linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: mengcong <mc@linux.vnet.ibm.com>
To: Al Viro <viro@ZenIV.linux.org.uk>
Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	Nick Piggin <npiggin@kernel.dk>,
	Dave Chinner <david@fromorbit.com>
Subject: Re: [PATCH] VFS: br_write_lock locks on possible CPUs other than online CPUs
Date: Mon, 19 Dec 2011 14:07:46 +0800	[thread overview]
Message-ID: <1324274866.25089.32.camel@mengcong> (raw)
In-Reply-To: <20111219050037.GO23662@dastard>

On Mon, 2011-12-19 at 16:00 +1100, Dave Chinner wrote:
> On Mon, Dec 19, 2011 at 04:11:42AM +0000, Al Viro wrote:
> > On Mon, Dec 19, 2011 at 11:36:15AM +0800, mengcong wrote:
> > > In a heavily loaded system, when frequently turning on and off CPUs, the
> > > kernel will detect soft-lockups on multiple CPUs. The detailed bug report
> > > is at https://lkml.org/lkml/2011/8/24/185.
> > > 
> > > The root cause is that brlock functions, i.e. br_write_lock() and
> > > br_write_unlock(), only locks/unlocks the per-CPU spinlock of CPUs that
> > > are online, which means, if one online CPU is locked and then goes
> > > offline, any later unlocking operation happens during its offline state
> > > will not touch it; and when it goes online again, it has the incorrect
> > > brlock state. This has been verified in current kernel.
> > > 
> > > I can reproduce this bug on the intact 3.1 kernel. After my patch applied, 
> > > I've ran an 8-hours long test(test script provided by the bug reporter), 
> > > and no soft lockup happened again.
> > 
> > Argh...  OK, that's seriously nasty.  I agree that this is broken, but
> > your patch makes br_write_lock() very costly on kernels build with
> > huge number of possible CPUs, even when it's run on a box with few
> > CPUs ;-/
> 
> I fixed this problem with the XFS per-cpu superblock counters
> (bit lock + counters in per-cpu structs) back in 2006. It basically
> uses a lglock-like local/global locking structure and iterates them
> using for_each_online_cpu().
> 
> I fixed it simply by registering a hotplug notifier and
> draining/reinitialising counters on the appropriate event under a
> global lock context. i.e. make CPU hotplug serialise against
> concurrent lock operations. See commit e8234a68 ("[XFS] Add support
> for hotplug CPUs...")
> 

how about to register a cpu hotplug notifier to align the brlock as what
Dave did?

> Cheers,
> 
> Dave.

  reply	other threads:[~2011-12-19  6:07 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-12-19  3:36 [PATCH] VFS: br_write_lock locks on possible CPUs other than online CPUs mengcong
2011-12-19  4:11 ` Al Viro
2011-12-19  5:00   ` Dave Chinner
2011-12-19  6:07     ` mengcong [this message]
2011-12-19  7:31 ` Srivatsa S. Bhat
2011-12-19  9:12   ` Stephen Boyd
2011-12-19 11:03     ` Srivatsa S. Bhat
2011-12-19 12:11       ` Al Viro
2011-12-19 20:23         ` Srivatsa S. Bhat
2011-12-19 20:52           ` Al Viro
2011-12-20  4:56             ` Srivatsa S. Bhat
2011-12-20  6:27               ` Al Viro
2011-12-20  7:28                 ` Srivatsa S. Bhat
2011-12-20  9:37                   ` mengcong
2011-12-20 10:36                     ` Srivatsa S. Bhat
2011-12-20 11:08                       ` Srivatsa S. Bhat
2011-12-20 12:50                         ` Srivatsa S. Bhat
2011-12-20 14:06                           ` Al Viro
2011-12-20 14:35                             ` Srivatsa S. Bhat
2011-12-20 17:59                               ` Al Viro
2011-12-20 19:12                                 ` Srivatsa S. Bhat
2011-12-20 19:58                                   ` Al Viro
2011-12-20 22:27                                     ` Dave Chinner
2011-12-20 23:31                                       ` Al Viro
2011-12-21 21:15                                     ` Srivatsa S. Bhat
2011-12-21 22:02                                       ` Al Viro
2011-12-21 22:12                                       ` Andrew Morton
2011-12-22  7:02                                         ` Al Viro
2011-12-22  7:20                                           ` Andrew Morton
2011-12-22  8:08                                             ` Al Viro
2011-12-22  8:17                                               ` Andi Kleen
2011-12-22  8:39                                                 ` Al Viro
2011-12-22  8:22                                             ` Andi Kleen
2011-12-20  7:30                 ` mengcong
2011-12-20  7:37                   ` Srivatsa S. Bhat
2011-12-19 23:56         ` Dave Chinner
2011-12-20  4:05           ` Al Viro

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=1324274866.25089.32.camel@mengcong \
    --to=mc@linux.vnet.ibm.com \
    --cc=david@fromorbit.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=npiggin@kernel.dk \
    --cc=viro@ZenIV.linux.org.uk \
    /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).