[llvm] [AMDGPU] Fix GCNRewritePartialRegUses pass: vector regclass is selected instead of scalar. (PR #69957)

[Valery Pykhtin via llvm-commits]

vpykhtin wrote:

> I do not know what is expected/common practice, I suspect either approach is valid. I have certainly seen people merge _main_ before, but typically I opt for a rebase and force push.

Thanks, done.

https://github.com/llvm/llvm-project/pull/69957