From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265740AbUADR7K (ORCPT ); Sun, 4 Jan 2004 12:59:10 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S265765AbUADR7K (ORCPT ); Sun, 4 Jan 2004 12:59:10 -0500 Received: from cpc3-hitc2-5-0-cust116.lutn.cable.ntl.com ([81.99.82.116]:55976 "EHLO zog.reactivated.net") by vger.kernel.org with ESMTP id S265740AbUADR7G (ORCPT ); Sun, 4 Jan 2004 12:59:06 -0500 Message-ID: <3FF8558A.2080308@reactivated.net> Date: Sun, 04 Jan 2004 18:03:54 +0000 From: Daniel Drake User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6b) Gecko/20031208 Thunderbird/0.4 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Herve Fache Cc: linux-kernel@vger.kernel.org Subject: Re: Linux hangs on nVidia nForce2 400 Ultra References: <20040104173624.19046.qmail@iname.com> In-Reply-To: <20040104173624.19046.qmail@iname.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi, Theres been some work done on this issue lately. Originally, my system only hung when I had IO-APIC and local APIC support compiled into the kernel. But I think other people had hanging problems even without this. You could first try using a 2.6-mm series kernel, as these include some patches which fix nforce2 problems for most people (but not me). The other option is to try Ross Dickinson's patches: http://marc.theaimsgroup.com/?l=linux-kernel&m=107199838022614&w=2 If you use Ross's patches, as well as the -mm fixes, you should revert out the nforce-disconnect-quirk patch to avoid your CPU getting needlessly hot. Personally, I have stability running 2.6.1-rc1-mm1 with the disconnect-quirk reverted out, Ross's patches, booting with parameter apic_tack=2. I have APIC/IOAPIC compiled in and working. Daniel. Herve Fache wrote: > Hi chaps! > > My system hangs (no oops, nothing) on disk access using either 2.4.23 or 2.6.0. A rather reliable way for me to > crash it is to, for example, copy the sources of the Linux kernel. > > It hanged once on CD-ROM access, which could lead to a more IDE-level problem. Also, the only time it did it in > another operating system (humm...) was after it crashed in Linux and I pressed reset (no shut down), which > makes me think it could the IDE controller [driver]'s fault... > > It seems I'm the only one on Earth to have this problem (according to Google), but if there was a way to track it > down I'd be happy to try. > > I have attached my 2.6.0 kernel config for info. > > Thanks! > Hervé.