This is an archive of the discontinued LLVM Phabricator instance.

[AMDGPU] Make flat_offset a 32-bit operand instead of 16-bits
ClosedPublic

Authored by matejam on Jan 25 2023, 8:01 AM.

Details

Summary

Change the type of offset in tablegen patterns and use AMDGPU::getNumFlatOffsetBits() in AMDGPUInstPrinter.

Diff Detail

Event Timeline

matejam created this revision.Jan 25 2023, 8:01 AM
Herald added a project: Restricted Project. · View Herald TranscriptJan 25 2023, 8:01 AM
matejam requested review of this revision.Jan 25 2023, 8:01 AM
arsenm accepted this revision.Jan 25 2023, 8:03 AM
This revision is now accepted and ready to land.Jan 25 2023, 8:03 AM
foad accepted this revision.Jan 25 2023, 8:07 AM
foad added inline comments.
llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUInstPrinter.cpp
144

Nit: sink this into the "else" part?

matejam updated this revision to Diff 492134.Jan 25 2023, 8:36 AM

Rebased and moved OffsetSize calculation into else statement.

foad accepted this revision.Jan 25 2023, 8:43 AM

Thanks!

llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUInstPrinter.cpp
144

You can remove the braces too now.

This revision was landed with ongoing or failed builds.Jan 25 2023, 8:52 AM
This revision was automatically updated to reflect the committed changes.