On Tue 03-04-18 21:29:52, Tetsuo Handa wrote: > Michal Hocko wrote: [...] > > Please be aware that we _do_ allocate in the exit path. I have a strong > > suspicion that even while fatal signal is pending. Do we really want > > fail those really easily. > > Does the exit path mean inside do_exit() ? If yes, fatal signals are already > cleared before reaching do_exit(). They usually are. But we can send a SIGKILL on an already killed task after it removed the previously deadly signal already AFAIR. Maybe I mis-remember of course. Signal handling code always makes my head explode and I tend to forget all the details. Anyway relying on fatal_signal_pending for some allocator semantic is just too subtle. -- Michal Hocko SUSE Labs
