From: "Paul E. McKenney" <paulmck@us.ibm.com>
To: Eric Dumazet <dada1@cosmosbay.com>
Cc: dipankar@in.ibm.com, Linus Torvalds <torvalds@osdl.org>,
Jean Delvare <khali@linux-fr.org>,
Serge Belyshev <belyshev@depni.sinp.msu.ru>,
LKML <linux-kernel@vger.kernel.org>,
Andrew Morton <akpm@osdl.org>,
Manfred Spraul <manfred@colorfullife.com>
Subject: Re: VFS: file-max limit 50044 reached
Date: Tue, 18 Oct 2005 09:22:23 -0700 [thread overview]
Message-ID: <20051018162222.GA1304@us.ibm.com> (raw)
In-Reply-To: <4354C476.40901@cosmosbay.com>
On Tue, Oct 18, 2005 at 11:46:30AM +0200, Eric Dumazet wrote:
> Paul E. McKenney a écrit :
> >
> >
> >>+/*
> >>+ * Should we directly call rcu_do_batch() here ?
> >>+ * if (unlikely(rdp->count > 10000))
> >>+ * rcu_do_batch(rdp);
> >>+ */
> >
> >
> >Good thing that the above is commented out! ;-)
> >
> >Doing this can result in self-deadlock, for example with the following:
> >
> > spin_lock(&mylock);
> > /* do some stuff. */
> > call_rcu(&p->rcu_head, my_rcu_callback);
> > /* do some more stuff. */
> > spin_unlock(&mylock);
> >
> >void my_rcu_callback(struct rcu_head *p)
> >{
> > spin_lock(&mylock);
> > /* self-deadlock via call_rcu() via rcu_do_batch()!!! */
> > spin_unlock(&mylock);
> >}
> >
> >
> > Thanx, Paul
>
> Thanks Paul for reminding us that call_rcu() should not ever call the
> callback function, as very well documented in Documentation/RCU/UP.txt
> (Example 3: Death by Deadlock)
>
> But is the same true for call_rcu_bh() ?
Yes, same rules for this aspect of call_rcu_bh() and call_rcu().
> I intentionally wrote the comment to remind readers that a low maxbatch can
> trigger OOM in case a CPU is filled by some kind of DOS (network IRQ flood
> for example, targeting the IP dst cache)
>
> To solve this problem, may be we could add a requirement to
> call_rcu_bh/callback functions : If they have to lock a spinlock, only use
> a spin_trylock() and make them returns a status (0 : sucessfull callback,
> 1: please requeue me)
>
> As most callback functions just kfree() some memory, most of OOM would be
> cleared.
>
> int my_rcu_callback(struct rcu_head *p)
> {
> if (!spin_trylock(&mylock))
> return 1; /* please call me later */
> /* do something here */
> ...
> spin_unlock(&mylock);
> return 0;
> }
>
> (Changes to rcu_do_batch() are left as an exercice :) )
Another approach that would keep the current easier-to-use semantics
would be to schedule a tasklet or workqueue to process the callbacks
in a safe context.
Thanx, Paul
next prev parent reply other threads:[~2005-10-18 16:21 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-10-15 13:19 VFS: file-max limit 50044 reached Serge Belyshev
2005-10-15 17:53 ` Serge Belyshev
2005-10-16 16:23 ` Dipankar Sarma
2005-10-16 18:51 ` Serge Belyshev
2005-10-16 18:56 ` Dipankar Sarma
2005-10-17 2:19 ` Linus Torvalds
2005-10-17 4:43 ` Serge Belyshev
2005-10-17 8:32 ` Jean Delvare
2005-10-17 8:46 ` Dipankar Sarma
2005-10-17 9:10 ` Eric Dumazet
2005-10-17 9:14 ` Christoph Hellwig
2005-10-17 9:25 ` Eric Dumazet
2005-10-17 10:32 ` Dipankar Sarma
2005-10-17 12:10 ` [RCU problem] was " Eric Dumazet
2005-10-17 12:31 ` linux-os (Dick Johnson)
2005-10-17 12:36 ` Dipankar Sarma
2005-10-17 13:28 ` Eric Dumazet
2005-10-17 13:33 ` Dipankar Sarma
2005-10-17 14:54 ` Eric Dumazet
2005-10-17 15:42 ` Linus Torvalds
2005-10-17 16:01 ` Eric Dumazet
2005-10-17 16:16 ` Linus Torvalds
2005-10-17 16:29 ` Dipankar Sarma
2005-10-17 18:01 ` Eric Dumazet
2005-10-17 18:31 ` Dipankar Sarma
2005-10-17 19:00 ` Linus Torvalds
2005-10-17 18:37 ` Linus Torvalds
2005-10-17 19:12 ` Eric Dumazet
2005-10-17 19:30 ` Linus Torvalds
2005-10-17 19:39 ` Eric Dumazet
2005-10-17 20:14 ` Linus Torvalds
2005-10-17 20:25 ` Christopher Friesen
2005-10-17 20:24 ` Dipankar Sarma
2005-10-18 15:55 ` Christopher Friesen
2005-10-17 20:38 ` Linus Torvalds
2005-10-17 20:33 ` Dipankar Sarma
2005-10-17 22:40 ` Linus Torvalds
2005-10-17 22:59 ` Paul E. McKenney
2005-10-18 9:46 ` Eric Dumazet
2005-10-18 16:22 ` Paul E. McKenney [this message]
2005-10-17 18:15 ` Dipankar Sarma
2005-10-17 18:40 ` Linus Torvalds
2005-10-17 16:23 ` Dipankar Sarma
2005-10-17 16:31 ` Lee Revell
2005-10-17 16:20 ` Dipankar Sarma
2005-10-17 2:34 ` Linus Torvalds
2005-10-17 3:54 ` Roland Dreier
2005-10-17 11:54 ` Dipankar Sarma
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=20051018162222.GA1304@us.ibm.com \
--to=paulmck@us.ibm.com \
--cc=akpm@osdl.org \
--cc=belyshev@depni.sinp.msu.ru \
--cc=dada1@cosmosbay.com \
--cc=dipankar@in.ibm.com \
--cc=khali@linux-fr.org \
--cc=linux-kernel@vger.kernel.org \
--cc=manfred@colorfullife.com \
--cc=torvalds@osdl.org \
/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