From 944061c4e152e9f66ffaaca5905253ba8012a4fa Mon Sep 17 00:00:00 2001
From: Nico Rieck <nico.rieck@gmail.com>
Date: Mon, 29 Jul 2013 13:07:06 +0000
Subject: [PATCH] Proper va_arg/va_copy lowering on win64

Win64 uses CharPtrBuiltinVaList instead of X86_64ABIBuiltinVaList like
other 64-bit targets.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187355 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/Target/X86/X86ISelLowering.cpp |  4 +-
 test/CodeGen/X86/win64_vararg.ll   | 60 ++++++++++++++++++++++++++++++
 2 files changed, 63 insertions(+), 1 deletion(-)

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index ad2d30891ed..04287c3e175 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -582,10 +582,12 @@ void X86TargetLowering::resetOperationActions() {
   // VASTART needs to be custom lowered to use the VarArgsFrameIndex
   setOperationAction(ISD::VASTART           , MVT::Other, Custom);
   setOperationAction(ISD::VAEND             , MVT::Other, Expand);
-  if (Subtarget->is64Bit()) {
+  if (Subtarget->is64Bit() && !Subtarget->isTargetWin64()) {
+    // TargetInfo::X86_64ABIBuiltinVaList
     setOperationAction(ISD::VAARG           , MVT::Other, Custom);
     setOperationAction(ISD::VACOPY          , MVT::Other, Custom);
   } else {
+    // TargetInfo::CharPtrBuiltinVaList
     setOperationAction(ISD::VAARG           , MVT::Other, Expand);
     setOperationAction(ISD::VACOPY          , MVT::Other, Expand);
   }
diff --git a/test/CodeGen/X86/win64_vararg.ll b/test/CodeGen/X86/win64_vararg.ll
index 74b7cead1ef..1a51b2a64a7 100644
--- a/test/CodeGen/X86/win64_vararg.ll
+++ b/test/CodeGen/X86/win64_vararg.ll
@@ -18,6 +18,7 @@ entry:
 }
 
 declare void @llvm.va_start(i8*) nounwind
+declare void @llvm.va_copy(i8*, i8*) nounwind
 
 ; CHECK-LABEL: f5:
 ; CHECK: pushq
@@ -51,3 +52,62 @@ entry:
   call void @llvm.va_start(i8* %ap1)
   ret i8* %ap1
 }
+
+; WinX86_64 uses char* for va_list. Verify that the correct amount of bytes
+; are copied using va_copy.
+
+; CHECK-LABEL: copy1:
+; CHECK: subq $16
+; CHECK: leaq 32(%rsp), [[REG_copy1:%[a-z]+]]
+; CHECK: movq [[REG_copy1]], 8(%rsp)
+; CHECK: movq [[REG_copy1]], (%rsp)
+; CHECK: addq $16
+; CHECK: ret
+define void @copy1(i64 %a0, ...) nounwind {
+entry:
+  %ap = alloca i8*, align 8
+  %cp = alloca i8*, align 8
+  %ap1 = bitcast i8** %ap to i8*
+  %cp1 = bitcast i8** %cp to i8*
+  call void @llvm.va_start(i8* %ap1)
+  call void @llvm.va_copy(i8* %cp1, i8* %ap1)
+  ret void
+}
+
+; CHECK-LABEL: copy4:
+; CHECK: subq $16
+; CHECK: leaq 56(%rsp), [[REG_copy4:%[a-z]+]]
+; CHECK: movq [[REG_copy4]], 8(%rsp)
+; CHECK: movq [[REG_copy4]], (%rsp)
+; CHECK: addq $16
+; CHECK: ret
+define void @copy4(i64 %a0, i64 %a1, i64 %a2, i64 %a3, ...) nounwind {
+entry:
+  %ap = alloca i8*, align 8
+  %cp = alloca i8*, align 8
+  %ap1 = bitcast i8** %ap to i8*
+  %cp1 = bitcast i8** %cp to i8*
+  call void @llvm.va_start(i8* %ap1)
+  call void @llvm.va_copy(i8* %cp1, i8* %ap1)
+  ret void
+}
+
+; CHECK-LABEL: arg4:
+; CHECK: pushq
+; va_start:
+; CHECK: leaq 48(%rsp), [[REG_arg4_1:%[a-z]+]]
+; CHECK: movq [[REG_arg4_1]], (%rsp)
+; va_arg:
+; CHECK: leaq 52(%rsp), [[REG_arg4_2:%[a-z]+]]
+; CHECK: movq [[REG_arg4_2]], (%rsp)
+; CHECK: movl 48(%rsp), %eax
+; CHECK: popq
+; CHECK: ret
+define i32 @arg4(i64 %a0, i64 %a1, i64 %a2, i64 %a3, ...) nounwind {
+entry:
+  %ap = alloca i8*, align 8
+  %ap1 = bitcast i8** %ap to i8*
+  call void @llvm.va_start(i8* %ap1)
+  %tmp = va_arg i8** %ap, i32
+  ret i32 %tmp
+}