Re: + oom-pm-oom-killed-task-cannot-escape-pm-suspend.patch added to -mm tree

[Oleg Nesterov <oleg@redhat.com>]

On 10/20, Michal Hocko wrote:
>
> On Fri 17-10-14 19:19:04, Oleg Nesterov wrote:
> > > @@ -504,11 +516,13 @@ void oom_kill_process(struct task_struct
> > >  			pr_err("Kill process %d (%s) sharing same memory\n",
> > >  				task_pid_nr(p), p->comm);
> > >  			task_unlock(p);
> > > +			atomic_inc(&oom_kills);
> >
> > Do we really need this? Can't freeze_processes() (ab)use oom_notify_list?
>
> I would really prefer not using oom_notify_list. It is just an ugly
> interface.

And to me oom_kills_count() is more ugly ;) But! of course this is
subjective, I am not going to insist.

> Reduce the race window by checking all tasks after OOM killer has been
> disabled. This is still not race free

Yes, thanks.

I only argued because this fact was not documented. And I agree that it
is hardly possible to close this race, and this patch makes the things
better.

I think this version is fine.

Oleg.