From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759582AbYAHPWM (ORCPT ); Tue, 8 Jan 2008 10:22:12 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754730AbYAHPV6 (ORCPT ); Tue, 8 Jan 2008 10:21:58 -0500 Received: from mail.gmx.net ([213.165.64.20]:59079 "HELO mail.gmx.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1754247AbYAHPV5 (ORCPT ); Tue, 8 Jan 2008 10:21:57 -0500 X-Authenticated: #14349625 X-Provags-ID: V01U2FsdGVkX1/NHkIiToLNIvL5uv8go4PgfFeibsoENQZwittuVf dYngtxvVLBOkv3 Subject: Re: [vm] writing to UDF DVD+RW (/dev/sr0) while under memory pressure: box ==> doorstop From: Mike Galbraith To: Andrew Morton Cc: LKML In-Reply-To: <20080108033801.40d0043a.akpm@linux-foundation.org> References: <1199447212.4529.13.camel@homer.simson.net> <1199612533.4384.54.camel@homer.simson.net> <1199642470.3927.12.camel@homer.simson.net> <20080106122954.d8f04c98.akpm@linux-foundation.org> <1199790316.4094.57.camel@homer.simson.net> <20080108033801.40d0043a.akpm@linux-foundation.org> Content-Type: text/plain Date: Tue, 08 Jan 2008 16:21:53 +0100 Message-Id: <1199805713.3571.12.camel@homer.simson.net> Mime-Version: 1.0 X-Mailer: Evolution 2.12.0 Content-Transfer-Encoding: 7bit X-Y-GMX-Trusted: 0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 2008-01-08 at 03:38 -0800, Andrew Morton wrote: > On Tue, 08 Jan 2008 12:05:16 +0100 Mike Galbraith wrote: > > Seeks on DVD drives are apparently far FAR more horrible that I ever > > imagined, so writing to a filesystem on a DVD+RW in effect constitutes > > the stupid thing I was worried about above. Make an image, populate it, > > write it instead... sure to be a truckload faster, and due to the VM > > connection, much safer too. Oh well. > > > > Well. From your earlier trace it appeared that something was causing > the filesystem to perform synchronous inode writes - sync_dirty_buffer() was > called. > > This will cause many more seeks than would occur if we were doing full > delayed writing, with obvious throughput implications. Yes, with UDF, the IO was _incredibly_ slow. With ext2, it was better, though still very bad. I tested with that other OS, and it gets ~same throughput with UDF as I got with ext2 (ick). UDF does udf_clear_inode() -> write_inode_now(inode, 1) I suppose I could try write_inode_now(inode, 0). Might unstick the box. -Mike