From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S271119AbTG1VwT (ORCPT ); Mon, 28 Jul 2003 17:52:19 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S271120AbTG1VwT (ORCPT ); Mon, 28 Jul 2003 17:52:19 -0400 Received: from amsfep16-int.chello.nl ([213.46.243.26]:46644 "EHLO amsfep16-int.chello.nl") by vger.kernel.org with ESMTP id S271119AbTG1VwR (ORCPT ); Mon, 28 Jul 2003 17:52:17 -0400 From: Ben Twijnstra Reply-To: bentw@chello.nl To: linux-kernel@vger.kernel.org Subject: Machine Check Exception in 2.6.0-test2 Date: Mon, 28 Jul 2003 23:52:15 +0200 User-Agent: KMail/1.5.2 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200307282352.15775.bentw@chello.nl> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi all, When running a fairly CPU and thread-hungry application under 2.5.75, 2.6.0-test1 and 2.6.0-test2 I'm getting the following error messages on the console: Jul 25 00:16:31 durak kernel: CPU 0: Machine Check Exception: 0000000000000004 Jul 25 00:16:31 durak kernel: Bank 0: b600000000000175 at 0000000025a0cff8 Jul 25 00:16:31 durak kernel: Kernel panic: CPU context corrupt Once I get this message, one of the threads becomes unkillable yet remains soaking up CPU cycles. This does not happen in the 2.4 kernels. How do I go about finding the source of the problem? The address indicated in the second line indicates that this happened somewhere in userspace (0x25a0cff8), but since this is a commercial package I don't have any symbolic info. This last address changes between runs, so I assume it's a virtual address. Any ideas welcome. Best regards, Ben