All of lore.kernel.org
 help / color / mirror / Atom feed
From: Oleg Nesterov <oleg@redhat.com>
To: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Konstantin Khlebnikov <khlebnikov@openvz.org>,
	KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
	KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
	"Rafael J. Wysocki" <rjw@sisk.pl>,
	Rusty Russell <rusty@rustcorp.com.au>,
	Tejun Heo <htejun@gmail.com>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [patch] oom: thaw threads if oom killed thread is frozen before deferring
Date: Fri, 30 Sep 2011 17:30:36 +0200	[thread overview]
Message-ID: <20110930153036.GA19095@redhat.com> (raw)
In-Reply-To: <20110929180021.GA27999@tiehlicka.suse.cz>

On 09/29, Michal Hocko wrote:
>
> On Thu 29-09-11 18:37:24, Oleg Nesterov wrote:
> >
> > Oh, I don't think so. For what? This doesn't close other races, and
> > in fact the fatal_signal_pending() this patch adds is itself racy,
> > SIGKILL can come in between.
>
> OK, I think I see your point. You mean that oom will send KILL after
> both fatal_signal_pending in refrigerator and signal_pending check in
> schedule, right?

No, schedule()->signal_pending_state(TASK_UNINTERRUPTIBLE) doesn't check
the signals. I simply meant

	if (fatal_signal_pending())
					// <--- SIGKILL from oom
		try_to_freeze();

> This is what the follow up fix from David is doing. Check frozen in
> select_bad_process if the task is TIF_MEMDIE and thaw the process.
>
> And it seems that the David's follow up fix is sufficient so let's leave
> refrigerator alone.

Agreed, afaics this should fix all races (although I didn't read the
whole discussion, perhaps I missed something else).

And in this case we do not even need to modify oom_kill_task/etc,
select_bad_process() will be called again and notice the frozen task
eventually. Afaics.

Or, as Tejun suggests, we can implement the race-free kill-even-if-frozen
later.

Oleg.


WARNING: multiple messages have this Message-ID (diff)
From: Oleg Nesterov <oleg@redhat.com>
To: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Konstantin Khlebnikov <khlebnikov@openvz.org>,
	KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
	KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
	"Rafael J. Wysocki" <rjw@sisk.pl>,
	Rusty Russell <rusty@rustcorp.com.au>,
	Tejun Heo <htejun@gmail.com>,
	linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: Re: [patch] oom: thaw threads if oom killed thread is frozen before deferring
Date: Fri, 30 Sep 2011 17:30:36 +0200	[thread overview]
Message-ID: <20110930153036.GA19095@redhat.com> (raw)
In-Reply-To: <20110929180021.GA27999@tiehlicka.suse.cz>

On 09/29, Michal Hocko wrote:
>
> On Thu 29-09-11 18:37:24, Oleg Nesterov wrote:
> >
> > Oh, I don't think so. For what? This doesn't close other races, and
> > in fact the fatal_signal_pending() this patch adds is itself racy,
> > SIGKILL can come in between.
>
> OK, I think I see your point. You mean that oom will send KILL after
> both fatal_signal_pending in refrigerator and signal_pending check in
> schedule, right?

No, schedule()->signal_pending_state(TASK_UNINTERRUPTIBLE) doesn't check
the signals. I simply meant

	if (fatal_signal_pending())
					// <--- SIGKILL from oom
		try_to_freeze();

> This is what the follow up fix from David is doing. Check frozen in
> select_bad_process if the task is TIF_MEMDIE and thaw the process.
>
> And it seems that the David's follow up fix is sufficient so let's leave
> refrigerator alone.

Agreed, afaics this should fix all races (although I didn't read the
whole discussion, perhaps I missed something else).

And in this case we do not even need to modify oom_kill_task/etc,
select_bad_process() will be called again and notice the frozen task
eventually. Afaics.

Or, as Tejun suggests, we can implement the race-free kill-even-if-frozen
later.

Oleg.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2011-09-30 15:34 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-09-27  8:09 [PATCH 0/2] oom: fix livelock when frozen task is selected Michal Hocko
2011-09-27  6:56 ` [PATCH 1/2] lguest: move process freezing before pending signals check Michal Hocko
2011-10-12  6:55   ` Michal Hocko
2011-10-12  6:55     ` Michal Hocko
2011-10-12 23:57     ` Rusty Russell
2011-10-12 23:57       ` Rusty Russell
2011-10-13  5:48       ` Michal Hocko
2011-10-13  5:48         ` Michal Hocko
2011-09-27  8:01 ` [PATCH 2/2] oom: do not live lock on frozen tasks Michal Hocko
2011-09-27 18:35   ` [patch] oom: thaw threads if oom killed thread is frozen before deferring David Rientjes
2011-09-27 18:35     ` David Rientjes
2011-09-28 10:44     ` Michal Hocko
2011-09-28 10:44       ` Michal Hocko
2011-09-29 11:51       ` Michal Hocko
2011-09-29 11:51         ` Michal Hocko
2011-09-29 12:05         ` Oleg Nesterov
2011-09-29 12:05           ` Oleg Nesterov
2011-09-29 13:02           ` Michal Hocko
2011-09-29 13:02             ` Michal Hocko
2011-09-29 16:32             ` Michal Hocko
2011-09-29 16:32               ` Michal Hocko
2011-09-29 16:37             ` Oleg Nesterov
2011-09-29 16:37               ` Oleg Nesterov
2011-09-29 18:00               ` Michal Hocko
2011-09-29 18:00                 ` Michal Hocko
2011-09-30  1:51                 ` Tejun Heo
2011-09-30  1:51                   ` Tejun Heo
2011-09-30  7:41                   ` Michal Hocko
2011-09-30  7:41                     ` Michal Hocko
2011-09-30  7:46                     ` Tejun Heo
2011-09-30  7:46                       ` Tejun Heo
2011-09-30  8:04                       ` Michal Hocko
2011-09-30  8:04                         ` Michal Hocko
2011-09-30 15:30                 ` Oleg Nesterov [this message]
2011-09-30 15:30                   ` Oleg Nesterov
2011-10-28 22:23   ` [PATCH 2/2] oom: do not live lock on frozen tasks Andrew Morton
2011-10-28 22:23     ` Andrew Morton
2011-10-29  9:01     ` Michal Hocko
2011-10-29  9:01       ` 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=20110930153036.GA19095@redhat.com \
    --to=oleg@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=htejun@gmail.com \
    --cc=kamezawa.hiroyu@jp.fujitsu.com \
    --cc=khlebnikov@openvz.org \
    --cc=kosaki.motohiro@jp.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.cz \
    --cc=rientjes@google.com \
    --cc=rjw@sisk.pl \
    --cc=rusty@rustcorp.com.au \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.