From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263478AbTIHSsW (ORCPT ); Mon, 8 Sep 2003 14:48:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263482AbTIHSsV (ORCPT ); Mon, 8 Sep 2003 14:48:21 -0400 Received: from chaos.analogic.com ([204.178.40.224]:1152 "EHLO chaos.analogic.com") by vger.kernel.org with ESMTP id S263478AbTIHSsU (ORCPT ); Mon, 8 Sep 2003 14:48:20 -0400 Date: Mon, 8 Sep 2003 14:48:10 -0400 (EDT) From: "Richard B. Johnson" X-X-Sender: root@chaos Reply-To: root@chaos.analogic.com To: Pavel Machek cc: Linux kernel Subject: Re: Some read-errors on floppys not reported on 2.4.22 In-Reply-To: <20030904171758.GR1358@openzaurus.ucw.cz> Message-ID: References: <20030904171758.GR1358@openzaurus.ucw.cz> 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 On Thu, 4 Sep 2003, Pavel Machek wrote: > Hi! > > > Success, even where there are lots of CRC errors that > > prematurely terminate the read: > > Can you find out if it works in 2.4.21? > -- > Pavel Okay. It works on 2.4.21, the code is identical to 2.4.20 which also works. On 2.4.22, bad CRC errors which terminate the read, don't always result in a bad return code. Some just return 0, which is treated like EOF in user-mode code. Application software has to add a call to fstat() to see if the bytes read were equal to the file size as a work-around. This step should not be required. Cheers, Dick Johnson Penguin : Linux version 2.4.22 on an i686 machine (794.73 BogoMips). Note 96.31% of all statistics are fiction.