[llvm] [AMDGPU][InstCombine] Optimize constant shuffle patterns (PR #192246)

[Pankaj Dwivedi via llvm-commits]

https://github.com/PankajDwivedi-25 approved this pull request.

LGTM as it incorporate all the changes intended in my patch along with other wave_shuffle lowering to VALU path instead ds.

Please wait for other reviewers approval too.

https://github.com/llvm/llvm-project/pull/192246