[clang] [llvm] [AMDGPU] Extend __builtin_amdgcn_ds_bpermute argument types (PR #153501)

[Shilei Tian via cfe-commits]

shiltian wrote:

I think what you proposed can be easily implemented as a function in a header, rather than as a compiler builtin. I also don't think this falls within the scope of a compiler builtin. You can probably just implement the logic in one of the compiler internal headers.

https://github.com/llvm/llvm-project/pull/153501