From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1762665AbXKIONY (ORCPT ); Fri, 9 Nov 2007 09:13:24 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1759912AbXKIONQ (ORCPT ); Fri, 9 Nov 2007 09:13:16 -0500 Received: from pat.uio.no ([129.240.10.15]:43465 "EHLO pat.uio.no" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759846AbXKIONQ (ORCPT ); Fri, 9 Nov 2007 09:13:16 -0500 Subject: Re: [OOPS] 2.6.23.1 in NFS From: Trond Myklebust To: Mark Hindley Cc: linux-kernel@vger.kernel.org, nfs@lists.sourceforge.net In-Reply-To: <20071109095558.GB29687@hindley.org.uk> References: <20071109095558.GB29687@hindley.org.uk> Content-Type: text/plain Date: Fri, 09 Nov 2007 09:16:14 -0500 Message-Id: <1194617774.7459.30.camel@heimdal.trondhjem.org> Mime-Version: 1.0 X-Mailer: Evolution 2.12.1 Content-Transfer-Encoding: 7bit X-UiO-Resend: resent X-UiO-ClamAV-Virus: No X-UiO-Spam-info: not spam, SpamAssassin (score=-0.2, required=12.0, autolearn=disabled, AWL=-0.160) X-UiO-Scanned: 459E35ADF549A6AEA593E9118E09489EBB210E9A X-UiO-SPAM-Test: remote_host: 129.240.10.9 spam_score: -1 maxlevel 200 minaction 2 bait 0 mail/h: 480 total 5031450 max/h 8345 blacklist 0 greylist 0 ratelimit 0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2007-11-09 at 09:55 +0000, Mark Hindley wrote: > I got this OOPS this morning on a K6. Sorry it is tainted by madwifi. > ath5k doesn't (yet!) support my card. Box is K6 200, headless, used as > firewall/router > > It looks like the same codepath as http://lkml.org/lkml/2007/11/6/275 > but my kernel was definitely compiled on a local disk. You are hitting the BUG_ON() in line 909, whereas Mathieu was hitting something further up in the same routine. However I agree that they look suspicious. Both look as if something is corrupting the rpc_task structure. In your case, a debugging tag that is only ever touched by the RPC code at the very start and very finish of the call is being changed, which points at something like a use-after-free issue. Hmm... I note that you are both running with the SLUB allocator. Any chance you can reproduce using SLAB (and with the SLAB debugging enabled)? Trond