6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-15 20:29:48 +00:00
Code Issues Projects Releases Wiki Activity
198d8baafb
llvm-6502/test/CodeGen/X86/win32_sret.ll
David Blaikie 198d8baafb [opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 
One of several parallel first steps to remove the target type of pointers,
replacing them with a single opaque pointer type.

This adds an explicit type parameter to the gep instruction so that when the
first parameter becomes an opaque pointer type, the type to gep through is
still available to the instructions.

* This doesn't modify gep operators, only instructions (operators will be
  handled separately)

* Textual IR changes only. Bitcode (including upgrade) and changing the
  in-memory representation will be in separate changes.

* geps of vectors are transformed as:
    getelementptr <4 x float*> %x, ...
  ->getelementptr float, <4 x float*> %x, ...
  Then, once the opaque pointer type is introduced, this will ultimately look
  like:
    getelementptr float, <4 x ptr> %x
  with the unambiguous interpretation that it is a vector of pointers to float.

* address spaces remain on the pointer, not the type:
    getelementptr float addrspace(1)* %x
  ->getelementptr float, float addrspace(1)* %x
  Then, eventually:
    getelementptr float, ptr addrspace(1) %x

Importantly, the massive amount of test case churn has been automated by
same crappy python code. I had to manually update a few test cases that
wouldn't fit the script's model (r228970,r229196,r229197,r229198). The
python script just massages stdin and writes the result to stdout, I
then wrapped that in a shell script to handle replacing files, then
using the usual find+xargs to migrate all the files.

update.py:
import fileinput
import sys
import re

ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")
normrep = re.compile(       r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")

def conv(match, line):
  if not match:
    return line
  line = match.groups()[0]
  if len(match.groups()[5]) == 0:
    line += match.groups()[2]
  line += match.groups()[3]
  line += ", "
  line += match.groups()[1]
  line += "\n"
  return line

for line in sys.stdin:
  if line.find("getelementptr ") == line.find("getelementptr inbounds"):
    if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("):
      line = conv(re.match(ibrep, line), line)
  elif line.find("getelementptr ") != line.find("getelementptr ("):
    line = conv(re.match(normrep, line), line)
  sys.stdout.write(line)

apply.sh:
for name in "$@"
do
  python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name"
  rm -f "$name.tmp"
done

The actual commands:
From llvm/src:
find test/ -name *.ll | xargs ./apply.sh
From llvm/src/tools/clang:
find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}"
From llvm/src/tools/polly:
find test/ -name *.ll | xargs ./apply.sh

After that, check-all (with llvm, clang, clang-tools-extra, lld,
compiler-rt, and polly all checked out).

The extra 'rm' in the apply.sh script is due to a few files in clang's test
suite using interesting unicode stuff that my python script was throwing
exceptions on. None of those files needed to be migrated, so it seemed
sufficient to ignore those cases.

Reviewers: rafael, dexonsmith, grosser

Differential Revision: http://reviews.llvm.org/D7636

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@230786 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-27 19:29:02 +00:00

244 lines
7.1 KiB
LLVM
Raw Blame History

; 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 -mcpu=core2 -mtriple=i686-pc-mingw32 | FileCheck %s -check-prefix=MINGW_X86
; RUN: llc < %s -mcpu=core2 -mtriple=i686-pc-cygwin | FileCheck %s -check-prefix=CYGWIN
; 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 -mcpu=core2 -O0 -mtriple=i686-pc-mingw32 | FileCheck %s -check-prefix=MINGW_X86
; RUN: llc < %s -mcpu=core2 -O0 -mtriple=i686-pc-cygwin | FileCheck %s -check-prefix=CYGWIN
; 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
; arguments are caller-cleanup like normal arguments.
define void @sret1(i8* sret %x) nounwind {
entry:
; WIN32-LABEL: _sret1:
; WIN32: movb $42, (%eax)
; WIN32-NOT: popl %eax
; WIN32: {{retl$}}
; MINGW_X86-LABEL: _sret1:
; MINGW_X86: {{retl$}}
; CYGWIN-LABEL: _sret1:
; CYGWIN: retl $4
; LINUX-LABEL: sret1:
; LINUX: retl $4
store i8 42, i8* %x, align 4
ret void
}
define void @sret2(i8* sret %x, i8 %y) nounwind {
entry:
; WIN32-LABEL: _sret2:
; WIN32: movb {{.*}}, (%eax)
; WIN32-NOT: popl %eax
; WIN32: {{retl$}}
; MINGW_X86-LABEL: _sret2:
; MINGW_X86: {{retl$}}
; CYGWIN-LABEL: _sret2:
; CYGWIN: retl $4
; LINUX-LABEL: sret2:
; LINUX: retl $4
store i8 %y, i8* %x
ret void
}
define void @sret3(i8* sret %x, i8* %y) nounwind {
entry:
; WIN32-LABEL: _sret3:
; WIN32: movb $42, (%eax)
; WIN32-NOT: movb $13, (%eax)
; WIN32-NOT: popl %eax
; WIN32: {{retl$}}
; MINGW_X86-LABEL: _sret3:
; MINGW_X86: {{retl$}}
; CYGWIN-LABEL: _sret3:
; CYGWIN: retl $4
; LINUX-LABEL: sret3:
; LINUX: retl $4
store i8 42, i8* %x
store i8 13, i8* %y
ret void
}
; PR15556
%struct.S4 = type { i32, i32, i32 }
define void @sret4(%struct.S4* noalias sret %agg.result) {
entry:
; WIN32-LABEL: _sret4:
; WIN32: movl $42, (%eax)
; WIN32-NOT: popl %eax
; WIN32: {{retl$}}
; MINGW_X86-LABEL: _sret4:
; MINGW_X86: {{retl$}}
; CYGWIN-LABEL: _sret4:
; CYGWIN: retl $4
; LINUX-LABEL: sret4:
; LINUX: retl $4
%x = getelementptr inbounds %struct.S4, %struct.S4* %agg.result, i32 0, i32 0
store i32 42, i32* %x, align 4
ret void
}
%struct.S5 = type { i32 }
%class.C5 = type { i8 }
define x86_thiscallcc void @"\01?foo@C5@@QAE?AUS5@@XZ"(%struct.S5* noalias sret %agg.result, %class.C5* %this) {
entry:
%this.addr = alloca %class.C5*, align 4
store %class.C5* %this, %class.C5** %this.addr, align 4
%this1 = load %class.C5** %this.addr
%x = getelementptr inbounds %struct.S5, %struct.S5* %agg.result, i32 0, i32 0
store i32 42, i32* %x, align 4
ret void
; WIN32-LABEL: {{^}}"?foo@C5@@QAE?AUS5@@XZ":
; MINGW_X86-LABEL: {{^}}"?foo@C5@@QAE?AUS5@@XZ":
; CYGWIN-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: retl $4
}
define void @call_foo5() {
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-LABEL: {{^}}_call_foo5:
; MINGW_X86-LABEL: {{^}}_call_foo5:
; CYGWIN-LABEL: {{^}}_call_foo5:
; LINUX-LABEL: {{^}}call_foo5:
; Load the address of the result and put it onto stack
; (through %ecx in the -O0 build).
; WIN32: leal {{[0-9]+}}(%esp), %e{{[a-d]}}x
; WIN32: movl %e{{[a-d]}}x, (%e{{([a-d]x)|(sp)}})
; The this pointer goes to ECX.
; WIN32-NEXT: leal {{[0-9]+}}(%esp), %ecx
; WIN32-NEXT: calll "?foo@C5@@QAE?AUS5@@XZ"
; 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:
; CYGWIN-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)}})
; CYGWIN: movl 8(%ebp), %eax
; CYGWIN: 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
; CYGWIN-NEXT: leal 8(%esp), %ecx
; CYGWIN-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*)
; Flipping the parameters at the IR level generates the same code.
%struct.test7 = type { i32, i32, i32 }
define void @test7_f(%struct.test7* %x) nounwind {
; WIN32-LABEL: _test7_f:
; MINGW_X86-LABEL: _test7_f:
; CYGWIN-LABEL: _test7_f:
; LINUX-LABEL: test7_f:
; The %x argument is moved to %ecx on all OSs. It will be the this pointer.
; WIN32: movl 8(%ebp), %ecx
; MINGW_X86: movl 8(%ebp), %ecx
; CYGWIN: movl 8(%ebp), %ecx
; The sret pointer is (%esp)
; WIN32: leal 8(%esp), %[[REG:e[a-d]x]]
; WIN32-NEXT: movl %[[REG]], (%e{{([a-d]x)|(sp)}})
; MINGW_X86: leal 8(%esp), %[[REG:e[a-d]x]]
; MINGW_X86-NEXT: movl %[[REG]], (%e{{([a-d]x)|(sp)}})
; CYGWIN: leal 8(%esp), %[[REG:e[a-d]x]]
; CYGWIN-NEXT: movl %[[REG]], (%e{{([a-d]x)|(sp)}})
%tmp = alloca %struct.test7, align 4
call x86_thiscallcc void @test7_g(%struct.test7* %x, %struct.test7* sret %tmp)
ret void
}
define x86_thiscallcc void @test7_g(%struct.test7* %in, %struct.test7* sret %out) {
%s = getelementptr %struct.test7, %struct.test7* %in, i32 0, i32 0
%d = getelementptr %struct.test7, %struct.test7* %out, i32 0, i32 0
%v = load i32* %s
store i32 %v, i32* %d
call void @clobber_eax()
ret void
; Make sure we return the second parameter in %eax.
; WIN32-LABEL: _test7_g:
; WIN32: calll _clobber_eax
; WIN32: movl {{.*}}, %eax
; WIN32: retl
}
declare void @clobber_eax()
; Test what happens if the first parameter has to be split by codegen.
; Realistically, no frontend will generate code like this, but here it is for
; completeness.
define void @test8_f(i64 inreg %a, i64* sret %out) {
store i64 %a, i64* %out
call void @clobber_eax()
ret void
; WIN32-LABEL: _test8_f:
; WIN32: movl {{[0-9]+}}(%esp), %[[out:[a-z]+]]
; WIN32-DAG: movl %edx, 4(%[[out]])
; WIN32-DAG: movl %eax, (%[[out]])
; WIN32: calll _clobber_eax
; WIN32: movl {{.*}}, %eax
; WIN32: retl
}
Reference in New Issue View Git Blame Copy Permalink