From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45268) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZMytv-0007Pg-Sq for qemu-devel@nongnu.org; Wed, 05 Aug 2015 09:38:37 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZMytq-0000hY-SQ for qemu-devel@nongnu.org; Wed, 05 Aug 2015 09:38:35 -0400 Received: from mail-ig0-f177.google.com ([209.85.213.177]:32823) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZMytq-0000hU-N3 for qemu-devel@nongnu.org; Wed, 05 Aug 2015 09:38:30 -0400 Received: by igbpg9 with SMTP id pg9so114578430igb.0 for ; Wed, 05 Aug 2015 06:38:29 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1438593754-30005-1-git-send-email-thibaut.collet@6wind.com> <1438593754-30005-2-git-send-email-thibaut.collet@6wind.com> Date: Wed, 5 Aug 2015 15:38:29 +0200 Message-ID: From: Thibaut Collet Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v5 1/2] vhost user: add support of live migration List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?UTF-8?B?TWFyYy1BbmRyw6kgTHVyZWF1?= Cc: Linhaifeng , "Michael S. Tsirkin" , Jason Wang , QEMU , Stefan Hajnoczi , Paolo Bonzini , changchun.ouyang@intel.com On Tue, Aug 4, 2015 at 7:50 PM, Marc-Andr=C3=A9 Lureau wrote: > Hi Thibaut > > On Mon, Aug 3, 2015 at 11:22 AM, Thibaut Collet > wrote: >> Some vhost user backends are able to support live migration. >> To provide this service the following features must be added: >> 1. Add the VIRTIO_NET_F_GUEST_ANNOUNCE capability to vhost-net when netd= ev >> backend is vhost-user. >> 2. Provide a nop receive callback to vhost-user. >> This callback is used only by qemu_announce_self after a migration to= send >> fake RARP to avoid network outage for peers talking to the migrated g= uest. >> - For guest with GUEST_ANNOUNCE capabilities these packets must be di= scarded, >> GARP are sent by the guest thanks the GUEST_ANNOUNCE. >> - For guest without GUEST_ANNOUNCE capabilities these packets are dis= carded >> too, migration termination is notified when the guest sends packets= . >> >> Signed-off-by: Thibaut Collet >> --- >> hw/net/vhost_net.c | 2 ++ >> net/vhost-user.c | 25 +++++++++++++++++++++++-- >> 2 files changed, 25 insertions(+), 2 deletions(-) >> >> diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c >> index c864237..9850520 100644 >> --- a/hw/net/vhost_net.c >> +++ b/hw/net/vhost_net.c >> @@ -85,6 +85,8 @@ static const int user_feature_bits[] =3D { >> VIRTIO_NET_F_CTRL_MAC_ADDR, >> VIRTIO_NET_F_CTRL_GUEST_OFFLOADS, >> >> + VIRTIO_NET_F_GUEST_ANNOUNCE, >> + >> VIRTIO_NET_F_MQ, >> >> VHOST_INVALID_FEATURE_BIT >> diff --git a/net/vhost-user.c b/net/vhost-user.c >> index 93dcecd..2290271 100644 >> --- a/net/vhost-user.c >> +++ b/net/vhost-user.c >> @@ -65,6 +65,28 @@ static void vhost_user_stop(VhostUserState *s) >> s->vhost_net =3D 0; >> } >> >> +static ssize_t vhost_user_receive(NetClientState *nc, const uint8_t *bu= f, >> + size_t size) >> +{ >> + /* A live migration is done. Display an error if the packet is not = a RARP. >> + * RARP are just discarded: guest is already notified about live mi= gration >> + * by the virtio-net NIC or by the vhost-user backend. >> + */ >> + if (size !=3D 60) { >> + static int display_trace =3D 1; >> + >> + if (display_trace) { >> + fprintf(stderr, >> + "Vhost user expects only RARP (size 60)." >> + "Receives unexpected packets with size %lu\n", >> + size); >> + fflush(stderr); >> + display_trace =3D 0; >> + } >> + } >> + return size; >> +} > > This warning appears during a dummy boot with vapp: qemu-system-x86_64 > -m 512 -object memory-backend-file,id=3Dmem,size=3D512M,mem-path=3D/huget= lbfs/,share=3Don > -numa node,memdev=3Dmem -netdev > vhost-user,id=3Dnet0,chardev=3Dchr0,vhostforce -device > virtio-net-pci,netdev=3Dnet0 -chardev > socket,id=3Dchr0,path=3D/tmp/vapp.sock. > > I think silentely returning 0 (without warning) is the right way to > keep the packets queued in the ring instead. But I must say I am quite > confused as this is triggered by a VIRTIO_PCI_QUEUE_NOTIFY ioport > write, and I fail to see how vhost-user handles it then without > eventfd kick... > >> static void vhost_user_cleanup(NetClientState *nc) >> { >> VhostUserState *s =3D DO_UPCAST(VhostUserState, nc, nc); >> @@ -90,6 +112,7 @@ static bool vhost_user_has_ufo(NetClientState *nc) >> static NetClientInfo net_vhost_user_info =3D { >> .type =3D NET_CLIENT_OPTIONS_KIND_VHOST_USER, >> .size =3D sizeof(VhostUserState), >> + .receive =3D vhost_user_receive, >> .cleanup =3D vhost_user_cleanup, >> .has_vnet_hdr =3D vhost_user_has_vnet_hdr, >> .has_ufo =3D vhost_user_has_ufo, >> @@ -143,8 +166,6 @@ static int net_vhost_user_init(NetClientState *peer,= const char *device, >> >> s =3D DO_UPCAST(VhostUserState, nc, nc); >> >> - /* We don't provide a receive callback */ >> - s->nc.receive_disabled =3D 1; >> s->chr =3D chr; >> >> qemu_chr_add_handlers(s->chr, NULL, NULL, net_vhost_user_event, s); >> -- >> 1.7.10.4 >> > > > > -- > Marc-Andr=C3=A9 Lureau Hi, I have reproduced the issue and the message receive by vhost-user is quite strange: this message is a virtio header plus a bootp/dhcp request (I do not understand why there are the virtio header) With a dummy boot, vhost user backend already receives 4 bootp/dhcp requests form the bios guest so I do not see any interest to transmit this request to vhostuser backend. I suggest to remove the warning and discard the dhcp/bootp request (it is useless to duplicate bootp/dhcp request) Thibaut.