[llvm] [AMDGPU] Do not use 64-bit literals with VOP3* encodings (PR #156602)

[Joe Nash via llvm-commits]

https://github.com/Sisyph commented:

It looks like the right thing to do.

Is it true that handling for the 64bit literals is already correct in codegen, so this patch would not change emitted code?

It is a little suspicious that all the affected test cases are vopc. Perhaps there is some additional handling for vop1/2 that is not present for vopc?

https://github.com/llvm/llvm-project/pull/156602