From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:52205) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fcYk1-0001N7-GR for qemu-devel@nongnu.org; Mon, 09 Jul 2018 12:10:22 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fcYjw-0008MB-BE for qemu-devel@nongnu.org; Mon, 09 Jul 2018 12:10:21 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:52306 helo=mx1.redhat.com) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fcYjw-0008LS-2R for qemu-devel@nongnu.org; Mon, 09 Jul 2018 12:10:16 -0400 Date: Mon, 9 Jul 2018 18:10:11 +0200 From: Cornelia Huck Message-ID: <20180709181011.3edc4193.cohuck@redhat.com> In-Reply-To: <20180709131419.GB6271@rkaganb.sw.ru> References: <20180629221907.3662-1-venu.busireddy@oracle.com> <20180702161404.GA2339@rkaganb.sw.ru> <449f1449-ddf6-cd95-976c-14d04d8d503a@oracle.com> <20180703095825.GC30904@rkaganb.sw.ru> <20180703142817.GA3088@vbusired-vm> <20180703165200.180c93bb.cohuck@redhat.com> <20180704141523.377cb66d.cohuck@redhat.com> <20180706155406.0587733c.cohuck@redhat.com> <20180709131419.GB6271@rkaganb.sw.ru> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [virtio-dev] Re: [PATCH v3 0/3] Use of unique identifier for pairing virtio and passthrough devices... List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Roman Kagan Cc: Siwei Liu , Venu Busireddy , si-wei liu , "Michael S . Tsirkin" , Marcel Apfelbaum , virtio-dev@lists.oasis-open.org, qemu-devel@nongnu.org On Mon, 9 Jul 2018 16:14:19 +0300 Roman Kagan wrote: > On Fri, Jul 06, 2018 at 03:54:06PM +0200, Cornelia Huck wrote: > > If we want to extend that to other device types, we can add the field > > in their config space; but I'd rather prefer a more generic "host > > relays config metainformation" approach. > > The problem with this approach is that it's too generic and poorly > scoped. As a result I guess it's likely to take ages to converge. I'd certainly not want to wait for it for this particular issue; however, I think there is value in having this (seeing that at least Open Stack and s390x DPM already make use of such a mechanism.) But we can still revisit this later on. (I'm not too sold on branching out to other device types in general, but there seems to be demand for it.) > > See for instance the discussion on including vm config metadata in qcow2 > images: I think it's similar in scope (or the absense thereof). > > Roman.