> I wonder if it makes sense to move ->cred_guard_mutex from task_struct > to signal_struct and thus make multiple-threads-inside-exec impossible. > Only one thread can win anyway. That probably makes sense. Note that cred_guard_mutex is also overloaded for ptrace_attach, so this would add some more serialization of attaches to threads in the same group. But as long as actual attachment serializes on tasklist_lock anyway, it doesn't make a material difference. (Even without that, it would presumably be the same debugger attaching serially to threads in the same group, so it wouldn't degrade anything in practice.) Thanks, Roland -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html