[llvm] [AMDGPU] Set DecoderNamespace consistently for GFX10+ MIMG instructions (PR #81881)

[Jay Foad via llvm-commits]

jayfoad wrote:

TBH I am not sure if there was any particular reason why some instructions use "AMDGPU" and some used arch-specific namespaces like "GFX10". But it was definitely inconsistent.

https://github.com/llvm/llvm-project/pull/81881