net: add rte prefix to ether structures
[dpdk.git] / drivers / net / ice / ice_ethdev.c
index f7c0b36..d09d172 100644 (file)
@@ -2,14 +2,23 @@
  * Copyright(c) 2018 Intel Corporation
  */
 
+#include <rte_string_fns.h>
 #include <rte_ethdev_pci.h>
 
+#include <stdio.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
 #include "base/ice_sched.h"
+#include "base/ice_flow.h"
+#include "base/ice_dcb.h"
 #include "ice_ethdev.h"
 #include "ice_rxtx.h"
 
 #define ICE_MAX_QP_NUM "max_queue_pair_num"
 #define ICE_DFLT_OUTER_TAG_TYPE ICE_AQ_VSI_OUTER_TAG_VLAN_9100
+#define ICE_DFLT_PKG_FILE "/lib/firmware/intel/ice/ddp/ice.pkg"
 
 int ice_logtype_init;
 int ice_logtype_driver;
@@ -23,7 +32,57 @@ static void ice_dev_info_get(struct rte_eth_dev *dev,
                             struct rte_eth_dev_info *dev_info);
 static int ice_link_update(struct rte_eth_dev *dev,
                           int wait_to_complete);
+static int ice_dev_set_link_up(struct rte_eth_dev *dev);
+static int ice_dev_set_link_down(struct rte_eth_dev *dev);
+
 static int ice_mtu_set(struct rte_eth_dev *dev, uint16_t mtu);
+static int ice_vlan_offload_set(struct rte_eth_dev *dev, int mask);
+static int ice_vlan_tpid_set(struct rte_eth_dev *dev,
+                            enum rte_vlan_type vlan_type,
+                            uint16_t tpid);
+static int ice_rss_reta_update(struct rte_eth_dev *dev,
+                              struct rte_eth_rss_reta_entry64 *reta_conf,
+                              uint16_t reta_size);
+static int ice_rss_reta_query(struct rte_eth_dev *dev,
+                             struct rte_eth_rss_reta_entry64 *reta_conf,
+                             uint16_t reta_size);
+static int ice_rss_hash_update(struct rte_eth_dev *dev,
+                              struct rte_eth_rss_conf *rss_conf);
+static int ice_rss_hash_conf_get(struct rte_eth_dev *dev,
+                                struct rte_eth_rss_conf *rss_conf);
+static void ice_promisc_enable(struct rte_eth_dev *dev);
+static void ice_promisc_disable(struct rte_eth_dev *dev);
+static void ice_allmulti_enable(struct rte_eth_dev *dev);
+static void ice_allmulti_disable(struct rte_eth_dev *dev);
+static int ice_vlan_filter_set(struct rte_eth_dev *dev,
+                              uint16_t vlan_id,
+                              int on);
+static int ice_macaddr_set(struct rte_eth_dev *dev,
+                          struct rte_ether_addr *mac_addr);
+static int ice_macaddr_add(struct rte_eth_dev *dev,
+                          struct rte_ether_addr *mac_addr,
+                          __rte_unused uint32_t index,
+                          uint32_t pool);
+static void ice_macaddr_remove(struct rte_eth_dev *dev, uint32_t index);
+static int ice_rx_queue_intr_enable(struct rte_eth_dev *dev,
+                                   uint16_t queue_id);
+static int ice_rx_queue_intr_disable(struct rte_eth_dev *dev,
+                                    uint16_t queue_id);
+static int ice_fw_version_get(struct rte_eth_dev *dev, char *fw_version,
+                             size_t fw_size);
+static int ice_vlan_pvid_set(struct rte_eth_dev *dev,
+                            uint16_t pvid, int on);
+static int ice_get_eeprom_length(struct rte_eth_dev *dev);
+static int ice_get_eeprom(struct rte_eth_dev *dev,
+                         struct rte_dev_eeprom_info *eeprom);
+static int ice_stats_get(struct rte_eth_dev *dev,
+                        struct rte_eth_stats *stats);
+static void ice_stats_reset(struct rte_eth_dev *dev);
+static int ice_xstats_get(struct rte_eth_dev *dev,
+                         struct rte_eth_xstat *xstats, unsigned int n);
+static int ice_xstats_get_names(struct rte_eth_dev *dev,
+                               struct rte_eth_xstat_name *xstats_names,
+                               unsigned int limit);
 
 static const struct rte_pci_id pci_id_ice_map[] = {
        { RTE_PCI_DEVICE(ICE_INTEL_VENDOR_ID, ICE_DEV_ID_E810C_BACKPLANE) },
@@ -38,6 +97,8 @@ static const struct eth_dev_ops ice_eth_dev_ops = {
        .dev_stop                     = ice_dev_stop,
        .dev_close                    = ice_dev_close,
        .dev_reset                    = ice_dev_reset,
+       .dev_set_link_up              = ice_dev_set_link_up,
+       .dev_set_link_down            = ice_dev_set_link_down,
        .rx_queue_start               = ice_rx_queue_start,
        .rx_queue_stop                = ice_rx_queue_stop,
        .tx_queue_start               = ice_tx_queue_start,
@@ -50,11 +111,117 @@ static const struct eth_dev_ops ice_eth_dev_ops = {
        .dev_supported_ptypes_get     = ice_dev_supported_ptypes_get,
        .link_update                  = ice_link_update,
        .mtu_set                      = ice_mtu_set,
+       .mac_addr_set                 = ice_macaddr_set,
+       .mac_addr_add                 = ice_macaddr_add,
+       .mac_addr_remove              = ice_macaddr_remove,
+       .vlan_filter_set              = ice_vlan_filter_set,
+       .vlan_offload_set             = ice_vlan_offload_set,
+       .vlan_tpid_set                = ice_vlan_tpid_set,
+       .reta_update                  = ice_rss_reta_update,
+       .reta_query                   = ice_rss_reta_query,
+       .rss_hash_update              = ice_rss_hash_update,
+       .rss_hash_conf_get            = ice_rss_hash_conf_get,
+       .promiscuous_enable           = ice_promisc_enable,
+       .promiscuous_disable          = ice_promisc_disable,
+       .allmulticast_enable          = ice_allmulti_enable,
+       .allmulticast_disable         = ice_allmulti_disable,
+       .rx_queue_intr_enable         = ice_rx_queue_intr_enable,
+       .rx_queue_intr_disable        = ice_rx_queue_intr_disable,
+       .fw_version_get               = ice_fw_version_get,
+       .vlan_pvid_set                = ice_vlan_pvid_set,
        .rxq_info_get                 = ice_rxq_info_get,
        .txq_info_get                 = ice_txq_info_get,
+       .get_eeprom_length            = ice_get_eeprom_length,
+       .get_eeprom                   = ice_get_eeprom,
        .rx_queue_count               = ice_rx_queue_count,
+       .rx_descriptor_status         = ice_rx_descriptor_status,
+       .tx_descriptor_status         = ice_tx_descriptor_status,
+       .stats_get                    = ice_stats_get,
+       .stats_reset                  = ice_stats_reset,
+       .xstats_get                   = ice_xstats_get,
+       .xstats_get_names             = ice_xstats_get_names,
+       .xstats_reset                 = ice_stats_reset,
+};
+
+/* store statistics names and its offset in stats structure */
+struct ice_xstats_name_off {
+       char name[RTE_ETH_XSTATS_NAME_SIZE];
+       unsigned int offset;
+};
+
+static const struct ice_xstats_name_off ice_stats_strings[] = {
+       {"rx_unicast_packets", offsetof(struct ice_eth_stats, rx_unicast)},
+       {"rx_multicast_packets", offsetof(struct ice_eth_stats, rx_multicast)},
+       {"rx_broadcast_packets", offsetof(struct ice_eth_stats, rx_broadcast)},
+       {"rx_dropped", offsetof(struct ice_eth_stats, rx_discards)},
+       {"rx_unknown_protocol_packets", offsetof(struct ice_eth_stats,
+               rx_unknown_protocol)},
+       {"tx_unicast_packets", offsetof(struct ice_eth_stats, tx_unicast)},
+       {"tx_multicast_packets", offsetof(struct ice_eth_stats, tx_multicast)},
+       {"tx_broadcast_packets", offsetof(struct ice_eth_stats, tx_broadcast)},
+       {"tx_dropped", offsetof(struct ice_eth_stats, tx_discards)},
+};
+
+#define ICE_NB_ETH_XSTATS (sizeof(ice_stats_strings) / \
+               sizeof(ice_stats_strings[0]))
+
+static const struct ice_xstats_name_off ice_hw_port_strings[] = {
+       {"tx_link_down_dropped", offsetof(struct ice_hw_port_stats,
+               tx_dropped_link_down)},
+       {"rx_crc_errors", offsetof(struct ice_hw_port_stats, crc_errors)},
+       {"rx_illegal_byte_errors", offsetof(struct ice_hw_port_stats,
+               illegal_bytes)},
+       {"rx_error_bytes", offsetof(struct ice_hw_port_stats, error_bytes)},
+       {"mac_local_errors", offsetof(struct ice_hw_port_stats,
+               mac_local_faults)},
+       {"mac_remote_errors", offsetof(struct ice_hw_port_stats,
+               mac_remote_faults)},
+       {"rx_len_errors", offsetof(struct ice_hw_port_stats,
+               rx_len_errors)},
+       {"tx_xon_packets", offsetof(struct ice_hw_port_stats, link_xon_tx)},
+       {"rx_xon_packets", offsetof(struct ice_hw_port_stats, link_xon_rx)},
+       {"tx_xoff_packets", offsetof(struct ice_hw_port_stats, link_xoff_tx)},
+       {"rx_xoff_packets", offsetof(struct ice_hw_port_stats, link_xoff_rx)},
+       {"rx_size_64_packets", offsetof(struct ice_hw_port_stats, rx_size_64)},
+       {"rx_size_65_to_127_packets", offsetof(struct ice_hw_port_stats,
+               rx_size_127)},
+       {"rx_size_128_to_255_packets", offsetof(struct ice_hw_port_stats,
+               rx_size_255)},
+       {"rx_size_256_to_511_packets", offsetof(struct ice_hw_port_stats,
+               rx_size_511)},
+       {"rx_size_512_to_1023_packets", offsetof(struct ice_hw_port_stats,
+               rx_size_1023)},
+       {"rx_size_1024_to_1522_packets", offsetof(struct ice_hw_port_stats,
+               rx_size_1522)},
+       {"rx_size_1523_to_max_packets", offsetof(struct ice_hw_port_stats,
+               rx_size_big)},
+       {"rx_undersized_errors", offsetof(struct ice_hw_port_stats,
+               rx_undersize)},
+       {"rx_oversize_errors", offsetof(struct ice_hw_port_stats,
+               rx_oversize)},
+       {"rx_mac_short_pkt_dropped", offsetof(struct ice_hw_port_stats,
+               mac_short_pkt_dropped)},
+       {"rx_fragmented_errors", offsetof(struct ice_hw_port_stats,
+               rx_fragments)},
+       {"rx_jabber_errors", offsetof(struct ice_hw_port_stats, rx_jabber)},
+       {"tx_size_64_packets", offsetof(struct ice_hw_port_stats, tx_size_64)},
+       {"tx_size_65_to_127_packets", offsetof(struct ice_hw_port_stats,
+               tx_size_127)},
+       {"tx_size_128_to_255_packets", offsetof(struct ice_hw_port_stats,
+               tx_size_255)},
+       {"tx_size_256_to_511_packets", offsetof(struct ice_hw_port_stats,
+               tx_size_511)},
+       {"tx_size_512_to_1023_packets", offsetof(struct ice_hw_port_stats,
+               tx_size_1023)},
+       {"tx_size_1024_to_1522_packets", offsetof(struct ice_hw_port_stats,
+               tx_size_1522)},
+       {"tx_size_1523_to_max_packets", offsetof(struct ice_hw_port_stats,
+               tx_size_big)},
 };
 
+#define ICE_NB_HW_PORT_XSTATS (sizeof(ice_hw_port_strings) / \
+               sizeof(ice_hw_port_strings[0]))
+
 static void
 ice_init_controlq_parameter(struct ice_hw *hw)
 {
@@ -319,26 +486,446 @@ ice_init_mac_address(struct rte_eth_dev *dev)
        struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
 
        if (!is_unicast_ether_addr
-               ((struct ether_addr *)hw->port_info[0].mac.lan_addr)) {
+               ((struct rte_ether_addr *)hw->port_info[0].mac.lan_addr)) {
                PMD_INIT_LOG(ERR, "Invalid MAC address");
                return -EINVAL;
        }
 
-       ether_addr_copy((struct ether_addr *)hw->port_info[0].mac.lan_addr,
-                       (struct ether_addr *)hw->port_info[0].mac.perm_addr);
+       ether_addr_copy((struct rte_ether_addr *)hw->port_info[0].mac.lan_addr,
+               (struct rte_ether_addr *)hw->port_info[0].mac.perm_addr);
 
-       dev->data->mac_addrs = rte_zmalloc(NULL, sizeof(struct ether_addr), 0);
+       dev->data->mac_addrs =
+               rte_zmalloc(NULL, sizeof(struct rte_ether_addr), 0);
        if (!dev->data->mac_addrs) {
                PMD_INIT_LOG(ERR,
                             "Failed to allocate memory to store mac address");
                return -ENOMEM;
        }
        /* store it to dev data */
-       ether_addr_copy((struct ether_addr *)hw->port_info[0].mac.perm_addr,
+       ether_addr_copy((struct rte_ether_addr *)hw->port_info[0].mac.perm_addr,
                        &dev->data->mac_addrs[0]);
        return 0;
 }
 
+/* Find out specific MAC filter */
+static struct ice_mac_filter *
+ice_find_mac_filter(struct ice_vsi *vsi, struct rte_ether_addr *macaddr)
+{
+       struct ice_mac_filter *f;
+
+       TAILQ_FOREACH(f, &vsi->mac_list, next) {
+               if (is_same_ether_addr(macaddr, &f->mac_info.mac_addr))
+                       return f;
+       }
+
+       return NULL;
+}
+
+static int
+ice_add_mac_filter(struct ice_vsi *vsi, struct rte_ether_addr *mac_addr)
+{
+       struct ice_fltr_list_entry *m_list_itr = NULL;
+       struct ice_mac_filter *f;
+       struct LIST_HEAD_TYPE list_head;
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       int ret = 0;
+
+       /* If it's added and configured, return */
+       f = ice_find_mac_filter(vsi, mac_addr);
+       if (f) {
+               PMD_DRV_LOG(INFO, "This MAC filter already exists.");
+               return 0;
+       }
+
+       INIT_LIST_HEAD(&list_head);
+
+       m_list_itr = (struct ice_fltr_list_entry *)
+               ice_malloc(hw, sizeof(*m_list_itr));
+       if (!m_list_itr) {
+               ret = -ENOMEM;
+               goto DONE;
+       }
+       ice_memcpy(m_list_itr->fltr_info.l_data.mac.mac_addr,
+                  mac_addr, ETH_ALEN, ICE_NONDMA_TO_NONDMA);
+       m_list_itr->fltr_info.src_id = ICE_SRC_ID_VSI;
+       m_list_itr->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+       m_list_itr->fltr_info.lkup_type = ICE_SW_LKUP_MAC;
+       m_list_itr->fltr_info.flag = ICE_FLTR_TX;
+       m_list_itr->fltr_info.vsi_handle = vsi->idx;
+
+       LIST_ADD(&m_list_itr->list_entry, &list_head);
+
+       /* Add the mac */
+       ret = ice_add_mac(hw, &list_head);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to add MAC filter");
+               ret = -EINVAL;
+               goto DONE;
+       }
+       /* Add the mac addr into mac list */
+       f = rte_zmalloc(NULL, sizeof(*f), 0);
+       if (!f) {
+               PMD_DRV_LOG(ERR, "failed to allocate memory");
+               ret = -ENOMEM;
+               goto DONE;
+       }
+       rte_memcpy(&f->mac_info.mac_addr, mac_addr, ETH_ADDR_LEN);
+       TAILQ_INSERT_TAIL(&vsi->mac_list, f, next);
+       vsi->mac_num++;
+
+       ret = 0;
+
+DONE:
+       rte_free(m_list_itr);
+       return ret;
+}
+
+static int
+ice_remove_mac_filter(struct ice_vsi *vsi, struct rte_ether_addr *mac_addr)
+{
+       struct ice_fltr_list_entry *m_list_itr = NULL;
+       struct ice_mac_filter *f;
+       struct LIST_HEAD_TYPE list_head;
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       int ret = 0;
+
+       /* Can't find it, return an error */
+       f = ice_find_mac_filter(vsi, mac_addr);
+       if (!f)
+               return -EINVAL;
+
+       INIT_LIST_HEAD(&list_head);
+
+       m_list_itr = (struct ice_fltr_list_entry *)
+               ice_malloc(hw, sizeof(*m_list_itr));
+       if (!m_list_itr) {
+               ret = -ENOMEM;
+               goto DONE;
+       }
+       ice_memcpy(m_list_itr->fltr_info.l_data.mac.mac_addr,
+                  mac_addr, ETH_ALEN, ICE_NONDMA_TO_NONDMA);
+       m_list_itr->fltr_info.src_id = ICE_SRC_ID_VSI;
+       m_list_itr->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+       m_list_itr->fltr_info.lkup_type = ICE_SW_LKUP_MAC;
+       m_list_itr->fltr_info.flag = ICE_FLTR_TX;
+       m_list_itr->fltr_info.vsi_handle = vsi->idx;
+
+       LIST_ADD(&m_list_itr->list_entry, &list_head);
+
+       /* remove the mac filter */
+       ret = ice_remove_mac(hw, &list_head);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to remove MAC filter");
+               ret = -EINVAL;
+               goto DONE;
+       }
+
+       /* Remove the mac addr from mac list */
+       TAILQ_REMOVE(&vsi->mac_list, f, next);
+       rte_free(f);
+       vsi->mac_num--;
+
+       ret = 0;
+DONE:
+       rte_free(m_list_itr);
+       return ret;
+}
+
+/* Find out specific VLAN filter */
+static struct ice_vlan_filter *
+ice_find_vlan_filter(struct ice_vsi *vsi, uint16_t vlan_id)
+{
+       struct ice_vlan_filter *f;
+
+       TAILQ_FOREACH(f, &vsi->vlan_list, next) {
+               if (vlan_id == f->vlan_info.vlan_id)
+                       return f;
+       }
+
+       return NULL;
+}
+
+static int
+ice_add_vlan_filter(struct ice_vsi *vsi, uint16_t vlan_id)
+{
+       struct ice_fltr_list_entry *v_list_itr = NULL;
+       struct ice_vlan_filter *f;
+       struct LIST_HEAD_TYPE list_head;
+       struct ice_hw *hw;
+       int ret = 0;
+
+       if (!vsi || vlan_id > ETHER_MAX_VLAN_ID)
+               return -EINVAL;
+
+       hw = ICE_VSI_TO_HW(vsi);
+
+       /* If it's added and configured, return. */
+       f = ice_find_vlan_filter(vsi, vlan_id);
+       if (f) {
+               PMD_DRV_LOG(INFO, "This VLAN filter already exists.");
+               return 0;
+       }
+
+       if (!vsi->vlan_anti_spoof_on && !vsi->vlan_filter_on)
+               return 0;
+
+       INIT_LIST_HEAD(&list_head);
+
+       v_list_itr = (struct ice_fltr_list_entry *)
+                     ice_malloc(hw, sizeof(*v_list_itr));
+       if (!v_list_itr) {
+               ret = -ENOMEM;
+               goto DONE;
+       }
+       v_list_itr->fltr_info.l_data.vlan.vlan_id = vlan_id;
+       v_list_itr->fltr_info.src_id = ICE_SRC_ID_VSI;
+       v_list_itr->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+       v_list_itr->fltr_info.lkup_type = ICE_SW_LKUP_VLAN;
+       v_list_itr->fltr_info.flag = ICE_FLTR_TX;
+       v_list_itr->fltr_info.vsi_handle = vsi->idx;
+
+       LIST_ADD(&v_list_itr->list_entry, &list_head);
+
+       /* Add the vlan */
+       ret = ice_add_vlan(hw, &list_head);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to add VLAN filter");
+               ret = -EINVAL;
+               goto DONE;
+       }
+
+       /* Add vlan into vlan list */
+       f = rte_zmalloc(NULL, sizeof(*f), 0);
+       if (!f) {
+               PMD_DRV_LOG(ERR, "failed to allocate memory");
+               ret = -ENOMEM;
+               goto DONE;
+       }
+       f->vlan_info.vlan_id = vlan_id;
+       TAILQ_INSERT_TAIL(&vsi->vlan_list, f, next);
+       vsi->vlan_num++;
+
+       ret = 0;
+
+DONE:
+       rte_free(v_list_itr);
+       return ret;
+}
+
+static int
+ice_remove_vlan_filter(struct ice_vsi *vsi, uint16_t vlan_id)
+{
+       struct ice_fltr_list_entry *v_list_itr = NULL;
+       struct ice_vlan_filter *f;
+       struct LIST_HEAD_TYPE list_head;
+       struct ice_hw *hw;
+       int ret = 0;
+
+       /**
+        * Vlan 0 is the generic filter for untagged packets
+        * and can't be removed.
+        */
+       if (!vsi || vlan_id == 0 || vlan_id > ETHER_MAX_VLAN_ID)
+               return -EINVAL;
+
+       hw = ICE_VSI_TO_HW(vsi);
+
+       /* Can't find it, return an error */
+       f = ice_find_vlan_filter(vsi, vlan_id);
+       if (!f)
+               return -EINVAL;
+
+       INIT_LIST_HEAD(&list_head);
+
+       v_list_itr = (struct ice_fltr_list_entry *)
+                     ice_malloc(hw, sizeof(*v_list_itr));
+       if (!v_list_itr) {
+               ret = -ENOMEM;
+               goto DONE;
+       }
+
+       v_list_itr->fltr_info.l_data.vlan.vlan_id = vlan_id;
+       v_list_itr->fltr_info.src_id = ICE_SRC_ID_VSI;
+       v_list_itr->fltr_info.fltr_act = ICE_FWD_TO_VSI;
+       v_list_itr->fltr_info.lkup_type = ICE_SW_LKUP_VLAN;
+       v_list_itr->fltr_info.flag = ICE_FLTR_TX;
+       v_list_itr->fltr_info.vsi_handle = vsi->idx;
+
+       LIST_ADD(&v_list_itr->list_entry, &list_head);
+
+       /* remove the vlan filter */
+       ret = ice_remove_vlan(hw, &list_head);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to remove VLAN filter");
+               ret = -EINVAL;
+               goto DONE;
+       }
+
+       /* Remove the vlan id from vlan list */
+       TAILQ_REMOVE(&vsi->vlan_list, f, next);
+       rte_free(f);
+       vsi->vlan_num--;
+
+       ret = 0;
+DONE:
+       rte_free(v_list_itr);
+       return ret;
+}
+
+static int
+ice_remove_all_mac_vlan_filters(struct ice_vsi *vsi)
+{
+       struct ice_mac_filter *m_f;
+       struct ice_vlan_filter *v_f;
+       int ret = 0;
+
+       if (!vsi || !vsi->mac_num)
+               return -EINVAL;
+
+       TAILQ_FOREACH(m_f, &vsi->mac_list, next) {
+               ret = ice_remove_mac_filter(vsi, &m_f->mac_info.mac_addr);
+               if (ret != ICE_SUCCESS) {
+                       ret = -EINVAL;
+                       goto DONE;
+               }
+       }
+
+       if (vsi->vlan_num == 0)
+               return 0;
+
+       TAILQ_FOREACH(v_f, &vsi->vlan_list, next) {
+               ret = ice_remove_vlan_filter(vsi, v_f->vlan_info.vlan_id);
+               if (ret != ICE_SUCCESS) {
+                       ret = -EINVAL;
+                       goto DONE;
+               }
+       }
+
+DONE:
+       return ret;
+}
+
+static int
+ice_vsi_config_qinq_insertion(struct ice_vsi *vsi, bool on)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       struct ice_vsi_ctx ctxt;
+       uint8_t qinq_flags;
+       int ret = 0;
+
+       /* Check if it has been already on or off */
+       if (vsi->info.valid_sections &
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_OUTER_TAG_VALID)) {
+               if (on) {
+                       if ((vsi->info.outer_tag_flags &
+                            ICE_AQ_VSI_OUTER_TAG_ACCEPT_HOST) ==
+                           ICE_AQ_VSI_OUTER_TAG_ACCEPT_HOST)
+                               return 0; /* already on */
+               } else {
+                       if (!(vsi->info.outer_tag_flags &
+                             ICE_AQ_VSI_OUTER_TAG_ACCEPT_HOST))
+                               return 0; /* already off */
+               }
+       }
+
+       if (on)
+               qinq_flags = ICE_AQ_VSI_OUTER_TAG_ACCEPT_HOST;
+       else
+               qinq_flags = 0;
+       /* clear global insertion and use per packet insertion */
+       vsi->info.outer_tag_flags &= ~(ICE_AQ_VSI_OUTER_TAG_INSERT);
+       vsi->info.outer_tag_flags &= ~(ICE_AQ_VSI_OUTER_TAG_ACCEPT_HOST);
+       vsi->info.outer_tag_flags |= qinq_flags;
+       /* use default vlan type 0x8100 */
+       vsi->info.outer_tag_flags &= ~(ICE_AQ_VSI_OUTER_TAG_TYPE_M);
+       vsi->info.outer_tag_flags |= ICE_DFLT_OUTER_TAG_TYPE <<
+                                    ICE_AQ_VSI_OUTER_TAG_TYPE_S;
+       (void)rte_memcpy(&ctxt.info, &vsi->info, sizeof(vsi->info));
+       ctxt.info.valid_sections =
+                       rte_cpu_to_le_16(ICE_AQ_VSI_PROP_OUTER_TAG_VALID);
+       ctxt.vsi_num = vsi->vsi_id;
+       ret = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+       if (ret) {
+               PMD_DRV_LOG(INFO,
+                           "Update VSI failed to %s qinq stripping",
+                           on ? "enable" : "disable");
+               return -EINVAL;
+       }
+
+       vsi->info.valid_sections |=
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_OUTER_TAG_VALID);
+
+       return ret;
+}
+
+static int
+ice_vsi_config_qinq_stripping(struct ice_vsi *vsi, bool on)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       struct ice_vsi_ctx ctxt;
+       uint8_t qinq_flags;
+       int ret = 0;
+
+       /* Check if it has been already on or off */
+       if (vsi->info.valid_sections &
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_OUTER_TAG_VALID)) {
+               if (on) {
+                       if ((vsi->info.outer_tag_flags &
+                            ICE_AQ_VSI_OUTER_TAG_MODE_M) ==
+                           ICE_AQ_VSI_OUTER_TAG_COPY)
+                               return 0; /* already on */
+               } else {
+                       if ((vsi->info.outer_tag_flags &
+                            ICE_AQ_VSI_OUTER_TAG_MODE_M) ==
+                           ICE_AQ_VSI_OUTER_TAG_NOTHING)
+                               return 0; /* already off */
+               }
+       }
+
+       if (on)
+               qinq_flags = ICE_AQ_VSI_OUTER_TAG_COPY;
+       else
+               qinq_flags = ICE_AQ_VSI_OUTER_TAG_NOTHING;
+       vsi->info.outer_tag_flags &= ~(ICE_AQ_VSI_OUTER_TAG_MODE_M);
+       vsi->info.outer_tag_flags |= qinq_flags;
+       /* use default vlan type 0x8100 */
+       vsi->info.outer_tag_flags &= ~(ICE_AQ_VSI_OUTER_TAG_TYPE_M);
+       vsi->info.outer_tag_flags |= ICE_DFLT_OUTER_TAG_TYPE <<
+                                    ICE_AQ_VSI_OUTER_TAG_TYPE_S;
+       (void)rte_memcpy(&ctxt.info, &vsi->info, sizeof(vsi->info));
+       ctxt.info.valid_sections =
+                       rte_cpu_to_le_16(ICE_AQ_VSI_PROP_OUTER_TAG_VALID);
+       ctxt.vsi_num = vsi->vsi_id;
+       ret = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+       if (ret) {
+               PMD_DRV_LOG(INFO,
+                           "Update VSI failed to %s qinq stripping",
+                           on ? "enable" : "disable");
+               return -EINVAL;
+       }
+
+       vsi->info.valid_sections |=
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_OUTER_TAG_VALID);
+
+       return ret;
+}
+
+static int
+ice_vsi_config_double_vlan(struct ice_vsi *vsi, int on)
+{
+       int ret;
+
+       ret = ice_vsi_config_qinq_stripping(vsi, on);
+       if (ret)
+               PMD_DRV_LOG(ERR, "Fail to set qinq stripping - %d", ret);
+
+       ret = ice_vsi_config_qinq_insertion(vsi, on);
+       if (ret)
+               PMD_DRV_LOG(ERR, "Fail to set qinq insertion - %d", ret);
+
+       return ret;
+}
+
 /* Enable IRQ0 */
 static void
 ice_pf_enable_irq0(struct ice_hw *hw)
@@ -547,6 +1134,9 @@ ice_setup_vsi(struct ice_pf *pf, enum ice_vsi_type type)
        struct ice_vsi *vsi = NULL;
        struct ice_vsi_ctx vsi_ctx;
        int ret;
+       struct rte_ether_addr broadcast = {
+               .addr_bytes = {0xff, 0xff, 0xff, 0xff, 0xff, 0xff} };
+       struct rte_ether_addr mac_addr;
        uint16_t max_txqs[ICE_MAX_TRAFFIC_CLASS] = { 0 };
        uint8_t tc_bitmap = 0x1;
 
@@ -565,6 +1155,12 @@ ice_setup_vsi(struct ice_pf *pf, enum ice_vsi_type type)
        TAILQ_INIT(&vsi->mac_list);
        TAILQ_INIT(&vsi->vlan_list);
 
+       /* Be sync with ETH_RSS_RETA_SIZE_x maximum value definition */
+       pf->hash_lut_size = hw->func_caps.common_cap.rss_table_size >
+                       ETH_RSS_RETA_SIZE_512 ? ETH_RSS_RETA_SIZE_512 :
+                       hw->func_caps.common_cap.rss_table_size;
+       pf->flags |= ICE_FLAG_RSS_AQ_CAPABLE;
+
        memset(&vsi_ctx, 0, sizeof(vsi_ctx));
        /* base_queue in used in queue mapping of VSI add/update command.
         * Suppose vsi->base_queue is 0 now, don't consider SRIOV, VMDQ
@@ -632,6 +1228,21 @@ ice_setup_vsi(struct ice_pf *pf, enum ice_vsi_type type)
        pf->vsis_allocated = vsi_ctx.vsis_allocd;
        pf->vsis_unallocated = vsi_ctx.vsis_unallocated;
 
+       /* MAC configuration */
+       rte_memcpy(pf->dev_addr.addr_bytes,
+                  hw->port_info->mac.perm_addr,
+                  ETH_ADDR_LEN);
+
+       rte_memcpy(&mac_addr, &pf->dev_addr, ETHER_ADDR_LEN);
+       ret = ice_add_mac_filter(vsi, &mac_addr);
+       if (ret != ICE_SUCCESS)
+               PMD_INIT_LOG(ERR, "Failed to add dflt MAC filter");
+
+       rte_memcpy(&mac_addr, &broadcast, ETHER_ADDR_LEN);
+       ret = ice_add_mac_filter(vsi, &mac_addr);
+       if (ret != ICE_SUCCESS)
+               PMD_INIT_LOG(ERR, "Failed to add MAC filter");
+
        /* At the beginning, only TC0. */
        /* What we need here is the maximam number of the TX queues.
         * Currently vsi->nb_qps means it.
@@ -650,6 +1261,21 @@ fail_mem:
        return NULL;
 }
 
+static int
+ice_send_driver_ver(struct ice_hw *hw)
+{
+       struct ice_driver_ver dv;
+
+       /* we don't have driver version use 0 for dummy */
+       dv.major_ver = 0;
+       dv.minor_ver = 0;
+       dv.build_ver = 0;
+       dv.subbuild_ver = 0;
+       strncpy((char *)dv.driver_string, "dpdk", sizeof(dv.driver_string));
+
+       return ice_aq_send_driver_ver(hw, &dv, NULL);
+}
+
 static int
 ice_pf_setup(struct ice_pf *pf)
 {
@@ -673,6 +1299,69 @@ ice_pf_setup(struct ice_pf *pf)
        return 0;
 }
 
+static int ice_load_pkg(struct rte_eth_dev *dev)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       const char *pkg_file = ICE_DFLT_PKG_FILE;
+       int err;
+       uint8_t *buf;
+       int buf_len;
+       FILE *file;
+       struct stat fstat;
+
+       file = fopen(pkg_file, "rb");
+       if (!file)  {
+               PMD_INIT_LOG(ERR, "failed to open file: %s\n", pkg_file);
+               return -1;
+       }
+
+       err = stat(pkg_file, &fstat);
+       if (err) {
+               PMD_INIT_LOG(ERR, "failed to get file stats\n");
+               fclose(file);
+               return err;
+       }
+
+       buf_len = fstat.st_size;
+       buf = rte_malloc(NULL, buf_len, 0);
+
+       if (!buf) {
+               PMD_INIT_LOG(ERR, "failed to allocate buf of size %d for package\n",
+                               buf_len);
+               fclose(file);
+               return -1;
+       }
+
+       err = fread(buf, buf_len, 1, file);
+       if (err != 1) {
+               PMD_INIT_LOG(ERR, "failed to read package data\n");
+               fclose(file);
+               err = -1;
+               goto fail_exit;
+       }
+
+       fclose(file);
+
+       err = ice_copy_and_init_pkg(hw, buf, buf_len);
+       if (err) {
+               PMD_INIT_LOG(ERR, "ice_copy_and_init_hw failed: %d\n", err);
+               goto fail_exit;
+       }
+       err = ice_init_hw_tbls(hw);
+       if (err) {
+               PMD_INIT_LOG(ERR, "ice_init_hw_tbls failed: %d\n", err);
+               goto fail_init_tbls;
+       }
+
+       return 0;
+
+fail_init_tbls:
+       rte_free(hw->pkg_copy);
+fail_exit:
+       rte_free(buf);
+       return err;
+}
+
 static int
 ice_dev_init(struct rte_eth_dev *dev)
 {
@@ -680,6 +1369,9 @@ ice_dev_init(struct rte_eth_dev *dev)
        struct rte_intr_handle *intr_handle;
        struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
        struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_adapter *ad =
+               ICE_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
+       struct ice_vsi *vsi;
        int ret;
 
        dev->dev_ops = &ice_eth_dev_ops;
@@ -711,6 +1403,13 @@ ice_dev_init(struct rte_eth_dev *dev)
                return -EINVAL;
        }
 
+       ret = ice_load_pkg(dev);
+       if (ret) {
+               PMD_INIT_LOG(WARNING, "Failed to load the DDP package,"
+                               "Entering Safe Mode");
+               ad->is_safe_mode = 1;
+       }
+
        PMD_INIT_LOG(INFO, "FW %d.%d.%05d API %d.%d",
                     hw->fw_maj_ver, hw->fw_min_ver, hw->fw_build,
                     hw->api_maj_ver, hw->api_min_ver);
@@ -735,6 +1434,21 @@ ice_dev_init(struct rte_eth_dev *dev)
                goto err_pf_setup;
        }
 
+       ret = ice_send_driver_ver(hw);
+       if (ret) {
+               PMD_INIT_LOG(ERR, "Failed to send driver version");
+               goto err_pf_setup;
+       }
+
+       vsi = pf->main_vsi;
+
+       /* Disable double vlan by default */
+       ice_vsi_config_double_vlan(vsi, FALSE);
+
+       ret = ice_aq_stop_lldp(hw, TRUE, NULL);
+       if (ret != ICE_SUCCESS)
+               PMD_INIT_LOG(DEBUG, "lldp has already stopped\n");
+
        /* register callback func to eal lib */
        rte_intr_callback_register(intr_handle,
                                   ice_interrupt_handler, dev);
@@ -770,6 +1484,8 @@ ice_release_vsi(struct ice_vsi *vsi)
 
        hw = ICE_VSI_TO_HW(vsi);
 
+       ice_remove_all_mac_vlan_filters(vsi);
+
        memset(&vsi_ctx, 0, sizeof(vsi_ctx));
 
        vsi_ctx.vsi_num = vsi->vsi_id;
@@ -785,11 +1501,40 @@ ice_release_vsi(struct ice_vsi *vsi)
        return 0;
 }
 
+static void
+ice_vsi_disable_queues_intr(struct ice_vsi *vsi)
+{
+       struct rte_eth_dev *dev = vsi->adapter->eth_dev;
+       struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
+       struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       uint16_t msix_intr, i;
+
+       /* disable interrupt and also clear all the exist config */
+       for (i = 0; i < vsi->nb_qps; i++) {
+               ICE_WRITE_REG(hw, QINT_TQCTL(vsi->base_queue + i), 0);
+               ICE_WRITE_REG(hw, QINT_RQCTL(vsi->base_queue + i), 0);
+               rte_wmb();
+       }
+
+       if (rte_intr_allow_others(intr_handle))
+               /* vfio-pci */
+               for (i = 0; i < vsi->nb_msix; i++) {
+                       msix_intr = vsi->msix_intr + i;
+                       ICE_WRITE_REG(hw, GLINT_DYN_CTL(msix_intr),
+                                     GLINT_DYN_CTL_WB_ON_ITR_M);
+               }
+       else
+               /* igb_uio */
+               ICE_WRITE_REG(hw, GLINT_DYN_CTL(0), GLINT_DYN_CTL_WB_ON_ITR_M);
+}
+
 static void
 ice_dev_stop(struct rte_eth_dev *dev)
 {
        struct rte_eth_dev_data *data = dev->data;
        struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *main_vsi = pf->main_vsi;
        struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
        struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
        uint16_t i;
@@ -806,9 +1551,14 @@ ice_dev_stop(struct rte_eth_dev *dev)
        for (i = 0; i < data->nb_tx_queues; i++)
                ice_tx_queue_stop(dev, i);
 
+       /* disable all queue interrupts */
+       ice_vsi_disable_queues_intr(main_vsi);
+
        /* Clear all queues and release mbufs */
        ice_clear_queues(dev);
 
+       ice_dev_set_link_down(dev);
+
        /* Clean datapath event and queue/vec mapping */
        rte_intr_efd_disable(intr_handle);
        if (intr_handle->intr_vec) {
@@ -825,6 +1575,13 @@ ice_dev_close(struct rte_eth_dev *dev)
        struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
        struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
 
+       /* Since stop will make link down, then the link event will be
+        * triggered, disable the irq firstly to avoid the port_infoe etc
+        * resources deallocation causing the interrupt service thread
+        * crash.
+        */
+       ice_pf_disable_irq0(hw);
+
        ice_dev_stop(dev);
 
        /* release all queue resource */
@@ -832,15 +1589,15 @@ ice_dev_close(struct rte_eth_dev *dev)
 
        ice_res_pool_destroy(&pf->msix_pool);
        ice_release_vsi(pf->main_vsi);
-
+       ice_sched_cleanup_all(hw);
+       rte_free(hw->port_info);
+       hw->port_info = NULL;
        ice_shutdown_all_ctrlq(hw);
 }
 
 static int
 ice_dev_uninit(struct rte_eth_dev *dev)
 {
-       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
-       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
        struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
        struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
 
@@ -856,15 +1613,10 @@ ice_dev_uninit(struct rte_eth_dev *dev)
        /* disable uio intr before callback unregister */
        rte_intr_disable(intr_handle);
 
-       /* register callback func to eal lib */
+       /* unregister callback func from eal lib */
        rte_intr_callback_unregister(intr_handle,
                                     ice_interrupt_handler, dev);
 
-       ice_release_vsi(pf->main_vsi);
-       ice_sched_cleanup_all(hw);
-       rte_free(hw->port_info);
-       ice_shutdown_all_ctrlq(hw);
-
        return 0;
 }
 
@@ -892,11 +1644,17 @@ static int ice_init_rss(struct ice_pf *pf)
        struct ice_aqc_get_set_rss_keys key;
        uint16_t i, nb_q;
        int ret = 0;
+       bool is_safe_mode = pf->adapter->is_safe_mode;
 
        rss_conf = &dev->data->dev_conf.rx_adv_conf.rss_conf;
        nb_q = dev->data->nb_rx_queues;
        vsi->rss_key_size = ICE_AQC_GET_SET_RSS_KEY_DATA_RSS_KEY_SIZE;
-       vsi->rss_lut_size = hw->func_caps.common_cap.rss_table_size;
+       vsi->rss_lut_size = pf->hash_lut_size;
+
+       if (is_safe_mode) {
+               PMD_DRV_LOG(WARNING, "RSS is not supported in safe mode\n");
+               return 0;
+       }
 
        if (!vsi->rss_key)
                vsi->rss_key = rte_zmalloc(NULL,
@@ -930,18 +1688,221 @@ static int ice_init_rss(struct ice_pf *pf)
        if (ret)
                return -EINVAL;
 
-       return 0;
-}
+       /* configure RSS for IPv4 with input set IPv4 src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV4,
+                             ICE_FLOW_SEG_HDR_IPV4);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s IPV4 rss flow fail %d", __func__, ret);
 
-static int
-ice_dev_start(struct rte_eth_dev *dev)
-{
-       struct rte_eth_dev_data *data = dev->data;
-       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
-       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
-       uint16_t nb_rxq = 0;
+       /* configure RSS for IPv6 with input set IPv6 src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV6,
+                             ICE_FLOW_SEG_HDR_IPV6);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s IPV6 rss flow fail %d", __func__, ret);
+
+       /* configure RSS for tcp6 with input set IPv6 src/dst, TCP src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_TCP_IPV6,
+                             ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_IPV6);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s TCP_IPV6 rss flow fail %d", __func__, ret);
+
+       /* configure RSS for udp6 with input set IPv6 src/dst, UDP src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_UDP_IPV6,
+                             ICE_FLOW_SEG_HDR_UDP | ICE_FLOW_SEG_HDR_IPV6);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s UDP_IPV6 rss flow fail %d", __func__, ret);
+
+       /* configure RSS for sctp6 with input set IPv6 src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV6,
+                             ICE_FLOW_SEG_HDR_SCTP | ICE_FLOW_SEG_HDR_IPV6);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s SCTP_IPV6 rss flow fail %d",
+                               __func__, ret);
+
+       /* configure RSS for tcp4 with input set IP src/dst, TCP src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_TCP_IPV4,
+                             ICE_FLOW_SEG_HDR_TCP | ICE_FLOW_SEG_HDR_IPV4);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s TCP_IPV4 rss flow fail %d", __func__, ret);
+
+       /* configure RSS for udp4 with input set IP src/dst, UDP src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_HASH_UDP_IPV4,
+                             ICE_FLOW_SEG_HDR_UDP | ICE_FLOW_SEG_HDR_IPV4);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s UDP_IPV4 rss flow fail %d", __func__, ret);
+
+       /* configure RSS for sctp4 with input set IP src/dst */
+       ret = ice_add_rss_cfg(hw, vsi->idx, ICE_FLOW_HASH_IPV4,
+                             ICE_FLOW_SEG_HDR_SCTP | ICE_FLOW_SEG_HDR_IPV4);
+       if (ret)
+               PMD_DRV_LOG(ERR, "%s SCTP_IPV4 rss flow fail %d",
+                               __func__, ret);
+
+       return 0;
+}
+
+static void
+__vsi_queues_bind_intr(struct ice_vsi *vsi, uint16_t msix_vect,
+                      int base_queue, int nb_queue)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       uint32_t val, val_tx;
+       int i;
+
+       for (i = 0; i < nb_queue; i++) {
+               /*do actual bind*/
+               val = (msix_vect & QINT_RQCTL_MSIX_INDX_M) |
+                     (0 < QINT_RQCTL_ITR_INDX_S) | QINT_RQCTL_CAUSE_ENA_M;
+               val_tx = (msix_vect & QINT_TQCTL_MSIX_INDX_M) |
+                        (0 < QINT_TQCTL_ITR_INDX_S) | QINT_TQCTL_CAUSE_ENA_M;
+
+               PMD_DRV_LOG(INFO, "queue %d is binding to vect %d",
+                           base_queue + i, msix_vect);
+               /* set ITR0 value */
+               ICE_WRITE_REG(hw, GLINT_ITR(0, msix_vect), 0x10);
+               ICE_WRITE_REG(hw, QINT_RQCTL(base_queue + i), val);
+               ICE_WRITE_REG(hw, QINT_TQCTL(base_queue + i), val_tx);
+       }
+}
+
+static void
+ice_vsi_queues_bind_intr(struct ice_vsi *vsi)
+{
+       struct rte_eth_dev *dev = vsi->adapter->eth_dev;
+       struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
+       struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       uint16_t msix_vect = vsi->msix_intr;
+       uint16_t nb_msix = RTE_MIN(vsi->nb_msix, intr_handle->nb_efd);
+       uint16_t queue_idx = 0;
+       int record = 0;
+       int i;
+
+       /* clear Rx/Tx queue interrupt */
+       for (i = 0; i < vsi->nb_used_qps; i++) {
+               ICE_WRITE_REG(hw, QINT_TQCTL(vsi->base_queue + i), 0);
+               ICE_WRITE_REG(hw, QINT_RQCTL(vsi->base_queue + i), 0);
+       }
+
+       /* PF bind interrupt */
+       if (rte_intr_dp_is_en(intr_handle)) {
+               queue_idx = 0;
+               record = 1;
+       }
+
+       for (i = 0; i < vsi->nb_used_qps; i++) {
+               if (nb_msix <= 1) {
+                       if (!rte_intr_allow_others(intr_handle))
+                               msix_vect = ICE_MISC_VEC_ID;
+
+                       /* uio mapping all queue to one msix_vect */
+                       __vsi_queues_bind_intr(vsi, msix_vect,
+                                              vsi->base_queue + i,
+                                              vsi->nb_used_qps - i);
+
+                       for (; !!record && i < vsi->nb_used_qps; i++)
+                               intr_handle->intr_vec[queue_idx + i] =
+                                       msix_vect;
+                       break;
+               }
+
+               /* vfio 1:1 queue/msix_vect mapping */
+               __vsi_queues_bind_intr(vsi, msix_vect,
+                                      vsi->base_queue + i, 1);
+
+               if (!!record)
+                       intr_handle->intr_vec[queue_idx + i] = msix_vect;
+
+               msix_vect++;
+               nb_msix--;
+       }
+}
+
+static void
+ice_vsi_enable_queues_intr(struct ice_vsi *vsi)
+{
+       struct rte_eth_dev *dev = vsi->adapter->eth_dev;
+       struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
+       struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       uint16_t msix_intr, i;
+
+       if (rte_intr_allow_others(intr_handle))
+               for (i = 0; i < vsi->nb_used_qps; i++) {
+                       msix_intr = vsi->msix_intr + i;
+                       ICE_WRITE_REG(hw, GLINT_DYN_CTL(msix_intr),
+                                     GLINT_DYN_CTL_INTENA_M |
+                                     GLINT_DYN_CTL_CLEARPBA_M |
+                                     GLINT_DYN_CTL_ITR_INDX_M |
+                                     GLINT_DYN_CTL_WB_ON_ITR_M);
+               }
+       else
+               ICE_WRITE_REG(hw, GLINT_DYN_CTL(0),
+                             GLINT_DYN_CTL_INTENA_M |
+                             GLINT_DYN_CTL_CLEARPBA_M |
+                             GLINT_DYN_CTL_ITR_INDX_M |
+                             GLINT_DYN_CTL_WB_ON_ITR_M);
+}
+
+static int
+ice_rxq_intr_setup(struct rte_eth_dev *dev)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
+       struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
+       struct ice_vsi *vsi = pf->main_vsi;
+       uint32_t intr_vector = 0;
+
+       rte_intr_disable(intr_handle);
+
+       /* check and configure queue intr-vector mapping */
+       if ((rte_intr_cap_multiple(intr_handle) ||
+            !RTE_ETH_DEV_SRIOV(dev).active) &&
+           dev->data->dev_conf.intr_conf.rxq != 0) {
+               intr_vector = dev->data->nb_rx_queues;
+               if (intr_vector > ICE_MAX_INTR_QUEUE_NUM) {
+                       PMD_DRV_LOG(ERR, "At most %d intr queues supported",
+                                   ICE_MAX_INTR_QUEUE_NUM);
+                       return -ENOTSUP;
+               }
+               if (rte_intr_efd_enable(intr_handle, intr_vector))
+                       return -1;
+       }
+
+       if (rte_intr_dp_is_en(intr_handle) && !intr_handle->intr_vec) {
+               intr_handle->intr_vec =
+               rte_zmalloc(NULL, dev->data->nb_rx_queues * sizeof(int),
+                           0);
+               if (!intr_handle->intr_vec) {
+                       PMD_DRV_LOG(ERR,
+                                   "Failed to allocate %d rx_queues intr_vec",
+                                   dev->data->nb_rx_queues);
+                       return -ENOMEM;
+               }
+       }
+
+       /* Map queues with MSIX interrupt */
+       vsi->nb_used_qps = dev->data->nb_rx_queues;
+       ice_vsi_queues_bind_intr(vsi);
+
+       /* Enable interrupts for all the queues */
+       ice_vsi_enable_queues_intr(vsi);
+
+       rte_intr_enable(intr_handle);
+
+       return 0;
+}
+
+static int
+ice_dev_start(struct rte_eth_dev *dev)
+{
+       struct rte_eth_dev_data *data = dev->data;
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       uint16_t nb_rxq = 0;
        uint16_t nb_txq, i;
-       int ret;
+       int mask, ret;
 
        /* program Tx queues' context in hardware */
        for (nb_txq = 0; nb_txq < data->nb_tx_queues; nb_txq++) {
@@ -968,6 +1929,27 @@ ice_dev_start(struct rte_eth_dev *dev)
        }
 
        ice_set_rx_function(dev);
+       ice_set_tx_function(dev);
+
+       mask = ETH_VLAN_STRIP_MASK | ETH_VLAN_FILTER_MASK |
+                       ETH_VLAN_EXTEND_MASK;
+       ret = ice_vlan_offload_set(dev, mask);
+       if (ret) {
+               PMD_INIT_LOG(ERR, "Unable to set VLAN offload");
+               goto rx_err;
+       }
+
+       /* enable Rx interrput and mapping Rx queue to interrupt vector */
+       if (ice_rxq_intr_setup(dev))
+               return -EIO;
+
+       /* Enable receiving broadcast packets and transmitting packets */
+       ret = ice_set_vsi_promisc(hw, vsi->idx,
+                                 ICE_PROMISC_BCAST_RX | ICE_PROMISC_BCAST_TX |
+                                 ICE_PROMISC_UCAST_TX | ICE_PROMISC_MCAST_TX,
+                                 0);
+       if (ret != ICE_SUCCESS)
+               PMD_DRV_LOG(INFO, "fail to set vsi broadcast");
 
        ret = ice_aq_set_event_mask(hw, hw->port_info->lport,
                                    ((u16)(ICE_AQ_LINK_EVENT_LINK_FAULT |
@@ -980,6 +1962,8 @@ ice_dev_start(struct rte_eth_dev *dev)
        if (ret != ICE_SUCCESS)
                PMD_DRV_LOG(WARNING, "Fail to set phy mask");
 
+       ice_dev_set_link_up(dev);
+
        /* Call get_link_info aq commond to enable/disable LSE */
        ice_link_update(dev, 0);
 
@@ -1028,6 +2012,9 @@ ice_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
        struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
        struct ice_vsi *vsi = pf->main_vsi;
        struct rte_pci_device *pci_dev = RTE_DEV_TO_PCI(dev->device);
+       bool is_safe_mode = pf->adapter->is_safe_mode;
+       u64 phy_type_low;
+       u64 phy_type_high;
 
        dev_info->min_rx_bufsize = ICE_BUF_SIZE_MIN;
        dev_info->max_rx_pktlen = ICE_FRAME_SIZE_MAX;
@@ -1035,26 +2022,44 @@ ice_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
        dev_info->max_tx_queues = vsi->nb_qps;
        dev_info->max_mac_addrs = vsi->max_macaddrs;
        dev_info->max_vfs = pci_dev->max_vfs;
+       dev_info->max_mtu = dev_info->max_rx_pktlen - ICE_ETH_OVERHEAD;
+       dev_info->min_mtu = ETHER_MIN_MTU;
 
        dev_info->rx_offload_capa =
-               DEV_RX_OFFLOAD_IPV4_CKSUM |
-               DEV_RX_OFFLOAD_UDP_CKSUM |
-               DEV_RX_OFFLOAD_TCP_CKSUM |
-               DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM |
+               DEV_RX_OFFLOAD_VLAN_STRIP |
                DEV_RX_OFFLOAD_JUMBO_FRAME |
-               DEV_RX_OFFLOAD_KEEP_CRC;
+               DEV_RX_OFFLOAD_KEEP_CRC |
+               DEV_RX_OFFLOAD_SCATTER |
+               DEV_RX_OFFLOAD_VLAN_FILTER;
        dev_info->tx_offload_capa =
-               DEV_TX_OFFLOAD_IPV4_CKSUM |
-               DEV_TX_OFFLOAD_UDP_CKSUM |
-               DEV_TX_OFFLOAD_TCP_CKSUM |
-               DEV_TX_OFFLOAD_SCTP_CKSUM |
-               DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM |
+               DEV_TX_OFFLOAD_VLAN_INSERT |
                DEV_TX_OFFLOAD_TCP_TSO |
-               DEV_TX_OFFLOAD_MULTI_SEGS;
+               DEV_TX_OFFLOAD_MULTI_SEGS |
+               DEV_TX_OFFLOAD_MBUF_FAST_FREE;
+       dev_info->flow_type_rss_offloads = 0;
+
+       if (!is_safe_mode) {
+               dev_info->rx_offload_capa |=
+                       DEV_RX_OFFLOAD_IPV4_CKSUM |
+                       DEV_RX_OFFLOAD_UDP_CKSUM |
+                       DEV_RX_OFFLOAD_TCP_CKSUM |
+                       DEV_RX_OFFLOAD_QINQ_STRIP |
+                       DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM |
+                       DEV_RX_OFFLOAD_VLAN_EXTEND;
+               dev_info->tx_offload_capa |=
+                       DEV_TX_OFFLOAD_QINQ_INSERT |
+                       DEV_TX_OFFLOAD_IPV4_CKSUM |
+                       DEV_TX_OFFLOAD_UDP_CKSUM |
+                       DEV_TX_OFFLOAD_TCP_CKSUM |
+                       DEV_TX_OFFLOAD_SCTP_CKSUM |
+                       DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM;
+               dev_info->flow_type_rss_offloads |= ICE_RSS_OFFLOAD_ALL;
+       }
+
        dev_info->rx_queue_offload_capa = 0;
        dev_info->tx_queue_offload_capa = 0;
 
-       dev_info->reta_size = hw->func_caps.common_cap.rss_table_size;
+       dev_info->reta_size = pf->hash_lut_size;
        dev_info->hash_key_size = (VSIQF_HKEY_MAX_INDEX + 1) * sizeof(uint32_t);
 
        dev_info->default_rxconf = (struct rte_eth_rxconf) {
@@ -1098,8 +2103,17 @@ ice_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
                               ETH_LINK_SPEED_5G |
                               ETH_LINK_SPEED_10G |
                               ETH_LINK_SPEED_20G |
-                              ETH_LINK_SPEED_25G |
-                              ETH_LINK_SPEED_40G;
+                              ETH_LINK_SPEED_25G;
+
+       phy_type_low = hw->port_info->phy.phy_type_low;
+       phy_type_high = hw->port_info->phy.phy_type_high;
+
+       if (ICE_PHY_TYPE_SUPPORT_50G(phy_type_low))
+               dev_info->speed_capa |= ETH_LINK_SPEED_50G;
+
+       if (ICE_PHY_TYPE_SUPPORT_100G_LOW(phy_type_low) ||
+                       ICE_PHY_TYPE_SUPPORT_100G_HIGH(phy_type_high))
+               dev_info->speed_capa |= ETH_LINK_SPEED_100G;
 
        dev_info->nb_rx_queues = dev->data->nb_rx_queues;
        dev_info->nb_tx_queues = dev->data->nb_tx_queues;
@@ -1141,7 +2155,7 @@ ice_atomic_write_link_status(struct rte_eth_dev *dev,
 }
 
 static int
-ice_link_update(struct rte_eth_dev *dev, __rte_unused int wait_to_complete)
+ice_link_update(struct rte_eth_dev *dev, int wait_to_complete)
 {
 #define CHECK_INTERVAL 100  /* 100ms */
 #define MAX_REPEAT_TIME 10  /* 1s (10 * 100ms) in total */
@@ -1210,6 +2224,12 @@ ice_link_update(struct rte_eth_dev *dev, __rte_unused int wait_to_complete)
        case ICE_AQ_LINK_SPEED_40GB:
                link.link_speed = ETH_SPEED_NUM_40G;
                break;
+       case ICE_AQ_LINK_SPEED_50GB:
+               link.link_speed = ETH_SPEED_NUM_50G;
+               break;
+       case ICE_AQ_LINK_SPEED_100GB:
+               link.link_speed = ETH_SPEED_NUM_100G;
+               break;
        case ICE_AQ_LINK_SPEED_UNKNOWN:
        default:
                PMD_DRV_LOG(ERR, "Unknown link speed");
@@ -1228,13 +2248,80 @@ out:
        return 0;
 }
 
+/* Force the physical link state by getting the current PHY capabilities from
+ * hardware and setting the PHY config based on the determined capabilities. If
+ * link changes, link event will be triggered because both the Enable Automatic
+ * Link Update and LESM Enable bits are set when setting the PHY capabilities.
+ */
+static enum ice_status
+ice_force_phys_link_state(struct ice_hw *hw, bool link_up)
+{
+       struct ice_aqc_set_phy_cfg_data cfg = { 0 };
+       struct ice_aqc_get_phy_caps_data *pcaps;
+       struct ice_port_info *pi;
+       enum ice_status status;
+
+       if (!hw || !hw->port_info)
+               return ICE_ERR_PARAM;
+
+       pi = hw->port_info;
+
+       pcaps = (struct ice_aqc_get_phy_caps_data *)
+               ice_malloc(hw, sizeof(*pcaps));
+       if (!pcaps)
+               return ICE_ERR_NO_MEMORY;
+
+       status = ice_aq_get_phy_caps(pi, false, ICE_AQC_REPORT_SW_CFG, pcaps,
+                                    NULL);
+       if (status)
+               goto out;
+
+       /* No change in link */
+       if (link_up == !!(pcaps->caps & ICE_AQC_PHY_EN_LINK) &&
+           link_up == !!(pi->phy.link_info.link_info & ICE_AQ_LINK_UP))
+               goto out;
+
+       cfg.phy_type_low = pcaps->phy_type_low;
+       cfg.phy_type_high = pcaps->phy_type_high;
+       cfg.caps = pcaps->caps | ICE_AQ_PHY_ENA_AUTO_LINK_UPDT;
+       cfg.low_power_ctrl = pcaps->low_power_ctrl;
+       cfg.eee_cap = pcaps->eee_cap;
+       cfg.eeer_value = pcaps->eeer_value;
+       cfg.link_fec_opt = pcaps->link_fec_options;
+       if (link_up)
+               cfg.caps |= ICE_AQ_PHY_ENA_LINK;
+       else
+               cfg.caps &= ~ICE_AQ_PHY_ENA_LINK;
+
+       status = ice_aq_set_phy_cfg(hw, pi->lport, &cfg, NULL);
+
+out:
+       ice_free(hw, pcaps);
+       return status;
+}
+
+static int
+ice_dev_set_link_up(struct rte_eth_dev *dev)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+       return ice_force_phys_link_state(hw, true);
+}
+
+static int
+ice_dev_set_link_down(struct rte_eth_dev *dev)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+       return ice_force_phys_link_state(hw, false);
+}
+
 static int
 ice_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
 {
        struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
        struct rte_eth_dev_data *dev_data = pf->dev_data;
-       uint32_t frame_size = mtu + ETHER_HDR_LEN
-                             + ETHER_CRC_LEN + ICE_VLAN_TAG_SIZE;
+       uint32_t frame_size = mtu + ICE_ETH_OVERHEAD;
 
        /* check if mtu is within the allowed range */
        if (mtu < ETHER_MIN_MTU || frame_size > ICE_FRAME_SIZE_MAX)
@@ -1260,6 +2347,1258 @@ ice_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
        return 0;
 }
 
+static int ice_macaddr_set(struct rte_eth_dev *dev,
+                          struct rte_ether_addr *mac_addr)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       struct ice_mac_filter *f;
+       uint8_t flags = 0;
+       int ret;
+
+       if (!is_valid_assigned_ether_addr(mac_addr)) {
+               PMD_DRV_LOG(ERR, "Tried to set invalid MAC address.");
+               return -EINVAL;
+       }
+
+       TAILQ_FOREACH(f, &vsi->mac_list, next) {
+               if (is_same_ether_addr(&pf->dev_addr, &f->mac_info.mac_addr))
+                       break;
+       }
+
+       if (!f) {
+               PMD_DRV_LOG(ERR, "Failed to find filter for default mac");
+               return -EIO;
+       }
+
+       ret = ice_remove_mac_filter(vsi, &f->mac_info.mac_addr);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to delete mac filter");
+               return -EIO;
+       }
+       ret = ice_add_mac_filter(vsi, mac_addr);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to add mac filter");
+               return -EIO;
+       }
+       memcpy(&pf->dev_addr, mac_addr, ETH_ADDR_LEN);
+
+       flags = ICE_AQC_MAN_MAC_UPDATE_LAA_WOL;
+       ret = ice_aq_manage_mac_write(hw, mac_addr->addr_bytes, flags, NULL);
+       if (ret != ICE_SUCCESS)
+               PMD_DRV_LOG(ERR, "Failed to set manage mac");
+
+       return 0;
+}
+
+/* Add a MAC address, and update filters */
+static int
+ice_macaddr_add(struct rte_eth_dev *dev,
+               struct rte_ether_addr *mac_addr,
+               __rte_unused uint32_t index,
+               __rte_unused uint32_t pool)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       int ret;
+
+       ret = ice_add_mac_filter(vsi, mac_addr);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR, "Failed to add MAC filter");
+               return -EINVAL;
+       }
+
+       return ICE_SUCCESS;
+}
+
+/* Remove a MAC address, and update filters */
+static void
+ice_macaddr_remove(struct rte_eth_dev *dev, uint32_t index)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       struct rte_eth_dev_data *data = dev->data;
+       struct rte_ether_addr *macaddr;
+       int ret;
+
+       macaddr = &data->mac_addrs[index];
+       ret = ice_remove_mac_filter(vsi, macaddr);
+       if (ret) {
+               PMD_DRV_LOG(ERR, "Failed to remove MAC filter");
+               return;
+       }
+}
+
+static int
+ice_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       int ret;
+
+       PMD_INIT_FUNC_TRACE();
+
+       if (on) {
+               ret = ice_add_vlan_filter(vsi, vlan_id);
+               if (ret < 0) {
+                       PMD_DRV_LOG(ERR, "Failed to add vlan filter");
+                       return -EINVAL;
+               }
+       } else {
+               ret = ice_remove_vlan_filter(vsi, vlan_id);
+               if (ret < 0) {
+                       PMD_DRV_LOG(ERR, "Failed to remove vlan filter");
+                       return -EINVAL;
+               }
+       }
+
+       return 0;
+}
+
+/* Configure vlan filter on or off */
+static int
+ice_vsi_config_vlan_filter(struct ice_vsi *vsi, bool on)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       struct ice_vsi_ctx ctxt;
+       uint8_t sec_flags, sw_flags2;
+       int ret = 0;
+
+       sec_flags = ICE_AQ_VSI_SEC_TX_VLAN_PRUNE_ENA <<
+                   ICE_AQ_VSI_SEC_TX_PRUNE_ENA_S;
+       sw_flags2 = ICE_AQ_VSI_SW_FLAG_RX_VLAN_PRUNE_ENA;
+
+       if (on) {
+               vsi->info.sec_flags |= sec_flags;
+               vsi->info.sw_flags2 |= sw_flags2;
+       } else {
+               vsi->info.sec_flags &= ~sec_flags;
+               vsi->info.sw_flags2 &= ~sw_flags2;
+       }
+       vsi->info.sw_id = hw->port_info->sw_id;
+       (void)rte_memcpy(&ctxt.info, &vsi->info, sizeof(vsi->info));
+       ctxt.info.valid_sections =
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_SW_VALID |
+                                ICE_AQ_VSI_PROP_SECURITY_VALID);
+       ctxt.vsi_num = vsi->vsi_id;
+
+       ret = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+       if (ret) {
+               PMD_DRV_LOG(INFO, "Update VSI failed to %s vlan rx pruning",
+                           on ? "enable" : "disable");
+               return -EINVAL;
+       } else {
+               vsi->info.valid_sections |=
+                       rte_cpu_to_le_16(ICE_AQ_VSI_PROP_SW_VALID |
+                                        ICE_AQ_VSI_PROP_SECURITY_VALID);
+       }
+
+       /* consist with other drivers, allow untagged packet when vlan filter on */
+       if (on)
+               ret = ice_add_vlan_filter(vsi, 0);
+       else
+               ret = ice_remove_vlan_filter(vsi, 0);
+
+       return 0;
+}
+
+static int
+ice_vsi_config_vlan_stripping(struct ice_vsi *vsi, bool on)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       struct ice_vsi_ctx ctxt;
+       uint8_t vlan_flags;
+       int ret = 0;
+
+       /* Check if it has been already on or off */
+       if (vsi->info.valid_sections &
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_VLAN_VALID)) {
+               if (on) {
+                       if ((vsi->info.vlan_flags &
+                            ICE_AQ_VSI_VLAN_EMOD_M) ==
+                           ICE_AQ_VSI_VLAN_EMOD_STR_BOTH)
+                               return 0; /* already on */
+               } else {
+                       if ((vsi->info.vlan_flags &
+                            ICE_AQ_VSI_VLAN_EMOD_M) ==
+                           ICE_AQ_VSI_VLAN_EMOD_NOTHING)
+                               return 0; /* already off */
+               }
+       }
+
+       if (on)
+               vlan_flags = ICE_AQ_VSI_VLAN_EMOD_STR_BOTH;
+       else
+               vlan_flags = ICE_AQ_VSI_VLAN_EMOD_NOTHING;
+       vsi->info.vlan_flags &= ~(ICE_AQ_VSI_VLAN_EMOD_M);
+       vsi->info.vlan_flags |= vlan_flags;
+       (void)rte_memcpy(&ctxt.info, &vsi->info, sizeof(vsi->info));
+       ctxt.info.valid_sections =
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_VLAN_VALID);
+       ctxt.vsi_num = vsi->vsi_id;
+       ret = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+       if (ret) {
+               PMD_DRV_LOG(INFO, "Update VSI failed to %s vlan stripping",
+                           on ? "enable" : "disable");
+               return -EINVAL;
+       }
+
+       vsi->info.valid_sections |=
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_VLAN_VALID);
+
+       return ret;
+}
+
+static int
+ice_vlan_offload_set(struct rte_eth_dev *dev, int mask)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       struct rte_eth_rxmode *rxmode;
+
+       rxmode = &dev->data->dev_conf.rxmode;
+       if (mask & ETH_VLAN_FILTER_MASK) {
+               if (rxmode->offloads & DEV_RX_OFFLOAD_VLAN_FILTER)
+                       ice_vsi_config_vlan_filter(vsi, TRUE);
+               else
+                       ice_vsi_config_vlan_filter(vsi, FALSE);
+       }
+
+       if (mask & ETH_VLAN_STRIP_MASK) {
+               if (rxmode->offloads & DEV_RX_OFFLOAD_VLAN_STRIP)
+                       ice_vsi_config_vlan_stripping(vsi, TRUE);
+               else
+                       ice_vsi_config_vlan_stripping(vsi, FALSE);
+       }
+
+       if (mask & ETH_VLAN_EXTEND_MASK) {
+               if (rxmode->offloads & DEV_RX_OFFLOAD_VLAN_EXTEND)
+                       ice_vsi_config_double_vlan(vsi, TRUE);
+               else
+                       ice_vsi_config_double_vlan(vsi, FALSE);
+       }
+
+       return 0;
+}
+
+static int
+ice_vlan_tpid_set(struct rte_eth_dev *dev,
+                 enum rte_vlan_type vlan_type,
+                 uint16_t tpid)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       uint64_t reg_r = 0, reg_w = 0;
+       uint16_t reg_id = 0;
+       int ret = 0;
+       int qinq = dev->data->dev_conf.rxmode.offloads &
+                  DEV_RX_OFFLOAD_VLAN_EXTEND;
+
+       switch (vlan_type) {
+       case ETH_VLAN_TYPE_OUTER:
+               if (qinq)
+                       reg_id = 3;
+               else
+                       reg_id = 5;
+               break;
+       case ETH_VLAN_TYPE_INNER:
+               if (qinq) {
+                       reg_id = 5;
+               } else {
+                       PMD_DRV_LOG(ERR,
+                                   "Unsupported vlan type in single vlan.");
+                       return -EINVAL;
+               }
+               break;
+       default:
+               PMD_DRV_LOG(ERR, "Unsupported vlan type %d", vlan_type);
+               return -EINVAL;
+       }
+       reg_r = ICE_READ_REG(hw, GL_SWT_L2TAGCTRL(reg_id));
+       PMD_DRV_LOG(DEBUG, "Debug read from ICE GL_SWT_L2TAGCTRL[%d]: "
+                   "0x%08"PRIx64"", reg_id, reg_r);
+
+       reg_w = reg_r & (~(GL_SWT_L2TAGCTRL_ETHERTYPE_M));
+       reg_w |= ((uint64_t)tpid << GL_SWT_L2TAGCTRL_ETHERTYPE_S);
+       if (reg_r == reg_w) {
+               PMD_DRV_LOG(DEBUG, "No need to write");
+               return 0;
+       }
+
+       ICE_WRITE_REG(hw, GL_SWT_L2TAGCTRL(reg_id), reg_w);
+       PMD_DRV_LOG(DEBUG, "Debug write 0x%08"PRIx64" to "
+                   "ICE GL_SWT_L2TAGCTRL[%d]", reg_w, reg_id);
+
+       return ret;
+}
+
+static int
+ice_get_rss_lut(struct ice_vsi *vsi, uint8_t *lut, uint16_t lut_size)
+{
+       struct ice_pf *pf = ICE_VSI_TO_PF(vsi);
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       int ret;
+
+       if (!lut)
+               return -EINVAL;
+
+       if (pf->flags & ICE_FLAG_RSS_AQ_CAPABLE) {
+               ret = ice_aq_get_rss_lut(hw, vsi->idx, TRUE,
+                                        lut, lut_size);
+               if (ret) {
+                       PMD_DRV_LOG(ERR, "Failed to get RSS lookup table");
+                       return -EINVAL;
+               }
+       } else {
+               uint64_t *lut_dw = (uint64_t *)lut;
+               uint16_t i, lut_size_dw = lut_size / 4;
+
+               for (i = 0; i < lut_size_dw; i++)
+                       lut_dw[i] = ICE_READ_REG(hw, PFQF_HLUT(i));
+       }
+
+       return 0;
+}
+
+static int
+ice_set_rss_lut(struct ice_vsi *vsi, uint8_t *lut, uint16_t lut_size)
+{
+       struct ice_pf *pf;
+       struct ice_hw *hw;
+       int ret;
+
+       if (!vsi || !lut)
+               return -EINVAL;
+
+       pf = ICE_VSI_TO_PF(vsi);
+       hw = ICE_VSI_TO_HW(vsi);
+
+       if (pf->flags & ICE_FLAG_RSS_AQ_CAPABLE) {
+               ret = ice_aq_set_rss_lut(hw, vsi->idx, TRUE,
+                                        lut, lut_size);
+               if (ret) {
+                       PMD_DRV_LOG(ERR, "Failed to set RSS lookup table");
+                       return -EINVAL;
+               }
+       } else {
+               uint64_t *lut_dw = (uint64_t *)lut;
+               uint16_t i, lut_size_dw = lut_size / 4;
+
+               for (i = 0; i < lut_size_dw; i++)
+                       ICE_WRITE_REG(hw, PFQF_HLUT(i), lut_dw[i]);
+
+               ice_flush(hw);
+       }
+
+       return 0;
+}
+
+static int
+ice_rss_reta_update(struct rte_eth_dev *dev,
+                   struct rte_eth_rss_reta_entry64 *reta_conf,
+                   uint16_t reta_size)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       uint16_t i, lut_size = pf->hash_lut_size;
+       uint16_t idx, shift;
+       uint8_t *lut;
+       int ret;
+
+       if (reta_size != ICE_AQC_GSET_RSS_LUT_TABLE_SIZE_128 &&
+           reta_size != ICE_AQC_GSET_RSS_LUT_TABLE_SIZE_512 &&
+           reta_size != ICE_AQC_GSET_RSS_LUT_TABLE_SIZE_2K) {
+               PMD_DRV_LOG(ERR,
+                           "The size of hash lookup table configured (%d)"
+                           "doesn't match the number hardware can "
+                           "supported (128, 512, 2048)",
+                           reta_size);
+               return -EINVAL;
+       }
+
+       /* It MUST use the current LUT size to get the RSS lookup table,
+        * otherwise if will fail with -100 error code.
+        */
+       lut = rte_zmalloc(NULL,  RTE_MAX(reta_size, lut_size), 0);
+       if (!lut) {
+               PMD_DRV_LOG(ERR, "No memory can be allocated");
+               return -ENOMEM;
+       }
+       ret = ice_get_rss_lut(pf->main_vsi, lut, lut_size);
+       if (ret)
+               goto out;
+
+       for (i = 0; i < reta_size; i++) {
+               idx = i / RTE_RETA_GROUP_SIZE;
+               shift = i % RTE_RETA_GROUP_SIZE;
+               if (reta_conf[idx].mask & (1ULL << shift))
+                       lut[i] = reta_conf[idx].reta[shift];
+       }
+       ret = ice_set_rss_lut(pf->main_vsi, lut, reta_size);
+       if (ret == 0 && lut_size != reta_size) {
+               PMD_DRV_LOG(INFO,
+                           "The size of hash lookup table is changed from (%d) to (%d)",
+                           lut_size, reta_size);
+               pf->hash_lut_size = reta_size;
+       }
+
+out:
+       rte_free(lut);
+
+       return ret;
+}
+
+static int
+ice_rss_reta_query(struct rte_eth_dev *dev,
+                  struct rte_eth_rss_reta_entry64 *reta_conf,
+                  uint16_t reta_size)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       uint16_t i, lut_size = pf->hash_lut_size;
+       uint16_t idx, shift;
+       uint8_t *lut;
+       int ret;
+
+       if (reta_size != lut_size) {
+               PMD_DRV_LOG(ERR,
+                           "The size of hash lookup table configured (%d)"
+                           "doesn't match the number hardware can "
+                           "supported (%d)",
+                           reta_size, lut_size);
+               return -EINVAL;
+       }
+
+       lut = rte_zmalloc(NULL, reta_size, 0);
+       if (!lut) {
+               PMD_DRV_LOG(ERR, "No memory can be allocated");
+               return -ENOMEM;
+       }
+
+       ret = ice_get_rss_lut(pf->main_vsi, lut, reta_size);
+       if (ret)
+               goto out;
+
+       for (i = 0; i < reta_size; i++) {
+               idx = i / RTE_RETA_GROUP_SIZE;
+               shift = i % RTE_RETA_GROUP_SIZE;
+               if (reta_conf[idx].mask & (1ULL << shift))
+                       reta_conf[idx].reta[shift] = lut[i];
+       }
+
+out:
+       rte_free(lut);
+
+       return ret;
+}
+
+static int
+ice_set_rss_key(struct ice_vsi *vsi, uint8_t *key, uint8_t key_len)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       int ret = 0;
+
+       if (!key || key_len == 0) {
+               PMD_DRV_LOG(DEBUG, "No key to be configured");
+               return 0;
+       } else if (key_len != (VSIQF_HKEY_MAX_INDEX + 1) *
+                  sizeof(uint32_t)) {
+               PMD_DRV_LOG(ERR, "Invalid key length %u", key_len);
+               return -EINVAL;
+       }
+
+       struct ice_aqc_get_set_rss_keys *key_dw =
+               (struct ice_aqc_get_set_rss_keys *)key;
+
+       ret = ice_aq_set_rss_key(hw, vsi->idx, key_dw);
+       if (ret) {
+               PMD_DRV_LOG(ERR, "Failed to configure RSS key via AQ");
+               ret = -EINVAL;
+       }
+
+       return ret;
+}
+
+static int
+ice_get_rss_key(struct ice_vsi *vsi, uint8_t *key, uint8_t *key_len)
+{
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       int ret;
+
+       if (!key || !key_len)
+               return -EINVAL;
+
+       ret = ice_aq_get_rss_key
+               (hw, vsi->idx,
+                (struct ice_aqc_get_set_rss_keys *)key);
+       if (ret) {
+               PMD_DRV_LOG(ERR, "Failed to get RSS key via AQ");
+               return -EINVAL;
+       }
+       *key_len = (VSIQF_HKEY_MAX_INDEX + 1) * sizeof(uint32_t);
+
+       return 0;
+}
+
+static int
+ice_rss_hash_update(struct rte_eth_dev *dev,
+                   struct rte_eth_rss_conf *rss_conf)
+{
+       enum ice_status status = ICE_SUCCESS;
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+
+       /* set hash key */
+       status = ice_set_rss_key(vsi, rss_conf->rss_key, rss_conf->rss_key_len);
+       if (status)
+               return status;
+
+       /* TODO: hash enable config, ice_add_rss_cfg */
+       return 0;
+}
+
+static int
+ice_rss_hash_conf_get(struct rte_eth_dev *dev,
+                     struct rte_eth_rss_conf *rss_conf)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+
+       ice_get_rss_key(vsi, rss_conf->rss_key,
+                       &rss_conf->rss_key_len);
+
+       /* TODO: default set to 0 as hf config is not supported now */
+       rss_conf->rss_hf = 0;
+       return 0;
+}
+
+static void
+ice_promisc_enable(struct rte_eth_dev *dev)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       enum ice_status status;
+       uint8_t pmask;
+
+       pmask = ICE_PROMISC_UCAST_RX | ICE_PROMISC_UCAST_TX |
+               ICE_PROMISC_MCAST_RX | ICE_PROMISC_MCAST_TX;
+
+       status = ice_set_vsi_promisc(hw, vsi->idx, pmask, 0);
+       if (status == ICE_ERR_ALREADY_EXISTS)
+               PMD_DRV_LOG(DEBUG, "Promisc mode has already been enabled");
+       else if (status != ICE_SUCCESS)
+               PMD_DRV_LOG(ERR, "Failed to enable promisc, err=%d", status);
+}
+
+static void
+ice_promisc_disable(struct rte_eth_dev *dev)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       enum ice_status status;
+       uint8_t pmask;
+
+       pmask = ICE_PROMISC_UCAST_RX | ICE_PROMISC_UCAST_TX |
+               ICE_PROMISC_MCAST_RX | ICE_PROMISC_MCAST_TX;
+
+       status = ice_clear_vsi_promisc(hw, vsi->idx, pmask, 0);
+       if (status != ICE_SUCCESS)
+               PMD_DRV_LOG(ERR, "Failed to clear promisc, err=%d", status);
+}
+
+static void
+ice_allmulti_enable(struct rte_eth_dev *dev)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       enum ice_status status;
+       uint8_t pmask;
+
+       pmask = ICE_PROMISC_MCAST_RX | ICE_PROMISC_MCAST_TX;
+
+       status = ice_set_vsi_promisc(hw, vsi->idx, pmask, 0);
+       if (status != ICE_SUCCESS)
+               PMD_DRV_LOG(ERR, "Failed to enable allmulti, err=%d", status);
+}
+
+static void
+ice_allmulti_disable(struct rte_eth_dev *dev)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       enum ice_status status;
+       uint8_t pmask;
+
+       if (dev->data->promiscuous == 1)
+               return; /* must remain in all_multicast mode */
+
+       pmask = ICE_PROMISC_MCAST_RX | ICE_PROMISC_MCAST_TX;
+
+       status = ice_clear_vsi_promisc(hw, vsi->idx, pmask, 0);
+       if (status != ICE_SUCCESS)
+               PMD_DRV_LOG(ERR, "Failed to clear allmulti, err=%d", status);
+}
+
+static int ice_rx_queue_intr_enable(struct rte_eth_dev *dev,
+                                   uint16_t queue_id)
+{
+       struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
+       struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       uint32_t val;
+       uint16_t msix_intr;
+
+       msix_intr = intr_handle->intr_vec[queue_id];
+
+       val = GLINT_DYN_CTL_INTENA_M | GLINT_DYN_CTL_CLEARPBA_M |
+             GLINT_DYN_CTL_ITR_INDX_M;
+       val &= ~GLINT_DYN_CTL_WB_ON_ITR_M;
+
+       ICE_WRITE_REG(hw, GLINT_DYN_CTL(msix_intr), val);
+       rte_intr_enable(&pci_dev->intr_handle);
+
+       return 0;
+}
+
+static int ice_rx_queue_intr_disable(struct rte_eth_dev *dev,
+                                    uint16_t queue_id)
+{
+       struct rte_pci_device *pci_dev = ICE_DEV_TO_PCI(dev);
+       struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       uint16_t msix_intr;
+
+       msix_intr = intr_handle->intr_vec[queue_id];
+
+       ICE_WRITE_REG(hw, GLINT_DYN_CTL(msix_intr), GLINT_DYN_CTL_WB_ON_ITR_M);
+
+       return 0;
+}
+
+static int
+ice_fw_version_get(struct rte_eth_dev *dev, char *fw_version, size_t fw_size)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       u32 full_ver;
+       u8 ver, patch;
+       u16 build;
+       int ret;
+
+       full_ver = hw->nvm.oem_ver;
+       ver = (u8)(full_ver >> 24);
+       build = (u16)((full_ver >> 8) & 0xffff);
+       patch = (u8)(full_ver & 0xff);
+
+       ret = snprintf(fw_version, fw_size,
+                       "%d.%d%d 0x%08x %d.%d.%d",
+                       ((hw->nvm.ver >> 12) & 0xf),
+                       ((hw->nvm.ver >> 4) & 0xff),
+                       (hw->nvm.ver & 0xf), hw->nvm.eetrack,
+                       ver, build, patch);
+
+       /* add the size of '\0' */
+       ret += 1;
+       if (fw_size < (u32)ret)
+               return ret;
+       else
+               return 0;
+}
+
+static int
+ice_vsi_vlan_pvid_set(struct ice_vsi *vsi, struct ice_vsi_vlan_pvid_info *info)
+{
+       struct ice_hw *hw;
+       struct ice_vsi_ctx ctxt;
+       uint8_t vlan_flags = 0;
+       int ret;
+
+       if (!vsi || !info) {
+               PMD_DRV_LOG(ERR, "invalid parameters");
+               return -EINVAL;
+       }
+
+       if (info->on) {
+               vsi->info.pvid = info->config.pvid;
+               /**
+                * If insert pvid is enabled, only tagged pkts are
+                * allowed to be sent out.
+                */
+               vlan_flags = ICE_AQ_VSI_PVLAN_INSERT_PVID |
+                            ICE_AQ_VSI_VLAN_MODE_UNTAGGED;
+       } else {
+               vsi->info.pvid = 0;
+               if (info->config.reject.tagged == 0)
+                       vlan_flags |= ICE_AQ_VSI_VLAN_MODE_TAGGED;
+
+               if (info->config.reject.untagged == 0)
+                       vlan_flags |= ICE_AQ_VSI_VLAN_MODE_UNTAGGED;
+       }
+       vsi->info.vlan_flags &= ~(ICE_AQ_VSI_PVLAN_INSERT_PVID |
+                                 ICE_AQ_VSI_VLAN_MODE_M);
+       vsi->info.vlan_flags |= vlan_flags;
+       memset(&ctxt, 0, sizeof(ctxt));
+       rte_memcpy(&ctxt.info, &vsi->info, sizeof(vsi->info));
+       ctxt.info.valid_sections =
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_VLAN_VALID);
+       ctxt.vsi_num = vsi->vsi_id;
+
+       hw = ICE_VSI_TO_HW(vsi);
+       ret = ice_update_vsi(hw, vsi->idx, &ctxt, NULL);
+       if (ret != ICE_SUCCESS) {
+               PMD_DRV_LOG(ERR,
+                           "update VSI for VLAN insert failed, err %d",
+                           ret);
+               return -EINVAL;
+       }
+
+       vsi->info.valid_sections |=
+               rte_cpu_to_le_16(ICE_AQ_VSI_PROP_VLAN_VALID);
+
+       return ret;
+}
+
+static int
+ice_vlan_pvid_set(struct rte_eth_dev *dev, uint16_t pvid, int on)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_vsi *vsi = pf->main_vsi;
+       struct rte_eth_dev_data *data = pf->dev_data;
+       struct ice_vsi_vlan_pvid_info info;
+       int ret;
+
+       memset(&info, 0, sizeof(info));
+       info.on = on;
+       if (info.on) {
+               info.config.pvid = pvid;
+       } else {
+               info.config.reject.tagged =
+                       data->dev_conf.txmode.hw_vlan_reject_tagged;
+               info.config.reject.untagged =
+                       data->dev_conf.txmode.hw_vlan_reject_untagged;
+       }
+
+       ret = ice_vsi_vlan_pvid_set(vsi, &info);
+       if (ret < 0) {
+               PMD_DRV_LOG(ERR, "Failed to set pvid.");
+               return -EINVAL;
+       }
+
+       return 0;
+}
+
+static int
+ice_get_eeprom_length(struct rte_eth_dev *dev)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+       /* Convert word count to byte count */
+       return hw->nvm.sr_words << 1;
+}
+
+static int
+ice_get_eeprom(struct rte_eth_dev *dev,
+              struct rte_dev_eeprom_info *eeprom)
+{
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       uint16_t *data = eeprom->data;
+       uint16_t first_word, last_word, nwords;
+       enum ice_status status = ICE_SUCCESS;
+
+       first_word = eeprom->offset >> 1;
+       last_word = (eeprom->offset + eeprom->length - 1) >> 1;
+       nwords = last_word - first_word + 1;
+
+       if (first_word >= hw->nvm.sr_words ||
+           last_word >= hw->nvm.sr_words) {
+               PMD_DRV_LOG(ERR, "Requested EEPROM bytes out of range.");
+               return -EINVAL;
+       }
+
+       eeprom->magic = hw->vendor_id | (hw->device_id << 16);
+
+       status = ice_read_sr_buf(hw, first_word, &nwords, data);
+       if (status) {
+               PMD_DRV_LOG(ERR, "EEPROM read failed.");
+               eeprom->length = sizeof(uint16_t) * nwords;
+               return -EIO;
+       }
+
+       return 0;
+}
+
+static void
+ice_stat_update_32(struct ice_hw *hw,
+                  uint32_t reg,
+                  bool offset_loaded,
+                  uint64_t *offset,
+                  uint64_t *stat)
+{
+       uint64_t new_data;
+
+       new_data = (uint64_t)ICE_READ_REG(hw, reg);
+       if (!offset_loaded)
+               *offset = new_data;
+
+       if (new_data >= *offset)
+               *stat = (uint64_t)(new_data - *offset);
+       else
+               *stat = (uint64_t)((new_data +
+                                   ((uint64_t)1 << ICE_32_BIT_WIDTH))
+                                  - *offset);
+}
+
+static void
+ice_stat_update_40(struct ice_hw *hw,
+                  uint32_t hireg,
+                  uint32_t loreg,
+                  bool offset_loaded,
+                  uint64_t *offset,
+                  uint64_t *stat)
+{
+       uint64_t new_data;
+
+       new_data = (uint64_t)ICE_READ_REG(hw, loreg);
+       new_data |= (uint64_t)(ICE_READ_REG(hw, hireg) & ICE_8_BIT_MASK) <<
+                   ICE_32_BIT_WIDTH;
+
+       if (!offset_loaded)
+               *offset = new_data;
+
+       if (new_data >= *offset)
+               *stat = new_data - *offset;
+       else
+               *stat = (uint64_t)((new_data +
+                                   ((uint64_t)1 << ICE_40_BIT_WIDTH)) -
+                                  *offset);
+
+       *stat &= ICE_40_BIT_MASK;
+}
+
+/* Get all the statistics of a VSI */
+static void
+ice_update_vsi_stats(struct ice_vsi *vsi)
+{
+       struct ice_eth_stats *oes = &vsi->eth_stats_offset;
+       struct ice_eth_stats *nes = &vsi->eth_stats;
+       struct ice_hw *hw = ICE_VSI_TO_HW(vsi);
+       int idx = rte_le_to_cpu_16(vsi->vsi_id);
+
+       ice_stat_update_40(hw, GLV_GORCH(idx), GLV_GORCL(idx),
+                          vsi->offset_loaded, &oes->rx_bytes,
+                          &nes->rx_bytes);
+       ice_stat_update_40(hw, GLV_UPRCH(idx), GLV_UPRCL(idx),
+                          vsi->offset_loaded, &oes->rx_unicast,
+                          &nes->rx_unicast);
+       ice_stat_update_40(hw, GLV_MPRCH(idx), GLV_MPRCL(idx),
+                          vsi->offset_loaded, &oes->rx_multicast,
+                          &nes->rx_multicast);
+       ice_stat_update_40(hw, GLV_BPRCH(idx), GLV_BPRCL(idx),
+                          vsi->offset_loaded, &oes->rx_broadcast,
+                          &nes->rx_broadcast);
+       /* exclude CRC bytes */
+       nes->rx_bytes -= (nes->rx_unicast + nes->rx_multicast +
+                         nes->rx_broadcast) * ETHER_CRC_LEN;
+
+       ice_stat_update_32(hw, GLV_RDPC(idx), vsi->offset_loaded,
+                          &oes->rx_discards, &nes->rx_discards);
+       /* GLV_REPC not supported */
+       /* GLV_RMPC not supported */
+       ice_stat_update_32(hw, GLSWID_RUPP(idx), vsi->offset_loaded,
+                          &oes->rx_unknown_protocol,
+                          &nes->rx_unknown_protocol);
+       ice_stat_update_40(hw, GLV_GOTCH(idx), GLV_GOTCL(idx),
+                          vsi->offset_loaded, &oes->tx_bytes,
+                          &nes->tx_bytes);
+       ice_stat_update_40(hw, GLV_UPTCH(idx), GLV_UPTCL(idx),
+                          vsi->offset_loaded, &oes->tx_unicast,
+                          &nes->tx_unicast);
+       ice_stat_update_40(hw, GLV_MPTCH(idx), GLV_MPTCL(idx),
+                          vsi->offset_loaded, &oes->tx_multicast,
+                          &nes->tx_multicast);
+       ice_stat_update_40(hw, GLV_BPTCH(idx), GLV_BPTCL(idx),
+                          vsi->offset_loaded,  &oes->tx_broadcast,
+                          &nes->tx_broadcast);
+       /* GLV_TDPC not supported */
+       ice_stat_update_32(hw, GLV_TEPC(idx), vsi->offset_loaded,
+                          &oes->tx_errors, &nes->tx_errors);
+       vsi->offset_loaded = true;
+
+       PMD_DRV_LOG(DEBUG, "************** VSI[%u] stats start **************",
+                   vsi->vsi_id);
+       PMD_DRV_LOG(DEBUG, "rx_bytes:            %"PRIu64"", nes->rx_bytes);
+       PMD_DRV_LOG(DEBUG, "rx_unicast:          %"PRIu64"", nes->rx_unicast);
+       PMD_DRV_LOG(DEBUG, "rx_multicast:        %"PRIu64"", nes->rx_multicast);
+       PMD_DRV_LOG(DEBUG, "rx_broadcast:        %"PRIu64"", nes->rx_broadcast);
+       PMD_DRV_LOG(DEBUG, "rx_discards:         %"PRIu64"", nes->rx_discards);
+       PMD_DRV_LOG(DEBUG, "rx_unknown_protocol: %"PRIu64"",
+                   nes->rx_unknown_protocol);
+       PMD_DRV_LOG(DEBUG, "tx_bytes:            %"PRIu64"", nes->tx_bytes);
+       PMD_DRV_LOG(DEBUG, "tx_unicast:          %"PRIu64"", nes->tx_unicast);
+       PMD_DRV_LOG(DEBUG, "tx_multicast:        %"PRIu64"", nes->tx_multicast);
+       PMD_DRV_LOG(DEBUG, "tx_broadcast:        %"PRIu64"", nes->tx_broadcast);
+       PMD_DRV_LOG(DEBUG, "tx_discards:         %"PRIu64"", nes->tx_discards);
+       PMD_DRV_LOG(DEBUG, "tx_errors:           %"PRIu64"", nes->tx_errors);
+       PMD_DRV_LOG(DEBUG, "************** VSI[%u] stats end ****************",
+                   vsi->vsi_id);
+}
+
+static void
+ice_read_stats_registers(struct ice_pf *pf, struct ice_hw *hw)
+{
+       struct ice_hw_port_stats *ns = &pf->stats; /* new stats */
+       struct ice_hw_port_stats *os = &pf->stats_offset; /* old stats */
+
+       /* Get statistics of struct ice_eth_stats */
+       ice_stat_update_40(hw, GLPRT_GORCH(hw->port_info->lport),
+                          GLPRT_GORCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.rx_bytes,
+                          &ns->eth.rx_bytes);
+       ice_stat_update_40(hw, GLPRT_UPRCH(hw->port_info->lport),
+                          GLPRT_UPRCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.rx_unicast,
+                          &ns->eth.rx_unicast);
+       ice_stat_update_40(hw, GLPRT_MPRCH(hw->port_info->lport),
+                          GLPRT_MPRCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.rx_multicast,
+                          &ns->eth.rx_multicast);
+       ice_stat_update_40(hw, GLPRT_BPRCH(hw->port_info->lport),
+                          GLPRT_BPRCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.rx_broadcast,
+                          &ns->eth.rx_broadcast);
+       ice_stat_update_32(hw, PRTRPB_RDPC,
+                          pf->offset_loaded, &os->eth.rx_discards,
+                          &ns->eth.rx_discards);
+
+       /* Workaround: CRC size should not be included in byte statistics,
+        * so subtract ETHER_CRC_LEN from the byte counter for each rx packet.
+        */
+       ns->eth.rx_bytes -= (ns->eth.rx_unicast + ns->eth.rx_multicast +
+                            ns->eth.rx_broadcast) * ETHER_CRC_LEN;
+
+       /* GLPRT_REPC not supported */
+       /* GLPRT_RMPC not supported */
+       ice_stat_update_32(hw, GLSWID_RUPP(hw->port_info->lport),
+                          pf->offset_loaded,
+                          &os->eth.rx_unknown_protocol,
+                          &ns->eth.rx_unknown_protocol);
+       ice_stat_update_40(hw, GLPRT_GOTCH(hw->port_info->lport),
+                          GLPRT_GOTCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.tx_bytes,
+                          &ns->eth.tx_bytes);
+       ice_stat_update_40(hw, GLPRT_UPTCH(hw->port_info->lport),
+                          GLPRT_UPTCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.tx_unicast,
+                          &ns->eth.tx_unicast);
+       ice_stat_update_40(hw, GLPRT_MPTCH(hw->port_info->lport),
+                          GLPRT_MPTCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.tx_multicast,
+                          &ns->eth.tx_multicast);
+       ice_stat_update_40(hw, GLPRT_BPTCH(hw->port_info->lport),
+                          GLPRT_BPTCL(hw->port_info->lport),
+                          pf->offset_loaded, &os->eth.tx_broadcast,
+                          &ns->eth.tx_broadcast);
+       ns->eth.tx_bytes -= (ns->eth.tx_unicast + ns->eth.tx_multicast +
+                            ns->eth.tx_broadcast) * ETHER_CRC_LEN;
+
+       /* GLPRT_TEPC not supported */
+
+       /* additional port specific stats */
+       ice_stat_update_32(hw, GLPRT_TDOLD(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_dropped_link_down,
+                          &ns->tx_dropped_link_down);
+       ice_stat_update_32(hw, GLPRT_CRCERRS(hw->port_info->lport),
+                          pf->offset_loaded, &os->crc_errors,
+                          &ns->crc_errors);
+       ice_stat_update_32(hw, GLPRT_ILLERRC(hw->port_info->lport),
+                          pf->offset_loaded, &os->illegal_bytes,
+                          &ns->illegal_bytes);
+       /* GLPRT_ERRBC not supported */
+       ice_stat_update_32(hw, GLPRT_MLFC(hw->port_info->lport),
+                          pf->offset_loaded, &os->mac_local_faults,
+                          &ns->mac_local_faults);
+       ice_stat_update_32(hw, GLPRT_MRFC(hw->port_info->lport),
+                          pf->offset_loaded, &os->mac_remote_faults,
+                          &ns->mac_remote_faults);
+
+       ice_stat_update_32(hw, GLPRT_RLEC(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_len_errors,
+                          &ns->rx_len_errors);
+
+       ice_stat_update_32(hw, GLPRT_LXONRXC(hw->port_info->lport),
+                          pf->offset_loaded, &os->link_xon_rx,
+                          &ns->link_xon_rx);
+       ice_stat_update_32(hw, GLPRT_LXOFFRXC(hw->port_info->lport),
+                          pf->offset_loaded, &os->link_xoff_rx,
+                          &ns->link_xoff_rx);
+       ice_stat_update_32(hw, GLPRT_LXONTXC(hw->port_info->lport),
+                          pf->offset_loaded, &os->link_xon_tx,
+                          &ns->link_xon_tx);
+       ice_stat_update_32(hw, GLPRT_LXOFFTXC(hw->port_info->lport),
+                          pf->offset_loaded, &os->link_xoff_tx,
+                          &ns->link_xoff_tx);
+       ice_stat_update_40(hw, GLPRT_PRC64H(hw->port_info->lport),
+                          GLPRT_PRC64L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_64,
+                          &ns->rx_size_64);
+       ice_stat_update_40(hw, GLPRT_PRC127H(hw->port_info->lport),
+                          GLPRT_PRC127L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_127,
+                          &ns->rx_size_127);
+       ice_stat_update_40(hw, GLPRT_PRC255H(hw->port_info->lport),
+                          GLPRT_PRC255L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_255,
+                          &ns->rx_size_255);
+       ice_stat_update_40(hw, GLPRT_PRC511H(hw->port_info->lport),
+                          GLPRT_PRC511L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_511,
+                          &ns->rx_size_511);
+       ice_stat_update_40(hw, GLPRT_PRC1023H(hw->port_info->lport),
+                          GLPRT_PRC1023L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_1023,
+                          &ns->rx_size_1023);
+       ice_stat_update_40(hw, GLPRT_PRC1522H(hw->port_info->lport),
+                          GLPRT_PRC1522L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_1522,
+                          &ns->rx_size_1522);
+       ice_stat_update_40(hw, GLPRT_PRC9522H(hw->port_info->lport),
+                          GLPRT_PRC9522L(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_size_big,
+                          &ns->rx_size_big);
+       ice_stat_update_32(hw, GLPRT_RUC(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_undersize,
+                          &ns->rx_undersize);
+       ice_stat_update_32(hw, GLPRT_RFC(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_fragments,
+                          &ns->rx_fragments);
+       ice_stat_update_32(hw, GLPRT_ROC(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_oversize,
+                          &ns->rx_oversize);
+       ice_stat_update_32(hw, GLPRT_RJC(hw->port_info->lport),
+                          pf->offset_loaded, &os->rx_jabber,
+                          &ns->rx_jabber);
+       ice_stat_update_40(hw, GLPRT_PTC64H(hw->port_info->lport),
+                          GLPRT_PTC64L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_64,
+                          &ns->tx_size_64);
+       ice_stat_update_40(hw, GLPRT_PTC127H(hw->port_info->lport),
+                          GLPRT_PTC127L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_127,
+                          &ns->tx_size_127);
+       ice_stat_update_40(hw, GLPRT_PTC255H(hw->port_info->lport),
+                          GLPRT_PTC255L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_255,
+                          &ns->tx_size_255);
+       ice_stat_update_40(hw, GLPRT_PTC511H(hw->port_info->lport),
+                          GLPRT_PTC511L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_511,
+                          &ns->tx_size_511);
+       ice_stat_update_40(hw, GLPRT_PTC1023H(hw->port_info->lport),
+                          GLPRT_PTC1023L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_1023,
+                          &ns->tx_size_1023);
+       ice_stat_update_40(hw, GLPRT_PTC1522H(hw->port_info->lport),
+                          GLPRT_PTC1522L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_1522,
+                          &ns->tx_size_1522);
+       ice_stat_update_40(hw, GLPRT_PTC9522H(hw->port_info->lport),
+                          GLPRT_PTC9522L(hw->port_info->lport),
+                          pf->offset_loaded, &os->tx_size_big,
+                          &ns->tx_size_big);
+
+       /* GLPRT_MSPDC not supported */
+       /* GLPRT_XEC not supported */
+
+       pf->offset_loaded = true;
+
+       if (pf->main_vsi)
+               ice_update_vsi_stats(pf->main_vsi);
+}
+
+/* Get all statistics of a port */
+static int
+ice_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       struct ice_hw_port_stats *ns = &pf->stats; /* new stats */
+
+       /* call read registers - updates values, now write them to struct */
+       ice_read_stats_registers(pf, hw);
+
+       stats->ipackets = pf->main_vsi->eth_stats.rx_unicast +
+                         pf->main_vsi->eth_stats.rx_multicast +
+                         pf->main_vsi->eth_stats.rx_broadcast -
+                         pf->main_vsi->eth_stats.rx_discards;
+       stats->opackets = ns->eth.tx_unicast +
+                         ns->eth.tx_multicast +
+                         ns->eth.tx_broadcast;
+       stats->ibytes   = pf->main_vsi->eth_stats.rx_bytes;
+       stats->obytes   = ns->eth.tx_bytes;
+       stats->oerrors  = ns->eth.tx_errors +
+                         pf->main_vsi->eth_stats.tx_errors;
+
+       /* Rx Errors */
+       stats->imissed  = ns->eth.rx_discards +
+                         pf->main_vsi->eth_stats.rx_discards;
+       stats->ierrors  = ns->crc_errors +
+                         ns->rx_undersize +
+                         ns->rx_oversize + ns->rx_fragments + ns->rx_jabber;
+
+       PMD_DRV_LOG(DEBUG, "*************** PF stats start *****************");
+       PMD_DRV_LOG(DEBUG, "rx_bytes:   %"PRIu64"", ns->eth.rx_bytes);
+       PMD_DRV_LOG(DEBUG, "rx_unicast: %"PRIu64"", ns->eth.rx_unicast);
+       PMD_DRV_LOG(DEBUG, "rx_multicast:%"PRIu64"", ns->eth.rx_multicast);
+       PMD_DRV_LOG(DEBUG, "rx_broadcast:%"PRIu64"", ns->eth.rx_broadcast);
+       PMD_DRV_LOG(DEBUG, "rx_discards:%"PRIu64"", ns->eth.rx_discards);
+       PMD_DRV_LOG(DEBUG, "vsi rx_discards:%"PRIu64"",
+                   pf->main_vsi->eth_stats.rx_discards);
+       PMD_DRV_LOG(DEBUG, "rx_unknown_protocol:  %"PRIu64"",
+                   ns->eth.rx_unknown_protocol);
+       PMD_DRV_LOG(DEBUG, "tx_bytes:   %"PRIu64"", ns->eth.tx_bytes);
+       PMD_DRV_LOG(DEBUG, "tx_unicast: %"PRIu64"", ns->eth.tx_unicast);
+       PMD_DRV_LOG(DEBUG, "tx_multicast:%"PRIu64"", ns->eth.tx_multicast);
+       PMD_DRV_LOG(DEBUG, "tx_broadcast:%"PRIu64"", ns->eth.tx_broadcast);
+       PMD_DRV_LOG(DEBUG, "tx_discards:%"PRIu64"", ns->eth.tx_discards);
+       PMD_DRV_LOG(DEBUG, "vsi tx_discards:%"PRIu64"",
+                   pf->main_vsi->eth_stats.tx_discards);
+       PMD_DRV_LOG(DEBUG, "tx_errors:          %"PRIu64"", ns->eth.tx_errors);
+
+       PMD_DRV_LOG(DEBUG, "tx_dropped_link_down:       %"PRIu64"",
+                   ns->tx_dropped_link_down);
+       PMD_DRV_LOG(DEBUG, "crc_errors: %"PRIu64"", ns->crc_errors);
+       PMD_DRV_LOG(DEBUG, "illegal_bytes:      %"PRIu64"",
+                   ns->illegal_bytes);
+       PMD_DRV_LOG(DEBUG, "error_bytes:        %"PRIu64"", ns->error_bytes);
+       PMD_DRV_LOG(DEBUG, "mac_local_faults:   %"PRIu64"",
+                   ns->mac_local_faults);
+       PMD_DRV_LOG(DEBUG, "mac_remote_faults:  %"PRIu64"",
+                   ns->mac_remote_faults);
+       PMD_DRV_LOG(DEBUG, "link_xon_rx:        %"PRIu64"", ns->link_xon_rx);
+       PMD_DRV_LOG(DEBUG, "link_xoff_rx:       %"PRIu64"", ns->link_xoff_rx);
+       PMD_DRV_LOG(DEBUG, "link_xon_tx:        %"PRIu64"", ns->link_xon_tx);
+       PMD_DRV_LOG(DEBUG, "link_xoff_tx:       %"PRIu64"", ns->link_xoff_tx);
+       PMD_DRV_LOG(DEBUG, "rx_size_64:         %"PRIu64"", ns->rx_size_64);
+       PMD_DRV_LOG(DEBUG, "rx_size_127:        %"PRIu64"", ns->rx_size_127);
+       PMD_DRV_LOG(DEBUG, "rx_size_255:        %"PRIu64"", ns->rx_size_255);
+       PMD_DRV_LOG(DEBUG, "rx_size_511:        %"PRIu64"", ns->rx_size_511);
+       PMD_DRV_LOG(DEBUG, "rx_size_1023:       %"PRIu64"", ns->rx_size_1023);
+       PMD_DRV_LOG(DEBUG, "rx_size_1522:       %"PRIu64"", ns->rx_size_1522);
+       PMD_DRV_LOG(DEBUG, "rx_size_big:        %"PRIu64"", ns->rx_size_big);
+       PMD_DRV_LOG(DEBUG, "rx_undersize:       %"PRIu64"", ns->rx_undersize);
+       PMD_DRV_LOG(DEBUG, "rx_fragments:       %"PRIu64"", ns->rx_fragments);
+       PMD_DRV_LOG(DEBUG, "rx_oversize:        %"PRIu64"", ns->rx_oversize);
+       PMD_DRV_LOG(DEBUG, "rx_jabber:          %"PRIu64"", ns->rx_jabber);
+       PMD_DRV_LOG(DEBUG, "tx_size_64:         %"PRIu64"", ns->tx_size_64);
+       PMD_DRV_LOG(DEBUG, "tx_size_127:        %"PRIu64"", ns->tx_size_127);
+       PMD_DRV_LOG(DEBUG, "tx_size_255:        %"PRIu64"", ns->tx_size_255);
+       PMD_DRV_LOG(DEBUG, "tx_size_511:        %"PRIu64"", ns->tx_size_511);
+       PMD_DRV_LOG(DEBUG, "tx_size_1023:       %"PRIu64"", ns->tx_size_1023);
+       PMD_DRV_LOG(DEBUG, "tx_size_1522:       %"PRIu64"", ns->tx_size_1522);
+       PMD_DRV_LOG(DEBUG, "tx_size_big:        %"PRIu64"", ns->tx_size_big);
+       PMD_DRV_LOG(DEBUG, "rx_len_errors:      %"PRIu64"", ns->rx_len_errors);
+       PMD_DRV_LOG(DEBUG, "************* PF stats end ****************");
+       return 0;
+}
+
+/* Reset the statistics */
+static void
+ice_stats_reset(struct rte_eth_dev *dev)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+       /* Mark PF and VSI stats to update the offset, aka "reset" */
+       pf->offset_loaded = false;
+       if (pf->main_vsi)
+               pf->main_vsi->offset_loaded = false;
+
+       /* read the stats, reading current register values into offset */
+       ice_read_stats_registers(pf, hw);
+}
+
+static uint32_t
+ice_xstats_calc_num(void)
+{
+       uint32_t num;
+
+       num = ICE_NB_ETH_XSTATS + ICE_NB_HW_PORT_XSTATS;
+
+       return num;
+}
+
+static int
+ice_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *xstats,
+              unsigned int n)
+{
+       struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
+       struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+       unsigned int i;
+       unsigned int count;
+       struct ice_hw_port_stats *hw_stats = &pf->stats;
+
+       count = ice_xstats_calc_num();
+       if (n < count)
+               return count;
+
+       ice_read_stats_registers(pf, hw);
+
+       if (!xstats)
+               return 0;
+
+       count = 0;
+
+       /* Get stats from ice_eth_stats struct */
+       for (i = 0; i < ICE_NB_ETH_XSTATS; i++) {
+               xstats[count].value =
+                       *(uint64_t *)((char *)&hw_stats->eth +
+                                     ice_stats_strings[i].offset);
+               xstats[count].id = count;
+               count++;
+       }
+
+       /* Get individiual stats from ice_hw_port struct */
+       for (i = 0; i < ICE_NB_HW_PORT_XSTATS; i++) {
+               xstats[count].value =
+                       *(uint64_t *)((char *)hw_stats +
+                                     ice_hw_port_strings[i].offset);
+               xstats[count].id = count;
+               count++;
+       }
+
+       return count;
+}
+
+static int ice_xstats_get_names(__rte_unused struct rte_eth_dev *dev,
+                               struct rte_eth_xstat_name *xstats_names,
+                               __rte_unused unsigned int limit)
+{
+       unsigned int count = 0;
+       unsigned int i;
+
+       if (!xstats_names)
+               return ice_xstats_calc_num();
+
+       /* Note: limit checked in rte_eth_xstats_names() */
+
+       /* Get stats from ice_eth_stats struct */
+       for (i = 0; i < ICE_NB_ETH_XSTATS; i++) {
+               strlcpy(xstats_names[count].name, ice_stats_strings[i].name,
+                       sizeof(xstats_names[count].name));
+               count++;
+       }
+
+       /* Get individiual stats from ice_hw_port struct */
+       for (i = 0; i < ICE_NB_HW_PORT_XSTATS; i++) {
+               strlcpy(xstats_names[count].name, ice_hw_port_strings[i].name,
+                       sizeof(xstats_names[count].name));
+               count++;
+       }
+
+       return count;
+}
+
 static int
 ice_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
              struct rte_pci_device *pci_dev)