[llvm] [AMDGPU] Enable overwrite ALU bit in sched.barrier mask (PR #160782)

[Jan Patrick Lehr via llvm-commits]

jplehr wrote:

I followed up internally on the request. Let me share that feedback:

The original ask seems to have been satisfied without any implementation work and instead an explanation about the behavior and a recommendation as to what mask to use.

So, from all the discussion we had, it seems that the likely correct path forward is to make the documentation better and keep the implementation as-is.

Thoughts on that?

https://github.com/llvm/llvm-project/pull/160782