From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Lalancette Subject: Re: kvm problems on new hardware Date: Fri, 30 Oct 2009 09:10:33 +0100 Message-ID: <4AEA9F79.4070607@redhat.com> References: <20091026100651.GA15566@dth.net> <4AE6C126.5030803@redhat.com> <1256639276.3889.29.camel@thinkpaddth> <4AE6CCA1.9000307@redhat.com> <1256841444.4117.16.camel@thinkpaddth> <20091029220740.GI13808@us.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Danny ter Haar , Avi Kivity , kvm@vger.kernel.org To: Ryan Harper Return-path: Received: from mx1.redhat.com ([209.132.183.28]:54646 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754097AbZJ3IKf (ORCPT ); Fri, 30 Oct 2009 04:10:35 -0400 In-Reply-To: <20091029220740.GI13808@us.ibm.com> Sender: kvm-owner@vger.kernel.org List-ID: Ryan Harper wrote: > * Danny ter Haar [2009-10-29 13:38]: >> Update: >> I compiled/installed 2.6.32-rc5-git3 on this machine >> >> I manually start kvm: >> >> vhost1:~# kvm -m 512 -cdrom /vz/template/iso/debian-503-amd64-netinst.iso >> >> The bootscreen comes up, i hit enter to install and i get these messages >> (copied from dmesg) >> >> handle_exception: unexpected, vectoring info 0x80000010 intr info 0x80000b0d >> handle_exception: unexpected, vectoring info 0x8000000d intr info 0x80000b0d >> [this line is repeated many times: >> dmesg |grep 0x80000b0d | wc -l >> 570 ] >> and finally before ending the kvm session i get: >> vmx_handle_exit: unexpected, valid vectoring info (0x8000000d) and exit reason is 0x80000021 > > I've seen that in a couple places. I don't think we have root cause, > but in at least one situation (running win2k3 with > 4G of ram) the > work around was to use: Actually, I'm not 100% sure, but I might have root-caused this. At least, the exit reason is the same exit reason I fixed. That exit reason essentially means "vmenter failed because of invalid guest fields". In the case I tracked down, it was because we were wrongly sign-extending the segment fields (instead of zero-extending them). Can either you or Danny: 1) Get the output from stderr of qemu when this happens? I believe qemu dumps the state of all of the guest fields when it's going to crash like this, and that can tell us if the GUEST_STATE is wrong. 2) Try the patch posted here: http://lkml.org/lkml/2009/10/28/201, and see if it helps? -- Chris Lalancette