},
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void
+
+    return a[0][0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n, int m>
+int tmain(T argc) {
+  T a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute parallel for collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = (T)0;
+    }
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int m = 2;
+  int a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute parallel for collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = 0;
+    }
+  }
+  return tmain<int, 10, 2>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: [[OMP_UB:%.omp.comb.ub]] = alloca i64,
+// CK2: store i64 {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_8({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i64* [[OMP_UB]],
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_{{[4|8]}}({{.+}}, {{.+}}, i32 34, {{.+}}, {{.+}},
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// discard loop variables not needed here
+// CK2: [[OMP_UB:%.omp.comb.ub]] = alloca i32,
+// CK2: store i32 {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[TPAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34, {{.+}}, {{.+}},
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_parallel_for_copyin_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_copyin_codegen.cpp
new file mode 100644
index 000000000000..0202a40002b5
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_copyin_codegen.cpp
@@ -0,0 +1,199 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+int x;
+#pragma omp threadprivate(x)
+
+template <typename T>
+T tmain() {
+  int a[2];
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(x)
+  for (int i = 0; i < 2; ++i) {
+    a[i] = x;
+  }
+  return T();
+}
+
+int main() {
+  int a[2];
+#ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(x)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]](
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    a[i] = x;
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca [2 x i{{[0-9]+}}]*,
+    // LAMBDA: [[X_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+
+    // LAMBDA:  [[X_REF:%.+]] = load {{.+}}, {{.+}} [[X_ADDR]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: [[X_VAL:%.+]] = load {{.+}}, {{.+}} [[X_REF]],
+    // LAMBDA: store {{.+}} [[X_VAL]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      a[i] = x;
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(x)
+  for (int i = 0; i < 2; ++i) {
+    a[i] = x;
+  }
+  return tmain<int>();
+  //return 0;
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[PAR_OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{%.+}} = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[X_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+
+// CHECK:  [[X_REF:%.+]] = load {{.+}}, {{.+}} [[X_ADDR]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK:  [[X_VAL:%.+]] = load {{.+}}, {{.+}} [[X_REF]],
+// CHECK: store {{.+}} [[X_VAL]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(
+// CHECK: call void @[[TOFFL1:.+]](
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]](
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL1:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[TPAR_OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// prev lb and ub
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+
+// CHECK: {{%.+}} = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[X_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// iter variables
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+
+// CHECK:  [[X_REF:%.+]] = load {{.+}}, {{.+}} [[X_ADDR]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK:  [[X_VAL:%.+]] = load {{.+}}, {{.+}} [[X_REF]],
+// CHECK: store {{.+}} [[X_VAL]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_copyin_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_copyin_messages.cpp
new file mode 100644
index 000000000000..30f3e75bf635
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_copyin_messages.cpp
@@ -0,0 +1,113 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 100 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+struct S1; // expected-note {{declared here}}
+class S2 {
+  mutable int a;
+
+public:
+  S2() : a(0) {}
+  S2 &operator=(S2 &s2) { return *this; }
+};
+class S3 {
+  int a;
+
+public:
+  S3() : a(0) {}
+  S3 &operator=(S3 &s3) { return *this; }
+};
+class S4 {
+  int a;
+  S4();
+  S4 &operator=(const S4 &s4); // expected-note {{implicitly declared private here}}
+
+public:
+  S4(int v) : a(v) {}
+};
+class S5 {
+  int a;
+  S5() : a(0) {}
+  S5 &operator=(const S5 &s5) { return *this; } // expected-note {{implicitly declared private here}}
+
+public:
+  S5(int v) : a(v) {}
+};
+template <class T>
+class ST {
+public:
+  static T s;
+};
+
+S2 k;
+S3 h;
+S4 l(3);
+S5 m(4);
+#pragma omp threadprivate(h, k, l, m)
+
+namespace A {
+double x;
+#pragma omp threadprivate(x)
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  int i;
+#pragma omp target
+#pragma omp teams distribute parallel for copyin // expected-error {{expected '(' after 'copyin'}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin( // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin() // expected-error {{expected expression}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(k // expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(h, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(l) // expected-error {{'operator=' is a private member of 'S4'}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(S1) // expected-error {{'S1' does not refer to a value}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(argv[1]) // expected-error {{expected variable name}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(i) // expected-error {{copyin variable must be threadprivate}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(m) // expected-error {{'operator=' is a private member of 'S5'}}
+  for (i = 0; i < argc; ++i)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for copyin(ST<int>::s, B::x) // expected-error {{copyin variable must be threadprivate}}
+  for (i = 0; i < argc; ++i)
+    foo();
+
+  return 0;
+}
diff --git a/test/OpenMP/teams_distribute_parallel_for_dist_schedule_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_dist_schedule_codegen.cpp
new file mode 100644
index 000000000000..dc48f166b4c0
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_dist_schedule_codegen.cpp
@@ -0,0 +1,263 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for dist_schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for dist_schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for dist_schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for dist_schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..fa99f9cfcf1b
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_firstprivate_codegen.cpp
@@ -0,0 +1,496 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+// CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute parallel for firstprivate(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}})
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for firstprivate(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}})
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]],
+    // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]],
+    // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // skip loop vars
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}})
+    // Skip global and bound tid vars, and prev lb and ub vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // skip loop vars
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP_PRIV:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]],
+
+    // use of private vars
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP_PRIV]]
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for firstprivate(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]],
+// CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]],
+// CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]],
+// CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[PAR_OUTL]]({{.+}})
+// Skip global and bound tid vars, and prev lb ub vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[TOFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]({{.+}})
+// CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]],
+// CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]],
+// CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]],
+// CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[TPAR_OUTL]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_firstprivate_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_firstprivate_messages.cpp
index e6314c210e8f..90e78defc5e2 100644
--- a/test/OpenMP/teams_distribute_parallel_for_firstprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_firstprivate_messages.cpp
@@ -144,8 +144,9 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for lastprivate(argc), firstprivate(argc) // OK
+#pragma omp teams distribute parallel for lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
   return 0;
diff --git a/test/OpenMP/teams_distribute_parallel_for_if_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_if_codegen.cpp
new file mode 100644
index 000000000000..bc40dca83fbe
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_if_codegen.cpp
@@ -0,0 +1,184 @@
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple %itanium_abi_triple -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+void fn1();
+void fn2();
+void fn3();
+void fn4();
+void fn5();
+void fn6();
+
+int Arg;
+
+// CHECK-LABEL: define {{.*}}void @{{.+}}gtid_test
+void gtid_test() {
+#pragma omp target
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+#pragma omp teams distribute parallel for
+  for(int i = 0 ; i < 100; i++) {}
+  // CHECK: define internal void [[OFFLOADING_FUN_0]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+  // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK:  call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_0:@.+]] to void
+  // CHECK: call void @__kmpc_for_static_fini(
+
+  // CHECK: define{{.+}} void [[OMP_OUTLINED_0]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: ret
+#pragma omp target
+#pragma omp teams distribute parallel for if (parallel: false)
+  for(int i = 0 ; i < 100; i++) {
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+  // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @__kmpc_serialized_parallel(
+  // CHECK: call void [[OMP_OUTLINED_1:@.+]](
+  // CHECK: call void @__kmpc_end_serialized_parallel(
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: define{{.+}} void [[OMP_OUTLINED_1]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @{{.+}}gtid_test
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: ret
+    gtid_test();
+  }
+}
+
+
+template <typename T>
+int tmain(T Arg) {
+#pragma omp target
+#pragma omp teams distribute parallel for if (true)
+  for(int i = 0 ; i < 100; i++) {
+    fn1();
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for if (false)
+  for(int i = 0 ; i < 100; i++) {
+    fn2();
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for if (parallel: Arg)
+  for(int i = 0 ; i < 100; i++) {
+    fn3();
+  }
+  return 0;
+}
+
+// CHECK-LABEL: define {{.*}}i{{[0-9]+}} @main()
+int main() {
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_2:@.+]](
+// CHECK: = call {{.*}}i{{.+}} @{{.+}}tmain
+#pragma omp target
+#pragma omp teams distribute parallel for if (true)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_0]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_2:@.+]] to void
+    // CHECK: call void @__kmpc_for_static_fini(
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_2]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn4
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    fn4();
+  }
+
+#pragma omp target
+#pragma omp teams distribute parallel for if (false)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_1]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void @__kmpc_serialized_parallel(
+    // CHECK: call void [[OMP_OUTLINED_3:@.+]](
+    // CHECK: call void @__kmpc_end_serialized_parallel(
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_3]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn5
+    // CHECK: call void @__kmpc_for_static_fini(
+    fn5();
+  }
+
+#pragma omp target
+#pragma omp teams distribute parallel for if (Arg)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_2]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}}* [[OMP_TEAMS_OUTLINED_2:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_2]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_4:@.+]] to void
+    // CHECK: call void @__kmpc_serialized_parallel(
+    // CHECK: call void [[OMP_OUTLINED_4:@.+]](
+    // CHECK: call void @__kmpc_end_serialized_parallel(
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_4]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn6
+    // CHECK: call void @__kmpc_for_static_fini(
+    fn6();
+  }
+
+  return tmain(Arg);
+}
+
+// CHECK-LABEL: define {{.+}} @{{.+}}tmain
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, void {{.+}}* [[T_OUTLINE_FUN_1:@.+]] to void
+// CHECK: call void @__kmpc_for_static_fini(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_1]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn1
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @__kmpc_serialized_parallel(
+// CHECK: call void [[T_OUTLINE_FUN_2:@.+]](
+// CHECK: call {{.*}}void @__kmpc_end_serialized_parallel(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_2]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn2
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, void {{.+}}* [[T_OUTLINE_FUN_3:@.+]] to void
+// CHECK: call {{.*}}void @__kmpc_serialized_parallel(
+// call void [[T_OUTLINE_FUN_3:@.+]](
+// CHECK: call {{.*}}void @__kmpc_end_serialized_parallel(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_3]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn3
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..501abb725b31
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_lastprivate_codegen.cpp
@@ -0,0 +1,584 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams distribute parallel for lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams distribute parallel for lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+
+      // LAMBDA: define{{.*}} internal{{.*}} void @[[LPAR_OUTL]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams distribute parallel for lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// gbl and bound tid vars, prev lb and ub vars
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void [[TPAR_OUTL:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid, and prev lb and ub vars
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_lastprivate_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_lastprivate_messages.cpp
index 5691af71f7cd..b58bac4fdfc3 100644
--- a/test/OpenMP/teams_distribute_parallel_for_lastprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -255,12 +255,14 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{firstprivate variable cannot be lastprivate}} expected-note@+2 {{defined as firstprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp teams distribute parallel for firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for lastprivate(n) firstprivate(n) // OK
+#pragma omp teams distribute parallel for lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/teams_distribute_parallel_for_linear_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_linear_messages.cpp
deleted file mode 100644
index b14fb2328b2b..000000000000
--- a/test/OpenMP/teams_distribute_parallel_for_linear_messages.cpp
+++ /dev/null
@@ -1,293 +0,0 @@
-// RUN: %clang_cc1 -verify -fopenmp %s
-
-namespace X {
-  int x;
-};
-
-struct B {
-  static int ib; // expected-note {{'B::ib' declared here}}
-  static int bfoo() { return 8; }
-};
-
-int bfoo() { return 4; }
-
-int z;
-const int C1 = 1;
-const int C2 = 2;
-void test_linear_colons()
-{
-  int B = 0;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B:bfoo())
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B:B::bfoo())
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(X::x : ::z)
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B,::z, X::x)
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(::z)
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B::bfoo()) // expected-error {{expected variable name}}
-  for (int i = 0; i < 10; ++i) ;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(B::ib,B:C1+C2)
-  for (int i = 0; i < 10; ++i) ;
-}
-
-template<int L, class T, class N> T test_template(T* arr, N num) {
-  N i;
-  T sum = (T)0;
-  T ind2 = - num * L; // expected-note {{'ind2' defined here}}
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(ind2:L) // expected-error {{argument of a linear clause should be of integral or pointer type}}
-  for (i = 0; i < num; ++i) {
-    T cur = arr[(int)ind2];
-    ind2 += L;
-    sum += cur;
-  }
-  return T();
-}
-
-template<int LEN> int test_warn() {
-  int ind2 = 0;
-  #pragma omp target
-  #pragma omp teams distribute parallel for linear(ind2:LEN) // expected-warning {{zero linear step (ind2 should probably be const)}}
-  for (int i = 0; i < 100; i++) {
-    ind2 += LEN;
-  }
-  return ind2;
-}
-
-struct S1; // expected-note 2 {{declared here}} expected-note 2 {{forward declaration of 'S1'}}
-extern S1 a;
-class S2 {
-  mutable int a;
-public:
-  S2():a(0) { }
-};
-const S2 b; // expected-note 2 {{'b' defined here}}
-const S2 ba[5];
-class S3 {
-  int a;
-public:
-  S3():a(0) { }
-};
-const S3 ca[5];
-class S4 {
-  int a;
-  S4();
-public:
-  S4(int v):a(v) { }
-};
-class S5 {
-  int a;
-  S5():a(0) {}
-public:
-  S5(int v):a(v) { }
-};
-
-S3 h;
-#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
-
-template<class I, class C> int foomain(I argc, C **argv) {
-  I e(4);
-  I g(5);
-  int i;
-  int &j = i;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear // expected-error {{expected '(' after 'linear'}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear ( // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear () // expected-error {{expected expression}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc : 5)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (S1) // expected-error {{'S1' does not refer to a value}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (a, b:B::ib) // expected-error {{linear variable with incomplete type 'S1'}} expected-error {{const-qualified variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argv[1]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(e, g)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target
-    #pragma omp teams distribute parallel for linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-  return 0;
-}
-
-namespace A {
-double x;
-#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
-}
-namespace C {
-using A::x;
-}
-
-int main(int argc, char **argv) {
-  double darr[100];
-  // expected-note@+1 {{in instantiation of function template specialization 'test_template<-4, double, int>' requested here}}
-  test_template<-4>(darr, 4);
-  // expected-note@+1 {{in instantiation of function template specialization 'test_warn<0>' requested here}}
-  test_warn<0>();
-
-  S4 e(4); // expected-note {{'e' defined here}}
-  S5 g(5); // expected-note {{'g' defined here}}
-  int i;
-  int &j = i;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear // expected-error {{expected '(' after 'linear'}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear ( // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear () // expected-error {{expected expression}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argc)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (S1) // expected-error {{'S1' does not refer to a value}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (a, b) // expected-error {{linear variable with incomplete type 'S1'}} expected-error {{const-qualified variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear (argv[1]) // expected-error {{expected variable name}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(e, g) // expected-error {{argument of a linear clause should be of integral or pointer type, not 'S4'}} expected-error {{argument of a linear clause should be of integral or pointer type, not 'S5'}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(h, C::x) // expected-error 2 {{threadprivate or thread local variable cannot be linear}}
-  for (int k = 0; k < argc; ++k) ++k;
-
-  #pragma omp parallel
-  {
-    int i;
-    #pragma omp target
-    #pragma omp teams distribute parallel for linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
-
-    #pragma omp target
-    #pragma omp teams distribute parallel for linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
-  }
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
-  return 0;
-}
-
diff --git a/test/OpenMP/teams_distribute_parallel_for_loop_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_loop_messages.cpp
index e0c315f513b7..a2f6f679d325 100644
--- a/test/OpenMP/teams_distribute_parallel_for_loop_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -691,8 +691,9 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for lastprivate(s) firstprivate(s)
+#pragma omp teams distribute parallel for lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/teams_distribute_parallel_for_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_messages.cpp
index c3536cd86167..9843b442e5b4 100644
--- a/test/OpenMP/teams_distribute_parallel_for_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_messages.cpp
@@ -9,6 +9,9 @@ static int pvt;
 #pragma omp teams distribute parallel for // expected-error {{unexpected OpenMP directive '#pragma omp teams distribute parallel for'}}
 
 int main(int argc, char **argv) {
+  #pragma omp target
+  #pragma omp teams distribute parallel for
+  f; // expected-error {{use of undeclared identifier 'f'}}
 #pragma omp target
 #pragma omp teams distribute parallel for { // expected-warning {{extra tokens at the end of '#pragma omp teams distribute parallel for' are ignored}}
   for (int i = 0; i < argc; ++i)
@@ -34,7 +37,7 @@ int main(int argc, char **argv) {
   for (int i = 0; i < argc; ++i)
     foo();
 #pragma omp target
-#pragma omp teams distribute parallel for
+#pragma omp teams distribute parallel for linear(argc) // expected-error {{unexpected OpenMP clause 'linear' in directive '#pragma omp teams distribute parallel for'}}
   for (int i = 0; i < argc; ++i)
     foo();
 // expected-warning@+2 {{extra tokens at the end of '#pragma omp teams distribute parallel for' are ignored}}
@@ -94,7 +97,7 @@ L1:
   }
 
 #pragma omp target
-#pragma omp teams distribute parallel for copyin(pvt) // expected-error {{unexpected OpenMP clause 'copyin' in directive '#pragma omp teams distribute parallel for'}}
+#pragma omp teams distribute parallel for copyin(pvt)
   for (int n = 0; n < 100; ++n) {}
 
   return 0;
@@ -107,3 +110,12 @@ void test_ordered() {
     ;
 }
 
+void test_cancel() {
+#pragma omp target
+#pragma omp teams distribute parallel for
+  for (int i = 0; i < 16; ++i) {
+#pragma omp cancel for
+    ;
+  }
+}
+
diff --git a/test/OpenMP/teams_distribute_parallel_for_num_threads_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_num_threads_codegen.cpp
new file mode 100644
index 000000000000..7c0ac461220d
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_num_threads_codegen.cpp
@@ -0,0 +1,120 @@
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple %itanium_abi_triple -fexceptions -fcxx-exceptions -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -fexceptions -fcxx-exceptions -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+typedef __INTPTR_TYPE__ intptr_t;
+
+// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[S_TY:%.+]] = type { [[INTPTR_T_TY:i[0-9]+]], [[INTPTR_T_TY]], [[INTPTR_T_TY]] }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_2:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+void foo();
+
+struct S {
+  intptr_t a, b, c;
+  S(intptr_t a) : a(a) {}
+  operator char() { return a; }
+  ~S() {}
+};
+
+template <typename T, int C>
+int tmain() {
+#pragma omp target
+#pragma omp teams distribute parallel for num_threads(C)
+  for (int i = 0; i < 100; i++)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for num_threads(T(23))
+  for (int i = 0; i < 100; i++)
+    foo();
+  return 0;
+}
+
+int main() {
+  S s(0);
+  char a = s;
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+// CHECK: invoke{{.+}} [[TMAIN_5:@.+]]()
+// CHECK: invoke{{.+}} [[TMAIN_1:@.+]]()
+#pragma omp target
+  // CHECK: define internal void [[OFFLOADING_FUN_0]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+#pragma omp teams distribute parallel for num_threads(2)
+  for (int i = 0; i < 100; i++) {
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+    // CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 2)
+    // CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+    foo();
+  }
+#pragma omp target
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+#pragma omp teams distribute parallel for num_threads(a)
+  for (int i = 0; i < 100; i++) {
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+    // CHECK-DAG: [[A_ADDR:%.+]] = alloca i8*,
+    // CHECK-DAG: [[A_REF:%.+]] = load i8*, i8** [[A_ADDR]],
+    // CHECK-DAG: [[A_VAL:%.+]] = load i8, i8* [[A_REF]],
+    // CHECK-DAG: [[A_EXT:%.+]] = sext i8 [[A_VAL]] to {{.+}}
+    // CHECK: call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 [[A_EXT]])
+    // CHECK: call {{.*}}void {{.*}} @__kmpc_fork_call(
+    foo();
+  }
+  return a + tmain<char, 5>() + tmain<S, 1>();
+}
+
+// tmain 5
+// CHECK-DAG: define {{.*}}i{{[0-9]+}} [[TMAIN_5]]()
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_1:@.+]](
+
+// tmain 1
+// CHECK-DAG: define {{.*}}i{{[0-9]+}} [[TMAIN_1]]()
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_2:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_3:@.+]](
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_0]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_0]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 5)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_1]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 23)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_2]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_2:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_2]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 1)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_3]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}}* [[T_OMP_TEAMS_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_3]]({{.+}}, {{.+}}, {{.+}} [[NUM_TH_CPT_IN:%.+]])
+// CHECK: [[NUM_TH_CPT:%.+]] = alloca i8*,
+// CHECK: store {{.+}} [[NUM_TH_CPT_IN]], {{.+}} [[NUM_TH_CPT]],
+// CHECK: [[NUM_TH_REF:%.+]] = load{{.+}}, {{.+}} [[NUM_TH_CPT]],
+// CHECK-DAG:   [[NUM_TH_VAL:%.+]] = load {{.+}}, {{.+}} [[NUM_TH_REF]],
+// CHECK-DAG:   [[NUM_TH_SEXT:%.+]] = sext i8 [[NUM_TH_VAL]] to {{.+}}
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 [[NUM_TH_SEXT]])
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp
new file mode 100644
index 000000000000..80a12738cab4
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_private_codegen.cpp
@@ -0,0 +1,332 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute parallel for private(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for private(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}})
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for private(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]()
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[PAR_OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[SIVAR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0,
+// CHECK: call void @[[TOFFL1:.+]]()
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL1:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[TPAR_OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// prev lb and ub
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// iter variables
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_proc_bind_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_proc_bind_codegen.cpp
new file mode 100644
index 000000000000..71e92bd5a555
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_proc_bind_codegen.cpp
@@ -0,0 +1,90 @@
+// add -fopenmp-targets
+
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+typedef __INTPTR_TYPE__ intptr_t;
+
+// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_2:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+void foo();
+
+struct S {
+  intptr_t a, b, c;
+  S(intptr_t a) : a(a) {}
+  operator char() { return a; }
+  ~S() {}
+};
+
+template <typename T>
+T tmain() {
+#pragma omp target
+#pragma omp teams distribute parallel for proc_bind(master)
+  for(int i = 0; i < 1000; i++) {}
+  return T();
+}
+
+int main() {
+  // CHECK-LABEL: @main
+#pragma omp target
+#pragma omp teams distribute parallel for proc_bind(spread)
+  for(int i = 0; i < 1000; i++) {}
+#pragma omp target
+#pragma omp teams distribute parallel for proc_bind(close)
+  for(int i = 0; i < 1000; i++) {}
+  return tmain<int>();
+}
+
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL1:@.+]]()
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL2:@.+]]()
+// CHECK: [[CALL_RET:%.+]] = call{{.+}} i32 [[TMAIN:@.+]]()
+// CHECK: ret i32 [[CALL_RET]]
+
+// CHECK: define{{.+}} void [[OFFL1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_1]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 4)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: define{{.+}} [[OFFL2]]()
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_1]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 3)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: define{{.+}} [[TMAIN]]()
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL3:@.+]]()
+
+// CHECK: define{{.+}} [[OFFL3]]()
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_3]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 2)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_reduction_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_reduction_codegen.cpp
new file mode 100644
index 000000000000..55ccfff99a3c
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_reduction_codegen.cpp
@@ -0,0 +1,345 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target
+#pragma omp teams distribute parallel for reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: br
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+
+    // Skip global and bound tid vars, and prev lb and ub vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // skip loop vars
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+     // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: br
+
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK-64: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: define internal void @[[PAR_OUTL]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars, and prev lb and ub
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// skip loop vars
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i{{64|32}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK-64: [[TVAR_CONV:%.+]] = bitcast{{.+}} [[TVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: define internal void @[[TPAR_OUTL]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars, and prev lb and ub vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// skip loop vars
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: br
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_reduction_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_reduction_messages.cpp
index 73661618d20c..889efd76cc3c 100644
--- a/test/OpenMP/teams_distribute_parallel_for_reduction_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_reduction_messages.cpp
@@ -19,9 +19,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/teams_distribute_parallel_for_schedule_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_schedule_codegen.cpp
new file mode 100644
index 000000000000..201a170ccd4a
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_schedule_codegen.cpp
@@ -0,0 +1,398 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL4:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for schedule(dynamic)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL5:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for schedule(dynamic, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 33,
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL4]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL4:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL4]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL4:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL4]]({{.+}})
+  // CK1: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+  // CK1: call {{.+}} @__kmpc_dispatch_next_4(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL5]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL5:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL5]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL5:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL5]]({{.+}})
+  // CK1: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+  // CK1: call {{.+}} @__kmpc_dispatch_next_4(
+  // CK1: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for schedule(dynamic)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for schedule(dynamic, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for dist_schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for schedule(dynamic)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for schedule(dynamic, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL4:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL5:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL4]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTL4:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL4]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL4:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL4]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+
+// CK2: define {{.*}}void @[[OFFL5]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTL5:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL5]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL5:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL5]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT4:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT5:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 33,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT4]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT4:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT4]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT4:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT4]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT5]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT5:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT5]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT5:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT5]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_parallel_for_shared_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_shared_messages.cpp
index 8edcd3664341..1a1af09aa205 100644
--- a/test/OpenMP/teams_distribute_parallel_for_shared_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_shared_messages.cpp
@@ -84,7 +84,7 @@ int main(int argc, char **argv) {
   #pragma omp teams distribute parallel for shared (S1) // expected-error {{'S1' does not refer to a value}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
-  #pragma omp teams distribute parallel for shared (a, b, c, d, f)
+  #pragma omp teams distribute parallel for shared (a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
   #pragma omp teams distribute parallel for shared (argv[1]) // expected-error {{expected variable name}}
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_ast_print.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_ast_print.cpp
index 26cc15822943..7573fdeafa31 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_ast_print.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_ast_print.cpp
@@ -29,9 +29,10 @@ public:
       ++this->a.a;
   }
   S7 &operator=(S7 &s) {
+    int k;
 #pragma omp target
-#pragma omp teams distribute parallel for simd private(a) private(this->a)
-    for (int k = 0; k < s.a.a; ++k)
+#pragma omp teams distribute parallel for simd private(a) private(this->a) linear(k)
+    for (k = 0; k < s.a.a; ++k)
       ++s.a.a;
 
     foo();
@@ -59,7 +60,7 @@ public:
 // CHECK: #pragma omp target
 // CHECK-NEXT: #pragma omp teams distribute parallel for simd private(this->a) private(this->a) private(T::a)
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for simd private(this->a) private(this->a)
+// CHECK-NEXT: #pragma omp teams distribute parallel for simd private(this->a) private(this->a) linear(k)
 // CHECK: #pragma omp target
 // CHECK-NEXT: #pragma omp teams distribute parallel for simd default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK: #pragma omp target
@@ -161,11 +162,11 @@ T tmain(T argc) {
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
-#pragma omp teams distribute parallel for simd simdlen(clen-1) linear(d)
+#pragma omp teams distribute parallel for simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for simd simdlen(clen - 1) linear(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
@@ -224,11 +225,11 @@ int main (int argc, char **argv) {
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
-#pragma omp teams distribute parallel for simd simdlen(clen-1) linear(d)
+#pragma omp teams distribute parallel for simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute parallel for simd simdlen(clen - 1) linear(d)
+// CHECK-NEXT: #pragma omp teams distribute parallel for simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_codegen.cpp
new file mode 100644
index 000000000000..00f2844735c9
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_codegen.cpp
@@ -0,0 +1,263 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+int a[100];
+
+// CK1: define {{.*}}i32 @{{.+}}teams_argument_globali(
+int teams_argument_global(int n){
+  int te = n / 128;
+  int th = 128;
+  // discard n_addr
+  // CK1: alloca i32,
+  // CK1: [[TE:%.+]] = alloca i32,
+  // CK1: [[TH:%.+]] = alloca i32,
+  // CK1: [[TE_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TE_PAR:%.+]] = load{{.+}}, {{.+}} [[TE_CAST]],
+  // CK1: [[TH_PAR:%.+]] = load{{.+}}, {{.+}} [[TH_CAST]],
+
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+
+  // CK1: call void @[[OFFL1:.+]](i{{32|64}} [[TE_PAR]], i{{32|64}} [[TH_PAR]],
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd num_teams(te), thread_limit(th) simdlen(64)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  int i;
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+  // CK1: call void @[[OFFL2:.+]](
+  #pragma omp target
+  {{{
+  #pragma omp teams distribute parallel for simd safelen(4) aligned(a) linear(i) 
+  for(i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  }}}
+  // outlined target regions
+  // CK1: define internal void @[[OFFL1]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], i{{32|64}} {{.+}}, {{.+}})
+  // CK1: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+  // CK1: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+  // CK1-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+  // CK1-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+  // CK1-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+  // CK1-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+  // CK1-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+  // CK1-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+  // CK1: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.+}} @__kmpc_fork_call(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]]({{.+}}, {{.+}})
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.+}} @__kmpc_fork_call(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  return a[0];
+}
+
+// CK1-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CK1-DAG: !{!"llvm.loop.vectorize.width", i32 4}
+// CK1-DAG: !{!"llvm.loop.vectorize.width", i32 64}
+
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+// CK2: define {{.*}}i32 @{{.+}}teams_local_argv(
+int teams_local_arg(void) {
+  int n = 100;
+  int a[n], i;
+
+  // CK2: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+  // CK2: call void @[[OFFL1:.+]](
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd safelen(4) aligned(a) linear(i)
+  for(i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  // outlined target region
+  // CK2: define internal void @[[OFFL1]]({{.+}}, {{.+}})
+  // CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK2: ret void
+
+  // CK2: define internal void @[[OUTL1]]({{.+}})
+  // CK2: call void @__kmpc_for_static_init_4(
+  // CK2: call void {{.+}} @__kmpc_fork_call(
+  // CK2: call void @__kmpc_for_static_fini(
+  // CK2: ret void
+
+  return a[0];
+}
+
+// CK2-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CK2-DAG: !{!"llvm.loop.vectorize.width", i32 4}
+
+#endif // CK2
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+#ifdef CK3
+
+// CK3: [[SSI:%.+]] = type { [{{.+}} x i32], float }
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK3: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+    int i;
+  // CK3: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+  // CK3: call void @[[OFFL1:.+]]([[SSI]]* %{{.+}})
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd safelen(4) aligned(a) linear(i)
+    for(i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+      // outlined target region
+  // CK3: define internal void @[[OFFL1]]([[SSI]]* {{.+}})
+  // CK3: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK3: ret void
+
+  // CK3: define internal void @[[OUTL1]]({{.+}})
+  // CK3: call void @__kmpc_for_static_init_4(
+  // CK3: call void {{.+}} @__kmpc_fork_call(
+  // CK3: call void @__kmpc_for_static_fini(
+  // CK3: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+
+// CK3-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CK3-DAG: !{!"llvm.loop.vectorize.width", i32 4}
+#endif // CK3
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+
+#ifdef CK4
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int te = n/128;
+  int th = 128;
+#pragma omp target
+#pragma omp teams distribute parallel for simd num_teams(te) thread_limit(th) simdlen(64)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n], i;
+#pragma omp target
+#pragma omp teams distribute parallel for simd safelen(4) aligned(a) linear(i)
+  for(i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK4:  define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+// CK4: call void @[[OFFL1:.+]]({{.+}})
+// CK4: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK4:  ret
+
+// CK4:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTL1]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void {{.+}} @__kmpc_fork_call(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+// CK4:  define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+// CK4: call void @[[OFFLT:.+]]({{.+}})
+// CK4:  ret
+// CK4-NEXT: }
+
+// CK4: define {{.*}}void @[[OFFLT]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], {{.+}})
+// CK4: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+// CK4: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+// CK4-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+// CK4-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+// CK4-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+// CK4-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+// CK4-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+// CK4-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+// CK4: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTLT]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void {{.+}} @__kmpc_fork_call(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+// CK4-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CK4-DAG: !{!"llvm.loop.vectorize.width", i32 4}
+// CK4-DAG: !{!"llvm.loop.vectorize.width", i32 64}
+
+#endif // CK4
+#endif
+
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_collapse_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_collapse_codegen.cpp
new file mode 100644
index 000000000000..91c451d49a08
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_collapse_codegen.cpp
@@ -0,0 +1,149 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X][Y];
+
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+    // CK1: call i32 @__tgt_target_teams(
+    // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd collapse(2)
+    for(int i = 0; i < X; i++) {
+      for(int j = 0; j < Y; j++) {
+	a[i][j] = (T)0;
+      }
+    }
+    // CK1: define internal void @[[OFFL1]](
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+    // CK1: ret void
+
+    // CK1: define internal void @[[OUTL1]]({{.+}})
+    // discard loop variables not needed here
+    // CK1: [[OMP_UB:%.omp.comb.ub]] = alloca i32,
+    // CK1: store i32 56087, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+    // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void
+
+    // CK1: define internal void @[[PAR_OUTL1]]({{.+}})
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34, {{.+}}, {{.+}},
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void
+
+    return a[0][0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+
+// CK4: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n, int m>
+int tmain(T argc) {
+  T a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = (T)0;
+    }
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int m = 2;
+  int a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = 0;
+    }
+  }
+  return tmain<int, 10, 2>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: [[OMP_UB:%.omp.comb.ub]] = alloca i64,
+// CK2: store i64 {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_8({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i64* [[OMP_UB]],
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_{{[4|8]}}({{.+}}, {{.+}}, i32 34, {{.+}}, {{.+}},
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// discard loop variables not needed here
+// CK2: [[OMP_UB:%.omp.comb.ub]] = alloca i32,
+// CK2: store i32 {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[TPAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34, {{.+}}, {{.+}},
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK4: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_dist_schedule_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_dist_schedule_codegen.cpp
new file mode 100644
index 000000000000..5c1f73066cd4
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_dist_schedule_codegen.cpp
@@ -0,0 +1,268 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd dist_schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd dist_schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+
+// CK1: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for simd
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd dist_schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for simd
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd dist_schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..1219742c752d
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_codegen.cpp
@@ -0,0 +1,501 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+// CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute parallel for simd firstprivate(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}})
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for simd firstprivate(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}})
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]],
+    // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]],
+    // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // skip loop vars
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}})
+    // Skip global and bound tid vars, and prev lb and ub vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // skip loop vars
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP_PRIV:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]],
+
+    // use of private vars
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP_PRIV]]
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+
+// LAMBDA: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for simd firstprivate(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]],
+// CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]],
+// CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]],
+// CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[PAR_OUTL]]({{.+}})
+// Skip global and bound tid vars, and prev lb ub vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+// CHECK: call void @[[TOFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]({{.+}})
+// CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]],
+// CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]],
+// CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]],
+// CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[TPAR_OUTL]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_messages.cpp
index cbbb313c06b3..668337db454b 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_firstprivate_messages.cpp
@@ -144,8 +144,9 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for simd firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for simd lastprivate(argc), firstprivate(argc) // OK
+#pragma omp teams distribute parallel for simd lastprivate(argc), firstprivate(argc)
   for (i = 0; i < argc; ++i) foo();
 
   return 0;
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_if_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_if_codegen.cpp
new file mode 100644
index 000000000000..6f2cef88cbbc
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_if_codegen.cpp
@@ -0,0 +1,187 @@
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple %itanium_abi_triple -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+void fn1();
+void fn2();
+void fn3();
+void fn4();
+void fn5();
+void fn6();
+
+int Arg;
+
+// CHECK-LABEL: define {{.*}}void @{{.+}}gtid_test
+void gtid_test() {
+#pragma omp target
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+#pragma omp teams distribute parallel for simd
+  for(int i = 0 ; i < 100; i++) {}
+  // CHECK: define internal void [[OFFLOADING_FUN_0]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+  // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK:  call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_0:@.+]] to void
+  // CHECK: call void @__kmpc_for_static_fini(
+
+  // CHECK: define{{.+}} void [[OMP_OUTLINED_0]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: ret
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (parallel: false)
+  for(int i = 0 ; i < 100; i++) {
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+  // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @__kmpc_serialized_parallel(
+  // CHECK: call void [[OMP_OUTLINED_1:@.+]](
+  // CHECK: call void @__kmpc_end_serialized_parallel(
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: define{{.+}} void [[OMP_OUTLINED_1]](
+  // CHECK: call void @__kmpc_for_static_init_4(
+  // CHECK: call void @{{.+}}gtid_test
+  // CHECK: call void @__kmpc_for_static_fini(
+  // CHECK: ret
+    gtid_test();
+  }
+}
+
+
+template <typename T>
+int tmain(T Arg) {
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (true)
+  for(int i = 0 ; i < 100; i++) {
+    fn1();
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (false)
+  for(int i = 0 ; i < 100; i++) {
+    fn2();
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (parallel: Arg)
+  for(int i = 0 ; i < 100; i++) {
+    fn3();
+  }
+  return 0;
+}
+
+// CHECK-LABEL: define {{.*}}i{{[0-9]+}} @main()
+int main() {
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_2:@.+]](
+// CHECK: = call {{.*}}i{{.+}} @{{.+}}tmain
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (true)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_0]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_2:@.+]] to void
+    // CHECK: call void @__kmpc_for_static_fini(
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_2]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn4
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    fn4();
+  }
+
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (false)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_1]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void @__kmpc_serialized_parallel(
+    // CHECK: call void [[OMP_OUTLINED_3:@.+]](
+    // CHECK: call void @__kmpc_end_serialized_parallel(
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_3]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn5
+    // CHECK: call void @__kmpc_for_static_fini(
+    fn5();
+  }
+
+#pragma omp target
+#pragma omp teams distribute parallel for simd if (Arg)
+  for(int i = 0 ; i < 100; i++) {
+    // CHECK: define internal void [[OFFLOADING_FUN_2]](
+    // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}}* [[OMP_TEAMS_OUTLINED_2:@.+]] to {{.+}})
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_2]](
+
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, {{.+}}* [[OMP_OUTLINED_4:@.+]] to void
+    // CHECK: call void @__kmpc_serialized_parallel(
+    // CHECK: call void [[OMP_OUTLINED_4:@.+]](
+    // CHECK: call void @__kmpc_end_serialized_parallel(
+    // CHECK: call void @__kmpc_for_static_fini(
+
+    // CHECK: define{{.+}} void [[OMP_OUTLINED_4]](
+    // CHECK: call void @__kmpc_for_static_init_4(
+    // CHECK: call {{.*}}void @{{.+}}fn6
+    // CHECK: call void @__kmpc_for_static_fini(
+    fn6();
+  }
+
+  return tmain(Arg);
+}
+
+// CHECK-LABEL: define {{.+}} @{{.+}}tmain
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, void {{.+}}* [[T_OUTLINE_FUN_1:@.+]] to void
+// CHECK: call void @__kmpc_for_static_fini(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_1]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn1
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @__kmpc_serialized_parallel(
+// CHECK: call void [[T_OUTLINE_FUN_2:@.+]](
+// CHECK: call {{.*}}void @__kmpc_end_serialized_parallel(
+// CHECK: call void @__kmpc_for_static_fini(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_2]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn2
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{.+}} 2, void {{.+}}* [[T_OUTLINE_FUN_3:@.+]] to void
+// CHECK: call {{.*}}void @__kmpc_serialized_parallel(
+// call void [[T_OUTLINE_FUN_3:@.+]](
+// CHECK: call {{.*}}void @__kmpc_end_serialized_parallel(
+
+// CHECK: define internal {{.*}}void [[T_OUTLINE_FUN_3]]
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call {{.*}}void @{{.+}}fn3
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..5df893d525f4
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_codegen.cpp
@@ -0,0 +1,599 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // skip final branch
+      // LAMBDA: br
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+
+      // LAMBDA: define{{.*}} internal{{.*}} void @[[LPAR_OUTL]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, {{.+}}, {{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+
+      // skip 'final' simd branch and block (checked as part of simd)
+      // LAMBDA: br
+
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+
+      [&]() {
+	// LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+	// LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+	g = 2;
+	g1 = 2;
+	svar = 4;
+	sfvar = 8.0;
+	// LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+	// LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+	// LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+	// LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+	// LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+	// LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+	// LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+	// LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+	// LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+	// LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+	// LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+	// LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+
+// LAMBDA: !{!"llvm.loop.vectorize.enable", i1 true}
+
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// gbl and bound tid vars, prev lb and ub vars
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// skip final branch and block
+// CHECK: br
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void [[TPAR_OUTL:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, {{.+}}, {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid, and prev lb and ub vars
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+
+// CHECK: call void @__kmpc_for_static_fini(
+
+// skip final branch and block
+// CHECK: br
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_messages.cpp
index fce59b924f67..204be4c95e91 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -255,12 +255,14 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for simd lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{firstprivate variable cannot be lastprivate}} expected-note@+2 {{defined as firstprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for simd firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp teams distribute parallel for simd firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for simd lastprivate(n) firstprivate(n) // OK
+#pragma omp teams distribute parallel for simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_linear_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_linear_messages.cpp
index 5b91a3c4cccc..c0a9baf916dd 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_linear_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_linear_messages.cpp
@@ -18,34 +18,42 @@ void test_linear_colons()
 {
   int B = 0;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(B:bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(B:B::bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(X::x : ::z)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(B,::z, X::x)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(::z)
   for (int i = 0; i < 10; ++i) ;
@@ -54,6 +62,7 @@ void test_linear_colons()
 #pragma omp teams distribute parallel for simd linear(B::bfoo()) // expected-error {{expected variable name}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(B::ib,B:C1+C2)
   for (int i = 0; i < 10; ++i) ;
@@ -76,6 +85,7 @@ template<int L, class T, class N> T test_template(T* arr, N num) {
 
 template<int LEN> int test_warn() {
   int ind2 = 0;
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
   #pragma omp target
   #pragma omp teams distribute parallel for simd linear(ind2:LEN) // expected-warning {{zero linear step (ind2 should probably be const)}}
   for (int i = 0; i < 100; i++) {
@@ -133,10 +143,12 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute parallel for simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
@@ -145,6 +157,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute parallel for simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear (argc : 5)
   for (int k = 0; k < argc; ++k) ++k;
@@ -161,6 +174,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute parallel for simd linear (argv[1]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(e, g)
   for (int k = 0; k < argc; ++k) ++k;
@@ -169,32 +183,11 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute parallel for simd linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear(i)
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target
-    #pragma omp teams distribute parallel for simd linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target
-#pragma omp teams distribute parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target
-#pragma omp teams distribute parallel for simd linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target
-#pragma omp teams distribute parallel for simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
   return 0;
 }
 
@@ -230,10 +223,12 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
@@ -242,6 +237,7 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd linear (argc)
   for (int k = 0; k < argc; ++k) ++k;
@@ -267,26 +263,6 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute parallel for simd linear(h, C::x) // expected-error 2 {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int i;
-    #pragma omp target
-    #pragma omp teams distribute parallel for simd linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
-
-    #pragma omp target
-    #pragma omp teams distribute parallel for simd linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
-  }
-
-#pragma omp target
-#pragma omp teams distribute parallel for simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute parallel for simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
   foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
   return 0;
 }
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_loop_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_loop_messages.cpp
index 03325623502f..66361b64eb84 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_loop_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -693,8 +693,9 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute parallel for simd lastprivate(s) firstprivate(s)
+#pragma omp teams distribute parallel for simd lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_messages.cpp
index 7876c38980ee..12ac8f9a9479 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_messages.cpp
@@ -9,6 +9,9 @@ static int pvt;
 #pragma omp teams distribute parallel for simd // expected-error {{unexpected OpenMP directive '#pragma omp teams distribute parallel for simd'}}
 
 int main(int argc, char **argv) {
+  #pragma omp target
+  #pragma omp teams distribute parallel for simd
+  f; // expected-error {{use of undeclared identifier 'f'}}
 #pragma omp target
 #pragma omp teams distribute parallel for simd { // expected-warning {{extra tokens at the end of '#pragma omp teams distribute parallel for simd' are ignored}}
   for (int i = 0; i < argc; ++i)
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_num_threads_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_num_threads_codegen.cpp
new file mode 100644
index 000000000000..93051af6161b
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_num_threads_codegen.cpp
@@ -0,0 +1,123 @@
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple %itanium_abi_triple -fexceptions -fcxx-exceptions -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple %itanium_abi_triple -fexceptions -fcxx-exceptions -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+typedef __INTPTR_TYPE__ intptr_t;
+
+// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[S_TY:%.+]] = type { [[INTPTR_T_TY:i[0-9]+]], [[INTPTR_T_TY]], [[INTPTR_T_TY]] }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_2:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+void foo();
+
+struct S {
+  intptr_t a, b, c;
+  S(intptr_t a) : a(a) {}
+  operator char() { return a; }
+  ~S() {}
+};
+
+template <typename T, int C>
+int tmain() {
+#pragma omp target
+#pragma omp teams distribute parallel for simd num_threads(C)
+  for (int i = 0; i < 100; i++)
+    foo();
+#pragma omp target
+#pragma omp teams distribute parallel for simd num_threads(T(23))
+  for (int i = 0; i < 100; i++)
+    foo();
+  return 0;
+}
+
+int main() {
+  S s(0);
+  char a = s;
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOADING_FUN_1:@.+]](
+// CHECK: invoke{{.+}} [[TMAIN_5:@.+]]()
+// CHECK: invoke{{.+}} [[TMAIN_1:@.+]]()
+#pragma omp target
+  // CHECK: define internal void [[OFFLOADING_FUN_0]](
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+#pragma omp teams distribute parallel for simd num_threads(2)
+  for (int i = 0; i < 100; i++) {
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_0]](
+    // CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 2)
+    // CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+    foo();
+  }
+#pragma omp target
+  // CHECK: define internal void [[OFFLOADING_FUN_1]](
+
+  // CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}}* [[OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+#pragma omp teams distribute parallel for simd num_threads(a)
+  for (int i = 0; i < 100; i++) {
+    // CHECK: define{{.+}} void [[OMP_TEAMS_OUTLINED_1]](
+    // CHECK-DAG: [[A_ADDR:%.+]] = alloca i8*,
+    // CHECK-DAG: [[A_REF:%.+]] = load i8*, i8** [[A_ADDR]],
+    // CHECK-DAG: [[A_VAL:%.+]] = load i8, i8* [[A_REF]],
+    // CHECK-DAG: [[A_EXT:%.+]] = sext i8 [[A_VAL]] to {{.+}}
+    // CHECK: call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 [[A_EXT]])
+    // CHECK: call {{.*}}void {{.*}} @__kmpc_fork_call(
+    foo();
+  }
+  return a + tmain<char, 5>() + tmain<S, 1>();
+}
+
+// tmain 5
+// CHECK-DAG: define {{.*}}i{{[0-9]+}} [[TMAIN_5]]()
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_0:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_1:@.+]](
+
+// tmain 1
+// CHECK-DAG: define {{.*}}i{{[0-9]+}} [[TMAIN_1]]()
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_2:@.+]](
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[T_OFFLOADING_FUN_3:@.+]](
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_0]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_0:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_0]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 5)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_1]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 23)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_2]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}}* [[T_OMP_TEAMS_OUTLINED_2:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_2]](
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 1)
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: define internal void [[T_OFFLOADING_FUN_3]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}}* [[T_OMP_TEAMS_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} void [[T_OMP_TEAMS_OUTLINED_3]]({{.+}}, {{.+}}, {{.+}} [[NUM_TH_CPT_IN:%.+]])
+// CHECK: [[NUM_TH_CPT:%.+]] = alloca i8*,
+// CHECK: store {{.+}} [[NUM_TH_CPT_IN]], {{.+}} [[NUM_TH_CPT]],
+// CHECK: [[NUM_TH_REF:%.+]] = load{{.+}}, {{.+}} [[NUM_TH_CPT]],
+// CHECK-DAG:   [[NUM_TH_VAL:%.+]] = load {{.+}}, {{.+}} [[NUM_TH_REF]],
+// CHECK-DAG:   [[NUM_TH_SEXT:%.+]] = sext i8 [[NUM_TH_VAL]] to {{.+}}
+// CHECK:       call {{.*}}void @__kmpc_push_num_threads([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 {{.+}}, i32 [[NUM_TH_SEXT]])
+// CHECK:       call {{.*}}void {{.*}} @__kmpc_fork_call(
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp
new file mode 100644
index 000000000000..a6dbd29a5b96
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_private_codegen.cpp
@@ -0,0 +1,336 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute parallel for simd private(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for simd private(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}})
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: ret void
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+
+// LAMBDA: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for simd private(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]()
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[PAR_OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[SIVAR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0,
+// CHECK: call void @[[TOFFL1:.+]]()
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.+}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL1:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define internal void @[[TPAR_OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// prev lb and ub
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// iter variables
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i32,
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_proc_bind_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_proc_bind_codegen.cpp
new file mode 100644
index 000000000000..7f95a877ab85
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_proc_bind_codegen.cpp
@@ -0,0 +1,93 @@
+// add -fopenmp-targets
+
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+typedef __INTPTR_TYPE__ intptr_t;
+
+// CHECK-DAG: [[IDENT_T_TY:%.+]] = type { i32, i32, i32, i32, i8* }
+// CHECK-DAG: [[STR:@.+]] = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00"
+// CHECK-DAG: [[DEF_LOC_2:@.+]] = private unnamed_addr constant [[IDENT_T_TY]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* [[STR]], i32 0, i32 0) }
+
+void foo();
+
+struct S {
+  intptr_t a, b, c;
+  S(intptr_t a) : a(a) {}
+  operator char() { return a; }
+  ~S() {}
+};
+
+template <typename T>
+T tmain() {
+#pragma omp target
+#pragma omp teams distribute parallel for simd proc_bind(master)
+  for(int i = 0; i < 1000; i++) {}
+  return T();
+}
+
+int main() {
+  // CHECK-LABEL: @main
+#pragma omp target
+#pragma omp teams distribute parallel for simd proc_bind(spread)
+  for(int i = 0; i < 1000; i++) {}
+#pragma omp target
+#pragma omp teams distribute parallel for simd proc_bind(close)
+  for(int i = 0; i < 1000; i++) {}
+  return tmain<int>();
+}
+
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL1:@.+]]()
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL2:@.+]]()
+// CHECK: [[CALL_RET:%.+]] = call{{.+}} i32 [[TMAIN:@.+]]()
+// CHECK: ret i32 [[CALL_RET]]
+
+// CHECK: define{{.+}} void [[OFFL1]](
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_1]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 4)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: define{{.+}} [[OFFL2]]()
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_1:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_1]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 3)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: define{{.+}} [[TMAIN]]()
+// CHECK: call {{.*}}@__tgt_target_teams({{.+}})
+// CHECK: call void [[OFFL3:@.+]]()
+
+// CHECK: define{{.+}} [[OFFL3]]()
+// CHECK: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, {{.+}}, {{.+}}* [[OMP_OUTLINED_3:@.+]] to {{.+}})
+
+// CHECK: define{{.+}} [[OMP_OUTLINED_3]](i32* {{.+}} [[GTID_IN:%.+]],
+// CHECK: [[GTID_ADDR:%.+]] = alloca i32*,
+// CHECK: store i32* [[GTID_IN]], i32** [[GTID_ADDR]],
+// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_ADDR]],
+// CHECK: [[GTID_VAL:%.+]] = load i32, i32* [[GTID_REF]],
+// CHECK: call {{.*}}void @__kmpc_push_proc_bind([[IDENT_T_TY]]* [[DEF_LOC_2]], i32 [[GTID_VAL]], i32 2)
+// CHECK: call {{.*}}void (%ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_reduction_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_reduction_codegen.cpp
new file mode 100644
index 000000000000..384ac4b5f32a
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_reduction_codegen.cpp
@@ -0,0 +1,351 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target
+#pragma omp teams distribute parallel for simd reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute parallel for simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[LPAR_OUTL:.+]] to
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: br
+
+    // LAMBDA: define internal void @[[LPAR_OUTL]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+
+    // Skip global and bound tid vars, and prev lb and ub vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // skip loop vars
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: br
+
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+
+// LAMBDA: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#else
+#pragma omp target
+#pragma omp teams distribute parallel for simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK-64: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: define internal void @[[PAR_OUTL]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars, and prev lb and ub
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// skip loop vars
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i{{64|32}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK-64: [[TVAR_CONV:%.+]] = bitcast{{.+}} [[TVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[TPAR_OUTL:.+]] to
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: define internal void @[[TPAR_OUTL]]({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars, and prev lb and ub vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// skip loop vars
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: alloca i32,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce_nowait({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce_nowait({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: br
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_reduction_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_reduction_messages.cpp
index 47494b94ff0e..b0522e8f69d2 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_reduction_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_reduction_messages.cpp
@@ -19,9 +19,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_schedule_codegen.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_schedule_codegen.cpp
new file mode 100644
index 000000000000..71e8ee8e8110
--- /dev/null
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_schedule_codegen.cpp
@@ -0,0 +1,402 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL4:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd schedule(dynamic)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL5:.+]](
+    #pragma omp target
+    #pragma omp teams distribute parallel for simd schedule(dynamic, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 33,
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL4]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL4:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL4]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL4:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL4]]({{.+}})
+  // CK1: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+  // CK1: call {{.+}} @__kmpc_dispatch_next_4(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL5]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL5:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL5]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL5:.+]] to
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[PAR_OUTL5]]({{.+}})
+  // CK1: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+  // CK1: call {{.+}} @__kmpc_dispatch_next_4(
+  // CK1: ret void
+
+  // CK1: !{!"llvm.loop.vectorize.enable", i1 true}
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for simd
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd schedule(dynamic)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd schedule(dynamic, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+  int m = 10;
+#pragma omp target
+#pragma omp teams distribute parallel for simd
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd dist_schedule(static, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd schedule(dynamic)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute parallel for simd schedule(dynamic, m)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL4:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL5:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL4]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTL4:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL4]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL4:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL4]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+
+// CK2: define {{.*}}void @[[OFFL5]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTL5:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL5]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTL5:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTL5]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT4:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT5:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT1:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT2:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 34,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT3:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 33,
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT4]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT4:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT4]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT4:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT4]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT5]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 {{.+}}, {{.+}} @[[OUTLT5:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT5]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4(
+// CK2: call void {{.*}} @__kmpc_fork_call({{.+}}, {{.+}}, {{.+}} @[[PAR_OUTLT5:.+]] to
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define internal void @[[PAR_OUTLT5]]({{.+}})
+// CK2: call void @__kmpc_dispatch_init_4({{.+}}, {{.+}}, i32 35,
+// CK2: call {{.+}} @__kmpc_dispatch_next_4(
+// CK2: ret void
+
+// CK2: !{!"llvm.loop.vectorize.enable", i1 true}
+
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_parallel_for_simd_shared_messages.cpp b/test/OpenMP/teams_distribute_parallel_for_simd_shared_messages.cpp
index f13146641a38..1089555889f0 100644
--- a/test/OpenMP/teams_distribute_parallel_for_simd_shared_messages.cpp
+++ b/test/OpenMP/teams_distribute_parallel_for_simd_shared_messages.cpp
@@ -84,7 +84,7 @@ int main(int argc, char **argv) {
   #pragma omp teams distribute parallel for simd shared (S1) // expected-error {{'S1' does not refer to a value}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
-  #pragma omp teams distribute parallel for simd shared (a, b, c, d, f)
+  #pragma omp teams distribute parallel for simd shared (a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
   #pragma omp teams distribute parallel for simd shared (argv[1]) // expected-error {{expected variable name}}
diff --git a/test/OpenMP/teams_distribute_private_codegen.cpp b/test/OpenMP/teams_distribute_private_codegen.cpp
new file mode 100644
index 000000000000..8e4a61cd7cb1
--- /dev/null
+++ b/test/OpenMP/teams_distribute_private_codegen.cpp
@@ -0,0 +1,236 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute private(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute private(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}})
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute private(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]()
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[SIVAR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0,
+// CHECK: call void @[[TOFFL1:.+]]()
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+#endif
diff --git a/test/OpenMP/teams_distribute_reduction_codegen.cpp b/test/OpenMP/teams_distribute_reduction_codegen.cpp
new file mode 100644
index 000000000000..961891da98d0
--- /dev/null
+++ b/test/OpenMP/teams_distribute_reduction_codegen.cpp
@@ -0,0 +1,217 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target
+#pragma omp teams distribute reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+    
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK-64: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i{{64|32}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK-64: [[TVAR_CONV:%.+]] = bitcast{{.+}} [[TVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+#endif
diff --git a/test/OpenMP/teams_distribute_reduction_messages.cpp b/test/OpenMP/teams_distribute_reduction_messages.cpp
index d0a6d296c0f5..f11236f0ce27 100644
--- a/test/OpenMP/teams_distribute_reduction_messages.cpp
+++ b/test/OpenMP/teams_distribute_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/teams_distribute_shared_messages.cpp b/test/OpenMP/teams_distribute_shared_messages.cpp
index b9e096ce1c24..ecf6f2e03673 100644
--- a/test/OpenMP/teams_distribute_shared_messages.cpp
+++ b/test/OpenMP/teams_distribute_shared_messages.cpp
@@ -84,7 +84,7 @@ int main(int argc, char **argv) {
   #pragma omp teams distribute shared (S1) // expected-error {{'S1' does not refer to a value}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
-  #pragma omp teams distribute shared (a, b, c, d, f)
+  #pragma omp teams distribute shared (a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
   #pragma omp teams distribute shared (argv[1]) // expected-error {{expected variable name}}
diff --git a/test/OpenMP/teams_distribute_simd_ast_print.cpp b/test/OpenMP/teams_distribute_simd_ast_print.cpp
index aaf6745774c5..e751413c4232 100644
--- a/test/OpenMP/teams_distribute_simd_ast_print.cpp
+++ b/test/OpenMP/teams_distribute_simd_ast_print.cpp
@@ -29,9 +29,10 @@ public:
       ++this->a.a;
   }
   S7 &operator=(S7 &s) {
+    int k;
 #pragma omp target
-#pragma omp teams distribute simd private(a) private(this->a)
-    for (int k = 0; k < s.a.a; ++k)
+#pragma omp teams distribute simd private(a) private(this->a) linear(k)
+    for (k = 0; k < s.a.a; ++k)
       ++s.a.a;
     return *this;
   }
@@ -56,7 +57,7 @@ public:
 // CHECK: #pragma omp target
 // CHECK-NEXT: #pragma omp teams distribute simd private(this->a) private(this->a) private(T::a)
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute simd private(this->a) private(this->a)
+// CHECK-NEXT: #pragma omp teams distribute simd private(this->a) private(this->a) linear(k)
 // CHECK: #pragma omp target
 // CHECK-NEXT: #pragma omp teams distribute simd default(none) private(b) firstprivate(argv) shared(d) reduction(+: c) reduction(max: e) num_teams(f) thread_limit(d)
 // CHECK: #pragma omp target
@@ -155,11 +156,11 @@ T tmain(T argc) {
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
-#pragma omp teams distribute simd simdlen(clen-1) linear(d)
+#pragma omp teams distribute simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute simd simdlen(clen - 1) linear(d)
+// CHECK-NEXT: #pragma omp teams distribute simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
@@ -218,11 +219,11 @@ int main (int argc, char **argv) {
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
-#pragma omp teams distribute simd simdlen(clen-1) linear(d)
+#pragma omp teams distribute simd simdlen(clen-1)
   for (int k = 0; k < 10; ++k)
     e += d + argc;
 // CHECK: #pragma omp target
-// CHECK-NEXT: #pragma omp teams distribute simd simdlen(clen - 1) linear(d)
+// CHECK-NEXT: #pragma omp teams distribute simd simdlen(clen - 1)
 // CHECK-NEXT: for (int k = 0; k < 10; ++k)
 // CHECK-NEXT: e += d + argc;
 #pragma omp target
diff --git a/test/OpenMP/teams_distribute_simd_codegen.cpp b/test/OpenMP/teams_distribute_simd_codegen.cpp
new file mode 100644
index 000000000000..cad5854a2f1e
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_codegen.cpp
@@ -0,0 +1,251 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+int a[100];
+
+// CK1: define {{.*}}i32 @{{.+}}teams_argument_globali(
+int teams_argument_global(int n) {
+  int i;
+  int te = n / 128;
+  int th = 128;
+  // discard n_addr and i
+  // CK1: alloca i32,
+  // CK1: alloca i32,
+  // CK1: [[TE:%.+]] = alloca i32,
+  // CK1: [[TH:%.+]] = alloca i32,
+  // CK1: [[TE_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_CAST:%.+]] = alloca i{{32|64}},
+  // CK1: [[TE_PAR:%.+]] = load{{.+}}, {{.+}} [[TE_CAST]],
+  // CK1: [[TH_PAR:%.+]] = load{{.+}}, {{.+}} [[TH_CAST]],
+
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 {{.+}}, i32 {{.+}})
+
+  // CK1: call void @[[OFFL1:.+]](i{{32|64}} [[TE_PAR]], i{{32|64}} [[TH_PAR]],
+  #pragma omp target
+  #pragma omp teams distribute simd num_teams(te), thread_limit(th) aligned(a) simdlen(16) linear(i)
+  for(i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  // CK1: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 2, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK1: call void @[[OFFL2:.+]](i{{64|32}} %{{.+}})
+  #pragma omp target
+  {{{
+  #pragma omp teams distribute simd safelen(32)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  }}}
+
+  // outlined target regions
+  // CK1: define internal void @[[OFFL1]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], [100 x i{{32|64}}]* {{.+}}, i{{32|64}} {{.+}}, {{.+}})
+  // CK1: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+  // CK1: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+  // CK1: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+  // CK1-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+  // CK1-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+  // CK1-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+  // CK1-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+  // CK1-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+  // CK1-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+  // CK1: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  // CK1: define internal void @[[OFFL2]]({{.+}}, {{.+}})
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 2, {{.+}} @[[OUTL2:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4(
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void
+
+  return a[0];
+}
+
+// CK1-DAG: !{!"llvm.loop.vectorize.width", i32 16}
+// CK1-DAG: !{!"llvm.loop.vectorize.enable", i1 true}
+// CK1-DAG: !{!"llvm.loop.vectorize.width", i32 32}
+
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+// CK2: define {{.*}}i32 @{{.+}}teams_local_argv(
+int teams_local_arg(void) {
+  int n = 100;
+  int a[n];
+
+  // CK2: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK2: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+  #pragma omp target
+  #pragma omp teams distribute simd
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+
+  // outlined target region
+  // CK2: define internal void @[[OFFL1]]({{.+}}, {{.+}})
+  // CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK2: ret void
+
+  // CK2: define internal void @[[OUTL1]]({{.+}})
+  // CK2: call void @__kmpc_for_static_init_4(
+  // CK2: call void @__kmpc_for_static_fini(
+  // CK2: ret void  
+
+  return a[0];
+}
+// CK2: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK2
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-64
+// RUN: %clang_cc1 -DCK3 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK3 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK3 --check-prefix CK3-32
+#ifdef CK3
+
+// CK3: [[SSI:%.+]] = type { [{{.+}} x i32], float }
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK3: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK3: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+  // CK3: call void @[[OFFL1:.+]]([[SSI]]* %{{.+}})
+    #pragma omp target
+    #pragma omp teams distribute simd
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+
+      // outlined target region
+  // CK3: define internal void @[[OFFL1]]([[SSI]]* {{.+}})
+  // CK3: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+  // CK3: ret void
+
+  // CK3: define internal void @[[OUTL1]]({{.+}})
+  // CK3: call void @__kmpc_for_static_init_4(
+  // CK3: call void @__kmpc_for_static_fini(
+  // CK3: ret void  
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+// CK3: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK3
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-64
+// RUN: %clang_cc1 -DCK4 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK4 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK4 --check-prefix CK4-32
+
+#ifdef CK4
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+  int te = n/128;
+  int th = 128;
+#pragma omp target
+#pragma omp teams distribute simd num_teams(te) thread_limit(th)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target
+#pragma omp teams distribute simd
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK4:  define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}, i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CK4: call void @[[OFFL1:.+]]({{.+}})
+// CK4: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK4:  ret
+
+// CK4:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTL1]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+// CK4:  define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK4:   call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 {{.+}}, i32 {{.+}})
+// CK4: call void @[[OFFLT:.+]]({{.+}})
+// CK4:  ret
+// CK4-NEXT: }
+
+// CK4: define {{.*}}void @[[OFFLT]](i{{32|64}} [[TE_ARG:%.+]], i{{32|64}} [[TH_ARG:%.+]], {{.+}})
+// CK4: [[TE_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: [[TH_ADDR:%.+]] = alloca i{{32|64}},
+// CK4: store{{.+}} [[TE_ARG]], {{.+}} [[TE_ADDR]],
+// CK4: store{{.+}} [[TH_ARG]], {{.+}} [[TH_ADDR]],
+// CK4-64: [[TE_CONV:%.+]] = bitcast{{.+}} [[TE_ADDR]] to
+// CK4-64: [[TH_CONV:%.+]] = bitcast{{.+}} [[TH_ADDR]] to
+// CK4-64: [[TE_VAL:%.+]] = load i32, i32* [[TE_CONV]],
+// CK4-64: [[TH_VAL:%.+]] = load i32, i32* [[TH_CONV]],
+// CK4-32: [[TE_VAL:%.+]] = load i32, i32* [[TE_ADDR]],
+// CK4-32: [[TH_VAL:%.+]] = load i32, i32* [[TH_ADDR]],
+// CK4: {{%.+}} = call i32 @__kmpc_push_num_teams({{.+}}, {{.+}}, i32 [[TE_VAL]], i32 [[TH_VAL]])
+// CK4: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT:.+]] to {{.+}}, {{.+}}, {{.+}})
+// CK4: ret void
+
+// CK4: define internal void @[[OUTLT]]({{.+}})
+// CK4: call void @__kmpc_for_static_init_4(
+// CK4: call void @__kmpc_for_static_fini(
+// CK4: ret void
+
+// CK4: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK4
+#endif
+
diff --git a/test/OpenMP/teams_distribute_simd_collapse_codegen.cpp b/test/OpenMP/teams_distribute_simd_collapse_codegen.cpp
new file mode 100644
index 000000000000..071d1672fad1
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_collapse_codegen.cpp
@@ -0,0 +1,131 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X][Y];
+
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+    
+    // CK1: call i32 @__tgt_target_teams(
+    // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute simd collapse(2)
+    for(int i = 0; i < X; i++) {
+      for(int j = 0; j < Y; j++) {
+	a[i][j] = (T)0;
+      }
+    }
+    // CK1: define internal void @[[OFFL1]](
+    // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+    // CK1: ret void
+    
+    // CK1: define internal void @[[OUTL1]]({{.+}})
+    // discard loop variables not needed here
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: = alloca i32,
+    // CK1: [[OMP_UB:%.+]] = alloca i32,
+    // CK1: store i32 56087, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+    // CK1: call void @__kmpc_for_static_fini(
+    // CK1: ret void  
+
+    return a[0][0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+
+// CK1: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n, int m>
+int tmain(T argc) {
+  T a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute simd collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = (T)0;
+    }
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int m = 2;
+  int a[n][m];
+  #pragma omp target
+  #pragma omp teams distribute simd collapse(2)
+  for(int i = 0; i < n; i++) {
+    for(int j = 0; j < m; j++) {
+      a[i][j] = 0;
+    }
+  }
+  return tmain<int, 10, 2>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i64,
+// CK2: store i64 {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_8({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i64* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// discard loop variables not needed here
+// CK2: [[OMP_UB:%.omp.ub]] = alloca i32,
+// CK2: store i32 {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92, {{.+}}, {{.+}}, i32* [[OMP_UB]],
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_simd_dist_schedule_codegen.cpp b/test/OpenMP/teams_distribute_simd_dist_schedule_codegen.cpp
new file mode 100644
index 000000000000..599e18f58a16
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_dist_schedule_codegen.cpp
@@ -0,0 +1,208 @@
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-64
+// RUN: %clang_cc1 -DCK1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK1 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK1 --check-prefix CK1-32
+#ifdef CK1
+
+template <typename T, int X, long long Y>
+struct SS{
+  T a[X];
+  float b;
+  // CK1: define {{.*}}i32 @{{.+}}foo{{.+}}(
+  int foo(void) {
+
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL1:.+]](
+    #pragma omp target
+    #pragma omp teams distribute simd
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL2:.+]](
+    #pragma omp target
+    #pragma omp teams distribute simd dist_schedule(static)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: call i32 @__tgt_target_teams(
+  // CK1: call void @[[OFFL3:.+]](
+    #pragma omp target
+    #pragma omp teams distribute simd dist_schedule(static, X/2)
+    for(int i = 0; i < X; i++) {
+      a[i] = (T)0;
+    }
+  // CK1: define internal void @[[OFFL1]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL1]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void  
+
+  // CK1: define internal void @[[OFFL2]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL2:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL2]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void  
+
+  // CK1: define internal void @[[OFFL3]](
+  // CK1: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL3:.+]] to {{.+}},
+  // CK1: ret void
+
+  // CK1: define internal void @[[OUTL3]]({{.+}})
+  // CK1: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+  // CK1: call void @__kmpc_for_static_fini(
+  // CK1: ret void  
+
+    return a[0];
+  }
+};
+
+int teams_template_struct(void) {
+  SS<int, 123, 456> V;
+  return V.foo();
+
+}
+// CK1: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK1
+
+// Test host codegen.
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-64
+// RUN: %clang_cc1 -DCK2 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCK2 -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CK2 --check-prefix CK2-32
+#ifdef CK2
+
+template <typename T, int n>
+int tmain(T argc) {
+  T a[n];
+#pragma omp target
+#pragma omp teams distribute simd
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+#pragma omp target
+#pragma omp teams distribute simd dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = (T)0;
+  }
+  return 0;
+}
+
+int main (int argc, char **argv) {
+  int n = 100;
+  int a[n];
+#pragma omp target
+#pragma omp teams distribute simd
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute simd dist_schedule(static)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+#pragma omp target
+#pragma omp teams distribute simd dist_schedule(static, n)
+  for(int i = 0; i < n; i++) {
+    a[i] = 0;
+  }
+  return tmain<int, 10>(argc);
+}
+
+// CK2: define {{.*}}i32 @{{[^,]+}}(i{{.+}}{{.+}} %[[ARGC:.+]], {{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFL3:.+]]({{.+}})
+// CK2: {{%.+}} = call{{.*}} i32 @[[TMAIN:.+]]({{.+}})
+// CK2: ret
+
+// CK2:  define {{.*}}void @[[OFFL1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}void @[[OFFL2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[OUTL2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFL3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[OUTL3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTL3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: define {{.*}}i32 @[[TMAIN]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT1:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT2:.+]]({{.+}})
+// CK2: call i32 @__tgt_target_teams(
+// CK2: call void @[[OFFLT3:.+]]({{.+}})
+// CK2:  ret
+// CK2-NEXT: }
+
+// CK2:  define {{.*}}void @[[OFFLT1]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT1:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT1]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT2]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT2:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT2]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 92
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2:  define {{.*}}void @[[OFFLT3]]({{.+}})
+// CK2: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTLT3:.+]] to {{.+}},
+// CK2: ret void
+
+// CK2: define internal void @[[OUTLT3]]({{.+}})
+// CK2: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, i32 91
+// CK2: call void @__kmpc_for_static_fini(
+// CK2: ret void
+
+// CK2: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif // CK2
+#endif // #ifndef HEADER
diff --git a/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp b/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp
new file mode 100644
index 000000000000..e503f5267f19
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_firstprivate_codegen.cpp
@@ -0,0 +1,337 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+// CHECK-DAG: [[ST_TY:%.+]] = type { i{{[0-9]+}}, i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute simd firstprivate(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 3, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](i{{64|32}} %{{.+}})
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute simd firstprivate(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}}, i{{64|32}} {{%.+}})
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: {{%.+}} = alloca i{{[0-9]+}},
+    // LAMBDA: [[G_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA-DAG: [[G_CAST_VAL:%.+]] = load{{.+}} [[G_CAST]],
+    // LAMBDA-DAG: [[G1_CAST_VAL:%.+]] = load{{.+}} [[G1_CAST]],
+    // LAMBDA-DAG: [[SIVAR_CAST_VAL:%.+]] = load{{.+}} [[SIVAR_CAST]],
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 3, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[G_CAST_VAL]], {{.+}} [[G1_CAST_VAL]], {{.+}} [[SIVAR_CAST_VAL]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[G_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_TMP:%.+]] = alloca i32*,
+    // skip loop vars
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[G1_ADDR]],
+    // LAMBDA-DAG: store {{.+}}, {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA-DAG: [[G_CONV:%.+]] = bitcast {{.+}} [[G_ADDR]] to
+    // LAMBDA-DAG: [[G1_CONV:%.+]] = bitcast {{.+}} [[G1_ADDR]] to
+    // LAMBDA-DAG: [[SIVAR_CONV:%.+]] = bitcast {{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA-DAG: store{{.+}} [[G1_CONV]], {{.+}} [[G1_TMP]],
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_CONV]],
+    // LAMBDA-DAG: [[G1:%.+]] = load{{.+}}, {{.+}}* [[G1_TMP]]
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_CONV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[G1_TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute simd firstprivate(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 5, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]({{.+}})
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[SIVAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[VEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_PRIV]],
+// CHECK-DAG: [[T_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_CAST]],
+// CHECK-DAG: [[S_ARR_TE_PAR:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_PRIV]],
+// CHECK-DAG: [[VAR_TE_PAR:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[VAR_PRIV]],
+// CHECK-DAG: [[SIVAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_CAST]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 5, {{.+}} @[[OUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[VEC_TE_PAR]], i{{[0-9]+}} [[T_VAR_TE_PAR]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_TE_PAR]], [[S_FLOAT_TY]]* [[VAR_TE_PAR]], i{{[0-9]+}} [[SIVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* {{.+}}, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* {{.+}}, [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[SIVAR_ADDR]],
+
+// T_VAR and SIVAR
+// CHECK-DAG-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK-DAG-64: [[CONV_SIVAR:%.+]] = bitcast i64* [[SIVAR_ADDR]] to i32*
+
+// preparation vars
+// CHECK-DAG: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-DAG: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] =  bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}}], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG-32: {{.+}} = {{.+}} [[SIVAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_SIVAR]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 4, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[TOFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]({{.+}})
+// CHECK: [[TT_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[TVEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[TS_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca [[S_INT_TY]]*,
+// CHECK: [[TT_VAR_CAST:%.+]] = alloca i{{[0-9]+}},
+
+// CHECK-DAG: [[TVEC_TE_PAR:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[TVEC_PRIV]],
+// CHECK-DAG: [[TT_VAR_TE_PAR:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[TT_VAR_CAST]],
+// CHECK-DAG: [[TS_ARR_TE_PAR:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[TS_ARR_PRIV]],
+// CHECK-DAG: [[TVAR_TE_PAR:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TVAR_PRIV]],
+
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}} @[[TOUTL1:.+]] to {{.+}}, [2 x i{{[0-9]+}}]* [[TVEC_TE_PAR]], i{{[0-9]+}} [[TT_VAR_TE_PAR]], [2 x [[S_INT_TY]]]* [[TS_ARR_TE_PAR]], [[S_INT_TY]]* [[TVAR_TE_PAR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_INT_TY]]*,
+// Skip temp vars for loop
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[AGG_TMP1:%.+]] = alloca [[ST_TY]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[AGG_TMP2:%.+]] = alloca [[ST_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// param copy
+// CHECK: store [2 x i{{[0-9]+}}]* {{.+}}, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}} {{.+}}, i{{[0-9]+}}* [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_INT_TY]]]* {{.+}}, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_INT_TY]]* {{.+}}, [[S_INT_TY]]** [[VAR_ADDR]],
+
+
+// T_VAR and preparation variables
+// CHECK: [[VEC_ADDR_VAL:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK-64: [[CONV_TVAR:%.+]] = bitcast i64* [[T_VAR_ADDR]] to i32*
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR]],
+
+// firstprivate vec(vec): copy from *_addr into priv1 and then from priv1 into priv2
+// CHECK-DAG: [[VEC_DEST_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8* 
+// CHECK-DAG: [[VEC_SRC:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_VAL]] to i8* 
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VEC_DEST_PRIV]], i8* [[VEC_SRC]], {{.+}})
+
+// firstprivate(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_ADDR_BGN:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]] to
+// CHECK-DAG: [[S_ARR_FIN:%.+]] = icmp{{.+}} [[S_ARR_PRIV_BGN]],
+// CHECK-DAG: [[S_ARR_SRC_COPY:%.+]] = phi{{.+}} [ [[S_ARR_ADDR_BGN]], {{.+}} ], [ [[S_ARR_SRC:%.+]], {{.+}} ]
+// CHECK-DAG: [[S_ARR_DST_COPY:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_DST:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_DST_COPY]], {{.+}} [[S_ARR_SRC_COPY]], {{.+}} [[AGG_TMP1]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP1]])
+// CHECK-DAG: [[S_ARR_DST]] = getelementptr {{.+}} [[S_ARR_DST_COPY]],
+// CHECK-DAG: [[S_ARR_SRC]] = getelementptr {{.+}} [[S_ARR_SRC_COPY]],
+
+// firstprivate(var)
+// CHECK-DAG: [[VAR_ADDR_REF:%.+]] = load{{.+}} [[VAR_ADDR]],
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]], {{.+}} [[VAR_ADDR_REF]], {{.+}} [[AGG_TMP2]])
+// CHECK-DAG: call void @{{.+}}({{.+}} [[AGG_TMP2]])
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV]], [[S_INT_TY]]** [[TMP]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG-32: {{.+}} = {{.+}} [[T_VAR_ADDR]]
+// CHECK-DAG-64: {{.+}} = {{.+}} [[CONV_TVAR]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/teams_distribute_simd_firstprivate_messages.cpp b/test/OpenMP/teams_distribute_simd_firstprivate_messages.cpp
index 324672c14129..bcc0e351c66e 100644
--- a/test/OpenMP/teams_distribute_simd_firstprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_firstprivate_messages.cpp
@@ -144,6 +144,7 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute simd firstprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
 #pragma omp teams distribute simd lastprivate(argc), firstprivate(argc) // OK
   for (i = 0; i < argc; ++i) foo();
diff --git a/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp b/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp
new file mode 100644
index 000000000000..3a18b75cf405
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_lastprivate_codegen.cpp
@@ -0,0 +1,371 @@
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-32
+
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1  -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1  -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a) {}
+  S() : f() {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+  #pragma omp target
+  #pragma omp teams distribute simd lastprivate(t_var, vec, s_arr, s_arr, var, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+int main() {
+  static int svar;
+  volatile double g;
+  volatile double &g1 = g;
+
+  #ifdef LAMBDA
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call{{.*}} void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    static float sfvar;
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i{{[0-9]+}} @__tgt_target_teams(
+    // LAMBDA: call void [[OFFLOADING_FUN:@.+]](
+
+    // LAMBDA: define{{.+}} void [[OFFLOADING_FUN]](
+    // LAMBDA: call {{.*}}void {{.+}} @__kmpc_fork_teams({{.+}}, i32 4, {{.+}}* [[OMP_OUTLINED:@.+]] to {{.+}})
+    #pragma omp target
+    #pragma omp teams distribute simd lastprivate(g, g1, svar, sfvar)
+    for (int i = 0; i < 2; ++i) {
+      // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_OUTLINED]](i32* noalias %{{.+}}, i32* noalias %{{.+}}, double*{{.+}} [[G_IN:%.+]], double*{{.+}} [[G1_IN:%.+]], i{{[0-9]+}}*{{.+}} [[SVAR_IN:%.+]], float*{{.+}} [[SFVAR_IN:%.+]])
+      // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[G1_PRIVATE_ADDR:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR:%.+]] = alloca float*,
+      // loop variables
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: {{.+}} = alloca i{{[0-9]+}},
+      // LAMBDA: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[G_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[G1_PRIVATE:%.+]] = alloca double,
+      // LAMBDA: [[TMP_G1_PRIVATE:%.+]] = alloca double*,
+      // LAMBDA: [[SVAR_PRIVATE:%.+]] = alloca i{{[0-9]+}},
+      // LAMBDA: [[SFVAR_PRIVATE:%.+]] = alloca float,
+      // LAMBDA: store double* [[G_IN]], double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_IN]], double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_IN]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: store float* [[SFVAR_IN]], float** [[SFVAR_PRIVATE_ADDR]],
+
+      // init private variables
+      // LAMBDA: [[G_IN_REF:%.+]] = load double*, double** [[G_PRIVATE_ADDR]],
+      // LAMBDA: [[SVAR_IN_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[SFVAR_IN_REF:%.+]] = load float*, float** [[SFVAR_PRIVATE_ADDR]],
+      // LAMBDA: [[G1_IN_REF:%.+]] = load double*, double** [[G1_PRIVATE_ADDR]],
+      // LAMBDA: store double* [[G1_PRIVATE]], double** [[TMP_G1_PRIVATE]],
+      g = 1;
+      g1 = 1;
+      svar = 3;
+      sfvar = 4.0;
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_init_4(
+      // LAMBDA: store double 1.0{{.+}}, double* [[G_PRIVATE]],
+      // LAMBDA: [[TMP_G1_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store{{.+}} double 1.0{{.+}}, double* [[TMP_G1_REF]],
+      // LAMBDA: store i{{[0-9]+}} 3, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store float 4.0{{.+}}, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: store double* [[G_PRIVATE]], double** [[G_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[TMP_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_PRIVATE_ADDR_FROM_TMP:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: store double* [[G1_PRIVATE_ADDR_FROM_TMP]], double** [[TMP_PRIVATE_ADDR_REF]],
+      // LAMBDA: [[SVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: store i{{[0-9]+}}* [[SVAR_PRIVATE]], i{{[0-9]+}}** [[SVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: [[SFVAR_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+      // LAMBDA: store float* [[SFVAR_PRIVATE]], float** [[SFVAR_PRIVATE_ADDR_REF]]
+      // LAMBDA: call{{.*}} void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]])
+      // LAMBDA: call {{.*}}void @__kmpc_for_static_fini(
+      // LAMBDA: store i32 2, i32* %
+      // LAMBDA: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+      // LAMBDA: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+      // LAMBDA: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+      // LAMBDA: [[OMP_LASTPRIV_BLOCK]]:
+      // LAMBDA: [[G_PRIV_VAL:%.+]] = load double, double* [[G_PRIVATE]],
+      // LAMBDA: store{{.*}} double [[G_PRIV_VAL]], double* [[G_IN_REF]],
+      // LAMBDA: [[TMP_G1_PRIV_REF:%.+]] = load double*, double** [[TMP_G1_PRIVATE]],
+      // LAMBDA: [[TMP_G1_PRIV_VAL:%.+]] = load double, double* [[TMP_G1_PRIV_REF]],
+      // LAMBDA: store{{.*}} double [[TMP_G1_PRIV_VAL]], double* [[G1_IN_REF]],
+
+      // LAMBDA: [[SVAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SVAR_PRIVATE]],
+      // LAMBDA: store i{{[0-9]+}} [[SVAR_PRIV_VAL]], i{{[0-9]+}}* [[SVAR_IN_REF]],
+      // LAMBDA: [[SFVAR_PRIV_VAL:%.+]] = load float, float* [[SFVAR_PRIVATE]],
+      // LAMBDA: store float [[SFVAR_PRIV_VAL]], float* [[SFVAR_IN_REF]],
+      // LAMBDA: br label %[[OMP_LASTPRIV_DONE]]
+      // LAMBDA: [[OMP_LASTPRIV_DONE]]:
+      // LAMBDA: ret
+      [&]() {
+        // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+        // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+        g = 2;
+        g1 = 2;
+        svar = 4;
+        sfvar = 8.0;
+        // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+        // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+        // LAMBDA: [[G_REF:%.+]] = load double*, double** [[G_PTR_REF]]
+        // LAMBDA: store double 2.0{{.+}}, double* [[G_REF]]
+
+        // LAMBDA: [[TMP_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+        // LAMBDA: [[G1_REF:%.+]] = load double*, double** [[TMP_PTR_REF]]
+        // LAMBDA: store double 2.0{{.+}}, double* [[G1_REF]],
+        // LAMBDA: [[SVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+        // LAMBDA: [[SVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_PTR_REF]]
+        // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SVAR_REF]]
+        // LAMBDA: [[SFVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 3
+        // LAMBDA: [[SFVAR_REF:%.+]] = load float*, float** [[SFVAR_PTR_REF]]
+        // LAMBDA: store float 8.0{{.+}}, float* [[SFVAR_REF]]
+      }();
+    }
+  }();
+  return 0;
+  #else
+  S<float> test;
+  int t_var = 0;
+  int vec[] = {1, 2};
+  S<float> s_arr[] = {1, 2};
+  S<float> &var = test;
+
+  #pragma omp target
+  #pragma omp teams distribute simd lastprivate(t_var, vec, s_arr, s_arr, var, var, svar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  int i;
+
+  return tmain<int>();
+  #endif
+}
+
+// CHECK: define{{.*}} i{{[0-9]+}} @main()
+// CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_FLOAT_TY]]]* {{.+}}, [[S_FLOAT_TY]]* {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: ret
+
+// CHECK: define{{.+}} [[OFFLOAD_FUN]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]*{{.+}} {{.+}}, [2 x [[S_FLOAT_TY]]]*{{.+}} {{.+}}, [[S_FLOAT_TY]]*{{.+}} {{.+}}, i{{[0-9]+}} {{.+}})
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(
+// CHECK: ret
+//
+// CHECK: define internal void [[OMP_OUTLINED:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN:%.+]], [2 x [[S_FLOAT_TY]]]*{{.+}} [[S_ARR_IN:%.+]], [[S_FLOAT_TY]]*{{.+}} [[VAR_IN:%.+]], i{{[0-9]+}}*{{.*}} [[S_VAR_IN:%.+]])
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR:%.+]] = alloca [2 x [[S_FLOAT_TY]]]*,
+// CHECK: [[VAR_ADDR:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[SVAR_ADDR:%.+]] = alloca i{{[0-9]+}}*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK: [[TMP_PRIV:%.+]] = alloca [[S_FLOAT_TY]]*,
+// CHECK: [[S_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN]], [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN]], i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: store [2 x [[S_FLOAT_TY]]]* [[S_ARR_IN]], [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: store [[S_FLOAT_TY]]* [[VAR_IN]], [[S_FLOAT_TY]]** [[VAR_ADDR]],
+// CHECK: store i{{[0-9]+}}* [[S_VAR_IN]], i{{[0-9]+}}** [[SVAR_ADDR]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR]],
+// CHECK: [[T_VAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR]],
+// CHECK: [[S_ARR_ADDR_REF:%.+]] = load [2 x [[S_FLOAT_TY]]]*, [2 x [[S_FLOAT_TY]]]** [[S_ARR_ADDR]],
+// CHECK: [[SVAR_ADDR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SVAR_ADDR]],
+// CHECK: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK:  [[VAR_ADDR_REF:%.+]] = load {{.+}}, {{.+}} [[VAR_ADDR]],
+// the distribute loop
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[T_VAR_PRIV_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: [[VEC_PTR:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL]], i{{[0-9]+}}* [[VEC_PTR]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[TMP_VAL:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_PTR]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST]], i8* [[TMP_VAL_BCAST]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]] to [[S_FLOAT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_FLOAT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_FLOAT_TY]], [[S_FLOAT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_FLOAT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_FLOAT_TY]]*, [[S_FLOAT_TY]]** [[TMP_PRIV]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[VAR_ADDR_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_FLOAT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: [[SVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[S_VAR_PRIV]],
+// CHECK: store i{{[0-9]+}} [[SVAR_VAL]], i{{[0-9]+}}* [[SVAR_ADDR_REF]],
+// CHECK: ret void
+
+// template tmain
+// CHECK: define{{.*}} i{{[0-9]+}} [[TMAIN_INT:@.+]]()
+// CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]],
+// CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]])
+// CHECK: call i{{[0-9]+}} @__tgt_target_teams(
+// CHECK: call void [[OFFLOAD_FUN_1:@.+]](i{{[0-9]+}} {{.+}}, [2 x i{{[0-9]+}}]* {{.+}}, [2 x [[S_INT_TY]]]* {{.+}}, [[S_INT_TY]]* {{.+}})
+// CHECK: ret
+
+
+// CHECK: define internal void [[OFFLOAD_FUN_1]](
+// CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 4,
+// CHECK: ret
+
+// CHECK: define internal void [[OMP_OUTLINED_1:@.+]](i{{[0-9]+}}* noalias [[GTID_ADDR1:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [2 x i{{[0-9]+}}]*{{.+}} [[VEC_IN1:%.+]], i{{[0-9]+}}*{{.+}} [[T_VAR_IN1:%.+]], [2 x [[S_INT_TY]]]*{{.+}} [[S_ARR_IN1:%.+]], [[S_INT_TY]]*{{.+}} [[VAR_IN1:%.+]])
+// skip alloca of global_tid and bound_tid
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: {{.+}} = alloca i{{[0-9]+}}*,
+// CHECK: [[VEC_ADDR1:%.+]] = alloca [2 x i{{[0-9]+}}]*,
+// CHECK: [[T_VAR_ADDR1:%.+]] = alloca i{{[0-9]+}}*,
+// CHECK: [[S_ARR_ADDR1:%.+]] = alloca [2 x [[S_INT_TY]]]*,
+// CHECK: [[VAR_ADDR1:%.+]] = alloca [[S_INT_TY]]*,
+// skip loop variables
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: {{.+}} = alloca i{{[0-9]+}},
+// CHECK: [[OMP_IS_LAST1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV1:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV1:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV1:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV1:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP_PRIV1:%.+]] = alloca [[S_INT_TY]]*,
+
+// skip init of bound and global tid
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// CHECK: store i{{[0-9]+}}* {{.*}},
+// copy from parameters to local address variables
+// CHECK: store [2 x i{{[0-9]+}}]* [[VEC_IN1]], [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: store i{{[0-9]+}}* [[T_VAR_IN1]], i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: store [2 x [[S_INT_TY]]]* [[S_ARR_IN1]], [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK: store [[S_INT_TY]]* [[VAR_IN1]], [[S_INT_TY]]** [[VAR_ADDR1]],
+
+// load content of local address variables
+// CHECK: [[VEC_ADDR_REF1:%.+]] = load [2 x i{{[0-9]+}}]*, [2 x i{{[0-9]+}}]** [[VEC_ADDR1]],
+// CHECK: [[T_VAR_ADDR_REF1:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[T_VAR_ADDR1]],
+// CHECK: [[S_ARR_ADDR_REF1:%.+]] = load [2 x [[S_INT_TY]]]*, [2 x [[S_INT_TY]]]** [[S_ARR_ADDR1]],
+// CHECK-DAG: store i{{[0-9]+}} 0, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[VAR_ADDR1_REF:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[VAR_ADDR1]],
+// CHECK-DAG: store [[S_INT_TY]]* [[VAR_PRIV1]], [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: call void @__kmpc_for_static_init_4(
+// assignment: vec[i] = t_var;
+// CHECK: [[IV_VAL1:%.+]] =
+// CHECK: [[T_VAR_PRIV_VAL1:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: [[VEC_PTR1:%.+]] = getelementptr inbounds [2 x i{{[0-9]+}}], [2 x i{{[0-9]+}}]* [[VEC_PRIV1]], i{{[0-9]+}} 0, i{{[0-9]+}} {{.+}}
+// CHECK:  store i{{[0-9]+}} [[T_VAR_PRIV_VAL1]], i{{[0-9]+}}* [[VEC_PTR1]],
+
+// assignment: s_arr[i] = var;
+// CHECK-DAG: [[S_ARR_PTR1:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]],
+// CHECK-DAG: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK-DAG: [[S_ARR_PTR_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_PTR1]] to i8*
+// CHECK-DAG: [[TMP_VAL_BCAST1:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK-DAG: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_PTR_BCAST1]], i8* [[TMP_VAL_BCAST1]],
+// CHECK: call void @__kmpc_for_static_fini(
+
+// lastprivates
+// CHECK: [[OMP_IS_LAST_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[OMP_IS_LAST1]],
+// CHECK: [[IS_LAST_IT:%.+]] = icmp ne i{{[0-9]+}} [[OMP_IS_LAST_VAL]], 0
+// CHECK: br i1 [[IS_LAST_IT]], label %[[OMP_LASTPRIV_BLOCK:.+]], label %[[OMP_LASTPRIV_DONE:.+]]
+
+// CHECK: [[OMP_LASTPRIV_BLOCK]]:
+// CHECK: [[T_VAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[T_VAR_PRIV1]],
+// CHECK: store i{{[0-9]+}} [[T_VAR_VAL]], i{{[0-9]+}}* [[T_VAR_ADDR_REF1]],
+// CHECK: [[BCAST_VEC_ADDR_REF:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_ADDR_REF1]] to i8*
+// CHECK: [[BCAST_VEC_PRIV:%.+]] = bitcast [2 x i{{[0-9]+}}]* [[VEC_PRIV1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[BCAST_VEC_ADDR_REF]], i8* [[BCAST_VEC_PRIV]],
+// CHECK: [[S_ARR_BEGIN:%.+]] = getelementptr inbounds [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_ADDR_REF]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+// CHECK: [[S_ARR_PRIV_BCAST:%.+]] = bitcast [2 x [[S_INT_TY]]]* [[S_ARR_PRIV1]] to [[S_INT_TY]]*
+// CHECK: [[S_ARR_BEGIN_GEP:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_BEGIN]], i{{[0-9]+}} 2
+// CHECK: [[S_ARR_IS_EMPTY:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_BEGIN]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[S_ARR_IS_EMPTY]], label %[[S_ARR_COPY_DONE:.+]], label %[[S_ARR_COPY_BLOCK:.+]]
+// CHECK: [[S_ARR_COPY_BLOCK]]:
+// CHECK: [[S_ARR_SRC_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_EL:%.+]] = phi [[S_INT_TY]]*{{.+}}
+// CHECK: [[S_ARR_DST_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_DST_EL]] to i8*
+// CHECK: [[S_ARR_SRC_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[S_ARR_SRC_EL]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[S_ARR_DST_BCAST]], i8* [[S_ARR_SRC_BCAST]]{{.+}})
+// CHECK: [[S_ARR_DST_NEXT:%.+]] = getelementptr [[S_INT_TY]], [[S_INT_TY]]* [[S_ARR_DST_EL]], i{{[0-9]+}} 1
+// CHECK: [[S_ARR_SRC_NEXT:%.+]] = getelementptr{{.+}}
+// CHECK: [[CPY_IS_FINISHED:%.+]] = icmp eq [[S_INT_TY]]* [[S_ARR_DST_NEXT]], [[S_ARR_BEGIN_GEP]]
+// CHECK: br i1 [[CPY_IS_FINISHED]], label %[[S_ARR_COPY_DONE]], label %[[S_ARR_COPY_BLOCK]]
+// CHECK: [[S_ARR_COPY_DONE]]:
+// CHECK: [[TMP_VAL1:%.+]] = load [[S_INT_TY]]*, [[S_INT_TY]]** [[TMP_PRIV1]],
+// CHECK: [[VAR_ADDR_REF_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[VAR_ADDR1_REF]] to i8*
+// CHECK: [[TMP_VAL1_BCAST:%.+]] = bitcast [[S_INT_TY]]* [[TMP_VAL1]] to i8*
+// CHECK: call void @llvm.memcpy.{{.+}}(i8* [[VAR_ADDR_REF_BCAST]], i8* [[TMP_VAL1_BCAST]],{{.+}})
+// CHECK: ret void
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/teams_distribute_simd_lastprivate_messages.cpp b/test/OpenMP/teams_distribute_simd_lastprivate_messages.cpp
index 4c715bf2ce30..51dbfef22933 100644
--- a/test/OpenMP/teams_distribute_simd_lastprivate_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_lastprivate_messages.cpp
@@ -18,14 +18,14 @@ public:
   const S2 &operator =(const S2&) const;
   S2 &operator =(const S2&);
   static float S2s; // expected-note {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note {{static data member is predetermined as shared}}
 };
-const float S2::S2sc = 0; // expected-note {{static data member is predetermined as shared}}
+const float S2::S2sc = 0;
 const S2 b;
 const S2 ba[5];
 class S3 {
   int a;
-  S3 &operator=(const S3 &s3); // expected-note 2 {{implicitly declared private here}}
+  S3 &operator=(const S3 &s3); // expected-note {{implicitly declared private here}}
 
 public:
   S3() : a(0) {}
@@ -52,7 +52,7 @@ public:
 };
 class S6 {
   int a;
-  S6() : a(0) {}
+  S6() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S6(const S6 &s6) : a(s6.a) {}
@@ -255,12 +255,14 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute simd lastprivate(j)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{firstprivate variable cannot be lastprivate}} expected-note@+2 {{defined as firstprivate}}
 #pragma omp target
-#pragma omp teams distribute simd firstprivate(m) lastprivate(m) // expected-error {{'operator=' is a private member of 'S3'}}
+#pragma omp teams distribute simd firstprivate(m) lastprivate(m)
   for (i = 0; i < argc; ++i) foo();
 
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute simd lastprivate(n) firstprivate(n) // OK
+#pragma omp teams distribute simd lastprivate(n) firstprivate(n) // expected-error {{calling a private constructor of class 'S6'}}
   for (i = 0; i < argc; ++i) foo();
 
   static int si;
diff --git a/test/OpenMP/teams_distribute_simd_linear_messages.cpp b/test/OpenMP/teams_distribute_simd_linear_messages.cpp
index 6bfdb16274a5..a53f0e62dda2 100644
--- a/test/OpenMP/teams_distribute_simd_linear_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_linear_messages.cpp
@@ -18,34 +18,42 @@ void test_linear_colons()
 {
   int B = 0;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(B:bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(B::ib:B:bfoo()) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(B:ib) // expected-error {{use of undeclared identifier 'ib'; did you mean 'B::ib'}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(z:B:ib) // expected-error {{unexpected ':' in nested name specifier; did you mean '::'?}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(B:B::bfoo())
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(X::x : ::z)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 3 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(B,::z, X::x)
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(::z)
   for (int i = 0; i < 10; ++i) ;
@@ -54,6 +62,7 @@ void test_linear_colons()
 #pragma omp teams distribute simd linear(B::bfoo()) // expected-error {{expected variable name}}
   for (int i = 0; i < 10; ++i) ;
 
+// expected-error@+2 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(B::ib,B:C1+C2)
   for (int i = 0; i < 10; ++i) ;
@@ -76,6 +85,7 @@ template<int L, class T, class N> T test_template(T* arr, N num) {
 
 template<int LEN> int test_warn() {
   int ind2 = 0;
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
   #pragma omp target
   #pragma omp teams distribute simd linear(ind2:LEN) // expected-warning {{zero linear step (ind2 should probably be const)}}
   for (int i = 0; i < 100; i++) {
@@ -133,10 +143,12 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
@@ -145,6 +157,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear (argc : 5)
   for (int k = 0; k < argc; ++k) ++k;
@@ -161,6 +174,7 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute simd linear (argv[1]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(e, g)
   for (int k = 0; k < argc; ++k) ++k;
@@ -169,32 +183,11 @@ template<class I, class C> int foomain(I argc, C **argv) {
 #pragma omp teams distribute simd linear(h) // expected-error {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear(i)
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int v = 0;
-    int i;
-    #pragma omp target
-    #pragma omp teams distribute simd linear(v:i)
-    for (int k = 0; k < argc; ++k) { i = k; v += i; }
-  }
-
-#pragma omp target
-#pragma omp teams distribute simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-  int v = 0;
-
-#pragma omp target
-#pragma omp teams distribute simd linear(v:j)
-  for (int k = 0; k < argc; ++k) { ++k; v += j; }
-
-#pragma omp target
-#pragma omp teams distribute simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
   return 0;
 }
 
@@ -230,10 +223,12 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute simd linear () // expected-error {{expected expression}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear (argc // expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear (argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
   for (int k = 0; k < argc; ++k) ++k;
@@ -242,6 +237,7 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute simd linear (argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name}}
   for (int k = 0; k < argc; ++k) ++k;
 
+// expected-error@+2 {{only loop iteration variables are allowed in 'linear' clause in distribute directives}}
 #pragma omp target
 #pragma omp teams distribute simd linear (argc)
   for (int k = 0; k < argc; ++k) ++k;
@@ -267,26 +263,6 @@ int main(int argc, char **argv) {
 #pragma omp teams distribute simd linear(h, C::x) // expected-error 2 {{threadprivate or thread local variable cannot be linear}}
   for (int k = 0; k < argc; ++k) ++k;
 
-  #pragma omp parallel
-  {
-    int i;
-    #pragma omp target
-    #pragma omp teams distribute simd linear(i)
-      for (int k = 0; k < argc; ++k) ++k;
-
-    #pragma omp target
-    #pragma omp teams distribute simd linear(i : 4)
-      for (int k = 0; k < argc; ++k) { ++k; i += 4; }
-  }
-
-#pragma omp target
-#pragma omp teams distribute simd linear(j)
-  for (int k = 0; k < argc; ++k) ++k;
-
-#pragma omp target
-#pragma omp teams distribute simd linear(i)
-  for (int k = 0; k < argc; ++k) ++k;
-
   foomain<int,char>(argc,argv); // expected-note {{in instantiation of function template specialization 'foomain<int, char>' requested here}}
   return 0;
 }
diff --git a/test/OpenMP/teams_distribute_simd_loop_messages.cpp b/test/OpenMP/teams_distribute_simd_loop_messages.cpp
index 81fbfe860a70..60e07a3db596 100644
--- a/test/OpenMP/teams_distribute_simd_loop_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_loop_messages.cpp
@@ -2,7 +2,7 @@
 
 class S {
   int a;
-  S() : a(0) {}
+  S() : a(0) {} // expected-note {{implicitly declared private here}}
 
 public:
   S(int v) : a(v) {}
@@ -693,8 +693,9 @@ void test_loop_eh() {
 
 void test_loop_firstprivate_lastprivate() {
   S s(4);
+// expected-error@+2 {{lastprivate variable cannot be firstprivate}} expected-note@+2 {{defined as lastprivate}}
 #pragma omp target
-#pragma omp teams distribute simd lastprivate(s) firstprivate(s)
+#pragma omp teams distribute simd lastprivate(s) firstprivate(s) // expected-error {{calling a private constructor of class 'S'}}
   for (int i = 0; i < 16; ++i)
     ;
 }
diff --git a/test/OpenMP/teams_distribute_simd_messages.cpp b/test/OpenMP/teams_distribute_simd_messages.cpp
index 9095caf38787..b00de796b0ed 100644
--- a/test/OpenMP/teams_distribute_simd_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_messages.cpp
@@ -9,6 +9,9 @@ static int pvt;
 #pragma omp teams distribute simd // expected-error {{unexpected OpenMP directive '#pragma omp teams distribute simd'}}
 
 int main(int argc, char **argv) {
+  #pragma omp target
+  #pragma omp teams distribute simd
+  f; // expected-error {{use of undeclared identifier 'f'}}
 #pragma omp target
 #pragma omp teams distribute simd { // expected-warning {{extra tokens at the end of '#pragma omp teams distribute simd' are ignored}}
   for (int i = 0; i < argc; ++i)
diff --git a/test/OpenMP/teams_distribute_simd_private_codegen.cpp b/test/OpenMP/teams_distribute_simd_private_codegen.cpp
new file mode 100644
index 000000000000..ca9a673e2441
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_private_codegen.cpp
@@ -0,0 +1,238 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+struct St {
+  int a, b;
+  St() : a(0), b(0) {}
+  St(const St &st) : a(st.a + st.b), b(0) {}
+  ~St() {}
+};
+
+volatile int g = 1212;
+volatile int &g1 = g;
+
+template <class T>
+struct S {
+  T f;
+  S(T a) : f(a + g) {}
+  S() : f(g) {}
+  S(const S &s, St t = St()) : f(s.f + t.a) {}
+  operator T() { return T(); }
+  ~S() {}
+};
+
+// CHECK-DAG: [[S_FLOAT_TY:%.+]] = type { float }
+// CHECK-DAG: [[S_INT_TY:%.+]] = type { i{{[0-9]+}} }
+
+template <typename T>
+T tmain() {
+  S<T> test;
+  T t_var = T();
+  T vec[] = {1, 2};
+  S<T> s_arr[] = {1, 2};
+  S<T> &var = test;
+#pragma omp target
+#pragma omp teams distribute simd private(t_var, vec, s_arr, var)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+  }
+  return T();
+}
+
+// CHECK-DAG: [[TEST:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> test;
+// CHECK-DAG: [[T_VAR:@.+]] = global i{{[0-9]+}} 333,
+int t_var = 333;
+// CHECK-DAG: [[VEC:@.+]] = global [2 x i{{[0-9]+}}] [i{{[0-9]+}} 1, i{{[0-9]+}} 2],
+int vec[] = {1, 2};
+// CHECK-DAG: [[S_ARR:@.+]] = global [2 x [[S_FLOAT_TY]]] zeroinitializer,
+S<float> s_arr[] = {1, 2};
+// CHECK-DAG: [[VAR:@.+]] = global [[S_FLOAT_TY]] zeroinitializer,
+S<float> var(3);
+// CHECK-DAG: [[SIVAR:@.+]] = internal global i{{[0-9]+}} 0,
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212,
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute simd private(g, g1, sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} {{%.+}})
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[LOUTL1:.+]] to {{.+}})
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}})
+    // Skip global, bound tid and loop vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: alloca i32,
+    // LAMBDA: [[G_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: [[G1_PRIV:%.+]] = alloca i{{[0-9]+}}
+    // LAMBDA: [[TMP:%.+]] = alloca i{{[0-9]+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+    // LAMBDA: store{{.+}} [[G1_PRIV]], {{.+}} [[TMP]],
+    
+    g = 1;
+    g1 = 1;
+    sivar = 2;
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G_PRIV]],
+    // LAMBDA-DAG: store{{.+}} 2, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[G1_REF:%.+]] = load{{.+}}, {{.+}} [[TMP]],
+    // LAMBDA-DAG: store{{.+}} 1, {{.+}} [[G1_REF]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+      g = 2;
+      g1 = 2;
+      sivar = 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]]
+      // LAMBDA: [[G1_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 1
+      // LAMBDA: [[G1_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G1_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G1_REF]]
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 2
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: store i{{[0-9]+}} 4, i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute simd private(t_var, vec, s_arr, var, sivar)
+  for (int i = 0; i < 2; ++i) {
+    vec[i] = t_var;
+    s_arr[i] = var;
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0, i8** null, i8** null, i{{64|32}}* null, i64* null, i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]]()
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[OUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK: {{.+}} = alloca i32,
+// CHECK-DAG: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK-DAG: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK-DAG: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_FLOAT_TY]]],
+// CHECK-DAG: [[VAR_PRIV:%.+]] = alloca [[S_FLOAT_TY]],
+// CHECK-DAG: [[SIVAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_FLOAT_TY]]], [2 x [[S_FLOAT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[SIVAR_PRIV]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 0,
+// CHECK: call void @[[TOFFL1:.+]]()
+// CHECK:  ret
+
+// CHECK: define {{.*}}void @[[TOFFL1]]()
+// CHECK: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 0, {{.+}} @[[TOUTL1:.+]] to {{.+}})
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}})
+// Skip global, bound tid and loop vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: alloca i{{[0-9]+}},
+// CHECK: [[T_VAR_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[VEC_PRIV:%.+]] = alloca [2 x i{{[0-9]+}}],
+// CHECK: [[S_ARR_PRIV:%.+]] = alloca [2 x [[S_INT_TY]]],
+// CHECK: [[VAR_PRIV:%.+]] = alloca [[S_INT_TY]],
+// CHECK: [[TMP:%.+]] = alloca [[S_INT_TY]]*,
+
+// private(s_arr)
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// CHECK-DAG: [[S_ARR_PRIV_BGN:%.+]] = getelementptr{{.*}} [2 x [[S_INT_TY]]], [2 x [[S_INT_TY]]]* [[S_ARR_PRIV]],
+// CHECK-DAG: [[S_ARR_PTR_ALLOC:%.+]] = phi{{.+}} [ [[S_ARR_PRIV_BGN]], {{.+}} ], [ [[S_ARR_NEXT:%.+]], {{.+}} ]
+// CHECK-DAG: call void @{{.+}}({{.+}} [[S_ARR_PTR_ALLOC]])
+// CHECK-DAG: [[S_ARR_NEXT]] = getelementptr {{.+}} [[S_ARR_PTR_ALLOC]],
+
+// private(var)
+// CHECK-DAG: call void @{{.+}}({{.+}} [[VAR_PRIV]])
+// CHECK-DAG: store{{.+}} [[VAR_PRIV]], {{.+}} [[TMP]]
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK-DAG: {{.+}} = {{.+}} [[T_VAR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[VEC_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[S_ARR_PRIV]]
+// CHECK-DAG: {{.+}} = {{.+}} [[TMP]]
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: ret void
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/teams_distribute_simd_reduction_codegen.cpp b/test/OpenMP/teams_distribute_simd_reduction_codegen.cpp
new file mode 100644
index 000000000000..3153cce3a923
--- /dev/null
+++ b/test/OpenMP/teams_distribute_simd_reduction_codegen.cpp
@@ -0,0 +1,218 @@
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
+// RUN: %clang_cc1 -DCHECK -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DCHECK -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
+
+// RUN: %clang_cc1 -DLAMBDA -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
+// RUN: %clang_cc1 -DLAMBDA -fopenmp -x c++  -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix LAMBDA --check-prefix LAMBDA-64
+
+// expected-no-diagnostics
+#ifndef HEADER
+#define HEADER
+
+template <typename T>
+T tmain() {
+  T t_var = T();
+  T vec[] = {1, 2};
+#pragma omp target
+#pragma omp teams distribute simd reduction(+: t_var)
+  for (int i = 0; i < 2; ++i) {
+    t_var += (T) i;
+  }
+  return T();
+}
+
+int main() {
+  static int sivar;
+#ifdef LAMBDA
+  // LAMBDA: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+  // LAMBDA-LABEL: @main
+  // LAMBDA: call void [[OUTER_LAMBDA:@.+]](
+  [&]() {
+    // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]](
+    // LAMBDA: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+    // LAMBDA: call void @[[LOFFL1:.+]](
+    // LAMBDA:  ret
+#pragma omp target
+#pragma omp teams distribute simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    // LAMBDA: define{{.*}} internal{{.*}} void @[[LOFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+    // LAMBDA: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[LOUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+    // LAMBDA: ret void
+
+    // LAMBDA: define internal void @[[LOUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+    // Skip global and bound tid vars
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: {{.+}} = alloca i32*,
+    // LAMBDA: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+    // LAMBDA: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+    // LAMBDA: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+    // LAMBDA: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+    // LAMBDA: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+    // LAMBDA: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+    
+    // LAMBDA: call void @__kmpc_for_static_init_4(
+    // LAMBDA: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA: call void [[INNER_LAMBDA:@.+]](
+    // LAMBDA: call void @__kmpc_for_static_fini(
+    // LAMBDA: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+    // LAMBDA: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+    // LAMBDA: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+    // LAMBDA: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+    // LAMBDA: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+    // LAMBDA: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+    // LAMBDA: {{.+}}, label %[[CASE1:.+]]
+    // LAMBDA: {{.+}}, label %[[CASE2:.+]]
+    // LAMBDA: ]
+    // LAMBDA: [[CASE1]]:
+    // LAMBDA-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+    // LAMBDA: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    // LAMBDA: [[CASE2]]:
+    // LAMBDA-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+    // LAMBDA-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+    // LAMBDA: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+    // LAMBDA: br
+    
+    sivar += i;
+
+    [&]() {
+      // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]])
+      // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]],
+
+      sivar += 4;
+      // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]]
+
+      // LAMBDA: [[SIVAR_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0
+      // LAMBDA: [[SIVAR_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[SIVAR_PTR_REF]]
+      // LAMBDA: [[SIVAR_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[SIVAR_REF]]
+      // LAMBDA: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], 4
+      // LAMBDA: store i{{[0-9]+}} [[SIVAR_INC]], i{{[0-9]+}}* [[SIVAR_REF]]
+    }();
+  }
+  }();
+  return 0;
+#else
+#pragma omp target
+#pragma omp teams distribute simd reduction(+: sivar)
+  for (int i = 0; i < 2; ++i) {
+    sivar += i;
+  }
+  return tmain<int>();
+#endif
+}
+
+// CHECK: [[RED_VAR:@.+]] = common global [8 x {{.+}}] zeroinitializer
+
+// CHECK: define {{.*}}i{{[0-9]+}} @main()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1, i8** %{{[^,]+}}, i8** %{{[^,]+}}, i{{64|32}}* {{.+}}@{{[^,]+}}, i32 0, i32 0), i64* {{.+}}@{{[^,]+}}, i32 0, i32 0), i32 0, i32 0)
+// CHECK: call void @[[OFFL1:.+]](i{{64|32}} %{{.+}})
+// CHECK: {{%.+}} = call{{.*}} i32 @[[TMAIN_INT:.+]]()
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[OFFL1]](i{{64|32}} [[SIVAR_ARG:%.+]])
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK-64: [[SIVAR_CONV:%.+]] = bitcast{{.+}} [[SIVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[OUTL1:.+]] to {{.+}}, {{.+}} [[SIVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[OUTL1]]({{.+}}, {{.+}}, {{.+}} [[SIVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[SIVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[SIVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[SIVAR_ARG]], {{.+}} [[SIVAR_ADDR]],
+// CHECK: [[SIVAR_REF:%.+]] = load{{.+}}, {{.+}} [[SIVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[SIVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[SIVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[SIVAR_PRIV]] to
+// CHECK: store{{.+}} [[SIVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[SIVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_REF]],
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[SIVAR_INC:%.+]] = add{{.+}} [[SIVAR_VAL]], [[SIVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[SIVAR_INC]], {{.+}} [[SIVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[SIVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[SIVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[SIVAR_REF]], {{.+}} [[SIVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+
+// CHECK: define{{.*}} i{{[0-9]+}} @[[TMAIN_INT]]()
+// CHECK: call i32 @__tgt_target_teams(i64 -1, i8* @{{[^,]+}}, i32 1,
+// CHECK: call void @[[TOFFL1:.+]]({{.+}})
+// CHECK:  ret
+
+// CHECK: define{{.*}} void @[[TOFFL1]](i{{64|32}} [[TVAR_ARG:%.+]])
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}},
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK-64: [[TVAR_CONV:%.+]] = bitcast{{.+}} [[TVAR_ADDR]] to
+// CHECK-64: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_CONV]])
+// CHECK-32: call void {{.+}} @__kmpc_fork_teams({{.+}}, i32 1, {{.+}} @[[TOUTL1:.+]] to {{.+}}, {{.+}} [[TVAR_ADDR]])
+// CHECK: ret void
+
+// CHECK: define internal void @[[TOUTL1]]({{.+}}, {{.+}}, {{.+}} [[TVAR_ARG:%.+]])
+// Skip global and bound tid vars
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: {{.+}} = alloca i32*,
+// CHECK: [[TVAR_ADDR:%.+]] = alloca i{{.+}}*,
+// CHECK: [[TVAR_PRIV:%.+]] = alloca i{{.+}},
+// CHECK: [[RED_LIST:%.+]] = alloca [1 x {{.+}}],
+// CHECK: store{{.+}} [[TVAR_ARG]], {{.+}} [[TVAR_ADDR]],
+// CHECK: [[TVAR_REF:%.+]] = load{{.+}}, {{.+}} [[TVAR_ADDR]]
+// CHECK: store{{.+}} 0, {{.+}} [[TVAR_PRIV]],
+
+// CHECK: call void @__kmpc_for_static_init_4(
+// CHECK: store{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: [[RED_LIST_GEP:%.+]] = getelementptr{{.+}} [[RED_LIST]],
+// CHECK: [[TVAR_PRIV_CAST:%.+]] = bitcast{{.+}} [[TVAR_PRIV]] to
+// CHECK: store{{.+}} [[TVAR_PRIV_CAST]], {{.+}} [[RED_LIST_GEP]],
+// CHECK: [[RED_LIST_BCAST:%.+]] = bitcast{{.+}} [[RED_LIST]] to
+// CHECK: [[K_RED_RET:%.+]] = call{{.+}} @__kmpc_reduce({{.+}}, {{.+}}, {{.+}}, {{.+}}, {{.+}} [[RED_LIST_BCAST]], {{.+}} [[RED_FUN:@.+]], {{.+}} [[RED_VAR]])
+// CHECK: switch{{.+}} [[K_RED_RET]], label{{.+}} [
+// CHECK: {{.+}}, label %[[CASE1:.+]]
+// CHECK: {{.+}}, label %[[CASE2:.+]]
+// CHECK: ]
+// CHECK: [[CASE1]]:
+// CHECK-DAG: [[TVAR_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_REF]],
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[TVAR_INC:%.+]] = add{{.+}} [[TVAR_VAL]], [[TVAR_PRIV_VAL]]
+// CHECK: store{{.+}} [[TVAR_INC]], {{.+}} [[TVAR_REF]],
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+// CHECK: [[CASE2]]:
+// CHECK-DAG: [[TVAR_PRIV_VAL:%.+]] = load{{.+}}, {{.+}} [[TVAR_PRIV]],
+// CHECK-DAG: [[ATOMIC_RES:%.+]] = atomicrmw add{{.+}} [[TVAR_REF]], {{.+}} [[TVAR_PRIV_VAL]]
+// CHECK: call void @__kmpc_end_reduce({{.+}}, {{.+}}, {{.+}} [[RED_VAR]])
+// CHECK: br
+
+// CHECK: !{!"llvm.loop.vectorize.enable", i1 true}
+#endif
diff --git a/test/OpenMP/teams_distribute_simd_reduction_messages.cpp b/test/OpenMP/teams_distribute_simd_reduction_messages.cpp
index 879ec688dcb5..09d22bd4bd81 100644
--- a/test/OpenMP/teams_distribute_simd_reduction_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_reduction_messages.cpp
@@ -19,9 +19,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/teams_distribute_simd_shared_messages.cpp b/test/OpenMP/teams_distribute_simd_shared_messages.cpp
index cd963e8469a8..6b06f5b6f203 100644
--- a/test/OpenMP/teams_distribute_simd_shared_messages.cpp
+++ b/test/OpenMP/teams_distribute_simd_shared_messages.cpp
@@ -84,7 +84,7 @@ int main(int argc, char **argv) {
   #pragma omp teams distribute simd shared (S1) // expected-error {{'S1' does not refer to a value}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
-  #pragma omp teams distribute simd shared (a, b, c, d, f)
+  #pragma omp teams distribute simd shared (a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   for (int j=0; j<100; j++) foo();
   #pragma omp target
   #pragma omp teams distribute simd shared (argv[1]) // expected-error {{expected variable name}}
diff --git a/test/OpenMP/teams_messages.cpp b/test/OpenMP/teams_messages.cpp
index 3a42a0e8dd6a..642fe73cac75 100644
--- a/test/OpenMP/teams_messages.cpp
+++ b/test/OpenMP/teams_messages.cpp
@@ -7,6 +7,9 @@ void foo() {
 
 int main(int argc, char **argv) {
   #pragma omp target
+  #pragma omp teams
+  f; // expected-error {{use of undeclared identifier 'f'}}
+  #pragma omp target
   #pragma omp teams { // expected-warning {{extra tokens at the end of '#pragma omp teams' are ignored}}
   foo();
   #pragma omp target
diff --git a/test/OpenMP/teams_private_codegen.cpp b/test/OpenMP/teams_private_codegen.cpp
index 1ba010fa5c3c..7ddf39c9b0ea 100644
--- a/test/OpenMP/teams_private_codegen.cpp
+++ b/test/OpenMP/teams_private_codegen.cpp
@@ -98,7 +98,7 @@ int main() {
 
   // lambda and target region in main
   // LAMBDA: define {{.+}} [[OUTER_LAMBDA]]([[CAP_TY]]* {{.+}})
-  // LAMBDA: call void @[[OMP_OFFLOADING:.+]](i{{[0-9]+}}* {{.+}}, i{{[0-9]+}} {{.+}}
+  // LAMBDA: call void @[[OMP_OFFLOADING:.+]]()
 
   // target region in struct constructor
   // LAMBDA: define{{.*}} void [[ST_CONSTR:@.+]]([[SS_TY]]* %this,
@@ -154,7 +154,7 @@ int main() {
 #pragma omp target
 #pragma omp teams private(g, sivar)
   {
-    // LAMBDA: define{{.+}} @[[OMP_OFFLOADING]](i{{[0-9]+}}* {{.+}} [[G_IN:%.+]], i{{[0-9]+}} [[SIVAR_IN:%.+]]
+    // LAMBDA: define{{.+}} @[[OMP_OFFLOADING]]()
     // LAMBDA: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 0, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*)* [[OMP_OUTLINED_1:@.+]] to void
     
     // LAMBDA: define {{.+}} [[OMP_OUTLINED_1]](i{{[0-9]+}}* {{.+}}, i{{[0-9]+}}* {{.+}}
@@ -196,14 +196,13 @@ int main() {
 // CHECK: define{{.*}} i{{[0-9]+}} @main()
 // CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]],
 // CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]])
-// CHECK: [[OFF_IN:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* {{%.+}},
-// CHECK: call void @[[OMP_OFFLOADING:.+]](i{{[0-9]+}} [[OFF_IN]]
+// CHECK: call void @[[OMP_OFFLOADING:.+]]()
 // CHECK: = call{{.*}} i{{.+}} [[TMAIN_INT:@.+]]()
 // CHECK: call void [[S_FLOAT_TY_DESTR:@.+]]([[S_FLOAT_TY]]*
 // CHECK: ret
 
 // target region in main function
-// CHECK: define{{.+}} @[[OMP_OFFLOADING]](i{{[0-9]+}}
+// CHECK: define{{.+}} @[[OMP_OFFLOADING]]()
 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_teams(%{{.+}}* @{{.+}}, i{{[0-9]+}} 0, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*)* [[OMP_OUTLINED:@.+]] to void
 
 // CHECK: define internal void [[OMP_OUTLINED]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}})
diff --git a/test/OpenMP/teams_reduction_messages.cpp b/test/OpenMP/teams_reduction_messages.cpp
index 9974c147b72c..7aef97730d9e 100644
--- a/test/OpenMP/teams_reduction_messages.cpp
+++ b/test/OpenMP/teams_reduction_messages.cpp
@@ -25,9 +25,9 @@ public:
   S2() : a(0) {}
   S2(S2 &s2) : a(s2.a) {}
   static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
-  static const float S2sc;
+  static const float S2sc; // expected-note 2 {{'S2sc' declared here}}
 };
-const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+const float S2::S2sc = 0;
 S2 b;                     // expected-note 3 {{'b' defined here}}
 const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
 class S3 {
diff --git a/test/OpenMP/teams_shared_messages.cpp b/test/OpenMP/teams_shared_messages.cpp
index e14ba1e32f47..6b013d099dce 100644
--- a/test/OpenMP/teams_shared_messages.cpp
+++ b/test/OpenMP/teams_shared_messages.cpp
@@ -84,7 +84,7 @@ int main(int argc, char **argv) {
   #pragma omp teams shared (S1) // expected-error {{'S1' does not refer to a value}}
   foo();
   #pragma omp target
-  #pragma omp teams shared (a, b, c, d, f)
+  #pragma omp teams shared (a, b, c, d, f) // expected-error {{incomplete type 'S1' where a complete type is required}}
   foo();
   #pragma omp target
   #pragma omp teams shared (argv[1]) // expected-error {{expected variable name}}
diff --git a/test/OpenMP/vla_crash.c b/test/OpenMP/vla_crash.c
index 50dcf068707b..0dcba5503e4c 100644
--- a/test/OpenMP/vla_crash.c
+++ b/test/OpenMP/vla_crash.c
@@ -1,5 +1,4 @@
 // RUN: %clang_cc1 -verify -triple powerpc64le-unknown-linux-gnu -fopenmp -x c -emit-llvm %s -o - | FileCheck %s
-// expected-no-diagnostics
 
 int a;
 
@@ -20,3 +19,24 @@ void foo() {
 
 // CHECK: define internal void [[OUTLINED]](i32* {{[^,]+}}, i32* {{[^,]+}}, i64 {{[^,]+}}, i32** {{[^,]+}}, i64 {{[^,]+}}, i32**** {{[^,]+}})
 
+// CHECK-LABEL: bar
+void bar(int n, int *a) {
+  // CHECK: [[N:%.+]] = alloca i32,
+  // CHECK: [[A:%.+]] = alloca i32*,
+  // CHECK: [[P:%.+]] = alloca i32*,
+  // CHECK: @__kmpc_global_thread_num
+  // CHECK: [[BC:%.+]] = bitcast i32** [[A]] to i32*
+  // CHECK: store i32* [[BC]], i32** [[P]],
+  // CHECK: call void @__kmpc_serialized_parallel
+  // CHECK: call void [[OUTLINED:@[^(]+]](i32* %{{[^,]+}}, i32* %{{[^,]+}}, i64 %{{[^,]+}}, i32** [[P]], i32** [[A]])
+  // CHECK: call void @__kmpc_end_serialized_parallel
+  // CHECK: ret void
+  // expected-warning@+1 {{incompatible pointer types initializing 'int (*)[n]' with an expression of type 'int **'}}
+  int(*p)[n] = &a;
+#pragma omp parallel if(0)
+  // expected-warning@+1 {{comparison of distinct pointer types ('int (*)[n]' and 'int **')}}
+  if (p == &a) {
+  }
+}
+
+// CHECK: define internal void [[OUTLINED]](i32* {{[^,]+}}, i32* {{[^,]+}}, i64 {{[^,]+}}, i32** {{[^,]+}}, i32** {{[^,]+}})