[llvm] [X86] Fold XOR of two vgf2p8affineqb instructions with same input (PR #179900)

[via llvm-commits]

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

----------------
bala-bhargav wrote:

Are you referring to PR #179103? That work was previously implemented for the vgf2p8affineqb instruction and the splat constant. Now, we are extending it to handle the two vgf2p8affineqb instructions.

https://github.com/llvm/llvm-project/pull/179900