efriedma-quic wrote: In the shadow-stack case, do we need to force a spill of the return address if there are objects that need a stack protector?
More generally, I think I'd like to tie the flag more strongly to the specific target feature that's providing the protection. If some target feature is providing a form of "stack protection", we need to make sure that feature is actually spilling the register in question in the right place. The implementation might be target-independent in some sense, but it's relying on invariants that the backend might not provide in the general case. https://github.com/llvm/llvm-project/pull/231651 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
