[Openmp-commits] [openmp] [OpenMP][AMDGPU] Adapt dynamic callstack sizes to HIP behavior (PR #74080)

[Jon Chesterfield via Openmp-commits]

JonChesterfield wrote:

Make a header under llvm/lib/Target/AMDGPU and include it from the backend and from here. Provided it's an inline function in the header there's no messing around with cmake. The llvm libs are already in scope for openmp runtime. I'd say trivial technical complexity and moderate chance of the backend developers not liking the patch spawning moderate political complexity.

https://github.com/llvm/llvm-project/pull/74080