linux-kernel - Re: [PATCH 0/2] oom, memcg: do not report racy no-eligible OOM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20190111113354.GD14956@dhcp22.suse.cz>
Date:   Fri, 11 Jan 2019 12:33:54 +0100
From:   Michal Hocko <mhocko@...nel.org>
To:     Tetsuo Handa <penguin-kernel@...ove.sakura.ne.jp>
Cc:     Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
        Johannes Weiner <hannes@...xchg.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 0/2] oom, memcg: do not report racy no-eligible OOM

On Fri 11-01-19 19:25:22, Tetsuo Handa wrote:
> On 2019/01/11 8:59, Tetsuo Handa wrote:
> > Michal Hocko wrote:
> >> On Wed 09-01-19 20:34:46, Tetsuo Handa wrote:
> >>> On 2019/01/09 20:03, Michal Hocko wrote:
> >>>> Tetsuo,
> >>>> can you confirm that these two patches are fixing the issue you have
> >>>> reported please?
> >>>>
> >>>
> >>> My patch fixes the issue better than your "[PATCH 2/2] memcg: do not
> >>> report racy no-eligible OOM tasks" does.
> >>
> >> OK, so we are stuck again. Hooray!
> > 
> > Andrew, will you pick up "[PATCH 3/2] memcg: Facilitate termination of memcg OOM victims." ?
> > Since mm-oom-marks-all-killed-tasks-as-oom-victims.patch does not call mark_oom_victim()
> > when task_will_free_mem() == true, memcg-do-not-report-racy-no-eligible-oom-tasks.patch
> > does not close the race whereas my patch closes the race better.
> > 
> 
> I confirmed that mm-oom-marks-all-killed-tasks-as-oom-victims.patch and
> memcg-do-not-report-racy-no-eligible-oom-tasks.patch are completely failing
> to fix the issue I am reporting. :-(

OK, this is really interesting. This means that we are racing
when marking all the tasks sharing the mm with the clone syscall.
Does fatal_signal_pending handle this better?
-- 
Michal Hocko
SUSE Labs