From mboxrd@z Thu Jan 1 00:00:00 1970 From: Cornelia Huck Subject: Re: [virtio-dev] Re: [Qemu-devel] [PATCH] qemu: Introduce VIRTIO_NET_F_STANDBY feature bit to virtio_net Date: Tue, 26 Jun 2018 18:03:16 +0200 Message-ID: <20180626180316.3723422f.cohuck@redhat.com> References: <20180623002628-mutt-send-email-mst@kernel.org> <20180623012406-mutt-send-email-mst@kernel.org> <20180626044650-mutt-send-email-mst@kernel.org> <20180626171732.5038f53f.cohuck@redhat.com> <20180626183706-mutt-send-email-mst@kernel.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT Cc: "Samudrala, Sridhar" , Siwei Liu , Alexander Duyck , virtio-dev@lists.oasis-open.org, aaron.f.brown@intel.com, Jiri Pirko , Jakub Kicinski , Netdev , qemu-devel@nongnu.org, virtualization@lists.linux-foundation.org, konrad.wilk@oracle.com, boris.ostrovsky@oracle.com, Joao Martins , Venu Busireddy , vijay.balakrishna@oracle.com To: "Michael S. Tsirkin" Return-path: Received: from mx3-rdu2.redhat.com ([66.187.233.73]:44244 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752524AbeFZQDW (ORCPT ); Tue, 26 Jun 2018 12:03:22 -0400 In-Reply-To: <20180626183706-mutt-send-email-mst@kernel.org> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 26 Jun 2018 18:38:51 +0300 "Michael S. Tsirkin" wrote: > On Tue, Jun 26, 2018 at 05:17:32PM +0200, Cornelia Huck wrote: > > On Tue, 26 Jun 2018 04:50:25 +0300 > > "Michael S. Tsirkin" wrote: > > > > > On Mon, Jun 25, 2018 at 10:54:09AM -0700, Samudrala, Sridhar wrote: > > > > > > > > Might not neccessarily be something wrong, but it's very limited to > > > > > > > > prohibit the MAC of VF from changing when enslaved by failover. > > > > > > > You mean guest changing MAC? I'm not sure why we prohibit that. > > > > > > I think Sridhar and Jiri might be better person to answer it. My > > > > > > impression was that sync'ing the MAC address change between all 3 > > > > > > devices is challenging, as the failover driver uses MAC address to > > > > > > match net_device internally. > > > > > > > > Yes. The MAC address is assigned by the hypervisor and it needs to manage the movement > > > > of the MAC between the PF and VF.  Allowing the guest to change the MAC will require > > > > synchronization between the hypervisor and the PF/VF drivers. Most of the VF drivers > > > > don't allow changing guest MAC unless it is a trusted VF. > > > > > > OK but it's a policy thing. Maybe it's a trusted VF. Who knows? > > > For example I can see host just > > > failing VIRTIO_NET_CTRL_MAC_ADDR_SET if it wants to block it. > > > I'm not sure why VIRTIO_NET_F_STANDBY has to block it in the guest. > > > > > > > So, what I get from this is that QEMU needs to be able to control all > > of standby, uuid, and mac to accommodate the different setups > > (respectively have libvirt/management software set it up). Is the host > > able to find out respectively define whether a VF is trusted? > > You do it with ip link I think but QEMU doesn't normally do this, > it relies on libvirt to poke at host kernel and supply the info. > Ok, that makes me conclude that we definitely need to involve the libvirt folks before we proceed further with defining QEMU interfaces.