From mboxrd@z Thu Jan 1 00:00:00 1970 From: Antonios Motakis Subject: Re: [Android-virt] [PATCH v5 11/13] ARM: KVM: Support SMP hosts Date: Mon, 19 Dec 2011 16:42:05 +0100 Message-ID: <4EEF5B4D.70501@virtualopensystems.com> References: <20111211102403.21693.6887.stgit@localhost> <20111211102529.21693.62306.stgit@localhost> <4EEED68D.5040307@virtualopensystems.com> <4EEF55EF.9050309@arm.com> <4EEF58A8.30004@virtualopensystems.com> <4EEF5A20.7060603@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Marc Zyngier , "tech@virtualopensystems.com" , "avi@redhat.com" , "android-virt@lists.cs.columbia.edu" , "kvm@vger.kernel.org" To: Christoffer Dall Return-path: Received: from mail-ey0-f174.google.com ([209.85.215.174]:39813 "EHLO mail-ey0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750734Ab1LSPmJ (ORCPT ); Mon, 19 Dec 2011 10:42:09 -0500 Received: by eaaj10 with SMTP id j10so4868331eaa.19 for ; Mon, 19 Dec 2011 07:42:08 -0800 (PST) In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: On 12/19/2011 04:40 PM, Christoffer Dall wrote: > On Mon, Dec 19, 2011 at 10:37 AM, Marc Zyngier wrote: >> On 19/12/11 15:30, Antonios Motakis wrote: >>> On 12/19/2011 04:19 PM, Marc Zyngier wrote: >>>> On 19/12/11 14:57, Christoffer Dall wrote: >>>>> >>>>> You should simply start booting a UP guest on an SMP host, see where >>>>> it crashes and start tracking it down. >>>> For the time being, I've yet to see UP guest crashing on SMP host. On >>>> the model, that is... >>>> >>>> >>> Last time I tried to run a guest in a 2 core model, it hanged and >>> crashed after a while. Anyway, I will investigate. So I gather critical >>> sections have been dealt with and it's just a matter of ironing bugs >>> right now? >> Depends when you tested. V4 was definitely unusable on SMP. >> >> With my patches merged in Christoffer's v5 release, the basics should be >> covered (correct SMP setup in the boot-wrapper and KVM, MPIDR >> virtualization). >> >> If you find any problem (and let's face it, I'm sure you will... ;-), >> I'll be happy to investigate. >> > I had the guest boot process crash on a page fault on the v5 series. > Consistently. I just didn't have time to investigate yet. I was also running v5, however I didn't record the exact crash behavior, since I assumed I was having the same results as everyone. I will look into it now.