diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index 11b592dbf16a..4fc45ee0c5d1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -6,6 +6,6 @@ mlx5_core-y :=	main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
 
 mlx5_core-$(CONFIG_MLX5_CORE_EN) += wq.o eswitch.o \
 		en_main.o en_fs.o en_ethtool.o en_tx.o en_rx.o \
-		en_txrx.o en_clock.o vxlan.o
+		en_txrx.o en_clock.o vxlan.o en_tc.o
 
 mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) +=  en_dcbnl.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index dbc2fb89e067..0f76d321030f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -43,6 +43,7 @@
 #include <linux/mlx5/port.h>
 #include <linux/mlx5/vport.h>
 #include <linux/mlx5/transobj.h>
+#include <linux/rhashtable.h>
 #include "wq.h"
 #include "mlx5_core.h"
 
@@ -527,8 +528,16 @@ struct mlx5e_flow_table {
 	struct mlx5_flow_group		**g;
 };
 
+struct mlx5e_tc_flow_table {
+	struct mlx5_flow_table		*t;
+
+	struct rhashtable_params        ht_params;
+	struct rhashtable               ht;
+};
+
 struct mlx5e_flow_tables {
 	struct mlx5_flow_namespace	*ns;
+	struct mlx5e_tc_flow_table	tc;
 	struct mlx5e_flow_table		vlan;
 	struct mlx5e_flow_table		main;
 };
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
index 80d81abc4820..d00a24203410 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
@@ -1041,7 +1041,7 @@ static int mlx5e_create_main_flow_table(struct mlx5e_priv *priv)
 	int err;
 
 	ft->num_groups = 0;
-	ft->t = mlx5_create_flow_table(priv->fts.ns, 0, MLX5E_MAIN_TABLE_SIZE);
+	ft->t = mlx5_create_flow_table(priv->fts.ns, 1, MLX5E_MAIN_TABLE_SIZE);
 
 	if (IS_ERR(ft->t)) {
 		err = PTR_ERR(ft->t);
@@ -1150,7 +1150,7 @@ static int mlx5e_create_vlan_flow_table(struct mlx5e_priv *priv)
 	int err;
 
 	ft->num_groups = 0;
-	ft->t = mlx5_create_flow_table(priv->fts.ns, 0, MLX5E_VLAN_TABLE_SIZE);
+	ft->t = mlx5_create_flow_table(priv->fts.ns, 1, MLX5E_VLAN_TABLE_SIZE);
 
 	if (IS_ERR(ft->t)) {
 		err = PTR_ERR(ft->t);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 19e5daeaa61d..ac5807803c84 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -30,9 +30,12 @@
  * SOFTWARE.
  */
 
+#include <net/tc_act/tc_gact.h>
+#include <net/pkt_cls.h>
 #include <linux/mlx5/fs.h>
 #include <net/vxlan.h>
 #include "en.h"
+#include "en_tc.h"
 #include "eswitch.h"
 #include "vxlan.h"
 
@@ -1883,7 +1886,25 @@ static int mlx5e_setup_tc(struct net_device *netdev, u8 tc)
 static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle,
 			      __be16 proto, struct tc_to_netdev *tc)
 {
-	if (handle != TC_H_ROOT || tc->type != TC_SETUP_MQPRIO)
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	if (TC_H_MAJ(handle) != TC_H_MAJ(TC_H_INGRESS))
+		goto mqprio;
+
+	switch (tc->type) {
+	case TC_SETUP_CLSFLOWER:
+		switch (tc->cls_flower->command) {
+		case TC_CLSFLOWER_REPLACE:
+			return mlx5e_configure_flower(priv, proto, tc->cls_flower);
+		case TC_CLSFLOWER_DESTROY:
+			return mlx5e_delete_flower(priv, tc->cls_flower);
+		}
+	default:
+		return -EOPNOTSUPP;
+	}
+
+mqprio:
+	if (tc->type != TC_SETUP_MQPRIO)
 		return -EINVAL;
 
 	return mlx5e_setup_tc(dev, tc->tc);
@@ -1968,6 +1989,13 @@ static int mlx5e_set_features(struct net_device *netdev,
 			mlx5e_disable_vlan_filter(priv);
 	}
 
+	if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) &&
+	    mlx5e_tc_num_filters(priv)) {
+		netdev_err(netdev,
+			   "Active offloaded tc filters, can't turn hw_tc_offload off\n");
+		return -EINVAL;
+	}
+
 	return err;
 }
 
@@ -2375,6 +2403,13 @@ static void mlx5e_build_netdev(struct net_device *netdev)
 	if (!priv->params.lro_en)
 		netdev->features  &= ~NETIF_F_LRO;
 
+#define FT_CAP(f) MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.f)
+	if (FT_CAP(flow_modify_en) &&
+	    FT_CAP(modify_root) &&
+	    FT_CAP(identified_miss_table_mode) &&
+	    FT_CAP(flow_table_modify))
+		priv->netdev->hw_features      |= NETIF_F_HW_TC;
+
 	netdev->features         |= NETIF_F_HIGHDMA;
 
 	netdev->priv_flags       |= IFF_UNICAST_FLT;
@@ -2496,6 +2531,10 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 
 	mlx5e_vxlan_init(priv);
 
+	err = mlx5e_tc_init(priv);
+	if (err)
+		goto err_destroy_flow_tables;
+
 #ifdef CONFIG_MLX5_CORE_EN_DCB
 	mlx5e_dcbnl_ieee_setets_core(priv, &priv->params.ets);
 #endif
@@ -2503,7 +2542,7 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 	err = register_netdev(netdev);
 	if (err) {
 		mlx5_core_err(mdev, "register_netdev failed, %d\n", err);
-		goto err_destroy_flow_tables;
+		goto err_tc_cleanup;
 	}
 
 	if (mlx5e_vxlan_allowed(mdev))
@@ -2514,6 +2553,9 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 
 	return priv;
 
+err_tc_cleanup:
+	mlx5e_tc_cleanup(priv);
+
 err_destroy_flow_tables:
 	mlx5e_destroy_flow_tables(priv);
 
@@ -2561,6 +2603,7 @@ static void mlx5e_destroy_netdev(struct mlx5_core_dev *mdev, void *vpriv)
 	mlx5e_disable_async_events(priv);
 	flush_scheduled_work();
 	unregister_netdev(netdev);
+	mlx5e_tc_cleanup(priv);
 	mlx5e_vxlan_cleanup(priv);
 	mlx5e_destroy_flow_tables(priv);
 	mlx5e_destroy_tirs(priv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 884ed19cded2..58d4e2f962c3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -35,6 +35,7 @@
 #include <linux/tcp.h>
 #include <net/busy_poll.h>
 #include "en.h"
+#include "en_tc.h"
 
 static inline bool mlx5e_rx_hw_stamp(struct mlx5e_tstamp *tstamp)
 {
@@ -224,6 +225,8 @@ static inline void mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
 	if (cqe_has_vlan(cqe))
 		__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
 				       be16_to_cpu(cqe->vlan_info));
+
+	skb->mark = be32_to_cpu(cqe->sop_drop_qpn) & MLX5E_TC_FLOW_ID_MASK;
 }
 
 int mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
new file mode 100644
index 000000000000..b3de09f13425
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -0,0 +1,429 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <net/flow_dissector.h>
+#include <net/pkt_cls.h>
+#include <net/tc_act/tc_gact.h>
+#include <net/tc_act/tc_skbedit.h>
+#include <linux/mlx5/fs.h>
+#include <linux/mlx5/device.h>
+#include <linux/rhashtable.h>
+#include "en.h"
+#include "en_tc.h"
+
+struct mlx5e_tc_flow {
+	struct rhash_head	node;
+	u64			cookie;
+	struct mlx5_flow_rule	*rule;
+};
+
+#define MLX5E_TC_FLOW_TABLE_NUM_ENTRIES 1024
+#define MLX5E_TC_FLOW_TABLE_NUM_GROUPS 4
+
+static struct mlx5_flow_rule *mlx5e_tc_add_flow(struct mlx5e_priv *priv,
+						u32 *match_c, u32 *match_v,
+						u32 action, u32 flow_tag)
+{
+	struct mlx5_flow_destination dest = {
+		.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE,
+		{.ft = priv->fts.vlan.t},
+	};
+	struct mlx5_flow_rule *rule;
+	bool table_created = false;
+
+	if (IS_ERR_OR_NULL(priv->fts.tc.t)) {
+		priv->fts.tc.t =
+			mlx5_create_auto_grouped_flow_table(priv->fts.ns, 0,
+							    MLX5E_TC_FLOW_TABLE_NUM_ENTRIES,
+							    MLX5E_TC_FLOW_TABLE_NUM_GROUPS);
+		if (IS_ERR(priv->fts.tc.t)) {
+			netdev_err(priv->netdev,
+				   "Failed to create tc offload table\n");
+			return ERR_CAST(priv->fts.tc.t);
+		}
+
+		table_created = true;
+	}
+
+	rule = mlx5_add_flow_rule(priv->fts.tc.t, MLX5_MATCH_OUTER_HEADERS,
+				  match_c, match_v,
+				  action, flow_tag,
+				  action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST ? &dest : NULL);
+
+	if (IS_ERR(rule) && table_created) {
+		mlx5_destroy_flow_table(priv->fts.tc.t);
+		priv->fts.tc.t = NULL;
+	}
+
+	return rule;
+}
+
+static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
+			      struct mlx5_flow_rule *rule)
+{
+	mlx5_del_flow_rule(rule);
+
+	if (!mlx5e_tc_num_filters(priv)) {
+		mlx5_destroy_flow_table(priv->fts.tc.t);
+		priv->fts.tc.t = NULL;
+	}
+}
+
+static int parse_cls_flower(struct mlx5e_priv *priv,
+			    u32 *match_c, u32 *match_v,
+			    struct tc_cls_flower_offload *f)
+{
+	void *headers_c = MLX5_ADDR_OF(fte_match_param, match_c, outer_headers);
+	void *headers_v = MLX5_ADDR_OF(fte_match_param, match_v, outer_headers);
+	u16 addr_type = 0;
+	u8 ip_proto = 0;
+
+	if (f->dissector->used_keys &
+	    ~(BIT(FLOW_DISSECTOR_KEY_CONTROL) |
+	      BIT(FLOW_DISSECTOR_KEY_BASIC) |
+	      BIT(FLOW_DISSECTOR_KEY_ETH_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_IPV4_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_IPV6_ADDRS) |
+	      BIT(FLOW_DISSECTOR_KEY_PORTS))) {
+		netdev_warn(priv->netdev, "Unsupported key used: 0x%x\n",
+			    f->dissector->used_keys);
+		return -EOPNOTSUPP;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_CONTROL)) {
+		struct flow_dissector_key_control *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->key);
+		addr_type = key->addr_type;
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_BASIC)) {
+		struct flow_dissector_key_basic *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->key);
+		struct flow_dissector_key_basic *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_BASIC,
+						  f->mask);
+		ip_proto = key->ip_proto;
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ethertype,
+			 ntohs(mask->n_proto));
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ethertype,
+			 ntohs(key->n_proto));
+
+		MLX5_SET(fte_match_set_lyr_2_4, headers_c, ip_protocol,
+			 mask->ip_proto);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, ip_protocol,
+			 key->ip_proto);
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_ETH_ADDRS)) {
+		struct flow_dissector_key_eth_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ETH_ADDRS,
+						  f->key);
+		struct flow_dissector_key_eth_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_ETH_ADDRS,
+						  f->mask);
+
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+					     dmac_47_16),
+				mask->dst);
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+					     dmac_47_16),
+				key->dst);
+
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+					     smac_47_16),
+				mask->src);
+		ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+					     smac_47_16),
+				key->src);
+	}
+
+	if (addr_type == FLOW_DISSECTOR_KEY_IPV4_ADDRS) {
+		struct flow_dissector_key_ipv4_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV4_ADDRS,
+						  f->key);
+		struct flow_dissector_key_ipv4_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV4_ADDRS,
+						  f->mask);
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       &mask->src, sizeof(mask->src));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    src_ipv4_src_ipv6.ipv4_layout.ipv4),
+		       &key->src, sizeof(key->src));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       &mask->dst, sizeof(mask->dst));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    dst_ipv4_dst_ipv6.ipv4_layout.ipv4),
+		       &key->dst, sizeof(key->dst));
+	}
+
+	if (addr_type == FLOW_DISSECTOR_KEY_IPV6_ADDRS) {
+		struct flow_dissector_key_ipv6_addrs *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV6_ADDRS,
+						  f->key);
+		struct flow_dissector_key_ipv6_addrs *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_IPV6_ADDRS,
+						  f->mask);
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &mask->src, sizeof(mask->src));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       &key->src, sizeof(key->src));
+
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_c,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &mask->dst, sizeof(mask->dst));
+		memcpy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, headers_v,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       &key->dst, sizeof(key->dst));
+	}
+
+	if (dissector_uses_key(f->dissector, FLOW_DISSECTOR_KEY_PORTS)) {
+		struct flow_dissector_key_ports *key =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_PORTS,
+						  f->key);
+		struct flow_dissector_key_ports *mask =
+			skb_flow_dissector_target(f->dissector,
+						  FLOW_DISSECTOR_KEY_PORTS,
+						  f->mask);
+		switch (ip_proto) {
+		case IPPROTO_TCP:
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 tcp_sport, ntohs(mask->src));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 tcp_sport, ntohs(key->src));
+
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 tcp_dport, ntohs(mask->dst));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 tcp_dport, ntohs(key->dst));
+			break;
+
+		case IPPROTO_UDP:
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 udp_sport, ntohs(mask->src));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 udp_sport, ntohs(key->src));
+
+			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
+				 udp_dport, ntohs(mask->dst));
+			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
+				 udp_dport, ntohs(key->dst));
+			break;
+		default:
+			netdev_err(priv->netdev,
+				   "Only UDP and TCP transport are supported\n");
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static int parse_tc_actions(struct mlx5e_priv *priv, struct tcf_exts *exts,
+			    u32 *action, u32 *flow_tag)
+{
+	const struct tc_action *a;
+
+	if (tc_no_actions(exts))
+		return -EINVAL;
+
+	*flow_tag = MLX5_FS_DEFAULT_FLOW_TAG;
+	*action = 0;
+
+	tc_for_each_action(a, exts) {
+		/* Only support a single action per rule */
+		if (*action)
+			return -EINVAL;
+
+		if (is_tcf_gact_shot(a)) {
+			*action |= MLX5_FLOW_CONTEXT_ACTION_DROP;
+			continue;
+		}
+
+		if (is_tcf_skbedit_mark(a)) {
+			u32 mark = tcf_skbedit_mark(a);
+
+			if (mark & ~MLX5E_TC_FLOW_ID_MASK) {
+				netdev_warn(priv->netdev, "Bad flow mark - only 16 bit is supported: 0x%x\n",
+					    mark);
+				return -EINVAL;
+			}
+
+			*flow_tag = mark;
+			*action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+			continue;
+		}
+
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int mlx5e_configure_flower(struct mlx5e_priv *priv, __be16 protocol,
+			   struct tc_cls_flower_offload *f)
+{
+	struct mlx5e_tc_flow_table *tc = &priv->fts.tc;
+	u32 *match_c;
+	u32 *match_v;
+	int err = 0;
+	u32 flow_tag;
+	u32 action;
+	struct mlx5e_tc_flow *flow;
+	struct mlx5_flow_rule *old = NULL;
+
+	flow = rhashtable_lookup_fast(&tc->ht, &f->cookie,
+				      tc->ht_params);
+	if (flow)
+		old = flow->rule;
+	else
+		flow = kzalloc(sizeof(*flow), GFP_KERNEL);
+
+	match_c = kzalloc(MLX5_ST_SZ_BYTES(fte_match_param), GFP_KERNEL);
+	match_v = kzalloc(MLX5_ST_SZ_BYTES(fte_match_param), GFP_KERNEL);
+	if (!match_c || !match_v || !flow) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	flow->cookie = f->cookie;
+
+	err = parse_cls_flower(priv, match_c, match_v, f);
+	if (err < 0)
+		goto err_free;
+
+	err = parse_tc_actions(priv, f->exts, &action, &flow_tag);
+	if (err < 0)
+		goto err_free;
+
+	err = rhashtable_insert_fast(&tc->ht, &flow->node,
+				     tc->ht_params);
+	if (err)
+		goto err_free;
+
+	flow->rule = mlx5e_tc_add_flow(priv, match_c, match_v, action,
+				       flow_tag);
+	if (IS_ERR(flow->rule)) {
+		err = PTR_ERR(flow->rule);
+		goto err_hash_del;
+	}
+
+	if (old)
+		mlx5e_tc_del_flow(priv, old);
+
+	goto out;
+
+err_hash_del:
+	rhashtable_remove_fast(&tc->ht, &flow->node, tc->ht_params);
+
+err_free:
+	if (!old)
+		kfree(flow);
+out:
+	kfree(match_c);
+	kfree(match_v);
+	return err;
+}
+
+int mlx5e_delete_flower(struct mlx5e_priv *priv,
+			struct tc_cls_flower_offload *f)
+{
+	struct mlx5e_tc_flow *flow;
+	struct mlx5e_tc_flow_table *tc = &priv->fts.tc;
+
+	flow = rhashtable_lookup_fast(&tc->ht, &f->cookie,
+				      tc->ht_params);
+	if (!flow)
+		return -EINVAL;
+
+	rhashtable_remove_fast(&tc->ht, &flow->node, tc->ht_params);
+
+	mlx5e_tc_del_flow(priv, flow->rule);
+
+	kfree(flow);
+
+	return 0;
+}
+
+static const struct rhashtable_params mlx5e_tc_flow_ht_params = {
+	.head_offset = offsetof(struct mlx5e_tc_flow, node),
+	.key_offset = offsetof(struct mlx5e_tc_flow, cookie),
+	.key_len = sizeof(((struct mlx5e_tc_flow *)0)->cookie),
+	.automatic_shrinking = true,
+};
+
+int mlx5e_tc_init(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tc_flow_table *tc = &priv->fts.tc;
+
+	tc->ht_params = mlx5e_tc_flow_ht_params;
+	return rhashtable_init(&tc->ht, &tc->ht_params);
+}
+
+static void _mlx5e_tc_del_flow(void *ptr, void *arg)
+{
+	struct mlx5e_tc_flow *flow = ptr;
+	struct mlx5e_priv *priv = arg;
+
+	mlx5e_tc_del_flow(priv, flow->rule);
+	kfree(flow);
+}
+
+void mlx5e_tc_cleanup(struct mlx5e_priv *priv)
+{
+	struct mlx5e_tc_flow_table *tc = &priv->fts.tc;
+
+	rhashtable_free_and_destroy(&tc->ht, _mlx5e_tc_del_flow, priv);
+
+	if (!IS_ERR_OR_NULL(priv->fts.tc.t)) {
+		mlx5_destroy_flow_table(priv->fts.tc.t);
+		priv->fts.tc.t = NULL;
+	}
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
new file mode 100644
index 000000000000..d677428dc10f
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.h
@@ -0,0 +1,51 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef __MLX5_EN_TC_H__
+#define __MLX5_EN_TC_H__
+
+#define MLX5E_TC_FLOW_ID_MASK 0x0000ffff
+
+int mlx5e_tc_init(struct mlx5e_priv *priv);
+void mlx5e_tc_cleanup(struct mlx5e_priv *priv);
+
+int mlx5e_configure_flower(struct mlx5e_priv *priv, __be16 protocol,
+			   struct tc_cls_flower_offload *f);
+int mlx5e_delete_flower(struct mlx5e_priv *priv,
+			struct tc_cls_flower_offload *f);
+
+static inline int mlx5e_tc_num_filters(struct mlx5e_priv *priv)
+{
+	return atomic_read(&priv->fts.tc.ht.nelems);
+}
+
+#endif /* __MLX5_EN_TC_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
index a9894d2e8e26..f46f1db0fc00 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
@@ -218,19 +218,22 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 				      match_value);
 	memcpy(in_match_value, &fte->val, MLX5_ST_SZ_BYTES(fte_match_param));
 
-	in_dests = MLX5_ADDR_OF(flow_context, in_flow_context, destination);
-	list_for_each_entry(dst, &fte->node.children, node.list) {
-		unsigned int id;
+	if (fte->action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		in_dests = MLX5_ADDR_OF(flow_context, in_flow_context, destination);
+		list_for_each_entry(dst, &fte->node.children, node.list) {
+			unsigned int id;
 
-		MLX5_SET(dest_format_struct, in_dests, destination_type,
-			 dst->dest_attr.type);
-		if (dst->dest_attr.type ==
-		    MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE)
-			id = dst->dest_attr.ft->id;
-		else
-			id = dst->dest_attr.tir_num;
-		MLX5_SET(dest_format_struct, in_dests, destination_id, id);
-		in_dests += MLX5_ST_SZ_BYTES(dest_format_struct);
+			MLX5_SET(dest_format_struct, in_dests, destination_type,
+				 dst->dest_attr.type);
+			if (dst->dest_attr.type ==
+			    MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE) {
+				id = dst->dest_attr.ft->id;
+			} else {
+				id = dst->dest_attr.tir_num;
+			}
+			MLX5_SET(dest_format_struct, in_dests, destination_id, id);
+			in_dests += MLX5_ST_SZ_BYTES(dest_format_struct);
+		}
 	}
 	memset(out, 0, sizeof(out));
 	err = mlx5_cmd_exec_check_status(dev, in, inlen, out,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 6f68dba8d7ed..e848d708d2b7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -73,8 +73,8 @@
 #define BY_PASS_MIN_LEVEL (KENREL_MIN_LEVEL + MLX5_BY_PASS_NUM_PRIOS +\
 			   LEFTOVERS_MAX_FT)
 
-#define KERNEL_MAX_FT 2
-#define KERNEL_NUM_PRIOS 1
+#define KERNEL_MAX_FT 3
+#define KERNEL_NUM_PRIOS 2
 #define KENREL_MIN_LEVEL 2
 
 struct node_caps {
@@ -360,8 +360,8 @@ static void del_rule(struct fs_node *node)
 	memcpy(match_value, fte->val, sizeof(fte->val));
 	fs_get_obj(ft, fg->node.parent);
 	list_del(&rule->node.list);
-	fte->dests_size--;
-	if (fte->dests_size) {
+	if ((fte->action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) &&
+	    --fte->dests_size) {
 		err = mlx5_cmd_update_fte(dev, ft,
 					  fg->id, fte);
 		if (err)
@@ -763,7 +763,8 @@ static struct mlx5_flow_rule *alloc_rule(struct mlx5_flow_destination *dest)
 		return NULL;
 
 	rule->node.type = FS_TYPE_FLOW_DEST;
-	memcpy(&rule->dest_attr, dest, sizeof(*dest));
+	if (dest)
+		memcpy(&rule->dest_attr, dest, sizeof(*dest));
 
 	return rule;
 }
@@ -785,8 +786,9 @@ static struct mlx5_flow_rule *add_rule_fte(struct fs_fte *fte,
 	/* Add dest to dests list- added as first element after the head */
 	tree_init_node(&rule->node, 1, del_rule);
 	list_add_tail(&rule->node.list, &fte->node.children);
-	fte->dests_size++;
-	if (fte->dests_size == 1)
+	if (dest)
+		fte->dests_size++;
+	if (fte->dests_size == 1 || !dest)
 		err = mlx5_cmd_create_fte(get_dev(&ft->node),
 					  ft, fg->id, fte);
 	else
@@ -802,7 +804,8 @@ static struct mlx5_flow_rule *add_rule_fte(struct fs_fte *fte,
 free_rule:
 	list_del(&rule->node.list);
 	kfree(rule);
-	fte->dests_size--;
+	if (dest)
+		fte->dests_size--;
 	return ERR_PTR(err);
 }
 
@@ -996,6 +999,9 @@ mlx5_add_flow_rule(struct mlx5_flow_table *ft,
 	struct mlx5_flow_group *g;
 	struct mlx5_flow_rule *rule;
 
+	if ((action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) && !dest)
+		return ERR_PTR(-EINVAL);
+
 	nested_lock_ref_node(&ft->node, FS_MUTEX_GRANDPARENT);
 	fs_for_each_fg(g, ft)
 		if (compare_match_criteria(g->mask.match_criteria_enable,
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index fd30cb545c45..41df0b450757 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -786,6 +786,7 @@ typedef u16 (*select_queue_fallback_t)(struct net_device *dev,
 enum {
 	TC_SETUP_MQPRIO,
 	TC_SETUP_CLSU32,
+	TC_SETUP_CLSFLOWER,
 };
 
 struct tc_cls_u32_offload;
@@ -795,6 +796,7 @@ struct tc_to_netdev {
 	union {
 		u8 tc;
 		struct tc_cls_u32_offload *cls_u32;
+		struct tc_cls_flower_offload *cls_flower;
 	};
 };
 
diff --git a/include/net/act_api.h b/include/net/act_api.h
index 342be6c5ab5c..2a19fe111c78 100644
--- a/include/net/act_api.h
+++ b/include/net/act_api.h
@@ -78,11 +78,6 @@ static inline void tcf_lastuse_update(struct tcf_t *tm)
 		tm->lastuse = now;
 }
 
-#ifdef CONFIG_NET_CLS_ACT
-
-#define ACT_P_CREATED 1
-#define ACT_P_DELETED 1
-
 struct tc_action {
 	void			*priv;
 	const struct tc_action_ops	*ops;
@@ -92,6 +87,11 @@ struct tc_action {
 	struct tcf_hashinfo	*hinfo;
 };
 
+#ifdef CONFIG_NET_CLS_ACT
+
+#define ACT_P_CREATED 1
+#define ACT_P_DELETED 1
+
 struct tc_action_ops {
 	struct list_head head;
 	char    kind[IFNAMSIZ];
@@ -171,5 +171,16 @@ int tcf_action_dump(struct sk_buff *skb, struct list_head *, int, int);
 int tcf_action_dump_old(struct sk_buff *skb, struct tc_action *a, int, int);
 int tcf_action_dump_1(struct sk_buff *skb, struct tc_action *a, int, int);
 int tcf_action_copy_stats(struct sk_buff *, struct tc_action *, int);
+
+#define tc_no_actions(_exts) \
+	(list_empty(&(_exts)->actions))
+
+#define tc_for_each_action(_a, _exts) \
+	list_for_each_entry(a, &(_exts)->actions, list)
+#else /* CONFIG_NET_CLS_ACT */
+
+#define tc_no_actions(_exts) true
+#define tc_for_each_action(_a, _exts) while (0)
+
 #endif /* CONFIG_NET_CLS_ACT */
 #endif
diff --git a/include/net/flow_dissector.h b/include/net/flow_dissector.h
index 8c8548cf5888..d3d60dccd19f 100644
--- a/include/net/flow_dissector.h
+++ b/include/net/flow_dissector.h
@@ -184,4 +184,17 @@ static inline bool flow_keys_have_l4(struct flow_keys *keys)
 
 u32 flow_hash_from_keys(struct flow_keys *keys);
 
+static inline bool dissector_uses_key(const struct flow_dissector *flow_dissector,
+				      enum flow_dissector_key_id key_id)
+{
+	return flow_dissector->used_keys & (1 << key_id);
+}
+
+static inline void *skb_flow_dissector_target(struct flow_dissector *flow_dissector,
+					      enum flow_dissector_key_id key_id,
+					      void *target_container)
+{
+	return ((char *)target_container) + flow_dissector->offset[key_id];
+}
+
 #endif
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index bea14eee373e..5b4e8f08b8f0 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -409,4 +409,18 @@ static inline bool tc_should_offload(struct net_device *dev, u32 flags)
 	return true;
 }
 
+enum tc_fl_command {
+	TC_CLSFLOWER_REPLACE,
+	TC_CLSFLOWER_DESTROY,
+};
+
+struct tc_cls_flower_offload {
+	enum tc_fl_command command;
+	u64 cookie;
+	struct flow_dissector *dissector;
+	struct fl_flow_key *mask;
+	struct fl_flow_key *key;
+	struct tcf_exts *exts;
+};
+
 #endif
diff --git a/include/net/tc_act/tc_gact.h b/include/net/tc_act/tc_gact.h
index 04a31830711b..93c520b83d10 100644
--- a/include/net/tc_act/tc_gact.h
+++ b/include/net/tc_act/tc_gact.h
@@ -16,9 +16,9 @@ struct tcf_gact {
 #define to_gact(a) \
 	container_of(a->priv, struct tcf_gact, common)
 
-#ifdef CONFIG_NET_CLS_ACT
 static inline bool is_tcf_gact_shot(const struct tc_action *a)
 {
+#ifdef CONFIG_NET_CLS_ACT
 	struct tcf_gact *gact;
 
 	if (a->ops && a->ops->type != TCA_ACT_GACT)
@@ -28,7 +28,7 @@ static inline bool is_tcf_gact_shot(const struct tc_action *a)
 	if (gact->tcf_action == TC_ACT_SHOT)
 		return true;
 
+#endif
 	return false;
 }
-#endif
 #endif /* __NET_TC_GACT_H */
diff --git a/include/net/tc_act/tc_skbedit.h b/include/net/tc_act/tc_skbedit.h
index 0df9a0db4a8e..b496d5ad7d42 100644
--- a/include/net/tc_act/tc_skbedit.h
+++ b/include/net/tc_act/tc_skbedit.h
@@ -20,6 +20,7 @@
 #define __NET_TC_SKBEDIT_H
 
 #include <net/act_api.h>
+#include <linux/tc_act/tc_skbedit.h>
 
 struct tcf_skbedit {
 	struct tcf_common	common;
@@ -32,4 +33,19 @@ struct tcf_skbedit {
 #define to_skbedit(a) \
 	container_of(a->priv, struct tcf_skbedit, common)
 
+/* Return true iff action is mark */
+static inline bool is_tcf_skbedit_mark(const struct tc_action *a)
+{
+#ifdef CONFIG_NET_CLS_ACT
+	if (a->ops && a->ops->type == TCA_ACT_SKBEDIT)
+		return to_skbedit(a)->flags == SKBEDIT_F_MARK;
+#endif
+	return false;
+}
+
+static inline u32 tcf_skbedit_mark(const struct tc_action *a)
+{
+	return to_skbedit(a)->mark;
+}
+
 #endif /* __NET_TC_SKBEDIT_H */
diff --git a/include/uapi/linux/pkt_cls.h b/include/uapi/linux/pkt_cls.h
index 9874f5680926..c43c5f78b9c4 100644
--- a/include/uapi/linux/pkt_cls.h
+++ b/include/uapi/linux/pkt_cls.h
@@ -417,6 +417,8 @@ enum {
 	TCA_FLOWER_KEY_TCP_DST,		/* be16 */
 	TCA_FLOWER_KEY_UDP_SRC,		/* be16 */
 	TCA_FLOWER_KEY_UDP_DST,		/* be16 */
+
+	TCA_FLOWER_FLAGS,
 	__TCA_FLOWER_MAX,
 };
 
diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
index 7c7b8739b8b8..a669dea146c6 100644
--- a/net/core/flow_dissector.c
+++ b/net/core/flow_dissector.c
@@ -19,25 +19,12 @@
 #include <net/flow_dissector.h>
 #include <scsi/fc/fc_fcoe.h>
 
-static bool dissector_uses_key(const struct flow_dissector *flow_dissector,
-			       enum flow_dissector_key_id key_id)
-{
-	return flow_dissector->used_keys & (1 << key_id);
-}
-
 static void dissector_set_key(struct flow_dissector *flow_dissector,
 			      enum flow_dissector_key_id key_id)
 {
 	flow_dissector->used_keys |= (1 << key_id);
 }
 
-static void *skb_flow_dissector_target(struct flow_dissector *flow_dissector,
-				       enum flow_dissector_key_id key_id,
-				       void *target_container)
-{
-	return ((char *) target_container) + flow_dissector->offset[key_id];
-}
-
 void skb_flow_dissector_init(struct flow_dissector *flow_dissector,
 			     const struct flow_dissector_key *key,
 			     unsigned int key_count)
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 95b021243233..25d87666bf1e 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -165,6 +165,51 @@ static void fl_destroy_filter(struct rcu_head *head)
 	kfree(f);
 }
 
+static void fl_hw_destroy_filter(struct tcf_proto *tp, u64 cookie)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_flower_offload offload = {0};
+	struct tc_to_netdev tc;
+
+	if (!tc_should_offload(dev, 0))
+		return;
+
+	offload.command = TC_CLSFLOWER_DESTROY;
+	offload.cookie = cookie;
+
+	tc.type = TC_SETUP_CLSFLOWER;
+	tc.cls_flower = &offload;
+
+	dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle, tp->protocol, &tc);
+}
+
+static void fl_hw_replace_filter(struct tcf_proto *tp,
+				 struct flow_dissector *dissector,
+				 struct fl_flow_key *mask,
+				 struct fl_flow_key *key,
+				 struct tcf_exts *actions,
+				 u64 cookie, u32 flags)
+{
+	struct net_device *dev = tp->q->dev_queue->dev;
+	struct tc_cls_flower_offload offload = {0};
+	struct tc_to_netdev tc;
+
+	if (!tc_should_offload(dev, flags))
+		return;
+
+	offload.command = TC_CLSFLOWER_REPLACE;
+	offload.cookie = cookie;
+	offload.dissector = dissector;
+	offload.mask = mask;
+	offload.key = key;
+	offload.exts = actions;
+
+	tc.type = TC_SETUP_CLSFLOWER;
+	tc.cls_flower = &offload;
+
+	dev->netdev_ops->ndo_setup_tc(dev, tp->q->handle, tp->protocol, &tc);
+}
+
 static bool fl_destroy(struct tcf_proto *tp, bool force)
 {
 	struct cls_fl_head *head = rtnl_dereference(tp->root);
@@ -174,6 +219,7 @@ static bool fl_destroy(struct tcf_proto *tp, bool force)
 		return false;
 
 	list_for_each_entry_safe(f, next, &head->filters, list) {
+		fl_hw_destroy_filter(tp, (u64)f);
 		list_del_rcu(&f->list);
 		call_rcu(&f->rcu, fl_destroy_filter);
 	}
@@ -459,6 +505,7 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 	struct cls_fl_filter *fnew;
 	struct nlattr *tb[TCA_FLOWER_MAX + 1];
 	struct fl_flow_mask mask = {};
+	u32 flags = 0;
 	int err;
 
 	if (!tca[TCA_OPTIONS])
@@ -486,6 +533,9 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 	}
 	fnew->handle = handle;
 
+	if (tb[TCA_FLOWER_FLAGS])
+		flags = nla_get_u32(tb[TCA_FLOWER_FLAGS]);
+
 	err = fl_set_parms(net, tp, fnew, &mask, base, tb, tca[TCA_RATE], ovr);
 	if (err)
 		goto errout;
@@ -498,9 +548,20 @@ static int fl_change(struct net *net, struct sk_buff *in_skb,
 				     head->ht_params);
 	if (err)
 		goto errout;
-	if (fold)
+
+	fl_hw_replace_filter(tp,
+			     &head->dissector,
+			     &mask.key,
+			     &fnew->key,
+			     &fnew->exts,
+			     (u64)fnew,
+			     flags);
+
+	if (fold) {
 		rhashtable_remove_fast(&head->ht, &fold->ht_node,
 				       head->ht_params);
+		fl_hw_destroy_filter(tp, (u64)fold);
+	}
 
 	*arg = (unsigned long) fnew;
 
@@ -527,6 +588,7 @@ static int fl_delete(struct tcf_proto *tp, unsigned long arg)
 	rhashtable_remove_fast(&head->ht, &f->ht_node,
 			       head->ht_params);
 	list_del_rcu(&f->list);
+	fl_hw_destroy_filter(tp, (u64)f);
 	tcf_unbind_filter(tp, &f->res);
 	call_rcu(&f->rcu, fl_destroy_filter);
 	return 0;