From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: aic7xxx timer handling bug Date: Mon, 09 Jan 2006 08:56:21 -0600 Message-ID: <1136818581.3364.3.camel@mulgrave> References: <20060109040606.118a0d33.akpm@osdl.org> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from stat9.steeleye.com ([209.192.50.41]:6272 "EHLO hancock.sc.steeleye.com") by vger.kernel.org with ESMTP id S964779AbWAIO4Z (ORCPT ); Mon, 9 Jan 2006 09:56:25 -0500 In-Reply-To: <20060109040606.118a0d33.akpm@osdl.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Andrew Morton Cc: linux-scsi@vger.kernel.org On Mon, 2006-01-09 at 04:06 -0800, Andrew Morton wrote: > While doing a binary search for a buggy patch (it was > gregkh-pci-x86-pci-domain-support-the-meat.patch, reported on > linux-kernel), I hit the below. This is a hard one to solve thanks to Justin and his idiot my driver knows best policies. The problem is essentially that recovery can take time and the mid-layer will get impatient and use a bigger hammer. Justin, therefore, tries to hold of the timeout until recovery has finished (which can apparently take forever in his driver). Just pulling out the timer adjustment might be a welcome cosmetic fix, but it won't solve the underlying problem: we'll have the mid-layer thinking the command is aborted and the aic driver still trying to process it. James