rjmccall added a comment. On the one hand, we can certainly just have different behavior on BPF targets if this is a common expectation there.
On the other hand, if there isn't a requirement to put `const volatile` objects in writable memory, then I'd rather not. If someone has specific expectations about the section a symbol ends up in that don't match the standard interpretation in the language, they really ought to enforce their expectations with a section attribute. The use case of volatile memory-mapped I/O needs something like that anyway (or something even worse). CHANGES SINCE LAST ACTION https://reviews.llvm.org/D131012/new/ https://reviews.llvm.org/D131012 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits