On Mon 28-01-19 10:26:16, Andrew Morton wrote: > On Fri, 25 Jan 2019 18:24:16 +0100 Michal Hocko <mhocko@xxxxxxxxxx> wrote: > > > > > out_of_memory > > > > select_bad_process # no task > > > > > > > > If Thread1 didn't race it would bail out from try_charge and force the > > > > charge. We can achieve the same by checking tsk_is_oom_victim inside the > > > > oom_lock and therefore close the race. > > > > > > > > [1] http://lkml.kernel.org/r/bb2074c0-34fe-8c2c-1c7d-db71338f1e7f@xxxxxxxxxxxxxxxxxxx > > > > Link: http://lkml.kernel.org/r/20190107143802.16847-3-mhocko@xxxxxxxxxx > > > > Signed-off-by: Michal Hocko <mhocko@xxxxxxxx> > > > > Cc: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> > > > > Cc: Johannes Weiner <hannes@xxxxxxxxxxx> > > > > Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> > > > > > > It looks like this problem is happening in production systems: > > > > > > https://www.spinics.net/lists/cgroups/msg21268.html > > > > > > where the threads don't exit because they are trapped writing out the > > > oom messages to a slow console (running the reproducer from this email > > > thread triggers the oom flooding). > > > > > > So IMO we should put this into 5.0 and add: > > > > Please note that Tetsuo has found out that this will not work with the > > CLONE_VM without CLONE_SIGHAND cases and his http://lkml.kernel.org/r/01370f70-e1f6-ebe4-b95e-0df21a0bc15e@xxxxxxxxxxxxxxxxxxx > > should handle this case as well. I've only had objections to the > > changelog but other than that the patch looked sensible to me. > > So I think you're saying that > > mm-oom-marks-all-killed-tasks-as-oom-victims.patch > and > memcg-do-not-report-racy-no-eligible-oom-tasks.patch > > should be dropped and that "[PATCH v2] memcg: killed threads should not > invoke memcg OOM killer" should be redone with some changelog > alterations and should be merged instead? Yup. -- Michal Hocko SUSE Labs
