[PATCH] D99429: [AMDGPU] Save WWM registers in functions

Sebastian Neubauer via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Mar 29 03:30:21 PDT 2021


sebastian-ne updated this revision to Diff 333805.
sebastian-ne marked 2 inline comments as done.
sebastian-ne added a comment.

> I'm not sure what this is setting (I'm also not a huge fan of stepping the iterator back to find the inserted instruction since it spreads the assumption that this only inserts one instruction)

Good point, I changed that to be an additional parameter of storeRegToStackSlot and loadRegFromStackSlot.

Change the SmallVector for WWMReservedRegs to a SmallDenseMap. Thanks for the suggestion.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D99429/new/

https://reviews.llvm.org/D99429

Files:
  llvm/lib/Target/AMDGPU/SIFrameLowering.cpp
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
  llvm/lib/Target/AMDGPU/SIInstrInfo.h
  llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h
  llvm/lib/Target/AMDGPU/SIPreAllocateWWMRegs.cpp
  llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
  llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D99429.333805.patch
Type: text/x-patch
Size: 57506 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210329/b43da8ae/attachment.bin>


More information about the llvm-commits mailing list