From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.fusionio.com ([66.114.96.30]:35177 "EHLO mx1.fusionio.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758615Ab2JZMhO (ORCPT ); Fri, 26 Oct 2012 08:37:14 -0400 Date: Fri, 26 Oct 2012 08:37:11 -0400 From: Chris Mason To: Jan Schmidt CC: Chris Mason , linux-btrfs Subject: Re: BUG in btrfs_set_item_key_safe Message-ID: <20121026123711.GC23729@shiny> References: <508A2F3F.3030500@jan-o-sch.net> <20121026121941.GB23729@shiny> <508A824E.805@jan-o-sch.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" In-Reply-To: <508A824E.805@jan-o-sch.net> Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Fri, Oct 26, 2012 at 06:30:06AM -0600, Jan Schmidt wrote: > On Fri, October 26, 2012 at 14:19 (+0200), Chris Mason wrote: > > On Fri, Oct 26, 2012 at 12:35:43AM -0600, Jan Schmidt wrote: > >> Hi, > >> > >> while running extensive qgroup and tree mod log tests I got to the following > >> BUG, which is probably not related to tree mod log: > > > > So our keys are out of order...are you able to reproduce this? > > No reproducer yet. My tests ran fine for about 15 hours before that happened. > I'll have an extensive amount of (tree mod log related) debug information if I > should get there again. > > My guess is that it's not related to tree mod log, though, as there are tree log > functions (fs/btrfs/tree-log.c) on the stack. I don't think rewinded buffers can > make it there. > > I'll call back when I'll hit this again. This is more likely to be memory corruption. I'd suggest turning on slab debug and CONFIG_DEBUG_PAGE_ALLOC...hopefully you'll get a faster crash. -chris