projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: a23cc72) | patch
[X86] Use OR32mi8Locked instead of LOCK_OR32mi8 in emitLockedStackOp.
authorCraig Topper <craig.topper@intel.com>
Wed, 15 May 2019 04:15:46 +0000 (04:15 +0000)
committerCraig Topper <craig.topper@intel.com>
Wed, 15 May 2019 04:15:46 +0000 (04:15 +0000)
commit384d46c0d5be4cece38d8f7448ff80e49c1ce3b8
tree7d862c460e54cabc2f449d1d17ee0c5864e9c18atree | snapshot
parenta23cc727d8155d390a04f4ccf989eda561c9fc06commit | diff
[X86] Use OR32mi8Locked instead of LOCK_OR32mi8 in emitLockedStackOp.

They encode the same way, but OR32mi8Locked sets hasUnmodeledSideEffects set
which should be stronger than the mayLoad/mayStore on LOCK_OR32mi8. I think
this makes sense since we are using it as a fence.

This also seems to hide the operation from the speculative load hardening pass
so I've reverted r360511.

llvm-svn: 360747
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/lib/Target/X86/X86SpeculativeLoadHardening.cpp diff | blob | history
llvm/test/CodeGen/X86/speculative-load-hardening.ll diff | blob | history
Domain: System / Toolchain;