From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga05.intel.com (mga05.intel.com [192.55.52.43]) by dpdk.org (Postfix) with ESMTP id A29D45901 for ; Tue, 6 Sep 2016 08:18:44 +0200 (CEST) Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga105.fm.intel.com with ESMTP; 05 Sep 2016 23:18:43 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.30,290,1470726000"; d="scan'208";a="757521014" Received: from yliu-dev.sh.intel.com (HELO yliu-dev) ([10.239.67.162]) by FMSMGA003.fm.intel.com with ESMTP; 05 Sep 2016 23:18:43 -0700 Date: Tue, 6 Sep 2016 14:30:49 +0800 From: Yuanhan Liu To: Jianfeng Tan Cc: dev@dpdk.org, zhihong.wang@intel.com, lining18@jd.com Message-ID: <20160906063049.GP30752@yliu-dev.sh.intel.com> References: <1470397003-5782-1-git-send-email-jianfeng.tan@intel.com> <1470397003-5782-2-git-send-email-jianfeng.tan@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1470397003-5782-2-git-send-email-jianfeng.tan@intel.com> User-Agent: Mutt/1.5.23 (2014-03-12) Subject: Re: [dpdk-dev] [PATCH 1/3] net/virtio_user: fix queue pair not enabled X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 06 Sep 2016 06:18:45 -0000 On Fri, Aug 05, 2016 at 11:36:41AM +0000, Jianfeng Tan wrote: > When virtio_user is used with OVS-DPDK (with mq disabled), it cannot > receive any packets. > > It's because when vhost provides VHOST_USER_GET_PROTOCOL_FEATURES, > all queue pairs are initialized in the disabled state. Quote > QEMU/docs/specs/vhost-user.txt: > If VHOST_USER_F_PROTOCOL_FEATURES has not been negotiated, the > ring is initialized in an enabled state. > If VHOST_USER_F_PROTOCOL_FEATURES has been negotiated, the ring > is initialized in a disabled state. > > In OVS-DPDK, all queue pairs are in the disabled state by default. > When used with QEMU, QEMU will set it as enabled in the process of > initialization. So the fix here is to include similar logic in > virtio_user. The behaviour is actually defined by virtio spec; the vhost-user just follows it (5.1.6.5.5 Automatic receive steering in multiqueue mode): Multiqueue is disabled by default. The driver enables multiqueue by executing the VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET command, specifying the number of the transmit and receive queues to be used up to max_virtqueue_pairs .... That is to say, it has nothing to do with OVS-DPDK; yet it's not caused by vhost-user. For QEMU, as the the virtio-net device emulator, it just follows the spec: it enables the 1st queue pair whenever MQ is enabled or not. Even the MQ __feature__ is enabled and queue=2 is given, only the 1st queue pair will be enabled. For linux virtio-net driver, you have to use ethtool to enable more queues if you want. However, for virtio-user, the yet another virtio-net device emulator, it didn't follow the rule, thus something went wrong. Put simply, you need make sure the 1st queue pair is enabled by default and the rest (if any, say when queues=2 is given) are disabled in the virtio-user initial stage (say, at virtio_user_dev_init()). --yliu > > Fixes: 37a7eb2ae816 ("net/virtio-user: add device emulation layer") > > Reported-by: Ning Li > Signed-off-by: Jianfeng Tan > --- > drivers/net/virtio/virtio_user/virtio_user_dev.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c b/drivers/net/virtio/virtio_user/virtio_user_dev.c > index 376c9cf..2c4e999 100644 > --- a/drivers/net/virtio/virtio_user/virtio_user_dev.c > +++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c > @@ -131,6 +131,13 @@ virtio_user_start_device(struct virtio_user_dev *dev) > } > } > > + /* As this feature is negotiated from the vhost, all queues are > + * initialized in the disabled state. For non-mq case, we enable > + * the 1st queue pair by default. > + */ > + if (dev->features & (1ull << VHOST_USER_GET_PROTOCOL_FEATURES)) > + vhost_user_enable_queue_pair(dev->vhostfd, 0, 1); > + > /* After setup all virtqueues, we need to set_features so that these > * features can be set into each virtqueue in vhost side. And before > * that, make sure VHOST_USER_F_PROTOCOL_FEATURES is added if mq is > -- > 2.7.4