From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Mon, 21 May 2001 05:40:02 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Mon, 21 May 2001 05:39:52 -0400 Received: from ppp0.ocs.com.au ([203.34.97.3]:57869 "HELO mail.ocs.com.au") by vger.kernel.org with SMTP id ; Mon, 21 May 2001 05:39:41 -0400 X-Mailer: exmh version 2.1.1 10/15/1999 From: Keith Owens To: kees cc: linux-kernel@vger.kernel.org Subject: Re: Hang with SMP 2.4.4 snd log In-Reply-To: Your message of "Mon, 21 May 2001 10:26:20 +0200." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Mon, 21 May 2001 19:39:23 +1000 Message-ID: <9363.990437963@ocs3.ocs-net> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 21 May 2001 10:26:20 +0200 (CEST), kees wrote: >I got a next hang with my SMP system, kdb log attached. Something strange >with the backtrace for CPU 0. Here is the first cut from the kdb log.. I do not trust either of those backtraces. There is no way to get from do_nmi to do_exit. The presence of "unknown" entries indicates that the cpu 0 back trace is bad. Also all the ebp pointers are suspect, they are way out of range for the task addresses. You could be looking at a stack overrun or just random corruption of kernel data. If you can reproduce the problem, set KDBDEBUG=0xff and bt. That will debug kdb and produce a lot of output. Send it to me, although I suspect it will just prove that you have stack corruption.