From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: dm-thinp bug Date: Wed, 27 Apr 2011 04:39:17 -0400 Message-ID: <20110427083917.GA23330@infradead.org> References: <20110426184742.GA13880@infradead.org> <1303892857.4679.2.camel@ubuntu> Reply-To: device-mapper development Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <1303892857.4679.2.camel@ubuntu> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: Joe Thornber Cc: Christoph Hellwig , dm-devel@redhat.com, ejt@redhat.com List-Id: dm-devel.ids On Wed, Apr 27, 2011 at 09:27:37AM +0100, Joe Thornber wrote: > On Tue, 2011-04-26 at 14:47 -0400, Christoph Hellwig wrote: > > I can reproduce a bug in virtio triggered by running under dm-thinp by > > just copying lots of files from my home directory into a freshly created > > filesystem on dm-thinp. > > Finally someone has tried using it! I'll look into this today. I actually used the earlier version you posted and didn't reproduce it there. I have a gut feeling it's related to barriers, as I probably just did a basic ext3 test back then. One thing that springs to mind is that both the thinp and multisnap targets don't appear to support empty barriers (that is REQ_FLUSH without a payload), because they don't set ti->num_flush_requests, but I'm not sure that's related.