Removes the need for the custom code in parseCustomOperand().
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Can you say a little bit more than "refine" (here and in other patches)? I can see you capitalised Sdwa and introduced a new SDWAOperand class but what's the bigger picture? I guess the win here is being able to remove some code from AMDGPUAsmParser.cpp?
Can you say a little bit more than "refine" (here and in other patches)? I can see you capitalised Sdwa and introduced a new SDWAOperand class but what's the bigger picture? I guess the win here is being able to remove some code from AMDGPUAsmParser.cpp?
Yes, same as with other similar patches this change removes the need for the custom code in parseCustomOperand(). Do you suggest the commit message to say that?
Yes I would like the commit message to have a brief description of the point of the change.
llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp | ||
---|---|---|
9045–9048 | Unrelated to the patch, but AsmParser violates 80 chars per line in many places. Maybe it needs to be just all clang-formatted. |
Unrelated to the patch, but AsmParser violates 80 chars per line in many places. Maybe it needs to be just all clang-formatted.