From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267621AbUG3FyG (ORCPT ); Fri, 30 Jul 2004 01:54:06 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267620AbUG3FyG (ORCPT ); Fri, 30 Jul 2004 01:54:06 -0400 Received: from ns.virtualhost.dk ([195.184.98.160]:29862 "EHLO virtualhost.dk") by vger.kernel.org with ESMTP id S267621AbUG3Fxk (ORCPT ); Fri, 30 Jul 2004 01:53:40 -0400 Date: Fri, 30 Jul 2004 07:53:33 +0200 From: Jens Axboe To: Robert Love Cc: "Bryan O'Sullivan" , Arjan van de Ven , Dave Jones , Edward Angelo Dayao , linux-kernel@vger.kernel.org Subject: Re: Recent 2.6 kernels can't read an entire ATAPI CD or DVD Message-ID: <20040730055333.GC7925@suse.de> References: <20040728053107.GB11690@suse.de> <20040728065319.GD11690@suse.de> <20040728145228.GA9316@redhat.com> <20040728145543.GB18846@devserv.devel.redhat.com> <20040728163353.GJ10377@suse.de> <20040728170507.GK10377@suse.de> <1091051858.13651.1.camel@camp4.serpentine.com> <20040729084928.GR10377@suse.de> <1091166553.1982.9.camel@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1091166553.1982.9.camel@localhost> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Jul 30 2004, Robert Love wrote: > On Thu, 2004-07-29 at 10:49 +0200, Jens Axboe wrote: > > > Looks pretty perfect, maybe it's read-ahead screwing it up. Try if > > setting hdparm -a0 /dev/hdc makes a difference. > > I am seeing similar errors[1] on later 2.6, too, with about 50% of my > audio CD collection. Some work, some do not. I see no pattern. > > Already tried disabling read-ahead, does not matter. Also tried a > different drive. read-ahead doesn't matter on ripping audio, just for fs work. The audio problems might be fixed in 2.6.8-rc1-mm1 + the bounce patch posted a few times (unfortunately 2.6.8-rc2-mm1 drops the patch for some reason). > Robert Love > > [1] > hdc: command error: status=0x51 { DriveReady SeekComplete Error } > hdc: command error: error=0x54 > end_request: I/O error, dev hdc, sector 8 > Buffer I/O error on device hdc, logical block 1 So this happens during ripping, or? The problem with the errors like the above is that they don't indicate a specific problem. I can't just look at the error and see what causes it, it could be a million things. -- Jens Axboe