This is an archive of the discontinued LLVM Phabricator instance.

[AMDGPU][MC][GFX11] AsmParser for op_sel for VOP3 dpp opcodes
ClosedPublic

Authored by Petar.Avramovic on Jul 14 2022, 6:07 AM.

Details

Summary

Parse op_sel for *_e64_dpp VOP3 opcodes.
Depends on D129637 and setting of VOP3_OPSEL in dpp pseudos.

Diff Detail

Event Timeline

Herald added a project: Restricted Project. · View Herald TranscriptJul 14 2022, 6:07 AM
Petar.Avramovic requested review of this revision.Jul 14 2022, 6:07 AM
Herald added a project: Restricted Project. · View Herald TranscriptJul 14 2022, 6:07 AM
Joe_Nash accepted this revision.Jul 14 2022, 7:33 AM

Overall looks good. Some minor comments.

llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
8029–8031

This is related to the existing behavior of the function, not your change, but this function really needs a documentation comment. Something like "Determines which bit DST_OP_SEL occupies in the op_sel operand according to the number of src operands present, then copies that bit into src0_modifiers"

8044

Weird semi-colon, clang-format please

8046

Move this check to the beginning of the function for early exit.

This revision is now accepted and ready to land.Jul 14 2022, 7:33 AM

Undo clang format.

Joe_Nash accepted this revision.Jul 14 2022, 8:26 AM
dp accepted this revision.Jul 14 2022, 9:12 AM

LGTM.

This revision was landed with ongoing or failed builds.Jul 18 2022, 6:09 AM
This revision was automatically updated to reflect the committed changes.