mbuf: mark old VLAN offload flags as deprecated
[dpdk.git] / drivers / net / cxgbe / cxgbe_ethdev.c
index 581a1f3..4758321 100644 (file)
 #include <rte_debug.h>
 #include <rte_pci.h>
 #include <rte_bus_pci.h>
-#include <rte_atomic.h>
 #include <rte_branch_prediction.h>
 #include <rte_memory.h>
 #include <rte_tailq.h>
 #include <rte_eal.h>
 #include <rte_alarm.h>
 #include <rte_ether.h>
-#include <rte_ethdev_driver.h>
-#include <rte_ethdev_pci.h>
+#include <ethdev_driver.h>
+#include <ethdev_pci.h>
 #include <rte_malloc.h>
 #include <rte_random.h>
 #include <rte_dev.h>
 
 #include "cxgbe.h"
 #include "cxgbe_pfvf.h"
+#include "cxgbe_flow.h"
 
 /*
  * Macros needed to support the PCI Device ID Table ...
 /*
  *... and the PCI ID Table itself ...
  */
-#include "t4_pci_id_tbl.h"
-
-#define CXGBE_TX_OFFLOADS (DEV_TX_OFFLOAD_VLAN_INSERT |\
-                          DEV_TX_OFFLOAD_IPV4_CKSUM |\
-                          DEV_TX_OFFLOAD_UDP_CKSUM |\
-                          DEV_TX_OFFLOAD_TCP_CKSUM |\
-                          DEV_TX_OFFLOAD_TCP_TSO)
-
-#define CXGBE_RX_OFFLOADS (DEV_RX_OFFLOAD_VLAN_STRIP |\
-                          DEV_RX_OFFLOAD_CRC_STRIP |\
-                          DEV_RX_OFFLOAD_IPV4_CKSUM |\
-                          DEV_RX_OFFLOAD_JUMBO_FRAME |\
-                          DEV_RX_OFFLOAD_UDP_CKSUM |\
-                          DEV_RX_OFFLOAD_TCP_CKSUM)
+#include "base/t4_pci_id_tbl.h"
 
 uint16_t cxgbe_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
                         uint16_t nb_pkts)
@@ -77,20 +64,25 @@ uint16_t cxgbe_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
        struct sge_eth_txq *txq = (struct sge_eth_txq *)tx_queue;
        uint16_t pkts_sent, pkts_remain;
        uint16_t total_sent = 0;
+       uint16_t idx = 0;
        int ret = 0;
 
-       CXGBE_DEBUG_TX(adapter, "%s: txq = %p; tx_pkts = %p; nb_pkts = %d\n",
-                      __func__, txq, tx_pkts, nb_pkts);
-
        t4_os_lock(&txq->txq_lock);
        /* free up desc from already completed tx */
        reclaim_completed_tx(&txq->q);
+       if (unlikely(!nb_pkts))
+               goto out_unlock;
+
+       rte_prefetch0(rte_pktmbuf_mtod(tx_pkts[0], volatile void *));
        while (total_sent < nb_pkts) {
                pkts_remain = nb_pkts - total_sent;
 
                for (pkts_sent = 0; pkts_sent < pkts_remain; pkts_sent++) {
-                       ret = t4_eth_xmit(txq, tx_pkts[total_sent + pkts_sent],
-                                         nb_pkts);
+                       idx = total_sent + pkts_sent;
+                       if ((idx + 1) < nb_pkts)
+                               rte_prefetch0(rte_pktmbuf_mtod(tx_pkts[idx + 1],
+                                                       volatile void *));
+                       ret = t4_eth_xmit(txq, tx_pkts[idx], nb_pkts);
                        if (ret < 0)
                                break;
                }
@@ -101,6 +93,7 @@ uint16_t cxgbe_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
                reclaim_completed_tx(&txq->q);
        }
 
+out_unlock:
        t4_os_unlock(&txq->txq_lock);
        return total_sent;
 }
@@ -111,22 +104,17 @@ uint16_t cxgbe_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
        struct sge_eth_rxq *rxq = (struct sge_eth_rxq *)rx_queue;
        unsigned int work_done;
 
-       CXGBE_DEBUG_RX(adapter, "%s: rxq->rspq.cntxt_id = %u; nb_pkts = %d\n",
-                      __func__, rxq->rspq.cntxt_id, nb_pkts);
-
        if (cxgbe_poll(&rxq->rspq, rx_pkts, (unsigned int)nb_pkts, &work_done))
                dev_err(adapter, "error in cxgbe poll\n");
 
-       CXGBE_DEBUG_RX(adapter, "%s: work_done = %u\n", __func__, work_done);
        return work_done;
 }
 
-void cxgbe_dev_info_get(struct rte_eth_dev *eth_dev,
+int cxgbe_dev_info_get(struct rte_eth_dev *eth_dev,
                        struct rte_eth_dev_info *device_info)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
-       int max_queues = adapter->sge.max_ethqsets / adapter->params.nports;
 
        static const struct rte_eth_desc_lim cxgbe_desc_lim = {
                .nb_max = CXGBE_MAX_RING_DESC_SIZE,
@@ -134,12 +122,10 @@ void cxgbe_dev_info_get(struct rte_eth_dev *eth_dev,
                .nb_align = 1,
        };
 
-       device_info->pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev);
-
        device_info->min_rx_bufsize = CXGBE_MIN_RX_BUFSIZE;
        device_info->max_rx_pktlen = CXGBE_MAX_RX_PKTLEN;
-       device_info->max_rx_queues = max_queues;
-       device_info->max_tx_queues = max_queues;
+       device_info->max_rx_queues = adapter->sge.max_ethqsets;
+       device_info->max_tx_queues = adapter->sge.max_ethqsets;
        device_info->max_mac_addrs = 1;
        /* XXX: For now we support one MAC/port */
        device_info->max_vfs = adapter->params.arch.vfcount;
@@ -158,136 +144,207 @@ void cxgbe_dev_info_get(struct rte_eth_dev *eth_dev,
        device_info->rx_desc_lim = cxgbe_desc_lim;
        device_info->tx_desc_lim = cxgbe_desc_lim;
        cxgbe_get_speed_caps(pi, &device_info->speed_capa);
+
+       return 0;
 }
 
-void cxgbe_dev_promiscuous_enable(struct rte_eth_dev *eth_dev)
+int cxgbe_dev_promiscuous_enable(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
+       int ret;
 
-       t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
-                     1, -1, 1, -1, false);
+       if (adapter->params.rawf_size != 0) {
+               ret = cxgbe_mpstcam_rawf_enable(pi);
+               if (ret < 0)
+                       return ret;
+       }
+
+       return t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
+                            1, -1, 1, -1, false);
 }
 
-void cxgbe_dev_promiscuous_disable(struct rte_eth_dev *eth_dev)
+int cxgbe_dev_promiscuous_disable(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
+       int ret;
 
-       t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
-                     0, -1, 1, -1, false);
+       if (adapter->params.rawf_size != 0) {
+               ret = cxgbe_mpstcam_rawf_disable(pi);
+               if (ret < 0)
+                       return ret;
+       }
+
+       return t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
+                            0, -1, 1, -1, false);
 }
 
-void cxgbe_dev_allmulticast_enable(struct rte_eth_dev *eth_dev)
+int cxgbe_dev_allmulticast_enable(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
 
        /* TODO: address filters ?? */
 
-       t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
-                     -1, 1, 1, -1, false);
+       return t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
+                            -1, 1, 1, -1, false);
 }
 
-void cxgbe_dev_allmulticast_disable(struct rte_eth_dev *eth_dev)
+int cxgbe_dev_allmulticast_disable(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
 
        /* TODO: address filters ?? */
 
-       t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
-                     -1, 0, 1, -1, false);
+       return t4_set_rxmode(adapter, adapter->mbox, pi->viid, -1,
+                            -1, 0, 1, -1, false);
 }
 
 int cxgbe_dev_link_update(struct rte_eth_dev *eth_dev,
-                         __rte_unused int wait_to_complete)
+                         int wait_to_complete)
+{
+       struct port_info *pi = eth_dev->data->dev_private;
+       unsigned int i, work_done, budget = 32;
+       struct link_config *lc = &pi->link_cfg;
+       struct adapter *adapter = pi->adapter;
+       struct rte_eth_link new_link = { 0 };
+       u8 old_link = pi->link_cfg.link_ok;
+       struct sge *s = &adapter->sge;
+
+       for (i = 0; i < CXGBE_LINK_STATUS_POLL_CNT; i++) {
+               if (!s->fw_evtq.desc)
+                       break;
+
+               cxgbe_poll(&s->fw_evtq, NULL, budget, &work_done);
+
+               /* Exit if link status changed or always forced up */
+               if (pi->link_cfg.link_ok != old_link ||
+                   cxgbe_force_linkup(adapter))
+                       break;
+
+               if (!wait_to_complete)
+                       break;
+
+               rte_delay_ms(CXGBE_LINK_STATUS_POLL_MS);
+       }
+
+       new_link.link_status = cxgbe_force_linkup(adapter) ?
+                              RTE_ETH_LINK_UP : pi->link_cfg.link_ok;
+       new_link.link_autoneg = (lc->link_caps & FW_PORT_CAP32_ANEG) ? 1 : 0;
+       new_link.link_duplex = RTE_ETH_LINK_FULL_DUPLEX;
+       new_link.link_speed = t4_fwcap_to_speed(lc->link_caps);
+
+       return rte_eth_linkstatus_set(eth_dev, &new_link);
+}
+
+/**
+ * Set device link up.
+ */
+int cxgbe_dev_set_link_up(struct rte_eth_dev *dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
+       unsigned int work_done, budget = 32;
        struct sge *s = &adapter->sge;
-       struct rte_eth_link *old_link = &eth_dev->data->dev_link;
-       unsigned int work_done, budget = 4;
+       int ret;
 
+       if (!s->fw_evtq.desc)
+               return -ENOMEM;
+
+       /* Flush all link events */
        cxgbe_poll(&s->fw_evtq, NULL, budget, &work_done);
-       if (old_link->link_status == pi->link_cfg.link_ok)
-               return -1;  /* link not changed */
 
-       eth_dev->data->dev_link.link_status = pi->link_cfg.link_ok;
-       eth_dev->data->dev_link.link_duplex = ETH_LINK_FULL_DUPLEX;
-       eth_dev->data->dev_link.link_speed = pi->link_cfg.speed;
+       /* If link already up, nothing to do */
+       if (pi->link_cfg.link_ok)
+               return 0;
+
+       ret = cxgbe_set_link_status(pi, true);
+       if (ret)
+               return ret;
 
-       /* link has changed */
+       cxgbe_dev_link_update(dev, 1);
        return 0;
 }
 
-int cxgbe_dev_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
+/**
+ * Set device link down.
+ */
+int cxgbe_dev_set_link_down(struct rte_eth_dev *dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
-       struct rte_eth_dev_info dev_info;
-       int err;
-       uint16_t new_mtu = mtu + ETHER_HDR_LEN + ETHER_CRC_LEN;
+       unsigned int work_done, budget = 32;
+       struct sge *s = &adapter->sge;
+       int ret;
 
-       cxgbe_dev_info_get(eth_dev, &dev_info);
+       if (!s->fw_evtq.desc)
+               return -ENOMEM;
 
-       /* Must accommodate at least ETHER_MIN_MTU */
-       if ((new_mtu < ETHER_MIN_MTU) || (new_mtu > dev_info.max_rx_pktlen))
-               return -EINVAL;
+       /* Flush all link events */
+       cxgbe_poll(&s->fw_evtq, NULL, budget, &work_done);
 
-       /* set to jumbo mode if needed */
-       if (new_mtu > ETHER_MAX_LEN)
-               eth_dev->data->dev_conf.rxmode.offloads |=
-                       DEV_RX_OFFLOAD_JUMBO_FRAME;
-       else
-               eth_dev->data->dev_conf.rxmode.offloads &=
-                       ~DEV_RX_OFFLOAD_JUMBO_FRAME;
+       /* If link already down, nothing to do */
+       if (!pi->link_cfg.link_ok)
+               return 0;
 
-       err = t4_set_rxmode(adapter, adapter->mbox, pi->viid, new_mtu, -1, -1,
-                           -1, -1, true);
-       if (!err)
-               eth_dev->data->dev_conf.rxmode.max_rx_pkt_len = new_mtu;
+       ret = cxgbe_set_link_status(pi, false);
+       if (ret)
+               return ret;
 
-       return err;
+       cxgbe_dev_link_update(dev, 0);
+       return 0;
+}
+
+int cxgbe_dev_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu)
+{
+       struct port_info *pi = eth_dev->data->dev_private;
+       struct adapter *adapter = pi->adapter;
+       uint16_t new_mtu = mtu + RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN;
+
+       return t4_set_rxmode(adapter, adapter->mbox, pi->viid, new_mtu, -1, -1,
+                           -1, -1, true);
 }
 
 /*
  * Stop device.
  */
-void cxgbe_dev_close(struct rte_eth_dev *eth_dev)
+int cxgbe_dev_close(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *temp_pi, *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
-       int i, dev_down = 0;
+       u8 i;
 
        CXGBE_FUNC_TRACE();
 
+       if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+               return 0;
+
        if (!(adapter->flags & FULL_INIT_DONE))
-               return;
+               return 0;
+
+       if (!pi->viid)
+               return 0;
 
        cxgbe_down(pi);
+       t4_sge_eth_release_queues(pi);
+       t4_free_vi(adapter, adapter->mbox, adapter->pf, 0, pi->viid);
+       pi->viid = 0;
 
-       /*
-        *  We clear queues only if both tx and rx path of the port
-        *  have been disabled
+       /* Free up the adapter-wide resources only after all the ports
+        * under this PF have been closed.
         */
-       t4_sge_eth_clear_queues(pi);
-
-       /*  See if all ports are down */
        for_each_port(adapter, i) {
-               pi = adap2pinfo(adapter, i);
-               /*
-                * Skip first port of the adapter since it will be closed
-                * by DPDK
-                */
-               if (i == 0)
-                       continue;
-               dev_down += (pi->eth_dev->data->dev_started == 0) ? 1 : 0;
+               temp_pi = adap2pinfo(adapter, i);
+               if (temp_pi->viid)
+                       return 0;
        }
 
-       /* If rest of the ports are stopped, then free up resources */
-       if (dev_down == (adapter->params.nports - 1))
-               cxgbe_close(adapter);
+       cxgbe_close(adapter);
+       rte_free(adapter);
+
+       return 0;
 }
 
 /* Start the device.
@@ -295,7 +352,8 @@ void cxgbe_dev_close(struct rte_eth_dev *eth_dev)
  */
 int cxgbe_dev_start(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
+       struct rte_eth_rxmode *rx_conf = &eth_dev->data->dev_conf.rxmode;
        struct adapter *adapter = pi->adapter;
        int err = 0, i;
 
@@ -316,9 +374,14 @@ int cxgbe_dev_start(struct rte_eth_dev *eth_dev)
                        goto out;
        }
 
+       if (rx_conf->offloads & RTE_ETH_RX_OFFLOAD_SCATTER)
+               eth_dev->data->scattered_rx = 1;
+       else
+               eth_dev->data->scattered_rx = 0;
+
        cxgbe_enable_rx_queues(pi);
 
-       err = setup_rss(pi);
+       err = cxgbe_setup_rss(pi);
        if (err)
                goto out;
 
@@ -334,7 +397,7 @@ int cxgbe_dev_start(struct rte_eth_dev *eth_dev)
                        goto out;
        }
 
-       err = link_start(pi);
+       err = cxgbe_link_start(pi);
        if (err)
                goto out;
 
@@ -345,15 +408,15 @@ out:
 /*
  * Stop device: disable rx and tx functions to allow for reconfiguring.
  */
-void cxgbe_dev_stop(struct rte_eth_dev *eth_dev)
+int cxgbe_dev_stop(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
 
        CXGBE_FUNC_TRACE();
 
        if (!(adapter->flags & FULL_INIT_DONE))
-               return;
+               return 0;
 
        cxgbe_down(pi);
 
@@ -362,47 +425,36 @@ void cxgbe_dev_stop(struct rte_eth_dev *eth_dev)
         *  have been disabled
         */
        t4_sge_eth_clear_queues(pi);
+       eth_dev->data->scattered_rx = 0;
+
+       return 0;
 }
 
 int cxgbe_dev_configure(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
-       uint64_t unsupported_offloads, configured_offloads;
        int err;
 
        CXGBE_FUNC_TRACE();
-       configured_offloads = eth_dev->data->dev_conf.rxmode.offloads;
-       if (!(configured_offloads & DEV_RX_OFFLOAD_CRC_STRIP)) {
-               dev_info(adapter, "can't disable hw crc strip\n");
-               configured_offloads |= DEV_RX_OFFLOAD_CRC_STRIP;
-       }
 
-       unsupported_offloads = configured_offloads & ~CXGBE_RX_OFFLOADS;
-       if (unsupported_offloads) {
-               dev_err(adapter, "Rx offloads 0x%" PRIx64 " are not supported. "
-                       "Supported:0x%" PRIx64 "\n",
-                       unsupported_offloads, (uint64_t)CXGBE_RX_OFFLOADS);
-               return -ENOTSUP;
-       }
-
-       configured_offloads = eth_dev->data->dev_conf.txmode.offloads;
-       unsupported_offloads = configured_offloads & ~CXGBE_TX_OFFLOADS;
-       if (unsupported_offloads) {
-               dev_err(adapter, "Tx offloads 0x%" PRIx64 " are not supported. "
-                       "Supported:0x%" PRIx64 "\n",
-                       unsupported_offloads, (uint64_t)CXGBE_TX_OFFLOADS);
-               return -ENOTSUP;
-       }
+       if (eth_dev->data->dev_conf.rxmode.mq_mode & RTE_ETH_MQ_RX_RSS_FLAG)
+               eth_dev->data->dev_conf.rxmode.offloads |=
+                       RTE_ETH_RX_OFFLOAD_RSS_HASH;
 
        if (!(adapter->flags & FW_QUEUE_BOUND)) {
-               err = setup_sge_fwevtq(adapter);
+               err = cxgbe_setup_sge_fwevtq(adapter);
                if (err)
                        return err;
                adapter->flags |= FW_QUEUE_BOUND;
+               if (is_pf4(adapter)) {
+                       err = cxgbe_setup_sge_ctrl_txq(adapter);
+                       if (err)
+                               return err;
+               }
        }
 
-       err = cfg_queue_count(eth_dev);
+       err = cxgbe_cfg_queue_count(eth_dev);
        if (err)
                return err;
 
@@ -442,31 +494,23 @@ int cxgbe_dev_tx_queue_stop(struct rte_eth_dev *eth_dev, uint16_t tx_queue_id)
 int cxgbe_dev_tx_queue_setup(struct rte_eth_dev *eth_dev,
                             uint16_t queue_idx, uint16_t nb_desc,
                             unsigned int socket_id,
-                            const struct rte_eth_txconf *tx_conf)
+                            const struct rte_eth_txconf *tx_conf __rte_unused)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        struct sge *s = &adapter->sge;
-       struct sge_eth_txq *txq = &s->ethtxq[pi->first_qset + queue_idx];
-       int err = 0;
        unsigned int temp_nb_desc;
-       uint64_t unsupported_offloads;
-
-       unsupported_offloads = tx_conf->offloads & ~CXGBE_TX_OFFLOADS;
-       if (unsupported_offloads) {
-               dev_err(adapter, "Tx offloads 0x%" PRIx64 " are not supported. "
-                       "Supported:0x%" PRIx64 "\n",
-                       unsupported_offloads, (uint64_t)CXGBE_TX_OFFLOADS);
-               return -ENOTSUP;
-       }
+       struct sge_eth_txq *txq;
+       int err = 0;
 
+       txq = &s->ethtxq[pi->first_txqset + queue_idx];
        dev_debug(adapter, "%s: eth_dev->data->nb_tx_queues = %d; queue_idx = %d; nb_desc = %d; socket_id = %d; pi->first_qset = %u\n",
                  __func__, eth_dev->data->nb_tx_queues, queue_idx, nb_desc,
-                 socket_id, pi->first_qset);
+                 socket_id, pi->first_txqset);
 
        /*  Free up the existing queue  */
        if (eth_dev->data->tx_queues[queue_idx]) {
-               cxgbe_dev_tx_queue_release(eth_dev->data->tx_queues[queue_idx]);
+               cxgbe_dev_tx_queue_release(eth_dev, queue_idx);
                eth_dev->data->tx_queues[queue_idx] = NULL;
        }
 
@@ -499,9 +543,9 @@ int cxgbe_dev_tx_queue_setup(struct rte_eth_dev *eth_dev,
        return err;
 }
 
-void cxgbe_dev_tx_queue_release(void *q)
+void cxgbe_dev_tx_queue_release(struct rte_eth_dev *eth_dev, uint16_t qid)
 {
-       struct sge_eth_txq *txq = (struct sge_eth_txq *)q;
+       struct sge_eth_txq *txq = eth_dev->data->tx_queues[qid];
 
        if (txq) {
                struct port_info *pi = (struct port_info *)
@@ -517,17 +561,16 @@ void cxgbe_dev_tx_queue_release(void *q)
 
 int cxgbe_dev_rx_queue_start(struct rte_eth_dev *eth_dev, uint16_t rx_queue_id)
 {
-       int ret;
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adap = pi->adapter;
-       struct sge_rspq *q;
+       struct sge_eth_rxq *rxq;
+       int ret;
 
        dev_debug(adapter, "%s: pi->port_id = %d; rx_queue_id = %d\n",
                  __func__, pi->port_id, rx_queue_id);
 
-       q = eth_dev->data->rx_queues[rx_queue_id];
-
-       ret = t4_sge_eth_rxq_start(adap, q);
+       rxq = eth_dev->data->rx_queues[rx_queue_id];
+       ret = t4_sge_eth_rxq_start(adap, rxq);
        if (ret == 0)
                eth_dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED;
 
@@ -536,16 +579,16 @@ int cxgbe_dev_rx_queue_start(struct rte_eth_dev *eth_dev, uint16_t rx_queue_id)
 
 int cxgbe_dev_rx_queue_stop(struct rte_eth_dev *eth_dev, uint16_t rx_queue_id)
 {
-       int ret;
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adap = pi->adapter;
-       struct sge_rspq *q;
+       struct sge_eth_rxq *rxq;
+       int ret;
 
        dev_debug(adapter, "%s: pi->port_id = %d; rx_queue_id = %d\n",
                  __func__, pi->port_id, rx_queue_id);
 
-       q = eth_dev->data->rx_queues[rx_queue_id];
-       ret = t4_sge_eth_rxq_stop(adap, q);
+       rxq = eth_dev->data->rx_queues[rx_queue_id];
+       ret = t4_sge_eth_rxq_stop(adap, rxq);
        if (ret == 0)
                eth_dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED;
 
@@ -555,41 +598,32 @@ int cxgbe_dev_rx_queue_stop(struct rte_eth_dev *eth_dev, uint16_t rx_queue_id)
 int cxgbe_dev_rx_queue_setup(struct rte_eth_dev *eth_dev,
                             uint16_t queue_idx, uint16_t nb_desc,
                             unsigned int socket_id,
-                            const struct rte_eth_rxconf *rx_conf,
+                            const struct rte_eth_rxconf *rx_conf __rte_unused,
                             struct rte_mempool *mp)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       unsigned int pkt_len = eth_dev->data->mtu + RTE_ETHER_HDR_LEN +
+               RTE_ETHER_CRC_LEN;
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
+       struct rte_eth_dev_info dev_info;
        struct sge *s = &adapter->sge;
-       struct sge_eth_rxq *rxq = &s->ethrxq[pi->first_qset + queue_idx];
-       int err = 0;
-       int msi_idx = 0;
        unsigned int temp_nb_desc;
-       struct rte_eth_dev_info dev_info;
-       unsigned int pkt_len = eth_dev->data->dev_conf.rxmode.max_rx_pkt_len;
-       uint64_t unsupported_offloads, configured_offloads;
-
-       configured_offloads = rx_conf->offloads;
-       if (!(configured_offloads & DEV_RX_OFFLOAD_CRC_STRIP)) {
-               dev_info(adapter, "can't disable hw crc strip\n");
-               configured_offloads |= DEV_RX_OFFLOAD_CRC_STRIP;
-       }
-
-       unsupported_offloads = configured_offloads & ~CXGBE_RX_OFFLOADS;
-       if (unsupported_offloads) {
-               dev_err(adapter, "Rx offloads 0x%" PRIx64 " are not supported. "
-                       "Supported:0x%" PRIx64 "\n",
-                       unsupported_offloads, (uint64_t)CXGBE_RX_OFFLOADS);
-               return -ENOTSUP;
-       }
+       int err = 0, msi_idx = 0;
+       struct sge_eth_rxq *rxq;
 
+       rxq = &s->ethrxq[pi->first_rxqset + queue_idx];
        dev_debug(adapter, "%s: eth_dev->data->nb_rx_queues = %d; queue_idx = %d; nb_desc = %d; socket_id = %d; mp = %p\n",
                  __func__, eth_dev->data->nb_rx_queues, queue_idx, nb_desc,
                  socket_id, mp);
 
-       cxgbe_dev_info_get(eth_dev, &dev_info);
+       err = cxgbe_dev_info_get(eth_dev, &dev_info);
+       if (err != 0) {
+               dev_err(adap, "%s: error during getting ethernet device info",
+                       __func__);
+               return err;
+       }
 
-       /* Must accommodate at least ETHER_MIN_MTU */
+       /* Must accommodate at least RTE_ETHER_MIN_MTU */
        if ((pkt_len < dev_info.min_rx_bufsize) ||
            (pkt_len > dev_info.max_rx_pktlen)) {
                dev_err(adap, "%s: max pkt len must be > %d and <= %d\n",
@@ -600,7 +634,7 @@ int cxgbe_dev_rx_queue_setup(struct rte_eth_dev *eth_dev,
 
        /*  Free up the existing queue  */
        if (eth_dev->data->rx_queues[queue_idx]) {
-               cxgbe_dev_rx_queue_release(eth_dev->data->rx_queues[queue_idx]);
+               cxgbe_dev_rx_queue_release(eth_dev, queue_idx);
                eth_dev->data->rx_queues[queue_idx] = NULL;
        }
 
@@ -624,19 +658,10 @@ int cxgbe_dev_rx_queue_setup(struct rte_eth_dev *eth_dev,
        }
 
        rxq->rspq.size = temp_nb_desc;
-       if ((&rxq->fl) != NULL)
-               rxq->fl.size = temp_nb_desc;
-
-       /* Set to jumbo mode if necessary */
-       if (pkt_len > ETHER_MAX_LEN)
-               eth_dev->data->dev_conf.rxmode.offloads |=
-                       DEV_RX_OFFLOAD_JUMBO_FRAME;
-       else
-               eth_dev->data->dev_conf.rxmode.offloads &=
-                       ~DEV_RX_OFFLOAD_JUMBO_FRAME;
+       rxq->fl.size = temp_nb_desc;
 
        err = t4_sge_alloc_rxq(adapter, &rxq->rspq, false, eth_dev, msi_idx,
-                              &rxq->fl, t4_ethrx_handler,
+                              &rxq->fl, NULL,
                               is_pf4(adapter) ?
                               t4_get_tp_ch_map(adapter, pi->tx_chan) : 0, mp,
                               queue_idx, socket_id);
@@ -647,14 +672,13 @@ int cxgbe_dev_rx_queue_setup(struct rte_eth_dev *eth_dev,
        return err;
 }
 
-void cxgbe_dev_rx_queue_release(void *q)
+void cxgbe_dev_rx_queue_release(struct rte_eth_dev *eth_dev, uint16_t qid)
 {
-       struct sge_eth_rxq *rxq = (struct sge_eth_rxq *)q;
-       struct sge_rspq *rq = &rxq->rspq;
+       struct sge_eth_rxq *rxq = eth_dev->data->rx_queues[qid];
 
-       if (rq) {
+       if (rxq) {
                struct port_info *pi = (struct port_info *)
-                                      (rq->eth_dev->data->dev_private);
+                                      (rxq->rspq.eth_dev->data->dev_private);
                struct adapter *adap = pi->adapter;
 
                dev_debug(adapter, "%s: pi->port_id = %d; rx_queue_id = %d\n",
@@ -670,7 +694,7 @@ void cxgbe_dev_rx_queue_release(void *q)
 static int cxgbe_dev_stats_get(struct rte_eth_dev *eth_dev,
                                struct rte_eth_stats *eth_stats)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        struct sge *s = &adapter->sge;
        struct port_stats ps;
@@ -693,104 +717,416 @@ static int cxgbe_dev_stats_get(struct rte_eth_dev *eth_dev,
        eth_stats->oerrors  = ps.tx_error_frames;
 
        for (i = 0; i < pi->n_rx_qsets; i++) {
-               struct sge_eth_rxq *rxq =
-                       &s->ethrxq[pi->first_qset + i];
+               struct sge_eth_rxq *rxq = &s->ethrxq[pi->first_rxqset + i];
 
-               eth_stats->q_ipackets[i] = rxq->stats.pkts;
-               eth_stats->q_ibytes[i] = rxq->stats.rx_bytes;
-               eth_stats->ipackets += eth_stats->q_ipackets[i];
-               eth_stats->ibytes += eth_stats->q_ibytes[i];
+               eth_stats->ipackets += rxq->stats.pkts;
+               eth_stats->ibytes += rxq->stats.rx_bytes;
        }
 
-       for (i = 0; i < pi->n_tx_qsets; i++) {
-               struct sge_eth_txq *txq =
-                       &s->ethtxq[pi->first_qset + i];
-
-               eth_stats->q_opackets[i] = txq->stats.pkts;
-               eth_stats->q_obytes[i] = txq->stats.tx_bytes;
-               eth_stats->q_errors[i] = txq->stats.mapping_err;
-       }
        return 0;
 }
 
 /*
  * Reset port statistics.
  */
-static void cxgbe_dev_stats_reset(struct rte_eth_dev *eth_dev)
+static int cxgbe_dev_stats_reset(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        struct sge *s = &adapter->sge;
        unsigned int i;
 
        cxgbe_stats_reset(pi);
        for (i = 0; i < pi->n_rx_qsets; i++) {
-               struct sge_eth_rxq *rxq =
-                       &s->ethrxq[pi->first_qset + i];
+               struct sge_eth_rxq *rxq = &s->ethrxq[pi->first_rxqset + i];
 
-               rxq->stats.pkts = 0;
-               rxq->stats.rx_bytes = 0;
+               memset(&rxq->stats, 0, sizeof(rxq->stats));
        }
        for (i = 0; i < pi->n_tx_qsets; i++) {
-               struct sge_eth_txq *txq =
-                       &s->ethtxq[pi->first_qset + i];
+               struct sge_eth_txq *txq = &s->ethtxq[pi->first_txqset + i];
+
+               memset(&txq->stats, 0, sizeof(txq->stats));
+       }
+
+       return 0;
+}
+
+/* Store extended statistics names and its offset in stats structure  */
+struct cxgbe_dev_xstats_name_off {
+       char name[RTE_ETH_XSTATS_NAME_SIZE];
+       unsigned int offset;
+};
+
+static const struct cxgbe_dev_xstats_name_off cxgbe_dev_rxq_stats_strings[] = {
+       {"packets", offsetof(struct sge_eth_rx_stats, pkts)},
+       {"bytes", offsetof(struct sge_eth_rx_stats, rx_bytes)},
+       {"checksum_offloads", offsetof(struct sge_eth_rx_stats, rx_cso)},
+       {"vlan_extractions", offsetof(struct sge_eth_rx_stats, vlan_ex)},
+       {"dropped_packets", offsetof(struct sge_eth_rx_stats, rx_drops)},
+};
+
+static const struct cxgbe_dev_xstats_name_off cxgbe_dev_txq_stats_strings[] = {
+       {"packets", offsetof(struct sge_eth_tx_stats, pkts)},
+       {"bytes", offsetof(struct sge_eth_tx_stats, tx_bytes)},
+       {"tso_requests", offsetof(struct sge_eth_tx_stats, tso)},
+       {"checksum_offloads", offsetof(struct sge_eth_tx_stats, tx_cso)},
+       {"vlan_insertions", offsetof(struct sge_eth_tx_stats, vlan_ins)},
+       {"packet_mapping_errors",
+        offsetof(struct sge_eth_tx_stats, mapping_err)},
+       {"coalesced_wrs", offsetof(struct sge_eth_tx_stats, coal_wr)},
+       {"coalesced_packets", offsetof(struct sge_eth_tx_stats, coal_pkts)},
+};
+
+static const struct cxgbe_dev_xstats_name_off cxgbe_dev_port_stats_strings[] = {
+       {"tx_bytes", offsetof(struct port_stats, tx_octets)},
+       {"tx_packets", offsetof(struct port_stats, tx_frames)},
+       {"tx_broadcast_packets", offsetof(struct port_stats, tx_bcast_frames)},
+       {"tx_multicast_packets", offsetof(struct port_stats, tx_mcast_frames)},
+       {"tx_unicast_packets", offsetof(struct port_stats, tx_ucast_frames)},
+       {"tx_error_packets", offsetof(struct port_stats, tx_error_frames)},
+       {"tx_size_64_packets", offsetof(struct port_stats, tx_frames_64)},
+       {"tx_size_65_to_127_packets",
+        offsetof(struct port_stats, tx_frames_65_127)},
+       {"tx_size_128_to_255_packets",
+        offsetof(struct port_stats, tx_frames_128_255)},
+       {"tx_size_256_to_511_packets",
+        offsetof(struct port_stats, tx_frames_256_511)},
+       {"tx_size_512_to_1023_packets",
+        offsetof(struct port_stats, tx_frames_512_1023)},
+       {"tx_size_1024_to_1518_packets",
+        offsetof(struct port_stats, tx_frames_1024_1518)},
+       {"tx_size_1519_to_max_packets",
+        offsetof(struct port_stats, tx_frames_1519_max)},
+       {"tx_drop_packets", offsetof(struct port_stats, tx_drop)},
+       {"tx_pause_frames", offsetof(struct port_stats, tx_pause)},
+       {"tx_ppp_pri0_packets", offsetof(struct port_stats, tx_ppp0)},
+       {"tx_ppp_pri1_packets", offsetof(struct port_stats, tx_ppp1)},
+       {"tx_ppp_pri2_packets", offsetof(struct port_stats, tx_ppp2)},
+       {"tx_ppp_pri3_packets", offsetof(struct port_stats, tx_ppp3)},
+       {"tx_ppp_pri4_packets", offsetof(struct port_stats, tx_ppp4)},
+       {"tx_ppp_pri5_packets", offsetof(struct port_stats, tx_ppp5)},
+       {"tx_ppp_pri6_packets", offsetof(struct port_stats, tx_ppp6)},
+       {"tx_ppp_pri7_packets", offsetof(struct port_stats, tx_ppp7)},
+       {"rx_bytes", offsetof(struct port_stats, rx_octets)},
+       {"rx_packets", offsetof(struct port_stats, rx_frames)},
+       {"rx_broadcast_packets", offsetof(struct port_stats, rx_bcast_frames)},
+       {"rx_multicast_packets", offsetof(struct port_stats, rx_mcast_frames)},
+       {"rx_unicast_packets", offsetof(struct port_stats, rx_ucast_frames)},
+       {"rx_too_long_packets", offsetof(struct port_stats, rx_too_long)},
+       {"rx_jabber_packets", offsetof(struct port_stats, rx_jabber)},
+       {"rx_fcs_error_packets", offsetof(struct port_stats, rx_fcs_err)},
+       {"rx_length_error_packets", offsetof(struct port_stats, rx_len_err)},
+       {"rx_symbol_error_packets",
+        offsetof(struct port_stats, rx_symbol_err)},
+       {"rx_short_packets", offsetof(struct port_stats, rx_runt)},
+       {"rx_size_64_packets", offsetof(struct port_stats, rx_frames_64)},
+       {"rx_size_65_to_127_packets",
+        offsetof(struct port_stats, rx_frames_65_127)},
+       {"rx_size_128_to_255_packets",
+        offsetof(struct port_stats, rx_frames_128_255)},
+       {"rx_size_256_to_511_packets",
+        offsetof(struct port_stats, rx_frames_256_511)},
+       {"rx_size_512_to_1023_packets",
+        offsetof(struct port_stats, rx_frames_512_1023)},
+       {"rx_size_1024_to_1518_packets",
+        offsetof(struct port_stats, rx_frames_1024_1518)},
+       {"rx_size_1519_to_max_packets",
+        offsetof(struct port_stats, rx_frames_1519_max)},
+       {"rx_pause_packets", offsetof(struct port_stats, rx_pause)},
+       {"rx_ppp_pri0_packets", offsetof(struct port_stats, rx_ppp0)},
+       {"rx_ppp_pri1_packets", offsetof(struct port_stats, rx_ppp1)},
+       {"rx_ppp_pri2_packets", offsetof(struct port_stats, rx_ppp2)},
+       {"rx_ppp_pri3_packets", offsetof(struct port_stats, rx_ppp3)},
+       {"rx_ppp_pri4_packets", offsetof(struct port_stats, rx_ppp4)},
+       {"rx_ppp_pri5_packets", offsetof(struct port_stats, rx_ppp5)},
+       {"rx_ppp_pri6_packets", offsetof(struct port_stats, rx_ppp6)},
+       {"rx_ppp_pri7_packets", offsetof(struct port_stats, rx_ppp7)},
+       {"rx_bg0_dropped_packets", offsetof(struct port_stats, rx_ovflow0)},
+       {"rx_bg1_dropped_packets", offsetof(struct port_stats, rx_ovflow1)},
+       {"rx_bg2_dropped_packets", offsetof(struct port_stats, rx_ovflow2)},
+       {"rx_bg3_dropped_packets", offsetof(struct port_stats, rx_ovflow3)},
+       {"rx_bg0_truncated_packets", offsetof(struct port_stats, rx_trunc0)},
+       {"rx_bg1_truncated_packets", offsetof(struct port_stats, rx_trunc1)},
+       {"rx_bg2_truncated_packets", offsetof(struct port_stats, rx_trunc2)},
+       {"rx_bg3_truncated_packets", offsetof(struct port_stats, rx_trunc3)},
+};
+
+static const struct cxgbe_dev_xstats_name_off
+cxgbevf_dev_port_stats_strings[] = {
+       {"tx_bytes", offsetof(struct port_stats, tx_octets)},
+       {"tx_broadcast_packets", offsetof(struct port_stats, tx_bcast_frames)},
+       {"tx_multicast_packets", offsetof(struct port_stats, tx_mcast_frames)},
+       {"tx_unicast_packets", offsetof(struct port_stats, tx_ucast_frames)},
+       {"tx_drop_packets", offsetof(struct port_stats, tx_drop)},
+       {"rx_broadcast_packets", offsetof(struct port_stats, rx_bcast_frames)},
+       {"rx_multicast_packets", offsetof(struct port_stats, rx_mcast_frames)},
+       {"rx_unicast_packets", offsetof(struct port_stats, rx_ucast_frames)},
+       {"rx_length_error_packets", offsetof(struct port_stats, rx_len_err)},
+};
+
+#define CXGBE_NB_RXQ_STATS RTE_DIM(cxgbe_dev_rxq_stats_strings)
+#define CXGBE_NB_TXQ_STATS RTE_DIM(cxgbe_dev_txq_stats_strings)
+#define CXGBE_NB_PORT_STATS RTE_DIM(cxgbe_dev_port_stats_strings)
+#define CXGBEVF_NB_PORT_STATS RTE_DIM(cxgbevf_dev_port_stats_strings)
+
+static u16 cxgbe_dev_xstats_count(struct port_info *pi)
+{
+       u16 count;
+
+       count = (pi->n_tx_qsets * CXGBE_NB_TXQ_STATS) +
+               (pi->n_rx_qsets * CXGBE_NB_RXQ_STATS);
+
+       if (is_pf4(pi->adapter) != 0)
+               count += CXGBE_NB_PORT_STATS;
+       else
+               count += CXGBEVF_NB_PORT_STATS;
+
+       return count;
+}
+
+static int cxgbe_dev_xstats(struct rte_eth_dev *dev,
+                           struct rte_eth_xstat_name *xstats_names,
+                           struct rte_eth_xstat *xstats, unsigned int size)
+{
+       const struct cxgbe_dev_xstats_name_off *xstats_str;
+       struct port_info *pi = dev->data->dev_private;
+       struct adapter *adap = pi->adapter;
+       struct sge *s = &adap->sge;
+       u16 count, i, qid, nstats;
+       struct port_stats ps;
+       u64 *stats_ptr;
+
+       count = cxgbe_dev_xstats_count(pi);
+       if (size < count)
+               return count;
+
+       if (is_pf4(adap) != 0) {
+               /* port stats for PF*/
+               cxgbe_stats_get(pi, &ps);
+               xstats_str = cxgbe_dev_port_stats_strings;
+               nstats = CXGBE_NB_PORT_STATS;
+       } else {
+               /* port stats for VF*/
+               cxgbevf_stats_get(pi, &ps);
+               xstats_str = cxgbevf_dev_port_stats_strings;
+               nstats = CXGBEVF_NB_PORT_STATS;
+       }
+
+       count = 0;
+       for (i = 0; i < nstats; i++, count++) {
+               if (xstats_names != NULL)
+                       snprintf(xstats_names[count].name,
+                                sizeof(xstats_names[count].name),
+                                "%s", xstats_str[i].name);
+               if (xstats != NULL) {
+                       stats_ptr = RTE_PTR_ADD(&ps,
+                                               xstats_str[i].offset);
+                       xstats[count].value = *stats_ptr;
+                       xstats[count].id = count;
+               }
+       }
+
+       /* per-txq stats */
+       xstats_str = cxgbe_dev_txq_stats_strings;
+       for (qid = 0; qid < pi->n_tx_qsets; qid++) {
+               struct sge_eth_txq *txq = &s->ethtxq[pi->first_txqset + qid];
+
+               for (i = 0; i < CXGBE_NB_TXQ_STATS; i++, count++) {
+                       if (xstats_names != NULL)
+                               snprintf(xstats_names[count].name,
+                                        sizeof(xstats_names[count].name),
+                                        "tx_q%u_%s",
+                                        qid, xstats_str[i].name);
+                       if (xstats != NULL) {
+                               stats_ptr = RTE_PTR_ADD(&txq->stats,
+                                                       xstats_str[i].offset);
+                               xstats[count].value = *stats_ptr;
+                               xstats[count].id = count;
+                       }
+               }
+       }
+
+       /* per-rxq stats */
+       xstats_str = cxgbe_dev_rxq_stats_strings;
+       for (qid = 0; qid < pi->n_rx_qsets; qid++) {
+               struct sge_eth_rxq *rxq = &s->ethrxq[pi->first_rxqset + qid];
+
+               for (i = 0; i < CXGBE_NB_RXQ_STATS; i++, count++) {
+                       if (xstats_names != NULL)
+                               snprintf(xstats_names[count].name,
+                                        sizeof(xstats_names[count].name),
+                                        "rx_q%u_%s",
+                                        qid, xstats_str[i].name);
+                       if (xstats != NULL) {
+                               stats_ptr = RTE_PTR_ADD(&rxq->stats,
+                                                       xstats_str[i].offset);
+                               xstats[count].value = *stats_ptr;
+                               xstats[count].id = count;
+                       }
+               }
+       }
+
+       return count;
+}
+
+/* Get port extended statistics by ID. */
+int cxgbe_dev_xstats_get_by_id(struct rte_eth_dev *dev,
+                              const uint64_t *ids, uint64_t *values,
+                              unsigned int n)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct rte_eth_xstat *xstats_copy;
+       u16 count, i;
+       int ret = 0;
+
+       count = cxgbe_dev_xstats_count(pi);
+       if (ids == NULL || values == NULL)
+               return count;
+
+       xstats_copy = rte_calloc(NULL, count, sizeof(*xstats_copy), 0);
+       if (xstats_copy == NULL)
+               return -ENOMEM;
+
+       cxgbe_dev_xstats(dev, NULL, xstats_copy, count);
+
+       for (i = 0; i < n; i++) {
+               if (ids[i] >= count) {
+                       ret = -EINVAL;
+                       goto out_err;
+               }
+               values[i] = xstats_copy[ids[i]].value;
+       }
+
+       ret = n;
+
+out_err:
+       rte_free(xstats_copy);
+       return ret;
+}
+
+/* Get names of port extended statistics by ID. */
+int cxgbe_dev_xstats_get_names_by_id(struct rte_eth_dev *dev,
+                                           const uint64_t *ids,
+                                           struct rte_eth_xstat_name *xnames,
+                                           unsigned int n)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct rte_eth_xstat_name *xnames_copy;
+       u16 count, i;
+       int ret = 0;
+
+       count = cxgbe_dev_xstats_count(pi);
+       if (ids == NULL || xnames == NULL)
+               return count;
+
+       xnames_copy = rte_calloc(NULL, count, sizeof(*xnames_copy), 0);
+       if (xnames_copy == NULL)
+               return -ENOMEM;
 
-               txq->stats.pkts = 0;
-               txq->stats.tx_bytes = 0;
-               txq->stats.mapping_err = 0;
+       cxgbe_dev_xstats(dev, xnames_copy, NULL, count);
+
+       for (i = 0; i < n; i++) {
+               if (ids[i] >= count) {
+                       ret = -EINVAL;
+                       goto out_err;
+               }
+               rte_strlcpy(xnames[i].name, xnames_copy[ids[i]].name,
+                           sizeof(xnames[i].name));
        }
+
+       ret = n;
+
+out_err:
+       rte_free(xnames_copy);
+       return ret;
+}
+
+/* Get port extended statistics. */
+int cxgbe_dev_xstats_get(struct rte_eth_dev *dev,
+                        struct rte_eth_xstat *xstats, unsigned int n)
+{
+       return cxgbe_dev_xstats(dev, NULL, xstats, n);
+}
+
+/* Get names of port extended statistics. */
+int cxgbe_dev_xstats_get_names(struct rte_eth_dev *dev,
+                              struct rte_eth_xstat_name *xstats_names,
+                              unsigned int n)
+{
+       return cxgbe_dev_xstats(dev, xstats_names, NULL, n);
+}
+
+/* Reset port extended statistics. */
+static int cxgbe_dev_xstats_reset(struct rte_eth_dev *dev)
+{
+       return cxgbe_dev_stats_reset(dev);
 }
 
 static int cxgbe_flow_ctrl_get(struct rte_eth_dev *eth_dev,
                               struct rte_eth_fc_conf *fc_conf)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct link_config *lc = &pi->link_cfg;
-       int rx_pause, tx_pause;
+       u8 rx_pause = 0, tx_pause = 0;
+       u32 caps = lc->link_caps;
 
-       fc_conf->autoneg = lc->fc & PAUSE_AUTONEG;
-       rx_pause = lc->fc & PAUSE_RX;
-       tx_pause = lc->fc & PAUSE_TX;
+       if (caps & FW_PORT_CAP32_ANEG)
+               fc_conf->autoneg = 1;
+
+       if (caps & FW_PORT_CAP32_FC_TX)
+               tx_pause = 1;
+
+       if (caps & FW_PORT_CAP32_FC_RX)
+               rx_pause = 1;
 
        if (rx_pause && tx_pause)
-               fc_conf->mode = RTE_FC_FULL;
+               fc_conf->mode = RTE_ETH_FC_FULL;
        else if (rx_pause)
-               fc_conf->mode = RTE_FC_RX_PAUSE;
+               fc_conf->mode = RTE_ETH_FC_RX_PAUSE;
        else if (tx_pause)
-               fc_conf->mode = RTE_FC_TX_PAUSE;
+               fc_conf->mode = RTE_ETH_FC_TX_PAUSE;
        else
-               fc_conf->mode = RTE_FC_NONE;
+               fc_conf->mode = RTE_ETH_FC_NONE;
        return 0;
 }
 
 static int cxgbe_flow_ctrl_set(struct rte_eth_dev *eth_dev,
                               struct rte_eth_fc_conf *fc_conf)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
-       struct adapter *adapter = pi->adapter;
+       struct port_info *pi = eth_dev->data->dev_private;
        struct link_config *lc = &pi->link_cfg;
+       u32 new_caps = lc->admin_caps;
+       u8 tx_pause = 0, rx_pause = 0;
+       int ret;
 
-       if (lc->pcaps & FW_PORT_CAP32_ANEG) {
-               if (fc_conf->autoneg)
-                       lc->requested_fc |= PAUSE_AUTONEG;
-               else
-                       lc->requested_fc &= ~PAUSE_AUTONEG;
+       if (fc_conf->mode == RTE_ETH_FC_FULL) {
+               tx_pause = 1;
+               rx_pause = 1;
+       } else if (fc_conf->mode == RTE_ETH_FC_TX_PAUSE) {
+               tx_pause = 1;
+       } else if (fc_conf->mode == RTE_ETH_FC_RX_PAUSE) {
+               rx_pause = 1;
        }
 
-       if (((fc_conf->mode & RTE_FC_FULL) == RTE_FC_FULL) ||
-           (fc_conf->mode & RTE_FC_RX_PAUSE))
-               lc->requested_fc |= PAUSE_RX;
-       else
-               lc->requested_fc &= ~PAUSE_RX;
+       ret = t4_set_link_pause(pi, fc_conf->autoneg, tx_pause,
+                               rx_pause, &new_caps);
+       if (ret != 0)
+               return ret;
 
-       if (((fc_conf->mode & RTE_FC_FULL) == RTE_FC_FULL) ||
-           (fc_conf->mode & RTE_FC_TX_PAUSE))
-               lc->requested_fc |= PAUSE_TX;
-       else
-               lc->requested_fc &= ~PAUSE_TX;
+       if (!fc_conf->autoneg) {
+               if (lc->pcaps & FW_PORT_CAP32_FORCE_PAUSE)
+                       new_caps |= FW_PORT_CAP32_FORCE_PAUSE;
+       } else {
+               new_caps &= ~FW_PORT_CAP32_FORCE_PAUSE;
+       }
+
+       if (new_caps != lc->admin_caps) {
+               ret = t4_link_l1cfg(pi, new_caps);
+               if (ret == 0)
+                       lc->admin_caps = new_caps;
+       }
 
-       return t4_link_l1cfg(adapter, adapter->mbox, pi->tx_chan,
-                            &pi->link_cfg);
+       return ret;
 }
 
 const uint32_t *
@@ -812,7 +1148,7 @@ cxgbe_dev_supported_ptypes_get(struct rte_eth_dev *eth_dev)
 static int cxgbe_dev_rss_hash_update(struct rte_eth_dev *dev,
                                     struct rte_eth_rss_conf *rss_conf)
 {
-       struct port_info *pi = (struct port_info *)(dev->data->dev_private);
+       struct port_info *pi = dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        int err;
 
@@ -842,7 +1178,7 @@ static int cxgbe_dev_rss_hash_update(struct rte_eth_dev *dev,
 static int cxgbe_dev_rss_hash_conf_get(struct rte_eth_dev *dev,
                                       struct rte_eth_rss_conf *rss_conf)
 {
-       struct port_info *pi = (struct port_info *)(dev->data->dev_private);
+       struct port_info *pi = dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        u64 rss_hf = 0;
        u64 flags = 0;
@@ -855,22 +1191,22 @@ static int cxgbe_dev_rss_hash_conf_get(struct rte_eth_dev *dev,
                return err;
 
        if (flags & F_FW_RSS_VI_CONFIG_CMD_IP6FOURTUPEN) {
-               rss_hf |= ETH_RSS_NONFRAG_IPV6_TCP;
+               rss_hf |= CXGBE_RSS_HF_TCP_IPV6_MASK;
                if (flags & F_FW_RSS_VI_CONFIG_CMD_UDPEN)
-                       rss_hf |= ETH_RSS_NONFRAG_IPV6_UDP;
+                       rss_hf |= CXGBE_RSS_HF_UDP_IPV6_MASK;
        }
 
        if (flags & F_FW_RSS_VI_CONFIG_CMD_IP6TWOTUPEN)
-               rss_hf |= ETH_RSS_IPV6;
+               rss_hf |= CXGBE_RSS_HF_IPV6_MASK;
 
        if (flags & F_FW_RSS_VI_CONFIG_CMD_IP4FOURTUPEN) {
-               rss_hf |= ETH_RSS_NONFRAG_IPV4_TCP;
+               rss_hf |= RTE_ETH_RSS_NONFRAG_IPV4_TCP;
                if (flags & F_FW_RSS_VI_CONFIG_CMD_UDPEN)
-                       rss_hf |= ETH_RSS_NONFRAG_IPV4_UDP;
+                       rss_hf |= RTE_ETH_RSS_NONFRAG_IPV4_UDP;
        }
 
        if (flags & F_FW_RSS_VI_CONFIG_CMD_IP4TWOTUPEN)
-               rss_hf |= ETH_RSS_IPV4;
+               rss_hf |= CXGBE_RSS_HF_IPV4_MASK;
 
        rss_conf->rss_hf = rss_hf;
 
@@ -889,6 +1225,69 @@ static int cxgbe_dev_rss_hash_conf_get(struct rte_eth_dev *dev,
        return 0;
 }
 
+static int cxgbe_dev_rss_reta_update(struct rte_eth_dev *dev,
+                                    struct rte_eth_rss_reta_entry64 *reta_conf,
+                                    uint16_t reta_size)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct adapter *adapter = pi->adapter;
+       u16 i, idx, shift, *rss;
+       int ret;
+
+       if (!(adapter->flags & FULL_INIT_DONE))
+               return -ENOMEM;
+
+       if (!reta_size || reta_size > pi->rss_size)
+               return -EINVAL;
+
+       rss = rte_calloc(NULL, pi->rss_size, sizeof(u16), 0);
+       if (!rss)
+               return -ENOMEM;
+
+       rte_memcpy(rss, pi->rss, pi->rss_size * sizeof(u16));
+       for (i = 0; i < reta_size; i++) {
+               idx = i / RTE_ETH_RETA_GROUP_SIZE;
+               shift = i % RTE_ETH_RETA_GROUP_SIZE;
+               if (!(reta_conf[idx].mask & (1ULL << shift)))
+                       continue;
+
+               rss[i] = reta_conf[idx].reta[shift];
+       }
+
+       ret = cxgbe_write_rss(pi, rss);
+       if (!ret)
+               rte_memcpy(pi->rss, rss, pi->rss_size * sizeof(u16));
+
+       rte_free(rss);
+       return ret;
+}
+
+static int cxgbe_dev_rss_reta_query(struct rte_eth_dev *dev,
+                                   struct rte_eth_rss_reta_entry64 *reta_conf,
+                                   uint16_t reta_size)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct adapter *adapter = pi->adapter;
+       u16 i, idx, shift;
+
+       if (!(adapter->flags & FULL_INIT_DONE))
+               return -ENOMEM;
+
+       if (!reta_size || reta_size > pi->rss_size)
+               return -EINVAL;
+
+       for (i = 0; i < reta_size; i++) {
+               idx = i / RTE_ETH_RETA_GROUP_SIZE;
+               shift = i % RTE_ETH_RETA_GROUP_SIZE;
+               if (!(reta_conf[idx].mask & (1ULL << shift)))
+                       continue;
+
+               reta_conf[idx].reta[shift] = pi->rss[i];
+       }
+
+       return 0;
+}
+
 static int cxgbe_get_eeprom_length(struct rte_eth_dev *dev)
 {
        RTE_SET_USED(dev);
@@ -951,7 +1350,7 @@ static int eeprom_wr_phys(struct adapter *adap, unsigned int phys_addr, u32 v)
 static int cxgbe_get_eeprom(struct rte_eth_dev *dev,
                            struct rte_dev_eeprom_info *e)
 {
-       struct port_info *pi = (struct port_info *)(dev->data->dev_private);
+       struct port_info *pi = dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        u32 i, err = 0;
        u8 *buf = rte_zmalloc(NULL, EEPROMSIZE, 0);
@@ -972,7 +1371,7 @@ static int cxgbe_get_eeprom(struct rte_eth_dev *dev,
 static int cxgbe_set_eeprom(struct rte_eth_dev *dev,
                            struct rte_dev_eeprom_info *eeprom)
 {
-       struct port_info *pi = (struct port_info *)(dev->data->dev_private);
+       struct port_info *pi = dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
        u8 *buf;
        int err = 0;
@@ -1030,7 +1429,7 @@ out:
 
 static int cxgbe_get_regs_len(struct rte_eth_dev *eth_dev)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
 
        return t4_get_regs_len(adapter) / sizeof(uint32_t);
@@ -1039,7 +1438,7 @@ static int cxgbe_get_regs_len(struct rte_eth_dev *eth_dev)
 static int cxgbe_get_regs(struct rte_eth_dev *eth_dev,
                          struct rte_dev_reg_info *regs)
 {
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = pi->adapter;
 
        regs->version = CHELSIO_CHIP_VERSION(adapter->params.chip) |
@@ -1058,20 +1457,163 @@ static int cxgbe_get_regs(struct rte_eth_dev *eth_dev,
        return 0;
 }
 
-void cxgbe_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *addr)
+int cxgbe_mac_addr_set(struct rte_eth_dev *dev, struct rte_ether_addr *addr)
 {
-       struct port_info *pi = (struct port_info *)(dev->data->dev_private);
-       struct adapter *adapter = pi->adapter;
+       struct port_info *pi = dev->data->dev_private;
        int ret;
 
-       ret = t4_change_mac(adapter, adapter->mbox, pi->viid,
-                           pi->xact_addr_filt, (u8 *)addr, true, true);
+       ret = cxgbe_mpstcam_modify(pi, (int)pi->xact_addr_filt, (u8 *)addr);
        if (ret < 0) {
                dev_err(adapter, "failed to set mac addr; err = %d\n",
                        ret);
-               return;
+               return ret;
        }
        pi->xact_addr_filt = ret;
+       return 0;
+}
+
+static int cxgbe_fec_get_capa_speed_to_fec(struct link_config *lc,
+                                          struct rte_eth_fec_capa *capa_arr)
+{
+       int num = 0;
+
+       if (lc->pcaps & FW_PORT_CAP32_SPEED_100G) {
+               if (capa_arr) {
+                       capa_arr[num].speed = RTE_ETH_SPEED_NUM_100G;
+                       capa_arr[num].capa = RTE_ETH_FEC_MODE_CAPA_MASK(NOFEC) |
+                                            RTE_ETH_FEC_MODE_CAPA_MASK(RS);
+               }
+               num++;
+       }
+
+       if (lc->pcaps & FW_PORT_CAP32_SPEED_50G) {
+               if (capa_arr) {
+                       capa_arr[num].speed = RTE_ETH_SPEED_NUM_50G;
+                       capa_arr[num].capa = RTE_ETH_FEC_MODE_CAPA_MASK(NOFEC) |
+                                            RTE_ETH_FEC_MODE_CAPA_MASK(BASER);
+               }
+               num++;
+       }
+
+       if (lc->pcaps & FW_PORT_CAP32_SPEED_25G) {
+               if (capa_arr) {
+                       capa_arr[num].speed = RTE_ETH_SPEED_NUM_25G;
+                       capa_arr[num].capa = RTE_ETH_FEC_MODE_CAPA_MASK(NOFEC) |
+                                            RTE_ETH_FEC_MODE_CAPA_MASK(BASER) |
+                                            RTE_ETH_FEC_MODE_CAPA_MASK(RS);
+               }
+               num++;
+       }
+
+       return num;
+}
+
+static int cxgbe_fec_get_capability(struct rte_eth_dev *dev,
+                                   struct rte_eth_fec_capa *speed_fec_capa,
+                                   unsigned int num)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct link_config *lc = &pi->link_cfg;
+       u8 num_entries;
+
+       if (!(lc->pcaps & V_FW_PORT_CAP32_FEC(M_FW_PORT_CAP32_FEC)))
+               return -EOPNOTSUPP;
+
+       num_entries = cxgbe_fec_get_capa_speed_to_fec(lc, NULL);
+       if (!speed_fec_capa || num < num_entries)
+               return num_entries;
+
+       return cxgbe_fec_get_capa_speed_to_fec(lc, speed_fec_capa);
+}
+
+static int cxgbe_fec_get(struct rte_eth_dev *dev, uint32_t *fec_capa)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct link_config *lc = &pi->link_cfg;
+       u32 fec_caps = 0, caps = lc->link_caps;
+
+       if (!(lc->pcaps & V_FW_PORT_CAP32_FEC(M_FW_PORT_CAP32_FEC)))
+               return -EOPNOTSUPP;
+
+       if (caps & FW_PORT_CAP32_FEC_RS)
+               fec_caps = RTE_ETH_FEC_MODE_CAPA_MASK(RS);
+       else if (caps & FW_PORT_CAP32_FEC_BASER_RS)
+               fec_caps = RTE_ETH_FEC_MODE_CAPA_MASK(BASER);
+       else
+               fec_caps = RTE_ETH_FEC_MODE_CAPA_MASK(NOFEC);
+
+       *fec_capa = fec_caps;
+       return 0;
+}
+
+static int cxgbe_fec_set(struct rte_eth_dev *dev, uint32_t fec_capa)
+{
+       struct port_info *pi = dev->data->dev_private;
+       u8 fec_rs = 0, fec_baser = 0, fec_none = 0;
+       struct link_config *lc = &pi->link_cfg;
+       u32 new_caps = lc->admin_caps;
+       int ret;
+
+       if (!(lc->pcaps & V_FW_PORT_CAP32_FEC(M_FW_PORT_CAP32_FEC)))
+               return -EOPNOTSUPP;
+
+       if (!fec_capa)
+               return -EINVAL;
+
+       if (fec_capa & RTE_ETH_FEC_MODE_CAPA_MASK(AUTO))
+               goto set_fec;
+
+       if (fec_capa & RTE_ETH_FEC_MODE_CAPA_MASK(NOFEC))
+               fec_none = 1;
+
+       if (fec_capa & RTE_ETH_FEC_MODE_CAPA_MASK(BASER))
+               fec_baser = 1;
+
+       if (fec_capa & RTE_ETH_FEC_MODE_CAPA_MASK(RS))
+               fec_rs = 1;
+
+set_fec:
+       ret = t4_set_link_fec(pi, fec_rs, fec_baser, fec_none, &new_caps);
+       if (ret != 0)
+               return ret;
+
+       if (lc->pcaps & FW_PORT_CAP32_FORCE_FEC)
+               new_caps |= FW_PORT_CAP32_FORCE_FEC;
+       else
+               new_caps &= ~FW_PORT_CAP32_FORCE_FEC;
+
+       if (new_caps != lc->admin_caps) {
+               ret = t4_link_l1cfg(pi, new_caps);
+               if (ret == 0)
+                       lc->admin_caps = new_caps;
+       }
+
+       return ret;
+}
+
+int cxgbe_fw_version_get(struct rte_eth_dev *dev, char *fw_version,
+                        size_t fw_size)
+{
+       struct port_info *pi = dev->data->dev_private;
+       struct adapter *adapter = pi->adapter;
+       int ret;
+
+       if (adapter->params.fw_vers == 0)
+               return -EIO;
+
+       ret = snprintf(fw_version, fw_size, "%u.%u.%u.%u",
+                      G_FW_HDR_FW_VER_MAJOR(adapter->params.fw_vers),
+                      G_FW_HDR_FW_VER_MINOR(adapter->params.fw_vers),
+                      G_FW_HDR_FW_VER_MICRO(adapter->params.fw_vers),
+                      G_FW_HDR_FW_VER_BUILD(adapter->params.fw_vers));
+       if (ret < 0)
+               return -EINVAL;
+
+       ret += 1;
+       if (fw_size < (size_t)ret)
+               return ret;
+
+       return 0;
 }
 
 static const struct eth_dev_ops cxgbe_eth_dev_ops = {
@@ -1086,6 +1628,8 @@ static const struct eth_dev_ops cxgbe_eth_dev_ops = {
        .dev_infos_get          = cxgbe_dev_info_get,
        .dev_supported_ptypes_get = cxgbe_dev_supported_ptypes_get,
        .link_update            = cxgbe_dev_link_update,
+       .dev_set_link_up        = cxgbe_dev_set_link_up,
+       .dev_set_link_down      = cxgbe_dev_set_link_down,
        .mtu_set                = cxgbe_dev_mtu_set,
        .tx_queue_setup         = cxgbe_dev_tx_queue_setup,
        .tx_queue_start         = cxgbe_dev_tx_queue_start,
@@ -1095,8 +1639,14 @@ static const struct eth_dev_ops cxgbe_eth_dev_ops = {
        .rx_queue_start         = cxgbe_dev_rx_queue_start,
        .rx_queue_stop          = cxgbe_dev_rx_queue_stop,
        .rx_queue_release       = cxgbe_dev_rx_queue_release,
+       .flow_ops_get           = cxgbe_dev_flow_ops_get,
        .stats_get              = cxgbe_dev_stats_get,
        .stats_reset            = cxgbe_dev_stats_reset,
+       .xstats_get             = cxgbe_dev_xstats_get,
+       .xstats_get_by_id       = cxgbe_dev_xstats_get_by_id,
+       .xstats_get_names       = cxgbe_dev_xstats_get_names,
+       .xstats_get_names_by_id = cxgbe_dev_xstats_get_names_by_id,
+       .xstats_reset           = cxgbe_dev_xstats_reset,
        .flow_ctrl_get          = cxgbe_flow_ctrl_get,
        .flow_ctrl_set          = cxgbe_flow_ctrl_set,
        .get_eeprom_length      = cxgbe_get_eeprom_length,
@@ -1106,6 +1656,12 @@ static const struct eth_dev_ops cxgbe_eth_dev_ops = {
        .rss_hash_update        = cxgbe_dev_rss_hash_update,
        .rss_hash_conf_get      = cxgbe_dev_rss_hash_conf_get,
        .mac_addr_set           = cxgbe_mac_addr_set,
+       .reta_update            = cxgbe_dev_rss_reta_update,
+       .reta_query             = cxgbe_dev_rss_reta_query,
+       .fec_get_capability     = cxgbe_fec_get_capability,
+       .fec_get                = cxgbe_fec_get,
+       .fec_set                = cxgbe_fec_set,
+       .fw_version_get         = cxgbe_fw_version_get,
 };
 
 /*
@@ -1115,7 +1671,7 @@ static const struct eth_dev_ops cxgbe_eth_dev_ops = {
 static int eth_cxgbe_dev_init(struct rte_eth_dev *eth_dev)
 {
        struct rte_pci_device *pci_dev;
-       struct port_info *pi = (struct port_info *)(eth_dev->data->dev_private);
+       struct port_info *pi = eth_dev->data->dev_private;
        struct adapter *adapter = NULL;
        char name[RTE_ETH_NAME_MAX_LEN];
        int err = 0;
@@ -1148,6 +1704,7 @@ static int eth_cxgbe_dev_init(struct rte_eth_dev *eth_dev)
                                        eth_dev->rx_pkt_burst;
                                rest_eth_dev->tx_pkt_burst =
                                        eth_dev->tx_pkt_burst;
+                               rte_eth_dev_probing_finish(rest_eth_dev);
                        }
                }
                return 0;
@@ -1169,6 +1726,8 @@ static int eth_cxgbe_dev_init(struct rte_eth_dev *eth_dev)
        adapter->eth_dev = eth_dev;
        pi->adapter = adapter;
 
+       cxgbe_process_devargs(adapter);
+
        err = cxgbe_probe(adapter);
        if (err) {
                dev_err(adapter, "%s: cxgbe probe failed with err %d\n",
@@ -1183,6 +1742,19 @@ out_free_adapter:
        return err;
 }
 
+static int eth_cxgbe_dev_uninit(struct rte_eth_dev *eth_dev)
+{
+       struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev);
+       uint16_t port_id;
+       int err = 0;
+
+       /* Free up other ports and all resources */
+       RTE_ETH_FOREACH_DEV_OF(port_id, &pci_dev->device)
+               err |= rte_eth_dev_close(port_id);
+
+       return err == 0 ? 0 : -EIO;
+}
+
 static int eth_cxgbe_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
        struct rte_pci_device *pci_dev)
 {
@@ -1192,7 +1764,7 @@ static int eth_cxgbe_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 
 static int eth_cxgbe_pci_remove(struct rte_pci_device *pci_dev)
 {
-       return rte_eth_dev_pci_generic_remove(pci_dev, NULL);
+       return rte_eth_dev_pci_generic_remove(pci_dev, eth_cxgbe_dev_uninit);
 }
 
 static struct rte_pci_driver rte_cxgbe_pmd = {
@@ -1205,3 +1777,10 @@ static struct rte_pci_driver rte_cxgbe_pmd = {
 RTE_PMD_REGISTER_PCI(net_cxgbe, rte_cxgbe_pmd);
 RTE_PMD_REGISTER_PCI_TABLE(net_cxgbe, cxgb4_pci_tbl);
 RTE_PMD_REGISTER_KMOD_DEP(net_cxgbe, "* igb_uio | uio_pci_generic | vfio-pci");
+RTE_PMD_REGISTER_PARAM_STRING(net_cxgbe,
+                             CXGBE_DEVARG_CMN_KEEP_OVLAN "=<0|1> "
+                             CXGBE_DEVARG_CMN_TX_MODE_LATENCY "=<0|1> "
+                             CXGBE_DEVARG_PF_FILTER_MODE "=<uint32> "
+                             CXGBE_DEVARG_PF_FILTER_MASK "=<uint32> ");
+RTE_LOG_REGISTER_DEFAULT(cxgbe_logtype, NOTICE);
+RTE_LOG_REGISTER_SUFFIX(cxgbe_mbox_logtype, mbox, NOTICE);