From: Oleg Nesterov <oleg@redhat.com>
To: Suren Baghdasaryan <surenb@google.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>,
Michal Hocko <mhocko@suse.com>,
Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
Christian Brauner <christian.brauner@ubuntu.com>,
Tim Murray <timmurray@google.com>,
mingo@kernel.org, Peter Zijlstra <peterz@infradead.org>,
Thomas Gleixner <tglx@linutronix.de>,
esyr@redhat.com, christian@kellner.me, areber@redhat.com,
Shakeel Butt <shakeelb@google.com>,
cyphar@cyphar.com, adobriyan@gmail.com,
Andrew Morton <akpm@linux-foundation.org>,
gladkov.alexey@gmail.com, Michel Lespinasse <walken@google.com>,
daniel.m.jordan@oracle.com, avagin@gmail.com,
bernd.edlinger@hotmail.de,
John Johansen <john.johansen@canonical.com>,
laoar.shao@gmail.com, Minchan Kim <minchan@kernel.org>,
kernel-team <kernel-team@android.com>,
LKML <linux-kernel@vger.kernel.org>,
linux-fsdevel@vger.kernel.org, linux-mm <linux-mm@kvack.org>
Subject: Re: [PATCH 1/1] mm, oom_adj: don't loop through tasks in __set_oom_adj when not necessary
Date: Fri, 21 Aug 2020 19:59:43 +0200 [thread overview]
Message-ID: <20200821175943.GD19445@redhat.com> (raw)
In-Reply-To: <20200821163300.GB19445@redhat.com>
On 08/21, Oleg Nesterov wrote:
>
> On 08/21, Suren Baghdasaryan wrote:
> >
> > On Fri, Aug 21, 2020 at 4:16 AM Oleg Nesterov <oleg@redhat.com> wrote:
> > >
> > > bool probably_has_other_mm_users(tsk)
> > > {
> > > return atomic_read_acquire(&tsk->mm->mm_users) >
> > > atomic_read(&tsk->signal->live);
> > > }
> > >
> > > The barrier implied by _acquire ensures that if we race with the exiting
> > > task and see the result of exit_mm()->mmput(mm), then we must also see
> > > the result of atomic_dec_and_test(signal->live).
> > >
> > > Either way, if we want to fix the race with clone(CLONE_VM) we need other
> > > changes.
> >
> > The way I understand this condition in __set_oom_adj() sync logic is
> > that we would be ok with false positives (when we loop unnecessarily)
> > but we can't tolerate false negatives (when oom_score_adj gets out of
> > sync).
>
> Yes,
>
> > With the clone(CLONE_VM) race not addressed we are allowing
> > false negatives and IMHO that's not acceptable because it creates a
> > possibility for userspace to get an inconsistent picture. When
> > developing the patch I did think about using (p->mm->mm_users >
> > p->signal->nr_threads) condition and had to reject it due to that
> > reason.
>
> Not sure I understand... I mean, the test_bit(MMF_PROC_SHARED) you propose
> is equally racy and we need copy_oom_score() at the end of copy_process()
> either way?
On a second thought I agree that probably_has_other_mm_users() above can't
work ;) Compared to the test_bit(MMF_PROC_SHARED) check it is not _equally_
racy, it adds _another_ race with clone(CLONE_VM).
Suppose a single-threaded process P does
clone(CLONE_VM); // creates the child C
// mm_users == 2; P->signal->live == 1;
clone(CLONE_THREAD | CLONE_VM);
// mm_users == 3; P->signal->live == 2;
the problem is that in theory clone(CLONE_THREAD | CLONE_VM) can increment
_both_ counters between atomic_read_acquire(mm_users) and atomic_read(live)
in probably_has_other_mm_users() so it can observe mm_users == live == 2.
Oleg.
next prev parent reply other threads:[~2020-08-21 18:00 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-20 0:20 [PATCH 1/1] mm, oom_adj: don't loop through tasks in __set_oom_adj when not necessary Suren Baghdasaryan
2020-08-20 5:56 ` Michal Hocko
2020-08-20 8:46 ` Christian Brauner
2020-08-20 9:09 ` Michal Hocko
2020-08-20 10:32 ` Christian Brauner
2020-08-20 11:14 ` Michal Hocko
2020-08-20 10:55 ` Oleg Nesterov
2020-08-20 11:13 ` Michal Hocko
2020-08-20 11:29 ` Michal Hocko
2020-08-20 11:41 ` Oleg Nesterov
2020-08-20 11:47 ` Christian Brauner
2020-08-20 11:30 ` Christian Brauner
2020-08-20 11:42 ` Michal Hocko
2020-08-20 12:41 ` Michal Hocko
2020-08-20 13:43 ` Christian Brauner
2020-08-20 12:34 ` Eric W. Biederman
2020-08-20 12:42 ` Michal Hocko
2020-08-20 12:45 ` Eric W. Biederman
2020-08-20 12:54 ` Eric W. Biederman
2020-08-20 13:26 ` Michal Hocko
2020-08-20 13:34 ` Christian Brauner
2020-08-20 13:48 ` Tetsuo Handa
2020-08-20 14:00 ` Christian Brauner
2020-08-20 14:15 ` Michal Hocko
2020-08-20 14:26 ` Tetsuo Handa
2020-08-20 14:34 ` Michal Hocko
2020-08-20 14:18 ` Tetsuo Handa
2020-08-20 14:49 ` Eric W. Biederman
2020-08-20 15:06 ` Christian Brauner
2020-08-20 15:56 ` Suren Baghdasaryan
2020-08-20 16:26 ` Michal Hocko
2020-08-20 16:29 ` Christian Brauner
2020-08-20 16:47 ` Suren Baghdasaryan
2020-08-21 4:39 ` Eric W. Biederman
2020-08-21 7:17 ` Michal Hocko
2020-08-21 11:15 ` Oleg Nesterov
2020-08-21 15:28 ` Suren Baghdasaryan
2020-08-21 16:06 ` Suren Baghdasaryan
2020-08-21 16:37 ` Oleg Nesterov
2020-08-21 17:22 ` Suren Baghdasaryan
2020-08-21 16:33 ` Oleg Nesterov
2020-08-21 17:59 ` Oleg Nesterov [this message]
2020-08-21 18:53 ` Suren Baghdasaryan
2020-08-24 20:03 ` Suren Baghdasaryan
2020-08-20 13:41 ` Eric W. Biederman
2020-08-20 14:04 ` Oleg Nesterov
2020-08-20 14:36 ` Oleg Nesterov
2020-08-20 15:06 ` Eric W. Biederman
2020-08-20 14:43 ` Eric W. Biederman
2020-08-20 14:12 ` Michal Hocko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200821175943.GD19445@redhat.com \
--to=oleg@redhat.com \
--cc=adobriyan@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=areber@redhat.com \
--cc=avagin@gmail.com \
--cc=bernd.edlinger@hotmail.de \
--cc=christian.brauner@ubuntu.com \
--cc=christian@kellner.me \
--cc=cyphar@cyphar.com \
--cc=daniel.m.jordan@oracle.com \
--cc=ebiederm@xmission.com \
--cc=esyr@redhat.com \
--cc=gladkov.alexey@gmail.com \
--cc=john.johansen@canonical.com \
--cc=kernel-team@android.com \
--cc=laoar.shao@gmail.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=minchan@kernel.org \
--cc=mingo@kernel.org \
--cc=penguin-kernel@i-love.sakura.ne.jp \
--cc=peterz@infradead.org \
--cc=shakeelb@google.com \
--cc=surenb@google.com \
--cc=tglx@linutronix.de \
--cc=timmurray@google.com \
--cc=walken@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).