mylai-mtk wrote: > It is useful to protect a binary with an -mcpu that doesn't support Zicfilp.
Generally I would agree that it's a good idea to protect a binary with an `-mcpu` that doesn't support Zicfilp, but to lift the check of Zicfilp would be misleading to naive compiler/RISC-V users who don't know that only with a system fully aware of Zicfilp, the protection of `-fcf-protection=branch` can be expected. I think from this perspective, compilers should at least warn the user if he asks for a feature that is known to be broken with his environment. @kito-cheng I need your input on this, since this would require a few updates on various PRs to the RISC-V specs. https://github.com/llvm/llvm-project/pull/109600 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
