From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261575AbTLDC6O (ORCPT ); Wed, 3 Dec 2003 21:58:14 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261681AbTLDC6O (ORCPT ); Wed, 3 Dec 2003 21:58:14 -0500 Received: from mail.netzentry.com ([157.22.10.66]:23558 "EHLO netzentry.com") by vger.kernel.org with ESMTP id S261575AbTLDC6M (ORCPT ); Wed, 3 Dec 2003 21:58:12 -0500 Message-ID: <3FCEA2AD.5040805@netzentry.com> Date: Wed, 03 Dec 2003 18:57:49 -0800 From: "b@netzentry.com" Reply-To: b@netzentry.com Organization: b@netzentry.com User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: recbo@nishanet.com CC: linux-kernel@vger.kernel.org Subject: Re: NForce2 pseudoscience stability testing (2.6.0-test11) Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org >Bob wrote: >Local APIC locked up with nforce2 and VIA, >impossibly serious with nforce2 and non-amd >offboard ide controller cards. BIOS flash made >problems go away. > >I experienced the lockups when using promise and >siig sis ide hd controller pci cards. I still had problems >with a 3ware card. > >Flashing the bios solved all problems. Now I run >both the via and nforce2 mboards with APIC and >Local APIC on in kernel. I'm running six ide drives, >four on a 3ware pci hd controller card using ide-scsi. > >I got sound working on nforce2, and nvidia ti4200 >agp8 vid card(nvidia drivers crash X but agpgart >with X "nv" instead of "nvidia" works in 2D well), >but not usb. The sound config problem was fixed >by "ln -s sound/dsp2 /dev/dsp". The apps only >look for /dev/dsp. > >-Bob Do you think that motherboard maker was really at fault or did they genuinely fix a grotesque error. Were any changes in the BIOS-change-list relevant to fixing up this APIC problem in Linux? Can you provide the following: - which motherboard - which bios revs (the broken one and the fixed one) - which kernel are you running (is it vanilla, from a dist, recompiled) - lspci - cat /proc/interrupts - dmesg - .config from kernel (if not stock from dist) I really, really hope this problem can be solved without a BIOS upgrade because getting board manufacturers to do anything is very difficult.