RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-23 13:02:52 +02:00
Code Issues Projects Releases Wiki Activity
2e9a461206
llvm-mirror/test/CodeGen/X86/win64_frame.ll

191 lines
5.3 KiB
LLVM
Raw Normal View History

[X86, Win64] Use a frame pointer if pushf is emitted A frame pointer must be used if stack pointer is modified after the prologue. LLVM will emit pushf/popf if we need to save/restore the FLAGS register, requiring us to have a frame pointer for the function. There is a small twist: this sequence might exist in user code via inline-assembly. For now, conservatively assume that such functions require a frame pointer. For real world justification, please see clang's implementation of __readeflags. This fixes PR25945. llvm-svn: 256456
2015-12-27 07:07:26 +01:00
; RUN: llc < %s -mtriple=x86_64-pc-win32 | FileCheck %s --check-prefix=CHECK --check-prefix=PUSHF
; RUN: llc < %s -mtriple=x86_64-pc-win32 -mattr=+sahf | FileCheck %s --check-prefix=SAHF
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
define i32 @f1(i32 %p1, i32 %p2, i32 %p3, i32 %p4, i32 %p5) "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f1:
; CHECK: movl 48(%rbp), %eax
ret i32 %p5
}
define void @f2(i32 %p, ...) "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f2:
; CHECK: .seh_stackalloc 8
X86: Use 'mov' instead of 'lea' in Win64 SEH prologues when possible 'mov' and 'lea' are equivalent when the displacement applied with 'lea' is zero. However, 'mov' should encode smaller. llvm-svn: 230269
2015-02-23 22:50:27 +01:00
; CHECK: movq %rsp, %rbp
Win64: Stack alignment constraints aren't applied during SET_FPREG Stack realignment occurs after the prolog, not during, for Win64. Because of this, don't factor in the maximum stack alignment when establishing a frame pointer. This fixes PR22572. llvm-svn: 230113
2015-02-21 02:04:47 +01:00
; CHECK: .seh_setframe 5, 0
; CHECK: movq %rdx, 32(%rbp)
; CHECK: leaq 32(%rbp), %rax
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
%ap = alloca i8, align 8
call void @llvm.va_start(i8* %ap)
ret void
}
define i8* @f3() "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f3:
X86: Use 'mov' instead of 'lea' in Win64 SEH prologues when possible 'mov' and 'lea' are equivalent when the displacement applied with 'lea' is zero. However, 'mov' should encode smaller. llvm-svn: 230269
2015-02-23 22:50:27 +01:00
; CHECK: movq %rsp, %rbp
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
; CHECK: .seh_setframe 5, 0
; CHECK: movq 8(%rbp), %rax
%ra = call i8* @llvm.returnaddress(i32 0)
ret i8* %ra
}
define i8* @f4() "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f4:
; CHECK: pushq %rbp
; CHECK: .seh_pushreg 5
; CHECK: subq $304, %rsp
; CHECK: .seh_stackalloc 304
; CHECK: leaq 128(%rsp), %rbp
; CHECK: .seh_setframe 5, 128
; CHECK: .seh_endprologue
; CHECK: movq 184(%rbp), %rax
alloca [300 x i8]
%ra = call i8* @llvm.returnaddress(i32 0)
ret i8* %ra
}
declare void @external(i8*)
define void @f5() "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f5:
; CHECK: subq $336, %rsp
; CHECK: .seh_stackalloc 336
; CHECK: leaq 128(%rsp), %rbp
; CHECK: .seh_setframe 5, 128
; CHECK: leaq -92(%rbp), %rcx
; CHECK: callq external
%a = alloca [300 x i8]
[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 llvm-svn: 230786
2015-02-27 20:29:02 +01:00
%gep = getelementptr [300 x i8], [300 x i8]* %a, i32 0, i32 0
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
call void @external(i8* %gep)
ret void
}
define void @f6(i32 %p, ...) "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f6:
; CHECK: subq $336, %rsp
; CHECK: .seh_stackalloc 336
; CHECK: leaq 128(%rsp), %rbp
; CHECK: .seh_setframe 5, 128
; CHECK: leaq -92(%rbp), %rcx
; CHECK: callq external
%a = alloca [300 x i8]
[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 llvm-svn: 230786
2015-02-27 20:29:02 +01:00
%gep = getelementptr [300 x i8], [300 x i8]* %a, i32 0, i32 0
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
call void @external(i8* %gep)
ret void
}
define i32 @f7(i32 %a, i32 %b, i32 %c, i32 %d, i32 %e) "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f7:
; CHECK: pushq %rbp
; CHECK: .seh_pushreg 5
Win64: Stack alignment constraints aren't applied during SET_FPREG Stack realignment occurs after the prolog, not during, for Win64. Because of this, don't factor in the maximum stack alignment when establishing a frame pointer. This fixes PR22572. llvm-svn: 230113
2015-02-21 02:04:47 +01:00
; CHECK: subq $304, %rsp
; CHECK: .seh_stackalloc 304
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
; CHECK: leaq 128(%rsp), %rbp
; CHECK: .seh_setframe 5, 128
Win64: Stack alignment constraints aren't applied during SET_FPREG Stack realignment occurs after the prolog, not during, for Win64. Because of this, don't factor in the maximum stack alignment when establishing a frame pointer. This fixes PR22572. llvm-svn: 230113
2015-02-21 02:04:47 +01:00
; CHECK: andq $-64, %rsp
; CHECK: movl 224(%rbp), %eax
; CHECK: leaq 176(%rbp), %rsp
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
alloca [300 x i8], align 64
ret i32 %e
}
define i32 @f8(i32 %a, i32 %b, i32 %c, i32 %d, i32 %e) "no-frame-pointer-elim"="true" {
; CHECK-LABEL: f8:
Win64: Stack alignment constraints aren't applied during SET_FPREG Stack realignment occurs after the prolog, not during, for Win64. Because of this, don't factor in the maximum stack alignment when establishing a frame pointer. This fixes PR22572. llvm-svn: 230113
2015-02-21 02:04:47 +01:00
; CHECK: subq $352, %rsp
; CHECK: .seh_stackalloc 352
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
; CHECK: leaq 128(%rsp), %rbp
; CHECK: .seh_setframe 5, 128
%alloca = alloca [300 x i8], align 64
; CHECK: andq $-64, %rsp
; CHECK: movq %rsp, %rbx
alloca i32, i32 %a
; CHECK: movl %ecx, %eax
Revert "[X86] Widen the 'AND' mask if doing so shrinks the encoding size" This reverts commit 245169 which miscompiles MultiSource/Applications/siod from LNT. llvm-svn: 245432
2015-08-19 13:35:10 +02:00
; CHECK: leaq 15(,%rax,4), %rcx
; CHECK: movabsq $34359738352, %rax
; CHECK: andq %rcx, %rax
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
; CHECK: callq __chkstk
; CHECK: subq %rax, %rsp
[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 llvm-svn: 230786
2015-02-27 20:29:02 +01:00
%gep = getelementptr [300 x i8], [300 x i8]* %alloca, i32 0, i32 0
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
call void @external(i8* %gep)
; CHECK: subq $32, %rsp
; CHECK: leaq (%rbx), %rcx
; CHECK: callq external
; CHECK: addq $32, %rsp
ret i32 %e
; CHECK: movl %esi, %eax
Win64: Stack alignment constraints aren't applied during SET_FPREG Stack realignment occurs after the prolog, not during, for Win64. Because of this, don't factor in the maximum stack alignment when establishing a frame pointer. This fixes PR22572. llvm-svn: 230113
2015-02-21 02:04:47 +01:00
; CHECK: leaq 224(%rbp), %rsp
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
}
[X86, Win64] Use a frame pointer if pushf is emitted A frame pointer must be used if stack pointer is modified after the prologue. LLVM will emit pushf/popf if we need to save/restore the FLAGS register, requiring us to have a frame pointer for the function. There is a small twist: this sequence might exist in user code via inline-assembly. For now, conservatively assume that such functions require a frame pointer. For real world justification, please see clang's implementation of __readeflags. This fixes PR25945. llvm-svn: 256456
2015-12-27 07:07:26 +01:00
define i64 @f9() {
entry:
; CHECK-LABEL: f9:
; CHECK: pushq %rbp
; CHECK: .seh_pushreg 5
; CHECK-NEXT: movq %rsp, %rbp
; CHECK: .seh_setframe 5, 0
; CHECK: .seh_endprologue
[X86] Add intrinsics for reading and writing to the flags register LLVM's targets need to know if stack pointer adjustments occur after the prologue. This is needed to correctly determine if the red-zone is appropriate to use or if a frame pointer is required. Normally, LLVM can figure this out very precisely by reasoning about the contents of the MachineFunction. There is an interesting corner case: inline assembly. The vast majority of inline assembly which will perform a push or pop is done so to pair up with pushf or popf as appropriate. Unfortunately, this inline assembly doesn't mark the stack pointer as clobbered because, well, it isn't. The stack pointer is decremented and then immediately incremented. Because of this, LLVM was changed in r256456 to conservatively assume that inline assembly contain a sequence of stack operations. This is unfortunate because the vast majority of inline assembly will not end up manipulating the stack pointer in any way at all. Instead, let's provide a more principled solution: an intrinsic. FWIW, other compilers (MSVC and GCC among them) also provide this functionality as an intrinsic. llvm-svn: 256685
2016-01-01 07:50:01 +01:00
%call = call i64 @llvm.x86.flags.read.u64()
[X86, Win64] Use a frame pointer if pushf is emitted A frame pointer must be used if stack pointer is modified after the prologue. LLVM will emit pushf/popf if we need to save/restore the FLAGS register, requiring us to have a frame pointer for the function. There is a small twist: this sequence might exist in user code via inline-assembly. For now, conservatively assume that such functions require a frame pointer. For real world justification, please see clang's implementation of __readeflags. This fixes PR25945. llvm-svn: 256456
2015-12-27 07:07:26 +01:00
; CHECK-NEXT: pushfq
; CHECK-NEXT: popq %rax
ret i64 %call
; CHECK-NEXT: popq %rbp
; CHECK-NEXT: retq
}
declare i64 @dummy()
define i64 @f10(i64* %foo, i64 %bar, i64 %baz) {
; CHECK-LABEL: f10:
; CHECK: pushq %rbp
; CHECK: .seh_pushreg 5
; CHECK: pushq %rsi
; CHECK: .seh_pushreg 6
; CHECK: pushq %rdi
; CHECK: .seh_pushreg 7
; CHECK: subq $32, %rsp
; CHECK: .seh_stackalloc 32
; CHECK: leaq 32(%rsp), %rbp
; CHECK: .seh_setframe 5, 32
; CHECK: .seh_endprologue
%cx = cmpxchg i64* %foo, i64 %bar, i64 %baz seq_cst seq_cst
; PUSHF: lock cmpxchgq
; PUSHF-NEXT: pushfq
; PUSHF-NEXT: popq %[[REG:.*]]
; SAHF: lock cmpxchgq
; SAHF-NEXT: seto %al
; SAHF-NEXT: lahf
%v = extractvalue { i64, i1 } %cx, 0
%p = extractvalue { i64, i1 } %cx, 1
%call = call i64 @dummy()
; PUSHF: callq dummy
; PUSHF-NEXT: pushq %[[REG]]
; PUSHF-NEXT: popfq
; SAHF: callq dummy
; SAHF-NEXT: pushq
; SAHF: addb $127, %al
; SAHF-NEXT: sahf
; SAHF-NEXT: popq
%sel = select i1 %p, i64 %call, i64 %bar
; CHECK-NEXT: cmovneq
ret i64 %sel
; CHECK-NEXT: addq $32, %rsp
; CHECK-NEXT: popq %rdi
; CHECK-NEXT: popq %rsi
; CHECK-NEXT: popq %rbp
}
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
declare i8* @llvm.returnaddress(i32) nounwind readnone
[X86] Add intrinsics for reading and writing to the flags register LLVM's targets need to know if stack pointer adjustments occur after the prologue. This is needed to correctly determine if the red-zone is appropriate to use or if a frame pointer is required. Normally, LLVM can figure this out very precisely by reasoning about the contents of the MachineFunction. There is an interesting corner case: inline assembly. The vast majority of inline assembly which will perform a push or pop is done so to pair up with pushf or popf as appropriate. Unfortunately, this inline assembly doesn't mark the stack pointer as clobbered because, well, it isn't. The stack pointer is decremented and then immediately incremented. Because of this, LLVM was changed in r256456 to conservatively assume that inline assembly contain a sequence of stack operations. This is unfortunate because the vast majority of inline assembly will not end up manipulating the stack pointer in any way at all. Instead, let's provide a more principled solution: an intrinsic. FWIW, other compilers (MSVC and GCC among them) also provide this functionality as an intrinsic. llvm-svn: 256685
2016-01-01 07:50:01 +01:00
declare i64 @llvm.x86.flags.read.u64()
X86: Emit an ABI compliant prologue and epilogue for Win64 Win64 has specific contraints on what valid prologues and epilogues look like. This constraint is born from the flexibility and descriptiveness of Win64's unwind opcodes. Prologues previously emitted by LLVM could not be represented by the unwind opcodes, preventing operations powered by stack unwinding to successfully work. Differential Revision: http://reviews.llvm.org/D7520 llvm-svn: 228641
2015-02-10 01:57:42 +01:00
declare void @llvm.va_start(i8*) nounwind
Reference in New Issue Copy Permalink