From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266677AbUBQVsA (ORCPT ); Tue, 17 Feb 2004 16:48:00 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266664AbUBQVpO (ORCPT ); Tue, 17 Feb 2004 16:45:14 -0500 Received: from fw.osdl.org ([65.172.181.6]:12767 "EHLO mail.osdl.org") by vger.kernel.org with ESMTP id S266652AbUBQVme (ORCPT ); Tue, 17 Feb 2004 16:42:34 -0500 Date: Tue, 17 Feb 2004 13:43:50 -0800 From: Andrew Morton To: Florian Schanda Cc: linux-kernel@vger.kernel.org Subject: Re: 99% System load Message-Id: <20040217134350.0da17784.akpm@osdl.org> In-Reply-To: <200402172116.18254.ma1flfs@bath.ac.uk> References: <200402111423.02217.ma1flfs@bath.ac.uk> <20040211234413.3d90df5d.akpm@osdl.org> <200402172116.18254.ma1flfs@bath.ac.uk> X-Mailer: Sylpheed version 0.9.7 (GTK+ 1.2.10; i586-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Florian Schanda wrote: > > Finally it happened again. I attached the output of that dmesg. > > > It would help if you could make a note of the PIDs of the hung processes, > > so they can be correlated with the sysrq output. > > The PIDs hanging are 3303, 3305, 3306, 3307. It _looks_ like all four CPUs are madly taking timer interrupts. Which tends to point at some APIC problem, failing to clear the interrupt source. But if that happened one wouldn't expect userspace to remain in a runnable state, and you say that you can still run commands. > An advance warning: I had the nvidia driver (yes, I know, closed source makes > debugging impossible) loaded, and before the trace begins, there are some > messages of that module: It is probably unrelated, but I'd suggest that you not use the nvidia driver for a while, verify that the lockup happens without it loaded.