From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753558Ab0G1GOc (ORCPT ); Wed, 28 Jul 2010 02:14:32 -0400 Received: from mail.vyatta.com ([76.74.103.46]:56865 "EHLO mail.vyatta.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753324Ab0G1GO3 (ORCPT ); Wed, 28 Jul 2010 02:14:29 -0400 Date: Tue, 27 Jul 2010 23:14:27 -0700 From: Stephen Hemminger To: Ben Greear Cc: linux-kernel@vger.kernel.org Subject: Re: 2.6.34.1+hacks: kernel deadlock related to file-IO? Message-ID: <20100727231427.1f3eb97b@nehalam> In-Reply-To: <4C4F767D.4020002@candelatech.com> References: <4C4F67B3.50101@candelatech.com> <20100727165146.71e46170@nehalam> <4C4F767D.4020002@candelatech.com> Organization: Vyatta X-Mailer: Claws Mail 3.7.5 (GTK+ 2.20.1; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 27 Jul 2010 17:14:53 -0700 Ben Greear wrote: > On 07/27/2010 04:51 PM, Stephen Hemminger wrote: > > On Tue, 27 Jul 2010 16:11:47 -0700 > > Ben Greear wrote: > > > >> NOTE: This is from a hacked 2.6.34.1 kernel, with > >> a module of my own devising loaded and in active use. > > > > Ben please don't ask LKML for help with problems when > > you have a binary only module loaded... > > Yeah, I know..but I spent all day hacking .34 enough to > run lockdep and get those backtraces..and they *might* be > valid. There's a small chance someone else has seen similar > things but not reported it for whatever reason, so maybe > they will chime in. > > And finally, I will attempt to reproduce with some official > kernels, and will report what I find one way or another. > > The .34 kernel has been a real bitch to get stable (You > know things must be bad when my name's on -stable patches), > so I wouldn't be surprised if I'm hitting another real bug. Understand your pain. The hardest problem is often finding a reproducible case. If the problem is not reproducible it is hard to add trace info or use bisection.