[llvm] [X86] Allow EVEX compression for VPMOV*2M + KMOV pattern (PR #175219)

[Qihan Cai via llvm-commits]

realqhc wrote:

I think `evex-to-vex-compress.mir` is more appropriate for logic related tests, so I put the positive and negative conversion tests there. 

https://github.com/llvm/llvm-project/pull/175219