In case this was missed, re-posting here.
Haven't found the message Daniel referenced, so don't know if it is
something XFS is doing to trigger this race condition or not.
-- Attached file included as plaintext by Ecartis --
-- Desc: Forwarded message - Re: MM patches against 2.5.31
Received: (qmail 12115 invoked from network); 29 Aug 2002 00:20:24 +0100
Received: from virgil.dstl.gov.uk (126.96.36.199) by taz.dera.gov.uk with
SMTP; 29 Aug 2002 00:20:24 +0100
Received: from relay.dera.gov.uk (not verified[188.8.131.52]) by
virgil.dstl.gov.uk with MailMarshal (4,2,5,0) id <B000064574>; Thu, 29
2002 00:20:26 +0100
X-Proxy: warlock.dstl.gov.uk protected by Firewall
Received: (qmail 11831 invoked by uid 501); 29 Aug 2002 00:20:27 +0100
Received: (qmail 11825 invoked from network); 29 Aug 2002 00:20:26 +0100
Received: from vger.kernel.org (184.108.40.206) by relay.dera.gov.uk with
SMTP; 29 Aug 2002 00:20:26 +0100
Received: (majordomo@xxxxxxxxxxxxxxx) by vger.kernel.org via listexpand id
<S319056AbSH1XNt>; Wed, 28 Aug 2002 19:13:49 -0400
Received: (majordomo@xxxxxxxxxxxxxxx) by vger.kernel.org id
<S319059AbSH1XNs>; Wed, 28 Aug 2002 19:13:48 -0400
Received: from dsl-213-023-022-149.arcor-ip.net ([220.127.116.11]:4045 "EHLO
starship") by vger.kernel.org with ESMTP id <S319056AbSH1XMk>; Wed, 28
2002 19:12:40 -0400
Received: from daniel by starship with local (Exim 3.35 #1 (Debian)) id
17kBkK-0002uI-00; Thu, 29 Aug 2002 00:57:08 +0200
Content-Type: text/plain; charset=US-ASCII
From: Daniel Phillips <phillips@xxxxxxxx>
To: Andrew Morton <akpm@xxxxxxxxxx>
Subject: Re: MM patches against 2.5.31
Date: Thu, 29 Aug 2002 00:57:06 +0200
X-Mailer: KMail [version 1.3.2]
Cc: Christian Ehrhardt <ehrhardt@xxxxxxxxxxxxxxxxxxxxx>, lkml
<linux-kernel@xxxxxxxxxxxxxxx>, "linux-mm@xxxxxxxxx" <linux-mm@xxxxxxxxx>
References: <3D644C70.6D100EA5@xxxxxxxxxx> <E17kAvf-0002tx-00@starship>
On Thursday 29 August 2002 00:39, Andrew Morton wrote:
> Daniel Phillips wrote:
> > ...
> > So there's no question that the race is lurking in 2.4. I noticed several
> > more paths besides the one above that look suspicious as well. The bottom
> > line is, 2.4 needs a fix along the lines of my suggestion or Christian's,
> > something that can actually be proved.
> > It's a wonder that this problem manifests so rarely in practice.
> I sort-of glanced through the 2.4 paths and it appears that in all of the
> places where it could do a page_cache_get/release, that would never happen
> because of other parts of the page state.
> Like: it can't be in pagecache, so we won't run writepage, and
> it can't have buffers, so we won't run try_to_release_page().
> Of course, I might have missed a path. And, well, generally: ugh.
I think it is happening. I just went sifting searching through the archives
on 'oops' and '2.4'. The first one I found was:
2.4.18-xfs (xfs related?) oops report
which fits the description nicely.
The race I showed actually causes the page->count to go negative, avoiding
a double free on a technicality. That doesn't make me feel much better about
it. Have you got a BUG_ON(!page_count(page)) in put_page_testzero? I think
we might see some action.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/