linux-kernel - Re: [3.13.2] list passed to list_sort() too long for efficiency.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20140220070907.GP13647@dastard>
Date:	Thu, 20 Feb 2014 18:09:07 +1100
From:	Dave Chinner <david@...morbit.com>
To:	Joe Perches <joe@...ches.com>
Cc:	pawel.sikora@...k.net, linux-kernel@...r.kernel.org
Subject: Re: [3.13.2] list passed to list_sort() too long for efficiency.

On Wed, Feb 12, 2014 at 10:07:46AM -0800, Joe Perches wrote:
> On Wed, 2014-02-12 at 18:59 +0100, Paweł Sikora wrote:
> > On Wednesday 12 of February 2014 09:46:26 Joe Perches wrote:
> > > On Wed, 2014-02-12 at 18:16 +0100, Paweł Sikora wrote:
> > > > [16855.582522] list passed to list_sort() too long for efficiency
> > > > 
> > > > could someone put some light on the last line?
> > > 
> > > Did you look at the list_sort function in lib/list_sort.c?
> > > 
> > > (granted the dmesg is harder to find that it should be
> > >  because the format is split)
> > > 
> > > It just means it's slower than it could otherwise be
> > 
> > thanks,
> > so without a stacktrace in dmesg we can't isolate a root of cause :/
> > 
> 
> Perhaps something like this would help:
> 
> Emit the caller of list_sort using %pS when the list is too
> long to be efficient.  Also emit the message if necessary
> once for each call.

Still not useful. For XFS, the list_sort caller isn't directly of
interest - it's further down the stack where the interesting
information is (e.g. caller flushing buffers from log recovery,
metadata writeback, quota shrinker, quotacheck, etc).  Only a full
stack trace will tell us what we need to know.

I think WARN_ON_ONCE() is probably a better solution here.

Cheers,

Dave.

-- 
Dave Chinner
david@...morbit.com
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/