[llvm] [AArch64] Utilize `XAR` for certain vector rotates (PR #137629)

[David Green via llvm-commits]

https://github.com/davemgreen commented:

I think this is looking OK. Can you add some tests for fixed-width and scalable types? Some of the fixed-width sizes do no have an instruction they can use, unless they start to use SVE instructions.

https://github.com/llvm/llvm-project/pull/137629