net/qede: reorganize filter code
authorShahed Shaikh <shahed.shaikh@cavium.com>
Sat, 8 Sep 2018 20:30:59 +0000 (13:30 -0700)
committerFerruh Yigit <ferruh.yigit@intel.com>
Thu, 27 Sep 2018 23:41:01 +0000 (01:41 +0200)
 - rename qede_fdir.c to qede_filter.c
 - move all filter code to qede_filter.c

Signed-off-by: Shahed Shaikh <shahed.shaikh@cavium.com>
drivers/net/qede/Makefile
drivers/net/qede/meson.build
drivers/net/qede/qede_ethdev.c
drivers/net/qede/qede_ethdev.h
drivers/net/qede/qede_fdir.c [deleted file]
drivers/net/qede/qede_filter.c [new file with mode: 0644]

index 488ca1d..2ecbd8d 100644 (file)
@@ -105,6 +105,6 @@ SRCS-$(CONFIG_RTE_LIBRTE_QEDE_PMD) += ecore_vf.c
 SRCS-$(CONFIG_RTE_LIBRTE_QEDE_PMD) += qede_ethdev.c
 SRCS-$(CONFIG_RTE_LIBRTE_QEDE_PMD) += qede_main.c
 SRCS-$(CONFIG_RTE_LIBRTE_QEDE_PMD) += qede_rxtx.c
-SRCS-$(CONFIG_RTE_LIBRTE_QEDE_PMD) += qede_fdir.c
+SRCS-$(CONFIG_RTE_LIBRTE_QEDE_PMD) += qede_filter.c
 
 include $(RTE_SDK)/mk/rte.lib.mk
index 6280073..12388a6 100644 (file)
@@ -6,7 +6,7 @@ objs = [base_objs]
 
 sources = files(
        'qede_ethdev.c',
-       'qede_fdir.c',
+       'qede_filter.c',
        'qede_main.c',
        'qede_rxtx.c',
 )
index 7bb52b1..361041e 100644 (file)
@@ -16,111 +16,6 @@ int qede_logtype_driver;
 static const struct qed_eth_ops *qed_ops;
 #define QEDE_SP_TIMER_PERIOD   10000 /* 100ms */
 
-/* VXLAN tunnel classification mapping */
-const struct _qede_udp_tunn_types {
-       uint16_t rte_filter_type;
-       enum ecore_filter_ucast_type qede_type;
-       enum ecore_tunn_clss qede_tunn_clss;
-       const char *string;
-} qede_tunn_types[] = {
-       {
-               ETH_TUNNEL_FILTER_OMAC,
-               ECORE_FILTER_MAC,
-               ECORE_TUNN_CLSS_MAC_VLAN,
-               "outer-mac"
-       },
-       {
-               ETH_TUNNEL_FILTER_TENID,
-               ECORE_FILTER_VNI,
-               ECORE_TUNN_CLSS_MAC_VNI,
-               "vni"
-       },
-       {
-               ETH_TUNNEL_FILTER_IMAC,
-               ECORE_FILTER_INNER_MAC,
-               ECORE_TUNN_CLSS_INNER_MAC_VLAN,
-               "inner-mac"
-       },
-       {
-               ETH_TUNNEL_FILTER_IVLAN,
-               ECORE_FILTER_INNER_VLAN,
-               ECORE_TUNN_CLSS_INNER_MAC_VLAN,
-               "inner-vlan"
-       },
-       {
-               ETH_TUNNEL_FILTER_OMAC | ETH_TUNNEL_FILTER_TENID,
-               ECORE_FILTER_MAC_VNI_PAIR,
-               ECORE_TUNN_CLSS_MAC_VNI,
-               "outer-mac and vni"
-       },
-       {
-               ETH_TUNNEL_FILTER_OMAC | ETH_TUNNEL_FILTER_IMAC,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "outer-mac and inner-mac"
-       },
-       {
-               ETH_TUNNEL_FILTER_OMAC | ETH_TUNNEL_FILTER_IVLAN,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "outer-mac and inner-vlan"
-       },
-       {
-               ETH_TUNNEL_FILTER_TENID | ETH_TUNNEL_FILTER_IMAC,
-               ECORE_FILTER_INNER_MAC_VNI_PAIR,
-               ECORE_TUNN_CLSS_INNER_MAC_VNI,
-               "vni and inner-mac",
-       },
-       {
-               ETH_TUNNEL_FILTER_TENID | ETH_TUNNEL_FILTER_IVLAN,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "vni and inner-vlan",
-       },
-       {
-               ETH_TUNNEL_FILTER_IMAC | ETH_TUNNEL_FILTER_IVLAN,
-               ECORE_FILTER_INNER_PAIR,
-               ECORE_TUNN_CLSS_INNER_MAC_VLAN,
-               "inner-mac and inner-vlan",
-       },
-       {
-               ETH_TUNNEL_FILTER_OIP,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "outer-IP"
-       },
-       {
-               ETH_TUNNEL_FILTER_IIP,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "inner-IP"
-       },
-       {
-               RTE_TUNNEL_FILTER_IMAC_IVLAN,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "IMAC_IVLAN"
-       },
-       {
-               RTE_TUNNEL_FILTER_IMAC_IVLAN_TENID,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "IMAC_IVLAN_TENID"
-       },
-       {
-               RTE_TUNNEL_FILTER_IMAC_TENID,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "IMAC_TENID"
-       },
-       {
-               RTE_TUNNEL_FILTER_OMAC_TENID_IMAC,
-               ECORE_FILTER_UNUSED,
-               MAX_ECORE_TUNN_CLSS,
-               "OMAC_TENID_IMAC"
-       },
-};
-
 struct rte_qede_xstats_name_off {
        char name[RTE_ETH_XSTATS_NAME_SIZE];
        uint64_t offset;
@@ -614,14 +509,6 @@ int qede_enable_tpa(struct rte_eth_dev *eth_dev, bool flg)
        return 0;
 }
 
-static void qede_set_ucast_cmn_params(struct ecore_filter_ucast *ucast)
-{
-       memset(ucast, 0, sizeof(struct ecore_filter_ucast));
-       ucast->is_rx_filter = true;
-       ucast->is_tx_filter = true;
-       /* ucast->assert_on_error = true; - For debug */
-}
-
 static int
 qed_configure_filter_rx_mode(struct rte_eth_dev *eth_dev,
                             enum qed_filter_rx_mode_type type)
@@ -660,167 +547,7 @@ qed_configure_filter_rx_mode(struct rte_eth_dev *eth_dev,
                        ECORE_SPQ_MODE_CB, NULL);
 }
 
-static int
-qede_tunnel_update(struct qede_dev *qdev,
-                  struct ecore_tunnel_info *tunn_info)
-{
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       enum _ecore_status_t rc = ECORE_INVAL;
-       struct ecore_hwfn *p_hwfn;
-       struct ecore_ptt *p_ptt;
-       int i;
-
-       for_each_hwfn(edev, i) {
-               p_hwfn = &edev->hwfns[i];
-               if (IS_PF(edev)) {
-                       p_ptt = ecore_ptt_acquire(p_hwfn);
-                       if (!p_ptt) {
-                               DP_ERR(p_hwfn, "Can't acquire PTT\n");
-                               return -EAGAIN;
-                       }
-               } else {
-                       p_ptt = NULL;
-               }
-
-               rc = ecore_sp_pf_update_tunn_cfg(p_hwfn, p_ptt,
-                               tunn_info, ECORE_SPQ_MODE_CB, NULL);
-               if (IS_PF(edev))
-                       ecore_ptt_release(p_hwfn, p_ptt);
-
-               if (rc != ECORE_SUCCESS)
-                       break;
-       }
-
-       return rc;
-}
-
-static int
-qede_vxlan_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
-                 bool enable)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       enum _ecore_status_t rc = ECORE_INVAL;
-       struct ecore_tunnel_info tunn;
-
-       if (qdev->vxlan.enable == enable)
-               return ECORE_SUCCESS;
-
-       memset(&tunn, 0, sizeof(struct ecore_tunnel_info));
-       tunn.vxlan.b_update_mode = true;
-       tunn.vxlan.b_mode_enabled = enable;
-       tunn.b_update_rx_cls = true;
-       tunn.b_update_tx_cls = true;
-       tunn.vxlan.tun_cls = clss;
-
-       tunn.vxlan_port.b_update_port = true;
-       tunn.vxlan_port.port = enable ? QEDE_VXLAN_DEF_PORT : 0;
-
-       rc = qede_tunnel_update(qdev, &tunn);
-       if (rc == ECORE_SUCCESS) {
-               qdev->vxlan.enable = enable;
-               qdev->vxlan.udp_port = (enable) ? QEDE_VXLAN_DEF_PORT : 0;
-               DP_INFO(edev, "vxlan is %s, UDP port = %d\n",
-                       enable ? "enabled" : "disabled", qdev->vxlan.udp_port);
-       } else {
-               DP_ERR(edev, "Failed to update tunn_clss %u\n",
-                      tunn.vxlan.tun_cls);
-       }
-
-       return rc;
-}
-
-static int
-qede_geneve_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
-                 bool enable)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       enum _ecore_status_t rc = ECORE_INVAL;
-       struct ecore_tunnel_info tunn;
-
-       memset(&tunn, 0, sizeof(struct ecore_tunnel_info));
-       tunn.l2_geneve.b_update_mode = true;
-       tunn.l2_geneve.b_mode_enabled = enable;
-       tunn.ip_geneve.b_update_mode = true;
-       tunn.ip_geneve.b_mode_enabled = enable;
-       tunn.l2_geneve.tun_cls = clss;
-       tunn.ip_geneve.tun_cls = clss;
-       tunn.b_update_rx_cls = true;
-       tunn.b_update_tx_cls = true;
-
-       tunn.geneve_port.b_update_port = true;
-       tunn.geneve_port.port = enable ? QEDE_GENEVE_DEF_PORT : 0;
-
-       rc = qede_tunnel_update(qdev, &tunn);
-       if (rc == ECORE_SUCCESS) {
-               qdev->geneve.enable = enable;
-               qdev->geneve.udp_port = (enable) ? QEDE_GENEVE_DEF_PORT : 0;
-               DP_INFO(edev, "GENEVE is %s, UDP port = %d\n",
-                       enable ? "enabled" : "disabled", qdev->geneve.udp_port);
-       } else {
-               DP_ERR(edev, "Failed to update tunn_clss %u\n",
-                      clss);
-       }
-
-       return rc;
-}
-
-static int
-qede_ipgre_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
-                 bool enable)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       enum _ecore_status_t rc = ECORE_INVAL;
-       struct ecore_tunnel_info tunn;
-
-       memset(&tunn, 0, sizeof(struct ecore_tunnel_info));
-       tunn.ip_gre.b_update_mode = true;
-       tunn.ip_gre.b_mode_enabled = enable;
-       tunn.ip_gre.tun_cls = clss;
-       tunn.ip_gre.tun_cls = clss;
-       tunn.b_update_rx_cls = true;
-       tunn.b_update_tx_cls = true;
-
-       rc = qede_tunnel_update(qdev, &tunn);
-       if (rc == ECORE_SUCCESS) {
-               qdev->ipgre.enable = enable;
-               DP_INFO(edev, "IPGRE is %s\n",
-                       enable ? "enabled" : "disabled");
-       } else {
-               DP_ERR(edev, "Failed to update tunn_clss %u\n",
-                      clss);
-       }
-
-       return rc;
-}
-
-static int
-qede_tunn_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
-                enum rte_eth_tunnel_type tunn_type, bool enable)
-{
-       int rc = -EINVAL;
-
-       switch (tunn_type) {
-       case RTE_TUNNEL_TYPE_VXLAN:
-               rc = qede_vxlan_enable(eth_dev, clss, enable);
-               break;
-       case RTE_TUNNEL_TYPE_GENEVE:
-               rc = qede_geneve_enable(eth_dev, clss, enable);
-               break;
-       case RTE_TUNNEL_TYPE_IP_IN_GRE:
-               rc = qede_ipgre_enable(eth_dev, clss, enable);
-               break;
-       default:
-               rc = -EINVAL;
-               break;
-       }
-
-       return rc;
-}
-
-static int
+int
 qede_ucast_filter(struct rte_eth_dev *eth_dev, struct ecore_filter_ucast *ucast,
                  bool add)
 {
@@ -941,7 +668,7 @@ static int qede_del_mcast_filters(struct rte_eth_dev *eth_dev)
        return 0;
 }
 
-static enum _ecore_status_t
+enum _ecore_status_t
 qede_mac_int_ops(struct rte_eth_dev *eth_dev, struct ecore_filter_ucast *ucast,
                 bool add)
 {
@@ -1033,7 +760,7 @@ qede_mac_addr_set(struct rte_eth_dev *eth_dev, struct ether_addr *mac_addr)
        return qede_mac_addr_add(eth_dev, mac_addr, 0, 0);
 }
 
-static void qede_config_accept_any_vlan(struct qede_dev *qdev, bool flg)
+void qede_config_accept_any_vlan(struct qede_dev *qdev, bool flg)
 {
        struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
        struct ecore_sp_vport_update_params params;
@@ -2567,414 +2294,6 @@ static int qede_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
        return 0;
 }
 
-static int
-qede_udp_dst_port_del(struct rte_eth_dev *eth_dev,
-                     struct rte_eth_udp_tunnel *tunnel_udp)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct ecore_tunnel_info tunn; /* @DPDK */
-       uint16_t udp_port;
-       int rc;
-
-       PMD_INIT_FUNC_TRACE(edev);
-
-       memset(&tunn, 0, sizeof(tunn));
-
-       switch (tunnel_udp->prot_type) {
-       case RTE_TUNNEL_TYPE_VXLAN:
-               if (qdev->vxlan.udp_port != tunnel_udp->udp_port) {
-                       DP_ERR(edev, "UDP port %u doesn't exist\n",
-                               tunnel_udp->udp_port);
-                       return ECORE_INVAL;
-               }
-               udp_port = 0;
-
-               tunn.vxlan_port.b_update_port = true;
-               tunn.vxlan_port.port = udp_port;
-
-               rc = qede_tunnel_update(qdev, &tunn);
-               if (rc != ECORE_SUCCESS) {
-                       DP_ERR(edev, "Unable to config UDP port %u\n",
-                              tunn.vxlan_port.port);
-                       return rc;
-               }
-
-               qdev->vxlan.udp_port = udp_port;
-               /* If the request is to delete UDP port and if the number of
-                * VXLAN filters have reached 0 then VxLAN offload can be be
-                * disabled.
-                */
-               if (qdev->vxlan.enable && qdev->vxlan.num_filters == 0)
-                       return qede_vxlan_enable(eth_dev,
-                                       ECORE_TUNN_CLSS_MAC_VLAN, false);
-
-               break;
-       case RTE_TUNNEL_TYPE_GENEVE:
-               if (qdev->geneve.udp_port != tunnel_udp->udp_port) {
-                       DP_ERR(edev, "UDP port %u doesn't exist\n",
-                               tunnel_udp->udp_port);
-                       return ECORE_INVAL;
-               }
-
-               udp_port = 0;
-
-               tunn.geneve_port.b_update_port = true;
-               tunn.geneve_port.port = udp_port;
-
-               rc = qede_tunnel_update(qdev, &tunn);
-               if (rc != ECORE_SUCCESS) {
-                       DP_ERR(edev, "Unable to config UDP port %u\n",
-                              tunn.vxlan_port.port);
-                       return rc;
-               }
-
-               qdev->vxlan.udp_port = udp_port;
-               /* If the request is to delete UDP port and if the number of
-                * GENEVE filters have reached 0 then GENEVE offload can be be
-                * disabled.
-                */
-               if (qdev->geneve.enable && qdev->geneve.num_filters == 0)
-                       return qede_geneve_enable(eth_dev,
-                                       ECORE_TUNN_CLSS_MAC_VLAN, false);
-
-               break;
-
-       default:
-               return ECORE_INVAL;
-       }
-
-       return 0;
-
-}
-static int
-qede_udp_dst_port_add(struct rte_eth_dev *eth_dev,
-                     struct rte_eth_udp_tunnel *tunnel_udp)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct ecore_tunnel_info tunn; /* @DPDK */
-       uint16_t udp_port;
-       int rc;
-
-       PMD_INIT_FUNC_TRACE(edev);
-
-       memset(&tunn, 0, sizeof(tunn));
-
-       switch (tunnel_udp->prot_type) {
-       case RTE_TUNNEL_TYPE_VXLAN:
-               if (qdev->vxlan.udp_port == tunnel_udp->udp_port) {
-                       DP_INFO(edev,
-                               "UDP port %u for VXLAN was already configured\n",
-                               tunnel_udp->udp_port);
-                       return ECORE_SUCCESS;
-               }
-
-               /* Enable VxLAN tunnel with default MAC/VLAN classification if
-                * it was not enabled while adding VXLAN filter before UDP port
-                * update.
-                */
-               if (!qdev->vxlan.enable) {
-                       rc = qede_vxlan_enable(eth_dev,
-                               ECORE_TUNN_CLSS_MAC_VLAN, true);
-                       if (rc != ECORE_SUCCESS) {
-                               DP_ERR(edev, "Failed to enable VXLAN "
-                                       "prior to updating UDP port\n");
-                               return rc;
-                       }
-               }
-               udp_port = tunnel_udp->udp_port;
-
-               tunn.vxlan_port.b_update_port = true;
-               tunn.vxlan_port.port = udp_port;
-
-               rc = qede_tunnel_update(qdev, &tunn);
-               if (rc != ECORE_SUCCESS) {
-                       DP_ERR(edev, "Unable to config UDP port %u for VXLAN\n",
-                              udp_port);
-                       return rc;
-               }
-
-               DP_INFO(edev, "Updated UDP port %u for VXLAN\n", udp_port);
-
-               qdev->vxlan.udp_port = udp_port;
-               break;
-       case RTE_TUNNEL_TYPE_GENEVE:
-               if (qdev->geneve.udp_port == tunnel_udp->udp_port) {
-                       DP_INFO(edev,
-                               "UDP port %u for GENEVE was already configured\n",
-                               tunnel_udp->udp_port);
-                       return ECORE_SUCCESS;
-               }
-
-               /* Enable GENEVE tunnel with default MAC/VLAN classification if
-                * it was not enabled while adding GENEVE filter before UDP port
-                * update.
-                */
-               if (!qdev->geneve.enable) {
-                       rc = qede_geneve_enable(eth_dev,
-                               ECORE_TUNN_CLSS_MAC_VLAN, true);
-                       if (rc != ECORE_SUCCESS) {
-                               DP_ERR(edev, "Failed to enable GENEVE "
-                                       "prior to updating UDP port\n");
-                               return rc;
-                       }
-               }
-               udp_port = tunnel_udp->udp_port;
-
-               tunn.geneve_port.b_update_port = true;
-               tunn.geneve_port.port = udp_port;
-
-               rc = qede_tunnel_update(qdev, &tunn);
-               if (rc != ECORE_SUCCESS) {
-                       DP_ERR(edev, "Unable to config UDP port %u for GENEVE\n",
-                              udp_port);
-                       return rc;
-               }
-
-               DP_INFO(edev, "Updated UDP port %u for GENEVE\n", udp_port);
-
-               qdev->geneve.udp_port = udp_port;
-               break;
-       default:
-               return ECORE_INVAL;
-       }
-
-       return 0;
-}
-
-static void qede_get_ecore_tunn_params(uint32_t filter, uint32_t *type,
-                                      uint32_t *clss, char *str)
-{
-       uint16_t j;
-       *clss = MAX_ECORE_TUNN_CLSS;
-
-       for (j = 0; j < RTE_DIM(qede_tunn_types); j++) {
-               if (filter == qede_tunn_types[j].rte_filter_type) {
-                       *type = qede_tunn_types[j].qede_type;
-                       *clss = qede_tunn_types[j].qede_tunn_clss;
-                       strcpy(str, qede_tunn_types[j].string);
-                       return;
-               }
-       }
-}
-
-static int
-qede_set_ucast_tunn_cmn_param(struct ecore_filter_ucast *ucast,
-                             const struct rte_eth_tunnel_filter_conf *conf,
-                             uint32_t type)
-{
-       /* Init commmon ucast params first */
-       qede_set_ucast_cmn_params(ucast);
-
-       /* Copy out the required fields based on classification type */
-       ucast->type = type;
-
-       switch (type) {
-       case ECORE_FILTER_VNI:
-               ucast->vni = conf->tenant_id;
-       break;
-       case ECORE_FILTER_INNER_VLAN:
-               ucast->vlan = conf->inner_vlan;
-       break;
-       case ECORE_FILTER_MAC:
-               memcpy(ucast->mac, conf->outer_mac.addr_bytes,
-                      ETHER_ADDR_LEN);
-       break;
-       case ECORE_FILTER_INNER_MAC:
-               memcpy(ucast->mac, conf->inner_mac.addr_bytes,
-                      ETHER_ADDR_LEN);
-       break;
-       case ECORE_FILTER_MAC_VNI_PAIR:
-               memcpy(ucast->mac, conf->outer_mac.addr_bytes,
-                       ETHER_ADDR_LEN);
-               ucast->vni = conf->tenant_id;
-       break;
-       case ECORE_FILTER_INNER_MAC_VNI_PAIR:
-               memcpy(ucast->mac, conf->inner_mac.addr_bytes,
-                       ETHER_ADDR_LEN);
-               ucast->vni = conf->tenant_id;
-       break;
-       case ECORE_FILTER_INNER_PAIR:
-               memcpy(ucast->mac, conf->inner_mac.addr_bytes,
-                       ETHER_ADDR_LEN);
-               ucast->vlan = conf->inner_vlan;
-       break;
-       default:
-               return -EINVAL;
-       }
-
-       return ECORE_SUCCESS;
-}
-
-static int
-_qede_tunn_filter_config(struct rte_eth_dev *eth_dev,
-                        const struct rte_eth_tunnel_filter_conf *conf,
-                        __attribute__((unused)) enum rte_filter_op filter_op,
-                        enum ecore_tunn_clss *clss,
-                        bool add)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct ecore_filter_ucast ucast = {0};
-       enum ecore_filter_ucast_type type;
-       uint16_t filter_type = 0;
-       char str[80];
-       int rc;
-
-       filter_type = conf->filter_type;
-       /* Determine if the given filter classification is supported */
-       qede_get_ecore_tunn_params(filter_type, &type, clss, str);
-       if (*clss == MAX_ECORE_TUNN_CLSS) {
-               DP_ERR(edev, "Unsupported filter type\n");
-               return -EINVAL;
-       }
-       /* Init tunnel ucast params */
-       rc = qede_set_ucast_tunn_cmn_param(&ucast, conf, type);
-       if (rc != ECORE_SUCCESS) {
-               DP_ERR(edev, "Unsupported Tunnel filter type 0x%x\n",
-               conf->filter_type);
-               return rc;
-       }
-       DP_INFO(edev, "Rule: \"%s\", op %d, type 0x%x\n",
-               str, filter_op, ucast.type);
-
-       ucast.opcode = add ? ECORE_FILTER_ADD : ECORE_FILTER_REMOVE;
-
-       /* Skip MAC/VLAN if filter is based on VNI */
-       if (!(filter_type & ETH_TUNNEL_FILTER_TENID)) {
-               rc = qede_mac_int_ops(eth_dev, &ucast, add);
-               if ((rc == 0) && add) {
-                       /* Enable accept anyvlan */
-                       qede_config_accept_any_vlan(qdev, true);
-               }
-       } else {
-               rc = qede_ucast_filter(eth_dev, &ucast, add);
-               if (rc == 0)
-                       rc = ecore_filter_ucast_cmd(edev, &ucast,
-                                           ECORE_SPQ_MODE_CB, NULL);
-       }
-
-       return rc;
-}
-
-static int
-qede_tunn_filter_config(struct rte_eth_dev *eth_dev,
-                       enum rte_filter_op filter_op,
-                       const struct rte_eth_tunnel_filter_conf *conf)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       enum ecore_tunn_clss clss = MAX_ECORE_TUNN_CLSS;
-       bool add;
-       int rc;
-
-       PMD_INIT_FUNC_TRACE(edev);
-
-       switch (filter_op) {
-       case RTE_ETH_FILTER_ADD:
-               add = true;
-               break;
-       case RTE_ETH_FILTER_DELETE:
-               add = false;
-               break;
-       default:
-               DP_ERR(edev, "Unsupported operation %d\n", filter_op);
-               return -EINVAL;
-       }
-
-       if (IS_VF(edev))
-               return qede_tunn_enable(eth_dev,
-                                       ECORE_TUNN_CLSS_MAC_VLAN,
-                                       conf->tunnel_type, add);
-
-       rc = _qede_tunn_filter_config(eth_dev, conf, filter_op, &clss, add);
-       if (rc != ECORE_SUCCESS)
-               return rc;
-
-       if (add) {
-               if (conf->tunnel_type == RTE_TUNNEL_TYPE_VXLAN) {
-                       qdev->vxlan.num_filters++;
-                       qdev->vxlan.filter_type = conf->filter_type;
-               } else { /* GENEVE */
-                       qdev->geneve.num_filters++;
-                       qdev->geneve.filter_type = conf->filter_type;
-               }
-
-               if (!qdev->vxlan.enable || !qdev->geneve.enable ||
-                   !qdev->ipgre.enable)
-                       return qede_tunn_enable(eth_dev, clss,
-                                               conf->tunnel_type,
-                                               true);
-       } else {
-               if (conf->tunnel_type == RTE_TUNNEL_TYPE_VXLAN)
-                       qdev->vxlan.num_filters--;
-               else /*GENEVE*/
-                       qdev->geneve.num_filters--;
-
-               /* Disable VXLAN if VXLAN filters become 0 */
-               if ((qdev->vxlan.num_filters == 0) ||
-                   (qdev->geneve.num_filters == 0))
-                       return qede_tunn_enable(eth_dev, clss,
-                                               conf->tunnel_type,
-                                               false);
-       }
-
-       return 0;
-}
-
-int qede_dev_filter_ctrl(struct rte_eth_dev *eth_dev,
-                        enum rte_filter_type filter_type,
-                        enum rte_filter_op filter_op,
-                        void *arg)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct rte_eth_tunnel_filter_conf *filter_conf =
-                       (struct rte_eth_tunnel_filter_conf *)arg;
-
-       switch (filter_type) {
-       case RTE_ETH_FILTER_TUNNEL:
-               switch (filter_conf->tunnel_type) {
-               case RTE_TUNNEL_TYPE_VXLAN:
-               case RTE_TUNNEL_TYPE_GENEVE:
-               case RTE_TUNNEL_TYPE_IP_IN_GRE:
-                       DP_INFO(edev,
-                               "Packet steering to the specified Rx queue"
-                               " is not supported with UDP tunneling");
-                       return(qede_tunn_filter_config(eth_dev, filter_op,
-                                                     filter_conf));
-               case RTE_TUNNEL_TYPE_TEREDO:
-               case RTE_TUNNEL_TYPE_NVGRE:
-               case RTE_L2_TUNNEL_TYPE_E_TAG:
-                       DP_ERR(edev, "Unsupported tunnel type %d\n",
-                               filter_conf->tunnel_type);
-                       return -EINVAL;
-               case RTE_TUNNEL_TYPE_NONE:
-               default:
-                       return 0;
-               }
-               break;
-       case RTE_ETH_FILTER_FDIR:
-               return qede_fdir_filter_conf(eth_dev, filter_op, arg);
-       case RTE_ETH_FILTER_NTUPLE:
-               return qede_ntuple_filter_conf(eth_dev, filter_op, arg);
-       case RTE_ETH_FILTER_MACVLAN:
-       case RTE_ETH_FILTER_ETHERTYPE:
-       case RTE_ETH_FILTER_FLEXIBLE:
-       case RTE_ETH_FILTER_SYN:
-       case RTE_ETH_FILTER_HASH:
-       case RTE_ETH_FILTER_L2_TUNNEL:
-       case RTE_ETH_FILTER_MAX:
-       default:
-               DP_ERR(edev, "Unsupported filter type %d\n",
-                       filter_type);
-               return -EINVAL;
-       }
-
-       return 0;
-}
-
 static const struct eth_dev_ops qede_eth_dev_ops = {
        .dev_configure = qede_dev_configure,
        .dev_infos_get = qede_dev_info_get,
index 6e9a5b4..d54f19b 100644 (file)
@@ -215,6 +215,15 @@ struct qede_dev {
        void *ethdev;
 };
 
+static inline void qede_set_ucast_cmn_params(struct ecore_filter_ucast *ucast)
+{
+       memset(ucast, 0, sizeof(struct ecore_filter_ucast));
+       ucast->is_rx_filter = true;
+       ucast->is_tx_filter = true;
+       /* ucast->assert_on_error = true; - For debug */
+}
+
+
 /* Non-static functions */
 int qede_config_rss(struct rte_eth_dev *eth_dev);
 
@@ -235,9 +244,6 @@ int qede_link_update(struct rte_eth_dev *eth_dev,
 int qede_dev_filter_ctrl(struct rte_eth_dev *dev, enum rte_filter_type type,
                         enum rte_filter_op op, void *arg);
 
-int qede_fdir_filter_conf(struct rte_eth_dev *eth_dev,
-                         enum rte_filter_op filter_op, void *arg);
-
 int qede_ntuple_filter_conf(struct rte_eth_dev *eth_dev,
                            enum rte_filter_op filter_op, void *arg);
 
@@ -255,5 +261,16 @@ int qede_activate_vport(struct rte_eth_dev *eth_dev, bool flg);
 int qede_update_mtu(struct rte_eth_dev *eth_dev, uint16_t mtu);
 
 int qede_enable_tpa(struct rte_eth_dev *eth_dev, bool flg);
-
+int qede_udp_dst_port_del(struct rte_eth_dev *eth_dev,
+                         struct rte_eth_udp_tunnel *tunnel_udp);
+int qede_udp_dst_port_add(struct rte_eth_dev *eth_dev,
+                         struct rte_eth_udp_tunnel *tunnel_udp);
+
+enum _ecore_status_t
+qede_mac_int_ops(struct rte_eth_dev *eth_dev, struct ecore_filter_ucast *ucast,
+                bool add);
+void qede_config_accept_any_vlan(struct qede_dev *qdev, bool flg);
+int qede_ucast_filter(struct rte_eth_dev *eth_dev,
+                     struct ecore_filter_ucast *ucast,
+                     bool add);
 #endif /* _QEDE_ETHDEV_H_ */
diff --git a/drivers/net/qede/qede_fdir.c b/drivers/net/qede/qede_fdir.c
deleted file mode 100644 (file)
index 83580d0..0000000
+++ /dev/null
@@ -1,470 +0,0 @@
-/* SPDX-License-Identifier: BSD-3-Clause
- * Copyright (c) 2017 Cavium Inc.
- * All rights reserved.
- * www.cavium.com
- */
-
-#include <rte_udp.h>
-#include <rte_tcp.h>
-#include <rte_sctp.h>
-#include <rte_errno.h>
-
-#include "qede_ethdev.h"
-
-#define IP_VERSION                             (0x40)
-#define IP_HDRLEN                              (0x5)
-#define QEDE_FDIR_IP_DEFAULT_VERSION_IHL       (IP_VERSION | IP_HDRLEN)
-#define QEDE_FDIR_TCP_DEFAULT_DATAOFF          (0x50)
-#define QEDE_FDIR_IPV4_DEF_TTL                 (64)
-
-/* Sum of length of header types of L2, L3, L4.
- * L2 : ether_hdr + vlan_hdr + vxlan_hdr
- * L3 : ipv6_hdr
- * L4 : tcp_hdr
- */
-#define QEDE_MAX_FDIR_PKT_LEN                  (86)
-
-#ifndef IPV6_ADDR_LEN
-#define IPV6_ADDR_LEN                          (16)
-#endif
-
-#define QEDE_VALID_FLOW(flow_type) \
-       ((flow_type) == RTE_ETH_FLOW_NONFRAG_IPV4_TCP   || \
-       (flow_type) == RTE_ETH_FLOW_NONFRAG_IPV4_UDP    || \
-       (flow_type) == RTE_ETH_FLOW_NONFRAG_IPV6_TCP    || \
-       (flow_type) == RTE_ETH_FLOW_NONFRAG_IPV6_UDP)
-
-/* Note: Flowdir support is only partial.
- * For ex: drop_queue, FDIR masks, flex_conf are not supported.
- * Parameters like pballoc/status fields are irrelevant here.
- */
-int qede_check_fdir_support(struct rte_eth_dev *eth_dev)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct rte_fdir_conf *fdir = &eth_dev->data->dev_conf.fdir_conf;
-
-       /* check FDIR modes */
-       switch (fdir->mode) {
-       case RTE_FDIR_MODE_NONE:
-               qdev->fdir_info.arfs.arfs_enable = false;
-               DP_INFO(edev, "flowdir is disabled\n");
-       break;
-       case RTE_FDIR_MODE_PERFECT:
-               if (ECORE_IS_CMT(edev)) {
-                       DP_ERR(edev, "flowdir is not supported in 100G mode\n");
-                       qdev->fdir_info.arfs.arfs_enable = false;
-                       return -ENOTSUP;
-               }
-               qdev->fdir_info.arfs.arfs_enable = true;
-               DP_INFO(edev, "flowdir is enabled\n");
-       break;
-       case RTE_FDIR_MODE_PERFECT_TUNNEL:
-       case RTE_FDIR_MODE_SIGNATURE:
-       case RTE_FDIR_MODE_PERFECT_MAC_VLAN:
-               DP_ERR(edev, "Unsupported flowdir mode %d\n", fdir->mode);
-               return -ENOTSUP;
-       }
-
-       return 0;
-}
-
-void qede_fdir_dealloc_resc(struct rte_eth_dev *eth_dev)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct qede_fdir_entry *tmp = NULL;
-
-       SLIST_FOREACH(tmp, &qdev->fdir_info.fdir_list_head, list) {
-               if (tmp) {
-                       if (tmp->mz)
-                               rte_memzone_free(tmp->mz);
-                       SLIST_REMOVE(&qdev->fdir_info.fdir_list_head, tmp,
-                                    qede_fdir_entry, list);
-                       rte_free(tmp);
-               }
-       }
-}
-
-static int
-qede_config_cmn_fdir_filter(struct rte_eth_dev *eth_dev,
-                           struct rte_eth_fdir_filter *fdir_filter,
-                           bool add)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       char mz_name[RTE_MEMZONE_NAMESIZE] = {0};
-       struct qede_fdir_entry *tmp = NULL;
-       struct qede_fdir_entry *fdir = NULL;
-       const struct rte_memzone *mz;
-       struct ecore_hwfn *p_hwfn;
-       enum _ecore_status_t rc;
-       uint16_t pkt_len;
-       void *pkt;
-
-       if (add) {
-               if (qdev->fdir_info.filter_count == QEDE_RFS_MAX_FLTR - 1) {
-                       DP_ERR(edev, "Reached max flowdir filter limit\n");
-                       return -EINVAL;
-               }
-               fdir = rte_malloc(NULL, sizeof(struct qede_fdir_entry),
-                                 RTE_CACHE_LINE_SIZE);
-               if (!fdir) {
-                       DP_ERR(edev, "Did not allocate memory for fdir\n");
-                       return -ENOMEM;
-               }
-       }
-       /* soft_id could have been used as memzone string, but soft_id is
-        * not currently used so it has no significance.
-        */
-       snprintf(mz_name, sizeof(mz_name) - 1, "%lx",
-                (unsigned long)rte_get_timer_cycles());
-       mz = rte_memzone_reserve_aligned(mz_name, QEDE_MAX_FDIR_PKT_LEN,
-                                        SOCKET_ID_ANY, 0, RTE_CACHE_LINE_SIZE);
-       if (!mz) {
-               DP_ERR(edev, "Failed to allocate memzone for fdir, err = %s\n",
-                      rte_strerror(rte_errno));
-               rc = -rte_errno;
-               goto err1;
-       }
-
-       pkt = mz->addr;
-       memset(pkt, 0, QEDE_MAX_FDIR_PKT_LEN);
-       pkt_len = qede_fdir_construct_pkt(eth_dev, fdir_filter, pkt,
-                                         &qdev->fdir_info.arfs);
-       if (pkt_len == 0) {
-               rc = -EINVAL;
-               goto err2;
-       }
-       DP_INFO(edev, "pkt_len = %u memzone = %s\n", pkt_len, mz_name);
-       if (add) {
-               SLIST_FOREACH(tmp, &qdev->fdir_info.fdir_list_head, list) {
-                       if (memcmp(tmp->mz->addr, pkt, pkt_len) == 0) {
-                               DP_INFO(edev, "flowdir filter exist\n");
-                               rc = 0;
-                               goto err2;
-                       }
-               }
-       } else {
-               SLIST_FOREACH(tmp, &qdev->fdir_info.fdir_list_head, list) {
-                       if (memcmp(tmp->mz->addr, pkt, pkt_len) == 0)
-                               break;
-               }
-               if (!tmp) {
-                       DP_ERR(edev, "flowdir filter does not exist\n");
-                       rc = -EEXIST;
-                       goto err2;
-               }
-       }
-       p_hwfn = ECORE_LEADING_HWFN(edev);
-       if (add) {
-               if (!qdev->fdir_info.arfs.arfs_enable) {
-                       /* Force update */
-                       eth_dev->data->dev_conf.fdir_conf.mode =
-                                               RTE_FDIR_MODE_PERFECT;
-                       qdev->fdir_info.arfs.arfs_enable = true;
-                       DP_INFO(edev, "Force enable flowdir in perfect mode\n");
-               }
-               /* Enable ARFS searcher with updated flow_types */
-               ecore_arfs_mode_configure(p_hwfn, p_hwfn->p_arfs_ptt,
-                                         &qdev->fdir_info.arfs);
-       }
-       /* configure filter with ECORE_SPQ_MODE_EBLOCK */
-       rc = ecore_configure_rfs_ntuple_filter(p_hwfn, NULL,
-                                              (dma_addr_t)mz->iova,
-                                              pkt_len,
-                                              fdir_filter->action.rx_queue,
-                                              0, add);
-       if (rc == ECORE_SUCCESS) {
-               if (add) {
-                       fdir->rx_queue = fdir_filter->action.rx_queue;
-                       fdir->pkt_len = pkt_len;
-                       fdir->mz = mz;
-                       SLIST_INSERT_HEAD(&qdev->fdir_info.fdir_list_head,
-                                         fdir, list);
-                       qdev->fdir_info.filter_count++;
-                       DP_INFO(edev, "flowdir filter added, count = %d\n",
-                               qdev->fdir_info.filter_count);
-               } else {
-                       rte_memzone_free(tmp->mz);
-                       SLIST_REMOVE(&qdev->fdir_info.fdir_list_head, tmp,
-                                    qede_fdir_entry, list);
-                       rte_free(tmp); /* the node deleted */
-                       rte_memzone_free(mz); /* temp node allocated */
-                       qdev->fdir_info.filter_count--;
-                       DP_INFO(edev, "Fdir filter deleted, count = %d\n",
-                               qdev->fdir_info.filter_count);
-               }
-       } else {
-               DP_ERR(edev, "flowdir filter failed, rc=%d filter_count=%d\n",
-                      rc, qdev->fdir_info.filter_count);
-       }
-
-       /* Disable ARFS searcher if there are no more filters */
-       if (qdev->fdir_info.filter_count == 0) {
-               memset(&qdev->fdir_info.arfs, 0,
-                      sizeof(struct ecore_arfs_config_params));
-               DP_INFO(edev, "Disabling flowdir\n");
-               qdev->fdir_info.arfs.arfs_enable = false;
-               ecore_arfs_mode_configure(p_hwfn, p_hwfn->p_arfs_ptt,
-                                         &qdev->fdir_info.arfs);
-       }
-       return 0;
-
-err2:
-       rte_memzone_free(mz);
-err1:
-       if (add)
-               rte_free(fdir);
-       return rc;
-}
-
-static int
-qede_fdir_filter_add(struct rte_eth_dev *eth_dev,
-                    struct rte_eth_fdir_filter *fdir,
-                    bool add)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-
-       if (!QEDE_VALID_FLOW(fdir->input.flow_type)) {
-               DP_ERR(edev, "invalid flow_type input\n");
-               return -EINVAL;
-       }
-
-       if (fdir->action.rx_queue >= QEDE_RSS_COUNT(qdev)) {
-               DP_ERR(edev, "invalid queue number %u\n",
-                      fdir->action.rx_queue);
-               return -EINVAL;
-       }
-
-       if (fdir->input.flow_ext.is_vf) {
-               DP_ERR(edev, "flowdir is not supported over VF\n");
-               return -EINVAL;
-       }
-
-       return qede_config_cmn_fdir_filter(eth_dev, fdir, add);
-}
-
-/* Fills the L3/L4 headers and returns the actual length  of flowdir packet */
-uint16_t
-qede_fdir_construct_pkt(struct rte_eth_dev *eth_dev,
-                       struct rte_eth_fdir_filter *fdir,
-                       void *buff,
-                       struct ecore_arfs_config_params *params)
-
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       uint16_t *ether_type;
-       uint8_t *raw_pkt;
-       struct rte_eth_fdir_input *input;
-       static uint8_t vlan_frame[] = {0x81, 0, 0, 0};
-       struct ipv4_hdr *ip;
-       struct ipv6_hdr *ip6;
-       struct udp_hdr *udp;
-       struct tcp_hdr *tcp;
-       uint16_t len;
-       static const uint8_t next_proto[] = {
-               [RTE_ETH_FLOW_NONFRAG_IPV4_TCP] = IPPROTO_TCP,
-               [RTE_ETH_FLOW_NONFRAG_IPV4_UDP] = IPPROTO_UDP,
-               [RTE_ETH_FLOW_NONFRAG_IPV6_TCP] = IPPROTO_TCP,
-               [RTE_ETH_FLOW_NONFRAG_IPV6_UDP] = IPPROTO_UDP,
-       };
-       raw_pkt = (uint8_t *)buff;
-       input = &fdir->input;
-       DP_INFO(edev, "flow_type %d\n", input->flow_type);
-
-       len =  2 * sizeof(struct ether_addr);
-       raw_pkt += 2 * sizeof(struct ether_addr);
-       if (input->flow_ext.vlan_tci) {
-               DP_INFO(edev, "adding VLAN header\n");
-               rte_memcpy(raw_pkt, vlan_frame, sizeof(vlan_frame));
-               rte_memcpy(raw_pkt + sizeof(uint16_t),
-                          &input->flow_ext.vlan_tci,
-                          sizeof(uint16_t));
-               raw_pkt += sizeof(vlan_frame);
-               len += sizeof(vlan_frame);
-       }
-       ether_type = (uint16_t *)raw_pkt;
-       raw_pkt += sizeof(uint16_t);
-       len += sizeof(uint16_t);
-
-       switch (input->flow_type) {
-       case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
-       case RTE_ETH_FLOW_NONFRAG_IPV4_UDP:
-               /* fill the common ip header */
-               ip = (struct ipv4_hdr *)raw_pkt;
-               *ether_type = rte_cpu_to_be_16(ETHER_TYPE_IPv4);
-               ip->version_ihl = QEDE_FDIR_IP_DEFAULT_VERSION_IHL;
-               ip->total_length = sizeof(struct ipv4_hdr);
-               ip->next_proto_id = input->flow.ip4_flow.proto ?
-                                   input->flow.ip4_flow.proto :
-                                   next_proto[input->flow_type];
-               ip->time_to_live = input->flow.ip4_flow.ttl ?
-                                  input->flow.ip4_flow.ttl :
-                                  QEDE_FDIR_IPV4_DEF_TTL;
-               ip->type_of_service = input->flow.ip4_flow.tos;
-               ip->dst_addr = input->flow.ip4_flow.dst_ip;
-               ip->src_addr = input->flow.ip4_flow.src_ip;
-               len += sizeof(struct ipv4_hdr);
-               params->ipv4 = true;
-
-               raw_pkt = (uint8_t *)buff;
-               /* UDP */
-               if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_UDP) {
-                       udp = (struct udp_hdr *)(raw_pkt + len);
-                       udp->dst_port = input->flow.udp4_flow.dst_port;
-                       udp->src_port = input->flow.udp4_flow.src_port;
-                       udp->dgram_len = sizeof(struct udp_hdr);
-                       len += sizeof(struct udp_hdr);
-                       /* adjust ip total_length */
-                       ip->total_length += sizeof(struct udp_hdr);
-                       params->udp = true;
-               } else { /* TCP */
-                       tcp = (struct tcp_hdr *)(raw_pkt + len);
-                       tcp->src_port = input->flow.tcp4_flow.src_port;
-                       tcp->dst_port = input->flow.tcp4_flow.dst_port;
-                       tcp->data_off = QEDE_FDIR_TCP_DEFAULT_DATAOFF;
-                       len += sizeof(struct tcp_hdr);
-                       /* adjust ip total_length */
-                       ip->total_length += sizeof(struct tcp_hdr);
-                       params->tcp = true;
-               }
-               break;
-       case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
-       case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
-               ip6 = (struct ipv6_hdr *)raw_pkt;
-               *ether_type = rte_cpu_to_be_16(ETHER_TYPE_IPv6);
-               ip6->proto = input->flow.ipv6_flow.proto ?
-                                       input->flow.ipv6_flow.proto :
-                                       next_proto[input->flow_type];
-               rte_memcpy(&ip6->src_addr, &input->flow.ipv6_flow.dst_ip,
-                          IPV6_ADDR_LEN);
-               rte_memcpy(&ip6->dst_addr, &input->flow.ipv6_flow.src_ip,
-                          IPV6_ADDR_LEN);
-               len += sizeof(struct ipv6_hdr);
-
-               raw_pkt = (uint8_t *)buff;
-               /* UDP */
-               if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_UDP) {
-                       udp = (struct udp_hdr *)(raw_pkt + len);
-                       udp->src_port = input->flow.udp6_flow.dst_port;
-                       udp->dst_port = input->flow.udp6_flow.src_port;
-                       len += sizeof(struct udp_hdr);
-                       params->udp = true;
-               } else { /* TCP */
-                       tcp = (struct tcp_hdr *)(raw_pkt + len);
-                       tcp->src_port = input->flow.tcp4_flow.src_port;
-                       tcp->dst_port = input->flow.tcp4_flow.dst_port;
-                       tcp->data_off = QEDE_FDIR_TCP_DEFAULT_DATAOFF;
-                       len += sizeof(struct tcp_hdr);
-                       params->tcp = true;
-               }
-               break;
-       default:
-               DP_ERR(edev, "Unsupported flow_type %u\n",
-                      input->flow_type);
-               return 0;
-       }
-
-       return len;
-}
-
-int
-qede_fdir_filter_conf(struct rte_eth_dev *eth_dev,
-                     enum rte_filter_op filter_op,
-                     void *arg)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct rte_eth_fdir_filter *fdir;
-       int ret;
-
-       fdir = (struct rte_eth_fdir_filter *)arg;
-       switch (filter_op) {
-       case RTE_ETH_FILTER_NOP:
-               /* Typically used to query flowdir support */
-               if (ECORE_IS_CMT(edev)) {
-                       DP_ERR(edev, "flowdir is not supported in 100G mode\n");
-                       return -ENOTSUP;
-               }
-               return 0; /* means supported */
-       case RTE_ETH_FILTER_ADD:
-               ret = qede_fdir_filter_add(eth_dev, fdir, 1);
-       break;
-       case RTE_ETH_FILTER_DELETE:
-               ret = qede_fdir_filter_add(eth_dev, fdir, 0);
-       break;
-       case RTE_ETH_FILTER_FLUSH:
-       case RTE_ETH_FILTER_UPDATE:
-       case RTE_ETH_FILTER_INFO:
-               return -ENOTSUP;
-       break;
-       default:
-               DP_ERR(edev, "unknown operation %u", filter_op);
-               ret = -EINVAL;
-       }
-
-       return ret;
-}
-
-int qede_ntuple_filter_conf(struct rte_eth_dev *eth_dev,
-                           enum rte_filter_op filter_op,
-                           void *arg)
-{
-       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
-       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
-       struct rte_eth_ntuple_filter *ntuple;
-       struct rte_eth_fdir_filter fdir_entry;
-       struct rte_eth_tcpv4_flow *tcpv4_flow;
-       struct rte_eth_udpv4_flow *udpv4_flow;
-       bool add = false;
-
-       switch (filter_op) {
-       case RTE_ETH_FILTER_NOP:
-               /* Typically used to query fdir support */
-               if (ECORE_IS_CMT(edev)) {
-                       DP_ERR(edev, "flowdir is not supported in 100G mode\n");
-                       return -ENOTSUP;
-               }
-               return 0; /* means supported */
-       case RTE_ETH_FILTER_ADD:
-               add = true;
-       break;
-       case RTE_ETH_FILTER_DELETE:
-       break;
-       case RTE_ETH_FILTER_INFO:
-       case RTE_ETH_FILTER_GET:
-       case RTE_ETH_FILTER_UPDATE:
-       case RTE_ETH_FILTER_FLUSH:
-       case RTE_ETH_FILTER_SET:
-       case RTE_ETH_FILTER_STATS:
-       case RTE_ETH_FILTER_OP_MAX:
-               DP_ERR(edev, "Unsupported filter_op %d\n", filter_op);
-               return -ENOTSUP;
-       }
-       ntuple = (struct rte_eth_ntuple_filter *)arg;
-       /* Internally convert ntuple to fdir entry */
-       memset(&fdir_entry, 0, sizeof(fdir_entry));
-       if (ntuple->proto == IPPROTO_TCP) {
-               fdir_entry.input.flow_type = RTE_ETH_FLOW_NONFRAG_IPV4_TCP;
-               tcpv4_flow = &fdir_entry.input.flow.tcp4_flow;
-               tcpv4_flow->ip.src_ip = ntuple->src_ip;
-               tcpv4_flow->ip.dst_ip = ntuple->dst_ip;
-               tcpv4_flow->ip.proto = IPPROTO_TCP;
-               tcpv4_flow->src_port = ntuple->src_port;
-               tcpv4_flow->dst_port = ntuple->dst_port;
-       } else {
-               fdir_entry.input.flow_type = RTE_ETH_FLOW_NONFRAG_IPV4_UDP;
-               udpv4_flow = &fdir_entry.input.flow.udp4_flow;
-               udpv4_flow->ip.src_ip = ntuple->src_ip;
-               udpv4_flow->ip.dst_ip = ntuple->dst_ip;
-               udpv4_flow->ip.proto = IPPROTO_TCP;
-               udpv4_flow->src_port = ntuple->src_port;
-               udpv4_flow->dst_port = ntuple->dst_port;
-       }
-
-       fdir_entry.action.rx_queue = ntuple->queue;
-
-       return qede_config_cmn_fdir_filter(eth_dev, &fdir_entry, add);
-}
diff --git a/drivers/net/qede/qede_filter.c b/drivers/net/qede/qede_filter.c
new file mode 100644 (file)
index 0000000..b8460a0
--- /dev/null
@@ -0,0 +1,1147 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright (c) 2017 Cavium Inc.
+ * All rights reserved.
+ * www.cavium.com
+ */
+
+#include <rte_udp.h>
+#include <rte_tcp.h>
+#include <rte_sctp.h>
+#include <rte_errno.h>
+
+#include "qede_ethdev.h"
+
+/* VXLAN tunnel classification mapping */
+const struct _qede_udp_tunn_types {
+       uint16_t rte_filter_type;
+       enum ecore_filter_ucast_type qede_type;
+       enum ecore_tunn_clss qede_tunn_clss;
+       const char *string;
+} qede_tunn_types[] = {
+       {
+               ETH_TUNNEL_FILTER_OMAC,
+               ECORE_FILTER_MAC,
+               ECORE_TUNN_CLSS_MAC_VLAN,
+               "outer-mac"
+       },
+       {
+               ETH_TUNNEL_FILTER_TENID,
+               ECORE_FILTER_VNI,
+               ECORE_TUNN_CLSS_MAC_VNI,
+               "vni"
+       },
+       {
+               ETH_TUNNEL_FILTER_IMAC,
+               ECORE_FILTER_INNER_MAC,
+               ECORE_TUNN_CLSS_INNER_MAC_VLAN,
+               "inner-mac"
+       },
+       {
+               ETH_TUNNEL_FILTER_IVLAN,
+               ECORE_FILTER_INNER_VLAN,
+               ECORE_TUNN_CLSS_INNER_MAC_VLAN,
+               "inner-vlan"
+       },
+       {
+               ETH_TUNNEL_FILTER_OMAC | ETH_TUNNEL_FILTER_TENID,
+               ECORE_FILTER_MAC_VNI_PAIR,
+               ECORE_TUNN_CLSS_MAC_VNI,
+               "outer-mac and vni"
+       },
+       {
+               ETH_TUNNEL_FILTER_OMAC | ETH_TUNNEL_FILTER_IMAC,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "outer-mac and inner-mac"
+       },
+       {
+               ETH_TUNNEL_FILTER_OMAC | ETH_TUNNEL_FILTER_IVLAN,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "outer-mac and inner-vlan"
+       },
+       {
+               ETH_TUNNEL_FILTER_TENID | ETH_TUNNEL_FILTER_IMAC,
+               ECORE_FILTER_INNER_MAC_VNI_PAIR,
+               ECORE_TUNN_CLSS_INNER_MAC_VNI,
+               "vni and inner-mac",
+       },
+       {
+               ETH_TUNNEL_FILTER_TENID | ETH_TUNNEL_FILTER_IVLAN,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "vni and inner-vlan",
+       },
+       {
+               ETH_TUNNEL_FILTER_IMAC | ETH_TUNNEL_FILTER_IVLAN,
+               ECORE_FILTER_INNER_PAIR,
+               ECORE_TUNN_CLSS_INNER_MAC_VLAN,
+               "inner-mac and inner-vlan",
+       },
+       {
+               ETH_TUNNEL_FILTER_OIP,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "outer-IP"
+       },
+       {
+               ETH_TUNNEL_FILTER_IIP,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "inner-IP"
+       },
+       {
+               RTE_TUNNEL_FILTER_IMAC_IVLAN,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "IMAC_IVLAN"
+       },
+       {
+               RTE_TUNNEL_FILTER_IMAC_IVLAN_TENID,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "IMAC_IVLAN_TENID"
+       },
+       {
+               RTE_TUNNEL_FILTER_IMAC_TENID,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "IMAC_TENID"
+       },
+       {
+               RTE_TUNNEL_FILTER_OMAC_TENID_IMAC,
+               ECORE_FILTER_UNUSED,
+               MAX_ECORE_TUNN_CLSS,
+               "OMAC_TENID_IMAC"
+       },
+};
+
+#define IP_VERSION                             (0x40)
+#define IP_HDRLEN                              (0x5)
+#define QEDE_FDIR_IP_DEFAULT_VERSION_IHL       (IP_VERSION | IP_HDRLEN)
+#define QEDE_FDIR_TCP_DEFAULT_DATAOFF          (0x50)
+#define QEDE_FDIR_IPV4_DEF_TTL                 (64)
+
+/* Sum of length of header types of L2, L3, L4.
+ * L2 : ether_hdr + vlan_hdr + vxlan_hdr
+ * L3 : ipv6_hdr
+ * L4 : tcp_hdr
+ */
+#define QEDE_MAX_FDIR_PKT_LEN                  (86)
+
+#ifndef IPV6_ADDR_LEN
+#define IPV6_ADDR_LEN                          (16)
+#endif
+
+static inline bool qede_valid_flow(uint16_t flow_type)
+{
+       return  ((flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_TCP) ||
+                (flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_UDP) ||
+                (flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_TCP) ||
+                (flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_UDP));
+}
+
+/* Note: Flowdir support is only partial.
+ * For ex: drop_queue, FDIR masks, flex_conf are not supported.
+ * Parameters like pballoc/status fields are irrelevant here.
+ */
+int qede_check_fdir_support(struct rte_eth_dev *eth_dev)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct rte_fdir_conf *fdir = &eth_dev->data->dev_conf.fdir_conf;
+
+       /* check FDIR modes */
+       switch (fdir->mode) {
+       case RTE_FDIR_MODE_NONE:
+               qdev->fdir_info.arfs.arfs_enable = false;
+               DP_INFO(edev, "flowdir is disabled\n");
+       break;
+       case RTE_FDIR_MODE_PERFECT:
+               if (ECORE_IS_CMT(edev)) {
+                       DP_ERR(edev, "flowdir is not supported in 100G mode\n");
+                       qdev->fdir_info.arfs.arfs_enable = false;
+                       return -ENOTSUP;
+               }
+               qdev->fdir_info.arfs.arfs_enable = true;
+               DP_INFO(edev, "flowdir is enabled\n");
+       break;
+       case RTE_FDIR_MODE_PERFECT_TUNNEL:
+       case RTE_FDIR_MODE_SIGNATURE:
+       case RTE_FDIR_MODE_PERFECT_MAC_VLAN:
+               DP_ERR(edev, "Unsupported flowdir mode %d\n", fdir->mode);
+               return -ENOTSUP;
+       }
+
+       return 0;
+}
+
+void qede_fdir_dealloc_resc(struct rte_eth_dev *eth_dev)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct qede_fdir_entry *tmp = NULL;
+
+       SLIST_FOREACH(tmp, &qdev->fdir_info.fdir_list_head, list) {
+               if (tmp) {
+                       if (tmp->mz)
+                               rte_memzone_free(tmp->mz);
+                       SLIST_REMOVE(&qdev->fdir_info.fdir_list_head, tmp,
+                                    qede_fdir_entry, list);
+                       rte_free(tmp);
+               }
+       }
+}
+
+static int
+qede_config_cmn_fdir_filter(struct rte_eth_dev *eth_dev,
+                           struct rte_eth_fdir_filter *fdir_filter,
+                           bool add)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       char mz_name[RTE_MEMZONE_NAMESIZE] = {0};
+       struct qede_fdir_entry *tmp = NULL;
+       struct qede_fdir_entry *fdir = NULL;
+       const struct rte_memzone *mz;
+       struct ecore_hwfn *p_hwfn;
+       enum _ecore_status_t rc;
+       uint16_t pkt_len;
+       void *pkt;
+
+       if (add) {
+               if (qdev->fdir_info.filter_count == QEDE_RFS_MAX_FLTR - 1) {
+                       DP_ERR(edev, "Reached max flowdir filter limit\n");
+                       return -EINVAL;
+               }
+               fdir = rte_malloc(NULL, sizeof(struct qede_fdir_entry),
+                                 RTE_CACHE_LINE_SIZE);
+               if (!fdir) {
+                       DP_ERR(edev, "Did not allocate memory for fdir\n");
+                       return -ENOMEM;
+               }
+       }
+       /* soft_id could have been used as memzone string, but soft_id is
+        * not currently used so it has no significance.
+        */
+       snprintf(mz_name, sizeof(mz_name) - 1, "%lx",
+                (unsigned long)rte_get_timer_cycles());
+       mz = rte_memzone_reserve_aligned(mz_name, QEDE_MAX_FDIR_PKT_LEN,
+                                        SOCKET_ID_ANY, 0, RTE_CACHE_LINE_SIZE);
+       if (!mz) {
+               DP_ERR(edev, "Failed to allocate memzone for fdir, err = %s\n",
+                      rte_strerror(rte_errno));
+               rc = -rte_errno;
+               goto err1;
+       }
+
+       pkt = mz->addr;
+       memset(pkt, 0, QEDE_MAX_FDIR_PKT_LEN);
+       pkt_len = qede_fdir_construct_pkt(eth_dev, fdir_filter, pkt,
+                                         &qdev->fdir_info.arfs);
+       if (pkt_len == 0) {
+               rc = -EINVAL;
+               goto err2;
+       }
+       DP_INFO(edev, "pkt_len = %u memzone = %s\n", pkt_len, mz_name);
+       if (add) {
+               SLIST_FOREACH(tmp, &qdev->fdir_info.fdir_list_head, list) {
+                       if (memcmp(tmp->mz->addr, pkt, pkt_len) == 0) {
+                               DP_INFO(edev, "flowdir filter exist\n");
+                               rc = 0;
+                               goto err2;
+                       }
+               }
+       } else {
+               SLIST_FOREACH(tmp, &qdev->fdir_info.fdir_list_head, list) {
+                       if (memcmp(tmp->mz->addr, pkt, pkt_len) == 0)
+                               break;
+               }
+               if (!tmp) {
+                       DP_ERR(edev, "flowdir filter does not exist\n");
+                       rc = -EEXIST;
+                       goto err2;
+               }
+       }
+       p_hwfn = ECORE_LEADING_HWFN(edev);
+       if (add) {
+               if (!qdev->fdir_info.arfs.arfs_enable) {
+                       /* Force update */
+                       eth_dev->data->dev_conf.fdir_conf.mode =
+                                               RTE_FDIR_MODE_PERFECT;
+                       qdev->fdir_info.arfs.arfs_enable = true;
+                       DP_INFO(edev, "Force enable flowdir in perfect mode\n");
+               }
+               /* Enable ARFS searcher with updated flow_types */
+               ecore_arfs_mode_configure(p_hwfn, p_hwfn->p_arfs_ptt,
+                                         &qdev->fdir_info.arfs);
+       }
+       /* configure filter with ECORE_SPQ_MODE_EBLOCK */
+       rc = ecore_configure_rfs_ntuple_filter(p_hwfn, NULL,
+                                              (dma_addr_t)mz->iova,
+                                              pkt_len,
+                                              fdir_filter->action.rx_queue,
+                                              0, add);
+       if (rc == ECORE_SUCCESS) {
+               if (add) {
+                       fdir->rx_queue = fdir_filter->action.rx_queue;
+                       fdir->pkt_len = pkt_len;
+                       fdir->mz = mz;
+                       SLIST_INSERT_HEAD(&qdev->fdir_info.fdir_list_head,
+                                         fdir, list);
+                       qdev->fdir_info.filter_count++;
+                       DP_INFO(edev, "flowdir filter added, count = %d\n",
+                               qdev->fdir_info.filter_count);
+               } else {
+                       rte_memzone_free(tmp->mz);
+                       SLIST_REMOVE(&qdev->fdir_info.fdir_list_head, tmp,
+                                    qede_fdir_entry, list);
+                       rte_free(tmp); /* the node deleted */
+                       rte_memzone_free(mz); /* temp node allocated */
+                       qdev->fdir_info.filter_count--;
+                       DP_INFO(edev, "Fdir filter deleted, count = %d\n",
+                               qdev->fdir_info.filter_count);
+               }
+       } else {
+               DP_ERR(edev, "flowdir filter failed, rc=%d filter_count=%d\n",
+                      rc, qdev->fdir_info.filter_count);
+       }
+
+       /* Disable ARFS searcher if there are no more filters */
+       if (qdev->fdir_info.filter_count == 0) {
+               memset(&qdev->fdir_info.arfs, 0,
+                      sizeof(struct ecore_arfs_config_params));
+               DP_INFO(edev, "Disabling flowdir\n");
+               qdev->fdir_info.arfs.arfs_enable = false;
+               ecore_arfs_mode_configure(p_hwfn, p_hwfn->p_arfs_ptt,
+                                         &qdev->fdir_info.arfs);
+       }
+       return 0;
+
+err2:
+       rte_memzone_free(mz);
+err1:
+       if (add)
+               rte_free(fdir);
+       return rc;
+}
+
+static int
+qede_fdir_filter_add(struct rte_eth_dev *eth_dev,
+                    struct rte_eth_fdir_filter *fdir,
+                    bool add)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+
+       if (!qede_valid_flow(fdir->input.flow_type)) {
+               DP_ERR(edev, "invalid flow_type input\n");
+               return -EINVAL;
+       }
+
+       if (fdir->action.rx_queue >= QEDE_RSS_COUNT(qdev)) {
+               DP_ERR(edev, "invalid queue number %u\n",
+                      fdir->action.rx_queue);
+               return -EINVAL;
+       }
+
+       if (fdir->input.flow_ext.is_vf) {
+               DP_ERR(edev, "flowdir is not supported over VF\n");
+               return -EINVAL;
+       }
+
+       return qede_config_cmn_fdir_filter(eth_dev, fdir, add);
+}
+
+/* Fills the L3/L4 headers and returns the actual length  of flowdir packet */
+uint16_t
+qede_fdir_construct_pkt(struct rte_eth_dev *eth_dev,
+                       struct rte_eth_fdir_filter *fdir,
+                       void *buff,
+                       struct ecore_arfs_config_params *params)
+
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       uint16_t *ether_type;
+       uint8_t *raw_pkt;
+       struct rte_eth_fdir_input *input;
+       static uint8_t vlan_frame[] = {0x81, 0, 0, 0};
+       struct ipv4_hdr *ip;
+       struct ipv6_hdr *ip6;
+       struct udp_hdr *udp;
+       struct tcp_hdr *tcp;
+       uint16_t len;
+       static const uint8_t next_proto[] = {
+               [RTE_ETH_FLOW_NONFRAG_IPV4_TCP] = IPPROTO_TCP,
+               [RTE_ETH_FLOW_NONFRAG_IPV4_UDP] = IPPROTO_UDP,
+               [RTE_ETH_FLOW_NONFRAG_IPV6_TCP] = IPPROTO_TCP,
+               [RTE_ETH_FLOW_NONFRAG_IPV6_UDP] = IPPROTO_UDP,
+       };
+       raw_pkt = (uint8_t *)buff;
+       input = &fdir->input;
+       DP_INFO(edev, "flow_type %d\n", input->flow_type);
+
+       len =  2 * sizeof(struct ether_addr);
+       raw_pkt += 2 * sizeof(struct ether_addr);
+       if (input->flow_ext.vlan_tci) {
+               DP_INFO(edev, "adding VLAN header\n");
+               rte_memcpy(raw_pkt, vlan_frame, sizeof(vlan_frame));
+               rte_memcpy(raw_pkt + sizeof(uint16_t),
+                          &input->flow_ext.vlan_tci,
+                          sizeof(uint16_t));
+               raw_pkt += sizeof(vlan_frame);
+               len += sizeof(vlan_frame);
+       }
+       ether_type = (uint16_t *)raw_pkt;
+       raw_pkt += sizeof(uint16_t);
+       len += sizeof(uint16_t);
+
+       switch (input->flow_type) {
+       case RTE_ETH_FLOW_NONFRAG_IPV4_TCP:
+       case RTE_ETH_FLOW_NONFRAG_IPV4_UDP:
+               /* fill the common ip header */
+               ip = (struct ipv4_hdr *)raw_pkt;
+               *ether_type = rte_cpu_to_be_16(ETHER_TYPE_IPv4);
+               ip->version_ihl = QEDE_FDIR_IP_DEFAULT_VERSION_IHL;
+               ip->total_length = sizeof(struct ipv4_hdr);
+               ip->next_proto_id = input->flow.ip4_flow.proto ?
+                                   input->flow.ip4_flow.proto :
+                                   next_proto[input->flow_type];
+               ip->time_to_live = input->flow.ip4_flow.ttl ?
+                                  input->flow.ip4_flow.ttl :
+                                  QEDE_FDIR_IPV4_DEF_TTL;
+               ip->type_of_service = input->flow.ip4_flow.tos;
+               ip->dst_addr = input->flow.ip4_flow.dst_ip;
+               ip->src_addr = input->flow.ip4_flow.src_ip;
+               len += sizeof(struct ipv4_hdr);
+               params->ipv4 = true;
+
+               raw_pkt = (uint8_t *)buff;
+               /* UDP */
+               if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_UDP) {
+                       udp = (struct udp_hdr *)(raw_pkt + len);
+                       udp->dst_port = input->flow.udp4_flow.dst_port;
+                       udp->src_port = input->flow.udp4_flow.src_port;
+                       udp->dgram_len = sizeof(struct udp_hdr);
+                       len += sizeof(struct udp_hdr);
+                       /* adjust ip total_length */
+                       ip->total_length += sizeof(struct udp_hdr);
+                       params->udp = true;
+               } else { /* TCP */
+                       tcp = (struct tcp_hdr *)(raw_pkt + len);
+                       tcp->src_port = input->flow.tcp4_flow.src_port;
+                       tcp->dst_port = input->flow.tcp4_flow.dst_port;
+                       tcp->data_off = QEDE_FDIR_TCP_DEFAULT_DATAOFF;
+                       len += sizeof(struct tcp_hdr);
+                       /* adjust ip total_length */
+                       ip->total_length += sizeof(struct tcp_hdr);
+                       params->tcp = true;
+               }
+               break;
+       case RTE_ETH_FLOW_NONFRAG_IPV6_TCP:
+       case RTE_ETH_FLOW_NONFRAG_IPV6_UDP:
+               ip6 = (struct ipv6_hdr *)raw_pkt;
+               *ether_type = rte_cpu_to_be_16(ETHER_TYPE_IPv6);
+               ip6->proto = input->flow.ipv6_flow.proto ?
+                                       input->flow.ipv6_flow.proto :
+                                       next_proto[input->flow_type];
+               rte_memcpy(&ip6->src_addr, &input->flow.ipv6_flow.dst_ip,
+                          IPV6_ADDR_LEN);
+               rte_memcpy(&ip6->dst_addr, &input->flow.ipv6_flow.src_ip,
+                          IPV6_ADDR_LEN);
+               len += sizeof(struct ipv6_hdr);
+
+               raw_pkt = (uint8_t *)buff;
+               /* UDP */
+               if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_UDP) {
+                       udp = (struct udp_hdr *)(raw_pkt + len);
+                       udp->src_port = input->flow.udp6_flow.dst_port;
+                       udp->dst_port = input->flow.udp6_flow.src_port;
+                       len += sizeof(struct udp_hdr);
+                       params->udp = true;
+               } else { /* TCP */
+                       tcp = (struct tcp_hdr *)(raw_pkt + len);
+                       tcp->src_port = input->flow.tcp4_flow.src_port;
+                       tcp->dst_port = input->flow.tcp4_flow.dst_port;
+                       tcp->data_off = QEDE_FDIR_TCP_DEFAULT_DATAOFF;
+                       len += sizeof(struct tcp_hdr);
+                       params->tcp = true;
+               }
+               break;
+       default:
+               DP_ERR(edev, "Unsupported flow_type %u\n",
+                      input->flow_type);
+               return 0;
+       }
+
+       return len;
+}
+
+static int
+qede_fdir_filter_conf(struct rte_eth_dev *eth_dev,
+                     enum rte_filter_op filter_op,
+                     void *arg)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct rte_eth_fdir_filter *fdir;
+       int ret;
+
+       fdir = (struct rte_eth_fdir_filter *)arg;
+       switch (filter_op) {
+       case RTE_ETH_FILTER_NOP:
+               /* Typically used to query flowdir support */
+               if (ECORE_IS_CMT(edev)) {
+                       DP_ERR(edev, "flowdir is not supported in 100G mode\n");
+                       return -ENOTSUP;
+               }
+               return 0; /* means supported */
+       case RTE_ETH_FILTER_ADD:
+               ret = qede_fdir_filter_add(eth_dev, fdir, 1);
+       break;
+       case RTE_ETH_FILTER_DELETE:
+               ret = qede_fdir_filter_add(eth_dev, fdir, 0);
+       break;
+       case RTE_ETH_FILTER_FLUSH:
+       case RTE_ETH_FILTER_UPDATE:
+       case RTE_ETH_FILTER_INFO:
+               return -ENOTSUP;
+       break;
+       default:
+               DP_ERR(edev, "unknown operation %u", filter_op);
+               ret = -EINVAL;
+       }
+
+       return ret;
+}
+
+int qede_ntuple_filter_conf(struct rte_eth_dev *eth_dev,
+                           enum rte_filter_op filter_op,
+                           void *arg)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct rte_eth_ntuple_filter *ntuple;
+       struct rte_eth_fdir_filter fdir_entry;
+       struct rte_eth_tcpv4_flow *tcpv4_flow;
+       struct rte_eth_udpv4_flow *udpv4_flow;
+       bool add = false;
+
+       switch (filter_op) {
+       case RTE_ETH_FILTER_NOP:
+               /* Typically used to query fdir support */
+               if (ECORE_IS_CMT(edev)) {
+                       DP_ERR(edev, "flowdir is not supported in 100G mode\n");
+                       return -ENOTSUP;
+               }
+               return 0; /* means supported */
+       case RTE_ETH_FILTER_ADD:
+               add = true;
+       break;
+       case RTE_ETH_FILTER_DELETE:
+       break;
+       case RTE_ETH_FILTER_INFO:
+       case RTE_ETH_FILTER_GET:
+       case RTE_ETH_FILTER_UPDATE:
+       case RTE_ETH_FILTER_FLUSH:
+       case RTE_ETH_FILTER_SET:
+       case RTE_ETH_FILTER_STATS:
+       case RTE_ETH_FILTER_OP_MAX:
+               DP_ERR(edev, "Unsupported filter_op %d\n", filter_op);
+               return -ENOTSUP;
+       }
+       ntuple = (struct rte_eth_ntuple_filter *)arg;
+       /* Internally convert ntuple to fdir entry */
+       memset(&fdir_entry, 0, sizeof(fdir_entry));
+       if (ntuple->proto == IPPROTO_TCP) {
+               fdir_entry.input.flow_type = RTE_ETH_FLOW_NONFRAG_IPV4_TCP;
+               tcpv4_flow = &fdir_entry.input.flow.tcp4_flow;
+               tcpv4_flow->ip.src_ip = ntuple->src_ip;
+               tcpv4_flow->ip.dst_ip = ntuple->dst_ip;
+               tcpv4_flow->ip.proto = IPPROTO_TCP;
+               tcpv4_flow->src_port = ntuple->src_port;
+               tcpv4_flow->dst_port = ntuple->dst_port;
+       } else {
+               fdir_entry.input.flow_type = RTE_ETH_FLOW_NONFRAG_IPV4_UDP;
+               udpv4_flow = &fdir_entry.input.flow.udp4_flow;
+               udpv4_flow->ip.src_ip = ntuple->src_ip;
+               udpv4_flow->ip.dst_ip = ntuple->dst_ip;
+               udpv4_flow->ip.proto = IPPROTO_TCP;
+               udpv4_flow->src_port = ntuple->src_port;
+               udpv4_flow->dst_port = ntuple->dst_port;
+       }
+
+       fdir_entry.action.rx_queue = ntuple->queue;
+
+       return qede_config_cmn_fdir_filter(eth_dev, &fdir_entry, add);
+}
+
+static int
+qede_tunnel_update(struct qede_dev *qdev,
+                  struct ecore_tunnel_info *tunn_info)
+{
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       enum _ecore_status_t rc = ECORE_INVAL;
+       struct ecore_hwfn *p_hwfn;
+       struct ecore_ptt *p_ptt;
+       int i;
+
+       for_each_hwfn(edev, i) {
+               p_hwfn = &edev->hwfns[i];
+               if (IS_PF(edev)) {
+                       p_ptt = ecore_ptt_acquire(p_hwfn);
+                       if (!p_ptt) {
+                               DP_ERR(p_hwfn, "Can't acquire PTT\n");
+                               return -EAGAIN;
+                       }
+               } else {
+                       p_ptt = NULL;
+               }
+
+               rc = ecore_sp_pf_update_tunn_cfg(p_hwfn, p_ptt,
+                               tunn_info, ECORE_SPQ_MODE_CB, NULL);
+               if (IS_PF(edev))
+                       ecore_ptt_release(p_hwfn, p_ptt);
+
+               if (rc != ECORE_SUCCESS)
+                       break;
+       }
+
+       return rc;
+}
+
+static int
+qede_vxlan_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
+                 bool enable)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       enum _ecore_status_t rc = ECORE_INVAL;
+       struct ecore_tunnel_info tunn;
+
+       if (qdev->vxlan.enable == enable)
+               return ECORE_SUCCESS;
+
+       memset(&tunn, 0, sizeof(struct ecore_tunnel_info));
+       tunn.vxlan.b_update_mode = true;
+       tunn.vxlan.b_mode_enabled = enable;
+       tunn.b_update_rx_cls = true;
+       tunn.b_update_tx_cls = true;
+       tunn.vxlan.tun_cls = clss;
+
+       tunn.vxlan_port.b_update_port = true;
+       tunn.vxlan_port.port = enable ? QEDE_VXLAN_DEF_PORT : 0;
+
+       rc = qede_tunnel_update(qdev, &tunn);
+       if (rc == ECORE_SUCCESS) {
+               qdev->vxlan.enable = enable;
+               qdev->vxlan.udp_port = (enable) ? QEDE_VXLAN_DEF_PORT : 0;
+               DP_INFO(edev, "vxlan is %s, UDP port = %d\n",
+                       enable ? "enabled" : "disabled", qdev->vxlan.udp_port);
+       } else {
+               DP_ERR(edev, "Failed to update tunn_clss %u\n",
+                      tunn.vxlan.tun_cls);
+       }
+
+       return rc;
+}
+
+static int
+qede_geneve_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
+                 bool enable)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       enum _ecore_status_t rc = ECORE_INVAL;
+       struct ecore_tunnel_info tunn;
+
+       memset(&tunn, 0, sizeof(struct ecore_tunnel_info));
+       tunn.l2_geneve.b_update_mode = true;
+       tunn.l2_geneve.b_mode_enabled = enable;
+       tunn.ip_geneve.b_update_mode = true;
+       tunn.ip_geneve.b_mode_enabled = enable;
+       tunn.l2_geneve.tun_cls = clss;
+       tunn.ip_geneve.tun_cls = clss;
+       tunn.b_update_rx_cls = true;
+       tunn.b_update_tx_cls = true;
+
+       tunn.geneve_port.b_update_port = true;
+       tunn.geneve_port.port = enable ? QEDE_GENEVE_DEF_PORT : 0;
+
+       rc = qede_tunnel_update(qdev, &tunn);
+       if (rc == ECORE_SUCCESS) {
+               qdev->geneve.enable = enable;
+               qdev->geneve.udp_port = (enable) ? QEDE_GENEVE_DEF_PORT : 0;
+               DP_INFO(edev, "GENEVE is %s, UDP port = %d\n",
+                       enable ? "enabled" : "disabled", qdev->geneve.udp_port);
+       } else {
+               DP_ERR(edev, "Failed to update tunn_clss %u\n",
+                      clss);
+       }
+
+       return rc;
+}
+
+static int
+qede_ipgre_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
+                 bool enable)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       enum _ecore_status_t rc = ECORE_INVAL;
+       struct ecore_tunnel_info tunn;
+
+       memset(&tunn, 0, sizeof(struct ecore_tunnel_info));
+       tunn.ip_gre.b_update_mode = true;
+       tunn.ip_gre.b_mode_enabled = enable;
+       tunn.ip_gre.tun_cls = clss;
+       tunn.ip_gre.tun_cls = clss;
+       tunn.b_update_rx_cls = true;
+       tunn.b_update_tx_cls = true;
+
+       rc = qede_tunnel_update(qdev, &tunn);
+       if (rc == ECORE_SUCCESS) {
+               qdev->ipgre.enable = enable;
+               DP_INFO(edev, "IPGRE is %s\n",
+                       enable ? "enabled" : "disabled");
+       } else {
+               DP_ERR(edev, "Failed to update tunn_clss %u\n",
+                      clss);
+       }
+
+       return rc;
+}
+
+int
+qede_udp_dst_port_del(struct rte_eth_dev *eth_dev,
+                     struct rte_eth_udp_tunnel *tunnel_udp)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct ecore_tunnel_info tunn; /* @DPDK */
+       uint16_t udp_port;
+       int rc;
+
+       PMD_INIT_FUNC_TRACE(edev);
+
+       memset(&tunn, 0, sizeof(tunn));
+
+       switch (tunnel_udp->prot_type) {
+       case RTE_TUNNEL_TYPE_VXLAN:
+               if (qdev->vxlan.udp_port != tunnel_udp->udp_port) {
+                       DP_ERR(edev, "UDP port %u doesn't exist\n",
+                               tunnel_udp->udp_port);
+                       return ECORE_INVAL;
+               }
+               udp_port = 0;
+
+               tunn.vxlan_port.b_update_port = true;
+               tunn.vxlan_port.port = udp_port;
+
+               rc = qede_tunnel_update(qdev, &tunn);
+               if (rc != ECORE_SUCCESS) {
+                       DP_ERR(edev, "Unable to config UDP port %u\n",
+                              tunn.vxlan_port.port);
+                       return rc;
+               }
+
+               qdev->vxlan.udp_port = udp_port;
+               /* If the request is to delete UDP port and if the number of
+                * VXLAN filters have reached 0 then VxLAN offload can be be
+                * disabled.
+                */
+               if (qdev->vxlan.enable && qdev->vxlan.num_filters == 0)
+                       return qede_vxlan_enable(eth_dev,
+                                       ECORE_TUNN_CLSS_MAC_VLAN, false);
+
+               break;
+       case RTE_TUNNEL_TYPE_GENEVE:
+               if (qdev->geneve.udp_port != tunnel_udp->udp_port) {
+                       DP_ERR(edev, "UDP port %u doesn't exist\n",
+                               tunnel_udp->udp_port);
+                       return ECORE_INVAL;
+               }
+
+               udp_port = 0;
+
+               tunn.geneve_port.b_update_port = true;
+               tunn.geneve_port.port = udp_port;
+
+               rc = qede_tunnel_update(qdev, &tunn);
+               if (rc != ECORE_SUCCESS) {
+                       DP_ERR(edev, "Unable to config UDP port %u\n",
+                              tunn.vxlan_port.port);
+                       return rc;
+               }
+
+               qdev->vxlan.udp_port = udp_port;
+               /* If the request is to delete UDP port and if the number of
+                * GENEVE filters have reached 0 then GENEVE offload can be be
+                * disabled.
+                */
+               if (qdev->geneve.enable && qdev->geneve.num_filters == 0)
+                       return qede_geneve_enable(eth_dev,
+                                       ECORE_TUNN_CLSS_MAC_VLAN, false);
+
+               break;
+
+       default:
+               return ECORE_INVAL;
+       }
+
+       return 0;
+}
+
+int
+qede_udp_dst_port_add(struct rte_eth_dev *eth_dev,
+                     struct rte_eth_udp_tunnel *tunnel_udp)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct ecore_tunnel_info tunn; /* @DPDK */
+       uint16_t udp_port;
+       int rc;
+
+       PMD_INIT_FUNC_TRACE(edev);
+
+       memset(&tunn, 0, sizeof(tunn));
+
+       switch (tunnel_udp->prot_type) {
+       case RTE_TUNNEL_TYPE_VXLAN:
+               if (qdev->vxlan.udp_port == tunnel_udp->udp_port) {
+                       DP_INFO(edev,
+                               "UDP port %u for VXLAN was already configured\n",
+                               tunnel_udp->udp_port);
+                       return ECORE_SUCCESS;
+               }
+
+               /* Enable VxLAN tunnel with default MAC/VLAN classification if
+                * it was not enabled while adding VXLAN filter before UDP port
+                * update.
+                */
+               if (!qdev->vxlan.enable) {
+                       rc = qede_vxlan_enable(eth_dev,
+                               ECORE_TUNN_CLSS_MAC_VLAN, true);
+                       if (rc != ECORE_SUCCESS) {
+                               DP_ERR(edev, "Failed to enable VXLAN "
+                                       "prior to updating UDP port\n");
+                               return rc;
+                       }
+               }
+               udp_port = tunnel_udp->udp_port;
+
+               tunn.vxlan_port.b_update_port = true;
+               tunn.vxlan_port.port = udp_port;
+
+               rc = qede_tunnel_update(qdev, &tunn);
+               if (rc != ECORE_SUCCESS) {
+                       DP_ERR(edev, "Unable to config UDP port %u for VXLAN\n",
+                              udp_port);
+                       return rc;
+               }
+
+               DP_INFO(edev, "Updated UDP port %u for VXLAN\n", udp_port);
+
+               qdev->vxlan.udp_port = udp_port;
+               break;
+       case RTE_TUNNEL_TYPE_GENEVE:
+               if (qdev->geneve.udp_port == tunnel_udp->udp_port) {
+                       DP_INFO(edev,
+                               "UDP port %u for GENEVE was already configured\n",
+                               tunnel_udp->udp_port);
+                       return ECORE_SUCCESS;
+               }
+
+               /* Enable GENEVE tunnel with default MAC/VLAN classification if
+                * it was not enabled while adding GENEVE filter before UDP port
+                * update.
+                */
+               if (!qdev->geneve.enable) {
+                       rc = qede_geneve_enable(eth_dev,
+                               ECORE_TUNN_CLSS_MAC_VLAN, true);
+                       if (rc != ECORE_SUCCESS) {
+                               DP_ERR(edev, "Failed to enable GENEVE "
+                                       "prior to updating UDP port\n");
+                               return rc;
+                       }
+               }
+               udp_port = tunnel_udp->udp_port;
+
+               tunn.geneve_port.b_update_port = true;
+               tunn.geneve_port.port = udp_port;
+
+               rc = qede_tunnel_update(qdev, &tunn);
+               if (rc != ECORE_SUCCESS) {
+                       DP_ERR(edev, "Unable to config UDP port %u for GENEVE\n",
+                              udp_port);
+                       return rc;
+               }
+
+               DP_INFO(edev, "Updated UDP port %u for GENEVE\n", udp_port);
+
+               qdev->geneve.udp_port = udp_port;
+               break;
+       default:
+               return ECORE_INVAL;
+       }
+
+       return 0;
+}
+
+static void qede_get_ecore_tunn_params(uint32_t filter, uint32_t *type,
+                                      uint32_t *clss, char *str)
+{
+       uint16_t j;
+       *clss = MAX_ECORE_TUNN_CLSS;
+
+       for (j = 0; j < RTE_DIM(qede_tunn_types); j++) {
+               if (filter == qede_tunn_types[j].rte_filter_type) {
+                       *type = qede_tunn_types[j].qede_type;
+                       *clss = qede_tunn_types[j].qede_tunn_clss;
+                       strcpy(str, qede_tunn_types[j].string);
+                       return;
+               }
+       }
+}
+
+static int
+qede_set_ucast_tunn_cmn_param(struct ecore_filter_ucast *ucast,
+                             const struct rte_eth_tunnel_filter_conf *conf,
+                             uint32_t type)
+{
+       /* Init commmon ucast params first */
+       qede_set_ucast_cmn_params(ucast);
+
+       /* Copy out the required fields based on classification type */
+       ucast->type = type;
+
+       switch (type) {
+       case ECORE_FILTER_VNI:
+               ucast->vni = conf->tenant_id;
+       break;
+       case ECORE_FILTER_INNER_VLAN:
+               ucast->vlan = conf->inner_vlan;
+       break;
+       case ECORE_FILTER_MAC:
+               memcpy(ucast->mac, conf->outer_mac.addr_bytes,
+                      ETHER_ADDR_LEN);
+       break;
+       case ECORE_FILTER_INNER_MAC:
+               memcpy(ucast->mac, conf->inner_mac.addr_bytes,
+                      ETHER_ADDR_LEN);
+       break;
+       case ECORE_FILTER_MAC_VNI_PAIR:
+               memcpy(ucast->mac, conf->outer_mac.addr_bytes,
+                       ETHER_ADDR_LEN);
+               ucast->vni = conf->tenant_id;
+       break;
+       case ECORE_FILTER_INNER_MAC_VNI_PAIR:
+               memcpy(ucast->mac, conf->inner_mac.addr_bytes,
+                       ETHER_ADDR_LEN);
+               ucast->vni = conf->tenant_id;
+       break;
+       case ECORE_FILTER_INNER_PAIR:
+               memcpy(ucast->mac, conf->inner_mac.addr_bytes,
+                       ETHER_ADDR_LEN);
+               ucast->vlan = conf->inner_vlan;
+       break;
+       default:
+               return -EINVAL;
+       }
+
+       return ECORE_SUCCESS;
+}
+
+static int
+_qede_tunn_filter_config(struct rte_eth_dev *eth_dev,
+                        const struct rte_eth_tunnel_filter_conf *conf,
+                        __attribute__((unused)) enum rte_filter_op filter_op,
+                        enum ecore_tunn_clss *clss,
+                        bool add)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct ecore_filter_ucast ucast = {0};
+       enum ecore_filter_ucast_type type;
+       uint16_t filter_type = 0;
+       char str[80];
+       int rc;
+
+       filter_type = conf->filter_type;
+       /* Determine if the given filter classification is supported */
+       qede_get_ecore_tunn_params(filter_type, &type, clss, str);
+       if (*clss == MAX_ECORE_TUNN_CLSS) {
+               DP_ERR(edev, "Unsupported filter type\n");
+               return -EINVAL;
+       }
+       /* Init tunnel ucast params */
+       rc = qede_set_ucast_tunn_cmn_param(&ucast, conf, type);
+       if (rc != ECORE_SUCCESS) {
+               DP_ERR(edev, "Unsupported Tunnel filter type 0x%x\n",
+               conf->filter_type);
+               return rc;
+       }
+       DP_INFO(edev, "Rule: \"%s\", op %d, type 0x%x\n",
+               str, filter_op, ucast.type);
+
+       ucast.opcode = add ? ECORE_FILTER_ADD : ECORE_FILTER_REMOVE;
+
+       /* Skip MAC/VLAN if filter is based on VNI */
+       if (!(filter_type & ETH_TUNNEL_FILTER_TENID)) {
+               rc = qede_mac_int_ops(eth_dev, &ucast, add);
+               if (rc == 0 && add) {
+                       /* Enable accept anyvlan */
+                       qede_config_accept_any_vlan(qdev, true);
+               }
+       } else {
+               rc = qede_ucast_filter(eth_dev, &ucast, add);
+               if (rc == 0)
+                       rc = ecore_filter_ucast_cmd(edev, &ucast,
+                                           ECORE_SPQ_MODE_CB, NULL);
+       }
+
+       return rc;
+}
+
+static int
+qede_tunn_enable(struct rte_eth_dev *eth_dev, uint8_t clss,
+                enum rte_eth_tunnel_type tunn_type, bool enable)
+{
+       int rc = -EINVAL;
+
+       switch (tunn_type) {
+       case RTE_TUNNEL_TYPE_VXLAN:
+               rc = qede_vxlan_enable(eth_dev, clss, enable);
+               break;
+       case RTE_TUNNEL_TYPE_GENEVE:
+               rc = qede_geneve_enable(eth_dev, clss, enable);
+               break;
+       case RTE_TUNNEL_TYPE_IP_IN_GRE:
+               rc = qede_ipgre_enable(eth_dev, clss, enable);
+               break;
+       default:
+               rc = -EINVAL;
+               break;
+       }
+
+       return rc;
+}
+
+static int
+qede_tunn_filter_config(struct rte_eth_dev *eth_dev,
+                       enum rte_filter_op filter_op,
+                       const struct rte_eth_tunnel_filter_conf *conf)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       enum ecore_tunn_clss clss = MAX_ECORE_TUNN_CLSS;
+       bool add;
+       int rc;
+
+       PMD_INIT_FUNC_TRACE(edev);
+
+       switch (filter_op) {
+       case RTE_ETH_FILTER_ADD:
+               add = true;
+               break;
+       case RTE_ETH_FILTER_DELETE:
+               add = false;
+               break;
+       default:
+               DP_ERR(edev, "Unsupported operation %d\n", filter_op);
+               return -EINVAL;
+       }
+
+       if (IS_VF(edev))
+               return qede_tunn_enable(eth_dev,
+                                       ECORE_TUNN_CLSS_MAC_VLAN,
+                                       conf->tunnel_type, add);
+
+       rc = _qede_tunn_filter_config(eth_dev, conf, filter_op, &clss, add);
+       if (rc != ECORE_SUCCESS)
+               return rc;
+
+       if (add) {
+               if (conf->tunnel_type == RTE_TUNNEL_TYPE_VXLAN) {
+                       qdev->vxlan.num_filters++;
+                       qdev->vxlan.filter_type = conf->filter_type;
+               } else { /* GENEVE */
+                       qdev->geneve.num_filters++;
+                       qdev->geneve.filter_type = conf->filter_type;
+               }
+
+               if (!qdev->vxlan.enable || !qdev->geneve.enable ||
+                   !qdev->ipgre.enable)
+                       return qede_tunn_enable(eth_dev, clss,
+                                               conf->tunnel_type,
+                                               true);
+       } else {
+               if (conf->tunnel_type == RTE_TUNNEL_TYPE_VXLAN)
+                       qdev->vxlan.num_filters--;
+               else /*GENEVE*/
+                       qdev->geneve.num_filters--;
+
+               /* Disable VXLAN if VXLAN filters become 0 */
+               if (qdev->vxlan.num_filters == 0 ||
+                   qdev->geneve.num_filters == 0)
+                       return qede_tunn_enable(eth_dev, clss,
+                                               conf->tunnel_type,
+                                               false);
+       }
+
+       return 0;
+}
+
+int qede_dev_filter_ctrl(struct rte_eth_dev *eth_dev,
+                        enum rte_filter_type filter_type,
+                        enum rte_filter_op filter_op,
+                        void *arg)
+{
+       struct qede_dev *qdev = QEDE_INIT_QDEV(eth_dev);
+       struct ecore_dev *edev = QEDE_INIT_EDEV(qdev);
+       struct rte_eth_tunnel_filter_conf *filter_conf =
+                       (struct rte_eth_tunnel_filter_conf *)arg;
+
+       switch (filter_type) {
+       case RTE_ETH_FILTER_TUNNEL:
+               switch (filter_conf->tunnel_type) {
+               case RTE_TUNNEL_TYPE_VXLAN:
+               case RTE_TUNNEL_TYPE_GENEVE:
+               case RTE_TUNNEL_TYPE_IP_IN_GRE:
+                       DP_INFO(edev,
+                               "Packet steering to the specified Rx queue"
+                               " is not supported with UDP tunneling");
+                       return(qede_tunn_filter_config(eth_dev, filter_op,
+                                                     filter_conf));
+               case RTE_TUNNEL_TYPE_TEREDO:
+               case RTE_TUNNEL_TYPE_NVGRE:
+               case RTE_L2_TUNNEL_TYPE_E_TAG:
+                       DP_ERR(edev, "Unsupported tunnel type %d\n",
+                               filter_conf->tunnel_type);
+                       return -EINVAL;
+               case RTE_TUNNEL_TYPE_NONE:
+               default:
+                       return 0;
+               }
+               break;
+       case RTE_ETH_FILTER_FDIR:
+               return qede_fdir_filter_conf(eth_dev, filter_op, arg);
+       case RTE_ETH_FILTER_NTUPLE:
+               return qede_ntuple_filter_conf(eth_dev, filter_op, arg);
+       case RTE_ETH_FILTER_MACVLAN:
+       case RTE_ETH_FILTER_ETHERTYPE:
+       case RTE_ETH_FILTER_FLEXIBLE:
+       case RTE_ETH_FILTER_SYN:
+       case RTE_ETH_FILTER_HASH:
+       case RTE_ETH_FILTER_L2_TUNNEL:
+       case RTE_ETH_FILTER_MAX:
+       default:
+               DP_ERR(edev, "Unsupported filter type %d\n",
+                       filter_type);
+               return -EINVAL;
+       }
+
+       return 0;
+}
+
+/* RTE_FLOW */