From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751356AbWF3HqA (ORCPT ); Fri, 30 Jun 2006 03:46:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751365AbWF3HqA (ORCPT ); Fri, 30 Jun 2006 03:46:00 -0400 Received: from mail.clusterfs.com ([206.168.112.78]:34194 "EHLO mail.clusterfs.com") by vger.kernel.org with ESMTP id S1751356AbWF3Hp7 (ORCPT ); Fri, 30 Jun 2006 03:45:59 -0400 Date: Fri, 30 Jun 2006 01:45:58 -0600 From: Andreas Dilger To: sho@tnes.nec.co.jp Cc: cmm@us.ibm.com, jitendra@linsyssoft.com, linux-kernel@vger.kernel.org, ext2-devel@lists.sourceforge.net Subject: Re: [RFC][PATCH]add ext3_fileblk_t for file relative offset Message-ID: <20060630074558.GP5318@schatzie.adilger.int> Mail-Followup-To: sho@tnes.nec.co.jp, cmm@us.ibm.com, jitendra@linsyssoft.com, linux-kernel@vger.kernel.org, ext2-devel@lists.sourceforge.net References: <20060628223109sho@rifu.tnes.nec.co.jp> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20060628223109sho@rifu.tnes.nec.co.jp> User-Agent: Mutt/1.4.1i X-GPG-Key: 1024D/0D35BED6 X-GPG-Fingerprint: 7A37 5D79 BF1B CECA D44F 8A29 A488 39F5 0D35 BED6 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Jun 28, 2006 22:31 +0900, sho@tnes.nec.co.jp wrote: > In Mingming's patch set, the single type `ext3_fsblk_t' deal with two > different values, which are file relative offset and filesystem > relative offset. I think it is confusing. > > Therefore I added "ext3_fileblk_t" for file relative offset. > It makes maintenance easier and code clearer if file size > needs to be larger. > ext3_fileblk_t is unsigned long which is the maximum size of > current type for file relative offset. I think this is a good idea and support its inclusion into the patch set. It also managed to find one (harmless) bug in the code. > int ext3_get_blocks_handle(handle_t *handle, struct inode *inode, > - sector_t iblock, unsigned long maxblocks, > + ext3_fileblk_t iblock, unsigned long maxblocks, iblock shouldn't have been a sector_t. Cheers, Andreas -- Andreas Dilger Principal Software Engineer Cluster File Systems, Inc.