From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262470AbVD2XXN (ORCPT ); Fri, 29 Apr 2005 19:23:13 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263058AbVD2XXN (ORCPT ); Fri, 29 Apr 2005 19:23:13 -0400 Received: from e32.co.us.ibm.com ([32.97.110.130]:33532 "EHLO e32.co.us.ibm.com") by vger.kernel.org with ESMTP id S262470AbVD2XXH (ORCPT ); Fri, 29 Apr 2005 19:23:07 -0400 Subject: Re: [Ext2-devel] [RFC] Adding multiple block allocation From: Mingming Cao Reply-To: cmm@us.ibm.com To: Badari Pulavarty Cc: suparna@in.ibm.com, Andrew Morton , "Stephen C. Tweedie" , linux-kernel , ext2-devel , linux-fsdevel@vger.kernel.org In-Reply-To: <427280C1.8090404@us.ibm.com> References: <1113220089.2164.52.camel@sisko.sctweedie.blueyonder.co.uk> <1113244710.4413.38.camel@localhost.localdomain> <1113249435.2164.198.camel@sisko.sctweedie.blueyonder.co.uk> <1113288087.4319.49.camel@localhost.localdomain> <1113304715.2404.39.camel@sisko.sctweedie.blueyonder.co.uk> <1113348434.4125.54.camel@dyn318043bld.beaverton.ibm.com> <1113388142.3019.12.camel@sisko.sctweedie.blueyonder.co.uk> <1114207837.7339.50.camel@localhost.localdomain> <1114659912.16933.5.camel@mindpipe> <1114715665.18996.29.camel@localhost.localdomain> <20050429135211.GA4539@in.ibm.com> <427280C1.8090404@us.ibm.com> Content-Type: text/plain Organization: IBM LTC Date: Fri, 29 Apr 2005 16:22:59 -0700 Message-Id: <1114816980.10473.90.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.0.2 (2.0.2-3) Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2005-04-29 at 11:45 -0700, Badari Pulavarty wrote: > I touch tested your patch earlier and seems to work fine. Lets integrate > Mingming's getblocks() patches with this and see if we get any benifit > from the whole effort. > I tried Suparna's mpage_writepages_getblocks patch with my ext3_get_blocks patch, seems to work fine, except that still only one block is allocated at a time. I got a little confused.... I did not see any delayed allocation code in your patch, I assume you have to update ext3_prepare_write to not call ext3_get_block, so that block allocation will be defered at ext3_writepages time. So without the delayed allocation part, the get_blocks in mpage_writepages is doing multiple blocks look up only, right? Mingming