From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261937AbUKHRbA (ORCPT ); Mon, 8 Nov 2004 12:31:00 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261971AbUKHROL (ORCPT ); Mon, 8 Nov 2004 12:14:11 -0500 Received: from smtp-4.vancouver.ipapp.com ([216.152.192.207]:59798 "EHLO envitech.com") by vger.kernel.org with ESMTP id S261937AbUKHQlh (ORCPT ); Mon, 8 Nov 2004 11:41:37 -0500 Message-ID: <418FA1DB.4060207@envitech.com> Date: Mon, 08 Nov 2004 11:42:03 -0500 From: Richard Brunelle Reply-To: rbrunelle@envitech.com Organization: Envitech Automation User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: Calibrating delay crash Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Rcpt-To: X-Country: CA Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org (please put me in CC when replying) I'm having trouble booting a PentiumMMX @ 266 MHz board with 64 Mbytes of RAM. The kernel crashes while in the main.c/calibrate_delay function. It never boot, sometimes it fails and reboot, sometimes a page fault occured. I tried with the kernels 2.4.21, 2.4.26 and 2.6.7 with the same results. I notice that if I return before the calibration loop, the system boots fine. This has been tried only for testing. void __init calibrate_delay(void) { unsigned long ticks, loopbit; int lps_precision = LPS_PREC; loops_per_jiffy = (1<<12); return; <<<<<<<<<<<<<<<<<<<<<<<<<<<<<< printk("Calibrating delay loop... "); .... I look in the archive and one post points to a memory problem, I ran memtest86 with no error found. Now I wonder what can cause this problem, is it a hardware issue/timing? What can I do to investigate more? How the calibrate_delay loop can crash the kernel? Thanks.