jroelofs wrote: > The checks on the tracing output test intentional aspects of the design that > mirror the VAPI state management of `setjmp`/`longjmp` on AIX and cannot be > observed otherwise.
Another option that you might consider: expose that state via a DWARF pseudo register, similar to how I fixed `UNW_AARCH64_RA_SIGN_STATE` in https://github.com/llvm/llvm-project/pull/209947/changes. That said, the pseudo would probably make more sense if you had CFI describing this, and the test you've got now seems good enough... so it may not be worth the hassle. Up to you. Either way, I think my concerns here have been addressed. https://github.com/llvm/llvm-project/pull/209306 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
