yxsamliu wrote: Overall I think this PR is sound, but could we add a short Clang doc section for `__amdgpu_named_workgroup_barrier_t`, probably in `clang/docs/AMDGPUSupport.rst`?
This type has unusual semantics, especially with wrapper types: the field itself cannot have a separate address space, so Clang treats the whole wrapper object as a named-barrier object. It would help to document the typical use case, allowed wrapper shapes, invalid uses, and how it is intended to be used with the named-barrier builtins. https://github.com/llvm/llvm-project/pull/207687 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
