--- /dev/null
+..  SPDX-License-Identifier: BSD-3-Clause
+    Copyright(c) 2018 Marvell International Ltd.
+    Copyright(c) 2018 Semihalf.
+    All rights reserved.
+
+MVNETA Poll Mode Driver
+=======================
+
+The MVNETA PMD (librte_pmd_mvneta) provides poll mode driver support
+for the Marvell NETA 1/2.5 Gbps adapter.
+
+Detailed information about SoCs that use PPv2 can be obtained here:
+
+* https://www.marvell.com/embedded-processors/armada-3700/
+
+.. Note::
+
+   Due to external dependencies, this driver is disabled by default. It must
+   be enabled manually by setting relevant configuration option manually.
+   Please refer to `Config File Options`_ section for further details.
+
+
+Features
+--------
+
+Features of the MVNETA PMD are:
+
+- Start/stop
+- tx/rx_queue_setup
+- Speed capabilities
+- Jumbo frame
+- CRC offload
+
+
+Limitations
+-----------
+
+- Flushing vlans added for filtering is not possible due to MUSDK missing
+  functionality. Current workaround is to reset board so that NETA has a
+  chance to start in a sane state.
+
+Prerequisites
+-------------
+
+- Custom Linux Kernel sources
+
+  .. code-block:: console
+
+     git clone https://github.com/MarvellEmbeddedProcessors/linux-marvell.git -b linux-4.4.120-armada-18.09
+
+
+- MUSDK (Marvell User-Space SDK) sources
+
+  .. code-block:: console
+
+     git clone https://github.com/MarvellEmbeddedProcessors/musdk-marvell.git -b musdk-armada-18.09
+
+  MUSDK is a light-weight library that provides direct access to Marvell's
+  NETA. Alternatively prebuilt MUSDK library can be
+  requested from `Marvell Extranet <https://extranet.marvell.com>`_. Once
+  approval has been granted, library can be found by typing ``musdk`` in
+  the search box.
+
+  MUSDK must be configured with the following features:
+
+  .. code-block:: console
+
+     --enable-pp2=no --enable-neta
+
+- DPDK environment
+
+  Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup
+  DPDK environment.
+
+Pre-Installation Configuration
+------------------------------
+
+Config File Options
+~~~~~~~~~~~~~~~~~~~
+
+The following options can be modified in the ``config`` file.
+
+- ``CONFIG_RTE_LIBRTE_MVNETA_PMD`` (default ``n``)
+
+    Toggle compilation of the librte_pmd_mvneta driver.
+
+Runtime options
+~~~~~~~~~~~~~~~
+
+The following ``devargs`` options can be enabled at runtime. They must
+be passed as part of EAL arguments.
+
+- ``iface`` (mandatory, with no default value)
+
+  The name of port (owned by MUSDK) that should be enabled in DPDK.
+  This options can be repeated resulting in a list of ports to be
+  enabled.  For instance below will enable ``eth0`` and ``eth1`` ports.
+
+.. code-block:: console
+
+   ./testpmd --vdev=net_mvneta,iface=eth0,iface=eth1 \
+    -c 3 -- -i --p 3 -a
+
+
+Building DPDK
+-------------
+
+Driver needs precompiled MUSDK library during compilation.
+
+.. code-block:: console
+
+   export CROSS_COMPILE=<toolchain>/bin/aarch64-linux-gnu-
+   ./bootstrap
+   ./configure --host=aarch64-linux-gnu --enable-pp2=no --enable-neta
+   make install
+
+MUSDK will be installed to `usr/local` under current directory.
+For the detailed build instructions please consult ``doc/musdk_get_started.txt``.
+
+Before the DPDK build process the environmental variable ``LIBMUSDK_PATH`` with
+the path to the MUSDK installation directory needs to be exported.
+
+.. code-block:: console
+
+   export LIBMUSDK_PATH=<musdk>/usr/local
+   export CROSS=aarch64-linux-gnu-
+   make config T=arm64-armv8a-linuxapp-gcc
+   sed -ri 's,(MVNETA_PMD=)n,\1y,' build/.config
+   make
+
+Usage Example
+-------------
+
+MVNETA PMD requires extra out of tree kernel modules to function properly.
+`musdk_uio` and `mv_neta_uio` sources are part of the MUSDK. Please consult
+``doc/musdk_get_started.txt`` for the detailed build instructions.
+
+.. code-block:: console
+
+   insmod musdk_uio.ko
+   insmod mv_neta_uio.ko
+
+Additionally interfaces used by DPDK application need to be put up:
+
+.. code-block:: console
+
+   ip link set eth0 up
+   ip link set eth1 up
+
+In order to run testpmd example application following command can be used:
+
+.. code-block:: console
+
+   ./testpmd --vdev=net_mvneta,iface=eth0,iface=eth1 -c 3 -- \
+     -i --p 3 -a --txd 256 --rxd 128 --rxq=1 --txq=1  --nb-cores=1
+
+
+In order to run l2fwd example application following command can be used:
+
+.. code-block:: console
+
+   ./l2fwd --vdev=net_mvneta,iface=eth0,iface=eth1 -c 3 -- -T 1 -p 3
 
--- /dev/null
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2018 Marvell International Ltd.
+ * Copyright(c) 2018 Semihalf.
+ * All rights reserved.
+ */
+
+#include <rte_ethdev_driver.h>
+#include <rte_kvargs.h>
+#include <rte_log.h>
+#include <rte_malloc.h>
+#include <rte_bus_vdev.h>
+
+#include <stdio.h>
+#include <fcntl.h>
+#include <linux/ethtool.h>
+#include <linux/sockios.h>
+#include <net/if.h>
+#include <net/if_arp.h>
+#include <sys/ioctl.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+
+#include <rte_mvep_common.h>
+
+#include "mvneta_ethdev.h"
+
+
+#define MVNETA_IFACE_NAME_ARG "iface"
+
+#define MVNETA_RX_OFFLOADS (DEV_RX_OFFLOAD_JUMBO_FRAME | \
+                         DEV_RX_OFFLOAD_CHECKSUM)
+
+/** Port Tx offloads capabilities */
+#define MVNETA_TX_OFFLOADS (DEV_TX_OFFLOAD_IPV4_CKSUM | \
+                         DEV_TX_OFFLOAD_UDP_CKSUM | \
+                         DEV_TX_OFFLOAD_TCP_CKSUM | \
+                         DEV_TX_OFFLOAD_MULTI_SEGS)
+
+#define MVNETA_PKT_SIZE_MAX (16382 - MV_MH_SIZE) /* 9700B */
+#define MVNETA_DEFAULT_MTU 1500
+
+#define MVNETA_MAC_ADDRS_MAX 256 /*16 UC, 256 IP, 256 MC/BC */
+/** Maximum length of a match string */
+#define MVNETA_MATCH_LEN 16
+
+int mvneta_logtype;
+
+static const char * const valid_args[] = {
+       MVNETA_IFACE_NAME_ARG,
+       NULL
+};
+
+struct mvneta_ifnames {
+       const char *names[NETA_NUM_ETH_PPIO];
+       int idx;
+};
+
+static int mvneta_dev_num;
+
+/**
+ * Deinitialize packet processor.
+ */
+static void
+mvneta_neta_deinit(void)
+{
+       neta_deinit();
+}
+
+/**
+ * Initialize packet processor.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+mvneta_neta_init(void)
+{
+       return neta_init();
+}
+
+/**
+ * Callback used by rte_kvargs_process() during argument parsing.
+ *
+ * @param key
+ *   Pointer to the parsed key (unused).
+ * @param value
+ *   Pointer to the parsed value.
+ * @param extra_args
+ *   Pointer to the extra arguments which contains address of the
+ *   table of pointers to parsed interface names.
+ *
+ * @return
+ *   Always 0.
+ */
+static int
+mvneta_ifnames_get(const char *key __rte_unused, const char *value,
+                void *extra_args)
+{
+       struct mvneta_ifnames *ifnames = extra_args;
+
+       ifnames->names[ifnames->idx++] = value;
+
+       return 0;
+}
+
+/**
+ * Ethernet device configuration.
+ *
+ * Prepare the driver for a given number of TX and RX queues and
+ * configure RSS if supported.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+mvneta_dev_configure(struct rte_eth_dev *dev)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+       struct neta_ppio_params *ppio_params;
+
+       if (dev->data->dev_conf.rxmode.mq_mode != ETH_MQ_RX_NONE) {
+               MVNETA_LOG(INFO, "Unsupported RSS and rx multi queue mode %d",
+                       dev->data->dev_conf.rxmode.mq_mode);
+               if (dev->data->nb_rx_queues > 1)
+                       return -EINVAL;
+       }
+
+       if (dev->data->dev_conf.rxmode.split_hdr_size) {
+               MVNETA_LOG(INFO, "Split headers not supported");
+               return -EINVAL;
+       }
+
+       if (dev->data->dev_conf.rxmode.offloads & DEV_RX_OFFLOAD_JUMBO_FRAME)
+               dev->data->mtu = dev->data->dev_conf.rxmode.max_rx_pkt_len -
+                                MRVL_NETA_ETH_HDRS_LEN;
+
+       if (dev->data->dev_conf.txmode.offloads & DEV_TX_OFFLOAD_MULTI_SEGS)
+               priv->multiseg = 1;
+
+       ppio_params = &priv->ppio_params;
+       ppio_params->outqs_params.num_outqs = dev->data->nb_tx_queues;
+       /* Default: 1 TC, no QoS supported. */
+       ppio_params->inqs_params.num_tcs = 1;
+       ppio_params->inqs_params.tcs_params[0].pkt_offset = MRVL_NETA_PKT_OFFS;
+       priv->ppio_id = dev->data->port_id;
+
+       return 0;
+}
+
+/**
+ * DPDK callback to get information about the device.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure (unused).
+ * @param info
+ *   Info structure output buffer.
+ */
+static void
+mvneta_dev_infos_get(struct rte_eth_dev *dev __rte_unused,
+                  struct rte_eth_dev_info *info)
+{
+       info->speed_capa = ETH_LINK_SPEED_10M |
+                          ETH_LINK_SPEED_100M |
+                          ETH_LINK_SPEED_1G |
+                          ETH_LINK_SPEED_2_5G;
+
+       info->max_rx_queues = MRVL_NETA_RXQ_MAX;
+       info->max_tx_queues = MRVL_NETA_TXQ_MAX;
+       info->max_mac_addrs = MVNETA_MAC_ADDRS_MAX;
+
+       info->rx_desc_lim.nb_max = MRVL_NETA_RXD_MAX;
+       info->rx_desc_lim.nb_min = MRVL_NETA_RXD_MIN;
+       info->rx_desc_lim.nb_align = MRVL_NETA_RXD_ALIGN;
+
+       info->tx_desc_lim.nb_max = MRVL_NETA_TXD_MAX;
+       info->tx_desc_lim.nb_min = MRVL_NETA_TXD_MIN;
+       info->tx_desc_lim.nb_align = MRVL_NETA_TXD_ALIGN;
+
+       info->rx_offload_capa = MVNETA_RX_OFFLOADS;
+       info->rx_queue_offload_capa = MVNETA_RX_OFFLOADS;
+
+       info->tx_offload_capa =  MVNETA_TX_OFFLOADS;
+       info->tx_queue_offload_capa =  MVNETA_TX_OFFLOADS;
+
+       /* By default packets are dropped if no descriptors are available */
+       info->default_rxconf.rx_drop_en = 1;
+       /* Deferred tx queue start is not supported */
+       info->default_txconf.tx_deferred_start = 0;
+       info->default_txconf.offloads = 0;
+
+       info->max_rx_pktlen = MVNETA_PKT_SIZE_MAX;
+}
+
+/**
+ * Return supported packet types.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure (unused).
+ *
+ * @return
+ *   Const pointer to the table with supported packet types.
+ */
+static const uint32_t *
+mvneta_dev_supported_ptypes_get(struct rte_eth_dev *dev __rte_unused)
+{
+       static const uint32_t ptypes[] = {
+               RTE_PTYPE_L2_ETHER,
+               RTE_PTYPE_L2_ETHER_VLAN,
+               RTE_PTYPE_L3_IPV4,
+               RTE_PTYPE_L3_IPV6,
+               RTE_PTYPE_L4_TCP,
+               RTE_PTYPE_L4_UDP
+       };
+
+       return ptypes;
+}
+
+/**
+ * DPDK callback to bring the link up.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+mvneta_dev_set_link_up(struct rte_eth_dev *dev)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+
+       if (!priv->ppio)
+               return 0;
+
+       return neta_ppio_enable(priv->ppio);
+}
+
+/**
+ * DPDK callback to bring the link down.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+mvneta_dev_set_link_down(struct rte_eth_dev *dev)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+
+       if (!priv->ppio)
+               return 0;
+
+       return neta_ppio_disable(priv->ppio);
+}
+
+/**
+ * DPDK callback to start the device.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ *
+ * @return
+ *   0 on success, negative errno value on failure.
+ */
+static int
+mvneta_dev_start(struct rte_eth_dev *dev)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+       char match[MVNETA_MATCH_LEN];
+       int ret = 0, i;
+
+       if (priv->ppio)
+               return mvneta_dev_set_link_up(dev);
+
+       snprintf(match, sizeof(match), "%s", dev->data->name);
+       priv->ppio_params.match = match;
+       priv->ppio_params.inqs_params.mtu = dev->data->mtu;
+
+       ret = neta_ppio_init(&priv->ppio_params, &priv->ppio);
+       if (ret) {
+               MVNETA_LOG(ERR, "Failed to init ppio");
+               return ret;
+       }
+       priv->ppio_id = priv->ppio->port_id;
+
+       /*
+        * In case there are some some stale uc/mc mac addresses flush them
+        * here. It cannot be done during mvneta_dev_close() as port information
+        * is already gone at that point (due to neta_ppio_deinit() in
+        * mvneta_dev_stop()).
+        */
+       if (!priv->uc_mc_flushed) {
+               ret = neta_ppio_flush_mac_addrs(priv->ppio, 0, 1);
+               if (ret) {
+                       MVNETA_LOG(ERR,
+                               "Failed to flush uc/mc filter list");
+                       goto out;
+               }
+               priv->uc_mc_flushed = 1;
+       }
+
+       ret = mvneta_dev_set_link_up(dev);
+       if (ret) {
+               MVNETA_LOG(ERR, "Failed to set link up");
+               goto out;
+       }
+
+       /* start tx queues */
+       for (i = 0; i < dev->data->nb_tx_queues; i++)
+               dev->data->tx_queue_state[i] = RTE_ETH_QUEUE_STATE_STARTED;
+
+       return 0;
+
+out:
+       MVNETA_LOG(ERR, "Failed to start device");
+       neta_ppio_deinit(priv->ppio);
+       return ret;
+}
+
+/**
+ * DPDK callback to stop the device.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ */
+static void
+mvneta_dev_stop(struct rte_eth_dev *dev)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+
+       if (!priv->ppio)
+               return;
+
+       mvneta_dev_set_link_down(dev);
+
+       neta_ppio_deinit(priv->ppio);
+
+       priv->ppio = NULL;
+}
+
+/**
+ * DPDK callback to close the device.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ */
+static void
+mvneta_dev_close(struct rte_eth_dev *dev)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+
+       if (priv->ppio)
+               mvneta_dev_stop(dev);
+}
+
+/**
+ * DPDK callback to set the primary MAC address.
+ *
+ * @param dev
+ *   Pointer to Ethernet device structure.
+ * @param mac_addr
+ *   MAC address to register.
+ */
+static int
+mvneta_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
+{
+       struct mvneta_priv *priv = dev->data->dev_private;
+       int ret;
+
+       if (!priv->ppio)
+               return -EINVAL;
+
+       ret = neta_ppio_set_mac_addr(priv->ppio, mac_addr->addr_bytes);
+       if (ret) {
+               char buf[ETHER_ADDR_FMT_SIZE];
+               ether_format_addr(buf, sizeof(buf), mac_addr);
+               MVNETA_LOG(ERR, "Failed to set mac to %s", buf);
+       }
+       return 0;
+}
+
+static const struct eth_dev_ops mvneta_ops = {
+       .dev_configure = mvneta_dev_configure,
+       .dev_start = mvneta_dev_start,
+       .dev_stop = mvneta_dev_stop,
+       .dev_set_link_up = mvneta_dev_set_link_up,
+       .dev_set_link_down = mvneta_dev_set_link_down,
+       .dev_close = mvneta_dev_close,
+       .mac_addr_set = mvneta_mac_addr_set,
+       .dev_infos_get = mvneta_dev_infos_get,
+       .dev_supported_ptypes_get = mvneta_dev_supported_ptypes_get,
+};
+
+/**
+ * Create device representing Ethernet port.
+ *
+ * @param name
+ *   Pointer to the port's name.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+mvneta_eth_dev_create(struct rte_vdev_device *vdev, const char *name)
+{
+       int ret, fd = socket(AF_INET, SOCK_DGRAM, 0);
+       struct rte_eth_dev *eth_dev;
+       struct mvneta_priv *priv;
+       struct ifreq req;
+
+       eth_dev = rte_eth_dev_allocate(name);
+       if (!eth_dev)
+               return -ENOMEM;
+
+       priv = rte_zmalloc_socket(name, sizeof(*priv), 0, rte_socket_id());
+       if (!priv) {
+               ret = -ENOMEM;
+               goto out_free_dev;
+       }
+
+       eth_dev->data->mac_addrs =
+               rte_zmalloc("mac_addrs",
+                           ETHER_ADDR_LEN * MVNETA_MAC_ADDRS_MAX, 0);
+       if (!eth_dev->data->mac_addrs) {
+               MVNETA_LOG(ERR, "Failed to allocate space for eth addrs");
+               ret = -ENOMEM;
+               goto out_free_priv;
+       }
+
+       memset(&req, 0, sizeof(req));
+       strcpy(req.ifr_name, name);
+       ret = ioctl(fd, SIOCGIFHWADDR, &req);
+       if (ret)
+               goto out_free_mac;
+
+       memcpy(eth_dev->data->mac_addrs[0].addr_bytes,
+              req.ifr_addr.sa_data, ETHER_ADDR_LEN);
+
+       eth_dev->data->kdrv = RTE_KDRV_NONE;
+       eth_dev->data->dev_private = priv;
+       eth_dev->device = &vdev->device;
+       eth_dev->dev_ops = &mvneta_ops;
+
+       rte_eth_dev_probing_finish(eth_dev);
+       return 0;
+out_free_mac:
+       rte_free(eth_dev->data->mac_addrs);
+out_free_priv:
+       rte_free(priv);
+out_free_dev:
+       rte_eth_dev_release_port(eth_dev);
+
+       return ret;
+}
+
+/**
+ * Cleanup previously created device representing Ethernet port.
+ *
+ * @param eth_dev
+ *   Pointer to the corresponding rte_eth_dev structure.
+ */
+static void
+mvneta_eth_dev_destroy(struct rte_eth_dev *eth_dev)
+{
+       rte_free(eth_dev->data->dev_private);
+       rte_free(eth_dev->data->mac_addrs);
+       rte_eth_dev_release_port(eth_dev);
+}
+
+/**
+ * Cleanup previously created device representing Ethernet port.
+ *
+ * @param name
+ *   Pointer to the port name.
+ */
+static void
+mvneta_eth_dev_destroy_name(const char *name)
+{
+       struct rte_eth_dev *eth_dev;
+
+       eth_dev = rte_eth_dev_allocated(name);
+       if (!eth_dev)
+               return;
+
+       mvneta_eth_dev_destroy(eth_dev);
+}
+
+/**
+ * DPDK callback to register the virtual device.
+ *
+ * @param vdev
+ *   Pointer to the virtual device.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+rte_pmd_mvneta_probe(struct rte_vdev_device *vdev)
+{
+       struct rte_kvargs *kvlist;
+       struct mvneta_ifnames ifnames;
+       int ret = -EINVAL;
+       uint32_t i, ifnum;
+       const char *params;
+
+       params = rte_vdev_device_args(vdev);
+       if (!params)
+               return -EINVAL;
+
+       kvlist = rte_kvargs_parse(params, valid_args);
+       if (!kvlist)
+               return -EINVAL;
+
+       ifnum = rte_kvargs_count(kvlist, MVNETA_IFACE_NAME_ARG);
+       if (ifnum > RTE_DIM(ifnames.names))
+               goto out_free_kvlist;
+
+       ifnames.idx = 0;
+       rte_kvargs_process(kvlist, MVNETA_IFACE_NAME_ARG,
+                          mvneta_ifnames_get, &ifnames);
+
+       /*
+        * The below system initialization should be done only once,
+        * on the first provided configuration file
+        */
+       if (mvneta_dev_num)
+               goto init_devices;
+
+       MVNETA_LOG(INFO, "Perform MUSDK initializations");
+
+       ret = rte_mvep_init(MVEP_MOD_T_NETA, kvlist);
+       if (ret)
+               goto out_free_kvlist;
+
+       ret = mvneta_neta_init();
+       if (ret) {
+               MVNETA_LOG(ERR, "Failed to init NETA!");
+               rte_mvep_deinit(MVEP_MOD_T_NETA);
+               goto out_free_kvlist;
+       }
+
+init_devices:
+       for (i = 0; i < ifnum; i++) {
+               MVNETA_LOG(INFO, "Creating %s", ifnames.names[i]);
+               ret = mvneta_eth_dev_create(vdev, ifnames.names[i]);
+               if (ret)
+                       goto out_cleanup;
+       }
+       mvneta_dev_num += ifnum;
+
+       rte_kvargs_free(kvlist);
+
+       return 0;
+out_cleanup:
+       for (; i > 0; i--)
+               mvneta_eth_dev_destroy_name(ifnames.names[i]);
+
+       if (mvneta_dev_num == 0) {
+               mvneta_neta_deinit();
+               rte_mvep_deinit(MVEP_MOD_T_NETA);
+       }
+out_free_kvlist:
+       rte_kvargs_free(kvlist);
+
+       return ret;
+}
+
+/**
+ * DPDK callback to remove virtual device.
+ *
+ * @param vdev
+ *   Pointer to the removed virtual device.
+ *
+ * @return
+ *   0 on success, negative error value otherwise.
+ */
+static int
+rte_pmd_mvneta_remove(struct rte_vdev_device *vdev)
+{
+       int i;
+       const char *name;
+
+       name = rte_vdev_device_name(vdev);
+       if (!name)
+               return -EINVAL;
+
+       MVNETA_LOG(INFO, "Removing %s", name);
+
+       RTE_ETH_FOREACH_DEV(i) {
+               if (rte_eth_devices[i].device != &vdev->device)
+                       continue;
+
+               mvneta_eth_dev_destroy(&rte_eth_devices[i]);
+               mvneta_dev_num--;
+       }
+
+       if (mvneta_dev_num == 0) {
+               MVNETA_LOG(INFO, "Perform MUSDK deinit");
+               mvneta_neta_deinit();
+               rte_mvep_deinit(MVEP_MOD_T_NETA);
+       }
+
+       return 0;
+}
+
+static struct rte_vdev_driver pmd_mvneta_drv = {
+       .probe = rte_pmd_mvneta_probe,
+       .remove = rte_pmd_mvneta_remove,
+};
+
+RTE_PMD_REGISTER_VDEV(net_mvneta, pmd_mvneta_drv);
+RTE_PMD_REGISTER_PARAM_STRING(net_mvneta, "iface=<ifc>");
+
+RTE_INIT(mvneta_init_log)
+{
+       mvneta_logtype = rte_log_register("pmd.net.mvneta");
+       if (mvneta_logtype >= 0)
+               rte_log_set_level(mvneta_logtype, RTE_LOG_NOTICE);
+}