From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:48776) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RUePd-0001ik-5o for qemu-devel@nongnu.org; Sun, 27 Nov 2011 08:04:54 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1RUePc-0000fP-82 for qemu-devel@nongnu.org; Sun, 27 Nov 2011 08:04:53 -0500 Received: from mx1.redhat.com ([209.132.183.28]:33059) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RUePb-0000fC-VJ for qemu-devel@nongnu.org; Sun, 27 Nov 2011 08:04:52 -0500 Message-ID: <4ED2356E.6070505@redhat.com> Date: Sun, 27 Nov 2011 15:04:46 +0200 From: Avi Kivity MIME-Version: 1.0 References: <20111124160355.7182f296@doriath> In-Reply-To: <20111124160355.7182f296@doriath> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [BUG 1.0] -cpu host breaks linux guest boot List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Luiz Capitulino , Anthony Liguori Cc: qemu-devel On 11/24/2011 08:03 PM, Luiz Capitulino wrote: > Avi, > > I'm getting the following panic very early on boot on a F16 VM (latest > qemu.git head 40897c9): > > [ 0.068018] Kernel panic - not syncing: Boot APIC ID in local APIC unexpected (-1 vs 0) > [ 0.070006] Pid: 1, comm: swapper Not tainted 3.1.0-7.fc16.x86_64 #1 > [ 0.071004] Call Trace: > [ 0.072035] [] panic+0x91/0x1a5 > [ 0.073027] [] native_smp_prepare_cpus+0x24c/0x300 > [ 0.075010] [] kernel_init+0x85/0x158 > [ 0.075993] [] ? schedule_tail+0x27/0x6c > [ 0.076017] [] kernel_thread_helper+0x4/0x10 > [ 0.077022] [] ? start_kernel+0x3b6/0x3b6 > [ 0.078009] [] ? gs_change+0x13/0x13 > > After some debugging and I found out that this only happens when '-cpu host' > is passed, and according to git bisect it's caused by the following commit: > > commit 66e3dd9282141b5ae75637c9676002cf3ceeb988 > Author: Avi Kivity > Date: Wed Nov 9 15:44:36 2011 +0200 > > i386: derive '-cpu host' from KVM_GET_SUPPORTED_CPUID > Strange, since it looks like cpuid 1 (ebx) returns the correct APIC ID. Anthony, I think we should revert that commit and refactor cpuid for 1.1. The logic is spread over too many places which makes it hard to reason about. -- error compiling committee.c: too many arguments to function