On Thu, 12 Mar 2020, Tetsuo Handa wrote: > > The cond_resched() here is needed if the iteration is lengthy depending on > > the number of descendant memcgs already. > > No. cond_resched() here will become no-op if CONFIG_PREEMPTION=y and current > thread has realtime priority. > It's helpful without CONFIG_PREEMPTION for excessively long memcg iterations to avoid starving need_resched. > > schedule_timeout_killable(1) does not make any guarantees that current > > will be scheduled after the victim or oom_reaper on UP systems. > > The point of schedule_timeout_*(1) is to guarantee that current thread > will yield CPU to other threads even if CONFIG_PREEMPTION=y and current > thread has realtime priority case. There is no guarantee that current > thread will be rescheduled immediately after a sleep is irrelevant. > > > > > If you have an alternate patch to try, we can test it. But since this > > cond_resched() is needed anyway, I'm not sure it will change the result. > > schedule_timeout_killable(1) is an alternate patch to try; I don't think > that this cond_resched() is needed anyway. > You are suggesting schedule_timeout_killable(1) in shrink_node_memcgs()?
