[llvm] [X86] Don't shrink VEX3 to VEX2 on a symbolic compare predicate (PR #213172)

[Zane Hambly via llvm-commits]

Zaneham wrote:

Added. I couldn't get it to crash from llvm-mc though, that one is only called from X86MCInstLower, so there's no test for it.

https://github.com/llvm/llvm-project/pull/213172