From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S264943AbUAaRUE (ORCPT ); Sat, 31 Jan 2004 12:20:04 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264949AbUAaRUE (ORCPT ); Sat, 31 Jan 2004 12:20:04 -0500 Received: from mail.parknet.co.jp ([210.171.160.6]:33549 "EHLO mail.parknet.co.jp") by vger.kernel.org with ESMTP id S264943AbUAaRUB (ORCPT ); Sat, 31 Jan 2004 12:20:01 -0500 To: "H. Peter Anvin" Cc: linux-kernel@vger.kernel.org Subject: Re: PATCH to access old-style FAT fs References: <20040126173949.GA788@frodo.local> <20040128202443.GA9246@frodo.local> <87bron7ppd.fsf@devron.myhome.or.jp> <20040129105624.GA1072@frodo.local> From: OGAWA Hirofumi Date: Sun, 01 Feb 2004 02:19:54 +0900 In-Reply-To: Message-ID: <87hdycujb9.fsf@devron.myhome.or.jp> User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org "H. Peter Anvin" writes: > This is of course dangerous, since it could create conflicts. > > Thus, I suggest we do write new 0x00 markers; even though it's not > required (it only matters if the filesystem is out of spec and > therefore by definition corrupt), it would help the Psion, and > wouldn't cause any problems for in-spec filesystems. Yes, that cluster is dangerous. However, repair should be fsck's job. IMO, for repair we shouldn't add unneeded I/O to normal path. -- OGAWA Hirofumi