From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1765858AbXLRAXh (ORCPT ); Mon, 17 Dec 2007 19:23:37 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1759701AbXLRAX1 (ORCPT ); Mon, 17 Dec 2007 19:23:27 -0500 Received: from ebiederm.dsl.xmission.com ([166.70.28.69]:41006 "EHLO ebiederm.dsl.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756724AbXLRAX0 (ORCPT ); Mon, 17 Dec 2007 19:23:26 -0500 From: ebiederm@xmission.com (Eric W. Biederman) To: Andi Kleen Cc: Neil Horman , Ben Woodard , Neil Horman , kexec@lists.infradead.org, Andi Kleen , linux-kernel@vger.kernel.org, hbabu@us.ibm.com, Yinghai Lu , hpa@zytor.com, tglx@linutronix.de, mingo@redhat.com Subject: Re: [PATCH] kexec: force x86_64 arches to boot kdump kernels on boot cpu References: <20071211192434.GD10999@hmsreliant.think-freely.org> <86802c440712111151t29acd38kf9fac8e41743f3e4@mail.gmail.com> <20071211205955.GF10999@hmsreliant.think-freely.org> <20071212142132.GC4889@bingen.suse.de> <20071212155515.GA29735@hmsendeavour.rdu.redhat.com> <20071212160722.GD4889@bingen.suse.de> <20071212202215.GB29735@hmsendeavour.rdu.redhat.com> <20071213143922.GA4674@hmsreliant.think-freely.org> <20071213151629.GB31269@one.firstfloor.org> Date: Mon, 17 Dec 2007 17:13:25 -0700 In-Reply-To: <20071213151629.GB31269@one.firstfloor.org> (Andi Kleen's message of "Thu, 13 Dec 2007 16:16:29 +0100") Message-ID: User-Agent: Gnus/5.110006 (No Gnus v0.6) Emacs/21.4 (gnu/linux) 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 Andi Kleen writes: > On Thu, Dec 13, 2007 at 09:39:22AM -0500, Neil Horman wrote: >> >> Ok, new patch attached, taking into account Andi's request for a cleaner > method > > Sorry for not noticing that earlier, but was there a specific reason this needs > to be an early quirk at all? kexec can only happen after the standard quirks > ran. > I think it should be fine as a standard "late" quirk. Just to document things. The important thing is this quirk happens before calibrate_delay(). Which is still before the normal pci subsystem gets initialized. So that seems to require an early_quirk as the pci subsystem is not initialized by that point. The only case we are likely to hit this is kdump because BIOS almost always boot us on the cpu with apic id == 0. However in the case of this bug if we happen to boot on a cpu with apic id >= 16 we won't be able to boot the linux kernel either, because calibrate_delay will fail. Eric