On Tue, Jul 21, 2026 at 12:54 AM Linus Walleij <[email protected]> wrote: > > On Wed, Jul 15, 2026 at 7:52 PM Kees Cook <[email protected]> wrote: > > > ARM32 GCC KCFI[1] violations currently show as generic "Oops - undefined > > instruction" errors, making debugging CFI failures difficult. Add a proper > > KCFI trap handler similar to the aarch64 implementation to provide clear > > CFI error messages, including the call target and the expected type. > > Reviewed-by: Linus Walleij <[email protected]> > > > Clang and GCC trap CFI failures differently on ARM32. Clang lowers > > its checks to a BKPT, handled via the breakpoint/prefetch-abort path > > in hw_breakpoint.c, which cannot recover the target or expected type > > and so must report via report_cfi_failure_noaddr(). GCC instead lowers > > KCFI checks to a UDF (undefined instruction) whose immediate encodes > > the registers involved, so the handler can decode it and call the full > > report_cfi_failure() with the target and expected type. > > Aha, that explains a few things. > > The UDF is a better approach for ARM I think (Russell will know best). > > As noted from earlier conversation the breakpoint usecase overloading > on ARM is a mess, and we would certainly like to avoid it if possible, > albeit I have a patch in Russell's patch tracker to paper it over. > > Do you think that long term we could switch Clang to do what > GCC does?
Yeah, I think that's reasonable. https://github.com/llvm/llvm-project/issues/211032 https://github.com/ClangBuiltLinux/linux/issues/2176 -- Thanks, ~Nick Desaulniers
