[llvm] [AMDGPU] Correctly insert s_nops for implicit read of SDWA (PR #100276)

[Jeffrey Byrnes via llvm-commits]

jrbyrnes wrote:

Added support for inline asm under the assumption that all inline asm have forwarded dest hazard, and all inline asm consume forward dest.

I think I may have misinterpreted the requests, and am not sure this is actually needed? I don't see why any user would use inline asm for VALU..

https://github.com/llvm/llvm-project/pull/100276