linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Oleg Nesterov <oleg@redhat.com>
To: David Rientjes <rientjes@google.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
	Hugh Dickins <hughd@google.com>,
	linux-mm@kvack.org, Andrey Vagin <avagin@openvz.org>
Subject: Re: [patch] oom: prevent unnecessary oom kills or kernel panics
Date: Wed, 9 Mar 2011 12:06:06 +0100	[thread overview]
Message-ID: <20110309110606.GA16719@redhat.com> (raw)
In-Reply-To: <alpine.DEB.2.00.1103081549530.27910@chino.kir.corp.google.com>

On 03/08, David Rientjes wrote:
>
> On Tue, 8 Mar 2011, Oleg Nesterov wrote:
>
> > > > By iterating over threads instead, it is possible to detect threads that
> > > > are exiting and nominate them for oom kill so they get access to memory
> > > > reserves.
> > >
> > > In fact, PF_EXITING is a sing of *THREAD* exiting, not process. Therefore
> > > PF_EXITING is not a sign of memory freeing in nearly future. If other
> > > CPUs don't try to free memory, prevent oom and waiting makes deadlock.
> >
> > I agree. I don't understand this patch.
> >
>
> Using for_each_process() does not consider threads that have failed to
> exit after the oom killed parent and, thus, we select another innocent
> task to kill when we're really just waiting for those threads to exit

How so? select_bad_process() checks TIF_MEMDIE and returns ERR_PTR()
if it is set.

And, exactly because we use for_each_process() we do not need to check
other threads. The main thread can't disappear until they all exit.

Imho TIF_MEMDIE is not perfect and should be replaced by MMF_, but this
is another story. Hmm... and in any case, currently TIF_MEMDIE is not
always used correctly, afaics.

> The end result is that without this patch, we sometimes unnecessarily
> panic (and "sometimes" is defined as "many machines" for us) when nothing
> else is eligible for kill within an oom cpuset yet doing a
> do_each_thread() over that cpuset shows threads of previously oom killed
> parent that have yet to exit.
>
> > > > @@ -324,7 +324,7 @@ static struct task_struct *select_bad_process(unsigned int *ppoints,
> > > >  		 * the process of exiting and releasing its resources.
> > > >  		 * Otherwise we could get an easy OOM deadlock.
> > > >  		 */
> > > > -		if (thread_group_empty(p) && (p->flags & PF_EXITING) && p->mm) {
> > > > +		if ((p->flags & PF_EXITING) && p->mm) {
> >
> > The previous check was not perfect, we know this.
> >
> > But with this patch applied, the simple program below disables oom-killer
> > completely. select_bad_process() can never succeed.
> >
>
> The program illustrates a problem that shouldn't be fixed in
> select_bad_process() but rather in oom_kill_process() when choosing an
> eligible child of the selected task to kill in place of its parent.

Can't understand. oom_kill_process() is never called exactly because
select_bad_process() is fooled.

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:[~2011-03-09 11:14 UTC|newest]

Thread overview: 53+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-03-01 19:09 [patch] oom: prevent unnecessary oom kills or kernel panics David Rientjes
2011-03-03  1:20 ` KOSAKI Motohiro
2011-03-03 19:53   ` David Rientjes
2011-03-06 11:14     ` KOSAKI Motohiro
2011-03-06 22:06       ` David Rientjes
2011-03-08  0:24         ` KOSAKI Motohiro
2011-03-08  2:01         ` KOSAKI Motohiro
2011-03-08 13:42   ` Oleg Nesterov
2011-03-08 23:57     ` David Rientjes
2011-03-09 10:36       ` KOSAKI Motohiro
2011-03-09 11:06       ` Oleg Nesterov [this message]
2011-03-09 20:32         ` David Rientjes
2011-03-10 12:05           ` Oleg Nesterov
2011-03-10 15:40             ` [PATCH 0/1] Was: " Oleg Nesterov
2011-03-10 15:41               ` [PATCH 1/1] oom_kill_task: mark every thread as TIF_MEMDIE Oleg Nesterov
2011-03-13  1:08                 ` David Rientjes
2011-03-10 16:36               ` [PATCH 0/1] select_bad_process: improve the PF_EXITING check Oleg Nesterov
2011-03-10 16:37                 ` [PATCH 1/1] " Oleg Nesterov
2011-03-10 16:40                 ` [PATCH 0/1] " Oleg Nesterov
2011-03-10 17:18                   ` [PATCH v2 " Oleg Nesterov
2011-03-10 17:19                     ` [PATCH v2 1/1] " Oleg Nesterov
2011-03-13  1:06             ` [patch] oom: prevent unnecessary oom kills or kernel panics David Rientjes
2011-03-09 23:19       ` Andrew Morton
2011-03-11 19:45         ` David Rientjes
2011-03-12 12:34           ` Oleg Nesterov
2011-03-12 13:43             ` [PATCH 0/3] oom: TIF_MEMDIE/PF_EXITING fixes Oleg Nesterov
2011-03-12 13:44               ` [PATCH 1/3] oom: oom_kill_task: mark every thread as TIF_MEMDIE Oleg Nesterov
2011-03-13  1:14                 ` David Rientjes
2011-03-12 13:44               ` [PATCH 2/3] oom: select_bad_process: improve the PF_EXITING check Oleg Nesterov
2011-03-12 13:44               ` [PATCH 3/3] oom: select_bad_process: use same_thread_group() Oleg Nesterov
2011-03-12 19:40               ` [PATCH 0/3] oom: TIF_MEMDIE/PF_EXITING fixes Hugh Dickins
2011-03-13  8:53                 ` KOSAKI Motohiro
2011-03-13 21:27                 ` Oleg Nesterov
2011-03-14 19:04                   ` [PATCH 0/3 for 2.6.38] oom: fixes Oleg Nesterov
2011-03-14 19:04                     ` [PATCH 1/3 for 2.6.38] oom: oom_kill_process: don't set TIF_MEMDIE if !p->mm Oleg Nesterov
2011-03-14 19:35                       ` Linus Torvalds
2011-03-14 20:31                         ` Oleg Nesterov
2011-03-14 20:32                         ` David Rientjes
2011-03-15 19:12                           ` Oleg Nesterov
2011-03-15 19:51                             ` David Rientjes
2011-03-14 20:22                       ` David Rientjes
2011-03-15 18:53                         ` Oleg Nesterov
2011-03-15 19:54                           ` David Rientjes
2011-03-15 21:16                             ` Oleg Nesterov
2011-03-14 19:05                     ` [PATCH 2/3 for 2.6.38] oom: select_bad_process: ignore TIF_MEMDIE zombies Oleg Nesterov
2011-03-14 20:50                       ` David Rientjes
2011-03-14 19:05                     ` [PATCH 3/3 for 2.6.38] oom: oom_kill_process: fix the child_points logic Oleg Nesterov
2011-03-14 20:41                       ` David Rientjes
2011-03-15 19:21                         ` Oleg Nesterov
2011-03-13 11:36               ` [PATCH 0/3] oom: TIF_MEMDIE/PF_EXITING fixes KOSAKI Motohiro
2011-03-13  1:11             ` [patch] oom: prevent unnecessary oom kills or kernel panics David Rientjes
2011-03-13  1:15               ` [patch -mm] oom: avoid deferring oom killer if exiting task is being traced David Rientjes
2011-03-14 17:40                 ` Oleg Nesterov

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=20110309110606.GA16719@redhat.com \
    --to=oleg@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=avagin@openvz.org \
    --cc=hughd@google.com \
    --cc=kamezawa.hiroyu@jp.fujitsu.com \
    --cc=kosaki.motohiro@jp.fujitsu.com \
    --cc=linux-mm@kvack.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).