1 files changed, 968 insertions, 366 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index e0adb604f461..fd4392999eee 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -48,6 +48,7 @@ struct mlx5e_sq_param {
 	u32                        sqc[MLX5_ST_SZ_DW(sqc)];
 	struct mlx5_wq_param       wq;
 	u16                        max_inline;
+	bool                       icosq;
 };
 
 struct mlx5e_cq_param {
@@ -59,8 +60,10 @@ struct mlx5e_cq_param {
 struct mlx5e_channel_param {
 	struct mlx5e_rq_param      rq;
 	struct mlx5e_sq_param      sq;
+	struct mlx5e_sq_param      icosq;
 	struct mlx5e_cq_param      rx_cq;
 	struct mlx5e_cq_param      tx_cq;
+	struct mlx5e_cq_param      icosq_cq;
 };
 
 static void mlx5e_update_carrier(struct mlx5e_priv *priv)
@@ -88,82 +91,15 @@ static void mlx5e_update_carrier_work(struct work_struct *work)
 	mutex_unlock(&priv->state_lock);
 }
 
-static void mlx5e_update_pport_counters(struct mlx5e_priv *priv)
-{
-	struct mlx5_core_dev *mdev = priv->mdev;
-	struct mlx5e_pport_stats *s = &priv->stats.pport;
-	u32 *in;
-	u32 *out;
-	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
-
-	in  = mlx5_vzalloc(sz);
-	out = mlx5_vzalloc(sz);
-	if (!in || !out)
-		goto free_out;
-
-	MLX5_SET(ppcnt_reg, in, local_port, 1);
-
-	MLX5_SET(ppcnt_reg, in, grp, MLX5_IEEE_802_3_COUNTERS_GROUP);
-	mlx5_core_access_reg(mdev, in, sz, out,
-			     sz, MLX5_REG_PPCNT, 0, 0);
-	memcpy(s->IEEE_802_3_counters,
-	       MLX5_ADDR_OF(ppcnt_reg, out, counter_set),
-	       sizeof(s->IEEE_802_3_counters));
-
-	MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2863_COUNTERS_GROUP);
-	mlx5_core_access_reg(mdev, in, sz, out,
-			     sz, MLX5_REG_PPCNT, 0, 0);
-	memcpy(s->RFC_2863_counters,
-	       MLX5_ADDR_OF(ppcnt_reg, out, counter_set),
-	       sizeof(s->RFC_2863_counters));
-
-	MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2819_COUNTERS_GROUP);
-	mlx5_core_access_reg(mdev, in, sz, out,
-			     sz, MLX5_REG_PPCNT, 0, 0);
-	memcpy(s->RFC_2819_counters,
-	       MLX5_ADDR_OF(ppcnt_reg, out, counter_set),
-	       sizeof(s->RFC_2819_counters));
-
-free_out:
-	kvfree(in);
-	kvfree(out);
-}
-
-void mlx5e_update_stats(struct mlx5e_priv *priv)
+static void mlx5e_update_sw_counters(struct mlx5e_priv *priv)
 {
-	struct mlx5_core_dev *mdev = priv->mdev;
-	struct mlx5e_vport_stats *s = &priv->stats.vport;
+	struct mlx5e_sw_stats *s = &priv->stats.sw;
 	struct mlx5e_rq_stats *rq_stats;
 	struct mlx5e_sq_stats *sq_stats;
-	u32 in[MLX5_ST_SZ_DW(query_vport_counter_in)];
-	u32 *out;
-	int outlen = MLX5_ST_SZ_BYTES(query_vport_counter_out);
-	u64 tx_offload_none;
+	u64 tx_offload_none = 0;
 	int i, j;
 
-	out = mlx5_vzalloc(outlen);
-	if (!out)
-		return;
-
-	/* Collect firts the SW counters and then HW for consistency */
-	s->rx_packets		= 0;
-	s->rx_bytes		= 0;
-	s->tx_packets		= 0;
-	s->tx_bytes		= 0;
-	s->tso_packets		= 0;
-	s->tso_bytes		= 0;
-	s->tso_inner_packets	= 0;
-	s->tso_inner_bytes	= 0;
-	s->tx_queue_stopped	= 0;
-	s->tx_queue_wake	= 0;
-	s->tx_queue_dropped	= 0;
-	s->tx_csum_inner	= 0;
-	tx_offload_none		= 0;
-	s->lro_packets		= 0;
-	s->lro_bytes		= 0;
-	s->rx_csum_none		= 0;
-	s->rx_csum_sw		= 0;
-	s->rx_wqe_err		= 0;
+	memset(s, 0, sizeof(*s));
 	for (i = 0; i < priv->params.num_channels; i++) {
 		rq_stats = &priv->channel[i]->rq.stats;
 
@@ -173,7 +109,13 @@ void mlx5e_update_stats(struct mlx5e_priv *priv)
 		s->lro_bytes	+= rq_stats->lro_bytes;
 		s->rx_csum_none	+= rq_stats->csum_none;
 		s->rx_csum_sw	+= rq_stats->csum_sw;
+		s->rx_csum_inner += rq_stats->csum_inner;
 		s->rx_wqe_err   += rq_stats->wqe_err;
+		s->rx_mpwqe_filler += rq_stats->mpwqe_filler;
+		s->rx_mpwqe_frag   += rq_stats->mpwqe_frag;
+		s->rx_buff_alloc_err += rq_stats->buff_alloc_err;
+		s->rx_cqe_compress_blks += rq_stats->cqe_compress_blks;
+		s->rx_cqe_compress_pkts += rq_stats->cqe_compress_pkts;
 
 		for (j = 0; j < priv->params.num_tc; j++) {
 			sq_stats = &priv->channel[i]->sq[j].stats;
@@ -192,7 +134,23 @@ void mlx5e_update_stats(struct mlx5e_priv *priv)
 		}
 	}
 
-	/* HW counters */
+	/* Update calculated offload counters */
+	s->tx_csum_offload = s->tx_packets - tx_offload_none - s->tx_csum_inner;
+	s->rx_csum_good    = s->rx_packets - s->rx_csum_none -
+			     s->rx_csum_sw;
+
+	s->link_down_events = MLX5_GET(ppcnt_reg,
+				priv->stats.pport.phy_counters,
+				counter_set.phys_layer_cntrs.link_down_events);
+}
+
+static void mlx5e_update_vport_counters(struct mlx5e_priv *priv)
+{
+	int outlen = MLX5_ST_SZ_BYTES(query_vport_counter_out);
+	u32 *out = (u32 *)priv->stats.vport.query_vport_out;
+	u32 in[MLX5_ST_SZ_DW(query_vport_counter_in)];
+	struct mlx5_core_dev *mdev = priv->mdev;
+
 	memset(in, 0, sizeof(in));
 
 	MLX5_SET(query_vport_counter_in, in, opcode,
@@ -202,56 +160,69 @@ void mlx5e_update_stats(struct mlx5e_priv *priv)
 
 	memset(out, 0, outlen);
 
-	if (mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen))
+	mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen);
+}
+
+static void mlx5e_update_pport_counters(struct mlx5e_priv *priv)
+{
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
+	int prio;
+	void *out;
+	u32 *in;
+
+	in = mlx5_vzalloc(sz);
+	if (!in)
 		goto free_out;
 
-#define MLX5_GET_CTR(p, x) \
-	MLX5_GET64(query_vport_counter_out, p, x)
-
-	s->rx_error_packets     =
-		MLX5_GET_CTR(out, received_errors.packets);
-	s->rx_error_bytes       =
-		MLX5_GET_CTR(out, received_errors.octets);
-	s->tx_error_packets     =
-		MLX5_GET_CTR(out, transmit_errors.packets);
-	s->tx_error_bytes       =
-		MLX5_GET_CTR(out, transmit_errors.octets);
-
-	s->rx_unicast_packets   =
-		MLX5_GET_CTR(out, received_eth_unicast.packets);
-	s->rx_unicast_bytes     =
-		MLX5_GET_CTR(out, received_eth_unicast.octets);
-	s->tx_unicast_packets   =
-		MLX5_GET_CTR(out, transmitted_eth_unicast.packets);
-	s->tx_unicast_bytes     =
-		MLX5_GET_CTR(out, transmitted_eth_unicast.octets);
-
-	s->rx_multicast_packets =
-		MLX5_GET_CTR(out, received_eth_multicast.packets);
-	s->rx_multicast_bytes   =
-		MLX5_GET_CTR(out, received_eth_multicast.octets);
-	s->tx_multicast_packets =
-		MLX5_GET_CTR(out, transmitted_eth_multicast.packets);
-	s->tx_multicast_bytes   =
-		MLX5_GET_CTR(out, transmitted_eth_multicast.octets);
-
-	s->rx_broadcast_packets =
-		MLX5_GET_CTR(out, received_eth_broadcast.packets);
-	s->rx_broadcast_bytes   =
-		MLX5_GET_CTR(out, received_eth_broadcast.octets);
-	s->tx_broadcast_packets =
-		MLX5_GET_CTR(out, transmitted_eth_broadcast.packets);
-	s->tx_broadcast_bytes   =
-		MLX5_GET_CTR(out, transmitted_eth_broadcast.octets);
+	MLX5_SET(ppcnt_reg, in, local_port, 1);
 
-	/* Update calculated offload counters */
-	s->tx_csum_offload = s->tx_packets - tx_offload_none - s->tx_csum_inner;
-	s->rx_csum_good    = s->rx_packets - s->rx_csum_none -
-			       s->rx_csum_sw;
+	out = pstats->IEEE_802_3_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_IEEE_802_3_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+
+	out = pstats->RFC_2863_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2863_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+
+	out = pstats->RFC_2819_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_RFC_2819_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+
+	out = pstats->phy_counters;
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_PHYSICAL_LAYER_COUNTERS_GROUP);
+	mlx5_core_access_reg(mdev, in, sz, out, sz, MLX5_REG_PPCNT, 0, 0);
+
+	MLX5_SET(ppcnt_reg, in, grp, MLX5_PER_PRIORITY_COUNTERS_GROUP);
+	for (prio = 0; prio < NUM_PPORT_PRIO; prio++) {
+		out = pstats->per_prio_counters[prio];
+		MLX5_SET(ppcnt_reg, in, prio_tc, prio);
+		mlx5_core_access_reg(mdev, in, sz, out, sz,
+				     MLX5_REG_PPCNT, 0, 0);
+	}
 
-	mlx5e_update_pport_counters(priv);
 free_out:
-	kvfree(out);
+	kvfree(in);
+}
+
+static void mlx5e_update_q_counter(struct mlx5e_priv *priv)
+{
+	struct mlx5e_qcounter_stats *qcnt = &priv->stats.qcnt;
+
+	if (!priv->q_counter)
+		return;
+
+	mlx5_core_query_out_of_buffer(priv->mdev, priv->q_counter,
+				      &qcnt->rx_out_of_buffer);
+}
+
+void mlx5e_update_stats(struct mlx5e_priv *priv)
+{
+	mlx5e_update_q_counter(priv);
+	mlx5e_update_vport_counters(priv);
+	mlx5e_update_pport_counters(priv);
+	mlx5e_update_sw_counters(priv);
 }
 
 static void mlx5e_update_stats_work(struct work_struct *work)
@@ -262,9 +233,8 @@ static void mlx5e_update_stats_work(struct work_struct *work)
 	mutex_lock(&priv->state_lock);
 	if (test_bit(MLX5E_STATE_OPENED, &priv->state)) {
 		mlx5e_update_stats(priv);
-		schedule_delayed_work(dwork,
-				      msecs_to_jiffies(
-					      MLX5E_UPDATE_STATS_INTERVAL));
+		queue_delayed_work(priv->wq, dwork,
+				   msecs_to_jiffies(MLX5E_UPDATE_STATS_INTERVAL));
 	}
 	mutex_unlock(&priv->state_lock);
 }
@@ -280,7 +250,7 @@ static void mlx5e_async_event(struct mlx5_core_dev *mdev, void *vpriv,
 	switch (event) {
 	case MLX5_DEV_EVENT_PORT_UP:
 	case MLX5_DEV_EVENT_PORT_DOWN:
-		schedule_work(&priv->update_carrier_work);
+		queue_work(priv->wq, &priv->update_carrier_work);
 		break;
 
 	default:
@@ -310,6 +280,7 @@ static int mlx5e_create_rq(struct mlx5e_channel *c,
 	struct mlx5_core_dev *mdev = priv->mdev;
 	void *rqc = param->rqc;
 	void *rqc_wq = MLX5_ADDR_OF(rqc, rqc, wq);
+	u32 byte_count;
 	int wq_sz;
 	int err;
 	int i;
@@ -324,32 +295,56 @@ static int mlx5e_create_rq(struct mlx5e_channel *c,
 	rq->wq.db = &rq->wq.db[MLX5_RCV_DBR];
 
 	wq_sz = mlx5_wq_ll_get_size(&rq->wq);
-	rq->skb = kzalloc_node(wq_sz * sizeof(*rq->skb), GFP_KERNEL,
-			       cpu_to_node(c->cpu));
-	if (!rq->skb) {
-		err = -ENOMEM;
-		goto err_rq_wq_destroy;
-	}
 
-	rq->wqe_sz = (priv->params.lro_en) ? priv->params.lro_wqe_sz :
-					     MLX5E_SW2HW_MTU(priv->netdev->mtu);
-	rq->wqe_sz = SKB_DATA_ALIGN(rq->wqe_sz + MLX5E_NET_IP_ALIGN);
+	switch (priv->params.rq_wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		rq->wqe_info = kzalloc_node(wq_sz * sizeof(*rq->wqe_info),
+					    GFP_KERNEL, cpu_to_node(c->cpu));
+		if (!rq->wqe_info) {
+			err = -ENOMEM;
+			goto err_rq_wq_destroy;
+		}
+		rq->handle_rx_cqe = mlx5e_handle_rx_cqe_mpwrq;
+		rq->alloc_wqe = mlx5e_alloc_rx_mpwqe;
+
+		rq->mpwqe_stride_sz = BIT(priv->params.mpwqe_log_stride_sz);
+		rq->mpwqe_num_strides = BIT(priv->params.mpwqe_log_num_strides);
+		rq->wqe_sz = rq->mpwqe_stride_sz * rq->mpwqe_num_strides;
+		byte_count = rq->wqe_sz;
+		break;
+	default: /* MLX5_WQ_TYPE_LINKED_LIST */
+		rq->skb = kzalloc_node(wq_sz * sizeof(*rq->skb), GFP_KERNEL,
+				       cpu_to_node(c->cpu));
+		if (!rq->skb) {
+			err = -ENOMEM;
+			goto err_rq_wq_destroy;
+		}
+		rq->handle_rx_cqe = mlx5e_handle_rx_cqe;
+		rq->alloc_wqe = mlx5e_alloc_rx_wqe;
+
+		rq->wqe_sz = (priv->params.lro_en) ?
+				priv->params.lro_wqe_sz :
+				MLX5E_SW2HW_MTU(priv->netdev->mtu);
+		rq->wqe_sz = SKB_DATA_ALIGN(rq->wqe_sz);
+		byte_count = rq->wqe_sz;
+		byte_count |= MLX5_HW_START_PADDING;
+	}
 
 	for (i = 0; i < wq_sz; i++) {
 		struct mlx5e_rx_wqe *wqe = mlx5_wq_ll_get_wqe(&rq->wq, i);
-		u32 byte_count = rq->wqe_sz - MLX5E_NET_IP_ALIGN;
 
-		wqe->data.lkey       = c->mkey_be;
-		wqe->data.byte_count =
-			cpu_to_be32(byte_count | MLX5_HW_START_PADDING);
+		wqe->data.byte_count = cpu_to_be32(byte_count);
 	}
 
+	rq->wq_type = priv->params.rq_wq_type;
 	rq->pdev    = c->pdev;
 	rq->netdev  = c->netdev;
 	rq->tstamp  = &priv->tstamp;
 	rq->channel = c;
 	rq->ix      = c->ix;
 	rq->priv    = c->priv;
+	rq->mkey_be = c->mkey_be;
+	rq->umr_mkey_be = cpu_to_be32(c->priv->umr_mkey.key);
 
 	return 0;
 
@@ -361,7 +356,14 @@ err_rq_wq_destroy:
 
 static void mlx5e_destroy_rq(struct mlx5e_rq *rq)
 {
-	kfree(rq->skb);
+	switch (rq->wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		kfree(rq->wqe_info);
+		break;
+	default: /* MLX5_WQ_TYPE_LINKED_LIST */
+		kfree(rq->skb);
+	}
+
 	mlx5_wq_destroy(&rq->wq_ctrl);
 }
 
@@ -390,6 +392,7 @@ static int mlx5e_enable_rq(struct mlx5e_rq *rq, struct mlx5e_rq_param *param)
 	MLX5_SET(rqc,  rqc, cqn,		rq->cq.mcq.cqn);
 	MLX5_SET(rqc,  rqc, state,		MLX5_RQC_STATE_RST);
 	MLX5_SET(rqc,  rqc, flush_in_error_en,	1);
+	MLX5_SET(rqc,  rqc, vsd, priv->params.vlan_strip_disable);
 	MLX5_SET(wq,   wq,  log_wq_pg_sz,	rq->wq_ctrl.buf.page_shift -
 						MLX5_ADAPTER_PAGE_SHIFT);
 	MLX5_SET64(wq, wq,  dbr_addr,		rq->wq_ctrl.db.dma);
@@ -404,7 +407,8 @@ static int mlx5e_enable_rq(struct mlx5e_rq *rq, struct mlx5e_rq_param *param)
 	return err;
 }
 
-static int mlx5e_modify_rq(struct mlx5e_rq *rq, int curr_state, int next_state)
+static int mlx5e_modify_rq_state(struct mlx5e_rq *rq, int curr_state,
+				 int next_state)
 {
 	struct mlx5e_channel *c = rq->channel;
 	struct mlx5e_priv *priv = c->priv;
@@ -432,6 +436,36 @@ static int mlx5e_modify_rq(struct mlx5e_rq *rq, int curr_state, int next_state)
 	return err;
 }
 
+static int mlx5e_modify_rq_vsd(struct mlx5e_rq *rq, bool vsd)
+{
+	struct mlx5e_channel *c = rq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *rqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_rq_in);
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	rqc = MLX5_ADDR_OF(modify_rq_in, in, ctx);
+
+	MLX5_SET(modify_rq_in, in, rq_state, MLX5_RQC_STATE_RDY);
+	MLX5_SET64(modify_rq_in, in, modify_bitmask, MLX5_RQ_BITMASK_VSD);
+	MLX5_SET(rqc, rqc, vsd, vsd);
+	MLX5_SET(rqc, rqc, state, MLX5_RQC_STATE_RDY);
+
+	err = mlx5_core_modify_rq(mdev, rq->rqn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
 static void mlx5e_disable_rq(struct mlx5e_rq *rq)
 {
 	mlx5_core_destroy_rq(rq->priv->mdev, rq->rqn);
@@ -458,6 +492,8 @@ static int mlx5e_open_rq(struct mlx5e_channel *c,
 			 struct mlx5e_rq_param *param,
 			 struct mlx5e_rq *rq)
 {
+	struct mlx5e_sq *sq = &c->icosq;
+	u16 pi = sq->pc & sq->wq.sz_m1;
 	int err;
 
 	err = mlx5e_create_rq(c, param, rq);
@@ -468,12 +504,15 @@ static int mlx5e_open_rq(struct mlx5e_channel *c,
 	if (err)
 		goto err_destroy_rq;
 
-	err = mlx5e_modify_rq(rq, MLX5_RQC_STATE_RST, MLX5_RQC_STATE_RDY);
+	err = mlx5e_modify_rq_state(rq, MLX5_RQC_STATE_RST, MLX5_RQC_STATE_RDY);
 	if (err)
 		goto err_disable_rq;
 
 	set_bit(MLX5E_RQ_STATE_POST_WQES_ENABLE, &rq->state);
-	mlx5e_send_nop(&c->sq[0], true); /* trigger mlx5e_post_rx_wqes() */
+
+	sq->ico_wqe_info[pi].opcode     = MLX5_OPCODE_NOP;
+	sq->ico_wqe_info[pi].num_wqebbs = 1;
+	mlx5e_send_nop(sq, true); /* trigger mlx5e_post_rx_wqes() */
 
 	return 0;
 
@@ -490,7 +529,7 @@ static void mlx5e_close_rq(struct mlx5e_rq *rq)
 	clear_bit(MLX5E_RQ_STATE_POST_WQES_ENABLE, &rq->state);
 	napi_synchronize(&rq->channel->napi); /* prevent mlx5e_post_rx_wqes */
 
-	mlx5e_modify_rq(rq, MLX5_RQC_STATE_RDY, MLX5_RQC_STATE_ERR);
+	mlx5e_modify_rq_state(rq, MLX5_RQC_STATE_RDY, MLX5_RQC_STATE_ERR);
 	while (!mlx5_wq_ll_is_empty(&rq->wq))
 		msleep(20);
 
@@ -539,7 +578,6 @@ static int mlx5e_create_sq(struct mlx5e_channel *c,
 
 	void *sqc = param->sqc;
 	void *sqc_wq = MLX5_ADDR_OF(sqc, sqc, wq);
-	int txq_ix;
 	int err;
 
 	err = mlx5_alloc_map_uar(mdev, &sq->uar, true);
@@ -567,8 +605,24 @@ static int mlx5e_create_sq(struct mlx5e_channel *c,
 	if (err)
 		goto err_sq_wq_destroy;
 
-	txq_ix = c->ix + tc * priv->params.num_channels;
-	sq->txq = netdev_get_tx_queue(priv->netdev, txq_ix);
+	if (param->icosq) {
+		u8 wq_sz = mlx5_wq_cyc_get_size(&sq->wq);
+
+		sq->ico_wqe_info = kzalloc_node(sizeof(*sq->ico_wqe_info) *
+						wq_sz,
+						GFP_KERNEL,
+						cpu_to_node(c->cpu));
+		if (!sq->ico_wqe_info) {
+			err = -ENOMEM;
+			goto err_free_sq_db;
+		}
+	} else {
+		int txq_ix;
+
+		txq_ix = c->ix + tc * priv->params.num_channels;
+		sq->txq = netdev_get_tx_queue(priv->netdev, txq_ix);
+		priv->txq_to_sq_map[txq_ix] = sq;
+	}
 
 	sq->pdev      = c->pdev;
 	sq->tstamp    = &priv->tstamp;
@@ -577,10 +631,12 @@ static int mlx5e_create_sq(struct mlx5e_channel *c,
 	sq->tc        = tc;
 	sq->edge      = (sq->wq.sz_m1 + 1) - MLX5_SEND_WQE_MAX_WQEBBS;
 	sq->bf_budget = MLX5E_SQ_BF_BUDGET;
-	priv->txq_to_sq_map[txq_ix] = sq;
 
 	return 0;
 
+err_free_sq_db:
+	mlx5e_free_sq_db(sq);
+
 err_sq_wq_destroy:
 	mlx5_wq_destroy(&sq->wq_ctrl);
 
@@ -595,6 +651,7 @@ static void mlx5e_destroy_sq(struct mlx5e_sq *sq)
 	struct mlx5e_channel *c = sq->channel;
 	struct mlx5e_priv *priv = c->priv;
 
+	kfree(sq->ico_wqe_info);
 	mlx5e_free_sq_db(sq);
 	mlx5_wq_destroy(&sq->wq_ctrl);
 	mlx5_unmap_free_uar(priv->mdev, &sq->uar);
@@ -623,10 +680,10 @@ static int mlx5e_enable_sq(struct mlx5e_sq *sq, struct mlx5e_sq_param *param)
 
 	memcpy(sqc, param->sqc, sizeof(param->sqc));
 
-	MLX5_SET(sqc,  sqc, tis_num_0,		priv->tisn[sq->tc]);
-	MLX5_SET(sqc,  sqc, cqn,		c->sq[sq->tc].cq.mcq.cqn);
+	MLX5_SET(sqc,  sqc, tis_num_0, param->icosq ? 0 : priv->tisn[sq->tc]);
+	MLX5_SET(sqc,  sqc, cqn,		sq->cq.mcq.cqn);
 	MLX5_SET(sqc,  sqc, state,		MLX5_SQC_STATE_RST);
-	MLX5_SET(sqc,  sqc, tis_lst_sz,		1);
+	MLX5_SET(sqc,  sqc, tis_lst_sz,		param->icosq ? 0 : 1);
 	MLX5_SET(sqc,  sqc, flush_in_error_en,	1);
 
 	MLX5_SET(wq,   wq, wq_type,       MLX5_WQ_TYPE_CYCLIC);
@@ -701,9 +758,11 @@ static int mlx5e_open_sq(struct mlx5e_channel *c,
 	if (err)
 		goto err_disable_sq;
 
-	set_bit(MLX5E_SQ_STATE_WAKE_TXQ_ENABLE, &sq->state);
-	netdev_tx_reset_queue(sq->txq);
-	netif_tx_start_queue(sq->txq);
+	if (sq->txq) {
+		set_bit(MLX5E_SQ_STATE_WAKE_TXQ_ENABLE, &sq->state);
+		netdev_tx_reset_queue(sq->txq);
+		netif_tx_start_queue(sq->txq);
+	}
 
 	return 0;
 
@@ -724,15 +783,19 @@ static inline void netif_tx_disable_queue(struct netdev_queue *txq)
 
 static void mlx5e_close_sq(struct mlx5e_sq *sq)
 {
-	clear_bit(MLX5E_SQ_STATE_WAKE_TXQ_ENABLE, &sq->state);
-	napi_synchronize(&sq->channel->napi); /* prevent netif_tx_wake_queue */
-	netif_tx_disable_queue(sq->txq);
+	if (sq->txq) {
+		clear_bit(MLX5E_SQ_STATE_WAKE_TXQ_ENABLE, &sq->state);
+		/* prevent netif_tx_wake_queue */
+		napi_synchronize(&sq->channel->napi);
+		netif_tx_disable_queue(sq->txq);
 
-	/* ensure hw is notified of all pending wqes */
-	if (mlx5e_sq_has_room_for(sq, 1))
-		mlx5e_send_nop(sq, true);
+		/* ensure hw is notified of all pending wqes */
+		if (mlx5e_sq_has_room_for(sq, 1))
+			mlx5e_send_nop(sq, true);
+
+		mlx5e_modify_sq(sq, MLX5_SQC_STATE_RDY, MLX5_SQC_STATE_ERR);
+	}
 
-	mlx5e_modify_sq(sq, MLX5_SQC_STATE_RDY, MLX5_SQC_STATE_ERR);
 	while (sq->cc != sq->pc) /* wait till sq is empty */
 		msleep(20);
 
@@ -986,10 +1049,14 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
 
 	netif_napi_add(netdev, &c->napi, mlx5e_napi_poll, 64);
 
-	err = mlx5e_open_tx_cqs(c, cparam);
+	err = mlx5e_open_cq(c, &cparam->icosq_cq, &c->icosq.cq, 0, 0);
 	if (err)
 		goto err_napi_del;
 
+	err = mlx5e_open_tx_cqs(c, cparam);
+	if (err)
+		goto err_close_icosq_cq;
+
 	err = mlx5e_open_cq(c, &cparam->rx_cq, &c->rq.cq,
 			    priv->params.rx_cq_moderation_usec,
 			    priv->params.rx_cq_moderation_pkts);
@@ -998,10 +1065,14 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
 
 	napi_enable(&c->napi);
 
-	err = mlx5e_open_sqs(c, cparam);
+	err = mlx5e_open_sq(c, 0, &cparam->icosq, &c->icosq);
 	if (err)
 		goto err_disable_napi;
 
+	err = mlx5e_open_sqs(c, cparam);
+	if (err)
+		goto err_close_icosq;
+
 	err = mlx5e_open_rq(c, &cparam->rq, &c->rq);
 	if (err)
 		goto err_close_sqs;
@@ -1014,6 +1085,9 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
 err_close_sqs:
 	mlx5e_close_sqs(c);
 
+err_close_icosq:
+	mlx5e_close_sq(&c->icosq);
+
 err_disable_napi:
 	napi_disable(&c->napi);
 	mlx5e_close_cq(&c->rq.cq);
@@ -1021,6 +1095,9 @@ err_disable_napi:
 err_close_tx_cqs:
 	mlx5e_close_tx_cqs(c);
 
+err_close_icosq_cq:
+	mlx5e_close_cq(&c->icosq.cq);
+
 err_napi_del:
 	netif_napi_del(&c->napi);
 	napi_hash_del(&c->napi);
@@ -1033,9 +1110,11 @@ static void mlx5e_close_channel(struct mlx5e_channel *c)
 {
 	mlx5e_close_rq(&c->rq);
 	mlx5e_close_sqs(c);
+	mlx5e_close_sq(&c->icosq);
 	napi_disable(&c->napi);
 	mlx5e_close_cq(&c->rq.cq);
 	mlx5e_close_tx_cqs(c);
+	mlx5e_close_cq(&c->icosq.cq);
 	netif_napi_del(&c->napi);
 
 	napi_hash_del(&c->napi);
@@ -1050,11 +1129,23 @@ static void mlx5e_build_rq_param(struct mlx5e_priv *priv,
 	void *rqc = param->rqc;
 	void *wq = MLX5_ADDR_OF(rqc, rqc, wq);
 
-	MLX5_SET(wq, wq, wq_type,          MLX5_WQ_TYPE_LINKED_LIST);
+	switch (priv->params.rq_wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		MLX5_SET(wq, wq, log_wqe_num_of_strides,
+			 priv->params.mpwqe_log_num_strides - 9);
+		MLX5_SET(wq, wq, log_wqe_stride_size,
+			 priv->params.mpwqe_log_stride_sz - 6);
+		MLX5_SET(wq, wq, wq_type, MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ);
+		break;
+	default: /* MLX5_WQ_TYPE_LINKED_LIST */
+		MLX5_SET(wq, wq, wq_type, MLX5_WQ_TYPE_LINKED_LIST);
+	}
+
 	MLX5_SET(wq, wq, end_padding_mode, MLX5_WQ_END_PAD_MODE_ALIGN);
 	MLX5_SET(wq, wq, log_wq_stride,    ilog2(sizeof(struct mlx5e_rx_wqe)));
 	MLX5_SET(wq, wq, log_wq_sz,        priv->params.log_rq_size);
 	MLX5_SET(wq, wq, pd,               priv->pdn);
+	MLX5_SET(rqc, rqc, counter_set_id, priv->q_counter);
 
 	param->wq.buf_numa_node = dev_to_node(&priv->mdev->pdev->dev);
 	param->wq.linear = 1;
@@ -1069,17 +1160,27 @@ static void mlx5e_build_drop_rq_param(struct mlx5e_rq_param *param)
 	MLX5_SET(wq, wq, log_wq_stride,    ilog2(sizeof(struct mlx5e_rx_wqe)));
 }
 
-static void mlx5e_build_sq_param(struct mlx5e_priv *priv,
-				 struct mlx5e_sq_param *param)
+static void mlx5e_build_sq_param_common(struct mlx5e_priv *priv,
+					struct mlx5e_sq_param *param)
 {
 	void *sqc = param->sqc;
 	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
 
-	MLX5_SET(wq, wq, log_wq_sz,     priv->params.log_sq_size);
 	MLX5_SET(wq, wq, log_wq_stride, ilog2(MLX5_SEND_WQE_BB));
 	MLX5_SET(wq, wq, pd,            priv->pdn);
 
 	param->wq.buf_numa_node = dev_to_node(&priv->mdev->pdev->dev);
+}
+
+static void mlx5e_build_sq_param(struct mlx5e_priv *priv,
+				 struct mlx5e_sq_param *param)
+{
+	void *sqc = param->sqc;
+	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	mlx5e_build_sq_param_common(priv, param);
+	MLX5_SET(wq, wq, log_wq_sz,     priv->params.log_sq_size);
+
 	param->max_inline = priv->params.tx_max_inline;
 }
 
@@ -1095,8 +1196,22 @@ static void mlx5e_build_rx_cq_param(struct mlx5e_priv *priv,
 				    struct mlx5e_cq_param *param)
 {
 	void *cqc = param->cqc;
+	u8 log_cq_size;
 
-	MLX5_SET(cqc, cqc, log_cq_size,  priv->params.log_rq_size);
+	switch (priv->params.rq_wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		log_cq_size = priv->params.log_rq_size +
+			priv->params.mpwqe_log_num_strides;
+		break;
+	default: /* MLX5_WQ_TYPE_LINKED_LIST */
+		log_cq_size = priv->params.log_rq_size;
+	}
+
+	MLX5_SET(cqc, cqc, log_cq_size, log_cq_size);
+	if (priv->params.rx_cqe_compress) {
+		MLX5_SET(cqc, cqc, mini_cqe_res_format, MLX5_CQE_FORMAT_CSUM);
+		MLX5_SET(cqc, cqc, cqe_comp_en, 1);
+	}
 
 	mlx5e_build_common_cq_param(priv, param);
 }
@@ -1106,25 +1221,52 @@ static void mlx5e_build_tx_cq_param(struct mlx5e_priv *priv,
 {
 	void *cqc = param->cqc;
 
-	MLX5_SET(cqc, cqc, log_cq_size,  priv->params.log_sq_size);
+	MLX5_SET(cqc, cqc, log_cq_size, priv->params.log_sq_size);
+
+	mlx5e_build_common_cq_param(priv, param);
+}
+
+static void mlx5e_build_ico_cq_param(struct mlx5e_priv *priv,
+				     struct mlx5e_cq_param *param,
+				     u8 log_wq_size)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, log_cq_size, log_wq_size);
 
 	mlx5e_build_common_cq_param(priv, param);
 }
 
-static void mlx5e_build_channel_param(struct mlx5e_priv *priv,
-				      struct mlx5e_channel_param *cparam)
+static void mlx5e_build_icosq_param(struct mlx5e_priv *priv,
+				    struct mlx5e_sq_param *param,
+				    u8 log_wq_size)
 {
-	memset(cparam, 0, sizeof(*cparam));
+	void *sqc = param->sqc;
+	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	mlx5e_build_sq_param_common(priv, param);
+
+	MLX5_SET(wq, wq, log_wq_sz, log_wq_size);
+	MLX5_SET(sqc, sqc, reg_umr, MLX5_CAP_ETH(priv->mdev, reg_umr_sq));
+
+	param->icosq = true;
+}
+
+static void mlx5e_build_channel_param(struct mlx5e_priv *priv, struct mlx5e_channel_param *cparam)
+{
+	u8 icosq_log_wq_sz = MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE;
 
 	mlx5e_build_rq_param(priv, &cparam->rq);
 	mlx5e_build_sq_param(priv, &cparam->sq);
+	mlx5e_build_icosq_param(priv, &cparam->icosq, icosq_log_wq_sz);
 	mlx5e_build_rx_cq_param(priv, &cparam->rx_cq);
 	mlx5e_build_tx_cq_param(priv, &cparam->tx_cq);
+	mlx5e_build_ico_cq_param(priv, &cparam->icosq_cq, icosq_log_wq_sz);
 }
 
 static int mlx5e_open_channels(struct mlx5e_priv *priv)
 {
-	struct mlx5e_channel_param cparam;
+	struct mlx5e_channel_param *cparam;
 	int nch = priv->params.num_channels;
 	int err = -ENOMEM;
 	int i;
@@ -1136,12 +1278,15 @@ static int mlx5e_open_channels(struct mlx5e_priv *priv)
 	priv->txq_to_sq_map = kcalloc(nch * priv->params.num_tc,
 				      sizeof(struct mlx5e_sq *), GFP_KERNEL);
 
-	if (!priv->channel || !priv->txq_to_sq_map)
+	cparam = kzalloc(sizeof(struct mlx5e_channel_param), GFP_KERNEL);
+
+	if (!priv->channel || !priv->txq_to_sq_map || !cparam)
 		goto err_free_txq_to_sq_map;
 
-	mlx5e_build_channel_param(priv, &cparam);
+	mlx5e_build_channel_param(priv, cparam);
+
 	for (i = 0; i < nch; i++) {
-		err = mlx5e_open_channel(priv, i, &cparam, &priv->channel[i]);
+		err = mlx5e_open_channel(priv, i, cparam, &priv->channel[i]);
 		if (err)
 			goto err_close_channels;
 	}
@@ -1152,6 +1297,7 @@ static int mlx5e_open_channels(struct mlx5e_priv *priv)
 			goto err_close_channels;
 	}
 
+	kfree(cparam);
 	return 0;
 
 err_close_channels:
@@ -1161,6 +1307,7 @@ err_close_channels:
 err_free_txq_to_sq_map:
 	kfree(priv->txq_to_sq_map);
 	kfree(priv->channel);
+	kfree(cparam);
 
 	return err;
 }
@@ -1200,48 +1347,36 @@ static void mlx5e_fill_indir_rqt_rqns(struct mlx5e_priv *priv, void *rqtc)
 
 	for (i = 0; i < MLX5E_INDIR_RQT_SIZE; i++) {
 		int ix = i;
+		u32 rqn;
 
 		if (priv->params.rss_hfunc == ETH_RSS_HASH_XOR)
 			ix = mlx5e_bits_invert(i, MLX5E_LOG_INDIR_RQT_SIZE);
 
 		ix = priv->params.indirection_rqt[ix];
-		MLX5_SET(rqtc, rqtc, rq_num[i],
-			 test_bit(MLX5E_STATE_OPENED, &priv->state) ?
-			 priv->channel[ix]->rq.rqn :
-			 priv->drop_rq.rqn);
+		rqn = test_bit(MLX5E_STATE_OPENED, &priv->state) ?
+				priv->channel[ix]->rq.rqn :
+				priv->drop_rq.rqn;
+		MLX5_SET(rqtc, rqtc, rq_num[i], rqn);
 	}
 }
 
-static void mlx5e_fill_rqt_rqns(struct mlx5e_priv *priv, void *rqtc,
-				enum mlx5e_rqt_ix rqt_ix)
+static void mlx5e_fill_direct_rqt_rqn(struct mlx5e_priv *priv, void *rqtc,
+				      int ix)
 {
+	u32 rqn = test_bit(MLX5E_STATE_OPENED, &priv->state) ?
+			priv->channel[ix]->rq.rqn :
+			priv->drop_rq.rqn;
 
-	switch (rqt_ix) {
-	case MLX5E_INDIRECTION_RQT:
-		mlx5e_fill_indir_rqt_rqns(priv, rqtc);
-
-		break;
-
-	default: /* MLX5E_SINGLE_RQ_RQT */
-		MLX5_SET(rqtc, rqtc, rq_num[0],
-			 test_bit(MLX5E_STATE_OPENED, &priv->state) ?
-			 priv->channel[0]->rq.rqn :
-			 priv->drop_rq.rqn);
-
-		break;
-	}
+	MLX5_SET(rqtc, rqtc, rq_num[0], rqn);
 }
 
-static int mlx5e_create_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
+static int mlx5e_create_rqt(struct mlx5e_priv *priv, int sz, int ix, u32 *rqtn)
 {
 	struct mlx5_core_dev *mdev = priv->mdev;
-	u32 *in;
 	void *rqtc;
 	int inlen;
-	int sz;
 	int err;
-
-	sz = (rqt_ix == MLX5E_SINGLE_RQ_RQT) ? 1 : MLX5E_INDIR_RQT_SIZE;
+	u32 *in;
 
 	inlen = MLX5_ST_SZ_BYTES(create_rqt_in) + sizeof(u32) * sz;
 	in = mlx5_vzalloc(inlen);
@@ -1253,26 +1388,73 @@ static int mlx5e_create_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
 	MLX5_SET(rqtc, rqtc, rqt_actual_size, sz);
 	MLX5_SET(rqtc, rqtc, rqt_max_size, sz);
 
-	mlx5e_fill_rqt_rqns(priv, rqtc, rqt_ix);
+	if (sz > 1) /* RSS */
+		mlx5e_fill_indir_rqt_rqns(priv, rqtc);
+	else
+		mlx5e_fill_direct_rqt_rqn(priv, rqtc, ix);
 
-	err = mlx5_core_create_rqt(mdev, in, inlen, &priv->rqtn[rqt_ix]);
+	err = mlx5_core_create_rqt(mdev, in, inlen, rqtn);
 
 	kvfree(in);
+	return err;
+}
+
+static void mlx5e_destroy_rqt(struct mlx5e_priv *priv, u32 rqtn)
+{
+	mlx5_core_destroy_rqt(priv->mdev, rqtn);
+}
+
+static int mlx5e_create_rqts(struct mlx5e_priv *priv)
+{
+	int nch = mlx5e_get_max_num_channels(priv->mdev);
+	u32 *rqtn;
+	int err;
+	int ix;
+
+	/* Indirect RQT */
+	rqtn = &priv->indir_rqtn;
+	err = mlx5e_create_rqt(priv, MLX5E_INDIR_RQT_SIZE, 0, rqtn);
+	if (err)
+		return err;
+
+	/* Direct RQTs */
+	for (ix = 0; ix < nch; ix++) {
+		rqtn = &priv->direct_tir[ix].rqtn;
+		err = mlx5e_create_rqt(priv, 1 /*size */, ix, rqtn);
+		if (err)
+			goto err_destroy_rqts;
+	}
+
+	return 0;
+
+err_destroy_rqts:
+	for (ix--; ix >= 0; ix--)
+		mlx5e_destroy_rqt(priv, priv->direct_tir[ix].rqtn);
+
+	mlx5e_destroy_rqt(priv, priv->indir_rqtn);
 
 	return err;
 }
 
-int mlx5e_redirect_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
+static void mlx5e_destroy_rqts(struct mlx5e_priv *priv)
+{
+	int nch = mlx5e_get_max_num_channels(priv->mdev);
+	int i;
+
+	for (i = 0; i < nch; i++)
+		mlx5e_destroy_rqt(priv, priv->direct_tir[i].rqtn);
+
+	mlx5e_destroy_rqt(priv, priv->indir_rqtn);
+}
+
+int mlx5e_redirect_rqt(struct mlx5e_priv *priv, u32 rqtn, int sz, int ix)
 {
 	struct mlx5_core_dev *mdev = priv->mdev;
-	u32 *in;
 	void *rqtc;
 	int inlen;
-	int sz;
+	u32 *in;
 	int err;
 
-	sz = (rqt_ix == MLX5E_SINGLE_RQ_RQT) ? 1 : MLX5E_INDIR_RQT_SIZE;
-
 	inlen = MLX5_ST_SZ_BYTES(modify_rqt_in) + sizeof(u32) * sz;
 	in = mlx5_vzalloc(inlen);
 	if (!in)
@@ -1281,27 +1463,31 @@ int mlx5e_redirect_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
 	rqtc = MLX5_ADDR_OF(modify_rqt_in, in, ctx);
 
 	MLX5_SET(rqtc, rqtc, rqt_actual_size, sz);
-
-	mlx5e_fill_rqt_rqns(priv, rqtc, rqt_ix);
+	if (sz > 1) /* RSS */
+		mlx5e_fill_indir_rqt_rqns(priv, rqtc);
+	else
+		mlx5e_fill_direct_rqt_rqn(priv, rqtc, ix);
 
 	MLX5_SET(modify_rqt_in, in, bitmask.rqn_list, 1);
 
-	err = mlx5_core_modify_rqt(mdev, priv->rqtn[rqt_ix], in, inlen);
+	err = mlx5_core_modify_rqt(mdev, rqtn, in, inlen);
 
 	kvfree(in);
 
 	return err;
 }
 
-static void mlx5e_destroy_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
-{
-	mlx5_core_destroy_rqt(priv->mdev, priv->rqtn[rqt_ix]);
-}
-
 static void mlx5e_redirect_rqts(struct mlx5e_priv *priv)
 {
-	mlx5e_redirect_rqt(priv, MLX5E_INDIRECTION_RQT);
-	mlx5e_redirect_rqt(priv, MLX5E_SINGLE_RQ_RQT);
+	u32 rqtn;
+	int ix;
+
+	rqtn = priv->indir_rqtn;
+	mlx5e_redirect_rqt(priv, rqtn, MLX5E_INDIR_RQT_SIZE, 0);
+	for (ix = 0; ix < priv->params.num_channels; ix++) {
+		rqtn = priv->direct_tir[ix].rqtn;
+		mlx5e_redirect_rqt(priv, rqtn, 1, ix);
+	}
 }
 
 static void mlx5e_build_tir_ctx_lro(void *tirc, struct mlx5e_priv *priv)
@@ -1346,6 +1532,7 @@ static int mlx5e_modify_tirs_lro(struct mlx5e_priv *priv)
 	int inlen;
 	int err;
 	int tt;
+	int ix;
 
 	inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
 	in = mlx5_vzalloc(inlen);
@@ -1357,23 +1544,32 @@ static int mlx5e_modify_tirs_lro(struct mlx5e_priv *priv)
 
 	mlx5e_build_tir_ctx_lro(tirc, priv);
 
-	for (tt = 0; tt < MLX5E_NUM_TT; tt++) {
-		err = mlx5_core_modify_tir(mdev, priv->tirn[tt], in, inlen);
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		err = mlx5_core_modify_tir(mdev, priv->indir_tirn[tt], in,
+					   inlen);
 		if (err)
-			break;
+			goto free_in;
 	}
 
+	for (ix = 0; ix < mlx5e_get_max_num_channels(mdev); ix++) {
+		err = mlx5_core_modify_tir(mdev, priv->direct_tir[ix].tirn,
+					   in, inlen);
+		if (err)
+			goto free_in;
+	}
+
+free_in:
 	kvfree(in);
 
 	return err;
 }
 
-static int mlx5e_refresh_tir_self_loopback_enable(struct mlx5_core_dev *mdev,
-						  u32 tirn)
+static int mlx5e_refresh_tirs_self_loopback_enable(struct mlx5e_priv *priv)
 {
 	void *in;
 	int inlen;
 	int err;
+	int i;
 
 	inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
 	in = mlx5_vzalloc(inlen);
@@ -1382,46 +1578,70 @@ static int mlx5e_refresh_tir_self_loopback_enable(struct mlx5_core_dev *mdev,
 
 	MLX5_SET(modify_tir_in, in, bitmask.self_lb_en, 1);
 
-	err = mlx5_core_modify_tir(mdev, tirn, in, inlen);
+	for (i = 0; i < MLX5E_NUM_INDIR_TIRS; i++) {
+		err = mlx5_core_modify_tir(priv->mdev, priv->indir_tirn[i], in,
+					   inlen);
+		if (err)
+			return err;
+	}
+
+	for (i = 0; i < priv->params.num_channels; i++) {
+		err = mlx5_core_modify_tir(priv->mdev,
+					   priv->direct_tir[i].tirn, in,
+					   inlen);
+		if (err)
+			return err;
+	}
 
 	kvfree(in);
 
-	return err;
+	return 0;
 }
 
-static int mlx5e_refresh_tirs_self_loopback_enable(struct mlx5e_priv *priv)
+static int mlx5e_set_mtu(struct mlx5e_priv *priv, u16 mtu)
 {
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 hw_mtu = MLX5E_SW2HW_MTU(mtu);
 	int err;
-	int i;
 
-	for (i = 0; i < MLX5E_NUM_TT; i++) {
-		err = mlx5e_refresh_tir_self_loopback_enable(priv->mdev,
-							     priv->tirn[i]);
-		if (err)
-			return err;
-	}
+	err = mlx5_set_port_mtu(mdev, hw_mtu, 1);
+	if (err)
+		return err;
 
+	/* Update vport context MTU */
+	mlx5_modify_nic_vport_mtu(mdev, hw_mtu);
 	return 0;
 }
 
+static void mlx5e_query_mtu(struct mlx5e_priv *priv, u16 *mtu)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u16 hw_mtu = 0;
+	int err;
+
+	err = mlx5_query_nic_vport_mtu(mdev, &hw_mtu);
+	if (err || !hw_mtu) /* fallback to port oper mtu */
+		mlx5_query_port_oper_mtu(mdev, &hw_mtu, 1);
+
+	*mtu = MLX5E_HW2SW_MTU(hw_mtu);
+}
+
 static int mlx5e_set_dev_port_mtu(struct net_device *netdev)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
-	struct mlx5_core_dev *mdev = priv->mdev;
-	int hw_mtu;
+	u16 mtu;
 	int err;
 
-	err = mlx5_set_port_mtu(mdev, MLX5E_SW2HW_MTU(netdev->mtu), 1);
+	err = mlx5e_set_mtu(priv, netdev->mtu);
 	if (err)
 		return err;
 
-	mlx5_query_port_oper_mtu(mdev, &hw_mtu, 1);
+	mlx5e_query_mtu(priv, &mtu);
+	if (mtu != netdev->mtu)
+		netdev_warn(netdev, "%s: VPort MTU %d is different than netdev mtu %d\n",
+			    __func__, mtu, netdev->mtu);
 
-	if (MLX5E_HW2SW_MTU(hw_mtu) != netdev->mtu)
-		netdev_warn(netdev, "%s: Port MTU %d is different than netdev mtu %d\n",
-			    __func__, MLX5E_HW2SW_MTU(hw_mtu), netdev->mtu);
-
-	netdev->mtu = MLX5E_HW2SW_MTU(hw_mtu);
+	netdev->mtu = mtu;
 	return 0;
 }
 
@@ -1478,8 +1698,11 @@ int mlx5e_open_locked(struct net_device *netdev)
 	mlx5e_redirect_rqts(priv);
 	mlx5e_update_carrier(priv);
 	mlx5e_timestamp_init(priv);
+#ifdef CONFIG_RFS_ACCEL
+	priv->netdev->rx_cpu_rmap = priv->mdev->rmap;
+#endif
 
-	schedule_delayed_work(&priv->update_stats_work, 0);
+	queue_delayed_work(priv->wq, &priv->update_stats_work, 0);
 
 	return 0;
 
@@ -1685,7 +1908,8 @@ static void mlx5e_destroy_tises(struct mlx5e_priv *priv)
 		mlx5e_destroy_tis(priv, tc);
 }
 
-static void mlx5e_build_tir_ctx(struct mlx5e_priv *priv, u32 *tirc, int tt)
+static void mlx5e_build_indir_tir_ctx(struct mlx5e_priv *priv, u32 *tirc,
+				      enum mlx5e_traffic_types tt)
 {
 	void *hfso = MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer);
 
@@ -1706,19 +1930,8 @@ static void mlx5e_build_tir_ctx(struct mlx5e_priv *priv, u32 *tirc, int tt)
 	mlx5e_build_tir_ctx_lro(tirc, priv);
 
 	MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT);
-
-	switch (tt) {
-	case MLX5E_TT_ANY:
-		MLX5_SET(tirc, tirc, indirect_table,
-			 priv->rqtn[MLX5E_SINGLE_RQ_RQT]);
-		MLX5_SET(tirc, tirc, rx_hash_fn, MLX5_RX_HASH_FN_INVERTED_XOR8);
-		break;
-	default:
-		MLX5_SET(tirc, tirc, indirect_table,
-			 priv->rqtn[MLX5E_INDIRECTION_RQT]);
-		mlx5e_build_tir_ctx_hash(tirc, priv);
-		break;
-	}
+	MLX5_SET(tirc, tirc, indirect_table, priv->indir_rqtn);
+	mlx5e_build_tir_ctx_hash(tirc, priv);
 
 	switch (tt) {
 	case MLX5E_TT_IPV4_TCP:
@@ -1798,64 +2011,107 @@ static void mlx5e_build_tir_ctx(struct mlx5e_priv *priv, u32 *tirc, int tt)
 		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
 			 MLX5_HASH_IP);
 		break;
+	default:
+		WARN_ONCE(true,
+			  "mlx5e_build_indir_tir_ctx: bad traffic type!\n");
 	}
 }
 
-static int mlx5e_create_tir(struct mlx5e_priv *priv, int tt)
+static void mlx5e_build_direct_tir_ctx(struct mlx5e_priv *priv, u32 *tirc,
+				       u32 rqtn)
 {
-	struct mlx5_core_dev *mdev = priv->mdev;
-	u32 *in;
+	MLX5_SET(tirc, tirc, transport_domain, priv->tdn);
+
+	mlx5e_build_tir_ctx_lro(tirc, priv);
+
+	MLX5_SET(tirc, tirc, disp_type, MLX5_TIRC_DISP_TYPE_INDIRECT);
+	MLX5_SET(tirc, tirc, indirect_table, rqtn);
+	MLX5_SET(tirc, tirc, rx_hash_fn, MLX5_RX_HASH_FN_INVERTED_XOR8);
+}
+
+static int mlx5e_create_tirs(struct mlx5e_priv *priv)
+{
+	int nch = mlx5e_get_max_num_channels(priv->mdev);
 	void *tirc;
 	int inlen;
+	u32 *tirn;
 	int err;
+	u32 *in;
+	int ix;
+	int tt;
 
 	inlen = MLX5_ST_SZ_BYTES(create_tir_in);
 	in = mlx5_vzalloc(inlen);
 	if (!in)
 		return -ENOMEM;
 
-	tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+	/* indirect tirs */
+	for (tt = 0; tt < MLX5E_NUM_INDIR_TIRS; tt++) {
+		memset(in, 0, inlen);
+		tirn = &priv->indir_tirn[tt];
+		tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+		mlx5e_build_indir_tir_ctx(priv, tirc, tt);
+		err = mlx5_core_create_tir(priv->mdev, in, inlen, tirn);
+		if (err)
+			goto err_destroy_tirs;
+	}
+
+	/* direct tirs */
+	for (ix = 0; ix < nch; ix++) {
+		memset(in, 0, inlen);
+		tirn = &priv->direct_tir[ix].tirn;
+		tirc = MLX5_ADDR_OF(create_tir_in, in, ctx);
+		mlx5e_build_direct_tir_ctx(priv, tirc,
+					   priv->direct_tir[ix].rqtn);
+		err = mlx5_core_create_tir(priv->mdev, in, inlen, tirn);
+		if (err)
+			goto err_destroy_ch_tirs;
+	}
+
+	kvfree(in);
+
+	return 0;
 
-	mlx5e_build_tir_ctx(priv, tirc, tt);
+err_destroy_ch_tirs:
+	for (ix--; ix >= 0; ix--)
+		mlx5_core_destroy_tir(priv->mdev, priv->direct_tir[ix].tirn);
 
-	err = mlx5_core_create_tir(mdev, in, inlen, &priv->tirn[tt]);
+err_destroy_tirs:
+	for (tt--; tt >= 0; tt--)
+		mlx5_core_destroy_tir(priv->mdev, priv->indir_tirn[tt]);
 
 	kvfree(in);
 
 	return err;
 }
 
-static void mlx5e_destroy_tir(struct mlx5e_priv *priv, int tt)
+static void mlx5e_destroy_tirs(struct mlx5e_priv *priv)
 {
-	mlx5_core_destroy_tir(priv->mdev, priv->tirn[tt]);
+	int nch = mlx5e_get_max_num_channels(priv->mdev);
+	int i;
+
+	for (i = 0; i < nch; i++)
+		mlx5_core_destroy_tir(priv->mdev, priv->direct_tir[i].tirn);
+
+	for (i = 0; i < MLX5E_NUM_INDIR_TIRS; i++)
+		mlx5_core_destroy_tir(priv->mdev, priv->indir_tirn[i]);
 }
 
-static int mlx5e_create_tirs(struct mlx5e_priv *priv)
+int mlx5e_modify_rqs_vsd(struct mlx5e_priv *priv, bool vsd)
 {
-	int err;
+	int err = 0;
 	int i;
 
-	for (i = 0; i < MLX5E_NUM_TT; i++) {
-		err = mlx5e_create_tir(priv, i);
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		return 0;
+
+	for (i = 0; i < priv->params.num_channels; i++) {
+		err = mlx5e_modify_rq_vsd(&priv->channel[i]->rq, vsd);
 		if (err)
-			goto err_destroy_tirs;
+			return err;
 	}
 
 	return 0;
-
-err_destroy_tirs:
-	for (i--; i >= 0; i--)
-		mlx5e_destroy_tir(priv, i);
-
-	return err;
-}
-
-static void mlx5e_destroy_tirs(struct mlx5e_priv *priv)
-{
-	int i;
-
-	for (i = 0; i < MLX5E_NUM_TT; i++)
-		mlx5e_destroy_tir(priv, i);
 }
 
 static int mlx5e_setup_tc(struct net_device *netdev, u8 tc)
@@ -1898,6 +2154,8 @@ static int mlx5e_ndo_setup_tc(struct net_device *dev, u32 handle,
 			return mlx5e_configure_flower(priv, proto, tc->cls_flower);
 		case TC_CLSFLOWER_DESTROY:
 			return mlx5e_delete_flower(priv, tc->cls_flower);
+		case TC_CLSFLOWER_STATS:
+			return mlx5e_stats_flower(priv, tc->cls_flower);
 		}
 	default:
 		return -EOPNOTSUPP;
@@ -1914,19 +2172,37 @@ static struct rtnl_link_stats64 *
 mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
 {
 	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_sw_stats *sstats = &priv->stats.sw;
 	struct mlx5e_vport_stats *vstats = &priv->stats.vport;
-
-	stats->rx_packets = vstats->rx_packets;
-	stats->rx_bytes   = vstats->rx_bytes;
-	stats->tx_packets = vstats->tx_packets;
-	stats->tx_bytes   = vstats->tx_bytes;
-	stats->multicast  = vstats->rx_multicast_packets +
-			    vstats->tx_multicast_packets;
-	stats->tx_errors  = vstats->tx_error_packets;
-	stats->rx_errors  = vstats->rx_error_packets;
-	stats->tx_dropped = vstats->tx_queue_dropped;
-	stats->rx_crc_errors = 0;
-	stats->rx_length_errors = 0;
+	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
+
+	stats->rx_packets = sstats->rx_packets;
+	stats->rx_bytes   = sstats->rx_bytes;
+	stats->tx_packets = sstats->tx_packets;
+	stats->tx_bytes   = sstats->tx_bytes;
+
+	stats->rx_dropped = priv->stats.qcnt.rx_out_of_buffer;
+	stats->tx_dropped = sstats->tx_queue_dropped;
+
+	stats->rx_length_errors =
+		PPORT_802_3_GET(pstats, a_in_range_length_errors) +
+		PPORT_802_3_GET(pstats, a_out_of_range_length_field) +
+		PPORT_802_3_GET(pstats, a_frame_too_long_errors);
+	stats->rx_crc_errors =
+		PPORT_802_3_GET(pstats, a_frame_check_sequence_errors);
+	stats->rx_frame_errors = PPORT_802_3_GET(pstats, a_alignment_errors);
+	stats->tx_aborted_errors = PPORT_2863_GET(pstats, if_out_discards);
+	stats->tx_carrier_errors =
+		PPORT_802_3_GET(pstats, a_symbol_error_during_carrier);
+	stats->rx_errors = stats->rx_length_errors + stats->rx_crc_errors +
+			   stats->rx_frame_errors;
+	stats->tx_errors = stats->tx_aborted_errors + stats->tx_carrier_errors;
+
+	/* vport multicast also counts packets that are dropped due to steering
+	 * or rx out of buffer
+	 */
+	stats->multicast =
+		VPORT_COUNTER_GET(vstats, received_eth_multicast.packets);
 
 	return stats;
 }
@@ -1935,7 +2211,7 @@ static void mlx5e_set_rx_mode(struct net_device *dev)
 {
 	struct mlx5e_priv *priv = netdev_priv(dev);
 
-	schedule_work(&priv->set_rx_mode_work);
+	queue_work(priv->wq, &priv->set_rx_mode_work);
 }
 
 static int mlx5e_set_mac(struct net_device *netdev, void *addr)
@@ -1950,71 +2226,180 @@ static int mlx5e_set_mac(struct net_device *netdev, void *addr)
 	ether_addr_copy(netdev->dev_addr, saddr->sa_data);
 	netif_addr_unlock_bh(netdev);
 
-	schedule_work(&priv->set_rx_mode_work);
+	queue_work(priv->wq, &priv->set_rx_mode_work);
 
 	return 0;
 }
 
-static int mlx5e_set_features(struct net_device *netdev,
-			      netdev_features_t features)
+#define MLX5E_SET_FEATURE(netdev, feature, enable)	\
+	do {						\
+		if (enable)				\
+			netdev->features |= feature;	\
+		else					\
+			netdev->features &= ~feature;	\
+	} while (0)
+
+typedef int (*mlx5e_feature_handler)(struct net_device *netdev, bool enable);
+
+static int set_feature_lro(struct net_device *netdev, bool enable)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
-	int err = 0;
-	netdev_features_t changes = features ^ netdev->features;
+	bool was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+	int err;
 
 	mutex_lock(&priv->state_lock);
 
-	if (changes & NETIF_F_LRO) {
-		bool was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
-
-		if (was_opened)
-			mlx5e_close_locked(priv->netdev);
-
-		priv->params.lro_en = !!(features & NETIF_F_LRO);
-		err = mlx5e_modify_tirs_lro(priv);
-		if (err)
-			mlx5_core_warn(priv->mdev, "lro modify failed, %d\n",
-				       err);
+	if (was_opened && (priv->params.rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST))
+		mlx5e_close_locked(priv->netdev);
 
-		if (was_opened)
-			err = mlx5e_open_locked(priv->netdev);
+	priv->params.lro_en = enable;
+	err = mlx5e_modify_tirs_lro(priv);
+	if (err) {
+		netdev_err(netdev, "lro modify failed, %d\n", err);
+		priv->params.lro_en = !enable;
 	}
 
+	if (was_opened && (priv->params.rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST))
+		mlx5e_open_locked(priv->netdev);
+
 	mutex_unlock(&priv->state_lock);
 
-	if (changes & NETIF_F_HW_VLAN_CTAG_FILTER) {
-		if (features & NETIF_F_HW_VLAN_CTAG_FILTER)
-			mlx5e_enable_vlan_filter(priv);
-		else
-			mlx5e_disable_vlan_filter(priv);
-	}
+	return err;
+}
+
+static int set_feature_vlan_filter(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (enable)
+		mlx5e_enable_vlan_filter(priv);
+	else
+		mlx5e_disable_vlan_filter(priv);
 
-	if ((changes & NETIF_F_HW_TC) && !(features & NETIF_F_HW_TC) &&
-	    mlx5e_tc_num_filters(priv)) {
+	return 0;
+}
+
+static int set_feature_tc_num_filters(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (!enable && mlx5e_tc_num_filters(priv)) {
 		netdev_err(netdev,
 			   "Active offloaded tc filters, can't turn hw_tc_offload off\n");
 		return -EINVAL;
 	}
 
+	return 0;
+}
+
+static int set_feature_rx_all(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_set_port_fcs(mdev, !enable);
+}
+
+static int set_feature_rx_vlan(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+
+	priv->params.vlan_strip_disable = !enable;
+	err = mlx5e_modify_rqs_vsd(priv, !enable);
+	if (err)
+		priv->params.vlan_strip_disable = enable;
+
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+#ifdef CONFIG_RFS_ACCEL
+static int set_feature_arfs(struct net_device *netdev, bool enable)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	if (enable)
+		err = mlx5e_arfs_enable(priv);
+	else
+		err = mlx5e_arfs_disable(priv);
+
 	return err;
 }
+#endif
+
+static int mlx5e_handle_feature(struct net_device *netdev,
+				netdev_features_t wanted_features,
+				netdev_features_t feature,
+				mlx5e_feature_handler feature_handler)
+{
+	netdev_features_t changes = wanted_features ^ netdev->features;
+	bool enable = !!(wanted_features & feature);
+	int err;
+
+	if (!(changes & feature))
+		return 0;
+
+	err = feature_handler(netdev, enable);
+	if (err) {
+		netdev_err(netdev, "%s feature 0x%llx failed err %d\n",
+			   enable ? "Enable" : "Disable", feature, err);
+		return err;
+	}
+
+	MLX5E_SET_FEATURE(netdev, feature, enable);
+	return 0;
+}
+
+static int mlx5e_set_features(struct net_device *netdev,
+			      netdev_features_t features)
+{
+	int err;
+
+	err  = mlx5e_handle_feature(netdev, features, NETIF_F_LRO,
+				    set_feature_lro);
+	err |= mlx5e_handle_feature(netdev, features,
+				    NETIF_F_HW_VLAN_CTAG_FILTER,
+				    set_feature_vlan_filter);
+	err |= mlx5e_handle_feature(netdev, features, NETIF_F_HW_TC,
+				    set_feature_tc_num_filters);
+	err |= mlx5e_handle_feature(netdev, features, NETIF_F_RXALL,
+				    set_feature_rx_all);
+	err |= mlx5e_handle_feature(netdev, features, NETIF_F_HW_VLAN_CTAG_RX,
+				    set_feature_rx_vlan);
+#ifdef CONFIG_RFS_ACCEL
+	err |= mlx5e_handle_feature(netdev, features, NETIF_F_NTUPLE,
+				    set_feature_arfs);
+#endif
+
+	return err ? -EINVAL : 0;
+}
+
+#define MXL5_HW_MIN_MTU 64
+#define MXL5E_MIN_MTU (MXL5_HW_MIN_MTU + ETH_FCS_LEN)
 
 static int mlx5e_change_mtu(struct net_device *netdev, int new_mtu)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
 	struct mlx5_core_dev *mdev = priv->mdev;
 	bool was_opened;
-	int max_mtu;
+	u16 max_mtu;
+	u16 min_mtu;
 	int err = 0;
 
 	mlx5_query_port_max_mtu(mdev, &max_mtu, 1);
 
 	max_mtu = MLX5E_HW2SW_MTU(max_mtu);
+	min_mtu = MLX5E_HW2SW_MTU(MXL5E_MIN_MTU);
 
-	if (new_mtu > max_mtu) {
+	if (new_mtu > max_mtu || new_mtu < min_mtu) {
 		netdev_err(netdev,
-			   "%s: Bad MTU (%d) > (%d) Max\n",
-			   __func__, new_mtu, max_mtu);
+			   "%s: Bad MTU (%d), valid range is: [%d..%d]\n",
+			   __func__, new_mtu, min_mtu, max_mtu);
 		return -EINVAL;
 	}
 
@@ -2063,6 +2448,21 @@ static int mlx5e_set_vf_vlan(struct net_device *dev, int vf, u16 vlan, u8 qos)
 					   vlan, qos);
 }
 
+static int mlx5e_set_vf_spoofchk(struct net_device *dev, int vf, bool setting)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_spoofchk(mdev->priv.eswitch, vf + 1, setting);
+}
+
+static int mlx5e_set_vf_trust(struct net_device *dev, int vf, bool setting)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	return mlx5_eswitch_set_vport_trust(mdev->priv.eswitch, vf + 1, setting);
+}
 static int mlx5_vport_link2ifla(u8 esw_link)
 {
 	switch (esw_link) {
@@ -2127,7 +2527,7 @@ static void mlx5e_add_vxlan_port(struct net_device *netdev,
 	if (!mlx5e_vxlan_allowed(priv->mdev))
 		return;
 
-	mlx5e_vxlan_add_port(priv, be16_to_cpu(port));
+	mlx5e_vxlan_queue_work(priv, sa_family, be16_to_cpu(port), 1);
 }
 
 static void mlx5e_del_vxlan_port(struct net_device *netdev,
@@ -2138,7 +2538,7 @@ static void mlx5e_del_vxlan_port(struct net_device *netdev,
 	if (!mlx5e_vxlan_allowed(priv->mdev))
 		return;
 
-	mlx5e_vxlan_del_port(priv, be16_to_cpu(port));
+	mlx5e_vxlan_queue_work(priv, sa_family, be16_to_cpu(port), 0);
 }
 
 static netdev_features_t mlx5e_vxlan_features_check(struct mlx5e_priv *priv,
@@ -2205,6 +2605,9 @@ static const struct net_device_ops mlx5e_netdev_ops_basic = {
 	.ndo_set_features        = mlx5e_set_features,
 	.ndo_change_mtu          = mlx5e_change_mtu,
 	.ndo_do_ioctl            = mlx5e_ioctl,
+#ifdef CONFIG_RFS_ACCEL
+	.ndo_rx_flow_steer	 = mlx5e_rx_flow_steer,
+#endif
 };
 
 static const struct net_device_ops mlx5e_netdev_ops_sriov = {
@@ -2224,8 +2627,13 @@ static const struct net_device_ops mlx5e_netdev_ops_sriov = {
 	.ndo_add_vxlan_port      = mlx5e_add_vxlan_port,
 	.ndo_del_vxlan_port      = mlx5e_del_vxlan_port,
 	.ndo_features_check      = mlx5e_features_check,
+#ifdef CONFIG_RFS_ACCEL
+	.ndo_rx_flow_steer	 = mlx5e_rx_flow_steer,
+#endif
 	.ndo_set_vf_mac          = mlx5e_set_vf_mac,
 	.ndo_set_vf_vlan         = mlx5e_set_vf_vlan,
+	.ndo_set_vf_spoofchk     = mlx5e_set_vf_spoofchk,
+	.ndo_set_vf_trust        = mlx5e_set_vf_trust,
 	.ndo_get_vf_config       = mlx5e_get_vf_config,
 	.ndo_set_vf_link_state   = mlx5e_set_vf_link_state,
 	.ndo_get_vf_stats        = mlx5e_get_vf_stats,
@@ -2283,25 +2691,121 @@ static void mlx5e_ets_init(struct mlx5e_priv *priv)
 }
 #endif
 
-void mlx5e_build_default_indir_rqt(u32 *indirection_rqt, int len,
+void mlx5e_build_default_indir_rqt(struct mlx5_core_dev *mdev,
+				   u32 *indirection_rqt, int len,
 				   int num_channels)
 {
+	int node = mdev->priv.numa_node;
+	int node_num_of_cores;
 	int i;
 
+	if (node == -1)
+		node = first_online_node;
+
+	node_num_of_cores = cpumask_weight(cpumask_of_node(node));
+
+	if (node_num_of_cores)
+		num_channels = min_t(int, num_channels, node_num_of_cores);
+
 	for (i = 0; i < len; i++)
 		indirection_rqt[i] = i % num_channels;
 }
 
+static bool mlx5e_check_fragmented_striding_rq_cap(struct mlx5_core_dev *mdev)
+{
+	return MLX5_CAP_GEN(mdev, striding_rq) &&
+		MLX5_CAP_GEN(mdev, umr_ptr_rlky) &&
+		MLX5_CAP_ETH(mdev, reg_umr_sq);
+}
+
+static int mlx5e_get_pci_bw(struct mlx5_core_dev *mdev, u32 *pci_bw)
+{
+	enum pcie_link_width width;
+	enum pci_bus_speed speed;
+	int err = 0;
+
+	err = pcie_get_minimum_link(mdev->pdev, &speed, &width);
+	if (err)
+		return err;
+
+	if (speed == PCI_SPEED_UNKNOWN || width == PCIE_LNK_WIDTH_UNKNOWN)
+		return -EINVAL;
+
+	switch (speed) {
+	case PCIE_SPEED_2_5GT:
+		*pci_bw = 2500 * width;
+		break;
+	case PCIE_SPEED_5_0GT:
+		*pci_bw = 5000 * width;
+		break;
+	case PCIE_SPEED_8_0GT:
+		*pci_bw = 8000 * width;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static bool cqe_compress_heuristic(u32 link_speed, u32 pci_bw)
+{
+	return (link_speed && pci_bw &&
+		(pci_bw < 40000) && (pci_bw < link_speed));
+}
+
 static void mlx5e_build_netdev_priv(struct mlx5_core_dev *mdev,
 				    struct net_device *netdev,
 				    int num_channels)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
+	u32 link_speed = 0;
+	u32 pci_bw = 0;
 
 	priv->params.log_sq_size           =
 		MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE;
-	priv->params.log_rq_size           =
-		MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE;
+	priv->params.rq_wq_type = mlx5e_check_fragmented_striding_rq_cap(mdev) ?
+		MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ :
+		MLX5_WQ_TYPE_LINKED_LIST;
+
+	/* set CQE compression */
+	priv->params.rx_cqe_compress_admin = false;
+	if (MLX5_CAP_GEN(mdev, cqe_compression) &&
+	    MLX5_CAP_GEN(mdev, vport_group_manager)) {
+		mlx5e_get_max_linkspeed(mdev, &link_speed);
+		mlx5e_get_pci_bw(mdev, &pci_bw);
+		mlx5_core_dbg(mdev, "Max link speed = %d, PCI BW = %d\n",
+			      link_speed, pci_bw);
+		priv->params.rx_cqe_compress_admin =
+			cqe_compress_heuristic(link_speed, pci_bw);
+	}
+
+	priv->params.rx_cqe_compress = priv->params.rx_cqe_compress_admin;
+
+	switch (priv->params.rq_wq_type) {
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+		priv->params.log_rq_size = MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE_MPW;
+		priv->params.mpwqe_log_stride_sz =
+			priv->params.rx_cqe_compress ?
+			MLX5_MPWRQ_LOG_STRIDE_SIZE_CQE_COMPRESS :
+			MLX5_MPWRQ_LOG_STRIDE_SIZE;
+		priv->params.mpwqe_log_num_strides = MLX5_MPWRQ_LOG_WQE_SZ -
+			priv->params.mpwqe_log_stride_sz;
+		priv->params.lro_en = true;
+		break;
+	default: /* MLX5_WQ_TYPE_LINKED_LIST */
+		priv->params.log_rq_size = MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE;
+	}
+
+	mlx5_core_info(mdev,
+		       "MLX5E: StrdRq(%d) RqSz(%ld) StrdSz(%ld) RxCqeCmprss(%d)\n",
+		       priv->params.rq_wq_type == MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ,
+		       BIT(priv->params.log_rq_size),
+		       BIT(priv->params.mpwqe_log_stride_sz),
+		       priv->params.rx_cqe_compress_admin);
+
+	priv->params.min_rx_wqes = mlx5_min_rx_wqes(priv->params.rq_wq_type,
+					    BIT(priv->params.log_rq_size));
 	priv->params.rx_cq_moderation_usec =
 		MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC;
 	priv->params.rx_cq_moderation_pkts =
@@ -2311,15 +2815,13 @@ static void mlx5e_build_netdev_priv(struct mlx5_core_dev *mdev,
 	priv->params.tx_cq_moderation_pkts =
 		MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_PKTS;
 	priv->params.tx_max_inline         = mlx5e_get_max_inline_cap(mdev);
-	priv->params.min_rx_wqes           =
-		MLX5E_PARAMS_DEFAULT_MIN_RX_WQES;
 	priv->params.num_tc                = 1;
 	priv->params.rss_hfunc             = ETH_RSS_HASH_XOR;
 
 	netdev_rss_key_fill(priv->params.toeplitz_hash_key,
 			    sizeof(priv->params.toeplitz_hash_key));
 
-	mlx5e_build_default_indir_rqt(priv->params.indirection_rqt,
+	mlx5e_build_default_indir_rqt(mdev, priv->params.indirection_rqt,
 				      MLX5E_INDIR_RQT_SIZE, num_channels);
 
 	priv->params.lro_wqe_sz            =
@@ -2356,6 +2858,8 @@ static void mlx5e_build_netdev(struct net_device *netdev)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
 	struct mlx5_core_dev *mdev = priv->mdev;
+	bool fcs_supported;
+	bool fcs_enabled;
 
 	SET_NETDEV_DEV(netdev, &mdev->pdev->dev);
 
@@ -2390,25 +2894,41 @@ static void mlx5e_build_netdev(struct net_device *netdev)
 	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_FILTER;
 
 	if (mlx5e_vxlan_allowed(mdev)) {
-		netdev->hw_features     |= NETIF_F_GSO_UDP_TUNNEL;
+		netdev->hw_features     |= NETIF_F_GSO_UDP_TUNNEL |
+					   NETIF_F_GSO_UDP_TUNNEL_CSUM |
+					   NETIF_F_GSO_PARTIAL;
 		netdev->hw_enc_features |= NETIF_F_IP_CSUM;
-		netdev->hw_enc_features |= NETIF_F_RXCSUM;
+		netdev->hw_enc_features |= NETIF_F_IPV6_CSUM;
 		netdev->hw_enc_features |= NETIF_F_TSO;
 		netdev->hw_enc_features |= NETIF_F_TSO6;
-		netdev->hw_enc_features |= NETIF_F_RXHASH;
 		netdev->hw_enc_features |= NETIF_F_GSO_UDP_TUNNEL;
+		netdev->hw_enc_features |= NETIF_F_GSO_UDP_TUNNEL_CSUM |
+					   NETIF_F_GSO_PARTIAL;
+		netdev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM;
 	}
 
+	mlx5_query_port_fcs(mdev, &fcs_supported, &fcs_enabled);
+
+	if (fcs_supported)
+		netdev->hw_features |= NETIF_F_RXALL;
+
 	netdev->features          = netdev->hw_features;
 	if (!priv->params.lro_en)
 		netdev->features  &= ~NETIF_F_LRO;
 
+	if (fcs_enabled)
+		netdev->features  &= ~NETIF_F_RXALL;
+
 #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;
+	    FT_CAP(flow_table_modify)) {
+		netdev->hw_features      |= NETIF_F_HW_TC;
+#ifdef CONFIG_RFS_ACCEL
+		netdev->hw_features	 |= NETIF_F_NTUPLE;
+#endif
+	}
 
 	netdev->features         |= NETIF_F_HIGHDMA;
 
@@ -2442,6 +2962,61 @@ static int mlx5e_create_mkey(struct mlx5e_priv *priv, u32 pdn,
 	return err;
 }
 
+static void mlx5e_create_q_counter(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5_core_alloc_q_counter(mdev, &priv->q_counter);
+	if (err) {
+		mlx5_core_warn(mdev, "alloc queue counter failed, %d\n", err);
+		priv->q_counter = 0;
+	}
+}
+
+static void mlx5e_destroy_q_counter(struct mlx5e_priv *priv)
+{
+	if (!priv->q_counter)
+		return;
+
+	mlx5_core_dealloc_q_counter(priv->mdev, priv->q_counter);
+}
+
+static int mlx5e_create_umr_mkey(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_create_mkey_mbox_in *in;
+	struct mlx5_mkey_seg *mkc;
+	int inlen = sizeof(*in);
+	u64 npages =
+		mlx5e_get_max_num_channels(mdev) * MLX5_CHANNEL_MAX_NUM_MTTS;
+	int err;
+
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	mkc = &in->seg;
+	mkc->status = MLX5_MKEY_STATUS_FREE;
+	mkc->flags = MLX5_PERM_UMR_EN |
+		     MLX5_PERM_LOCAL_READ |
+		     MLX5_PERM_LOCAL_WRITE |
+		     MLX5_ACCESS_MODE_MTT;
+
+	mkc->qpn_mkey7_0 = cpu_to_be32(0xffffff << 8);
+	mkc->flags_pd = cpu_to_be32(priv->pdn);
+	mkc->len = cpu_to_be64(npages << PAGE_SHIFT);
+	mkc->xlt_oct_size = cpu_to_be32(mlx5e_get_mtt_octw(npages));
+	mkc->log2_page_size = PAGE_SHIFT;
+
+	err = mlx5_core_create_mkey(mdev, &priv->umr_mkey, in, inlen, NULL,
+				    NULL, NULL);
+
+	kvfree(in);
+
+	return err;
+}
+
 static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 {
 	struct net_device *netdev;
@@ -2467,10 +3042,14 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 
 	priv = netdev_priv(netdev);
 
+	priv->wq = create_singlethread_workqueue("mlx5e");
+	if (!priv->wq)
+		goto err_free_netdev;
+
 	err = mlx5_alloc_map_uar(mdev, &priv->cq_uar, false);
 	if (err) {
 		mlx5_core_err(mdev, "alloc_map uar failed, %d\n", err);
-		goto err_free_netdev;
+		goto err_destroy_wq;
 	}
 
 	err = mlx5_core_alloc_pd(mdev, &priv->pdn);
@@ -2491,10 +3070,16 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 		goto err_dealloc_transport_domain;
 	}
 
+	err = mlx5e_create_umr_mkey(priv);
+	if (err) {
+		mlx5_core_err(mdev, "create umr mkey failed, %d\n", err);
+		goto err_destroy_mkey;
+	}
+
 	err = mlx5e_create_tises(priv);
 	if (err) {
 		mlx5_core_warn(mdev, "create tises failed, %d\n", err);
-		goto err_destroy_mkey;
+		goto err_destroy_umr_mkey;
 	}
 
 	err = mlx5e_open_drop_rq(priv);
@@ -2503,37 +3088,33 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 		goto err_destroy_tises;
 	}
 
-	err = mlx5e_create_rqt(priv, MLX5E_INDIRECTION_RQT);
+	err = mlx5e_create_rqts(priv);
 	if (err) {
-		mlx5_core_warn(mdev, "create rqt(INDIR) failed, %d\n", err);
+		mlx5_core_warn(mdev, "create rqts failed, %d\n", err);
 		goto err_close_drop_rq;
 	}
 
-	err = mlx5e_create_rqt(priv, MLX5E_SINGLE_RQ_RQT);
-	if (err) {
-		mlx5_core_warn(mdev, "create rqt(SINGLE) failed, %d\n", err);
-		goto err_destroy_rqt_indir;
-	}
-
 	err = mlx5e_create_tirs(priv);
 	if (err) {
 		mlx5_core_warn(mdev, "create tirs failed, %d\n", err);
-		goto err_destroy_rqt_single;
+		goto err_destroy_rqts;
 	}
 
-	err = mlx5e_create_flow_tables(priv);
+	err = mlx5e_create_flow_steering(priv);
 	if (err) {
-		mlx5_core_warn(mdev, "create flow tables failed, %d\n", err);
+		mlx5_core_warn(mdev, "create flow steering failed, %d\n", err);
 		goto err_destroy_tirs;
 	}
 
-	mlx5e_init_eth_addr(priv);
+	mlx5e_create_q_counter(priv);
+
+	mlx5e_init_l2_addr(priv);
 
 	mlx5e_vxlan_init(priv);
 
 	err = mlx5e_tc_init(priv);
 	if (err)
-		goto err_destroy_flow_tables;
+		goto err_dealloc_q_counters;
 
 #ifdef CONFIG_MLX5_CORE_EN_DCB
 	mlx5e_dcbnl_ieee_setets_core(priv, &priv->params.ets);
@@ -2545,28 +3126,29 @@ static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
 		goto err_tc_cleanup;
 	}
 
-	if (mlx5e_vxlan_allowed(mdev))
+	if (mlx5e_vxlan_allowed(mdev)) {
+		rtnl_lock();
 		vxlan_get_rx_port(netdev);
+		rtnl_unlock();
+	}
 
 	mlx5e_enable_async_events(priv);
-	schedule_work(&priv->set_rx_mode_work);
+	queue_work(priv->wq, &priv->set_rx_mode_work);
 
 	return priv;
 
 err_tc_cleanup:
 	mlx5e_tc_cleanup(priv);
 
-err_destroy_flow_tables:
-	mlx5e_destroy_flow_tables(priv);
+err_dealloc_q_counters:
+	mlx5e_destroy_q_counter(priv);
+	mlx5e_destroy_flow_steering(priv);
 
 err_destroy_tirs:
 	mlx5e_destroy_tirs(priv);
 
-err_destroy_rqt_single:
-	mlx5e_destroy_rqt(priv, MLX5E_SINGLE_RQ_RQT);
-
-err_destroy_rqt_indir:
-	mlx5e_destroy_rqt(priv, MLX5E_INDIRECTION_RQT);
+err_destroy_rqts:
+	mlx5e_destroy_rqts(priv);
 
 err_close_drop_rq:
 	mlx5e_close_drop_rq(priv);
@@ -2574,6 +3156,9 @@ err_close_drop_rq:
 err_destroy_tises:
 	mlx5e_destroy_tises(priv);
 
+err_destroy_umr_mkey:
+	mlx5_core_destroy_mkey(mdev, &priv->umr_mkey);
+
 err_destroy_mkey:
 	mlx5_core_destroy_mkey(mdev, &priv->mkey);
 
@@ -2586,6 +3171,9 @@ err_dealloc_pd:
 err_unmap_free_uar:
 	mlx5_unmap_free_uar(mdev, &priv->cq_uar);
 
+err_destroy_wq:
+	destroy_workqueue(priv->wq);
+
 err_free_netdev:
 	free_netdev(netdev);
 
@@ -2599,23 +3187,37 @@ static void mlx5e_destroy_netdev(struct mlx5_core_dev *mdev, void *vpriv)
 
 	set_bit(MLX5E_STATE_DESTROYING, &priv->state);
 
-	schedule_work(&priv->set_rx_mode_work);
+	queue_work(priv->wq, &priv->set_rx_mode_work);
 	mlx5e_disable_async_events(priv);
-	flush_scheduled_work();
-	unregister_netdev(netdev);
+	flush_workqueue(priv->wq);
+	if (test_bit(MLX5_INTERFACE_STATE_SHUTDOWN, &mdev->intf_state)) {
+		netif_device_detach(netdev);
+		mutex_lock(&priv->state_lock);
+		if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+			mlx5e_close_locked(netdev);
+		mutex_unlock(&priv->state_lock);
+	} else {
+		unregister_netdev(netdev);
+	}
+
 	mlx5e_tc_cleanup(priv);
 	mlx5e_vxlan_cleanup(priv);
-	mlx5e_destroy_flow_tables(priv);
+	mlx5e_destroy_q_counter(priv);
+	mlx5e_destroy_flow_steering(priv);
 	mlx5e_destroy_tirs(priv);
-	mlx5e_destroy_rqt(priv, MLX5E_SINGLE_RQ_RQT);
-	mlx5e_destroy_rqt(priv, MLX5E_INDIRECTION_RQT);
+	mlx5e_destroy_rqts(priv);
 	mlx5e_close_drop_rq(priv);
 	mlx5e_destroy_tises(priv);
+	mlx5_core_destroy_mkey(priv->mdev, &priv->umr_mkey);
 	mlx5_core_destroy_mkey(priv->mdev, &priv->mkey);
 	mlx5_core_dealloc_transport_domain(priv->mdev, priv->tdn);
 	mlx5_core_dealloc_pd(priv->mdev, priv->pdn);
 	mlx5_unmap_free_uar(priv->mdev, &priv->cq_uar);
-	free_netdev(netdev);
+	cancel_delayed_work_sync(&priv->update_stats_work);
+	destroy_workqueue(priv->wq);
+
+	if (!test_bit(MLX5_INTERFACE_STATE_SHUTDOWN, &mdev->intf_state))
+		free_netdev(netdev);
 }
 
 static void *mlx5e_get_netdev(void *vpriv)