aaron.ballman added a comment. In D131307#3709494 <https://reviews.llvm.org/D131307#3709494>, @JonChesterfield wrote:
> Did some digging here. The function hsa_agent_get_info takes an argument of > type hsa_agent_info_t, which has declared values in the range [0 24]. The > implementation of that (in amd_gpu_agent fwiw) casts that argument to a > size_t and then switches on it, checking those declared values and a bunch of > extensions. This is used to provide vendor extensions through a > vendor-agnostic interface. > > This seems to be a case where C and C++ have diverged. Yes, they've always been divergent in this area, it's only thanks to the magic of constexpr that anyone really notices now. > As far as I can tell, C thinks an enum is an int, and anything that fits in > an int can be stored in one and retrieved later. It's a bit more complicated than that, unfortunately. But this is kind of the gist of it. > C23 lets one specify the underlying type. C++ evidently thinks the value > stored must be within [min max] of the declaration, which is at least more > flexible than requiring it be one in the declaration. Correct, C++ uses the minimum-sized bit-field that can hold all of the values of the enumeration. > So I think the fix here is to change hsa_agent_info_t to include > `HSA_AGENT_INFO_UNUSED_INCREASE_RANGE_OF_TYPE = INT32_MAX` so the vendor > extensions remain accessible. It's a header that is usable from C and C++ so > it needs to do something conforming to both. Does that sound right? Yes, that's how I'd solve the problem. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D131307/new/ https://reviews.llvm.org/D131307 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
