From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932794AbYEGTIA (ORCPT ); Wed, 7 May 2008 15:08:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1761554AbYEGTHW (ORCPT ); Wed, 7 May 2008 15:07:22 -0400 Received: from www.tglx.de ([62.245.132.106]:39598 "EHLO www.tglx.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759340AbYEGTHT (ORCPT ); Wed, 7 May 2008 15:07:19 -0400 Date: Wed, 7 May 2008 21:06:57 +0200 (CEST) From: Thomas Gleixner To: Mirco Tischler cc: Andi Kleen , Ingo Molnar , lkml Subject: Re: [BUG]:WARNING: at arch/x86/kernel/rtc.c:122 in 2.6.26-rc1 In-Reply-To: <1210183536.28099.10.camel@mtlp.tischlernetz> Message-ID: References: <1210183536.28099.10.camel@mtlp.tischlernetz> User-Agent: Alpine 1.10 (LFD 962 2008-03-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 7 May 2008, Mirco Tischler wrote: > Hello > I've triggered this warning in 2.6.26-rc1-git from 6th of may: > ... > [11677.490019] CPU1 is down > [11677.490149] Extended CMOS year: 2000 > [11677.490149] x86 PAT enabled: cpu 0, old 0x7040600070406, new > 0x7010600070106 > [11677.490747] Back to C! > [11677.490952] microcode: CPU0 updated from revision 0x92 to 0x92, date > = 03132007 > [11677.491065] ------------[ cut here ]------------ > [11677.491065] WARNING: at arch/x86/kernel/rtc.c:122 mach_get_cmos_time > +0x9e/0x14e() > [11677.491065] [] mach_get_cmos_time+0x9e/0x14e > [11677.491065] [] read_persistent_clock+0x1b/0x35 > [11677.491065] [] timekeeping_resume+0xb/0x120 Interesting. That's in the resume path. > I wasn't able to reproduce the warning yet, but because the commit > message introducing the WARN_ON says: This is a WARN_ON_ONCE so you only see it once per boot. Resume does not count :) You might change that WARN_ON_ONCE to WARN_ON to check whether it happens more than once. > > [ tglx@linutronix.de: changed to WARN_ON - we want to know that! > > If no one reports it we can remove the complete if (RTC_ALWAYS_BCD) > > magic, which has RTC_ALWAYS_BCD defined to 1 since Linux 1.0 ... ] > I'm reporting it anyway now. Thanks. That's why I put the WARN_ON there. The backtrace attracts attention and it's also an interesting info that this did not happen on boot, where we read out the RTC as well. Can you please provide some details about the system ? Thanks, tglx