[dpdk-dev] [PATCH v2 2/2] vhost: Add VHOST PMD

Tetsuya Mukawa mukawa at igel.co.jp
Mon Nov 2 04:58:57 CET 2015


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 vhost messages will be handled only when a port is started. So start
a port first, then invoke QEMU.

The PMD has 2 parameters.
 - iface:  The parameter is used to specify a path to connect to a
           virtio-net device.
 - queues: The parameter is used to specify the number of the queues
           virtio-net device has.
           (Default: 1)

Here is an example.
$ ./testpmd -c f -n 4 --vdev 'eth_vhost0,iface=/tmp/sock0,queues=1' -- -i

To connect above testpmd, here is qemu command example.

$ qemu-system-x86_64 \
        <snip>
        -chardev socket,id=chr0,path=/tmp/sock0 \
        -netdev vhost-user,id=net0,chardev=chr0,vhostforce,queues=1 \
        -device virtio-net-pci,netdev=net0

Signed-off-by: Tetsuya Mukawa <mukawa at igel.co.jp>
---
 config/common_linuxapp                      |   6 +
 doc/guides/nics/index.rst                   |   1 +
 doc/guides/nics/vhost.rst                   |  82 +++
 doc/guides/rel_notes/release_2_2.rst        |   2 +
 drivers/net/Makefile                        |   4 +
 drivers/net/vhost/Makefile                  |  62 +++
 drivers/net/vhost/rte_eth_vhost.c           | 765 ++++++++++++++++++++++++++++
 drivers/net/vhost/rte_eth_vhost.h           |  65 +++
 drivers/net/vhost/rte_pmd_vhost_version.map |   8 +
 mk/rte.app.mk                               |   8 +-
 10 files changed, 1002 insertions(+), 1 deletion(-)
 create mode 100644 doc/guides/nics/vhost.rst
 create mode 100644 drivers/net/vhost/Makefile
 create mode 100644 drivers/net/vhost/rte_eth_vhost.c
 create mode 100644 drivers/net/vhost/rte_eth_vhost.h
 create mode 100644 drivers/net/vhost/rte_pmd_vhost_version.map

diff --git a/config/common_linuxapp b/config/common_linuxapp
index c1d4bbd..fd103e7 100644
--- a/config/common_linuxapp
+++ b/config/common_linuxapp
@@ -457,6 +457,12 @@ CONFIG_RTE_LIBRTE_VHOST_NUMA=n
 CONFIG_RTE_LIBRTE_VHOST_DEBUG=n
 
 #
+# Compile vhost PMD
+# To compile, CONFIG_RTE_LIBRTE_VHOST should be enabled.
+#
+CONFIG_RTE_LIBRTE_PMD_VHOST=y
+
+#
 #Compile Xen domain0 support
 #
 CONFIG_RTE_LIBRTE_XEN_DOM0=n
diff --git a/doc/guides/nics/index.rst b/doc/guides/nics/index.rst
index 2d4936d..57d1041 100644
--- a/doc/guides/nics/index.rst
+++ b/doc/guides/nics/index.rst
@@ -47,6 +47,7 @@ Network Interface Controller Drivers
     mlx4
     mlx5
     virtio
+    vhost
     vmxnet3
     pcap_ring
 
diff --git a/doc/guides/nics/vhost.rst b/doc/guides/nics/vhost.rst
new file mode 100644
index 0000000..2ec8d79
--- /dev/null
+++ b/doc/guides/nics/vhost.rst
@@ -0,0 +1,82 @@
+..  BSD LICENSE
+    Copyright(c) 2015 IGEL Co., Ltd.. All rights reserved.
+    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 copyright
+    notice, this list of conditions and the following disclaimer in
+    the documentation and/or other materials provided with the
+    distribution.
+    * Neither the name of IGEL Co., Ltd. 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.
+
+Poll Mode Driver that wraps vhost library
+=========================================
+
+This PMD is a thin wrapper of the DPDK vhost library.
+The User can handle virtqueues as one of normal DPDK port.
+
+Vhost Implementation in DPDK
+----------------------------
+
+Please refer to Chapter "Vhost Library" of Programmer's Guide to know detail of vhost.
+
+Features and Limitations of vhost PMD
+-------------------------------------
+
+In this release, the vhost PMD provides the basic functionality of packet reception and transmission.
+
+*   It provides the function to convert port_id to a pointer of virtio_net device.
+    It allows the user to use vhost library with the PMD in parallel.
+
+*   It has multiple queues support.
+
+*   It supports Port Hotplug functionality.
+
+*   Don't need to stop RX/TX, when the user wants to stop a guest or a virtio-net driver on guest.
+
+Vhost PMD with testpmd application
+----------------------------------
+
+This section demonstrates vhost PMD with testpmd DPDK sample application.
+
+#.  Launch the testpmd with vhost PMD:
+
+    .. code-block:: console
+
+        ./testpmd -c f -n 4 --vdev 'eth_vhost0,iface=/tmp/sock0,queues=1' -- -i
+
+    Other basic DPDK preparations like hugepage enabling here.
+    Please refer to the *DPDK Getting Started Guide* for detailed instructions.
+
+#.  Launch the QEMU:
+
+    .. code-block:: console
+
+       qemu-system-x86_64 <snip>
+                   -chardev socket,id=chr0,path=/tmp/sock0 \
+                   -netdev vhost-user,id=net0,chardev=chr0,vhostforce,queues=1 \
+                   -device virtio-net-pci,netdev=net0
+
+    This command generates one virtio-net device for QEMU.
+    Once device is recognized by guest, The user can handle the device as normal
+    virtio-net device.
+    When initialization processes between virtio-net driver and vhost library are done, Port status of the testpmd will be linked up.
diff --git a/doc/guides/rel_notes/release_2_2.rst b/doc/guides/rel_notes/release_2_2.rst
index 429dfe6..466c1de 100644
--- a/doc/guides/rel_notes/release_2_2.rst
+++ b/doc/guides/rel_notes/release_2_2.rst
@@ -58,6 +58,8 @@ New Features
 * **Added port hotplug support to xenvirt.**
 
 
+* **Added vhost PMD.**
+
 * **Removed the PCI device from vdev PMD's.**
 
   * This change required modifications to librte_ether and all vdev and pdev PMD's.
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 6da1ce2..66eb63d 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -50,5 +50,9 @@ DIRS-$(CONFIG_RTE_LIBRTE_VIRTIO_PMD) += virtio
 DIRS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) += vmxnet3
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_XENVIRT) += xenvirt
 
+ifeq ($(CONFIG_RTE_LIBRTE_VHOST),y)
+DIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += 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..8186a80
--- /dev/null
+++ b/drivers/net/vhost/Makefile
@@ -0,0 +1,62 @@
+#   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 copyright
+#       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 = librte_pmd_vhost.a
+
+CFLAGS += -O3
+CFLAGS += $(WERROR_FLAGS)
+
+EXPORT_MAP := rte_pmd_vhost_version.map
+
+LIBABIVER := 1
+
+#
+# all source are stored in SRCS-y
+#
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += rte_eth_vhost.c
+
+#
+# Export include files
+#
+SYMLINK-y-include += rte_eth_vhost.h
+
+# this lib depends upon:
+DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += lib/librte_mbuf
+DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += lib/librte_ether
+DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += lib/librte_kvargs
+DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += lib/librte_vhost
+
+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..5e6da9a
--- /dev/null
+++ b/drivers/net/vhost/rte_eth_vhost.c
@@ -0,0 +1,765 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2015 IGEL Co., Ltd.
+ *   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 copyright
+ *       notice, this list of conditions and the following disclaimer in
+ *       the documentation and/or other materials provided with the
+ *       distribution.
+ *     * Neither the name of IGEL Co.,Ltd. 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 <unistd.h>
+#include <pthread.h>
+
+#include <rte_mbuf.h>
+#include <rte_ethdev.h>
+#include <rte_malloc.h>
+#include <rte_memcpy.h>
+#include <rte_dev.h>
+#include <rte_kvargs.h>
+#include <rte_virtio_net.h>
+
+#include "rte_eth_vhost.h"
+
+#define ETH_VHOST_IFACE_ARG		"iface"
+#define ETH_VHOST_QUEUES_ARG		"queues"
+
+static const char *drivername = "VHOST PMD";
+
+static const char *valid_arguments[] = {
+	ETH_VHOST_IFACE_ARG,
+	ETH_VHOST_QUEUES_ARG,
+	NULL
+};
+
+static struct ether_addr base_eth_addr = {
+	.addr_bytes = {
+		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_atomic32_t allow_queuing;
+	rte_atomic32_t while_queuing;
+	uint64_t rx_pkts;
+	uint64_t tx_pkts;
+	uint64_t err_pkts;
+};
+
+struct pmd_internal {
+	TAILQ_ENTRY(pmd_internal) next;
+	char *dev_name;
+	char *iface_name;
+	unsigned nb_rx_queues;
+	unsigned nb_tx_queues;
+
+	struct vhost_queue *rx_vhost_queues[RTE_MAX_QUEUES_PER_PORT];
+	struct vhost_queue *tx_vhost_queues[RTE_MAX_QUEUES_PER_PORT];
+
+	volatile uint16_t once;
+	pthread_t session_th;
+};
+
+TAILQ_HEAD(pmd_internal_head, pmd_internal);
+static struct pmd_internal_head internals_list =
+	TAILQ_HEAD_INITIALIZER(internals_list);
+
+static pthread_mutex_t internal_list_lock = PTHREAD_MUTEX_INITIALIZER;
+
+static struct rte_eth_link pmd_link = {
+		.link_speed = 10000,
+		.link_duplex = ETH_LINK_FULL_DUPLEX,
+		.link_status = 0
+};
+
+static uint16_t
+eth_vhost_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
+{
+	struct vhost_queue *r = q;
+	uint16_t nb_rx = 0;
+
+	if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0))
+		return 0;
+
+	rte_atomic32_set(&r->while_queuing, 1);
+
+	if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0))
+		goto out;
+
+	/* Dequeue packets from guest TX queue */
+	nb_rx = (uint16_t)rte_vhost_dequeue_burst(r->device,
+			VIRTIO_TXQ, r->mb_pool, bufs, nb_bufs);
+
+	r->rx_pkts += nb_rx;
+
+out:
+	rte_atomic32_set(&r->while_queuing, 0);
+
+	return nb_rx;
+}
+
+static uint16_t
+eth_vhost_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs)
+{
+	struct vhost_queue *r = q;
+	uint16_t i, nb_tx = 0;
+
+	if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0))
+		return 0;
+
+	rte_atomic32_set(&r->while_queuing, 1);
+
+	if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0))
+		goto out;
+
+	/* Enqueue packets to guest RX queue */
+	nb_tx = (uint16_t)rte_vhost_enqueue_burst(r->device,
+			VIRTIO_RXQ, bufs, nb_bufs);
+
+	r->tx_pkts += nb_tx;
+	r->err_pkts += nb_bufs - nb_tx;
+
+	for (i = 0; likely(i < nb_tx); i++)
+		rte_pktmbuf_free(bufs[i]);
+
+out:
+	rte_atomic32_set(&r->while_queuing, 0);
+
+	return nb_tx;
+}
+
+static int
+eth_dev_configure(struct rte_eth_dev *dev __rte_unused) { return 0; }
+
+static inline struct pmd_internal *
+find_internal_resource(char *ifname)
+{
+	int found = 0;
+	struct pmd_internal *internal;
+
+	if (ifname == NULL)
+		return NULL;
+
+	pthread_mutex_lock(&internal_list_lock);
+
+	TAILQ_FOREACH(internal, &internals_list, next) {
+		if (!strcmp(internal->iface_name, ifname)) {
+			found = 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 == NULL) {
+		RTE_LOG(INFO, PMD, "Invalid argument\n");
+		return -1;
+	}
+
+	internal = find_internal_resource(dev->ifname);
+	if (internal == NULL) {
+		RTE_LOG(INFO, PMD, "Invalid device name\n");
+		return -1;
+	}
+
+	if ((dev->virt_qp_nb < internal->nb_rx_queues) ||
+			(dev->virt_qp_nb < internal->nb_tx_queues)) {
+		RTE_LOG(INFO, PMD, "Not enough queues\n");
+		return -1;
+	}
+
+	eth_dev = rte_eth_dev_allocated(internal->dev_name);
+	if (eth_dev == NULL) {
+		RTE_LOG(INFO, PMD, "Failed to find a ethdev\n");
+		return -1;
+	}
+
+	for (i = 0; i < internal->nb_rx_queues; i++) {
+		vq = internal->rx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		vq->device = dev;
+		vq->internal = internal;
+	}
+	for (i = 0; i < internal->nb_tx_queues; i++) {
+		vq = internal->tx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		vq->device = dev;
+		vq->internal = internal;
+	}
+
+	dev->flags |= VIRTIO_DEV_RUNNING;
+	dev->pmd_priv = eth_dev;
+	eth_dev->data->dev_link.link_status = 1;
+
+	for (i = 0; i < internal->nb_rx_queues; i++) {
+		vq = internal->rx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		rte_atomic32_set(&vq->allow_queuing, 1);
+	}
+	for (i = 0; i < internal->nb_tx_queues; i++) {
+		vq = internal->tx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		rte_atomic32_set(&vq->allow_queuing, 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 == NULL) {
+		RTE_LOG(INFO, PMD, "Invalid argument\n");
+		return;
+	}
+
+	eth_dev = (struct rte_eth_dev *)dev->pmd_priv;
+	if (eth_dev == NULL) {
+		RTE_LOG(INFO, PMD, "Failed to find a ethdev\n");
+		return;
+	}
+
+	internal = eth_dev->data->dev_private;
+
+	/* Wait until rx/tx_pkt_burst stops accessing vhost device */
+	for (i = 0; i < internal->nb_rx_queues; i++) {
+		vq = internal->rx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		rte_atomic32_set(&vq->allow_queuing, 0);
+		while (rte_atomic32_read(&vq->while_queuing))
+			rte_pause();
+	}
+	for (i = 0; i < internal->nb_tx_queues; i++) {
+		vq = internal->tx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		rte_atomic32_set(&vq->allow_queuing, 0);
+		while (rte_atomic32_read(&vq->while_queuing))
+			rte_pause();
+	}
+
+	eth_dev->data->dev_link.link_status = 0;
+
+	dev->pmd_priv = NULL;
+	dev->flags &= ~VIRTIO_DEV_RUNNING;
+
+	for (i = 0; i < internal->nb_rx_queues; i++) {
+		vq = internal->rx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		vq->device = NULL;
+	}
+	for (i = 0; i < internal->nb_tx_queues; i++) {
+		vq = internal->tx_vhost_queues[i];
+		if (vq == NULL)
+			continue;
+		vq->device = NULL;
+	}
+
+	RTE_LOG(INFO, PMD, "Connection closed\n");
+}
+
+static void *vhost_driver_session(void *param __rte_unused)
+{
+	static struct virtio_net_device_ops *vhost_ops;
+
+	vhost_ops = rte_zmalloc(NULL, sizeof(*vhost_ops), 0);
+	if (vhost_ops == NULL)
+		rte_panic("Can't allocate memory\n");
+
+	/* set vhost arguments */
+	vhost_ops->new_device = new_device;
+	vhost_ops->destroy_device = destroy_device;
+	if (rte_vhost_driver_pmd_callback_register(vhost_ops) < 0)
+		rte_panic("Can't register callbacks\n");
+
+	/* start event handling */
+	rte_vhost_driver_session_start();
+
+	rte_free(vhost_ops);
+	pthread_exit(0);
+}
+
+static void vhost_driver_session_start(struct pmd_internal *internal)
+{
+	int ret;
+
+	ret = pthread_create(&internal->session_th,
+			NULL, vhost_driver_session, NULL);
+	if (ret)
+		rte_panic("Can't create a thread\n");
+}
+
+static void vhost_driver_session_stop(struct pmd_internal *internal)
+{
+	int ret;
+
+	ret = pthread_cancel(internal->session_th);
+	if (ret)
+		rte_panic("Can't cancel the thread\n");
+
+	ret = pthread_join(internal->session_th, NULL);
+	if (ret)
+		rte_panic("Can't join the thread\n");
+}
+
+static int
+eth_dev_start(struct rte_eth_dev *dev)
+{
+	int ret;
+	struct pmd_internal *internal = dev->data->dev_private;
+
+	if (rte_atomic16_cmpset(&internal->once, 0, 1)) {
+		ret = rte_vhost_driver_register(internal->iface_name);
+		if (ret)
+			return ret;
+
+		vhost_driver_session_start(internal);
+	}
+	return 0;
+}
+
+static void
+eth_dev_stop(struct rte_eth_dev *dev)
+{
+	struct pmd_internal *internal = dev->data->dev_private;
+
+	if (rte_atomic16_cmpset(&internal->once, 1, 0)) {
+		rte_vhost_driver_unregister(internal->iface_name);
+		vhost_driver_session_stop(internal);
+	}
+}
+
+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,
+		   const struct rte_eth_rxconf *rx_conf __rte_unused,
+		   struct rte_mempool *mb_pool)
+{
+	struct pmd_internal *internal = dev->data->dev_private;
+	struct vhost_queue *vq;
+
+	if (internal->rx_vhost_queues[rx_queue_id] != NULL)
+		rte_free(internal->rx_vhost_queues[rx_queue_id]);
+
+	vq = rte_zmalloc_socket(NULL, sizeof(struct vhost_queue),
+			RTE_CACHE_LINE_SIZE, socket_id);
+	if (vq == NULL) {
+		RTE_LOG(ERR, PMD, "Failed to allocate memory for rx queue\n");
+		return -ENOMEM;
+	}
+
+	vq->mb_pool = mb_pool;
+	internal->rx_vhost_queues[rx_queue_id] = vq;
+	dev->data->rx_queues[rx_queue_id] = vq;
+	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,
+		   const struct rte_eth_txconf *tx_conf __rte_unused)
+{
+	struct pmd_internal *internal = dev->data->dev_private;
+	struct vhost_queue *vq;
+
+	if (internal->tx_vhost_queues[tx_queue_id] != NULL)
+		rte_free(internal->tx_vhost_queues[tx_queue_id]);
+
+	vq = rte_zmalloc_socket(NULL, sizeof(struct vhost_queue),
+			RTE_CACHE_LINE_SIZE, socket_id);
+	if (vq == NULL) {
+		RTE_LOG(ERR, PMD, "Failed to allocate memory for tx queue\n");
+		return -ENOMEM;
+	}
+
+	internal->tx_vhost_queues[tx_queue_id] = vq;
+	dev->data->tx_queues[tx_queue_id] = vq;
+	return 0;
+}
+
+
+static void
+eth_dev_info(struct rte_eth_dev *dev,
+	     struct rte_eth_dev_info *dev_info)
+{
+	struct pmd_internal *internal = dev->data->dev_private;
+
+	dev_info->driver_name = drivername;
+	dev_info->max_mac_addrs = 1;
+	dev_info->max_rx_pktlen = (uint32_t)-1;
+	dev_info->max_rx_queues = (uint16_t)internal->nb_rx_queues;
+	dev_info->max_tx_queues = (uint16_t)internal->nb_tx_queues;
+	dev_info->min_rx_bufsize = 0;
+}
+
+static void
+eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_stats)
+{
+	unsigned i;
+	unsigned long rx_total = 0, tx_total = 0, tx_err_total = 0;
+	const struct pmd_internal *internal = dev->data->dev_private;
+
+	for (i = 0; i < RTE_ETHDEV_QUEUE_STAT_CNTRS &&
+	     i < internal->nb_rx_queues; i++) {
+		if (internal->rx_vhost_queues[i] == NULL)
+			continue;
+		igb_stats->q_ipackets[i] = internal->rx_vhost_queues[i]->rx_pkts;
+		rx_total += igb_stats->q_ipackets[i];
+	}
+
+	for (i = 0; i < RTE_ETHDEV_QUEUE_STAT_CNTRS &&
+	     i < internal->nb_tx_queues; i++) {
+		if (internal->tx_vhost_queues[i] == NULL)
+			continue;
+		igb_stats->q_opackets[i] = internal->tx_vhost_queues[i]->tx_pkts;
+		igb_stats->q_errors[i] = internal->tx_vhost_queues[i]->err_pkts;
+		tx_total += igb_stats->q_opackets[i];
+		tx_err_total += igb_stats->q_errors[i];
+	}
+
+	igb_stats->ipackets = rx_total;
+	igb_stats->opackets = tx_total;
+	igb_stats->oerrors = tx_err_total;
+}
+
+static void
+eth_stats_reset(struct rte_eth_dev *dev)
+{
+	unsigned i;
+	struct pmd_internal *internal = dev->data->dev_private;
+
+	for (i = 0; i < internal->nb_rx_queues; i++) {
+		if (internal->rx_vhost_queues[i] == NULL)
+			continue;
+		internal->rx_vhost_queues[i]->rx_pkts = 0;
+	}
+	for (i = 0; i < internal->nb_tx_queues; i++) {
+		if (internal->tx_vhost_queues[i] == NULL)
+			continue;
+		internal->tx_vhost_queues[i]->tx_pkts = 0;
+		internal->tx_vhost_queues[i]->err_pkts = 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 = {
+	.dev_start = eth_dev_start,
+	.dev_stop = eth_dev_stop,
+	.dev_configure = eth_dev_configure,
+	.dev_infos_get = eth_dev_info,
+	.rx_queue_setup = eth_rx_queue_setup,
+	.tx_queue_setup = eth_tx_queue_setup,
+	.rx_queue_release = eth_queue_release,
+	.tx_queue_release = eth_queue_release,
+	.link_update = eth_link_update,
+	.stats_get = eth_stats_get,
+	.stats_reset = eth_stats_reset,
+};
+
+static int
+eth_dev_vhost_create(const char *name, int index,
+		     char *iface_name,
+		     int16_t queues,
+		     const unsigned numa_node)
+{
+	struct rte_eth_dev_data *data = NULL;
+	struct pmd_internal *internal = NULL;
+	struct rte_eth_dev *eth_dev = NULL;
+	struct ether_addr *eth_addr = NULL;
+
+	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 = rte_zmalloc_socket(name, sizeof(*data), 0, numa_node);
+	if (data == NULL)
+		goto error;
+
+	internal = rte_zmalloc_socket(name, sizeof(*internal), 0, numa_node);
+	if (internal == NULL)
+		goto error;
+
+	eth_addr = rte_zmalloc_socket(name, sizeof(*eth_addr), 0, numa_node);
+	if (eth_addr == NULL)
+		goto error;
+	*eth_addr = base_eth_addr;
+	eth_addr->addr_bytes[5] = index;
+
+	/* reserve an ethdev entry */
+	eth_dev = rte_eth_dev_allocate(name, RTE_ETH_DEV_VIRTUAL);
+	if (eth_dev == NULL)
+		goto error;
+
+	/* now put it all together
+	 * - store queue data in internal,
+	 * - store numa_node info in ethdev data
+	 * - point eth_dev_data to internals
+	 * - and point eth_dev structure to new eth_dev_data structure
+	 */
+	internal->nb_rx_queues = queues;
+	internal->nb_tx_queues = queues;
+	internal->dev_name = strdup(name);
+	if (internal->dev_name == NULL)
+		goto error;
+	internal->iface_name = strdup(iface_name);
+	if (internal->iface_name == NULL)
+		goto error;
+
+	pthread_mutex_lock(&internal_list_lock);
+	TAILQ_INSERT_TAIL(&internals_list, internal, next);
+	pthread_mutex_unlock(&internal_list_lock);
+
+	data->dev_private = internal;
+	data->port_id = eth_dev->data->port_id;
+	memmove(data->name, eth_dev->data->name, sizeof(data->name));
+	data->nb_rx_queues = queues;
+	data->nb_tx_queues = queues;
+	data->dev_link = pmd_link;
+	data->mac_addrs = 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 = data;
+	eth_dev->dev_ops = &ops;
+	eth_dev->driver = NULL;
+	eth_dev->data->dev_flags = RTE_ETH_DEV_DETACHABLE;
+	eth_dev->data->kdrv = RTE_KDRV_NONE;
+	eth_dev->data->drv_name = internal->dev_name;
+	eth_dev->data->numa_node = numa_node;
+
+	/* finally assign rx and tx ops */
+	eth_dev->rx_pkt_burst = eth_vhost_rx;
+	eth_dev->tx_pkt_burst = eth_vhost_tx;
+
+	return data->port_id;
+
+error:
+	rte_free(data);
+	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 = extra_args;
+
+	if (value == NULL)
+		return -1;
+
+	*iface_name = value;
+
+	return 0;
+}
+
+static inline int
+open_queues(const char *key __rte_unused, const char *value, void *extra_args)
+{
+	uint16_t *q = extra_args;
+
+	if ((value == NULL) || (extra_args == NULL))
+		return -EINVAL;
+
+	*q = (uint16_t)strtoul(value, NULL, 0);
+	if ((*q == USHRT_MAX) && (errno == ERANGE))
+		return -1;
+
+	if (*q > RTE_MAX_QUEUES_PER_PORT)
+		return -1;
+
+	return 0;
+}
+
+static int
+rte_pmd_vhost_devinit(const char *name, const char *params)
+{
+	struct rte_kvargs *kvlist = NULL;
+	int ret = 0;
+	int index;
+	char *iface_name;
+	uint16_t queues;
+
+	RTE_LOG(INFO, PMD, "Initializing pmd_vhost for %s\n", name);
+
+	kvlist = rte_kvargs_parse(params, valid_arguments);
+	if (kvlist == NULL)
+		return -1;
+
+	if (strlen(name) < strlen("eth_vhost"))
+		return -1;
+
+	index = strtol(name + strlen("eth_vhost"), NULL, 0);
+	if (errno == ERANGE)
+		return -1;
+
+	if (rte_kvargs_count(kvlist, ETH_VHOST_IFACE_ARG) == 1) {
+		ret = rte_kvargs_process(kvlist, ETH_VHOST_IFACE_ARG,
+					 &open_iface, &iface_name);
+		if (ret < 0)
+			goto out_free;
+	}
+
+	if (rte_kvargs_count(kvlist, ETH_VHOST_QUEUES_ARG) == 1) {
+		ret = rte_kvargs_process(kvlist, ETH_VHOST_QUEUES_ARG,
+					 &open_queues, &queues);
+		if (ret < 0)
+			goto out_free;
+
+	} else
+		queues = 1;
+
+	eth_dev_vhost_create(name, index,
+			iface_name, queues, 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 = NULL;
+	struct pmd_internal *internal;
+	unsigned int i;
+
+	RTE_LOG(INFO, PMD, "Un-Initializing pmd_vhost for %s\n", name);
+
+	if (name == NULL)
+		return -EINVAL;
+
+	/* find an ethdev entry */
+	eth_dev = rte_eth_dev_allocated(name);
+	if (eth_dev == NULL)
+		return -ENODEV;
+
+	internal = 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->mac_addrs);
+	rte_free(eth_dev->data);
+
+	for (i = 0; i < internal->nb_rx_queues; i++) {
+		if (internal->rx_vhost_queues[i] != NULL)
+			rte_free(internal->rx_vhost_queues[i]);
+	}
+	for (i = 0; i < internal->nb_tx_queues; i++) {
+		if (internal->tx_vhost_queues[i] != NULL)
+			rte_free(internal->tx_vhost_queues[i]);
+	}
+	rte_free(internal);
+
+	rte_eth_dev_release_port(eth_dev);
+
+	return 0;
+}
+
+static struct rte_driver pmd_vhost_drv = {
+	.name = "eth_vhost",
+	.type = PMD_VDEV,
+	.init = rte_pmd_vhost_devinit,
+	.uninit = rte_pmd_vhost_devuninit,
+};
+
+struct
+virtio_net *rte_eth_vhost_portid2vdev(uint16_t port_id)
+{
+	struct rte_eth_dev *eth_dev;
+
+	if (rte_eth_dev_is_valid_port(port_id) == 0)
+		return NULL;
+
+	eth_dev = &rte_eth_devices[port_id];
+	if (strncmp("eth_vhost", eth_dev->data->drv_name,
+				strlen("eth_vhost")) == 0) {
+		struct pmd_internal *internal;
+		struct vhost_queue *vq;
+
+		internal = eth_dev->data->dev_private;
+		vq = internal->rx_vhost_queues[0];
+		if ((vq != NULL) && (vq->device != NULL))
+			return vq->device;
+	}
+
+	return NULL;
+}
+
+PMD_REGISTER_DRIVER(pmd_vhost_drv);
diff --git a/drivers/net/vhost/rte_eth_vhost.h b/drivers/net/vhost/rte_eth_vhost.h
new file mode 100644
index 0000000..22a880f
--- /dev/null
+++ b/drivers/net/vhost/rte_eth_vhost.h
@@ -0,0 +1,65 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2015 IGEL Co., Ltd.
+ *   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 copyright
+ *       notice, this list of conditions and the following disclaimer in
+ *       the documentation and/or other materials provided with the
+ *       distribution.
+ *     * Neither the name of IGEL Co., Ltd. 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.
+ */
+
+#ifndef _RTE_ETH_AF_PACKET_H_
+#define _RTE_ETH_AF_PACKET_H_
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <rte_virtio_net.h>
+
+/**
+ * The function convert specified port_id to virtio device structure.
+ * The retured device can be used for vhost library APIs.
+ * To use vhost library APIs and vhost PMD parallely, below API should
+ * not be called, because the API will be called by vhost PMD.
+ * - rte_vhost_driver_session_start()
+ * Once a device is managed by vhost PMD, below API should not be called.
+ * - rte_vhost_driver_unregister()
+ * To unregister the device, call Port Hotplug APIs.
+ *
+ * @param port_id
+ *  port number
+ * @return
+ *  virtio net device structure corresponding to the specified port
+ *  NULL will be returned in error cases.
+ */
+struct virtio_net *rte_eth_vhost_portid2vdev(uint16_t port_id);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
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..bf0361a
--- /dev/null
+++ b/drivers/net/vhost/rte_pmd_vhost_version.map
@@ -0,0 +1,8 @@
+DPDK_2.2 {
+
+	global:
+
+	rte_eth_vhost_portid2vdev;
+
+	local: *;
+};
diff --git a/mk/rte.app.mk b/mk/rte.app.mk
index 724efa7..1af4bb3 100644
--- a/mk/rte.app.mk
+++ b/mk/rte.app.mk
@@ -148,7 +148,13 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_PCAP)       += -lrte_pmd_pcap
 _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_AF_PACKET)  += -lrte_pmd_af_packet
 _LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_NULL)       += -lrte_pmd_null
 
-endif # ! $(CONFIG_RTE_BUILD_SHARED_LIB)
+ifeq ($(CONFIG_RTE_LIBRTE_VHOST),y)
+
+_LDLIBS-$(CONFIG_RTE_LIBRTE_PMD_VHOST)      += -lrte_pmd_vhost
+
+endif # ! $(CONFIG_RTE_LIBRTE_VHOST)
+
+endif # $(CONFIG_RTE_BUILD_SHARED_LIB)
 
 endif # ! CONFIG_RTE_BUILD_COMBINE_LIBS
 
-- 
2.1.4



More information about the dev mailing list