From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756654Ab0E0Gyy (ORCPT ); Thu, 27 May 2010 02:54:54 -0400 Received: from mga05.intel.com ([192.55.52.89]:51024 "EHLO fmsmga101.fm.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1754089Ab0E0Gyx (ORCPT ); Thu, 27 May 2010 02:54:53 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.53,310,1272870000"; d="scan'208";a="802305318" Message-ID: <4BFE1735.9080806@linux.intel.com> Date: Thu, 27 May 2010 08:54:45 +0200 From: Andi Kleen User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.9.1.9) Gecko/20100317 Thunderbird/3.0.4 MIME-Version: 1.0 To: Huang Ying CC: Jin Dongming , LKLM , Hidetoshi Seto Subject: Re: [Patch-next] Remove notify_die in do_machine_check functioin References: <4BFDDBA9.4010702@np.css.fujitsu.com> <1274930481.3444.258.camel@yhuang-dev.sh.intel.com> In-Reply-To: <1274930481.3444.258.camel@yhuang-dev.sh.intel.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > I have heard about that on some machine, some hardware error output pin > of chipset may be linked with some input pin of CPU which can cause MCE. Yes that happens. > That is, MCE is used to report some chipset errors too. I think that is > why notify_die is called in do_machine_check. Simply removing notify_die > is not good for these machines. In general deciding what to do on a MCE is rather complicated and probably too much for any die handler. > Maybe we should fix the notifier user instead. Which notifier user > consumes the DIE_NMI notification? Yes. It would be good to find out which user it is. Perhaps gdb? One approach would be to give it a different type (DIE_MCE) But today we don't really need it. notify_die() is primarily for debuggers of all kinds, and I never liked the idea to call a debugger on a machine check. So I would be ok with just removing the call. -Andi