From: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
To: Kirill Tkhai <ktkhai@virtuozzo.com>,
Michal Hocko <mhocko@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>,
David Rientjes <rientjes@google.com>,
linux-mm@kvack.org, Peter Zijlstra <peterz@infradead.org>
Subject: Re: [PATCH] memcg: killed threads should not invoke memcg OOM killer
Date: Fri, 28 Dec 2018 20:00:50 +0900 [thread overview]
Message-ID: <16e155a5-5eff-a165-bbab-7219674683bf@i-love.sakura.ne.jp> (raw)
In-Reply-To: <6a52dc15-3e0a-5469-3a68-c7922a52a2d3@virtuozzo.com>
On 2018/12/28 19:22, Kirill Tkhai wrote:
>> @@ -1389,8 +1389,13 @@ static bool mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask,
>> };
>> bool ret;
>>
>> - mutex_lock(&oom_lock);
>> - ret = out_of_memory(&oc);
>> + if (mutex_lock_killable(&oom_lock))
>> + return true;
>> + /*
>> + * A few threads which were not waiting at mutex_lock_killable() can
>> + * fail to bail out. Therefore, check again after holding oom_lock.
>> + */
>> + ret = fatal_signal_pending(current) || out_of_memory(&oc);
>
> This fatal_signal_pending() check has a sense because of
> it's possible, a killed task is waking up slowly, and it
> returns from schedule(), when there are no more waiters
> for a lock.
Thanks. Michal thinks that mutex_lock_killable() would be sufficient
( https://lkml.kernel.org/r/20181107100810.GA27423@dhcp22.suse.cz ) but
I can confirm that mutex_lock_killable() is not sufficient when I test
using a VM with 8 CPUs. Thus, I'd like to keep this fatal_signal_pending()
check.
>
> Why not make this approach generic, and add a check into
> __mutex_lock_common() after schedule_preempt_disabled()
> instead of this? This will handle all the places like
> that at once.
>
> (The only adding a check is not enough for __mutex_lock_common(),
> since mutex code will require to wake next waiter also. So,
> you will need a couple of changes in mutex code).
I think that we should not assume that everybody is ready for making
mutex_lock_killable() to return -EINTR if fatal_signal_pending() is
true, and that adding below version would be a safer choice.
int __sched mutex_lock_unless_killed(struct mutex *lock)
{
const int ret = mutex_lock_killable(lock);
if (ret)
return ret;
if (fatal_signale_pending(current)) {
mutex_unlock(lock);
return -EINTR;
}
return 0;
}
next prev parent reply other threads:[~2018-12-28 11:01 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-26 10:13 [PATCH] memcg: killed threads should not invoke memcg OOM killer Tetsuo Handa
2018-12-28 10:22 ` Kirill Tkhai
2018-12-28 11:00 ` Tetsuo Handa [this message]
2018-12-28 11:28 ` Kirill Tkhai
2019-01-06 6:02 ` Tetsuo Handa
2019-01-06 6:02 ` Tetsuo Handa
2019-01-07 11:41 ` Michal Hocko
2019-01-07 13:07 ` Tetsuo Handa
2019-01-07 13:37 ` Michal Hocko
2019-01-07 14:20 ` Tetsuo Handa
2019-01-09 10:56 ` Tetsuo Handa
2019-01-09 10:56 ` Tetsuo Handa
2019-01-15 10:17 ` [PATCH v2] " Tetsuo Handa
2019-01-15 11:55 ` 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=16e155a5-5eff-a165-bbab-7219674683bf@i-love.sakura.ne.jp \
--to=penguin-kernel@i-love.sakura.ne.jp \
--cc=akpm@linux-foundation.org \
--cc=hannes@cmpxchg.org \
--cc=ktkhai@virtuozzo.com \
--cc=linux-mm@kvack.org \
--cc=mhocko@kernel.org \
--cc=peterz@infradead.org \
--cc=rientjes@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).