[llvm] [X86] Fold shift into GF2P8AFFINEQB instruction (PR #180019)

[Simon Pilgrim via llvm-commits]

================

----------------
RKSimon wrote:

we normally wait for legalisation to turn intrinsics into X86ISD::GF2P8AFFINEQB nodes - why do you need to do this earlier?

https://github.com/llvm/llvm-project/pull/180019