[llvm] [AMDGPU] Implement moveToVALU for S_CSELECT_B64 (PR #70352)

[Piotr Sobczak via llvm-commits]

piotrAMD wrote:

Looks good as a natural extension of 32-bit select lowering. 

Just wondering if there is still something left to implement - I can see there were more test changes in [D109159](https://reviews.llvm.org/D109159) than here and #/69703. But maybe we already handle these cases.

https://github.com/llvm/llvm-project/pull/70352