[llvm] [X86] Use RORX over SHR imm (PR #77964)

[Bryce Wilson via llvm-commits]

Bryce-MW wrote:

I think that I am going to close this for now. Doing this during instruction selection is nice because of load folding at least but it makes it harder to find situations where the optimization can be done, especially if I started adding other similar transformations. I think what I'd like to do is instead work on doing things like this in the copy flags lowering stage.

https://github.com/llvm/llvm-project/pull/77964