[llvm] [AMDGPU][MC] Disallow op_sel in v_dot4 and v_dot8 with 4- or 8-bit packed data (PR #100485)

[Matt Arsenault via llvm-commits]

arsenm wrote:

We should run this by hardware. If I were to guess, this is yet another case where they didn't validate it works so did not document it as supported.

https://github.com/llvm/llvm-project/pull/100485