[llvm] [AMDGPU] Optimize away v_readfirstlane_b32 on SGPR input (PR #151033)

[Josh Hutton via llvm-commits]

JoshHuttonCode wrote:

If this does not need to be solved, then should #125950 be marked as resolved without this PR?

I can include the IR test as described in the issue.

https://github.com/llvm/llvm-project/pull/151033