From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: CPU softlockup due to smp_call_function() Date: Thu, 05 Apr 2012 15:51:00 +0300 Message-ID: <4F7D9534.90604@redhat.com> References: <4F7D8F16.9090904@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Peter Zijlstra , Thomas Gleixner , Ingo Molnar , "H. Peter Anvin" , Dave Jones , kvm@vger.kernel.org, "linux-kernel@vger.kernel.org List" To: Sasha Levin Return-path: Received: from mx1.redhat.com ([209.132.183.28]:50202 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752549Ab2DEMvO (ORCPT ); Thu, 5 Apr 2012 08:51:14 -0400 In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: On 04/05/2012 03:32 PM, Sasha Levin wrote: > > > > It would be good to enhance smp_call_function_*() to do this > > automatically when it happens - it's spinning there anyway, so it might > > as well count the iterations and NMI the lagging cpu if it waits for too > > long. > > What do you think about modifying the softlockup detector to NMI all > CPUs if it's going to panic because it detected a lockup? Fine by me. The triggering cpu should be dumped first though. We lose knowledge of exactly which cpu is hung, but it should usually be easily seen from the traces (the exceptional case would be a busy server that is handling a lot of IPIs at the moment we trigger the traces). -- error compiling committee.c: too many arguments to function