From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932561AbVHSFy4 (ORCPT ); Fri, 19 Aug 2005 01:54:56 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932565AbVHSFyz (ORCPT ); Fri, 19 Aug 2005 01:54:55 -0400 Received: from mail.dvmed.net ([216.237.124.58]:12507 "EHLO mail.dvmed.net") by vger.kernel.org with ESMTP id S932561AbVHSFyy (ORCPT ); Fri, 19 Aug 2005 01:54:54 -0400 Message-ID: <43057421.5090306@pobox.com> Date: Fri, 19 Aug 2005 01:54:41 -0400 From: Jeff Garzik User-Agent: Mozilla Thunderbird 1.0.6-1.1.fc4 (X11/20050720) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Tejun Heo CC: linux-ide@vger.kernel.org, linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org, Jens Axboe , Alan Cox Subject: Re: libata error handling References: <20050729050654.GA10413@havoc.gtf.org> <20050807054850.GA13335@htj.dyndns.org> <430556BF.5070004@pobox.com> <430570B5.60109@gmail.com> In-Reply-To: <430570B5.60109@gmail.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Score: 0.0 (/) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Tejun Heo wrote: > Heh... Maybe I'm just reluctant to let go of my patches. Anyways, I'll > now stand down and see how things go and try to help. Note that my email simply describes a long term target. For the short term, and perhaps medium term, libata will continue to use ->eh_strategy_handler(). Given Mark's messages, my own knowledge, and other reports, there continues to be room for improvement in the current EH code. In general, we need to distinguish between PCI bus errors, SATA bus errors, and ATA device errors, and handle each error class appropriately. In the SCSI layer, ->eh_strategy_handler() or no, this will likely consist of taking the SCSI device offline and dealing with the error(s). Jeff