From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga03.intel.com (mga03.intel.com [134.134.136.65]) by dpdk.org (Postfix) with ESMTP id A6E058E9F for ; Mon, 31 Aug 2015 07:15:37 +0200 (CEST) Received: from orsmga003.jf.intel.com ([10.7.209.27]) by orsmga103.jf.intel.com with ESMTP; 30 Aug 2015 22:15:36 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.17,438,1437462000"; d="scan'208";a="635233865" Received: from pgsmsx106.gar.corp.intel.com ([10.221.44.98]) by orsmga003.jf.intel.com with ESMTP; 30 Aug 2015 22:15:34 -0700 Received: from shsmsx152.ccr.corp.intel.com (10.239.6.52) by PGSMSX106.gar.corp.intel.com (10.221.44.98) with Microsoft SMTP Server (TLS) id 14.3.224.2; Mon, 31 Aug 2015 13:14:30 +0800 Received: from shsmsx102.ccr.corp.intel.com ([169.254.2.206]) by SHSMSX152.ccr.corp.intel.com ([169.254.6.182]) with mapi id 14.03.0224.002; Mon, 31 Aug 2015 13:14:28 +0800 From: "Ouyang, Changchun" To: Tetsuya Mukawa , "dev@dpdk.org" Thread-Topic: [dpdk-dev] [RFC PATCH] vhost: Add VHOST PMD Thread-Index: AQHQ4UDP90rzg30UEE2Y6d031aQkhZ4ldMFQ Date: Mon, 31 Aug 2015 05:14:28 +0000 Message-ID: References: <1440732101-18704-1-git-send-email-mukawa@igel.co.jp> <1440732101-18704-2-git-send-email-mukawa@igel.co.jp> In-Reply-To: <1440732101-18704-2-git-send-email-mukawa@igel.co.jp> Accept-Language: zh-CN, en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.239.127.40] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Cc: "ann.zhuangyanying@huawei.com" , "Liu, Yuanhan" Subject: Re: [dpdk-dev] [RFC PATCH] vhost: Add VHOST PMD 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: Mon, 31 Aug 2015 05:15:38 -0000 > -----Original Message----- > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Tetsuya Mukawa > Sent: Friday, August 28, 2015 11:22 AM > To: dev@dpdk.org > Cc: ann.zhuangyanying@huawei.com > Subject: [dpdk-dev] [RFC PATCH] vhost: Add VHOST PMD >=20 > The patch introduces a new PMD. This PMD is implemented as thin wrapper > of librte_vhost. It means librte_vhost is also needed to compile the PMD. > The PMD can have 'iface' parameter like below to specify a path to connec= t > to a virtio-net device. >=20 > $ ./testpmd -c f -n 4 --vdev 'eth_vhost0,iface=3D/tmp/sock0' -- -i >=20 > To connect above testpmd, here is qemu command example. >=20 > $ qemu-system-x86_64 \ > > -chardev socket,id=3Dchr0,path=3D/tmp/sock0 \ > -netdev vhost-user,id=3Dnet0,chardev=3Dchr0,vhostforce \ > -device virtio-net-pci,netdev=3Dnet0 >=20 > Signed-off-by: Tetsuya Mukawa > --- > config/common_linuxapp | 6 + > drivers/net/Makefile | 4 + > drivers/net/vhost/Makefile | 61 +++ > drivers/net/vhost/rte_eth_vhost.c | 639 > ++++++++++++++++++++++++++++ > drivers/net/vhost/rte_pmd_vhost_version.map | 4 + > mk/rte.app.mk | 8 +- > 6 files changed, 721 insertions(+), 1 deletion(-) create mode 100644 > drivers/net/vhost/Makefile create mode 100644 > drivers/net/vhost/rte_eth_vhost.c create mode 100644 > drivers/net/vhost/rte_pmd_vhost_version.map >=20 > diff --git a/config/common_linuxapp b/config/common_linuxapp index > 0de43d5..7310240 100644 > --- a/config/common_linuxapp > +++ b/config/common_linuxapp > @@ -446,6 +446,12 @@ CONFIG_RTE_LIBRTE_VHOST_NUMA=3Dn > CONFIG_RTE_LIBRTE_VHOST_DEBUG=3Dn >=20 > # > +# Compile vhost PMD > +# To compile, CONFIG_RTE_LIBRTE_VHOST should be enabled. > +# > +CONFIG_RTE_LIBRTE_PMD_VHOST=3Dy > + > +# > #Compile Xen domain0 support > # > CONFIG_RTE_LIBRTE_XEN_DOM0=3Dn > diff --git a/drivers/net/Makefile b/drivers/net/Makefile index > 5ebf963..e46a38e 100644 > --- a/drivers/net/Makefile > +++ b/drivers/net/Makefile > @@ -49,5 +49,9 @@ DIRS-$(CONFIG_RTE_LIBRTE_VIRTIO_PMD) +=3D virtio > DIRS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) +=3D vmxnet3 > DIRS-$(CONFIG_RTE_LIBRTE_PMD_XENVIRT) +=3D xenvirt >=20 > +ifeq ($(CONFIG_RTE_LIBRTE_VHOST),y) > +DIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) +=3D vhost endif # > +$(CONFIG_RTE_LIBRTE_VHOST) > + > include $(RTE_SDK)/mk/rte.sharelib.mk > include $(RTE_SDK)/mk/rte.subdir.mk > diff --git a/drivers/net/vhost/Makefile b/drivers/net/vhost/Makefile new > file mode 100644 index 0000000..018edde > --- /dev/null > +++ b/drivers/net/vhost/Makefile > @@ -0,0 +1,61 @@ > +# BSD LICENSE > +# > +# Copyright (c) 2010-2015 Intel Corporation. > +# All rights reserved. > +# > +# Redistribution and use in source and binary forms, with or without > +# modification, are permitted provided that the following conditions > +# are met: > +# > +# * Redistributions of source code must retain the above copyright > +# notice, this list of conditions and the following disclaimer. > +# * Redistributions in binary form must reproduce the above copyrigh= t > +# notice, this list of conditions and the following disclaimer in > +# the documentation and/or other materials provided with the > +# distribution. > +# * Neither the name of Intel corporation nor the names of its > +# contributors may be used to endorse or promote products derived > +# from this software without specific prior written permission. > +# > +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND > CONTRIBUTORS > +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT > NOT > +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND > FITNESS FOR > +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE > COPYRIGHT > +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, > INCIDENTAL, > +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT > NOT > +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS > OF USE, > +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED > AND ON ANY > +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR > TORT > +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF > THE USE > +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH > DAMAGE. > + > +include $(RTE_SDK)/mk/rte.vars.mk > + > +# > +# library name > +# > +LIB =3D librte_pmd_vhost.a > + > +CFLAGS +=3D -O3 > +CFLAGS +=3D $(WERROR_FLAGS) > + > +EXPORT_MAP :=3D rte_pmd_vhost_version.map > + > +LIBABIVER :=3D 1 > + > +# > +# all source are stored in SRCS-y > +# > +SRCS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) +=3D rte_eth_vhost.c > + > +# > +# Export include files > +# > +SYMLINK-y-include +=3D > + > +# this lib depends upon: > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) +=3D lib/librte_mbuf > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) +=3D lib/librte_ether > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) +=3D lib/librte_kvargs > + > +include $(RTE_SDK)/mk/rte.lib.mk > diff --git a/drivers/net/vhost/rte_eth_vhost.c > b/drivers/net/vhost/rte_eth_vhost.c > new file mode 100644 > index 0000000..961d120 > --- /dev/null > +++ b/drivers/net/vhost/rte_eth_vhost.c > @@ -0,0 +1,639 @@ > +/*- > + * BSD LICENSE > + * > + * Copyright (c) 2010-2015 Intel Corporation. > + * All rights reserved. > + * > + * Redistribution and use in source and binary forms, with or without > + * modification, are permitted provided that the following conditions > + * are met: > + * > + * * Redistributions of source code must retain the above copyright > + * notice, this list of conditions and the following disclaimer. > + * * Redistributions in binary form must reproduce the above copyrig= ht > + * notice, this list of conditions and the following disclaimer in > + * the documentation and/or other materials provided with the > + * distribution. > + * * Neither the name of Intel Corporation nor the names of its > + * contributors may be used to endorse or promote products derived > + * from this software without specific prior written permission. > + * > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND > CONTRIBUTORS > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT > NOT > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND > FITNESS FOR > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE > COPYRIGHT > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, > INCIDENTAL, > + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT > NOT > + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS > OF USE, > + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED > AND ON ANY > + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR > TORT > + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF > THE USE > + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH > DAMAGE. > + */ > +#include > +#include > + > +#include > +#include > +#include > +#include > +#include > +#include > +#include > + > +#define ETH_VHOST_IFACE_ARG "iface" > + > +static const char *drivername =3D "VHOST PMD"; > + > +static const char *valid_arguments[] =3D { > + ETH_VHOST_IFACE_ARG, > + NULL > +}; > + > +static struct ether_addr base_eth_addr =3D { > + .addr_bytes =3D{ > + 0x56 /* V */, > + 0x48 /* H */, > + 0x4F /* O */, > + 0x53 /* S */, > + 0x54 /* T */, > + 0x00 > + } > +}; > + > +struct vhost_queue { > + struct virtio_net *device; > + struct pmd_internal *internal; > + struct rte_mempool *mb_pool; > + rte_atomic64_t rx_pkts; > + rte_atomic64_t tx_pkts; > + rte_atomic64_t err_pkts; > + rte_atomic16_t rx_executing; > + rte_atomic16_t tx_executing; > +}; > + > +struct pmd_internal { > + TAILQ_ENTRY(pmd_internal) next; > + char *dev_name; > + char *iface_name; > + unsigned nb_rx_queues; > + unsigned nb_tx_queues; > + rte_atomic16_t xfer; > + > + struct vhost_queue > rx_vhost_queues[RTE_PMD_RING_MAX_RX_RINGS]; > + struct vhost_queue > tx_vhost_queues[RTE_PMD_RING_MAX_TX_RINGS]; > +}; Need consider how the vhost multiple queues implements here. You can refer to the patch set I sent before. > + > +TAILQ_HEAD(pmd_internal_head, pmd_internal); static struct > +pmd_internal_head internals_list =3D > + TAILQ_HEAD_INITIALIZER(internals_list); > + > +static pthread_mutex_t internal_list_lock =3D PTHREAD_MUTEX_INITIALIZER; > + > +static struct rte_eth_link pmd_link =3D { > + .link_speed =3D 10000, > + .link_duplex =3D ETH_LINK_FULL_DUPLEX, > + .link_status =3D 0 > +}; > + > +static uint16_t > +eth_vhost_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs) { > + struct vhost_queue *r =3D q; > + uint16_t nb_rx; > + > + if (unlikely(r->internal =3D=3D NULL)) > + return 0; > + > + if (unlikely(rte_atomic16_read(&r->internal->xfer) =3D=3D 0)) > + return 0; > + > + rte_atomic16_set(&r->rx_executing, 1); > + > + if (unlikely(rte_atomic16_read(&r->internal->xfer) =3D=3D 0)) > + goto out; > + > + nb_rx =3D (uint16_t)rte_vhost_dequeue_burst(r->device, > + VIRTIO_TXQ, r->mb_pool, bufs, nb_bufs); Logically correct here,=20 But it would be better to have more clear description why need use VIRTIO_T= XQ for vhost_rx function. It increases readability. :-) > + > + rte_atomic64_add(&(r->rx_pkts), nb_rx); > + > +out: > + rte_atomic16_set(&r->rx_executing, 0); > + > + return nb_rx; > +} > + > +static uint16_t > +eth_vhost_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs) { > + struct vhost_queue *r =3D q; > + uint16_t i, nb_tx; > + > + if (unlikely(r->internal =3D=3D NULL)) > + return 0; > + > + if (unlikely(rte_atomic16_read(&r->internal->xfer) =3D=3D 0)) > + return 0; > + > + rte_atomic16_set(&r->tx_executing, 1); > + > + if (unlikely(rte_atomic16_read(&r->internal->xfer) =3D=3D 0)) > + goto out; > + > + nb_tx =3D (uint16_t)rte_vhost_enqueue_burst(r->device, > + VIRTIO_RXQ, bufs, nb_bufs); > + > + rte_atomic64_add(&(r->tx_pkts), nb_tx); > + rte_atomic64_add(&(r->err_pkts), nb_bufs - nb_tx); > + > + for (i =3D 0; likely(i < nb_tx); i++) > + rte_pktmbuf_free(bufs[i]); > + > +out: > + rte_atomic16_set(&r->tx_executing, 0); > + > + return nb_tx; > +} > + > +static int > +eth_dev_configure(struct rte_eth_dev *dev __rte_unused) { return 0; } > + > +static int > +eth_dev_start(struct rte_eth_dev *dev) > +{ > + struct pmd_internal *internal =3D dev->data->dev_private; > + > + return rte_vhost_driver_register(internal->iface_name); > +} > + > +static void > +eth_dev_stop(struct rte_eth_dev *dev) > +{ > + struct pmd_internal *internal =3D dev->data->dev_private; > + > + rte_vhost_driver_unregister(internal->iface_name); > + return; > +} > + > +static int > +eth_rx_queue_setup(struct rte_eth_dev *dev,uint16_t rx_queue_id, > + uint16_t nb_rx_desc __rte_unused, > + unsigned int socket_id __rte_unused, > + const struct rte_eth_rxconf *rx_conf > __rte_unused, > + struct rte_mempool *mb_pool) > +{ > + struct pmd_internal *internal =3D dev->data->dev_private; > + internal->rx_vhost_queues[rx_queue_id].mb_pool =3D mb_pool; > + dev->data->rx_queues[rx_queue_id] =3D &internal- > >rx_vhost_queues[rx_queue_id]; > + return 0; > +} > + > +static int > +eth_tx_queue_setup(struct rte_eth_dev *dev, uint16_t tx_queue_id, > + uint16_t nb_tx_desc __rte_unused, > + unsigned int socket_id __rte_unused, > + const struct rte_eth_txconf *tx_conf > __rte_unused) { > + struct pmd_internal *internal =3D dev->data->dev_private; > + dev->data->tx_queues[tx_queue_id] =3D &internal- > >tx_vhost_queues[tx_queue_id]; > + return 0; > +} > + > + > +static void > +eth_dev_info(struct rte_eth_dev *dev, > + struct rte_eth_dev_info *dev_info) > +{ > + struct pmd_internal *internal =3D dev->data->dev_private; > + dev_info->driver_name =3D drivername; > + dev_info->max_mac_addrs =3D 1; > + dev_info->max_rx_pktlen =3D (uint32_t)-1; > + dev_info->max_rx_queues =3D (uint16_t)internal->nb_rx_queues; > + dev_info->max_tx_queues =3D (uint16_t)internal->nb_tx_queues; > + dev_info->min_rx_bufsize =3D 0; > + dev_info->pci_dev =3D dev->pci_dev; > +} > + > +static void > +eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats) > +{ > + unsigned i; > + unsigned long rx_total =3D 0, tx_total =3D 0, tx_err_total =3D 0; > + const struct pmd_internal *internal =3D dev->data->dev_private; > + > + for (i =3D 0; i < RTE_ETHDEV_QUEUE_STAT_CNTRS && > + i < internal->nb_rx_queues; i++) { > + igb_stats->q_ipackets[i] =3D internal- > >rx_vhost_queues[i].rx_pkts.cnt; > + rx_total +=3D igb_stats->q_ipackets[i]; > + } > + > + for (i =3D 0; i < RTE_ETHDEV_QUEUE_STAT_CNTRS && > + i < internal->nb_tx_queues; i++) { > + igb_stats->q_opackets[i] =3D internal- > >tx_vhost_queues[i].tx_pkts.cnt; > + igb_stats->q_errors[i] =3D internal- > >tx_vhost_queues[i].err_pkts.cnt; > + tx_total +=3D igb_stats->q_opackets[i]; > + tx_err_total +=3D igb_stats->q_errors[i]; > + } > + > + igb_stats->ipackets =3D rx_total; > + igb_stats->opackets =3D tx_total; > + igb_stats->oerrors =3D tx_err_total; > +} > + > +static void > +eth_stats_reset(struct rte_eth_dev *dev) { > + unsigned i; > + struct pmd_internal *internal =3D dev->data->dev_private; > + for (i =3D 0; i < internal->nb_rx_queues; i++) > + internal->rx_vhost_queues[i].rx_pkts.cnt =3D 0; > + for (i =3D 0; i < internal->nb_tx_queues; i++) { > + internal->tx_vhost_queues[i].tx_pkts.cnt =3D 0; > + internal->tx_vhost_queues[i].err_pkts.cnt =3D 0; > + } > +} > + > +static void > +eth_queue_release(void *q __rte_unused) { ; } static int > +eth_link_update(struct rte_eth_dev *dev __rte_unused, > + int wait_to_complete __rte_unused) { return 0; } > + > +static const struct eth_dev_ops ops =3D { > + .dev_start =3D eth_dev_start, > + .dev_stop =3D eth_dev_stop, > + .dev_configure =3D eth_dev_configure, > + .dev_infos_get =3D eth_dev_info, > + .rx_queue_setup =3D eth_rx_queue_setup, > + .tx_queue_setup =3D eth_tx_queue_setup, > + .rx_queue_release =3D eth_queue_release, > + .tx_queue_release =3D eth_queue_release, > + .link_update =3D eth_link_update, > + .stats_get =3D eth_stats_get, > + .stats_reset =3D eth_stats_reset, > +}; > + > +static struct eth_driver rte_vhost_pmd =3D { > + .pci_drv =3D { > + .name =3D "rte_vhost_pmd", > + .drv_flags =3D RTE_PCI_DRV_DETACHABLE, > + }, > +}; > + > +static struct rte_pci_id id_table; > + > +static inline struct pmd_internal * > +find_internal_resource(char *ifname) > +{ > + int found =3D 0; > + struct pmd_internal *internal; > + > + if (ifname =3D=3D NULL) > + return NULL; > + > + pthread_mutex_lock(&internal_list_lock); > + > + TAILQ_FOREACH(internal, &internals_list, next) { > + if (!strcmp(internal->iface_name, ifname)) { > + found =3D 1; > + break; > + } > + } > + > + pthread_mutex_unlock(&internal_list_lock); > + > + if (!found) > + return NULL; > + > + return internal; > +} > + > +static int > +new_device(struct virtio_net *dev) > +{ > + struct rte_eth_dev *eth_dev; > + struct pmd_internal *internal; > + struct vhost_queue *vq; > + unsigned i; > + > + if (dev =3D=3D NULL) { > + RTE_LOG(INFO, PMD, "invalid argument\n"); > + return -1; > + } > + > + internal =3D find_internal_resource(dev->ifname); > + if (internal =3D=3D NULL) { > + RTE_LOG(INFO, PMD, "invalid device name\n"); > + return -1; > + } > + > + eth_dev =3D rte_eth_dev_allocated(internal->dev_name); > + if (eth_dev =3D=3D NULL) { > + RTE_LOG(INFO, PMD, "failuer to find ethdev\n"); > + return -1; > + } > + > + internal =3D eth_dev->data->dev_private; > + > + for (i =3D 0; i < internal->nb_rx_queues; i++) { > + vq =3D &internal->rx_vhost_queues[i]; > + vq->device =3D dev; > + vq->internal =3D internal; > + } > + for (i =3D 0; i < internal->nb_tx_queues; i++) { > + vq =3D &internal->tx_vhost_queues[i]; > + vq->device =3D dev; > + vq->internal =3D internal; > + } > + > + dev->flags |=3D VIRTIO_DEV_RUNNING; > + dev->priv =3D eth_dev; > + > + eth_dev->data->dev_link.link_status =3D 1; > + rte_atomic16_set(&internal->xfer, 1); > + > + RTE_LOG(INFO, PMD, "New connection established\n"); > + > + return 0; > +} > + > +static void > +destroy_device(volatile struct virtio_net *dev) { > + struct rte_eth_dev *eth_dev; > + struct pmd_internal *internal; > + struct vhost_queue *vq; > + unsigned i; > + > + if (dev =3D=3D NULL) { > + RTE_LOG(INFO, PMD, "invalid argument\n"); > + return; > + } > + > + eth_dev =3D (struct rte_eth_dev *)dev->priv; > + if (eth_dev =3D=3D NULL) { > + RTE_LOG(INFO, PMD, "failuer to find a ethdev\n"); > + return; > + } > + > + internal =3D eth_dev->data->dev_private; > + > + /* Wait until rx/tx_pkt_burst stops accesing vhost device */ > + rte_atomic16_set(&internal->xfer, 0); > + for (i =3D 0; i < internal->nb_rx_queues; i++) { > + vq =3D &internal->rx_vhost_queues[i]; > + while (rte_atomic16_read(&vq->rx_executing)) > + rte_pause(); > + } > + for (i =3D 0; i < internal->nb_tx_queues; i++) { > + vq =3D &internal->tx_vhost_queues[i]; > + while (rte_atomic16_read(&vq->tx_executing)) > + rte_pause(); > + } > + > + eth_dev->data->dev_link.link_status =3D 0; > + > + dev->priv =3D NULL; > + dev->flags &=3D ~VIRTIO_DEV_RUNNING; > + > + for (i =3D 0; i < internal->nb_rx_queues; i++) { > + vq =3D &internal->rx_vhost_queues[i]; > + vq->device =3D NULL; > + } > + for (i =3D 0; i < internal->nb_tx_queues; i++) { > + vq =3D &internal->tx_vhost_queues[i]; > + vq->device =3D NULL; > + } > + > + RTE_LOG(INFO, PMD, "Connection closed\n"); > + > + return; > +} > + > +static void *vhost_driver_session(void *param __rte_unused) { > + static struct virtio_net_device_ops *vhost_ops =3D NULL; > + vhost_ops =3D rte_zmalloc(NULL, sizeof(*vhost_ops), 0); > + if (vhost_ops =3D=3D NULL) > + rte_panic("Can't allocate memory\n"); > + > + /* set vhost arguments */ > + vhost_ops->new_device =3D new_device; > + vhost_ops->destroy_device =3D destroy_device; > + if (rte_vhost_driver_callback_register(vhost_ops) < 0) > + rte_panic("Can't register callbacks\n"); > + > + /* start event handling */ > + rte_vhost_driver_session_start(); It should be called after rte_vhost_driver_register,=20 But rte_vhost_driver_session_start is called when dev_init, Error here? > + > + rte_free(vhost_ops); > + pthread_exit(0); > +} > + > +static pthread_once_t once_cont =3D PTHREAD_ONCE_INIT; static pthread_t > +session_th; > + > +static void vhost_driver_session_start(void) { > + int ret; > + > + ret =3D pthread_create(&session_th, NULL, vhost_driver_session, > NULL); > + if (ret) > + rte_panic("Can't create a thread\n"); } > + > +static int > +eth_dev_vhost_create(const char *name, int index, > + char *iface_name, > + const unsigned numa_node) > +{ > + struct rte_eth_dev_data *data =3D NULL; > + struct rte_pci_device *pci_dev =3D NULL; > + struct pmd_internal *internal =3D NULL; > + struct rte_eth_dev *eth_dev =3D NULL; > + struct ether_addr *eth_addr =3D NULL; > + uint16_t nb_rx_queues =3D 1; > + uint16_t nb_tx_queues =3D 1; > + > + RTE_LOG(INFO, PMD, "Creating VHOST-USER backend on numa > socket %u\n", > + numa_node); > + > + /* now do all data allocation - for eth_dev structure, dummy pci > driver > + * and internal (private) data > + */ > + data =3D rte_zmalloc_socket(name, sizeof(*data), 0, numa_node); > + if (data =3D=3D NULL) > + goto error; > + > + pci_dev =3D rte_zmalloc_socket(name, sizeof(*pci_dev), 0, > numa_node); > + if (pci_dev =3D=3D NULL) > + goto error; > + > + internal =3D rte_zmalloc_socket(name, sizeof(*internal), 0, > numa_node); > + if (internal =3D=3D NULL) > + goto error; > + > + eth_addr =3D rte_zmalloc_socket(name, sizeof(*eth_addr), 0, > numa_node); > + if (eth_addr =3D=3D NULL) > + goto error; > + *eth_addr =3D base_eth_addr; > + eth_addr->addr_bytes[5] =3D index; > + > + /* reserve an ethdev entry */ > + eth_dev =3D rte_eth_dev_allocate(name, RTE_ETH_DEV_VIRTUAL); > + if (eth_dev =3D=3D NULL) > + goto error; > + > + /* now put it all together > + * - store queue data in internal, > + * - store numa_node info in pci_driver > + * - point eth_dev_data to internal and pci_driver > + * - and point eth_dev structure to new eth_dev_data structure > + */ > + internal->nb_rx_queues =3D nb_rx_queues; > + internal->nb_tx_queues =3D nb_tx_queues; > + internal->dev_name =3D strdup(name); > + if (internal->dev_name =3D=3D NULL) > + goto error; > + internal->iface_name =3D strdup(iface_name); > + if (internal->iface_name =3D=3D NULL) > + goto error; > + > + pthread_mutex_lock(&internal_list_lock); > + TAILQ_INSERT_TAIL(&internals_list, internal, next); > + pthread_mutex_unlock(&internal_list_lock); > + > + rte_vhost_pmd.pci_drv.name =3D drivername; > + rte_vhost_pmd.pci_drv.id_table =3D &id_table; > + > + pci_dev->numa_node =3D numa_node; > + pci_dev->driver =3D &rte_vhost_pmd.pci_drv; > + > + data->dev_private =3D internal; > + data->port_id =3D eth_dev->data->port_id; > + memmove(data->name, eth_dev->data->name, sizeof(data- > >name)); > + data->nb_rx_queues =3D (uint16_t)nb_rx_queues; > + data->nb_tx_queues =3D (uint16_t)nb_tx_queues; > + data->dev_link =3D pmd_link; > + data->mac_addrs =3D eth_addr; > + > + /* > + * We'll replace the 'data' originally allocated by eth_dev. So the > + * vhost PMD resources won't be shared between multi processes. > + */ > + eth_dev->data =3D data; > + eth_dev->driver =3D &rte_vhost_pmd; > + eth_dev->dev_ops =3D &ops; > + eth_dev->pci_dev =3D pci_dev; > + > + /* finally assign rx and tx ops */ > + eth_dev->rx_pkt_burst =3D eth_vhost_rx; > + eth_dev->tx_pkt_burst =3D eth_vhost_tx; > + > + /* start vhost driver session. It should be called only once */ > + pthread_once(&once_cont, vhost_driver_session_start); > + > + return data->port_id; > + > +error: > + rte_free(data); > + rte_free(pci_dev); > + rte_free(internal); > + rte_free(eth_addr); > + > + return -1; > +} > + > +static inline int > +open_iface(const char *key __rte_unused, const char *value, void > +*extra_args) { > + const char **iface_name =3D extra_args; > + > + if (value =3D=3D NULL) > + return -1; > + > + *iface_name =3D value; > + > + return 0; > +} > + > +static int > +rte_pmd_vhost_devinit(const char *name, const char *params) { > + struct rte_kvargs *kvlist =3D NULL; > + int ret =3D 0; > + int index; > + char *iface_name; > + > + RTE_LOG(INFO, PMD, "Initializing pmd_vhost for %s\n", name); > + > + kvlist =3D rte_kvargs_parse(params, valid_arguments); > + if (kvlist =3D=3D NULL) > + return -1; > + > + if (strlen(name) < strlen("eth_vhost")) > + return -1; > + > + index =3D strtol(name + strlen("eth_vhost"), NULL, 0); > + if (errno =3D=3D ERANGE) > + return -1; > + > + if (rte_kvargs_count(kvlist, ETH_VHOST_IFACE_ARG) =3D=3D 1) { > + ret =3D rte_kvargs_process(kvlist, ETH_VHOST_IFACE_ARG, > + &open_iface, &iface_name); > + if (ret < 0) > + goto out_free; > + > + eth_dev_vhost_create(name, index, iface_name, > rte_socket_id()); > + } > + > +out_free: > + rte_kvargs_free(kvlist); > + return ret; > +} > + > +static int > +rte_pmd_vhost_devuninit(const char *name) { > + struct rte_eth_dev *eth_dev =3D NULL; > + struct pmd_internal *internal; > + > + RTE_LOG(INFO, PMD, "Un-Initializing pmd_vhost for %s\n", name); > + > + if (name =3D=3D NULL) > + return -EINVAL; > + > + /* find an ethdev entry */ > + eth_dev =3D rte_eth_dev_allocated(name); > + if (eth_dev =3D=3D NULL) > + return -ENODEV; > + > + internal =3D eth_dev->data->dev_private; > + > + pthread_mutex_lock(&internal_list_lock); > + TAILQ_REMOVE(&internals_list, internal, next); > + pthread_mutex_unlock(&internal_list_lock); > + > + eth_dev_stop(eth_dev); > + > + if ((internal) && (internal->dev_name)) > + free(internal->dev_name); > + if ((internal) && (internal->iface_name)) > + free(internal->iface_name); > + rte_free(eth_dev->data->dev_private); > + rte_free(eth_dev->data); > + rte_free(eth_dev->pci_dev); > + > + rte_eth_dev_release_port(eth_dev); > + return 0; > +} > + > +static struct rte_driver pmd_vhost_drv =3D { > + .name =3D "eth_vhost", > + .type =3D PMD_VDEV, > + .init =3D rte_pmd_vhost_devinit, > + .uninit =3D rte_pmd_vhost_devuninit, > +}; > + > +PMD_REGISTER_DRIVER(pmd_vhost_drv); > diff --git a/drivers/net/vhost/rte_pmd_vhost_version.map > b/drivers/net/vhost/rte_pmd_vhost_version.map > new file mode 100644 > index 0000000..5151684 > --- /dev/null > +++ b/drivers/net/vhost/rte_pmd_vhost_version.map > @@ -0,0 +1,4 @@ > +DPDK_2.2 { > + > + local: *; > +}; > diff --git a/mk/rte.app.mk b/mk/rte.app.mk index 3871205..1c42fb1 100644 > --- a/mk/rte.app.mk > +++ b/mk/rte.app.mk > @@ -144,7 +144,13 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_PCAP) +=3D - > lrte_pmd_pcap > _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_AF_PACKET) +=3D - > lrte_pmd_af_packet > _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_NULL) +=3D -lrte_pmd_null >=20 > -endif # ! $(CONFIG_RTE_BUILD_SHARED_LIB) > +ifeq ($(CONFIG_RTE_LIBRTE_VHOST),y) > + > +_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) +=3D -lrte_pmd_vhost > + > +endif # ! $(CONFIG_RTE_LIBRTE_VHOST) > + > +endif # $(CONFIG_RTE_BUILD_SHARED_LIB) >=20 > endif # ! CONFIG_RTE_BUILD_COMBINE_LIBS >=20 > -- > 2.1.4