[X86] Missed Fold: `vgf2p8affineqb(x | C, M, imm8)` => `vgf2p8affineqb(x, M & ~C, (u8)vgf2p8affineqb(C, M, imm8))`

Removing the same bit within each matrix's row acts like masking away the corresponding source bit. Using the following identity `(x & ~C) ^ C = x | C`, ORs by splatted 8-bit constants can be folded by masking the matrix and setting the immediate to be the affine transformation of the constant in the same manner (see #179606 & #191325).

Only works if the 64-bit matrix is the same for all lanes as they all use the same immediate.

```asm
setLowerBitRev_src:
por xmm0, xmmword ptr [rip + .LCPI0_0]
gf2p8affineqb xmm0, xmmword ptr [rip + .LCPI0_1], 0
ret
```

```asm
setLowerBitRev_tgt:
gf2p8affineqb xmm0, xmmword ptr [rip + .LCPI1_0], 240
ret
```

https://godbolt.org/z/ed7E45b31

Start by comparing the identities and examples in #179606 and #191325, then reproduce the source and target assembly on the linked Godbolt example. The work is done when the compiler recognizes this vgf2p8affineqb transformation, including the condition that the 64-bit matrix is identical across lanes, and produces the target form.

Written by the indexing model from the issue text.