[llvm] [AMDGPU] CodeGen for GFX12 8/16-bit SMEM loads (PR #77633)

[Jay Foad via llvm-commits]

jayfoad wrote:

> > I'm not sure whether `SITargetLowering::widenLoad` needs any update. What optimization is it trying to perform?
> 
> It's trying to turn 8 and 16-bit scalar loads into SMEM eligible 32-bit loads

I've added a TODO there too (a3fc0f9d2b73b2bf6afbf420d204ab30253ebf33).

https://github.com/llvm/llvm-project/pull/77633