net/ice: support ESP/AH/L2TP
[dpdk.git] / drivers / net / ice / ice_switch_filter.c
index 6fb6348..9c87a16 100644 (file)
  * Copyright(c) 2019 Intel Corporation
  */
 
+#include <sys/queue.h>
+#include <stdio.h>
+#include <errno.h>
+#include <stdint.h>
+#include <string.h>
+#include <unistd.h>
+#include <stdarg.h>
+#include <rte_debug.h>
+#include <rte_ether.h>
+#include <rte_ethdev_driver.h>
+#include <rte_log.h>
+#include <rte_malloc.h>
+#include <rte_eth_ctrl.h>
+#include <rte_tailq.h>
+#include <rte_flow_driver.h>
+#include <rte_flow.h>
+#include "base/ice_type.h"
+#include "base/ice_switch.h"
+#include "ice_logs.h"
+#include "ice_ethdev.h"
+#include "ice_generic_flow.h"
+
+
+#define MAX_QGRP_NUM_TYPE 7
+
+#define ICE_SW_INSET_ETHER ( \
+       ICE_INSET_DMAC | ICE_INSET_SMAC | ICE_INSET_ETHERTYPE)
+#define ICE_SW_INSET_MAC_VLAN ( \
+               ICE_INSET_DMAC | ICE_INSET_SMAC | ICE_INSET_ETHERTYPE | \
+               ICE_INSET_VLAN_OUTER)
+#define ICE_SW_INSET_MAC_IPV4 ( \
+       ICE_INSET_DMAC | ICE_INSET_IPV4_DST | ICE_INSET_IPV4_SRC | \
+       ICE_INSET_IPV4_PROTO | ICE_INSET_IPV4_TTL | ICE_INSET_IPV4_TOS)
+#define ICE_SW_INSET_MAC_IPV4_TCP ( \
+       ICE_INSET_DMAC | ICE_INSET_IPV4_DST | ICE_INSET_IPV4_SRC | \
+       ICE_INSET_IPV4_TTL | ICE_INSET_IPV4_TOS | \
+       ICE_INSET_TCP_DST_PORT | ICE_INSET_TCP_SRC_PORT)
+#define ICE_SW_INSET_MAC_IPV4_UDP ( \
+       ICE_INSET_DMAC | ICE_INSET_IPV4_DST | ICE_INSET_IPV4_SRC | \
+       ICE_INSET_IPV4_TTL | ICE_INSET_IPV4_TOS | \
+       ICE_INSET_UDP_DST_PORT | ICE_INSET_UDP_SRC_PORT)
+#define ICE_SW_INSET_MAC_IPV6 ( \
+       ICE_INSET_DMAC | ICE_INSET_IPV6_DST | ICE_INSET_IPV6_SRC | \
+       ICE_INSET_IPV6_TC | ICE_INSET_IPV6_HOP_LIMIT | \
+       ICE_INSET_IPV6_NEXT_HDR)
+#define ICE_SW_INSET_MAC_IPV6_TCP ( \
+       ICE_INSET_DMAC | ICE_INSET_IPV6_DST | ICE_INSET_IPV6_SRC | \
+       ICE_INSET_IPV6_HOP_LIMIT | ICE_INSET_IPV6_TC | \
+       ICE_INSET_TCP_DST_PORT | ICE_INSET_TCP_SRC_PORT)
+#define ICE_SW_INSET_MAC_IPV6_UDP ( \
+       ICE_INSET_DMAC | ICE_INSET_IPV6_DST | ICE_INSET_IPV6_SRC | \
+       ICE_INSET_IPV6_HOP_LIMIT | ICE_INSET_IPV6_TC | \
+       ICE_INSET_UDP_DST_PORT | ICE_INSET_UDP_SRC_PORT)
+#define ICE_SW_INSET_DIST_NVGRE_IPV4 ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_DMAC | ICE_INSET_TUN_NVGRE_TNI | ICE_INSET_IPV4_DST)
+#define ICE_SW_INSET_DIST_VXLAN_IPV4 ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_DMAC | ICE_INSET_TUN_VXLAN_VNI | ICE_INSET_IPV4_DST)
+#define ICE_SW_INSET_DIST_NVGRE_IPV4_TCP ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_TCP_SRC_PORT | ICE_INSET_TUN_TCP_DST_PORT | \
+       ICE_INSET_TUN_DMAC | ICE_INSET_TUN_NVGRE_TNI | ICE_INSET_IPV4_DST)
+#define ICE_SW_INSET_DIST_NVGRE_IPV4_UDP ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_UDP_SRC_PORT | ICE_INSET_TUN_UDP_DST_PORT | \
+       ICE_INSET_TUN_DMAC | ICE_INSET_TUN_NVGRE_TNI | ICE_INSET_IPV4_DST)
+#define ICE_SW_INSET_DIST_VXLAN_IPV4_TCP ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_TCP_SRC_PORT | ICE_INSET_TUN_TCP_DST_PORT | \
+       ICE_INSET_TUN_DMAC | ICE_INSET_TUN_VXLAN_VNI | ICE_INSET_IPV4_DST)
+#define ICE_SW_INSET_DIST_VXLAN_IPV4_UDP ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_UDP_SRC_PORT | ICE_INSET_TUN_UDP_DST_PORT | \
+       ICE_INSET_TUN_DMAC | ICE_INSET_TUN_VXLAN_VNI | ICE_INSET_IPV4_DST)
+#define ICE_SW_INSET_PERM_TUNNEL_IPV4 ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_IPV4_PROTO | ICE_INSET_TUN_IPV4_TOS)
+#define ICE_SW_INSET_PERM_TUNNEL_IPV4_TCP ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_TCP_SRC_PORT | ICE_INSET_TUN_TCP_DST_PORT | \
+       ICE_INSET_TUN_IPV4_TOS)
+#define ICE_SW_INSET_PERM_TUNNEL_IPV4_UDP ( \
+       ICE_INSET_TUN_IPV4_SRC | ICE_INSET_TUN_IPV4_DST | \
+       ICE_INSET_TUN_UDP_SRC_PORT | ICE_INSET_TUN_UDP_DST_PORT | \
+       ICE_INSET_TUN_IPV4_TOS)
+#define ICE_SW_INSET_MAC_PPPOE  ( \
+       ICE_INSET_VLAN_OUTER | ICE_INSET_VLAN_INNER | \
+       ICE_INSET_DMAC | ICE_INSET_ETHERTYPE | ICE_INSET_PPPOE_SESSION)
+#define ICE_SW_INSET_MAC_PPPOE_PROTO  ( \
+       ICE_INSET_VLAN_OUTER | ICE_INSET_VLAN_INNER | \
+       ICE_INSET_DMAC | ICE_INSET_ETHERTYPE | ICE_INSET_PPPOE_SESSION | \
+       ICE_INSET_PPPOE_PROTO)
+
+struct sw_meta {
+       struct ice_adv_lkup_elem *list;
+       uint16_t lkups_num;
+       struct ice_adv_rule_info rule_info;
+};
+
+static struct ice_flow_parser ice_switch_dist_parser_os;
+static struct ice_flow_parser ice_switch_dist_parser_comms;
+static struct ice_flow_parser ice_switch_perm_parser;
+
+static struct
+ice_pattern_match_item ice_switch_pattern_dist_comms[] = {
+       {pattern_ethertype,
+                       ICE_SW_INSET_ETHER, ICE_INSET_NONE},
+       {pattern_ethertype_vlan,
+                       ICE_SW_INSET_MAC_VLAN, ICE_INSET_NONE},
+       {pattern_eth_ipv4,
+                       ICE_SW_INSET_MAC_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp,
+                       ICE_SW_INSET_MAC_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_tcp,
+                       ICE_SW_INSET_MAC_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv6,
+                       ICE_SW_INSET_MAC_IPV6, ICE_INSET_NONE},
+       {pattern_eth_ipv6_udp,
+                       ICE_SW_INSET_MAC_IPV6_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv6_tcp,
+                       ICE_SW_INSET_MAC_IPV6_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4,
+                       ICE_SW_INSET_DIST_VXLAN_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4_udp,
+                       ICE_SW_INSET_DIST_VXLAN_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4_tcp,
+                       ICE_SW_INSET_DIST_VXLAN_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4,
+                       ICE_SW_INSET_DIST_NVGRE_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4_udp,
+                       ICE_SW_INSET_DIST_NVGRE_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4_tcp,
+                       ICE_SW_INSET_DIST_NVGRE_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_pppoed,
+                       ICE_SW_INSET_MAC_PPPOE, ICE_INSET_NONE},
+       {pattern_eth_vlan_pppoed,
+                       ICE_SW_INSET_MAC_PPPOE, ICE_INSET_NONE},
+       {pattern_eth_pppoes,
+                       ICE_SW_INSET_MAC_PPPOE, ICE_INSET_NONE},
+       {pattern_eth_vlan_pppoes,
+                       ICE_SW_INSET_MAC_PPPOE, ICE_INSET_NONE},
+       {pattern_eth_pppoes_proto,
+                       ICE_SW_INSET_MAC_PPPOE_PROTO, ICE_INSET_NONE},
+       {pattern_eth_vlan_pppoes_proto,
+                       ICE_SW_INSET_MAC_PPPOE_PROTO, ICE_INSET_NONE},
+       {pattern_eth_ipv6_esp,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+       {pattern_eth_ipv6_ah,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+       {pattern_eth_ipv6_l2tp,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+};
+
+static struct
+ice_pattern_match_item ice_switch_pattern_dist_os[] = {
+       {pattern_ethertype,
+                       ICE_SW_INSET_ETHER, ICE_INSET_NONE},
+       {pattern_ethertype_vlan,
+                       ICE_SW_INSET_MAC_VLAN, ICE_INSET_NONE},
+       {pattern_eth_arp,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+       {pattern_eth_ipv4,
+                       ICE_SW_INSET_MAC_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp,
+                       ICE_SW_INSET_MAC_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_tcp,
+                       ICE_SW_INSET_MAC_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv6,
+                       ICE_SW_INSET_MAC_IPV6, ICE_INSET_NONE},
+       {pattern_eth_ipv6_udp,
+                       ICE_SW_INSET_MAC_IPV6_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv6_tcp,
+                       ICE_SW_INSET_MAC_IPV6_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4,
+                       ICE_SW_INSET_DIST_VXLAN_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4_udp,
+                       ICE_SW_INSET_DIST_VXLAN_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4_tcp,
+                       ICE_SW_INSET_DIST_VXLAN_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4,
+                       ICE_SW_INSET_DIST_NVGRE_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4_udp,
+                       ICE_SW_INSET_DIST_NVGRE_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4_tcp,
+                       ICE_SW_INSET_DIST_NVGRE_IPV4_TCP, ICE_INSET_NONE},
+};
+
+static struct
+ice_pattern_match_item ice_switch_pattern_perm[] = {
+       {pattern_ethertype_vlan,
+                       ICE_SW_INSET_MAC_VLAN, ICE_INSET_NONE},
+       {pattern_eth_ipv4,
+                       ICE_SW_INSET_MAC_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp,
+                       ICE_SW_INSET_MAC_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_tcp,
+                       ICE_SW_INSET_MAC_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv6,
+                       ICE_SW_INSET_MAC_IPV6, ICE_INSET_NONE},
+       {pattern_eth_ipv6_udp,
+                       ICE_SW_INSET_MAC_IPV6_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv6_tcp,
+                       ICE_SW_INSET_MAC_IPV6_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4,
+                       ICE_SW_INSET_PERM_TUNNEL_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4_udp,
+                       ICE_SW_INSET_PERM_TUNNEL_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_udp_vxlan_eth_ipv4_tcp,
+                       ICE_SW_INSET_PERM_TUNNEL_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4,
+                       ICE_SW_INSET_PERM_TUNNEL_IPV4, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4_udp,
+                       ICE_SW_INSET_PERM_TUNNEL_IPV4_UDP, ICE_INSET_NONE},
+       {pattern_eth_ipv4_nvgre_eth_ipv4_tcp,
+                       ICE_SW_INSET_PERM_TUNNEL_IPV4_TCP, ICE_INSET_NONE},
+       {pattern_eth_ipv6_esp,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+       {pattern_eth_ipv6_ah,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+       {pattern_eth_ipv6_l2tp,
+                       ICE_INSET_NONE, ICE_INSET_NONE},
+};
+
+static int
+ice_switch_create(struct ice_adapter *ad,
+               struct rte_flow *flow,
+               void *meta,
+               struct rte_flow_error *error)
+{
+       int ret = 0;
+       struct ice_pf *pf = &ad->pf;
+       struct ice_hw *hw = ICE_PF_TO_HW(pf);
+       struct ice_rule_query_data rule_added = {0};
+       struct ice_rule_query_data *filter_ptr;
+       struct ice_adv_lkup_elem *list =
+               ((struct sw_meta *)meta)->list;
+       uint16_t lkups_cnt =
+               ((struct sw_meta *)meta)->lkups_num;
+       struct ice_adv_rule_info *rule_info =
+               &((struct sw_meta *)meta)->rule_info;
+
+       if (lkups_cnt > ICE_MAX_CHAIN_WORDS) {
+               rte_flow_error_set(error, EINVAL,
+                       RTE_FLOW_ERROR_TYPE_ITEM, NULL,
+                       "item number too large for rule");
+               goto error;
+       }
+       if (!list) {
+               rte_flow_error_set(error, EINVAL,
+                       RTE_FLOW_ERROR_TYPE_ITEM, NULL,
+                       "lookup list should not be NULL");
+               goto error;
+       }
+       ret = ice_add_adv_rule(hw, list, lkups_cnt, rule_info, &rule_added);
+       if (!ret) {
+               filter_ptr = rte_zmalloc("ice_switch_filter",
+                       sizeof(struct ice_rule_query_data), 0);
+               if (!filter_ptr) {
+                       rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                                  "No memory for ice_switch_filter");
+                       goto error;
+               }
+               flow->rule = filter_ptr;
+               rte_memcpy(filter_ptr,
+                       &rule_added,
+                       sizeof(struct ice_rule_query_data));
+       } else {
+               rte_flow_error_set(error, EINVAL,
+                       RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                       "switch filter create flow fail");
+               goto error;
+       }
+
+       rte_free(list);
+       rte_free(meta);
+       return 0;
+
+error:
+       rte_free(list);
+       rte_free(meta);
+
+       return -rte_errno;
+}
+
+static int
+ice_switch_destroy(struct ice_adapter *ad,
+               struct rte_flow *flow,
+               struct rte_flow_error *error)
+{
+       struct ice_hw *hw = &ad->hw;
+       int ret;
+       struct ice_rule_query_data *filter_ptr;
+
+       filter_ptr = (struct ice_rule_query_data *)
+               flow->rule;
+
+       if (!filter_ptr) {
+               rte_flow_error_set(error, EINVAL,
+                       RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                       "no such flow"
+                       " create by switch filter");
+               return -rte_errno;
+       }
+
+       ret = ice_rem_adv_rule_by_id(hw, filter_ptr);
+       if (ret) {
+               rte_flow_error_set(error, EINVAL,
+                       RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                       "fail to destroy switch filter rule");
+               return -rte_errno;
+       }
+
+       rte_free(filter_ptr);
+       return ret;
+}
+
+static void
+ice_switch_filter_rule_free(struct rte_flow *flow)
+{
+       rte_free(flow->rule);
+}
+
+static uint64_t
+ice_switch_inset_get(const struct rte_flow_item pattern[],
+               struct rte_flow_error *error,
+               struct ice_adv_lkup_elem *list,
+               uint16_t *lkups_num,
+               enum ice_sw_tunnel_type *tun_type)
+{
+       const struct rte_flow_item *item = pattern;
+       enum rte_flow_item_type item_type;
+       const struct rte_flow_item_eth *eth_spec, *eth_mask;
+       const struct rte_flow_item_ipv4 *ipv4_spec, *ipv4_mask;
+       const struct rte_flow_item_ipv6 *ipv6_spec, *ipv6_mask;
+       const struct rte_flow_item_tcp *tcp_spec, *tcp_mask;
+       const struct rte_flow_item_udp *udp_spec, *udp_mask;
+       const struct rte_flow_item_sctp *sctp_spec, *sctp_mask;
+       const struct rte_flow_item_nvgre *nvgre_spec, *nvgre_mask;
+       const struct rte_flow_item_vxlan *vxlan_spec, *vxlan_mask;
+       const struct rte_flow_item_vlan *vlan_spec, *vlan_mask;
+       const struct rte_flow_item_pppoe *pppoe_spec, *pppoe_mask;
+       const struct rte_flow_item_pppoe_proto_id *pppoe_proto_spec,
+                               *pppoe_proto_mask;
+       const struct rte_flow_item_esp *esp_spec, *esp_mask;
+       const struct rte_flow_item_ah *ah_spec, *ah_mask;
+       const struct rte_flow_item_l2tpv3oip *l2tp_spec, *l2tp_mask;
+       uint64_t input_set = ICE_INSET_NONE;
+       uint16_t j, t = 0;
+       uint16_t tunnel_valid = 0;
+       uint16_t pppoe_valid = 0;
+       uint16_t ipv6_valiad = 0;
+
+
+       for (item = pattern; item->type !=
+                       RTE_FLOW_ITEM_TYPE_END; item++) {
+               if (item->last) {
+                       rte_flow_error_set(error, EINVAL,
+                                       RTE_FLOW_ERROR_TYPE_ITEM,
+                                       item,
+                                       "Not support range");
+                       return 0;
+               }
+               item_type = item->type;
+
+               switch (item_type) {
+               case RTE_FLOW_ITEM_TYPE_ETH:
+                       eth_spec = item->spec;
+                       eth_mask = item->mask;
+                       if (eth_spec && eth_mask) {
+                               const uint8_t *a = eth_mask->src.addr_bytes;
+                               const uint8_t *b = eth_mask->dst.addr_bytes;
+                               for (j = 0; j < RTE_ETHER_ADDR_LEN; j++) {
+                                       if (a[j] && tunnel_valid) {
+                                               input_set |=
+                                                       ICE_INSET_TUN_SMAC;
+                                               break;
+                                       } else if (a[j]) {
+                                               input_set |=
+                                                       ICE_INSET_SMAC;
+                                               break;
+                                       }
+                               }
+                               for (j = 0; j < RTE_ETHER_ADDR_LEN; j++) {
+                                       if (b[j] && tunnel_valid) {
+                                               input_set |=
+                                                       ICE_INSET_TUN_DMAC;
+                                               break;
+                                       } else if (b[j]) {
+                                               input_set |=
+                                                       ICE_INSET_DMAC;
+                                               break;
+                                       }
+                               }
+                               if (eth_mask->type)
+                                       input_set |= ICE_INSET_ETHERTYPE;
+                               list[t].type = (tunnel_valid  == 0) ?
+                                       ICE_MAC_OFOS : ICE_MAC_IL;
+                               struct ice_ether_hdr *h;
+                               struct ice_ether_hdr *m;
+                               uint16_t i = 0;
+                               h = &list[t].h_u.eth_hdr;
+                               m = &list[t].m_u.eth_hdr;
+                               for (j = 0; j < RTE_ETHER_ADDR_LEN; j++) {
+                                       if (eth_mask->src.addr_bytes[j]) {
+                                               h->src_addr[j] =
+                                               eth_spec->src.addr_bytes[j];
+                                               m->src_addr[j] =
+                                               eth_mask->src.addr_bytes[j];
+                                               i = 1;
+                                       }
+                                       if (eth_mask->dst.addr_bytes[j]) {
+                                               h->dst_addr[j] =
+                                               eth_spec->dst.addr_bytes[j];
+                                               m->dst_addr[j] =
+                                               eth_mask->dst.addr_bytes[j];
+                                               i = 1;
+                                       }
+                               }
+                               if (i)
+                                       t++;
+                               if (eth_mask->type) {
+                                       list[t].type = ICE_ETYPE_OL;
+                                       list[t].h_u.ethertype.ethtype_id =
+                                               eth_spec->type;
+                                       list[t].m_u.ethertype.ethtype_id =
+                                               eth_mask->type;
+                                       t++;
+                               }
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_IPV4:
+                       ipv4_spec = item->spec;
+                       ipv4_mask = item->mask;
+                       if (ipv4_spec && ipv4_mask) {
+                               /* Check IPv4 mask and update input set */
+                               if (ipv4_mask->hdr.version_ihl ||
+                                       ipv4_mask->hdr.total_length ||
+                                       ipv4_mask->hdr.packet_id ||
+                                       ipv4_mask->hdr.hdr_checksum) {
+                                       rte_flow_error_set(error, EINVAL,
+                                                  RTE_FLOW_ERROR_TYPE_ITEM,
+                                                  item,
+                                                  "Invalid IPv4 mask.");
+                                       return 0;
+                               }
+
+                               if (tunnel_valid) {
+                                       if (ipv4_mask->hdr.type_of_service)
+                                               input_set |=
+                                                       ICE_INSET_TUN_IPV4_TOS;
+                                       if (ipv4_mask->hdr.src_addr)
+                                               input_set |=
+                                                       ICE_INSET_TUN_IPV4_SRC;
+                                       if (ipv4_mask->hdr.dst_addr)
+                                               input_set |=
+                                                       ICE_INSET_TUN_IPV4_DST;
+                                       if (ipv4_mask->hdr.time_to_live)
+                                               input_set |=
+                                                       ICE_INSET_TUN_IPV4_TTL;
+                                       if (ipv4_mask->hdr.next_proto_id)
+                                               input_set |=
+                                               ICE_INSET_TUN_IPV4_PROTO;
+                               } else {
+                                       if (ipv4_mask->hdr.src_addr)
+                                               input_set |= ICE_INSET_IPV4_SRC;
+                                       if (ipv4_mask->hdr.dst_addr)
+                                               input_set |= ICE_INSET_IPV4_DST;
+                                       if (ipv4_mask->hdr.time_to_live)
+                                               input_set |= ICE_INSET_IPV4_TTL;
+                                       if (ipv4_mask->hdr.next_proto_id)
+                                               input_set |=
+                                               ICE_INSET_IPV4_PROTO;
+                                       if (ipv4_mask->hdr.type_of_service)
+                                               input_set |=
+                                                       ICE_INSET_IPV4_TOS;
+                               }
+                               list[t].type = (tunnel_valid  == 0) ?
+                                       ICE_IPV4_OFOS : ICE_IPV4_IL;
+                               if (ipv4_mask->hdr.src_addr) {
+                                       list[t].h_u.ipv4_hdr.src_addr =
+                                               ipv4_spec->hdr.src_addr;
+                                       list[t].m_u.ipv4_hdr.src_addr =
+                                               ipv4_mask->hdr.src_addr;
+                               }
+                               if (ipv4_mask->hdr.dst_addr) {
+                                       list[t].h_u.ipv4_hdr.dst_addr =
+                                               ipv4_spec->hdr.dst_addr;
+                                       list[t].m_u.ipv4_hdr.dst_addr =
+                                               ipv4_mask->hdr.dst_addr;
+                               }
+                               if (ipv4_mask->hdr.time_to_live) {
+                                       list[t].h_u.ipv4_hdr.time_to_live =
+                                               ipv4_spec->hdr.time_to_live;
+                                       list[t].m_u.ipv4_hdr.time_to_live =
+                                               ipv4_mask->hdr.time_to_live;
+                               }
+                               if (ipv4_mask->hdr.next_proto_id) {
+                                       list[t].h_u.ipv4_hdr.protocol =
+                                               ipv4_spec->hdr.next_proto_id;
+                                       list[t].m_u.ipv4_hdr.protocol =
+                                               ipv4_mask->hdr.next_proto_id;
+                               }
+                               if (ipv4_mask->hdr.type_of_service) {
+                                       list[t].h_u.ipv4_hdr.tos =
+                                               ipv4_spec->hdr.type_of_service;
+                                       list[t].m_u.ipv4_hdr.tos =
+                                               ipv4_mask->hdr.type_of_service;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_IPV6:
+                       ipv6_spec = item->spec;
+                       ipv6_mask = item->mask;
+                       ipv6_valiad = 1;
+                       if (ipv6_spec && ipv6_mask) {
+                               if (ipv6_mask->hdr.payload_len) {
+                                       rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid IPv6 mask");
+                                       return 0;
+                               }
+
+                               for (j = 0; j < ICE_IPV6_ADDR_LENGTH; j++) {
+                                       if (ipv6_mask->hdr.src_addr[j] &&
+                                               tunnel_valid) {
+                                               input_set |=
+                                               ICE_INSET_TUN_IPV6_SRC;
+                                               break;
+                                       } else if (ipv6_mask->hdr.src_addr[j]) {
+                                               input_set |= ICE_INSET_IPV6_SRC;
+                                               break;
+                                       }
+                               }
+                               for (j = 0; j < ICE_IPV6_ADDR_LENGTH; j++) {
+                                       if (ipv6_mask->hdr.dst_addr[j] &&
+                                               tunnel_valid) {
+                                               input_set |=
+                                               ICE_INSET_TUN_IPV6_DST;
+                                               break;
+                                       } else if (ipv6_mask->hdr.dst_addr[j]) {
+                                               input_set |= ICE_INSET_IPV6_DST;
+                                               break;
+                                       }
+                               }
+                               if (ipv6_mask->hdr.proto &&
+                                       tunnel_valid)
+                                       input_set |=
+                                               ICE_INSET_TUN_IPV6_NEXT_HDR;
+                               else if (ipv6_mask->hdr.proto)
+                                       input_set |=
+                                               ICE_INSET_IPV6_NEXT_HDR;
+                               if (ipv6_mask->hdr.hop_limits &&
+                                       tunnel_valid)
+                                       input_set |=
+                                               ICE_INSET_TUN_IPV6_HOP_LIMIT;
+                               else if (ipv6_mask->hdr.hop_limits)
+                                       input_set |=
+                                               ICE_INSET_IPV6_HOP_LIMIT;
+                               if ((ipv6_mask->hdr.vtc_flow &
+                                               rte_cpu_to_be_32
+                                               (RTE_IPV6_HDR_TC_MASK)) &&
+                                       tunnel_valid)
+                                       input_set |=
+                                                       ICE_INSET_TUN_IPV6_TC;
+                               else if (ipv6_mask->hdr.vtc_flow &
+                                               rte_cpu_to_be_32
+                                               (RTE_IPV6_HDR_TC_MASK))
+                                       input_set |= ICE_INSET_IPV6_TC;
+
+                               list[t].type = (tunnel_valid  == 0) ?
+                                       ICE_IPV6_OFOS : ICE_IPV6_IL;
+                               struct ice_ipv6_hdr *f;
+                               struct ice_ipv6_hdr *s;
+                               f = &list[t].h_u.ipv6_hdr;
+                               s = &list[t].m_u.ipv6_hdr;
+                               for (j = 0; j < ICE_IPV6_ADDR_LENGTH; j++) {
+                                       if (ipv6_mask->hdr.src_addr[j]) {
+                                               f->src_addr[j] =
+                                               ipv6_spec->hdr.src_addr[j];
+                                               s->src_addr[j] =
+                                               ipv6_mask->hdr.src_addr[j];
+                                       }
+                                       if (ipv6_mask->hdr.dst_addr[j]) {
+                                               f->dst_addr[j] =
+                                               ipv6_spec->hdr.dst_addr[j];
+                                               s->dst_addr[j] =
+                                               ipv6_mask->hdr.dst_addr[j];
+                                       }
+                               }
+                               if (ipv6_mask->hdr.proto) {
+                                       f->next_hdr =
+                                               ipv6_spec->hdr.proto;
+                                       s->next_hdr =
+                                               ipv6_mask->hdr.proto;
+                               }
+                               if (ipv6_mask->hdr.hop_limits) {
+                                       f->hop_limit =
+                                               ipv6_spec->hdr.hop_limits;
+                                       s->hop_limit =
+                                               ipv6_mask->hdr.hop_limits;
+                               }
+                               if (ipv6_mask->hdr.vtc_flow &
+                                               rte_cpu_to_be_32
+                                               (RTE_IPV6_HDR_TC_MASK)) {
+                                       struct ice_le_ver_tc_flow vtf;
+                                       vtf.u.fld.version = 0;
+                                       vtf.u.fld.flow_label = 0;
+                                       vtf.u.fld.tc = (rte_be_to_cpu_32
+                                               (ipv6_spec->hdr.vtc_flow) &
+                                                       RTE_IPV6_HDR_TC_MASK) >>
+                                                       RTE_IPV6_HDR_TC_SHIFT;
+                                       f->be_ver_tc_flow = CPU_TO_BE32(vtf.u.val);
+                                       vtf.u.fld.tc = (rte_be_to_cpu_32
+                                               (ipv6_mask->hdr.vtc_flow) &
+                                                       RTE_IPV6_HDR_TC_MASK) >>
+                                                       RTE_IPV6_HDR_TC_SHIFT;
+                                       s->be_ver_tc_flow = CPU_TO_BE32(vtf.u.val);
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_UDP:
+                       udp_spec = item->spec;
+                       udp_mask = item->mask;
+                       if (udp_spec && udp_mask) {
+                               /* Check UDP mask and update input set*/
+                               if (udp_mask->hdr.dgram_len ||
+                                   udp_mask->hdr.dgram_cksum) {
+                                       rte_flow_error_set(error, EINVAL,
+                                                  RTE_FLOW_ERROR_TYPE_ITEM,
+                                                  item,
+                                                  "Invalid UDP mask");
+                                       return 0;
+                               }
+
+                               if (tunnel_valid) {
+                                       if (udp_mask->hdr.src_port)
+                                               input_set |=
+                                               ICE_INSET_TUN_UDP_SRC_PORT;
+                                       if (udp_mask->hdr.dst_port)
+                                               input_set |=
+                                               ICE_INSET_TUN_UDP_DST_PORT;
+                               } else {
+                                       if (udp_mask->hdr.src_port)
+                                               input_set |=
+                                               ICE_INSET_UDP_SRC_PORT;
+                                       if (udp_mask->hdr.dst_port)
+                                               input_set |=
+                                               ICE_INSET_UDP_DST_PORT;
+                               }
+                               if (*tun_type == ICE_SW_TUN_VXLAN &&
+                                               tunnel_valid == 0)
+                                       list[t].type = ICE_UDP_OF;
+                               else
+                                       list[t].type = ICE_UDP_ILOS;
+                               if (udp_mask->hdr.src_port) {
+                                       list[t].h_u.l4_hdr.src_port =
+                                               udp_spec->hdr.src_port;
+                                       list[t].m_u.l4_hdr.src_port =
+                                               udp_mask->hdr.src_port;
+                               }
+                               if (udp_mask->hdr.dst_port) {
+                                       list[t].h_u.l4_hdr.dst_port =
+                                               udp_spec->hdr.dst_port;
+                                       list[t].m_u.l4_hdr.dst_port =
+                                               udp_mask->hdr.dst_port;
+                               }
+                                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_TCP:
+                       tcp_spec = item->spec;
+                       tcp_mask = item->mask;
+                       if (tcp_spec && tcp_mask) {
+                               /* Check TCP mask and update input set */
+                               if (tcp_mask->hdr.sent_seq ||
+                                       tcp_mask->hdr.recv_ack ||
+                                       tcp_mask->hdr.data_off ||
+                                       tcp_mask->hdr.tcp_flags ||
+                                       tcp_mask->hdr.rx_win ||
+                                       tcp_mask->hdr.cksum ||
+                                       tcp_mask->hdr.tcp_urp) {
+                                       rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid TCP mask");
+                                       return 0;
+                               }
+
+                               if (tunnel_valid) {
+                                       if (tcp_mask->hdr.src_port)
+                                               input_set |=
+                                               ICE_INSET_TUN_TCP_SRC_PORT;
+                                       if (tcp_mask->hdr.dst_port)
+                                               input_set |=
+                                               ICE_INSET_TUN_TCP_DST_PORT;
+                               } else {
+                                       if (tcp_mask->hdr.src_port)
+                                               input_set |=
+                                               ICE_INSET_TCP_SRC_PORT;
+                                       if (tcp_mask->hdr.dst_port)
+                                               input_set |=
+                                               ICE_INSET_TCP_DST_PORT;
+                               }
+                               list[t].type = ICE_TCP_IL;
+                               if (tcp_mask->hdr.src_port) {
+                                       list[t].h_u.l4_hdr.src_port =
+                                               tcp_spec->hdr.src_port;
+                                       list[t].m_u.l4_hdr.src_port =
+                                               tcp_mask->hdr.src_port;
+                               }
+                               if (tcp_mask->hdr.dst_port) {
+                                       list[t].h_u.l4_hdr.dst_port =
+                                               tcp_spec->hdr.dst_port;
+                                       list[t].m_u.l4_hdr.dst_port =
+                                               tcp_mask->hdr.dst_port;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_SCTP:
+                       sctp_spec = item->spec;
+                       sctp_mask = item->mask;
+                       if (sctp_spec && sctp_mask) {
+                               /* Check SCTP mask and update input set */
+                               if (sctp_mask->hdr.cksum) {
+                                       rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid SCTP mask");
+                                       return 0;
+                               }
+
+                               if (tunnel_valid) {
+                                       if (sctp_mask->hdr.src_port)
+                                               input_set |=
+                                               ICE_INSET_TUN_SCTP_SRC_PORT;
+                                       if (sctp_mask->hdr.dst_port)
+                                               input_set |=
+                                               ICE_INSET_TUN_SCTP_DST_PORT;
+                               } else {
+                                       if (sctp_mask->hdr.src_port)
+                                               input_set |=
+                                               ICE_INSET_SCTP_SRC_PORT;
+                                       if (sctp_mask->hdr.dst_port)
+                                               input_set |=
+                                               ICE_INSET_SCTP_DST_PORT;
+                               }
+                               list[t].type = ICE_SCTP_IL;
+                               if (sctp_mask->hdr.src_port) {
+                                       list[t].h_u.sctp_hdr.src_port =
+                                               sctp_spec->hdr.src_port;
+                                       list[t].m_u.sctp_hdr.src_port =
+                                               sctp_mask->hdr.src_port;
+                               }
+                               if (sctp_mask->hdr.dst_port) {
+                                       list[t].h_u.sctp_hdr.dst_port =
+                                               sctp_spec->hdr.dst_port;
+                                       list[t].m_u.sctp_hdr.dst_port =
+                                               sctp_mask->hdr.dst_port;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_VXLAN:
+                       vxlan_spec = item->spec;
+                       vxlan_mask = item->mask;
+                       /* Check if VXLAN item is used to describe protocol.
+                        * If yes, both spec and mask should be NULL.
+                        * If no, both spec and mask shouldn't be NULL.
+                        */
+                       if ((!vxlan_spec && vxlan_mask) ||
+                           (vxlan_spec && !vxlan_mask)) {
+                               rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid VXLAN item");
+                               return 0;
+                       }
+
+                       tunnel_valid = 1;
+                       if (vxlan_spec && vxlan_mask) {
+                               list[t].type = ICE_VXLAN;
+                               if (vxlan_mask->vni[0] ||
+                                       vxlan_mask->vni[1] ||
+                                       vxlan_mask->vni[2]) {
+                                       list[t].h_u.tnl_hdr.vni =
+                                               (vxlan_spec->vni[2] << 16) |
+                                               (vxlan_spec->vni[1] << 8) |
+                                               vxlan_spec->vni[0];
+                                       list[t].m_u.tnl_hdr.vni =
+                                               (vxlan_mask->vni[2] << 16) |
+                                               (vxlan_mask->vni[1] << 8) |
+                                               vxlan_mask->vni[0];
+                                       input_set |=
+                                               ICE_INSET_TUN_VXLAN_VNI;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_NVGRE:
+                       nvgre_spec = item->spec;
+                       nvgre_mask = item->mask;
+                       /* Check if NVGRE item is used to describe protocol.
+                        * If yes, both spec and mask should be NULL.
+                        * If no, both spec and mask shouldn't be NULL.
+                        */
+                       if ((!nvgre_spec && nvgre_mask) ||
+                           (nvgre_spec && !nvgre_mask)) {
+                               rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid NVGRE item");
+                               return 0;
+                       }
+                       tunnel_valid = 1;
+                       if (nvgre_spec && nvgre_mask) {
+                               list[t].type = ICE_NVGRE;
+                               if (nvgre_mask->tni[0] ||
+                                       nvgre_mask->tni[1] ||
+                                       nvgre_mask->tni[2]) {
+                                       list[t].h_u.nvgre_hdr.tni_flow =
+                                               (nvgre_spec->tni[2] << 16) |
+                                               (nvgre_spec->tni[1] << 8) |
+                                               nvgre_spec->tni[0];
+                                       list[t].m_u.nvgre_hdr.tni_flow =
+                                               (nvgre_mask->tni[2] << 16) |
+                                               (nvgre_mask->tni[1] << 8) |
+                                               nvgre_mask->tni[0];
+                                       input_set |=
+                                               ICE_INSET_TUN_NVGRE_TNI;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_VLAN:
+                       vlan_spec = item->spec;
+                       vlan_mask = item->mask;
+                       /* Check if VLAN item is used to describe protocol.
+                        * If yes, both spec and mask should be NULL.
+                        * If no, both spec and mask shouldn't be NULL.
+                        */
+                       if ((!vlan_spec && vlan_mask) ||
+                           (vlan_spec && !vlan_mask)) {
+                               rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid VLAN item");
+                               return 0;
+                       }
+                       if (vlan_spec && vlan_mask) {
+                               list[t].type = ICE_VLAN_OFOS;
+                               if (vlan_mask->tci) {
+                                       list[t].h_u.vlan_hdr.vlan =
+                                               vlan_spec->tci;
+                                       list[t].m_u.vlan_hdr.vlan =
+                                               vlan_mask->tci;
+                                       input_set |= ICE_INSET_VLAN_OUTER;
+                               }
+                               if (vlan_mask->inner_type) {
+                                       list[t].h_u.vlan_hdr.type =
+                                               vlan_spec->inner_type;
+                                       list[t].m_u.vlan_hdr.type =
+                                               vlan_mask->inner_type;
+                                       input_set |= ICE_INSET_VLAN_OUTER;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_PPPOED:
+               case RTE_FLOW_ITEM_TYPE_PPPOES:
+                       pppoe_spec = item->spec;
+                       pppoe_mask = item->mask;
+                       /* Check if PPPoE item is used to describe protocol.
+                        * If yes, both spec and mask should be NULL.
+                        * If no, both spec and mask shouldn't be NULL.
+                        */
+                       if ((!pppoe_spec && pppoe_mask) ||
+                               (pppoe_spec && !pppoe_mask)) {
+                               rte_flow_error_set(error, EINVAL,
+                                       RTE_FLOW_ERROR_TYPE_ITEM,
+                                       item,
+                                       "Invalid pppoe item");
+                               return 0;
+                       }
+                       if (pppoe_spec && pppoe_mask) {
+                               /* Check pppoe mask and update input set */
+                               if (pppoe_mask->length ||
+                                       pppoe_mask->code ||
+                                       pppoe_mask->version_type) {
+                                       rte_flow_error_set(error, EINVAL,
+                                               RTE_FLOW_ERROR_TYPE_ITEM,
+                                               item,
+                                               "Invalid pppoe mask");
+                                       return 0;
+                               }
+                               list[t].type = ICE_PPPOE;
+                               if (pppoe_mask->session_id) {
+                                       list[t].h_u.pppoe_hdr.session_id =
+                                               pppoe_spec->session_id;
+                                       list[t].m_u.pppoe_hdr.session_id =
+                                               pppoe_mask->session_id;
+                                       input_set |= ICE_INSET_PPPOE_SESSION;
+                               }
+                               t++;
+                               pppoe_valid = 1;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_PPPOE_PROTO_ID:
+                       pppoe_proto_spec = item->spec;
+                       pppoe_proto_mask = item->mask;
+                       /* Check if PPPoE optional proto_id item
+                        * is used to describe protocol.
+                        * If yes, both spec and mask should be NULL.
+                        * If no, both spec and mask shouldn't be NULL.
+                        */
+                       if ((!pppoe_proto_spec && pppoe_proto_mask) ||
+                               (pppoe_proto_spec && !pppoe_proto_mask)) {
+                               rte_flow_error_set(error, EINVAL,
+                                       RTE_FLOW_ERROR_TYPE_ITEM,
+                                       item,
+                                       "Invalid pppoe proto item");
+                               return 0;
+                       }
+                       if (pppoe_proto_spec && pppoe_proto_mask) {
+                               if (pppoe_valid)
+                                       t--;
+                               list[t].type = ICE_PPPOE;
+                               if (pppoe_proto_mask->proto_id) {
+                                       list[t].h_u.pppoe_hdr.ppp_prot_id =
+                                               pppoe_proto_spec->proto_id;
+                                       list[t].m_u.pppoe_hdr.ppp_prot_id =
+                                               pppoe_proto_mask->proto_id;
+                                       input_set |= ICE_INSET_PPPOE_PROTO;
+                               }
+                               t++;
+                       }
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_ESP:
+                       esp_spec = item->spec;
+                       esp_mask = item->mask;
+                       if (esp_spec || esp_mask) {
+                               rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid esp item");
+                               return -ENOTSUP;
+                       }
+                       if (ipv6_valiad)
+                               *tun_type = ICE_SW_TUN_PROFID_IPV6_ESP;
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_AH:
+                       ah_spec = item->spec;
+                       ah_mask = item->mask;
+                       if (ah_spec || ah_mask) {
+                               rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid ah item");
+                               return -ENOTSUP;
+                       }
+                       if (ipv6_valiad)
+                               *tun_type = ICE_SW_TUN_PROFID_IPV6_AH;
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_L2TPV3OIP:
+                       l2tp_spec = item->spec;
+                       l2tp_mask = item->mask;
+                       if (l2tp_spec || l2tp_mask) {
+                               rte_flow_error_set(error, EINVAL,
+                                          RTE_FLOW_ERROR_TYPE_ITEM,
+                                          item,
+                                          "Invalid l2tp item");
+                               return -ENOTSUP;
+                       }
+                       if (ipv6_valiad)
+                               *tun_type = ICE_SW_TUN_PROFID_MAC_IPV6_L2TPV3;
+                       break;
+
+               case RTE_FLOW_ITEM_TYPE_VOID:
+                       break;
+
+               default:
+                       rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_ITEM, pattern,
+                                  "Invalid pattern item.");
+                       goto out;
+               }
+       }
+
+       *lkups_num = t;
+
+       return input_set;
+out:
+       return 0;
+}
+
+static int
+ice_switch_parse_dcf_action(const struct rte_flow_action *actions,
+                           struct rte_flow_error *error,
+                           struct ice_adv_rule_info *rule_info)
+{
+       const struct rte_flow_action_vf *act_vf;
+       const struct rte_flow_action *action;
+       enum rte_flow_action_type action_type;
+
+       for (action = actions; action->type !=
+                               RTE_FLOW_ACTION_TYPE_END; action++) {
+               action_type = action->type;
+               switch (action_type) {
+               case RTE_FLOW_ACTION_TYPE_VF:
+                       rule_info->sw_act.fltr_act = ICE_FWD_TO_VSI;
+                       act_vf = action->conf;
+                       rule_info->sw_act.vsi_handle = act_vf->id;
+                       break;
+               default:
+                       rte_flow_error_set(error,
+                                          EINVAL, RTE_FLOW_ERROR_TYPE_ACTION,
+                                          actions,
+                                          "Invalid action type or queue number");
+                       return -rte_errno;
+               }
+       }
+
+       rule_info->sw_act.src = rule_info->sw_act.vsi_handle;
+       rule_info->rx = 1;
+       rule_info->priority = 5;
+
+       return 0;
+}
+
+static int
+ice_switch_parse_action(struct ice_pf *pf,
+               const struct rte_flow_action *actions,
+               struct rte_flow_error *error,
+               struct ice_adv_rule_info *rule_info)
+{
+       struct ice_vsi *vsi = pf->main_vsi;
+       struct rte_eth_dev *dev = pf->adapter->eth_dev;
+       const struct rte_flow_action_queue *act_q;
+       const struct rte_flow_action_rss *act_qgrop;
+       uint16_t base_queue, i;
+       const struct rte_flow_action *action;
+       enum rte_flow_action_type action_type;
+       uint16_t valid_qgrop_number[MAX_QGRP_NUM_TYPE] = {
+                2, 4, 8, 16, 32, 64, 128};
+
+       base_queue = pf->base_queue + vsi->base_queue;
+       for (action = actions; action->type !=
+                       RTE_FLOW_ACTION_TYPE_END; action++) {
+               action_type = action->type;
+               switch (action_type) {
+               case RTE_FLOW_ACTION_TYPE_RSS:
+                       act_qgrop = action->conf;
+                       rule_info->sw_act.fltr_act =
+                               ICE_FWD_TO_QGRP;
+                       rule_info->sw_act.fwd_id.q_id =
+                               base_queue + act_qgrop->queue[0];
+                       for (i = 0; i < MAX_QGRP_NUM_TYPE; i++) {
+                               if (act_qgrop->queue_num ==
+                                       valid_qgrop_number[i])
+                                       break;
+                       }
+                       if (i == MAX_QGRP_NUM_TYPE)
+                               goto error;
+                       if ((act_qgrop->queue[0] +
+                               act_qgrop->queue_num) >
+                               dev->data->nb_rx_queues)
+                               goto error;
+                       for (i = 0; i < act_qgrop->queue_num - 1; i++)
+                               if (act_qgrop->queue[i + 1] !=
+                                       act_qgrop->queue[i] + 1)
+                                       goto error;
+                       rule_info->sw_act.qgrp_size =
+                               act_qgrop->queue_num;
+                       break;
+               case RTE_FLOW_ACTION_TYPE_QUEUE:
+                       act_q = action->conf;
+                       if (act_q->index >= dev->data->nb_rx_queues)
+                               goto error;
+                       rule_info->sw_act.fltr_act =
+                               ICE_FWD_TO_Q;
+                       rule_info->sw_act.fwd_id.q_id =
+                               base_queue + act_q->index;
+                       break;
+
+               case RTE_FLOW_ACTION_TYPE_DROP:
+                       rule_info->sw_act.fltr_act =
+                               ICE_DROP_PACKET;
+                       break;
+
+               case RTE_FLOW_ACTION_TYPE_VOID:
+                       break;
+
+               default:
+                       goto error;
+               }
+       }
+
+       rule_info->sw_act.vsi_handle = vsi->idx;
+       rule_info->rx = 1;
+       rule_info->sw_act.src = vsi->idx;
+       rule_info->priority = 5;
+
+       return 0;
+
+error:
+       rte_flow_error_set(error,
+               EINVAL, RTE_FLOW_ERROR_TYPE_ACTION,
+               actions,
+               "Invalid action type or queue number");
+       return -rte_errno;
+}
+
+static int
+ice_switch_check_action(const struct rte_flow_action *actions,
+                           struct rte_flow_error *error)
+{
+       const struct rte_flow_action *action;
+       enum rte_flow_action_type action_type;
+       uint16_t actions_num = 0;
+
+       for (action = actions; action->type !=
+                               RTE_FLOW_ACTION_TYPE_END; action++) {
+               action_type = action->type;
+               switch (action_type) {
+               case RTE_FLOW_ACTION_TYPE_VF:
+               case RTE_FLOW_ACTION_TYPE_RSS:
+               case RTE_FLOW_ACTION_TYPE_QUEUE:
+               case RTE_FLOW_ACTION_TYPE_DROP:
+                       actions_num++;
+                       break;
+               case RTE_FLOW_ACTION_TYPE_VOID:
+                       continue;
+               default:
+                       rte_flow_error_set(error,
+                                          EINVAL, RTE_FLOW_ERROR_TYPE_ACTION,
+                                          actions,
+                                          "Invalid action type");
+                       return -rte_errno;
+               }
+       }
+
+       if (actions_num > 1) {
+               rte_flow_error_set(error,
+                                  EINVAL, RTE_FLOW_ERROR_TYPE_ACTION,
+                                  actions,
+                                  "Invalid action number");
+               return -rte_errno;
+       }
+
+       return 0;
+}
+
+static bool
+ice_is_profile_rule(enum ice_sw_tunnel_type tun_type)
+{
+       switch (tun_type) {
+       case ICE_SW_TUN_PROFID_IPV6_ESP:
+       case ICE_SW_TUN_PROFID_IPV6_AH:
+       case ICE_SW_TUN_PROFID_MAC_IPV6_L2TPV3:
+               return true;
+       default:
+               break;
+       }
+
+       return false;
+}
+
+static int
+ice_switch_parse_pattern_action(struct ice_adapter *ad,
+               struct ice_pattern_match_item *array,
+               uint32_t array_len,
+               const struct rte_flow_item pattern[],
+               const struct rte_flow_action actions[],
+               void **meta,
+               struct rte_flow_error *error)
+{
+       struct ice_pf *pf = &ad->pf;
+       uint64_t inputset = 0;
+       int ret = 0;
+       struct sw_meta *sw_meta_ptr = NULL;
+       struct ice_adv_rule_info rule_info;
+       struct ice_adv_lkup_elem *list = NULL;
+       uint16_t lkups_num = 0;
+       const struct rte_flow_item *item = pattern;
+       uint16_t item_num = 0;
+       enum ice_sw_tunnel_type tun_type =
+               ICE_SW_TUN_AND_NON_TUN;
+       struct ice_pattern_match_item *pattern_match_item = NULL;
+
+       for (; item->type != RTE_FLOW_ITEM_TYPE_END; item++) {
+               item_num++;
+               if (item->type == RTE_FLOW_ITEM_TYPE_VXLAN)
+                       tun_type = ICE_SW_TUN_VXLAN;
+               if (item->type == RTE_FLOW_ITEM_TYPE_NVGRE)
+                       tun_type = ICE_SW_TUN_NVGRE;
+               if (item->type == RTE_FLOW_ITEM_TYPE_PPPOED ||
+                               item->type == RTE_FLOW_ITEM_TYPE_PPPOES)
+                       tun_type = ICE_SW_TUN_PPPOE;
+               if (item->type == RTE_FLOW_ITEM_TYPE_ETH) {
+                       const struct rte_flow_item_eth *eth_mask;
+                       if (item->mask)
+                               eth_mask = item->mask;
+                       else
+                               continue;
+                       if (eth_mask->type == UINT16_MAX)
+                               tun_type = ICE_SW_TUN_AND_NON_TUN;
+               }
+               /* reserve one more memory slot for ETH which may
+                * consume 2 lookup items.
+                */
+               if (item->type == RTE_FLOW_ITEM_TYPE_ETH)
+                       item_num++;
+       }
+
+       list = rte_zmalloc(NULL, item_num * sizeof(*list), 0);
+       if (!list) {
+               rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                                  "No memory for PMD internal items");
+               return -rte_errno;
+       }
+
+       sw_meta_ptr =
+               rte_zmalloc(NULL, sizeof(*sw_meta_ptr), 0);
+       if (!sw_meta_ptr) {
+               rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                                  "No memory for sw_pattern_meta_ptr");
+               goto error;
+       }
+
+       pattern_match_item =
+               ice_search_pattern_match_item(pattern, array, array_len, error);
+       if (!pattern_match_item) {
+               rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                                  "Invalid input pattern");
+               goto error;
+       }
+
+       inputset = ice_switch_inset_get
+               (pattern, error, list, &lkups_num, &tun_type);
+       if ((!inputset && !ice_is_profile_rule(tun_type)) ||
+               (inputset & ~pattern_match_item->input_set_mask)) {
+               rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_ITEM_SPEC,
+                                  pattern,
+                                  "Invalid input set");
+               goto error;
+       }
+
+       rule_info.tun_type = tun_type;
+
+       ret = ice_switch_check_action(actions, error);
+       if (ret) {
+               rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                                  "Invalid input action number");
+               goto error;
+       }
+
+       if (ad->hw.dcf_enabled)
+               ret = ice_switch_parse_dcf_action(actions, error, &rule_info);
+       else
+               ret = ice_switch_parse_action(pf, actions, error, &rule_info);
+
+       if (ret) {
+               rte_flow_error_set(error, EINVAL,
+                                  RTE_FLOW_ERROR_TYPE_HANDLE, NULL,
+                                  "Invalid input action");
+               goto error;
+       }
+
+       if (meta) {
+               *meta = sw_meta_ptr;
+               ((struct sw_meta *)*meta)->list = list;
+               ((struct sw_meta *)*meta)->lkups_num = lkups_num;
+               ((struct sw_meta *)*meta)->rule_info = rule_info;
+       } else {
+               rte_free(list);
+               rte_free(sw_meta_ptr);
+       }
+
+       rte_free(pattern_match_item);
+
+       return 0;
+
+error:
+       rte_free(list);
+       rte_free(sw_meta_ptr);
+       rte_free(pattern_match_item);
+
+       return -rte_errno;
+}
+
+static int
+ice_switch_query(struct ice_adapter *ad __rte_unused,
+               struct rte_flow *flow __rte_unused,
+               struct rte_flow_query_count *count __rte_unused,
+               struct rte_flow_error *error)
+{
+       rte_flow_error_set(error, EINVAL,
+               RTE_FLOW_ERROR_TYPE_HANDLE,
+               NULL,
+               "count action not supported by switch filter");
+
+       return -rte_errno;
+}
+
+static int
+ice_switch_init(struct ice_adapter *ad)
+{
+       int ret = 0;
+       struct ice_flow_parser *dist_parser;
+       struct ice_flow_parser *perm_parser = &ice_switch_perm_parser;
+
+       if (ad->active_pkg_type == ICE_PKG_TYPE_COMMS)
+               dist_parser = &ice_switch_dist_parser_comms;
+       else if (ad->active_pkg_type == ICE_PKG_TYPE_OS_DEFAULT)
+               dist_parser = &ice_switch_dist_parser_os;
+       else
+               return -EINVAL;
+
+       if (ad->devargs.pipe_mode_support)
+               ret = ice_register_parser(perm_parser, ad);
+       else
+               ret = ice_register_parser(dist_parser, ad);
+       return ret;
+}
+
+static void
+ice_switch_uninit(struct ice_adapter *ad)
+{
+       struct ice_flow_parser *dist_parser;
+       struct ice_flow_parser *perm_parser = &ice_switch_perm_parser;
+
+       if (ad->active_pkg_type == ICE_PKG_TYPE_COMMS)
+               dist_parser = &ice_switch_dist_parser_comms;
+       else
+               dist_parser = &ice_switch_dist_parser_os;
+
+       if (ad->devargs.pipe_mode_support)
+               ice_unregister_parser(perm_parser, ad);
+       else
+               ice_unregister_parser(dist_parser, ad);
+}
+
+static struct
+ice_flow_engine ice_switch_engine = {
+       .init = ice_switch_init,
+       .uninit = ice_switch_uninit,
+       .create = ice_switch_create,
+       .destroy = ice_switch_destroy,
+       .query_count = ice_switch_query,
+       .free = ice_switch_filter_rule_free,
+       .type = ICE_FLOW_ENGINE_SWITCH,
+};
+
+static struct
+ice_flow_parser ice_switch_dist_parser_os = {
+       .engine = &ice_switch_engine,
+       .array = ice_switch_pattern_dist_os,
+       .array_len = RTE_DIM(ice_switch_pattern_dist_os),
+       .parse_pattern_action = ice_switch_parse_pattern_action,
+       .stage = ICE_FLOW_STAGE_DISTRIBUTOR,
+};
+
+static struct
+ice_flow_parser ice_switch_dist_parser_comms = {
+       .engine = &ice_switch_engine,
+       .array = ice_switch_pattern_dist_comms,
+       .array_len = RTE_DIM(ice_switch_pattern_dist_comms),
+       .parse_pattern_action = ice_switch_parse_pattern_action,
+       .stage = ICE_FLOW_STAGE_DISTRIBUTOR,
+};
+
+static struct
+ice_flow_parser ice_switch_perm_parser = {
+       .engine = &ice_switch_engine,
+       .array = ice_switch_pattern_perm,
+       .array_len = RTE_DIM(ice_switch_pattern_perm),
+       .parse_pattern_action = ice_switch_parse_pattern_action,
+       .stage = ICE_FLOW_STAGE_PERMISSION,
+};
+
+RTE_INIT(ice_sw_engine_init)
+{
+       struct ice_flow_engine *engine = &ice_switch_engine;
+       ice_register_flow_engine(engine);
+}