From: Jan Kara Subject: Re: locking typo in ext4_mb_add_n_trim() Date: Tue, 31 Mar 2009 15:00:32 +0200 Message-ID: <20090331130032.GE20127@atrey.karlin.mff.cuni.cz> References: <49CCDFDB.9000109@redhat.com> <20090327233634.GI5176@mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Eric Sandeen , Dan Carpenter , adilger@sun.com, linux-ext4@vger.kernel.org To: Theodore Tso Return-path: Received: from atrey.karlin.mff.cuni.cz ([195.113.26.193]:41124 "EHLO atrey.karlin.mff.cuni.cz" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754023AbZCaNAe (ORCPT ); Tue, 31 Mar 2009 09:00:34 -0400 Content-Disposition: inline In-Reply-To: <20090327233634.GI5176@mit.edu> Sender: linux-ext4-owner@vger.kernel.org List-ID: > On Fri, Mar 27, 2009 at 09:16:59AM -0500, Eric Sandeen wrote: > > > > although I wonder why we don't trip over this in spinlock debugging > > (seems like it'd lead to a double unlock at times) I wonder if we can > > tie this to any other bugs we've seen. > > I was wondering if it could be tied to the "rm -rf" soft lockup hang.... > > I wonder if vendor kernels (specifically, Ubuntu in this case) disable > spinlock debugging, which is why we wouldn't have seen the double > unlock warning. (Or maybe it happened earlier in the log, and users > didn't notice it). Don't know about Ubuntu but SUSE has definitely turned off spinlock debugging (I guess because it may cost some performance) and most other debug stuff. Honza -- Jan Kara SuSE CR Labs