From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jim Paris Subject: Re: Disk stuck in error recovery loop with AHCI Date: Sat, 24 Feb 2007 03:00:39 -0500 Message-ID: <20070224080039.GA25323@jim.sh> References: <20070221052022.GA15964@jim.sh> <20070223072826.GA2763@jim.sh> <45DFBD9A.6090204@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from NEUROSIS.MIT.EDU ([18.95.3.133]:36123 "EHLO neurosis.jim.sh" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933422AbXBXIAl (ORCPT ); Sat, 24 Feb 2007 03:00:41 -0500 Content-Disposition: inline In-Reply-To: <45DFBD9A.6090204@gmail.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Tejun Heo Cc: linux-ide@vger.kernel.org Tejun Heo wrote: > Jim Paris wrote: > > Still, it seems that some improvements could be made to the EH when > > this sort of thing happens. For example, after "speed down requested > > but no transfer mode left" a few times in a row, maybe it would make > > sense to just fail the disk and give up. That would have allowed > > higher layers like MD to recover. > > That's SD retrying sector-by-sector a large request as libata doesn't > report proper failed sector info. Please try 2.6.20. md will fail the > drive after a few errors. Hi Tejun, It was 2.6.20. The system was stuck repeating EH for about half an hour before I physically pulled the disk to get MD to fail the device and to bring the system back to a responsive state. > Patches to improve EH behavior further went into 2.6.21 devel tree and > some are pending. Great, I've hopefully eliminated the cause of my errors but I'll give that a try if I run into this again. -jim