[llvm] Reapply [AMDGPU] Avoid resource propagation for recursion through multiple functions (PR #112251)

[Matt Arsenault via llvm-commits]

https://github.com/arsenm approved this pull request.

lgtm but please continue working on flattening the expression instead of going to the worst case 

https://github.com/llvm/llvm-project/pull/112251