From: Alasdair G Kergon Subject: Re: Re: dm: bounce_pfn limit added Date: Tue, 30 Oct 2007 23:26:17 +0000 Message-ID: <20071030232617.GJ10006@agk.fab.redhat.com> References: <47257E4B.4080904@sw.ru> <20071030131138.3ba59731.akpm@linux-foundation.org> Reply-To: device-mapper development Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20071030131138.3ba59731.akpm@linux-foundation.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: Vasily Averin Cc: linux-kernel@vger.kernel.org, dm-devel@redhat.com, Jens Axboe , Andrew Morton , stable@kernel.org, devel@openvz.org List-Id: dm-devel.ids On Tue, Oct 30, 2007 at 01:11:38PM -0700, Andrew Morton wrote: > On Mon, 29 Oct 2007 09:31:39 +0300 > Vasily Averin wrote: > > Device mapper uses its own bounce_pfn that may differ from one on underlying > > device. In that way dm can build incorrect requests that contain sg elements > > greater than underlying device is able to handle. > > This is the cause of slab corruption in i2o layer, occurred on i386 arch when > > very long direct IO requests are addressed to dm-over-i2o device. Interesting. When this patch was originally proposed a couple of years ago, we did this instead: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=daef265f1590cf3e6de989d074041a280c82d58b;hp=00d6da9b4d6707b808481372537adb0fb38f99b3 ) DM doesn't need to bounce bio's on its own, but the block layer defaults ) to that in blk_queue_make_request(). The lower level drivers should ) bounce ios themselves, that is what they need to do if not layered below ) dm anyways. Alasdair -- agk@redhat.com