From: Tejun Heo <htejun@gmail.com>
To: Oleg Nesterov <oleg@tv-sign.ru>
Cc: Andrew Morton <akpm@linux-foundation.org>,
David Chinner <dgc@sgi.com>, David Howells <dhowells@redhat.com>,
Gautham Shenoy <ego@in.ibm.com>, Jarek Poplawski <jarkao2@o2.pl>,
Ingo Molnar <mingo@elte.hu>,
Srivatsa Vaddagiri <vatsa@in.ibm.com>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH] make cancel_rearming_delayed_work() reliable
Date: Fri, 11 May 2007 17:19:01 +0200 [thread overview]
Message-ID: <46448965.7070500@gmail.com> (raw)
In-Reply-To: <20070511134714.GA191@tv-sign.ru>
Oleg Nesterov wrote:
>> and this. After grabbing cwq lock, compare it to get_wq_data() first,
>> if they are the same, both are using the same lock so there's no
>> reason for the barriers. If they are different, it doesn't matter
>> anyway. We need to retry the locking.
>
> I think this is not right. The problem is that work->data (its
> WORK_STRUCT_WQ_DATA_MASK part, cwq) could be changed _without_ holding
> the old cwq->lock.
>
> Suppose that CPU_0 does queue_delayed_work(dwork). We start the timer,
> work->data points to cwq_0. CPU_DEAD comes, the timer migrates to
> CPU_1, but work->data was not changed yet.
>
>> retry:
>> cwq = get_sw_data(work);
>> if (!cwq)
>> return ret;
>>
>> spin_lock_irq(&cwq->lock);
>> if (unlikely(cwq != get_wq_data(work))) {
>> /* oops wrong cwq */
>> spin_unlock_irq(&cwq->lock);
>> goto retry; /* or just return 0; */
>> }
>
> dwork->timer fires, delayed_work_timer_fn() changes work->data to cwq_1
> and queues this work on cwq_1.
>
>> if (!list_empty(&work->entry)) {
>> list_del_init(&work->entry);
>
> Oops! we are holding cwq_0->lock, but modify cwq_1->worklist.
>
> Actually, we have the same problem with a plain queue_work() if _cpu_down()
> comes at "right" time.
>
> However, I agree, this smp_wmb() in insert_work() should die. We can
> introduce "smp_mb__before_spinlock()" (no-op on x86 at least) to kill it.
Yeah, right, we allow cwq pointer to change without holding the lock.
Although I still think that is where we should fix the problem. Taking
down CPU is a cold cold path. We can afford a lot of overhead there.
IMHO, if we can do that, it would be far better than memory barrier
dance which tends to be difficult to understand and thus prove/maintain
correctness. I'll think about it more.
>>> + * It is possible to use this function if the work re-queues itself. It can
>>> + * cancel the work even if it migrates to another workqueue, however in that
>>> + * case it only garantees that work->func() has completed on the last queued
>>> + * workqueue.
>> We first prevent requeueing from happening; then, wait for each cwq to
>> finish the work, so I think we're guaranteed that they're finished on
>> all cpus. Otherwise, the 'sync' part isn't too useful as it means all
>> rearming tasks might be running on completion of cancel_work_sync().
>
> Yes, sure, you are right. What I meant was:
>
> struct workqueue_struct *WQ1, *WQ1;
>
> void work_func(struct work_struct *self)
> {
> // migrate on another workqueue
> queue_work(WQ2, self);
>
> do_something();
> }
>
> queue_work(WQ1, work);
>
> Now, cancel_work_sync() can't guarantee that this work has finished
> its execution on WQ1.
Right again, I misunderstood that the migration was between cwqs.
Things definitely won't work if it's between different wqs.
Thanks.
--
tejun
next prev parent reply other threads:[~2007-05-11 15:19 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-05-03 20:42 Oleg Nesterov
2007-05-04 1:15 ` Andrew Morton
2007-05-04 17:09 ` Oleg Nesterov
2007-05-05 21:32 ` [PATCH] make-cancel_rearming_delayed_work-reliable-fix Oleg Nesterov
2007-05-07 10:31 ` Jarek Poplawski
2007-05-07 10:34 ` Oleg Nesterov
2007-05-07 11:55 ` Anton Vorontsov
2007-05-07 11:33 ` Oleg Nesterov
2007-05-08 9:16 ` Jarek Poplawski
2007-05-08 12:02 ` Oleg Nesterov
2007-05-08 13:07 ` Jarek Poplawski
2007-05-08 7:15 ` [PATCH] make cancel_rearming_delayed_work() reliable Jarek Poplawski
2007-05-08 12:31 ` Oleg Nesterov
2007-05-08 13:56 ` Jarek Poplawski
2007-05-08 14:05 ` Oleg Nesterov
2007-05-08 14:32 ` Jarek Poplawski
2007-05-08 14:12 ` Jarek Poplawski
2007-05-11 13:55 ` Tejun Heo
2007-05-11 13:47 ` Oleg Nesterov
2007-05-11 15:19 ` Tejun Heo [this message]
2007-05-11 14:53 ` Oleg Nesterov
2007-05-12 5:50 ` Tejun Heo
2007-05-13 19:27 ` Oleg Nesterov
2007-05-13 20:16 ` Tejun Heo
2007-05-13 21:25 ` Oleg Nesterov
2007-05-14 19:44 ` Oleg Nesterov
2007-05-15 8:26 ` Tejun Heo
2007-05-15 13:09 ` Jarek Poplawski
2007-05-15 22:08 ` Oleg Nesterov
2007-05-16 5:21 ` Jarek Poplawski
2007-05-15 22:00 ` Oleg Nesterov
2007-05-16 11:25 ` Tejun Heo
2007-05-16 18:52 ` Oleg Nesterov
2007-05-17 9:36 ` Tejun Heo
2007-05-18 7:35 ` Jarek Poplawski
2007-05-18 8:13 ` Jarek Poplawski
2007-05-18 13:33 ` Tejun Heo
2007-05-21 7:00 ` Jarek Poplawski
2007-05-21 8:59 ` Tejun Heo
2007-05-21 10:10 ` Jarek Poplawski
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=46448965.7070500@gmail.com \
--to=htejun@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=dgc@sgi.com \
--cc=dhowells@redhat.com \
--cc=ego@in.ibm.com \
--cc=jarkao2@o2.pl \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=oleg@tv-sign.ru \
--cc=vatsa@in.ibm.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
Powered by JetHome