[llvm] [AMDGPU] Select 64-bit moves (PR #70395)

[Jay Foad via llvm-commits]

https://github.com/jayfoad approved this pull request.

> [AMDGPU] Select 64-bit moves
> 
> This allows folding of 64-bit operands if fit into 32-bit. Fixes #67781

Description could be improved: "Select 64-bit immediate moves", "... of 64-bit immediate operands if they can be encoded as 32-bit literals"?

Anyway LGTM, thanks.

https://github.com/llvm/llvm-project/pull/70395