[llvm-bugs] [Bug 50165] New: PEI clobbers live EFLAGS when using "probe-stack"="inline-asm"
via llvm-bugs
llvm-bugs at lists.llvm.org
Thu Apr 29 01:54:11 PDT 2021
https://bugs.llvm.org/show_bug.cgi?id=50165
Bug ID: 50165
Summary: PEI clobbers live EFLAGS when using
"probe-stack"="inline-asm"
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Backend: X86
Assignee: unassignedbugs at nondot.org
Reporter: tomasz.miasko at gmail.com
CC: craig.topper at gmail.com, llvm-bugs at lists.llvm.org,
llvm-dev at redking.me.uk, pengfei.wang at intel.com,
spatel+llvm at rotateright.com
Created attachment 24813
--> https://bugs.llvm.org/attachment.cgi?id=24813&action=edit
positdiv32 LLVM IR
When building attached llvm IR the eflags used by
liveins: $edi, $eflags, $esi
renamable $r15b = SETCCr 8, implicit $eflags
are clobbered after prologue insertion:
$rsp = frame-setup SUB64ri8 $rsp(tied-def 0), 8, implicit-def dead $eflags
renamable $r15b = SETCCr 8, implicit $eflags
Before Prologue/Epilogue Insertion & Frame Finalization (prologepilog)
bb.4.bb16.i:
; predecessors: %bb.2
successors: %bb.8(0x40000000), %bb.5(0x40000000); %bb.8(50.00%),
%bb.5(50.00%)
liveins: $edi, $eflags, $esi
renamable $r15b = SETCCr 8, implicit $eflags
TEST32rr renamable $esi, renamable $esi, implicit-def $eflags
renamable $r14b = SETCCr 8, implicit $eflags
renamable $edx = COPY renamable $edi, implicit-def $rdx
renamable $edx = NEG32r renamable $edx(tied-def 0), implicit-def $eflags,
implicit killed $rdx, implicit-def $rdx
renamable $edx = CMOV32rr renamable $edx(tied-def 0), killed renamable $edi,
12, implicit $eflags, implicit killed $rdx, implicit-def $rdx
renamable $eax = COPY renamable $esi, implicit-def $rax
renamable $eax = NEG32r renamable $eax(tied-def 0), implicit-def $eflags,
implicit killed $rax, implicit-def $rax
renamable $eax = CMOV32rr renamable $eax(tied-def 0), killed renamable $esi,
12, implicit $eflags, implicit killed $rax, implicit-def $rax
renamable $ecx = LEA64_32r $noreg, 4, renamable $rdx, 0, $noreg, implicit-def
$rcx
TEST32ri renamable $edx, 1073741824, implicit-def $eflags, implicit killed
$rdx
JCC_1 %bb.8, 4, implicit $eflags
JMP_1 %bb.5
After Prologue/Epilogue Insertion & Frame Finalization (prologepilog):
bb.4.bb16.i:
; predecessors: %bb.2
successors: %bb.8(0x40000000), %bb.5(0x40000000); %bb.8(50.00%),
%bb.5(50.00%)
liveins: $edi, $eflags, $esi, $rbp, $r15, $r14, $rbx
frame-setup PUSH64r killed $rbp, implicit-def $rsp, implicit $rsp
frame-setup PUSH64r killed $r15, implicit-def $rsp, implicit $rsp
frame-setup PUSH64r killed $r14, implicit-def $rsp, implicit $rsp
frame-setup PUSH64r killed $rbx, implicit-def $rsp, implicit $rsp
$rsp = frame-setup SUB64ri8 $rsp(tied-def 0), 8, implicit-def dead $eflags
renamable $r15b = SETCCr 8, implicit $eflags
TEST32rr renamable $esi, renamable $esi, implicit-def $eflags
renamable $r14b = SETCCr 8, implicit $eflags
renamable $edx = COPY renamable $edi, implicit-def $rdx
renamable $edx = NEG32r renamable $edx(tied-def 0), implicit-def $eflags,
implicit killed $rdx, implicit-def $rdx
renamable $edx = CMOV32rr renamable $edx(tied-def 0), killed renamable $edi,
12, implicit $eflags, implicit killed $rdx, implicit-def $rdx
renamable $eax = COPY renamable $esi, implicit-def $rax
renamable $eax = NEG32r renamable $eax(tied-def 0), implicit-def $eflags,
implicit killed $rax, implicit-def $rax
renamable $eax = CMOV32rr renamable $eax(tied-def 0), killed renamable $esi,
12, implicit $eflags, implicit killed $rax, implicit-def $rax
renamable $ecx = LEA64_32r $noreg, 4, renamable $rdx, 0, $noreg, implicit-def
$rcx
TEST32ri renamable $edx, 1073741824, implicit-def $eflags, implicit killed
$rdx
JCC_1 %bb.8, 4, implicit $eflags
JMP_1 %bb.5
Originally reported in https://github.com/rust-lang/rust/issues/84667
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20210429/f1d16da8/attachment-0001.html>
More information about the llvm-bugs
mailing list