[llvm] [X86] Fold nested VGF2P8AFFINEQB instructions (PR #195210)

[via llvm-commits]

WalterKruger wrote:

I failed to update a different regression test, which was improved by this fold. It was a 8-bit `bit_reverse(lshr(x, 1))` which was folded together. I have verified that the folded matrix is correct.

@RKSimon can you please rerun the CI?

https://github.com/llvm/llvm-project/pull/195210