Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751129AbWEWOiO (ORCPT ); Tue, 23 May 2006 10:38:14 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750892AbWEWOiO (ORCPT ); Tue, 23 May 2006 10:38:14 -0400 Received: from cantor2.suse.de ([195.135.220.15]:19929 "EHLO mx2.suse.de") by vger.kernel.org with ESMTP id S1751042AbWEWOiN (ORCPT ); Tue, 23 May 2006 10:38:13 -0400 To: Mikael Pettersson Cc: linux-kernel@vger.kernel.org, mmlnx@us.ibm.com Subject: Re: nmi_watchdog default setting on i386 and x86_64 References: <44724DE3.2000209@us.ibm.com> <17523.6413.711397.401340@alkaid.it.uu.se> From: Andi Kleen Date: 23 May 2006 16:37:57 +0200 In-Reply-To: <17523.6413.711397.401340@alkaid.it.uu.se> Message-ID: User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1017 Lines: 19 Mikael Pettersson writes: > Mike Mason writes: > > Does anybody know the reasoning behind having nmi_watchdog turned off by > > default on i386 and on by default on x86_64. I've heard that i386 had > > problems with false positives in the past, but that local apic watchdog > > may make that concern obsolete. > > On i386 the problems are mainly hardware and BIOS. In particular, > lots of Dell laptops have capable hardware but broken BIOSen that > hang the machines if we try to enable anything sending performance > counter interrupts via the local APIC. AFAIK that trouble was mostly when you forced the local APIC on against the wishes of the BIOS. That was always a dumb idea and gladly Linux doesn't try that by default anymore. -Andi - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/