diff options
Diffstat (limited to 'test/CodeGen/X86/win32_sret.ll')
| -rw-r--r-- | test/CodeGen/X86/win32_sret.ll | 98 |
1 files changed, 66 insertions, 32 deletions
diff --git a/test/CodeGen/X86/win32_sret.ll b/test/CodeGen/X86/win32_sret.ll index a24963a..d8ecd44 100644 --- a/test/CodeGen/X86/win32_sret.ll +++ b/test/CodeGen/X86/win32_sret.ll @@ -1,11 +1,11 @@ ; We specify -mcpu explicitly to avoid instruction reordering that happens on ; some setups (e.g., Atom) from affecting the output. ; RUN: llc < %s -mcpu=core2 -mtriple=i686-pc-win32 | FileCheck %s -check-prefix=WIN32 -; RUN: llc < %s -mtriple=i686-pc-mingw32 | FileCheck %s -check-prefix=MINGW_X86 -; RUN: llc < %s -mtriple=i386-pc-linux | FileCheck %s -check-prefix=LINUX +; RUN: llc < %s -mcpu=core2 -mtriple=i686-pc-mingw32 | FileCheck %s -check-prefix=MINGW_X86 +; RUN: llc < %s -mcpu=core2 -mtriple=i386-pc-linux | FileCheck %s -check-prefix=LINUX ; RUN: llc < %s -mcpu=core2 -O0 -mtriple=i686-pc-win32 | FileCheck %s -check-prefix=WIN32 -; RUN: llc < %s -O0 -mtriple=i686-pc-mingw32 | FileCheck %s -check-prefix=MINGW_X86 -; RUN: llc < %s -O0 -mtriple=i386-pc-linux | FileCheck %s -check-prefix=LINUX +; RUN: llc < %s -mcpu=core2 -O0 -mtriple=i686-pc-mingw32 | FileCheck %s -check-prefix=MINGW_X86 +; RUN: llc < %s -mcpu=core2 -O0 -mtriple=i386-pc-linux | FileCheck %s -check-prefix=LINUX ; The SysV ABI used by most Unixes and Mingw on x86 specifies that an sret pointer ; is callee-cleanup. However, in MSVC's cdecl calling convention, sret pointer @@ -13,16 +13,16 @@ define void @sret1(i8* sret %x) nounwind { entry: -; WIN32: sret1 +; WIN32-LABEL: _sret1: ; WIN32: movb $42, (%eax) ; WIN32-NOT: popl %eax -; WIN32: {{ret$}} +; WIN32: {{retl$}} -; MINGW_X86: sret1 -; MINGW_X86: ret $4 +; MINGW_X86-LABEL: _sret1: +; MINGW_X86: {{retl$}} -; LINUX: sret1 -; LINUX: ret $4 +; LINUX-LABEL: sret1: +; LINUX: retl $4 store i8 42, i8* %x, align 4 ret void @@ -30,16 +30,16 @@ entry: define void @sret2(i8* sret %x, i8 %y) nounwind { entry: -; WIN32: sret2 +; WIN32-LABEL: _sret2: ; WIN32: movb {{.*}}, (%eax) ; WIN32-NOT: popl %eax -; WIN32: {{ret$}} +; WIN32: {{retl$}} -; MINGW_X86: sret2 -; MINGW_X86: ret $4 +; MINGW_X86-LABEL: _sret2: +; MINGW_X86: {{retl$}} -; LINUX: sret2 -; LINUX: ret $4 +; LINUX-LABEL: sret2: +; LINUX: retl $4 store i8 %y, i8* %x ret void @@ -47,17 +47,17 @@ entry: define void @sret3(i8* sret %x, i8* %y) nounwind { entry: -; WIN32: sret3 +; WIN32-LABEL: _sret3: ; WIN32: movb $42, (%eax) ; WIN32-NOT: movb $13, (%eax) ; WIN32-NOT: popl %eax -; WIN32: {{ret$}} +; WIN32: {{retl$}} -; MINGW_X86: sret3 -; MINGW_X86: ret $4 +; MINGW_X86-LABEL: _sret3: +; MINGW_X86: {{retl$}} -; LINUX: sret3 -; LINUX: ret $4 +; LINUX-LABEL: sret3: +; LINUX: retl $4 store i8 42, i8* %x store i8 13, i8* %y @@ -69,16 +69,16 @@ entry: define void @sret4(%struct.S4* noalias sret %agg.result) { entry: -; WIN32: sret4 +; WIN32-LABEL: _sret4: ; WIN32: movl $42, (%eax) ; WIN32-NOT: popl %eax -; WIN32: {{ret$}} +; WIN32: {{retl$}} -; MINGW_X86: sret4 -; MINGW_X86: ret $4 +; MINGW_X86-LABEL: _sret4: +; MINGW_X86: {{retl$}} -; LINUX: sret4 -; LINUX: ret $4 +; LINUX-LABEL: sret4: +; LINUX: retl $4 %x = getelementptr inbounds %struct.S4* %agg.result, i32 0, i32 0 store i32 42, i32* %x, align 4 @@ -96,14 +96,16 @@ entry: %x = getelementptr inbounds %struct.S5* %agg.result, i32 0, i32 0 store i32 42, i32* %x, align 4 ret void -; WIN32: {{^}}"?foo@C5@@QAE?AUS5@@XZ": +; WIN32-LABEL: {{^}}"?foo@C5@@QAE?AUS5@@XZ": +; MINGW_X86-LABEL: {{^}}"?foo@C5@@QAE?AUS5@@XZ": +; LINUX-LABEL: {{^}}"?foo@C5@@QAE?AUS5@@XZ": ; The address of the return structure is passed as an implicit parameter. ; In the -O0 build, %eax is spilled at the beginning of the function, hence we ; should match both 4(%esp) and 8(%esp). ; WIN32: {{[48]}}(%esp), %eax ; WIN32: movl $42, (%eax) -; WIN32: ret $4 +; WIN32: retl $4 } define void @call_foo5() { @@ -111,7 +113,10 @@ entry: %c = alloca %class.C5, align 1 %s = alloca %struct.S5, align 4 call x86_thiscallcc void @"\01?foo@C5@@QAE?AUS5@@XZ"(%struct.S5* sret %s, %class.C5* %c) -; WIN32: {{^}}_call_foo5: +; WIN32-LABEL: {{^}}_call_foo5: +; MINGW_X86-LABEL: {{^}}_call_foo5: +; LINUX-LABEL: {{^}}call_foo5: + ; Load the address of the result and put it onto stack ; (through %ecx in the -O0 build). @@ -121,6 +126,35 @@ entry: ; The this pointer goes to ECX. ; WIN32-NEXT: leal {{[0-9]+}}(%esp), %ecx ; WIN32-NEXT: calll "?foo@C5@@QAE?AUS5@@XZ" -; WIN32: ret +; WIN32: retl + ret void +} + + +%struct.test6 = type { i32, i32, i32 } +define void @test6_f(%struct.test6* %x) nounwind { +; WIN32-LABEL: _test6_f: +; MINGW_X86-LABEL: _test6_f: +; LINUX-LABEL: test6_f: + +; The %x argument is moved to %ecx. It will be the this pointer. +; WIN32: movl 8(%ebp), %ecx + +; The %x argument is moved to (%esp). It will be the this pointer. With -O0 +; we copy esp to ecx and use (ecx) instead of (esp). +; MINGW_X86: movl 8(%ebp), %eax +; MINGW_X86: movl %eax, (%e{{([a-d]x)|(sp)}}) + +; The sret pointer is (%esp) +; WIN32: leal 8(%esp), %[[REG:e[a-d]x]] +; WIN32-NEXT: movl %[[REG]], (%e{{([a-d]x)|(sp)}}) + +; The sret pointer is %ecx +; MINGW_X86-NEXT: leal 8(%esp), %ecx +; MINGW_X86-NEXT: calll _test6_g + + %tmp = alloca %struct.test6, align 4 + call x86_thiscallcc void @test6_g(%struct.test6* sret %tmp, %struct.test6* %x) ret void } +declare x86_thiscallcc void @test6_g(%struct.test6* sret, %struct.test6*) |