On Thu 14-04-16 14:01:06, Michal Hocko wrote: [...] > diff --git a/mm/oom_kill.c b/mm/oom_kill.c > index 716759e3eaab..d5a4d08f2031 100644 > --- a/mm/oom_kill.c > +++ b/mm/oom_kill.c > @@ -286,6 +286,13 @@ enum oom_scan_t oom_scan_process_thread(struct oom_control *oc, > return OOM_SCAN_CONTINUE; > > /* > + * mm of this task has already been reaped so it doesn't make any > + * sense to select it as a new oom victim. > + */ > + if (test_bit(MMF_OOM_REAPED, &task->mm->flags)) > + return OOM_SCAN_CONTINUE; This will have to move to oom_badness to where we check for OOM_SCORE_ADJ_MIN to catch the case where we try to sacrifice a child... In the meantime I have generated a full patch and will repost it with other oom reaper follow ups sometimes next week. -- Michal Hocko SUSE Labs -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>