From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Oleg Nesterov <oleg@tv-sign.ru>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
Ingo Molnar <mingo@elte.hu>,
Christoph Hellwig <hch@infradead.org>,
Andrew Morton <akpm@osdl.org>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/7] barrier: a scalable synchonisation barrier
Date: Wed, 31 Jan 2007 22:48:21 +0100 [thread overview]
Message-ID: <1170280101.10924.36.camel@lappy> (raw)
In-Reply-To: <20070131211340.GA171@tv-sign.ru>
On Thu, 2007-02-01 at 00:13 +0300, Oleg Nesterov wrote:
> On 01/31, Paul E. McKenney wrote:
> >
> > On Sun, Jan 28, 2007 at 04:24:35PM +0100, Ingo Molnar wrote:
> > >
> > > * Christoph Hellwig <hch@infradead.org> wrote:
> > >
> > > > On Sun, Jan 28, 2007 at 12:51:21PM +0100, Peter Zijlstra wrote:
> > > > > This barrier thing is constructed so that it will not write in the
> > > > > sync() condition (the hot path) when there are no active lock
> > > > > sections; thus avoiding cacheline bouncing. -- I'm just not sure how
> > > > > this will work out in relation to PI. We might track those in the
> > > > > barrier scope and boost those by the max prio of the blockers.
> > > >
> > > > Is this really needed? We seem to grow new funky locking algorithms
> > > > exponentially, while people already have a hard time understanding the
> > > > existing ones.
> > >
> > > yes, it's needed.
> >
> > Would it be possible to come up with something common between this primitive
> > and the one that Oleg Nesterov put together for Jens Axboe?
> >
> > http://lkml.org/lkml/2006/11/29/330
> >
> > Oleg's approach acquires a lock on the update side, which Peter would
> > not want in the uncontended case -- but perhaps there is some way to
> > make Oleg's approach be able to safely test both counters so as to
> > avoid acquiring the lock if there are no readers.
> >
> > Oleg, any chance of this working? I believe it does, but have not
> > thought it through fully.
>
> I think no. From the quick reading, barrier_sync() and qrcu/srcu are
> quite different. Consider:
>
> barrier_lock()
>
> barrier_sync();
>
> barrier_unlock();
> ... wake up ...
> barrier_lock();
>
> schedule again
>
> The last "schedule again" would be a BUG for qrcu/srcu, but probably
> it is ok for barrier_sync().
Yes, that would be ok.
> It looks like barrier_sync() is more a
> rw semaphore biased to readers.
Indeed, the locked sections are designed to be the rare case.
> A couple of minor off-topic notes,
>
> +static inline void barrier_unlock(struct barrier *b)
> +{
> + smp_wmb();
> + if (atomic_dec_and_test(&b->count))
> + __wake_up(&b->wait, TASK_INTERRUPTIBLE|TASK_UNINTERRUPTIBLE, 0, b);
>
> This is wake_up_all(&b->wait), yes? I don't undestans why key == b, it could be NULL.
>
> +static inline void barrier_sync(struct barrier *b)
> +{
> + might_sleep();
> +
> + if (unlikely(atomic_read(&b->count))) {
> + DEFINE_WAIT(wait);
> + prepare_to_wait(&b->wait, &wait, TASK_UNINTERRUPTIBLE);
> + while (atomic_read(&b->count))
> + schedule();
> + finish_wait(&b->wait, &wait);
> + }
> +}
>
> This should be open-coded wait_event(), but wrong! With the scenario above this
> can hang forever! because the first wake_up removes the task from the &b->wait.
This would be me struggling with the waitqueue API, its all a tad
confusing at first look.
next prev parent reply other threads:[~2007-01-31 21:48 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-01-28 11:51 [PATCH 0/7] breaking the global file_list_lock Peter Zijlstra
2007-01-28 11:51 ` [PATCH 1/7] lockdep: lock_set_subclass - reset a held locks subclass Peter Zijlstra
2007-01-28 11:51 ` [PATCH 3/7] barrier: a scalable synchonisation barrier Peter Zijlstra
2007-01-28 14:39 ` Christoph Hellwig
2007-01-28 15:24 ` Ingo Molnar
2007-01-28 15:34 ` Christoph Hellwig
2007-01-31 19:12 ` Paul E. McKenney
2007-01-31 21:13 ` Oleg Nesterov
2007-01-31 21:30 ` Oleg Nesterov
2007-01-31 21:48 ` Peter Zijlstra [this message]
2007-01-31 23:32 ` Paul E. McKenney
2007-02-01 0:03 ` Peter Zijlstra
2007-02-01 0:48 ` Paul E. McKenney
2007-02-01 16:00 ` Oleg Nesterov
2007-02-01 21:38 ` Paul E. McKenney
2007-02-02 11:56 ` Oleg Nesterov
2007-02-02 12:01 ` Peter Zijlstra
2007-02-02 17:13 ` Paul E. McKenney
2007-02-03 16:38 ` Oleg Nesterov
2007-02-04 0:23 ` Paul E. McKenney
2007-02-04 3:24 ` Alan Stern
2007-02-04 5:46 ` Paul E. McKenney
2007-01-28 11:51 ` [PATCH 4/7] fs: break the file_list_lock for sb->s_files Peter Zijlstra
2007-01-28 14:43 ` Christoph Hellwig
2007-01-28 15:21 ` Ingo Molnar
2007-01-28 15:30 ` Christoph Hellwig
2007-01-28 15:32 ` Peter Zijlstra
2007-01-28 15:36 ` Christoph Hellwig
2007-01-28 15:44 ` Ingo Molnar
2007-01-28 16:25 ` Bill Huey
2007-01-28 11:51 ` [PATCH 5/7] fs: restore previous sb->s_files iteration semantics Peter Zijlstra
2007-01-28 11:51 ` [PATCH 6/7] schedule_on_each_cpu_wq() Peter Zijlstra
2007-01-28 11:51 ` [PATCH 7/7] fs: fixup filevec_add_drain_all Peter Zijlstra
2007-01-28 12:16 ` [PATCH 8/7] fs: free_write_pipe() fix Peter Zijlstra
2007-01-28 14:43 ` [PATCH 0/7] breaking the global file_list_lock Christoph Hellwig
2007-01-28 15:11 ` Christoph Hellwig
2007-01-28 15:29 ` Peter Zijlstra
2007-01-28 15:33 ` Christoph Hellwig
2007-01-29 13:32 ` Stephen Smalley
2007-01-29 18:02 ` Christoph Hellwig
2007-01-28 15:24 ` Ingo Molnar
2007-01-28 16:52 ` Martin J. Bligh
2007-01-28 17:04 ` lockmeter Christoph Hellwig
2007-01-28 17:38 ` lockmeter Martin J. Bligh
2007-01-28 18:01 ` lockmeter Bill Huey
2007-01-28 19:26 ` lockmeter Ingo Molnar
2007-01-28 21:17 ` lockmeter Ingo Molnar
2007-01-29 5:27 ` lockmeter Bill Huey
2007-01-29 10:26 ` lockmeter Bill Huey
2007-01-29 1:08 ` lockmeter Arjan van de Ven
2007-01-29 1:12 ` lockmeter Martin J. Bligh
2007-01-28 18:41 ` [PATCH 0/7] breaking the global file_list_lock Ingo Molnar
2007-01-28 20:38 ` Christoph Hellwig
2007-01-28 21:05 ` Ingo Molnar
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=1170280101.10924.36.camel@lappy \
--to=a.p.zijlstra@chello.nl \
--cc=akpm@osdl.org \
--cc=hch@infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=oleg@tv-sign.ru \
--cc=paulmck@linux.vnet.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;
as well as URLs for NNTP newsgroup(s).