From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alan Cox Subject: Re: [PATCH 11/16] libata-eh-fw: implement new EH scheduling via timeout Date: Sat, 29 Apr 2006 22:13:12 +0100 Message-ID: <1146345193.3302.32.camel@localhost.localdomain> References: <1144762974340-git-send-email-htejun@gmail.com> <443D8106.4000607@pobox.com> <443DBA20.9050508@gmail.com> <443DC2E8.8000005@pobox.com> <443DC751.8090209@gmail.com> <4450AC03.303@pobox.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from outpipe-village-512-1.bc.nu ([81.2.110.250]:9447 "EHLO lxorguk.ukuu.org.uk") by vger.kernel.org with ESMTP id S1750770AbWD2VDe (ORCPT ); Sat, 29 Apr 2006 17:03:34 -0400 In-Reply-To: <4450AC03.303@pobox.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Jeff Garzik Cc: Tejun Heo , axboe@suse.de, albertcc@tw.ibm.com, lkosewsk@gmail.com, linux-ide@vger.kernel.org On Iau, 2006-04-27 at 07:33 -0400, Jeff Garzik wrote: > Well overall I want to minimize the change from existing behavior, since > there is enough change as it is. In current libata and drivers/ide, it > can continue after a properly indicated DMA error without doing a reset. Sometimes. But the smarts in it also break some hardware horribly because the properly indicated DMA error doesn't neccessarily leave the state machines on the chip sane, and worse it sometimes does resets then recovery and thats ugly. > And I just think its rude, to reset hardware when it doesn't need > resetting. Its sorta like the Windows mentality -- "just reboot it, > that will fix stuff" Linux shouldn't be rude to hardware, when it need > not be :) Windows generally uses a 7 second timeout btw.