From mboxrd@z Thu Jan 1 00:00:00 1970 From: Theodore Ts'o Subject: Re: [PATCH] xfstests-bld: exclude ext4 defrag tests from unsupported configurations Date: Sun, 14 Jun 2015 21:14:33 -0400 Message-ID: <20150615011433.GA15793@thunk.org> References: <20150610181319.GA23530@localhost.localdomain> <20150610192810.GB24181@thunk.org> <20150611200051.GA505@localhost.localdomain> <20150612002421.GB505@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-ext4@vger.kernel.org To: Eric Whitney Return-path: Received: from imap.thunk.org ([74.207.234.97]:35024 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752472AbbFOBOf (ORCPT ); Sun, 14 Jun 2015 21:14:35 -0400 Content-Disposition: inline In-Reply-To: <20150612002421.GB505@localhost.localdomain> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Jun 11, 2015 at 08:24:21PM -0400, Eric Whitney wrote: > Spoke too soon - hit what looks like the old problem on ARM running 4.1-rc7 > on the ninth trial run. Various components of the test are left in the 'D' > state after a familiar oops - kernel BUG at fs/jbd2/transaction.c:2150 while > truncating. > > I'll see if I can still get the problem to appear on x86_64. I was able to reproduce this on x86_64, after running a kernel build on the host as an antogonist. I'm guessing that the race window is pretty narrow, so that's why it wasn't triggering for me easily until I had some additional I/O traffic to slow things down further. After some investigation, it looks like this bug has been around since 2008(!). I'll send out a fix. - Ted