Merge llvm 3.5.0 release from ^/vendor/llvm/dist, resolve conflicts, and

preserve our customizations, where necessary.

1 files changed, 51 insertions, 12 deletions

diff --git a/contrib/llvm/lib/Target/X86/X86CallingConv.td b/contrib/llvm/lib/Target/X86/X86CallingConv.td
index a78b5c0a7967..86c01bd64647 100644
--- a/contrib/llvm/lib/Target/X86/X86CallingConv.td
+++ b/contrib/llvm/lib/Target/X86/X86CallingConv.td
@@ -52,7 +52,7 @@ def RetCC_X86Common : CallingConv<[
   // 512-bit vectors are returned in ZMM0 and ZMM1, when they fit. ZMM2 and ZMM3
   // can only be used by ABI non-compliant code. This vector type is only
   // supported while using the AVX-512 target feature.
-  CCIfType<[v16i32, v8i64, v16f32, v8f64],
+  CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64],
             CCAssignToReg<[ZMM0,ZMM1,ZMM2,ZMM3]>>,
 
   // MMX vector types are always returned in MM0. If the target doesn't have
@@ -252,7 +252,7 @@ def CC_X86_64_C : CallingConv<[
                                          YMM4, YMM5, YMM6, YMM7]>>>>,
 
   // The first 8 512-bit vector arguments are passed in ZMM registers.
-  CCIfNotVarArg<CCIfType<[v16i32, v8i64, v16f32, v8f64],
+  CCIfNotVarArg<CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64],
             CCIfSubtarget<"hasAVX512()",
             CCAssignToReg<[ZMM0, ZMM1, ZMM2, ZMM3, ZMM4, ZMM5, ZMM6, ZMM7]>>>>,
 
@@ -357,9 +357,16 @@ def CC_X86_64_WebKit_JS : CallingConv<[
   // Promote i8/i16 arguments to i32.
   CCIfType<[i8, i16], CCPromoteToType<i32>>,
 
-  // Integer/FP values are always stored in stack slots that are 8 bytes in size
-  // and 8-byte aligned.
-  CCIfType<[i32, i64, f32, f64], CCAssignToStack<8, 8>>
+  // Only the first integer argument is passed in register.
+  CCIfType<[i32], CCAssignToReg<[EAX]>>,
+  CCIfType<[i64], CCAssignToReg<[RAX]>>,
+
+  // The remaining integer arguments are passed on the stack. 32bit integer and
+  // floating-point arguments are aligned to 4 byte and stored in 4 byte slots.
+  // 64bit integer and floating-point arguments are aligned to 8 byte and stored
+  // in 8 byte stack slots.
+  CCIfType<[i32, f32], CCAssignToStack<4, 4>>,
+  CCIfType<[i64, f64], CCAssignToStack<8, 8>>
 ]>;
 
 // No explicit register is specified for the AnyReg calling convention. The
@@ -453,18 +460,34 @@ def CC_X86_32_FastCall : CallingConv<[
   CCDelegateTo<CC_X86_32_Common>
 ]>;
 
-def CC_X86_32_ThisCall : CallingConv<[
+def CC_X86_32_ThisCall_Common : CallingConv<[
+  // The first integer argument is passed in ECX
+  CCIfType<[i32], CCAssignToReg<[ECX]>>,
+
+  // Otherwise, same as everything else.
+  CCDelegateTo<CC_X86_32_Common>
+]>;
+
+def CC_X86_32_ThisCall_Mingw : CallingConv<[
+  // Promote i8/i16 arguments to i32.
+  CCIfType<[i8, i16], CCPromoteToType<i32>>,
+
+  CCDelegateTo<CC_X86_32_ThisCall_Common>
+]>;
+
+def CC_X86_32_ThisCall_Win : CallingConv<[
   // Promote i8/i16 arguments to i32.
   CCIfType<[i8, i16], CCPromoteToType<i32>>,
 
   // Pass sret arguments indirectly through stack.
   CCIfSRet<CCAssignToStack<4, 4>>,
 
-  // The first integer argument is passed in ECX
-  CCIfType<[i32], CCAssignToReg<[ECX]>>,
+  CCDelegateTo<CC_X86_32_ThisCall_Common>
+]>;
 
-  // Otherwise, same as everything else.
-  CCDelegateTo<CC_X86_32_Common>
+def CC_X86_32_ThisCall : CallingConv<[
+  CCIfSubtarget<"isTargetCygMing()", CCDelegateTo<CC_X86_32_ThisCall_Mingw>>,
+  CCDelegateTo<CC_X86_32_ThisCall_Win>
 ]>;
 
 def CC_X86_32_FastCC : CallingConv<[
@@ -597,10 +620,26 @@ def CSR_64EHRet : CalleeSavedRegs<(add RAX, RDX, CSR_64)>;
 def CSR_Win64 : CalleeSavedRegs<(add RBX, RBP, RDI, RSI, R12, R13, R14, R15,
                                      (sequence "XMM%u", 6, 15))>;
 
-def CSR_MostRegs_64 : CalleeSavedRegs<(add RBX, RCX, RDX, RSI, RDI, R8, R9, R10,
+// All GPRs - except r11
+def CSR_64_RT_MostRegs : CalleeSavedRegs<(add CSR_64, RAX, RCX, RDX, RSI, RDI,
+                                              R8, R9, R10, RSP)>;
+
+// All registers - except r11
+def CSR_64_RT_AllRegs     : CalleeSavedRegs<(add CSR_64_RT_MostRegs,
+                                                 (sequence "XMM%u", 0, 15))>;
+def CSR_64_RT_AllRegs_AVX : CalleeSavedRegs<(add CSR_64_RT_MostRegs,
+                                                 (sequence "YMM%u", 0, 15))>;
+
+def CSR_64_MostRegs : CalleeSavedRegs<(add RBX, RCX, RDX, RSI, RDI, R8, R9, R10,
                                            R11, R12, R13, R14, R15, RBP,
                                            (sequence "XMM%u", 0, 15))>;
 
+def CSR_64_AllRegs     : CalleeSavedRegs<(add CSR_64_MostRegs, RAX, RSP,
+                                              (sequence "XMM%u", 16, 31))>;
+def CSR_64_AllRegs_AVX : CalleeSavedRegs<(sub (add CSR_64_MostRegs, RAX, RSP,
+                                                   (sequence "YMM%u", 0, 31)),
+                                              (sequence "XMM%u", 0, 15))>;
+
 // Standard C + YMM6-15
 def CSR_Win64_Intel_OCL_BI_AVX : CalleeSavedRegs<(add RBX, RBP, RDI, RSI, R12,
                                                   R13, R14, R15,
@@ -618,6 +657,6 @@ def CSR_64_Intel_OCL_BI       : CalleeSavedRegs<(add CSR_64,
 def CSR_64_Intel_OCL_BI_AVX    : CalleeSavedRegs<(add CSR_64,
                                                   (sequence "YMM%u", 8, 15))>;
 
-def CSR_64_Intel_OCL_BI_AVX512    : CalleeSavedRegs<(add CSR_64,
+def CSR_64_Intel_OCL_BI_AVX512 : CalleeSavedRegs<(add RBX, RDI, RSI, R14, R15,
                                                   (sequence "ZMM%u", 16, 31),
                                                   K4, K5, K6, K7)>;