Vendor import of llvm RELEASE_360/rc2 tag r227651 (effectively, 3.6.0 RC2):vendor/llvm/llvm-release_360-r227651

https://llvm.org/svn/llvm-project/llvm/tags/RELEASE_360/rc2@227651

9 files changed, 156 insertions, 41 deletions

diff --git a/test/CodeGen/R600/basic-loop.ll b/test/CodeGen/R600/basic-loop.ll
index 72737ae273e6..9d0509b38d8a 100644
--- a/test/CodeGen/R600/basic-loop.ll
+++ b/test/CodeGen/R600/basic-loop.ll
@@ -1,4 +1,3 @@
-; XFAIL: *
 ; RUN: llc -O0 -verify-machineinstrs -march=amdgcn -mcpu=SI < %s | FileCheck %s
 
 ; CHECK-LABEL: {{^}}test_loop:
diff --git a/test/CodeGen/R600/ctpop.ll b/test/CodeGen/R600/ctpop.ll
index a47bc876cb96..c64f443ad697 100644
--- a/test/CodeGen/R600/ctpop.ll
+++ b/test/CodeGen/R600/ctpop.ll
@@ -24,8 +24,7 @@ define void @s_ctpop_i32(i32 addrspace(1)* noalias %out, i32 %val) nounwind {
 ; XXX - Why 0 in register?
 ; FUNC-LABEL: {{^}}v_ctpop_i32:
 ; SI: buffer_load_dword [[VAL:v[0-9]+]],
-; SI: v_mov_b32_e32 [[VZERO:v[0-9]+]], 0
-; SI: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], [[VAL]], [[VZERO]]
+; SI: v_bcnt_u32_b32_e64 [[RESULT:v[0-9]+]], [[VAL]], 0
 ; SI: buffer_store_dword [[RESULT]],
 ; SI: s_endpgm
 
@@ -40,8 +39,7 @@ define void @v_ctpop_i32(i32 addrspace(1)* noalias %out, i32 addrspace(1)* noali
 ; FUNC-LABEL: {{^}}v_ctpop_add_chain_i32:
 ; SI: buffer_load_dword [[VAL0:v[0-9]+]],
 ; SI: buffer_load_dword [[VAL1:v[0-9]+]],
-; SI: v_mov_b32_e32 [[VZERO:v[0-9]+]], 0
-; SI: v_bcnt_u32_b32_e32 [[MIDRESULT:v[0-9]+]], [[VAL1]], [[VZERO]]
+; SI: v_bcnt_u32_b32_e64 [[MIDRESULT:v[0-9]+]], [[VAL1]], 0
 ; SI-NEXT: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], [[VAL0]], [[MIDRESULT]]
 ; SI: buffer_store_dword [[RESULT]],
 ; SI: s_endpgm
@@ -73,8 +71,8 @@ define void @v_ctpop_add_sgpr_i32(i32 addrspace(1)* noalias %out, i32 addrspace(
 }
 
 ; FUNC-LABEL: {{^}}v_ctpop_v2i32:
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
 ; SI: s_endpgm
 
 ; EG: BCNT_INT
@@ -87,10 +85,10 @@ define void @v_ctpop_v2i32(<2 x i32> addrspace(1)* noalias %out, <2 x i32> addrs
 }
 
 ; FUNC-LABEL: {{^}}v_ctpop_v4i32:
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
 ; SI: s_endpgm
 
 ; EG: BCNT_INT
@@ -105,14 +103,14 @@ define void @v_ctpop_v4i32(<4 x i32> addrspace(1)* noalias %out, <4 x i32> addrs
 }
 
 ; FUNC-LABEL: {{^}}v_ctpop_v8i32:
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
 ; SI: s_endpgm
 
 ; EG: BCNT_INT
@@ -131,22 +129,22 @@ define void @v_ctpop_v8i32(<8 x i32> addrspace(1)* noalias %out, <8 x i32> addrs
 }
 
 ; FUNC-LABEL: {{^}}v_ctpop_v16i32:
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
-; SI: v_bcnt_u32_b32_e32
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
+; SI: v_bcnt_u32_b32_e64
 ; SI: s_endpgm
 
 ; EG: BCNT_INT
diff --git a/test/CodeGen/R600/ctpop64.ll b/test/CodeGen/R600/ctpop64.ll
index 8dfe571d3477..9758ac96ea9b 100644
--- a/test/CodeGen/R600/ctpop64.ll
+++ b/test/CodeGen/R600/ctpop64.ll
@@ -21,8 +21,7 @@ define void @s_ctpop_i64(i32 addrspace(1)* noalias %out, i64 %val) nounwind {
 
 ; FUNC-LABEL: {{^}}v_ctpop_i64:
 ; SI: buffer_load_dwordx2 v{{\[}}[[LOVAL:[0-9]+]]:[[HIVAL:[0-9]+]]{{\]}},
-; SI: v_mov_b32_e32 [[VZERO:v[0-9]+]], 0
-; SI: v_bcnt_u32_b32_e32 [[MIDRESULT:v[0-9]+]], v[[LOVAL]], [[VZERO]]
+; SI: v_bcnt_u32_b32_e64 [[MIDRESULT:v[0-9]+]], v[[LOVAL]], 0
 ; SI-NEXT: v_bcnt_u32_b32_e32 [[RESULT:v[0-9]+]], v[[HIVAL]], [[MIDRESULT]]
 ; SI: buffer_store_dword [[RESULT]],
 ; SI: s_endpgm
diff --git a/test/CodeGen/R600/ds_read2st64.ll b/test/CodeGen/R600/ds_read2st64.ll
index 24834af20404..efd875e93176 100644
--- a/test/CodeGen/R600/ds_read2st64.ll
+++ b/test/CodeGen/R600/ds_read2st64.ll
@@ -65,8 +65,8 @@ define void @simple_read2st64_f32_max_offset(float addrspace(1)* %out, float add
 
 ; SI-LABEL: @simple_read2st64_f32_over_max_offset
 ; SI-NOT: ds_read2st64_b32
-; SI: ds_read_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:256
 ; SI: v_add_i32_e32 [[BIGADD:v[0-9]+]], 0x10000, {{v[0-9]+}}
+; SI: ds_read_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:256
 ; SI: ds_read_b32 {{v[0-9]+}}, [[BIGADD]]
 ; SI: s_endpgm
 define void @simple_read2st64_f32_over_max_offset(float addrspace(1)* %out, float addrspace(3)* %lds) #0 {
@@ -197,8 +197,8 @@ define void @simple_read2st64_f64_max_offset(double addrspace(1)* %out, double a
 
 ; SI-LABEL: @simple_read2st64_f64_over_max_offset
 ; SI-NOT: ds_read2st64_b64
-; SI: ds_read_b64 {{v\[[0-9]+:[0-9]+\]}}, {{v[0-9]+}} offset:512
 ; SI: v_add_i32_e32 [[BIGADD:v[0-9]+]], 0x10000, {{v[0-9]+}}
+; SI: ds_read_b64 {{v\[[0-9]+:[0-9]+\]}}, {{v[0-9]+}} offset:512
 ; SI: ds_read_b64 {{v\[[0-9]+:[0-9]+\]}}, [[BIGADD]]
 ; SI: s_endpgm
 define void @simple_read2st64_f64_over_max_offset(double addrspace(1)* %out, double addrspace(3)* %lds) #0 {
diff --git a/test/CodeGen/R600/fp_to_sint.ll b/test/CodeGen/R600/fp_to_sint.ll
index 35cfb03d39b4..d76e8a341c6f 100644
--- a/test/CodeGen/R600/fp_to_sint.ll
+++ b/test/CodeGen/R600/fp_to_sint.ll
@@ -1,16 +1,27 @@
 ; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck %s --check-prefix=EG --check-prefix=FUNC
 ; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs < %s | FileCheck %s --check-prefix=SI --check-prefix=FUNC
 
+declare float @llvm.fabs.f32(float) #0
+
 ; FUNC-LABEL: {{^}}fp_to_sint_i32:
 ; EG: FLT_TO_INT {{\** *}}T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
 ; SI: v_cvt_i32_f32_e32
 ; SI: s_endpgm
-define void @fp_to_sint_i32 (i32 addrspace(1)* %out, float %in) {
+define void @fp_to_sint_i32(i32 addrspace(1)* %out, float %in) {
   %conv = fptosi float %in to i32
   store i32 %conv, i32 addrspace(1)* %out
   ret void
 }
 
+; FUNC-LABEL: {{^}}fp_to_sint_i32_fabs:
+; SI: v_cvt_i32_f32_e64 v{{[0-9]+}}, |s{{[0-9]+}}|{{$}}
+define void @fp_to_sint_i32_fabs(i32 addrspace(1)* %out, float %in) {
+  %in.fabs = call float @llvm.fabs.f32(float %in) #0
+  %conv = fptosi float %in.fabs to i32
+  store i32 %conv, i32 addrspace(1)* %out
+  ret void
+}
+
 ; FUNC-LABEL: {{^}}fp_to_sint_v2i32:
 ; EG: FLT_TO_INT {{\** *}}T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
 ; EG: FLT_TO_INT {{\** *}}T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
@@ -214,3 +225,5 @@ define void @fp_to_sint_v4i64(<4 x i64> addrspace(1)* %out, <4 x float> %x) {
   store <4 x i64> %conv, <4 x i64> addrspace(1)* %out
   ret void
 }
+
+attributes #0 = { nounwind readnone }
diff --git a/test/CodeGen/R600/hsa.ll b/test/CodeGen/R600/hsa.ll
index 2e79866362ac..5ce3beaa16c0 100644
--- a/test/CodeGen/R600/hsa.ll
+++ b/test/CodeGen/R600/hsa.ll
@@ -1,6 +1,8 @@
 ; RUN: llc < %s -mtriple=r600--amdhsa -mcpu=kaveri | FileCheck --check-prefix=HSA %s
 
 ; HSA: {{^}}simple:
+; HSA: .section        .hsa.version
+; HSA-NEXT: .ascii  "HSA Code Unit:0.0:AMD:0.1:GFX8.1:0"
 ; Make sure we are setting the ATC bit:
 ; HSA: s_mov_b32 s[[HI:[0-9]]], 0x100f000
 ; HSA: buffer_store_dword v{{[0-9]+}}, s[0:[[HI]]], 0
diff --git a/test/CodeGen/R600/misaligned-load.ll b/test/CodeGen/R600/misaligned-load.ll
new file mode 100644
index 000000000000..6290ca09d502
--- /dev/null
+++ b/test/CodeGen/R600/misaligned-load.ll
@@ -0,0 +1,18 @@
+; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI %s
+
+; SI: @byte_aligned_load64
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: ds_read_u8
+; SI: s_endpgm
+define void @byte_aligned_load64(i64 addrspace(1)* %out, i64 addrspace(3)* %in) {
+entry:
+  %0 = load i64 addrspace(3)* %in, align 1
+  store i64 %0, i64 addrspace(1)* %out
+  ret void
+}
diff --git a/test/CodeGen/R600/scratch-buffer.ll b/test/CodeGen/R600/scratch-buffer.ll
new file mode 100644
index 000000000000..740328a495da
--- /dev/null
+++ b/test/CodeGen/R600/scratch-buffer.ll
@@ -0,0 +1,86 @@
+; RUN: llc -verify-machineinstrs -march=amdgcn -mcpu=SI < %s | FileCheck %s
+
+; When a frame index offset is more than 12-bits, make sure we don't store
+; it in mubuf's offset field.
+
+; Also, make sure we use the same register for storing the scratch buffer addresss
+; for both stores. This register is allocated by the register scavenger, so we
+; should be able to reuse the same regiser for each scratch buffer access.
+
+; CHECK-LABEL: {{^}}legal_offset_fi:
+; CHECK: v_mov_b32_e32 [[OFFSET:v[0-9]+]], 0{{$}}
+; CHECK: buffer_store_dword v{{[0-9]+}}, [[OFFSET]], s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen
+; CHECK: v_mov_b32_e32 [[OFFSET]], 0x8000
+; CHECK: buffer_store_dword v{{[0-9]+}}, [[OFFSET]], s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen{{$}}
+
+define void @legal_offset_fi(i32 addrspace(1)* %out, i32 %cond, i32 %if_offset, i32 %else_offset) {
+entry:
+  %scratch0 = alloca [8192 x i32]
+  %scratch1 = alloca [8192 x i32]
+
+  %scratchptr0 = getelementptr [8192 x i32]* %scratch0, i32 0, i32 0
+  store i32 1, i32* %scratchptr0
+
+  %scratchptr1 = getelementptr [8192 x i32]* %scratch1, i32 0, i32 0
+  store i32 2, i32* %scratchptr1
+
+  %cmp = icmp eq i32 %cond, 0
+  br i1 %cmp, label %if, label %else
+
+if:
+  %if_ptr = getelementptr [8192 x i32]* %scratch0, i32 0, i32 %if_offset
+  %if_value = load i32* %if_ptr
+  br label %done
+
+else:
+  %else_ptr = getelementptr [8192 x i32]* %scratch1, i32 0, i32 %else_offset
+  %else_value = load i32* %else_ptr
+  br label %done
+
+done:
+  %value = phi i32 [%if_value, %if], [%else_value, %else]
+  store i32 %value, i32 addrspace(1)* %out
+  ret void
+
+  ret void
+
+}
+
+; CHECK-LABEL: {{^}}legal_offset_fi_offset
+; CHECK: buffer_store_dword v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen
+; CHECK: v_add_i32_e32 [[OFFSET:v[0-9]+]], 0x8000
+; CHECK: buffer_store_dword v{{[0-9]+}}, [[OFFSET]], s[{{[0-9]+}}:{{[0-9]+}}], s{{[0-9]+}} offen{{$}}
+
+define void @legal_offset_fi_offset(i32 addrspace(1)* %out, i32 %cond, i32 addrspace(1)* %offsets, i32 %if_offset, i32 %else_offset) {
+entry:
+  %scratch0 = alloca [8192 x i32]
+  %scratch1 = alloca [8192 x i32]
+
+  %offset0 = load i32 addrspace(1)* %offsets
+  %scratchptr0 = getelementptr [8192 x i32]* %scratch0, i32 0, i32 %offset0
+  store i32 %offset0, i32* %scratchptr0
+
+  %offsetptr1 = getelementptr i32 addrspace(1)* %offsets, i32 1
+  %offset1 = load i32 addrspace(1)* %offsetptr1
+  %scratchptr1 = getelementptr [8192 x i32]* %scratch1, i32 0, i32 %offset1
+  store i32 %offset1, i32* %scratchptr1
+
+  %cmp = icmp eq i32 %cond, 0
+  br i1 %cmp, label %if, label %else
+
+if:
+  %if_ptr = getelementptr [8192 x i32]* %scratch0, i32 0, i32 %if_offset
+  %if_value = load i32* %if_ptr
+  br label %done
+
+else:
+  %else_ptr = getelementptr [8192 x i32]* %scratch1, i32 0, i32 %else_offset
+  %else_value = load i32* %else_ptr
+  br label %done
+
+done:
+  %value = phi i32 [%if_value, %if], [%else_value, %else]
+  store i32 %value, i32 addrspace(1)* %out
+  ret void
+}
+
diff --git a/test/CodeGen/R600/si-triv-disjoint-mem-access.ll b/test/CodeGen/R600/si-triv-disjoint-mem-access.ll
index b2f4a9ff05e1..f6dcb388248a 100644
--- a/test/CodeGen/R600/si-triv-disjoint-mem-access.ll
+++ b/test/CodeGen/R600/si-triv-disjoint-mem-access.ll
@@ -51,8 +51,8 @@ define void @no_reorder_local_load_volatile_global_store_local_load(i32 addrspac
 
 ; FUNC-LABEL: @no_reorder_barrier_local_load_global_store_local_load
 ; CI: ds_read_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:4
-; CI: buffer_store_dword
 ; CI: ds_read_b32 {{v[0-9]+}}, {{v[0-9]+}} offset:8
+; CI: buffer_store_dword
 define void @no_reorder_barrier_local_load_global_store_local_load(i32 addrspace(1)* %out, i32 addrspace(1)* %gptr) #0 {
   %ptr0 = load i32 addrspace(3)* addrspace(3)* @stored_lds_ptr, align 4