From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kiszka Subject: Re: [PATCH v2] kvm: Disable MSI/MSI-X in assigned device reset path Date: Mon, 16 Apr 2012 18:38:22 +0200 Message-ID: <4F8C4AFE.8010309@siemens.com> References: <20120405033450.25661.68810.stgit@bling.home> <1334584997.3112.8.camel@bling.home> <20120416150640.GB13713@redhat.com> <4F8C364F.6090901@siemens.com> <20120416160823.GA18914@redhat.com> <4F8C450F.9030706@siemens.com> <20120416163624.GD18914@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Alex Williamson , "kvm@vger.kernel.org" , Avi Kivity , Marcelo Tosatti , "jbaron@redhat.com" To: "Michael S. Tsirkin" Return-path: Received: from goliath.siemens.de ([192.35.17.28]:20171 "EHLO goliath.siemens.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752516Ab2DPQi0 (ORCPT ); Mon, 16 Apr 2012 12:38:26 -0400 In-Reply-To: <20120416163624.GD18914@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On 2012-04-16 18:36, Michael S. Tsirkin wrote: > On Mon, Apr 16, 2012 at 06:13:03PM +0200, Jan Kiszka wrote: >> On 2012-04-16 18:08, Michael S. Tsirkin wrote: >>> On Mon, Apr 16, 2012 at 05:10:07PM +0200, Jan Kiszka wrote: >>>> On 2012-04-16 17:06, Michael S. Tsirkin wrote: >>>>> On Mon, Apr 16, 2012 at 08:03:17AM -0600, Alex Williamson wrote: >>>>>> The discussion on this patch seems to have fizzled, with no clear short >>>>>> term solution. >>>>> >>>>> I think we are in concensus, it's just that there are >>>>> multiple bugs still left to fix. >>>>> >>>>> First, we need to prevent guest from touching command >>>>> register except for the bus master bit. Something like >>>> >>>> INTx is harmless as well. Denying access breaks masking for the guest. >>> >>> We currently mark intx as emulated, no? >> >> If we use the mask also in the kernel for IRQ sharing, we *monitor* it. >> >>> Anyway - care fixing it properly? >> >> There is nothing to fix. >> >> Jan > > Guest access to command register is a bug. > E.g. it has no business enabling SERR as SERR crashes > host when triggered. Maybe a misunderstanding. I agree that other bits are not guest business. But INTx should be excluded from this filtering and handled as is, i.e. passed through. Jan -- Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux