On 2020/07/14 9:15, Tetsuo Handa wrote: > On 2020/07/14 4:05, Michal Hocko wrote: >> Another alternative would be to check mem_cgroup_margin after the lock >> is taken but it would be better to keep in sync with the global case as >> much as possible unless there is a good reason to differ. >> > > If mem_cgroup_margin() is something what zone_watermark_ok() from > get_page_from_freelist() will do, checking mem_cgroup_margin() with > oom_lock held will make more keep in sync with the global case > (i.e. calling __alloc_pages_may_oom() with oom_lock held). > I meant: (i.e. calling get_page_from_freelist() from __alloc_pages_may_oom() with oom_lock held).
