[AMDGPU] Remove Code Object V2

[Pierre-vh]

Summary

Code Object V2 has been deprecated for more than a year now. We can safely remove it from LLVM.

• [clang] Remove support for the -mcode-object-version=2 option.
• [lld] Remove/refactor tests that were still using COV2
• [llvm] Update AMDGPUUsage.rst
  • Code Object V2 docs are left for informational purposes because those code objects may still be supported by the runtime/loaders for a while.
• [AMDGPU] Remove COV2 emission capabilities.
• [AMDGPU] Remove MetadataStreamerYamlV2 which was only used by COV2
• [AMDGPU] Update all tests that were still using COV2
  • They are either deleted or ported directly to code object v4 (as v3 is also planned to be removed soon).

Depends on D145671

[kzhuravl]

llvm/docs/AMDGPUUsage.rst

1387

emission->generation?

2655

emission->generation?

14605

ditto

llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp

749–750

Nice to see this finally go

[yaxunl]

LGTM on clang side.

[arsenm]

Is clover still relying on the cov2 support?

clang/include/clang/Basic/TargetOptions.h

85

I wouldn't remove the enum field, just add a comment that emission is unsupported or something

llvm/test/MC/AMDGPU/hsa-gfx10.s

3

I thought we were still going to be able to read old objects

llvm/tools/llvm-readobj/ELFDumper.cpp

5437–5440 ↗

(On Diff #504984)

This looks like a separate change?

[Pierre-vh]

Address comments, split some change in D146119

llvm/test/MC/AMDGPU/hsa-gfx10.s

3

I think llvm-readobj uses all of the MC/Target infrastructure so if we remove emission, we also remove reading, no?

I'm actually not sure if we plan to let readobj/readelf read COV2 object files, it's an interesting question

llvm/tools/llvm-readobj/ELFDumper.cpp

5437–5440 ↗

(On Diff #504984)

Moved to D146119

[Pierre-vh]

Rebase

[Pierre-vh]

Rebase

[Pierre-vh]

ping