thurstond wrote: > Agreed, so what should be the best way to approach this issue currently. > Would it be better to fold this into MSan, remove it from UBSan, or something > else you suggest?
I suggest folding it into MSan. There is somewhat of a precedent with `-fsanitize-memory-param-retval`, which optionally treats function parameters/return values as sinks. Analogously, there could be an `-fsanitize-memory-local-address-never-taken` (or something like that) that enables the additional sink behavior. The tricky part is that probably needs some information to be plumbed from the clang front-end through to MSan via the IR. https://github.com/llvm/llvm-project/pull/207529 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
