https://bugs.kde.org/show_bug.cgi?id=452274
--- Comment #5 from Paul Floyd <pjfl...@wanadoo.fr> --- OK thanks. In that case I have an idea. Back to https://github.com/paulfloyd/freebsd_valgrind/issues/137, the flow is signal raised -> pth_sighandler -> user sighandler -> return to pth_sighandler -> call sigreturn -> sigreturn fails -> annoying kernel message -> return to Valgrind trampiline -> syscall fake_sigreturn -> back to VEX CPU but what would happen if "sigreturn fails" doesn't fail? That's really bad news and will almost certainly result in a crash. We need it to fail in order to get back to fake_sigreturn. I'll try to see what happens if I do something like set bit 4 of mc_flags in the synthesized ucontext to ensure that sigreturn fails in the mc_flags check (which is the least-worst option at the moment). -- You are receiving this mail because: You are watching all bug changes.
