From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1030500AbWF1KLp (ORCPT ); Wed, 28 Jun 2006 06:11:45 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932567AbWF1KLp (ORCPT ); Wed, 28 Jun 2006 06:11:45 -0400 Received: from outpipe-village-512-1.bc.nu ([81.2.110.250]:4228 "EHLO lxorguk.ukuu.org.uk") by vger.kernel.org with ESMTP id S932538AbWF1KLo (ORCPT ); Wed, 28 Jun 2006 06:11:44 -0400 Subject: Re: [PATCH] watchdog: add pretimeout ioctl From: Alan Cox To: Corey Minyard Cc: Linux Kernel , Andrew Morton , OpenIPMI Developers , Wim Van Sebroeck In-Reply-To: <44A1D5E5.3020903@acm.org> References: <20060627182225.GD10805@localdomain> <1151434785.32186.56.camel@localhost.localdomain> <44A1D5E5.3020903@acm.org> Content-Type: text/plain Content-Transfer-Encoding: 7bit Date: Wed, 28 Jun 2006 11:27:18 +0100 Message-Id: <1151490438.15166.8.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.6.2 (2.6.2-1.fc5.5) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Ar Maw, 2006-06-27 am 20:05 -0500, ysgrifennodd Corey Minyard: > since even if the machine is hard-locked the NMI will come through. Don't bet on that. There are a considerable number of crash stall cases where NMI will not be delivered, plus you can disable NMI if you are really sneaky and want to. The usual case NMI fails is crashed video cards. The PCI transaction stalls forever on many boards and the CPU therefore never finishes the instruction and never services the NMI Alan