These changes are the raw update to linux-4.4.6-rt14. Kernel sources

are taken from kernel.org, and rt patch from the rt wiki download page.

During the rebasing, the following patch collided:

Force tick interrupt and get rid of softirq magic(I70131fb85).

Collisions have been removed because its logic was found on the
source already.

Change-Id: I7f57a4081d9deaa0d9ccfc41a6c8daccdee3b769
Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>

75 files changed, 23590 insertions, 1108 deletions

diff --git a/kernel/drivers/net/ethernet/mellanox/Kconfig b/kernel/drivers/net/ethernet/mellanox/Kconfig
index 8cf7563a8..d54701047 100644
--- a/kernel/drivers/net/ethernet/mellanox/Kconfig
+++ b/kernel/drivers/net/ethernet/mellanox/Kconfig
@@ -7,9 +7,7 @@ config NET_VENDOR_MELLANOX
 	default y
 	depends on PCI
 	---help---
-	  If you have a network (Ethernet) card belonging to this class, say Y
-	  and read the Ethernet-HOWTO, available from
-	  <http://www.tldp.org/docs.html#howto>.
+	  If you have a network (Ethernet) card belonging to this class, say Y.
 
 	  Note that the answer to this question doesn't directly affect the
 	  kernel: saying N will just cause the configurator to skip all
@@ -20,5 +18,6 @@ if NET_VENDOR_MELLANOX
 
 source "drivers/net/ethernet/mellanox/mlx4/Kconfig"
 source "drivers/net/ethernet/mellanox/mlx5/core/Kconfig"
+source "drivers/net/ethernet/mellanox/mlxsw/Kconfig"
 
 endif # NET_VENDOR_MELLANOX
diff --git a/kernel/drivers/net/ethernet/mellanox/Makefile b/kernel/drivers/net/ethernet/mellanox/Makefile
index 38fe32ef5..2e2a5ec50 100644
--- a/kernel/drivers/net/ethernet/mellanox/Makefile
+++ b/kernel/drivers/net/ethernet/mellanox/Makefile
@@ -4,3 +4,4 @@
 
 obj-$(CONFIG_MLX4_CORE) += mlx4/
 obj-$(CONFIG_MLX5_CORE) += mlx5/core/
+obj-$(CONFIG_MLXSW_CORE) += mlxsw/
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/cmd.c b/kernel/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 529ef0594..d48d57934 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -49,6 +49,7 @@
 #include "mlx4.h"
 #include "fw.h"
 #include "fw_qos.h"
+#include "mlx4_stats.h"
 
 #define CMD_POLL_TOKEN 0xffff
 #define INBOX_MASK	0xffffffffffffff00ULL
@@ -685,6 +686,7 @@ static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
 {
 	struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
 	struct mlx4_cmd_context *context;
+	long ret_wait;
 	int err = 0;
 
 	down(&cmd->event_sem);
@@ -710,8 +712,20 @@ static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
 	if (err)
 		goto out_reset;
 
-	if (!wait_for_completion_timeout(&context->done,
-					 msecs_to_jiffies(timeout))) {
+	if (op == MLX4_CMD_SENSE_PORT) {
+		ret_wait =
+			wait_for_completion_interruptible_timeout(&context->done,
+								  msecs_to_jiffies(timeout));
+		if (ret_wait < 0) {
+			context->fw_status = 0;
+			context->out_param = 0;
+			context->result = 0;
+		}
+	} else {
+		ret_wait = (long)wait_for_completion_timeout(&context->done,
+							     msecs_to_jiffies(timeout));
+	}
+	if (!ret_wait) {
 		mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
 			  op);
 		if (op == MLX4_CMD_NOP) {
@@ -882,7 +896,7 @@ static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
 {
 	struct ib_smp *smp = inbox->buf;
 	u32 index;
-	u8 port;
+	u8 port, slave_port;
 	u8 opcode_modifier;
 	u16 *table;
 	int err;
@@ -894,7 +908,8 @@ static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
 	__be32 slave_cap_mask;
 	__be64 slave_node_guid;
 
-	port = vhcr->in_modifier;
+	slave_port = vhcr->in_modifier;
+	port = mlx4_slave_convert_port(dev, slave, slave_port);
 
 	/* network-view bit is for driver use only, and should not be passed to FW */
 	opcode_modifier = vhcr->op_modifier & ~0x8; /* clear netw view bit */
@@ -930,8 +945,9 @@ static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
 			if (smp->attr_id == IB_SMP_ATTR_PORT_INFO) {
 				/*get the slave specific caps:*/
 				/*do the command */
+				smp->attr_mod = cpu_to_be32(port);
 				err = mlx4_cmd_box(dev, inbox->dma, outbox->dma,
-					    vhcr->in_modifier, opcode_modifier,
+					    port, opcode_modifier,
 					    vhcr->op, MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
 				/* modify the response for slaves */
 				if (!err && slave != mlx4_master_func_num(dev)) {
@@ -975,7 +991,7 @@ static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
 			}
 			if (smp->attr_id == IB_SMP_ATTR_NODE_INFO) {
 				err = mlx4_cmd_box(dev, inbox->dma, outbox->dma,
-					     vhcr->in_modifier, opcode_modifier,
+					     port, opcode_modifier,
 					     vhcr->op, MLX4_CMD_TIME_CLASS_C, MLX4_CMD_NATIVE);
 				if (!err) {
 					slave_node_guid =  mlx4_get_slave_node_guid(dev, slave);
@@ -994,7 +1010,7 @@ static int mlx4_MAD_IFC_wrapper(struct mlx4_dev *dev, int slave,
 		if (!(smp->mgmt_class == IB_MGMT_CLASS_SUBN_LID_ROUTED &&
 		      smp->method == IB_MGMT_METHOD_GET) || network_view) {
 			mlx4_err(dev, "Unprivileged slave %d is trying to execute a Subnet MGMT MAD, class 0x%x, method 0x%x, view=%s for attr 0x%x. Rejecting\n",
-				 slave, smp->method, smp->mgmt_class,
+				 slave, smp->mgmt_class, smp->method,
 				 network_view ? "Network" : "Host",
 				 be16_to_cpu(smp->attr_id));
 			return -EPERM;
@@ -2382,7 +2398,7 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
 			}
 		}
 
-		memset(&priv->mfunc.master.cmd_eqe, 0, dev->caps.eqe_size);
+		memset(&priv->mfunc.master.cmd_eqe, 0, sizeof(struct mlx4_eqe));
 		priv->mfunc.master.cmd_eqe.type = MLX4_EVENT_TYPE_CMD;
 		INIT_WORK(&priv->mfunc.master.comm_work,
 			  mlx4_master_comm_channel);
@@ -2915,7 +2931,7 @@ int mlx4_set_vf_mac(struct mlx4_dev *dev, int port, int vf, u64 mac)
 	port = mlx4_slaves_closest_port(dev, slave, port);
 	s_info = &priv->mfunc.master.vf_admin[slave].vport[port];
 	s_info->mac = mac;
-	mlx4_info(dev, "default mac on vf %d port %d to %llX will take afect only after vf restart\n",
+	mlx4_info(dev, "default mac on vf %d port %d to %llX will take effect only after vf restart\n",
 		  vf, port, s_info->mac);
 	return 0;
 }
@@ -3164,6 +3180,92 @@ int mlx4_set_vf_link_state(struct mlx4_dev *dev, int port, int vf, int link_stat
 }
 EXPORT_SYMBOL_GPL(mlx4_set_vf_link_state);
 
+int mlx4_get_counter_stats(struct mlx4_dev *dev, int counter_index,
+			   struct mlx4_counter *counter_stats, int reset)
+{
+	struct mlx4_cmd_mailbox *mailbox = NULL;
+	struct mlx4_counter *tmp_counter;
+	int err;
+	u32 if_stat_in_mod;
+
+	if (!counter_stats)
+		return -EINVAL;
+
+	if (counter_index == MLX4_SINK_COUNTER_INDEX(dev))
+		return 0;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+
+	memset(mailbox->buf, 0, sizeof(struct mlx4_counter));
+	if_stat_in_mod = counter_index;
+	if (reset)
+		if_stat_in_mod |= MLX4_QUERY_IF_STAT_RESET;
+	err = mlx4_cmd_box(dev, 0, mailbox->dma,
+			   if_stat_in_mod, 0,
+			   MLX4_CMD_QUERY_IF_STAT,
+			   MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+	if (err) {
+		mlx4_dbg(dev, "%s: failed to read statistics for counter index %d\n",
+			 __func__, counter_index);
+		goto if_stat_out;
+	}
+	tmp_counter = (struct mlx4_counter *)mailbox->buf;
+	counter_stats->counter_mode = tmp_counter->counter_mode;
+	if (counter_stats->counter_mode == 0) {
+		counter_stats->rx_frames =
+			cpu_to_be64(be64_to_cpu(counter_stats->rx_frames) +
+				    be64_to_cpu(tmp_counter->rx_frames));
+		counter_stats->tx_frames =
+			cpu_to_be64(be64_to_cpu(counter_stats->tx_frames) +
+				    be64_to_cpu(tmp_counter->tx_frames));
+		counter_stats->rx_bytes =
+			cpu_to_be64(be64_to_cpu(counter_stats->rx_bytes) +
+				    be64_to_cpu(tmp_counter->rx_bytes));
+		counter_stats->tx_bytes =
+			cpu_to_be64(be64_to_cpu(counter_stats->tx_bytes) +
+				    be64_to_cpu(tmp_counter->tx_bytes));
+	}
+
+if_stat_out:
+	mlx4_free_cmd_mailbox(dev, mailbox);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_counter_stats);
+
+int mlx4_get_vf_stats(struct mlx4_dev *dev, int port, int vf_idx,
+		      struct ifla_vf_stats *vf_stats)
+{
+	struct mlx4_counter tmp_vf_stats;
+	int slave;
+	int err = 0;
+
+	if (!vf_stats)
+		return -EINVAL;
+
+	if (!mlx4_is_master(dev))
+		return -EPROTONOSUPPORT;
+
+	slave = mlx4_get_slave_indx(dev, vf_idx);
+	if (slave < 0)
+		return -EINVAL;
+
+	port = mlx4_slaves_closest_port(dev, slave, port);
+	err = mlx4_calc_vf_counters(dev, slave, port, &tmp_vf_stats);
+	if (!err && tmp_vf_stats.counter_mode == 0) {
+		vf_stats->rx_packets = be64_to_cpu(tmp_vf_stats.rx_frames);
+		vf_stats->tx_packets = be64_to_cpu(tmp_vf_stats.tx_frames);
+		vf_stats->rx_bytes = be64_to_cpu(tmp_vf_stats.rx_bytes);
+		vf_stats->tx_bytes = be64_to_cpu(tmp_vf_stats.tx_bytes);
+	}
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_vf_stats);
+
 int mlx4_vf_smi_enabled(struct mlx4_dev *dev, int slave, int port)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
@@ -3197,6 +3299,12 @@ int mlx4_vf_set_enable_smi_admin(struct mlx4_dev *dev, int slave, int port,
 				 int enabled)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_active_ports actv_ports = mlx4_get_active_ports(
+			&priv->dev, slave);
+	int min_port = find_first_bit(actv_ports.ports,
+				      priv->dev.caps.num_ports) + 1;
+	int max_port = min_port - 1 +
+		bitmap_weight(actv_ports.ports, priv->dev.caps.num_ports);
 
 	if (slave == mlx4_master_func_num(dev))
 		return 0;
@@ -3206,6 +3314,11 @@ int mlx4_vf_set_enable_smi_admin(struct mlx4_dev *dev, int slave, int port,
 	    enabled < 0 || enabled > 1)
 		return -EINVAL;
 
+	if (min_port == max_port && dev->caps.num_ports > 1) {
+		mlx4_info(dev, "SMI access disallowed for single ported VFs\n");
+		return -EPROTONOSUPPORT;
+	}
+
 	priv->mfunc.master.vf_admin[slave].enable_smi[port] = enabled;
 	return 0;
 }
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/cq.c b/kernel/drivers/net/ethernet/mellanox/mlx4/cq.c
index e71f31387..3348e646d 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/cq.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/cq.c
@@ -292,7 +292,7 @@ int mlx4_cq_alloc(struct mlx4_dev *dev, int nent,
 	u64 mtt_addr;
 	int err;
 
-	if (vector > dev->caps.num_comp_vectors + dev->caps.comp_pool)
+	if (vector >= dev->caps.num_comp_vectors)
 		return -EINVAL;
 
 	cq->vector = vector;
@@ -319,7 +319,7 @@ int mlx4_cq_alloc(struct mlx4_dev *dev, int nent,
 		cq_context->flags  |= cpu_to_be32(1 << 19);
 
 	cq_context->logsize_usrpage = cpu_to_be32((ilog2(nent) << 24) | uar->index);
-	cq_context->comp_eqn	    = priv->eq_table.eq[vector].eqn;
+	cq_context->comp_eqn	    = priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].eqn;
 	cq_context->log_page_size   = mtt->page_shift - MLX4_ICM_PAGE_SHIFT;
 
 	mtt_addr = mlx4_mtt_addr(dev, mtt);
@@ -339,11 +339,11 @@ int mlx4_cq_alloc(struct mlx4_dev *dev, int nent,
 	init_completion(&cq->free);
 	cq->comp = mlx4_add_cq_to_tasklet;
 	cq->tasklet_ctx.priv =
-		&priv->eq_table.eq[cq->vector].tasklet_ctx;
+		&priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].tasklet_ctx;
 	INIT_LIST_HEAD(&cq->tasklet_ctx.list);
 
 
-	cq->irq = priv->eq_table.eq[cq->vector].irq;
+	cq->irq = priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].irq;
 	return 0;
 
 err_radix:
@@ -368,7 +368,10 @@ void mlx4_cq_free(struct mlx4_dev *dev, struct mlx4_cq *cq)
 	if (err)
 		mlx4_warn(dev, "HW2SW_CQ failed (%d) for CQN %06x\n", err, cq->cqn);
 
-	synchronize_irq(priv->eq_table.eq[cq->vector].irq);
+	synchronize_irq(priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(cq->vector)].irq);
+	if (priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(cq->vector)].irq !=
+	    priv->eq_table.eq[MLX4_EQ_ASYNC].irq)
+		synchronize_irq(priv->eq_table.eq[MLX4_EQ_ASYNC].irq);
 
 	spin_lock_irq(&cq_table->lock);
 	radix_tree_delete(&cq_table->tree, cq->cqn);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_clock.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_clock.c
index 8a083d73e..1494997c4 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_clock.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_clock.c
@@ -236,23 +236,43 @@ static const struct ptp_clock_info mlx4_en_ptp_clock_info = {
 	.enable		= mlx4_en_phc_enable,
 };
 
+#define MLX4_EN_WRAP_AROUND_SEC	10ULL
+
+/* This function calculates the max shift that enables the user range
+ * of MLX4_EN_WRAP_AROUND_SEC values in the cycles register.
+ */
+static u32 freq_to_shift(u16 freq)
+{
+	u32 freq_khz = freq * 1000;
+	u64 max_val_cycles = freq_khz * 1000 * MLX4_EN_WRAP_AROUND_SEC;
+	u64 max_val_cycles_rounded = is_power_of_2(max_val_cycles + 1) ?
+		max_val_cycles : roundup_pow_of_two(max_val_cycles) - 1;
+	/* calculate max possible multiplier in order to fit in 64bit */
+	u64 max_mul = div_u64(0xffffffffffffffffULL, max_val_cycles_rounded);
+
+	/* This comes from the reverse of clocksource_khz2mult */
+	return ilog2(div_u64(max_mul * freq_khz, 1000000));
+}
+
 void mlx4_en_init_timestamp(struct mlx4_en_dev *mdev)
 {
 	struct mlx4_dev *dev = mdev->dev;
 	unsigned long flags;
 	u64 ns, zero = 0;
 
+	/* mlx4_en_init_timestamp is called for each netdev.
+	 * mdev->ptp_clock is common for all ports, skip initialization if
+	 * was done for other port.
+	 */
+	if (mdev->ptp_clock)
+		return;
+
 	rwlock_init(&mdev->clock_lock);
 
 	memset(&mdev->cycles, 0, sizeof(mdev->cycles));
 	mdev->cycles.read = mlx4_en_read_clock;
 	mdev->cycles.mask = CLOCKSOURCE_MASK(48);
-	/* Using shift to make calculation more accurate. Since current HW
-	 * clock frequency is 427 MHz, and cycles are given using a 48 bits
-	 * register, the biggest shift when calculating using u64, is 14
-	 * (max_cycles * multiplier < 2^64)
-	 */
-	mdev->cycles.shift = 14;
+	mdev->cycles.shift = freq_to_shift(dev->caps.hca_core_clock);
 	mdev->cycles.mult =
 		clocksource_khz2mult(1000 * dev->caps.hca_core_clock, mdev->cycles.shift);
 	mdev->nominal_c_mult = mdev->cycles.mult;
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_cq.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_cq.c
index 22da4d0d0..eb8a4988d 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_cq.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_cq.c
@@ -66,6 +66,7 @@ int mlx4_en_create_cq(struct mlx4_en_priv *priv,
 
 	cq->ring = ring;
 	cq->is_tx = mode;
+	cq->vector = mdev->dev->caps.num_comp_vectors;
 
 	/* Allocate HW buffers on provided NUMA node.
 	 * dev->numa_node is used in mtt range allocation flow.
@@ -99,14 +100,8 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
 {
 	struct mlx4_en_dev *mdev = priv->mdev;
 	int err = 0;
-	char name[25];
 	int timestamp_en = 0;
-	struct cpu_rmap *rmap =
-#ifdef CONFIG_RFS_ACCEL
-		priv->dev->rx_cpu_rmap;
-#else
-		NULL;
-#endif
+	bool assigned_eq = false;
 
 	cq->dev = mdev->pndev[priv->port];
 	cq->mcq.set_ci_db  = cq->wqres.db.db;
@@ -116,23 +111,19 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
 	memset(cq->buf, 0, cq->buf_size);
 
 	if (cq->is_tx == RX) {
-		if (mdev->dev->caps.comp_pool) {
-			if (!cq->vector) {
-				sprintf(name, "%s-%d", priv->dev->name,
-					cq->ring);
-				/* Set IRQ for specific name (per ring) */
-				if (mlx4_assign_eq(mdev->dev, name, rmap,
-						   &cq->vector)) {
-					cq->vector = (cq->ring + 1 + priv->port)
-					    % mdev->dev->caps.num_comp_vectors;
-					mlx4_warn(mdev, "Failed assigning an EQ to %s, falling back to legacy EQ's\n",
-						  name);
-				}
-
+		if (!mlx4_is_eq_vector_valid(mdev->dev, priv->port,
+					     cq->vector)) {
+			cq->vector = cpumask_first(priv->rx_ring[cq->ring]->affinity_mask);
+
+			err = mlx4_assign_eq(mdev->dev, priv->port,
+					     &cq->vector);
+			if (err) {
+				mlx4_err(mdev, "Failed assigning an EQ to CQ vector %d\n",
+					 cq->vector);
+				goto free_eq;
 			}
-		} else {
-			cq->vector = (cq->ring + 1 + priv->port) %
-				mdev->dev->caps.num_comp_vectors;
+
+			assigned_eq = true;
 		}
 
 		cq->irq_desc =
@@ -159,7 +150,7 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
 			    &mdev->priv_uar, cq->wqres.db.dma, &cq->mcq,
 			    cq->vector, 0, timestamp_en);
 	if (err)
-		return err;
+		goto free_eq;
 
 	cq->mcq.comp  = cq->is_tx ? mlx4_en_tx_irq : mlx4_en_rx_irq;
 	cq->mcq.event = mlx4_en_cq_event;
@@ -168,13 +159,6 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
 		netif_napi_add(cq->dev, &cq->napi, mlx4_en_poll_tx_cq,
 			       NAPI_POLL_WEIGHT);
 	} else {
-		struct mlx4_en_rx_ring *ring = priv->rx_ring[cq->ring];
-
-		err = irq_set_affinity_hint(cq->mcq.irq,
-					    ring->affinity_mask);
-		if (err)
-			mlx4_warn(mdev, "Failed setting affinity hint\n");
-
 		netif_napi_add(cq->dev, &cq->napi, mlx4_en_poll_rx_cq, 64);
 		napi_hash_add(&cq->napi);
 	}
@@ -182,6 +166,12 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq,
 	napi_enable(&cq->napi);
 
 	return 0;
+
+free_eq:
+	if (assigned_eq)
+		mlx4_release_eq(mdev->dev, cq->vector);
+	cq->vector = mdev->dev->caps.num_comp_vectors;
+	return err;
 }
 
 void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq **pcq)
@@ -191,9 +181,9 @@ void mlx4_en_destroy_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq **pcq)
 
 	mlx4_en_unmap_buffer(&cq->wqres.buf);
 	mlx4_free_hwq_res(mdev->dev, &cq->wqres, cq->buf_size);
-	if (priv->mdev->dev->caps.comp_pool && cq->vector) {
+	if (mlx4_is_eq_vector_valid(mdev->dev, priv->port, cq->vector) &&
+	    cq->is_tx == RX)
 		mlx4_release_eq(priv->mdev->dev, cq->vector);
-	}
 	cq->vector = 0;
 	cq->buf_size = 0;
 	cq->buf = NULL;
@@ -207,7 +197,6 @@ void mlx4_en_deactivate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq)
 	if (!cq->is_tx) {
 		napi_hash_del(&cq->napi);
 		synchronize_rcu();
-		irq_set_affinity_hint(cq->mcq.irq, NULL);
 	}
 	netif_napi_del(&cq->napi);
 
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
index a2ddf3d75..ddb554188 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_ethtool.c
@@ -95,13 +95,11 @@ mlx4_en_get_drvinfo(struct net_device *dev, struct ethtool_drvinfo *drvinfo)
 		(u16) (mdev->dev->caps.fw_ver & 0xffff));
 	strlcpy(drvinfo->bus_info, pci_name(mdev->dev->persist->pdev),
 		sizeof(drvinfo->bus_info));
-	drvinfo->n_stats = 0;
-	drvinfo->regdump_len = 0;
-	drvinfo->eedump_len = 0;
 }
 
 static const char mlx4_en_priv_flags[][ETH_GSTRING_LEN] = {
 	"blueflame",
+	"phv-bit"
 };
 
 static const char main_strings[][ETH_GSTRING_LEN] = {
@@ -119,6 +117,12 @@ static const char main_strings[][ETH_GSTRING_LEN] = {
 	"queue_stopped", "wake_queue", "tx_timeout", "rx_alloc_failed",
 	"rx_csum_good", "rx_csum_none", "rx_csum_complete", "tx_chksum_offload",
 
+	/* pf statistics */
+	"pf_rx_packets",
+	"pf_rx_bytes",
+	"pf_tx_packets",
+	"pf_tx_bytes",
+
 	/* priority flow control statistics rx */
 	"rx_pause_prio_0", "rx_pause_duration_prio_0",
 	"rx_pause_transition_prio_0",
@@ -368,6 +372,11 @@ static void mlx4_en_get_ethtool_stats(struct net_device *dev,
 		if (bitmap_iterator_test(&it))
 			data[index++] = ((unsigned long *)&priv->port_stats)[i];
 
+	for (i = 0; i < NUM_PF_STATS; i++, bitmap_iterator_inc(&it))
+		if (bitmap_iterator_test(&it))
+			data[index++] =
+				((unsigned long *)&priv->pf_stats)[i];
+
 	for (i = 0; i < NUM_FLOW_PRIORITY_STATS_RX;
 	     i++, bitmap_iterator_inc(&it))
 		if (bitmap_iterator_test(&it))
@@ -448,6 +457,12 @@ static void mlx4_en_get_strings(struct net_device *dev,
 				strcpy(data + (index++) * ETH_GSTRING_LEN,
 				       main_strings[strings]);
 
+		for (i = 0; i < NUM_PF_STATS; i++, strings++,
+		     bitmap_iterator_inc(&it))
+			if (bitmap_iterator_test(&it))
+				strcpy(data + (index++) * ETH_GSTRING_LEN,
+				       main_strings[strings]);
+
 		for (i = 0; i < NUM_FLOW_STATS; i++, strings++,
 		     bitmap_iterator_inc(&it))
 			if (bitmap_iterator_test(&it))
@@ -1780,35 +1795,49 @@ static int mlx4_en_get_ts_info(struct net_device *dev,
 static int mlx4_en_set_priv_flags(struct net_device *dev, u32 flags)
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = priv->mdev;
 	bool bf_enabled_new = !!(flags & MLX4_EN_PRIV_FLAGS_BLUEFLAME);
 	bool bf_enabled_old = !!(priv->pflags & MLX4_EN_PRIV_FLAGS_BLUEFLAME);
+	bool phv_enabled_new = !!(flags & MLX4_EN_PRIV_FLAGS_PHV);
+	bool phv_enabled_old = !!(priv->pflags & MLX4_EN_PRIV_FLAGS_PHV);
 	int i;
+	int ret = 0;
 
-	if (bf_enabled_new == bf_enabled_old)
-		return 0; /* Nothing to do */
+	if (bf_enabled_new != bf_enabled_old) {
+		if (bf_enabled_new) {
+			bool bf_supported = true;
 
-	if (bf_enabled_new) {
-		bool bf_supported = true;
+			for (i = 0; i < priv->tx_ring_num; i++)
+				bf_supported &= priv->tx_ring[i]->bf_alloced;
 
-		for (i = 0; i < priv->tx_ring_num; i++)
-			bf_supported &= priv->tx_ring[i]->bf_alloced;
+			if (!bf_supported) {
+				en_err(priv, "BlueFlame is not supported\n");
+				return -EINVAL;
+			}
 
-		if (!bf_supported) {
-			en_err(priv, "BlueFlame is not supported\n");
-			return -EINVAL;
+			priv->pflags |= MLX4_EN_PRIV_FLAGS_BLUEFLAME;
+		} else {
+			priv->pflags &= ~MLX4_EN_PRIV_FLAGS_BLUEFLAME;
 		}
 
-		priv->pflags |= MLX4_EN_PRIV_FLAGS_BLUEFLAME;
-	} else {
-		priv->pflags &= ~MLX4_EN_PRIV_FLAGS_BLUEFLAME;
-	}
-
-	for (i = 0; i < priv->tx_ring_num; i++)
-		priv->tx_ring[i]->bf_enabled = bf_enabled_new;
+		for (i = 0; i < priv->tx_ring_num; i++)
+			priv->tx_ring[i]->bf_enabled = bf_enabled_new;
 
-	en_info(priv, "BlueFlame %s\n",
-		bf_enabled_new ?  "Enabled" : "Disabled");
+		en_info(priv, "BlueFlame %s\n",
+			bf_enabled_new ?  "Enabled" : "Disabled");
+	}
 
+	if (phv_enabled_new != phv_enabled_old) {
+		ret = set_phv_bit(mdev->dev, priv->port, (int)phv_enabled_new);
+		if (ret)
+			return ret;
+		else if (phv_enabled_new)
+			priv->pflags |= MLX4_EN_PRIV_FLAGS_PHV;
+		else
+			priv->pflags &= ~MLX4_EN_PRIV_FLAGS_PHV;
+		en_info(priv, "PHV bit %s\n",
+			phv_enabled_new ?  "Enabled" : "Disabled");
+	}
 	return 0;
 }
 
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_main.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_main.c
index 913b716ed..e0ec280a7 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_main.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_main.c
@@ -123,6 +123,28 @@ void mlx4_en_update_loopback_state(struct net_device *dev,
 	 */
 	if (mlx4_is_mfunc(priv->mdev->dev) || priv->validate_loopback)
 		priv->flags |= MLX4_EN_FLAG_ENABLE_HW_LOOPBACK;
+
+	mutex_lock(&priv->mdev->state_lock);
+	if (priv->mdev->dev->caps.flags2 &
+	    MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB &&
+	    priv->rss_map.indir_qp.qpn) {
+		int i;
+		int err = 0;
+		int loopback = !!(features & NETIF_F_LOOPBACK);
+
+		for (i = 0; i < priv->rx_ring_num; i++) {
+			int ret;
+
+			ret = mlx4_en_change_mcast_lb(priv,
+						      &priv->rss_map.qps[i],
+						      loopback);
+			if (!err)
+				err = ret;
+		}
+		if (err)
+			mlx4_warn(priv->mdev, "failed to change mcast loopback\n");
+	}
+	mutex_unlock(&priv->mdev->state_lock);
 }
 
 static int mlx4_en_get_profile(struct mlx4_en_dev *mdev)
@@ -210,9 +232,6 @@ static void mlx4_en_remove(struct mlx4_dev *dev, void *endev_ptr)
 		if (mdev->pndev[i])
 			mlx4_en_destroy_netdev(mdev->pndev[i]);
 
-	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
-		mlx4_en_remove_timestamp(mdev);
-
 	flush_workqueue(mdev->workqueue);
 	destroy_workqueue(mdev->workqueue);
 	(void) mlx4_mr_free(dev, &mdev->mr);
@@ -224,6 +243,26 @@ static void mlx4_en_remove(struct mlx4_dev *dev, void *endev_ptr)
 	kfree(mdev);
 }
 
+static void mlx4_en_activate(struct mlx4_dev *dev, void *ctx)
+{
+	int i;
+	struct mlx4_en_dev *mdev = ctx;
+
+	/* Create a netdev for each port */
+	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
+		mlx4_info(mdev, "Activating port:%d\n", i);
+		if (mlx4_en_init_netdev(mdev, i, &mdev->profile.prof[i]))
+			mdev->pndev[i] = NULL;
+	}
+
+	/* register notifier */
+	mdev->nb.notifier_call = mlx4_en_netdev_event;
+	if (register_netdevice_notifier(&mdev->nb)) {
+		mdev->nb.notifier_call = NULL;
+		mlx4_err(mdev, "Failed to create notifier\n");
+	}
+}
+
 static void *mlx4_en_add(struct mlx4_dev *dev)
 {
 	struct mlx4_en_dev *mdev;
@@ -278,10 +317,6 @@ static void *mlx4_en_add(struct mlx4_dev *dev)
 	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH)
 		mdev->port_cnt++;
 
-	/* Initialize time stamp mechanism */
-	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
-		mlx4_en_init_timestamp(mdev);
-
 	/* Set default number of RX rings*/
 	mlx4_en_set_num_rx_rings(mdev);
 
@@ -297,21 +332,6 @@ static void *mlx4_en_add(struct mlx4_dev *dev)
 	mutex_init(&mdev->state_lock);
 	mdev->device_up = true;
 
-	/* Setup ports */
-
-	/* Create a netdev for each port */
-	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
-		mlx4_info(mdev, "Activating port:%d\n", i);
-		if (mlx4_en_init_netdev(mdev, i, &mdev->profile.prof[i]))
-			mdev->pndev[i] = NULL;
-	}
-	/* register notifier */
-	mdev->nb.notifier_call = mlx4_en_netdev_event;
-	if (register_netdevice_notifier(&mdev->nb)) {
-		mdev->nb.notifier_call = NULL;
-		mlx4_err(mdev, "Failed to create notifier\n");
-	}
-
 	return mdev;
 
 err_mr:
@@ -335,6 +355,7 @@ static struct mlx4_interface mlx4_en_interface = {
 	.event		= mlx4_en_event,
 	.get_dev	= mlx4_en_get_netdev,
 	.protocol	= MLX4_PROT_ETH,
+	.activate	= mlx4_en_activate,
 };
 
 static void mlx4_en_verify_params(void)
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index a5a0b8420..67e9633ea 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -573,10 +573,8 @@ static int mlx4_en_get_qp(struct mlx4_en_priv *priv)
 {
 	struct mlx4_en_dev *mdev = priv->mdev;
 	struct mlx4_dev *dev = mdev->dev;
-	struct mlx4_mac_entry *entry;
 	int index = 0;
 	int err = 0;
-	u64 reg_id = 0;
 	int *qpn = &priv->base_qpn;
 	u64 mac = mlx4_mac_to_u64(priv->dev->dev_addr);
 
@@ -600,44 +598,11 @@ static int mlx4_en_get_qp(struct mlx4_en_priv *priv)
 	en_dbg(DRV, priv, "Reserved qp %d\n", *qpn);
 	if (err) {
 		en_err(priv, "Failed to reserve qp for mac registration\n");
-		goto qp_err;
+		mlx4_unregister_mac(dev, priv->port, mac);
+		return err;
 	}
 
-	err = mlx4_en_uc_steer_add(priv, priv->dev->dev_addr, qpn, &reg_id);
-	if (err)
-		goto steer_err;
-
-	err = mlx4_en_tunnel_steer_add(priv, priv->dev->dev_addr, *qpn,
-				       &priv->tunnel_reg_id);
-	if (err)
-		goto tunnel_err;
-
-	entry = kmalloc(sizeof(*entry), GFP_KERNEL);
-	if (!entry) {
-		err = -ENOMEM;
-		goto alloc_err;
-	}
-	memcpy(entry->mac, priv->dev->dev_addr, sizeof(entry->mac));
-	memcpy(priv->current_mac, entry->mac, sizeof(priv->current_mac));
-	entry->reg_id = reg_id;
-
-	hlist_add_head_rcu(&entry->hlist,
-			   &priv->mac_hash[entry->mac[MLX4_EN_MAC_HASH_IDX]]);
-
 	return 0;
-
-alloc_err:
-	if (priv->tunnel_reg_id)
-		mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
-tunnel_err:
-	mlx4_en_uc_steer_release(priv, priv->dev->dev_addr, *qpn, reg_id);
-
-steer_err:
-	mlx4_qp_release_range(dev, *qpn, 1);
-
-qp_err:
-	mlx4_unregister_mac(dev, priv->port, mac);
-	return err;
 }
 
 static void mlx4_en_put_qp(struct mlx4_en_priv *priv)
@@ -645,39 +610,13 @@ static void mlx4_en_put_qp(struct mlx4_en_priv *priv)
 	struct mlx4_en_dev *mdev = priv->mdev;
 	struct mlx4_dev *dev = mdev->dev;
 	int qpn = priv->base_qpn;
-	u64 mac;
 
 	if (dev->caps.steering_mode == MLX4_STEERING_MODE_A0) {
-		mac = mlx4_mac_to_u64(priv->dev->dev_addr);
+		u64 mac = mlx4_mac_to_u64(priv->dev->dev_addr);
 		en_dbg(DRV, priv, "Registering MAC: %pM for deleting\n",
 		       priv->dev->dev_addr);
 		mlx4_unregister_mac(dev, priv->port, mac);
 	} else {
-		struct mlx4_mac_entry *entry;
-		struct hlist_node *tmp;
-		struct hlist_head *bucket;
-		unsigned int i;
-
-		for (i = 0; i < MLX4_EN_MAC_HASH_SIZE; ++i) {
-			bucket = &priv->mac_hash[i];
-			hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
-				mac = mlx4_mac_to_u64(entry->mac);
-				en_dbg(DRV, priv, "Registering MAC: %pM for deleting\n",
-				       entry->mac);
-				mlx4_en_uc_steer_release(priv, entry->mac,
-							 qpn, entry->reg_id);
-
-				mlx4_unregister_mac(dev, priv->port, mac);
-				hlist_del_rcu(&entry->hlist);
-				kfree_rcu(entry, rcu);
-			}
-		}
-
-		if (priv->tunnel_reg_id) {
-			mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
-			priv->tunnel_reg_id = 0;
-		}
-
 		en_dbg(DRV, priv, "Releasing qp: port %d, qpn %d\n",
 		       priv->port, qpn);
 		mlx4_qp_release_range(dev, qpn, 1);
@@ -1283,6 +1222,75 @@ static void mlx4_en_netpoll(struct net_device *dev)
 }
 #endif
 
+static int mlx4_en_set_rss_steer_rules(struct mlx4_en_priv *priv)
+{
+	u64 reg_id;
+	int err = 0;
+	int *qpn = &priv->base_qpn;
+	struct mlx4_mac_entry *entry;
+
+	err = mlx4_en_uc_steer_add(priv, priv->dev->dev_addr, qpn, &reg_id);
+	if (err)
+		return err;
+
+	err = mlx4_en_tunnel_steer_add(priv, priv->dev->dev_addr, *qpn,
+				       &priv->tunnel_reg_id);
+	if (err)
+		goto tunnel_err;
+
+	entry = kmalloc(sizeof(*entry), GFP_KERNEL);
+	if (!entry) {
+		err = -ENOMEM;
+		goto alloc_err;
+	}
+
+	memcpy(entry->mac, priv->dev->dev_addr, sizeof(entry->mac));
+	memcpy(priv->current_mac, entry->mac, sizeof(priv->current_mac));
+	entry->reg_id = reg_id;
+	hlist_add_head_rcu(&entry->hlist,
+			   &priv->mac_hash[entry->mac[MLX4_EN_MAC_HASH_IDX]]);
+
+	return 0;
+
+alloc_err:
+	if (priv->tunnel_reg_id)
+		mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+
+tunnel_err:
+	mlx4_en_uc_steer_release(priv, priv->dev->dev_addr, *qpn, reg_id);
+	return err;
+}
+
+static void mlx4_en_delete_rss_steer_rules(struct mlx4_en_priv *priv)
+{
+	u64 mac;
+	unsigned int i;
+	int qpn = priv->base_qpn;
+	struct hlist_head *bucket;
+	struct hlist_node *tmp;
+	struct mlx4_mac_entry *entry;
+
+	for (i = 0; i < MLX4_EN_MAC_HASH_SIZE; ++i) {
+		bucket = &priv->mac_hash[i];
+		hlist_for_each_entry_safe(entry, tmp, bucket, hlist) {
+			mac = mlx4_mac_to_u64(entry->mac);
+			en_dbg(DRV, priv, "Registering MAC:%pM for deleting\n",
+			       entry->mac);
+			mlx4_en_uc_steer_release(priv, entry->mac,
+						 qpn, entry->reg_id);
+
+			mlx4_unregister_mac(priv->mdev->dev, priv->port, mac);
+			hlist_del_rcu(&entry->hlist);
+			kfree_rcu(entry, rcu);
+		}
+	}
+
+	if (priv->tunnel_reg_id) {
+		mlx4_flow_detach(priv->mdev->dev, priv->tunnel_reg_id);
+		priv->tunnel_reg_id = 0;
+	}
+}
+
 static void mlx4_en_tx_timeout(struct net_device *dev)
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
@@ -1597,6 +1605,9 @@ int mlx4_en_start_port(struct net_device *dev)
 	}
 	mdev->mac_removed[priv->port] = 0;
 
+	priv->counter_index =
+			mlx4_get_default_counter_index(mdev->dev, priv->port);
+
 	err = mlx4_en_config_rss_steer(priv);
 	if (err) {
 		en_err(priv, "Failed configuring rss steering\n");
@@ -1681,6 +1692,11 @@ int mlx4_en_start_port(struct net_device *dev)
 		goto tx_err;
 	}
 
+	/* Set Unicast and VXLAN steering rules */
+	if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0 &&
+	    mlx4_en_set_rss_steer_rules(priv))
+		mlx4_warn(mdev, "Failed setting steering rules\n");
+
 	/* Attach rx QP to bradcast address */
 	eth_broadcast_addr(&mc_list[10]);
 	mc_list[5] = priv->port; /* needed for B0 steering support */
@@ -1755,6 +1771,7 @@ void mlx4_en_stop_port(struct net_device *dev, int detach)
 
 	/* Set port as not active */
 	priv->port_up = false;
+	priv->counter_index = MLX4_SINK_COUNTER_INDEX(mdev->dev);
 
 	/* Promsicuous mode */
 	if (mdev->dev->caps.steering_mode ==
@@ -1827,6 +1844,9 @@ void mlx4_en_stop_port(struct net_device *dev, int detach)
 	for (i = 0; i < priv->tx_ring_num; i++)
 		mlx4_en_free_tx_buf(dev, priv->tx_ring[i]);
 
+	if (mdev->dev->caps.steering_mode != MLX4_STEERING_MODE_A0)
+		mlx4_en_delete_rss_steer_rules(priv);
+
 	/* Free RSS qps */
 	mlx4_en_release_rss_steer(priv);
 
@@ -1891,6 +1911,7 @@ static void mlx4_en_clear_stats(struct net_device *dev)
 	       sizeof(priv->rx_priority_flowstats));
 	memset(&priv->tx_priority_flowstats, 0,
 	       sizeof(priv->tx_priority_flowstats));
+	memset(&priv->pf_stats, 0, sizeof(priv->pf_stats));
 
 	for (i = 0; i < priv->tx_ring_num; i++) {
 		priv->tx_ring[i]->bytes = 0;
@@ -1954,7 +1975,6 @@ void mlx4_en_free_resources(struct mlx4_en_priv *priv)
 	int i;
 
 #ifdef CONFIG_RFS_ACCEL
-	free_irq_cpu_rmap(priv->dev->rx_cpu_rmap);
 	priv->dev->rx_cpu_rmap = NULL;
 #endif
 
@@ -2008,11 +2028,7 @@ int mlx4_en_alloc_resources(struct mlx4_en_priv *priv)
 	}
 
 #ifdef CONFIG_RFS_ACCEL
-	if (priv->mdev->dev->caps.comp_pool) {
-		priv->dev->rx_cpu_rmap = alloc_irq_cpu_rmap(priv->mdev->dev->caps.comp_pool);
-		if (!priv->dev->rx_cpu_rmap)
-			goto err;
-	}
+	priv->dev->rx_cpu_rmap = mlx4_get_cpu_rmap(priv->mdev->dev, priv->port);
 #endif
 
 	return 0;
@@ -2056,6 +2072,9 @@ void mlx4_en_destroy_netdev(struct net_device *dev)
 	/* flush any pending task for this netdev */
 	flush_workqueue(mdev->workqueue);
 
+	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
+		mlx4_en_remove_timestamp(mdev);
+
 	/* Detach the netdev so tasks would not attempt to access it */
 	mutex_lock(&mdev->state_lock);
 	mdev->pndev[priv->port] = NULL;
@@ -2184,6 +2203,25 @@ static int mlx4_en_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
 	}
 }
 
+static netdev_features_t mlx4_en_fix_features(struct net_device *netdev,
+					      netdev_features_t features)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(netdev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	/* Since there is no support for separate RX C-TAG/S-TAG vlan accel
+	 * enable/disable make sure S-TAG flag is always in same state as
+	 * C-TAG.
+	 */
+	if (features & NETIF_F_HW_VLAN_CTAG_RX &&
+	    !(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN))
+		features |= NETIF_F_HW_VLAN_STAG_RX;
+	else
+		features &= ~NETIF_F_HW_VLAN_STAG_RX;
+
+	return features;
+}
+
 static int mlx4_en_set_features(struct net_device *netdev,
 		netdev_features_t features)
 {
@@ -2218,6 +2256,10 @@ static int mlx4_en_set_features(struct net_device *netdev,
 		en_info(priv, "Turn %s TX vlan strip offload\n",
 			(features & NETIF_F_HW_VLAN_CTAG_TX) ? "ON" : "OFF");
 
+	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_HW_VLAN_STAG_TX))
+		en_info(priv, "Turn %s TX S-VLAN strip offload\n",
+			(features & NETIF_F_HW_VLAN_STAG_TX) ? "ON" : "OFF");
+
 	if (DEV_FEATURE_CHANGED(netdev, features, NETIF_F_LOOPBACK)) {
 		en_info(priv, "Turn %s loopback\n",
 			(features & NETIF_F_LOOPBACK) ? "ON" : "OFF");
@@ -2288,6 +2330,15 @@ static int mlx4_en_set_vf_link_state(struct net_device *dev, int vf, int link_st
 	return mlx4_set_vf_link_state(mdev->dev, en_priv->port, vf, link_state);
 }
 
+static int mlx4_en_get_vf_stats(struct net_device *dev, int vf,
+				struct ifla_vf_stats *vf_stats)
+{
+	struct mlx4_en_priv *en_priv = netdev_priv(dev);
+	struct mlx4_en_dev *mdev = en_priv->mdev;
+
+	return mlx4_get_vf_stats(mdev->dev, en_priv->port, vf, vf_stats);
+}
+
 #define PORT_ID_BYTE_LEN 8
 static int mlx4_en_get_phys_port_id(struct net_device *dev,
 				    struct netdev_phys_item_id *ppid)
@@ -2330,8 +2381,6 @@ out:
 	/* set offloads */
 	priv->dev->hw_enc_features |= NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
 				      NETIF_F_TSO | NETIF_F_GSO_UDP_TUNNEL;
-	priv->dev->hw_features |= NETIF_F_GSO_UDP_TUNNEL;
-	priv->dev->features    |= NETIF_F_GSO_UDP_TUNNEL;
 }
 
 static void mlx4_en_del_vxlan_offloads(struct work_struct *work)
@@ -2342,8 +2391,6 @@ static void mlx4_en_del_vxlan_offloads(struct work_struct *work)
 	/* unset offloads */
 	priv->dev->hw_enc_features &= ~(NETIF_F_IP_CSUM | NETIF_F_RXCSUM |
 				      NETIF_F_TSO | NETIF_F_GSO_UDP_TUNNEL);
-	priv->dev->hw_features &= ~NETIF_F_GSO_UDP_TUNNEL;
-	priv->dev->features    &= ~NETIF_F_GSO_UDP_TUNNEL;
 
 	ret = mlx4_SET_PORT_VXLAN(priv->mdev->dev, priv->port,
 				  VXLAN_STEER_BY_OUTER_MAC, 0);
@@ -2451,6 +2498,7 @@ static const struct net_device_ops mlx4_netdev_ops = {
 	.ndo_poll_controller	= mlx4_en_netpoll,
 #endif
 	.ndo_set_features	= mlx4_en_set_features,
+	.ndo_fix_features	= mlx4_en_fix_features,
 	.ndo_setup_tc		= mlx4_en_setup_tc,
 #ifdef CONFIG_RFS_ACCEL
 	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
@@ -2485,11 +2533,13 @@ static const struct net_device_ops mlx4_netdev_ops_master = {
 	.ndo_set_vf_rate	= mlx4_en_set_vf_rate,
 	.ndo_set_vf_spoofchk	= mlx4_en_set_vf_spoofchk,
 	.ndo_set_vf_link_state	= mlx4_en_set_vf_link_state,
+	.ndo_get_vf_stats       = mlx4_en_get_vf_stats,
 	.ndo_get_vf_config	= mlx4_en_get_vf_config,
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= mlx4_en_netpoll,
 #endif
 	.ndo_set_features	= mlx4_en_set_features,
+	.ndo_fix_features	= mlx4_en_fix_features,
 	.ndo_setup_tc		= mlx4_en_setup_tc,
 #ifdef CONFIG_RFS_ACCEL
 	.ndo_rx_flow_steer	= mlx4_en_filter_rfs,
@@ -2682,7 +2732,7 @@ void mlx4_en_update_pfc_stats_bitmap(struct mlx4_dev *dev,
 				     u8 rx_ppp, u8 rx_pause,
 				     u8 tx_ppp, u8 tx_pause)
 {
-	int last_i = NUM_MAIN_STATS + NUM_PORT_STATS;
+	int last_i = NUM_MAIN_STATS + NUM_PORT_STATS + NUM_PF_STATS;
 
 	if (!mlx4_is_slave(dev) &&
 	    (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_FLOWSTATS_EN)) {
@@ -2744,6 +2794,11 @@ void mlx4_en_set_stats_bitmap(struct mlx4_dev *dev,
 	bitmap_set(stats_bitmap->bitmap, last_i, NUM_PORT_STATS);
 	last_i += NUM_PORT_STATS;
 
+	if (mlx4_is_master(dev))
+		bitmap_set(stats_bitmap->bitmap, last_i,
+			   NUM_PF_STATS);
+	last_i += NUM_PF_STATS;
+
 	mlx4_en_update_pfc_stats_bitmap(dev, stats_bitmap,
 					rx_ppp, rx_pause,
 					tx_ppp, tx_pause);
@@ -2760,7 +2815,6 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	struct mlx4_en_priv *priv;
 	int i;
 	int err;
-	u64 mac_u64;
 
 	dev = alloc_etherdev_mqs(sizeof(struct mlx4_en_priv),
 				 MAX_TX_RINGS, MAX_RX_RINGS);
@@ -2779,6 +2833,7 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 
 	priv = netdev_priv(dev);
 	memset(priv, 0, sizeof(struct mlx4_en_priv));
+	priv->counter_index = MLX4_SINK_COUNTER_INDEX(mdev->dev);
 	spin_lock_init(&priv->stats_lock);
 	INIT_WORK(&priv->rx_mode_task, mlx4_en_do_set_rx_mode);
 	INIT_WORK(&priv->watchdog_task, mlx4_en_restart);
@@ -2851,17 +2906,17 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	dev->addr_len = ETH_ALEN;
 	mlx4_en_u64_to_mac(dev->dev_addr, mdev->dev->caps.def_mac[priv->port]);
 	if (!is_valid_ether_addr(dev->dev_addr)) {
-		if (mlx4_is_slave(priv->mdev->dev)) {
-			eth_hw_addr_random(dev);
-			en_warn(priv, "Assigned random MAC address %pM\n", dev->dev_addr);
-			mac_u64 = mlx4_mac_to_u64(dev->dev_addr);
-			mdev->dev->caps.def_mac[priv->port] = mac_u64;
-		} else {
-			en_err(priv, "Port: %d, invalid mac burned: %pM, quiting\n",
-			       priv->port, dev->dev_addr);
-			err = -EINVAL;
-			goto out;
-		}
+		en_err(priv, "Port: %d, invalid mac burned: %pM, quiting\n",
+		       priv->port, dev->dev_addr);
+		err = -EINVAL;
+		goto out;
+	} else if (mlx4_is_slave(priv->mdev->dev) &&
+		   (priv->mdev->dev->port_random_macs & 1 << priv->port)) {
+		/* Random MAC was assigned in mlx4_slave_cap
+		 * in mlx4_core module
+		 */
+		dev->addr_assign_type |= NET_ADDR_RANDOM;
+		en_warn(priv, "Assigned random MAC address %pM\n", dev->dev_addr);
 	}
 
 	memcpy(priv->current_mac, dev->dev_addr, sizeof(priv->current_mac));
@@ -2915,6 +2970,27 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	dev->hw_features |= NETIF_F_LOOPBACK |
 			NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX;
 
+	if (!(mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN)) {
+		dev->features |= NETIF_F_HW_VLAN_STAG_RX |
+			NETIF_F_HW_VLAN_STAG_FILTER;
+		dev->hw_features |= NETIF_F_HW_VLAN_STAG_RX;
+	}
+
+	if (mlx4_is_slave(mdev->dev)) {
+		int phv;
+
+		err = get_phv_bit(mdev->dev, port, &phv);
+		if (!err && phv) {
+			dev->hw_features |= NETIF_F_HW_VLAN_STAG_TX;
+			priv->pflags |= MLX4_EN_PRIV_FLAGS_PHV;
+		}
+	} else {
+		if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PHV_EN &&
+		    !(mdev->dev->caps.flags2 &
+		      MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN))
+			dev->hw_features |= NETIF_F_HW_VLAN_STAG_TX;
+	}
+
 	if (mdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)
 		dev->hw_features |= NETIF_F_RXFCS;
 
@@ -2940,6 +3016,11 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 		priv->rss_hash_fn = ETH_RSS_HASH_TOP;
 	}
 
+	if (mdev->dev->caps.tunnel_offload_mode == MLX4_TUNNEL_OFFLOAD_MODE_VXLAN) {
+		dev->hw_features |= NETIF_F_GSO_UDP_TUNNEL;
+		dev->features    |= NETIF_F_GSO_UDP_TUNNEL;
+	}
+
 	mdev->pndev[port] = dev;
 	mdev->upper[port] = NULL;
 
@@ -2981,9 +3062,12 @@ int mlx4_en_init_netdev(struct mlx4_en_dev *mdev, int port,
 	}
 	queue_delayed_work(mdev->workqueue, &priv->stats_task, STATS_DELAY);
 
+	/* Initialize time stamp mechanism */
 	if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS)
-		queue_delayed_work(mdev->workqueue, &priv->service_task,
-				   SERVICE_TASK_DELAY);
+		mlx4_en_init_timestamp(mdev);
+
+	queue_delayed_work(mdev->workqueue, &priv->service_task,
+			   SERVICE_TASK_DELAY);
 
 	mlx4_en_set_stats_bitmap(mdev->dev, &priv->stats_bitmap,
 				 mdev->profile.prof[priv->port].rx_ppp,
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_port.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_port.c
index 0a56f010c..3904b5fc0 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_port.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_port.c
@@ -149,6 +149,7 @@ static unsigned long en_stats_adder(__be64 *start, __be64 *next, int num)
 
 int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
 {
+	struct mlx4_counter tmp_counter_stats;
 	struct mlx4_en_stat_out_mbox *mlx4_en_stats;
 	struct mlx4_en_stat_out_flow_control_mbox *flowstats;
 	struct mlx4_en_priv *priv = netdev_priv(mdev->pndev[port]);
@@ -156,7 +157,7 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
 	struct mlx4_cmd_mailbox *mailbox;
 	u64 in_mod = reset << 8 | port;
 	int err;
-	int i;
+	int i, counter_index;
 
 	mailbox = mlx4_alloc_cmd_mailbox(mdev->dev);
 	if (IS_ERR(mailbox))
@@ -202,6 +203,20 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
 		priv->port_stats.tso_packets       += ring->tso_packets;
 		priv->port_stats.xmit_more         += ring->xmit_more;
 	}
+	if (mlx4_is_master(mdev->dev)) {
+		stats->rx_packets = en_stats_adder(&mlx4_en_stats->RTOT_prio_0,
+						   &mlx4_en_stats->RTOT_prio_1,
+						   NUM_PRIORITIES);
+		stats->tx_packets = en_stats_adder(&mlx4_en_stats->TTOT_prio_0,
+						   &mlx4_en_stats->TTOT_prio_1,
+						   NUM_PRIORITIES);
+		stats->rx_bytes = en_stats_adder(&mlx4_en_stats->ROCT_prio_0,
+						 &mlx4_en_stats->ROCT_prio_1,
+						 NUM_PRIORITIES);
+		stats->tx_bytes = en_stats_adder(&mlx4_en_stats->TOCT_prio_0,
+						 &mlx4_en_stats->TOCT_prio_1,
+						 NUM_PRIORITIES);
+	}
 
 	/* net device stats */
 	stats->rx_errors = be64_to_cpu(mlx4_en_stats->PCS) +
@@ -223,11 +238,11 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
 	stats->collisions = 0;
 	stats->rx_dropped = be32_to_cpu(mlx4_en_stats->RDROP);
 	stats->rx_length_errors = be32_to_cpu(mlx4_en_stats->RdropLength);
-	stats->rx_over_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
+	stats->rx_over_errors = 0;
 	stats->rx_crc_errors = be32_to_cpu(mlx4_en_stats->RCRC);
 	stats->rx_frame_errors = 0;
 	stats->rx_fifo_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
-	stats->rx_missed_errors = be32_to_cpu(mlx4_en_stats->RdropOvflw);
+	stats->rx_missed_errors = 0;
 	stats->tx_aborted_errors = 0;
 	stats->tx_carrier_errors = 0;
 	stats->tx_fifo_errors = 0;
@@ -296,6 +311,11 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
 
 	spin_unlock_bh(&priv->stats_lock);
 
+	memset(&tmp_counter_stats, 0, sizeof(tmp_counter_stats));
+	counter_index = mlx4_get_default_counter_index(mdev->dev, port);
+	err = mlx4_get_counter_stats(mdev->dev, counter_index,
+				     &tmp_counter_stats, reset);
+
 	/* 0xffs indicates invalid value */
 	memset(mailbox->buf, 0xff, sizeof(*flowstats) * MLX4_NUM_PRIORITIES);
 
@@ -314,6 +334,13 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset)
 
 	spin_lock_bh(&priv->stats_lock);
 
+	if (tmp_counter_stats.counter_mode == 0) {
+		priv->pf_stats.rx_bytes   = be64_to_cpu(tmp_counter_stats.rx_bytes);
+		priv->pf_stats.tx_bytes   = be64_to_cpu(tmp_counter_stats.tx_bytes);
+		priv->pf_stats.rx_packets = be64_to_cpu(tmp_counter_stats.rx_frames);
+		priv->pf_stats.tx_packets = be64_to_cpu(tmp_counter_stats.tx_frames);
+	}
+
 	for (i = 0; i < MLX4_NUM_PRIORITIES; i++)	{
 		priv->rx_priority_flowstats[i].rx_pause =
 			be64_to_cpu(flowstats[i].rx_pause);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_resources.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_resources.c
index 34f2fdf4f..12aab5a65 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_resources.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_resources.c
@@ -66,9 +66,18 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
 		context->pri_path.sched_queue |= user_prio << 3;
 		context->pri_path.feup = MLX4_FEUP_FORCE_ETH_UP;
 	}
-	context->pri_path.counter_index = 0xff;
+	context->pri_path.counter_index = priv->counter_index;
 	context->cqn_send = cpu_to_be32(cqn);
 	context->cqn_recv = cpu_to_be32(cqn);
+	if (!rss &&
+	    (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_LB_SRC_CHK) &&
+	    context->pri_path.counter_index !=
+			    MLX4_SINK_COUNTER_INDEX(mdev->dev)) {
+		/* disable multicast loopback to qp with same counter */
+		if (!(dev->features & NETIF_F_LOOPBACK))
+			context->pri_path.fl |= MLX4_FL_ETH_SRC_CHECK_MC_LB;
+		context->pri_path.control |= MLX4_CTRL_ETH_SRC_CHECK_IF_COUNTER;
+	}
 	context->db_rec_addr = cpu_to_be64(priv->res.db.dma << 2);
 	if (!(dev->features & NETIF_F_HW_VLAN_CTAG_RX))
 		context->param3 |= cpu_to_be32(1 << 30);
@@ -80,6 +89,22 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
 	}
 }
 
+int mlx4_en_change_mcast_lb(struct mlx4_en_priv *priv, struct mlx4_qp *qp,
+			    int loopback)
+{
+	int ret;
+	struct mlx4_update_qp_params qp_params;
+
+	memset(&qp_params, 0, sizeof(qp_params));
+	if (!loopback)
+		qp_params.flags = MLX4_UPDATE_QP_PARAMS_FLAGS_ETH_CHECK_MC_LB;
+
+	ret = mlx4_update_qp(priv->mdev->dev, qp->qpn,
+			     MLX4_UPDATE_QP_ETH_SRC_CHECK_MC_LB,
+			     &qp_params);
+
+	return ret;
+}
 
 int mlx4_en_map_buffer(struct mlx4_buf *buf)
 {
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_rx.c
index 80aac2010..e7a5000aa 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_rx.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_rx.c
@@ -246,7 +246,6 @@ static int mlx4_en_prepare_rx_desc(struct mlx4_en_priv *priv,
 
 static inline bool mlx4_en_is_ring_empty(struct mlx4_en_rx_ring *ring)
 {
-	BUG_ON((u32)(ring->prod - ring->cons) > ring->actual_size);
 	return ring->prod == ring->cons;
 }
 
@@ -337,15 +336,10 @@ void mlx4_en_set_num_rx_rings(struct mlx4_en_dev *mdev)
 	struct mlx4_dev *dev = mdev->dev;
 
 	mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_ETH) {
-		if (!dev->caps.comp_pool)
-			num_of_eqs = max_t(int, MIN_RX_RINGS,
-					   min_t(int,
-						 dev->caps.num_comp_vectors,
-						 DEF_RX_RINGS));
-		else
-			num_of_eqs = min_t(int, MAX_MSIX_P_PORT,
-					   dev->caps.comp_pool/
-					   dev->caps.num_ports) - 1;
+		num_of_eqs = max_t(int, MIN_RX_RINGS,
+				   min_t(int,
+					 mlx4_get_eqs_per_port(mdev->dev, i),
+					 DEF_RX_RINGS));
 
 		num_rx_rings = mlx4_low_memory_profile() ? MIN_RX_RINGS :
 			min_t(int, num_of_eqs,
@@ -731,7 +725,7 @@ static int check_csum(struct mlx4_cqe *cqe, struct sk_buff *skb, void *va,
 
 	hw_checksum = csum_unfold((__force __sum16)cqe->checksum);
 
-	if (cqe->vlan_my_qpn & cpu_to_be32(MLX4_CQE_VLAN_PRESENT_MASK) &&
+	if (cqe->vlan_my_qpn & cpu_to_be32(MLX4_CQE_CVLAN_PRESENT_MASK) &&
 	    !(dev_features & NETIF_F_HW_VLAN_CTAG_RX)) {
 		hw_checksum = get_fixed_vlan_csum(hw_checksum, hdr);
 		hdr += sizeof(struct vlan_hdr);
@@ -912,17 +906,25 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 				gro_skb->csum_level = 1;
 
 			if ((cqe->vlan_my_qpn &
-			    cpu_to_be32(MLX4_CQE_VLAN_PRESENT_MASK)) &&
+			    cpu_to_be32(MLX4_CQE_CVLAN_PRESENT_MASK)) &&
 			    (dev->features & NETIF_F_HW_VLAN_CTAG_RX)) {
 				u16 vid = be16_to_cpu(cqe->sl_vid);
 
 				__vlan_hwaccel_put_tag(gro_skb, htons(ETH_P_8021Q), vid);
+			} else if ((be32_to_cpu(cqe->vlan_my_qpn) &
+				  MLX4_CQE_SVLAN_PRESENT_MASK) &&
+				 (dev->features & NETIF_F_HW_VLAN_STAG_RX)) {
+				__vlan_hwaccel_put_tag(gro_skb,
+						       htons(ETH_P_8021AD),
+						       be16_to_cpu(cqe->sl_vid));
 			}
 
 			if (dev->features & NETIF_F_RXHASH)
 				skb_set_hash(gro_skb,
 					     be32_to_cpu(cqe->immed_rss_invalid),
-					     PKT_HASH_TYPE_L3);
+					     (ip_summed == CHECKSUM_UNNECESSARY) ?
+						PKT_HASH_TYPE_L4 :
+						PKT_HASH_TYPE_L3);
 
 			skb_record_rx_queue(gro_skb, cq->ring);
 			skb_mark_napi_id(gro_skb, &cq->napi);
@@ -968,12 +970,19 @@ int mlx4_en_process_rx_cq(struct net_device *dev, struct mlx4_en_cq *cq, int bud
 		if (dev->features & NETIF_F_RXHASH)
 			skb_set_hash(skb,
 				     be32_to_cpu(cqe->immed_rss_invalid),
-				     PKT_HASH_TYPE_L3);
+				     (ip_summed == CHECKSUM_UNNECESSARY) ?
+					PKT_HASH_TYPE_L4 :
+					PKT_HASH_TYPE_L3);
 
 		if ((be32_to_cpu(cqe->vlan_my_qpn) &
-		    MLX4_CQE_VLAN_PRESENT_MASK) &&
+		    MLX4_CQE_CVLAN_PRESENT_MASK) &&
 		    (dev->features & NETIF_F_HW_VLAN_CTAG_RX))
 			__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), be16_to_cpu(cqe->sl_vid));
+		else if ((be32_to_cpu(cqe->vlan_my_qpn) &
+			  MLX4_CQE_SVLAN_PRESENT_MASK) &&
+			 (dev->features & NETIF_F_HW_VLAN_STAG_RX))
+			__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021AD),
+					       be16_to_cpu(cqe->sl_vid));
 
 		if (ring->hwtstamp_rx_filter == HWTSTAMP_FILTER_ALL) {
 			timestamp = mlx4_en_get_cqe_ts(cqe);
@@ -1038,13 +1047,15 @@ int mlx4_en_poll_rx_cq(struct napi_struct *napi, int budget)
 
 	/* If we used up all the quota - we're probably not done yet... */
 	if (done == budget) {
-		int cpu_curr;
 		const struct cpumask *aff;
+		struct irq_data *idata;
+		int cpu_curr;
 
 		INC_PERF_COUNTER(priv->pstats.napi_quota);
 
 		cpu_curr = smp_processor_id();
-		aff = irq_desc_get_irq_data(cq->irq_desc)->affinity;
+		idata = irq_desc_get_irq_data(cq->irq_desc);
+		aff = irq_data_get_affinity_mask(idata);
 
 		if (likely(cpumask_test_cpu(cpu_curr, aff)))
 			return budget;
@@ -1071,7 +1082,10 @@ static const int frag_sizes[] = {
 void mlx4_en_calc_rx_buf(struct net_device *dev)
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
-	int eff_mtu = dev->mtu + ETH_HLEN + VLAN_HLEN;
+	/* VLAN_HLEN is added twice,to support skb vlan tagged with multiple
+	 * headers. (For example: ETH_P_8021Q and ETH_P_8021AD).
+	 */
+	int eff_mtu = dev->mtu + ETH_HLEN + (2 * VLAN_HLEN);
 	int buf_size = 0;
 	int i = 0;
 
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/kernel/drivers/net/ethernet/mellanox/mlx4/en_tx.c
index c10d98f6a..4421bf546 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/en_tx.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/en_tx.c
@@ -718,6 +718,7 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 	u32 index, bf_index;
 	__be32 op_own;
 	u16 vlan_tag = 0;
+	u16 vlan_proto = 0;
 	int i_frag;
 	int lso_header_size;
 	void *fragptr = NULL;
@@ -750,9 +751,10 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 		goto tx_drop;
 	}
 
-	if (skb_vlan_tag_present(skb))
+	if (skb_vlan_tag_present(skb)) {
 		vlan_tag = skb_vlan_tag_get(skb);
-
+		vlan_proto = be16_to_cpu(skb->vlan_proto);
+	}
 
 	netdev_txq_bql_enqueue_prefetchw(ring->tx_queue);
 
@@ -958,8 +960,13 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
 		ring->bf.offset ^= ring->bf.buf_size;
 	} else {
 		tx_desc->ctrl.vlan_tag = cpu_to_be16(vlan_tag);
-		tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_VLAN *
-			!!skb_vlan_tag_present(skb);
+		if (vlan_proto == ETH_P_8021AD)
+			tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_SVLAN;
+		else if (vlan_proto == ETH_P_8021Q)
+			tx_desc->ctrl.ins_vlan = MLX4_WQE_CTRL_INS_CVLAN;
+		else
+			tx_desc->ctrl.ins_vlan = 0;
+
 		tx_desc->ctrl.fence_size = real_size;
 
 		/* Ensure new descriptor hits memory
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/eq.c b/kernel/drivers/net/ethernet/mellanox/mlx4/eq.c
index 983b1d512..603d1c3d3 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/eq.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/eq.c
@@ -145,7 +145,7 @@ void mlx4_gen_slave_eqe(struct work_struct *work)
 	struct mlx4_slave_event_eq *slave_eq = &mfunc->master.slave_eq;
 	struct mlx4_eqe *eqe;
 	u8 slave;
-	int i;
+	int i, phys_port, slave_port;
 
 	for (eqe = next_slave_event_eqe(slave_eq); eqe;
 	      eqe = next_slave_event_eqe(slave_eq)) {
@@ -154,9 +154,20 @@ void mlx4_gen_slave_eqe(struct work_struct *work)
 		/* All active slaves need to receive the event */
 		if (slave == ALL_SLAVES) {
 			for (i = 0; i <= dev->persist->num_vfs; i++) {
+				phys_port = 0;
+				if (eqe->type == MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT &&
+				    eqe->subtype == MLX4_DEV_PMC_SUBTYPE_PORT_INFO) {
+					phys_port  = eqe->event.port_mgmt_change.port;
+					slave_port = mlx4_phys_to_slave_port(dev, i, phys_port);
+					if (slave_port < 0) /* VF doesn't have this port */
+						continue;
+					eqe->event.port_mgmt_change.port = slave_port;
+				}
 				if (mlx4_GEN_EQE(dev, i, eqe))
 					mlx4_warn(dev, "Failed to generate event for slave %d\n",
 						  i);
+				if (phys_port)
+					eqe->event.port_mgmt_change.port = phys_port;
 			}
 		} else {
 			if (mlx4_GEN_EQE(dev, slave, eqe))
@@ -185,7 +196,7 @@ static void slave_event(struct mlx4_dev *dev, u8 slave, struct mlx4_eqe *eqe)
 		return;
 	}
 
-	memcpy(s_eqe, eqe, dev->caps.eqe_size - 1);
+	memcpy(s_eqe, eqe, sizeof(struct mlx4_eqe) - 1);
 	s_eqe->slave_id = slave;
 	/* ensure all information is written before setting the ownersip bit */
 	dma_wmb();
@@ -210,6 +221,22 @@ static void mlx4_slave_event(struct mlx4_dev *dev, int slave,
 	slave_event(dev, slave, eqe);
 }
 
+#if defined(CONFIG_SMP)
+static void mlx4_set_eq_affinity_hint(struct mlx4_priv *priv, int vec)
+{
+	int hint_err;
+	struct mlx4_dev *dev = &priv->dev;
+	struct mlx4_eq *eq = &priv->eq_table.eq[vec];
+
+	if (!eq->affinity_mask || cpumask_empty(eq->affinity_mask))
+		return;
+
+	hint_err = irq_set_affinity_hint(eq->irq, eq->affinity_mask);
+	if (hint_err)
+		mlx4_warn(dev, "irq_set_affinity_hint failed, err %d\n", hint_err);
+}
+#endif
+
 int mlx4_gen_pkey_eqe(struct mlx4_dev *dev, int slave, u8 port)
 {
 	struct mlx4_eqe eqe;
@@ -224,7 +251,7 @@ int mlx4_gen_pkey_eqe(struct mlx4_dev *dev, int slave, u8 port)
 
 	eqe.type = MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT;
 	eqe.subtype = MLX4_DEV_PMC_SUBTYPE_PKEY_TABLE;
-	eqe.event.port_mgmt_change.port = port;
+	eqe.event.port_mgmt_change.port = mlx4_phys_to_slave_port(dev, slave, port);
 
 	return mlx4_GEN_EQE(dev, slave, &eqe);
 }
@@ -241,7 +268,7 @@ int mlx4_gen_guid_change_eqe(struct mlx4_dev *dev, int slave, u8 port)
 
 	eqe.type = MLX4_EVENT_TYPE_PORT_MNG_CHG_EVENT;
 	eqe.subtype = MLX4_DEV_PMC_SUBTYPE_GUID_INFO;
-	eqe.event.port_mgmt_change.port = port;
+	eqe.event.port_mgmt_change.port = mlx4_phys_to_slave_port(dev, slave, port);
 
 	return mlx4_GEN_EQE(dev, slave, &eqe);
 }
@@ -251,6 +278,7 @@ int mlx4_gen_port_state_change_eqe(struct mlx4_dev *dev, int slave, u8 port,
 				   u8 port_subtype_change)
 {
 	struct mlx4_eqe eqe;
+	u8 slave_port = mlx4_phys_to_slave_port(dev, slave, port);
 
 	/*don't send if we don't have the that slave */
 	if (dev->persist->num_vfs < slave)
@@ -259,7 +287,7 @@ int mlx4_gen_port_state_change_eqe(struct mlx4_dev *dev, int slave, u8 port,
 
 	eqe.type = MLX4_EVENT_TYPE_PORT_CHANGE;
 	eqe.subtype = port_subtype_change;
-	eqe.event.port_change.port = cpu_to_be32(port << 28);
+	eqe.event.port_change.port = cpu_to_be32(slave_port << 28);
 
 	mlx4_dbg(dev, "%s: sending: %d to slave: %d on port: %d\n", __func__,
 		 port_subtype_change, slave, port);
@@ -589,6 +617,10 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
 						if (SLAVE_PORT_GEN_EVENT_DOWN ==  gen_event) {
 							if (i == mlx4_master_func_num(dev))
 								continue;
+							eqe->event.port_change.port =
+								cpu_to_be32(
+								(be32_to_cpu(eqe->event.port_change.port) & 0xFFFFFFF)
+								| (mlx4_phys_to_slave_port(dev, i, port) << 28));
 							mlx4_slave_event(dev, i, eqe);
 						}
 					}
@@ -879,8 +911,8 @@ static int mlx4_num_eq_uar(struct mlx4_dev *dev)
 	 * we need to map, take the difference of highest index and
 	 * the lowest index we'll use and add 1.
 	 */
-	return (dev->caps.num_comp_vectors + 1 + dev->caps.reserved_eqs +
-		 dev->caps.comp_pool)/4 - dev->caps.reserved_eqs/4 + 1;
+	return (dev->caps.num_comp_vectors + 1 + dev->caps.reserved_eqs) / 4 -
+		dev->caps.reserved_eqs / 4 + 1;
 }
 
 static void __iomem *mlx4_get_eq_uar(struct mlx4_dev *dev, struct mlx4_eq *eq)
@@ -1069,32 +1101,21 @@ static void mlx4_free_eq(struct mlx4_dev *dev,
 static void mlx4_free_irqs(struct mlx4_dev *dev)
 {
 	struct mlx4_eq_table *eq_table = &mlx4_priv(dev)->eq_table;
-	struct mlx4_priv *priv = mlx4_priv(dev);
-	int	i, vec;
+	int	i;
 
 	if (eq_table->have_irq)
 		free_irq(dev->persist->pdev->irq, dev);
 
 	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
 		if (eq_table->eq[i].have_irq) {
+			free_cpumask_var(eq_table->eq[i].affinity_mask);
+#if defined(CONFIG_SMP)
+			irq_set_affinity_hint(eq_table->eq[i].irq, NULL);
+#endif
 			free_irq(eq_table->eq[i].irq, eq_table->eq + i);
 			eq_table->eq[i].have_irq = 0;
 		}
 
-	for (i = 0; i < dev->caps.comp_pool; i++) {
-		/*
-		 * Freeing the assigned irq's
-		 * all bits should be 0, but we need to validate
-		 */
-		if (priv->msix_ctl.pool_bm & 1ULL << i) {
-			/* NO need protecting*/
-			vec = dev->caps.num_comp_vectors + 1 + i;
-			free_irq(priv->eq_table.eq[vec].irq,
-				 &priv->eq_table.eq[vec]);
-		}
-	}
-
-
 	kfree(eq_table->irq_names);
 }
 
@@ -1175,76 +1196,73 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
 	}
 
 	priv->eq_table.irq_names =
-		kmalloc(MLX4_IRQNAME_SIZE * (dev->caps.num_comp_vectors + 1 +
-					     dev->caps.comp_pool),
+		kmalloc(MLX4_IRQNAME_SIZE * (dev->caps.num_comp_vectors + 1),
 			GFP_KERNEL);
 	if (!priv->eq_table.irq_names) {
 		err = -ENOMEM;
-		goto err_out_bitmap;
+		goto err_out_clr_int;
 	}
 
-	for (i = 0; i < dev->caps.num_comp_vectors; ++i) {
-		err = mlx4_create_eq(dev, dev->caps.num_cqs -
-					  dev->caps.reserved_cqs +
-					  MLX4_NUM_SPARE_EQE,
-				     (dev->flags & MLX4_FLAG_MSI_X) ? i : 0,
-				     &priv->eq_table.eq[i]);
-		if (err) {
-			--i;
-			goto err_out_unmap;
-		}
-	}
-
-	err = mlx4_create_eq(dev, MLX4_NUM_ASYNC_EQE + MLX4_NUM_SPARE_EQE,
-			     (dev->flags & MLX4_FLAG_MSI_X) ? dev->caps.num_comp_vectors : 0,
-			     &priv->eq_table.eq[dev->caps.num_comp_vectors]);
-	if (err)
-		goto err_out_comp;
-
-	/*if additional completion vectors poolsize is 0 this loop will not run*/
-	for (i = dev->caps.num_comp_vectors + 1;
-	      i < dev->caps.num_comp_vectors + dev->caps.comp_pool + 1; ++i) {
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i) {
+		if (i == MLX4_EQ_ASYNC) {
+			err = mlx4_create_eq(dev,
+					     MLX4_NUM_ASYNC_EQE + MLX4_NUM_SPARE_EQE,
+					     0, &priv->eq_table.eq[MLX4_EQ_ASYNC]);
+		} else {
+			struct mlx4_eq	*eq = &priv->eq_table.eq[i];
+#ifdef CONFIG_RFS_ACCEL
+			int port = find_first_bit(eq->actv_ports.ports,
+						  dev->caps.num_ports) + 1;
+
+			if (port <= dev->caps.num_ports) {
+				struct mlx4_port_info *info =
+					&mlx4_priv(dev)->port[port];
+
+				if (!info->rmap) {
+					info->rmap = alloc_irq_cpu_rmap(
+						mlx4_get_eqs_per_port(dev, port));
+					if (!info->rmap) {
+						mlx4_warn(dev, "Failed to allocate cpu rmap\n");
+						err = -ENOMEM;
+						goto err_out_unmap;
+					}
+				}
 
-		err = mlx4_create_eq(dev, dev->caps.num_cqs -
-					  dev->caps.reserved_cqs +
-					  MLX4_NUM_SPARE_EQE,
-				     (dev->flags & MLX4_FLAG_MSI_X) ? i : 0,
-				     &priv->eq_table.eq[i]);
-		if (err) {
-			--i;
-			goto err_out_unmap;
+				err = irq_cpu_rmap_add(
+					info->rmap, eq->irq);
+				if (err)
+					mlx4_warn(dev, "Failed adding irq rmap\n");
+			}
+#endif
+			err = mlx4_create_eq(dev, dev->caps.num_cqs -
+						  dev->caps.reserved_cqs +
+						  MLX4_NUM_SPARE_EQE,
+					     (dev->flags & MLX4_FLAG_MSI_X) ?
+					     i + 1 - !!(i > MLX4_EQ_ASYNC) : 0,
+					     eq);
 		}
+		if (err)
+			goto err_out_unmap;
 	}
 
-
 	if (dev->flags & MLX4_FLAG_MSI_X) {
 		const char *eq_name;
 
-		for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i) {
-			if (i < dev->caps.num_comp_vectors) {
-				snprintf(priv->eq_table.irq_names +
-					 i * MLX4_IRQNAME_SIZE,
-					 MLX4_IRQNAME_SIZE,
-					 "mlx4-comp-%d@pci:%s", i,
-					 pci_name(dev->persist->pdev));
-			} else {
-				snprintf(priv->eq_table.irq_names +
-					 i * MLX4_IRQNAME_SIZE,
-					 MLX4_IRQNAME_SIZE,
-					 "mlx4-async@pci:%s",
-					 pci_name(dev->persist->pdev));
-			}
+		snprintf(priv->eq_table.irq_names +
+			 MLX4_EQ_ASYNC * MLX4_IRQNAME_SIZE,
+			 MLX4_IRQNAME_SIZE,
+			 "mlx4-async@pci:%s",
+			 pci_name(dev->persist->pdev));
+		eq_name = priv->eq_table.irq_names +
+			MLX4_EQ_ASYNC * MLX4_IRQNAME_SIZE;
 
-			eq_name = priv->eq_table.irq_names +
-				  i * MLX4_IRQNAME_SIZE;
-			err = request_irq(priv->eq_table.eq[i].irq,
-					  mlx4_msi_x_interrupt, 0, eq_name,
-					  priv->eq_table.eq + i);
-			if (err)
-				goto err_out_async;
+		err = request_irq(priv->eq_table.eq[MLX4_EQ_ASYNC].irq,
+				  mlx4_msi_x_interrupt, 0, eq_name,
+				  priv->eq_table.eq + MLX4_EQ_ASYNC);
+		if (err)
+			goto err_out_unmap;
 
-			priv->eq_table.eq[i].have_irq = 1;
-		}
+		priv->eq_table.eq[MLX4_EQ_ASYNC].have_irq = 1;
 	} else {
 		snprintf(priv->eq_table.irq_names,
 			 MLX4_IRQNAME_SIZE,
@@ -1253,36 +1271,38 @@ int mlx4_init_eq_table(struct mlx4_dev *dev)
 		err = request_irq(dev->persist->pdev->irq, mlx4_interrupt,
 				  IRQF_SHARED, priv->eq_table.irq_names, dev);
 		if (err)
-			goto err_out_async;
+			goto err_out_unmap;
 
 		priv->eq_table.have_irq = 1;
 	}
 
 	err = mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0,
-			  priv->eq_table.eq[dev->caps.num_comp_vectors].eqn);
+			  priv->eq_table.eq[MLX4_EQ_ASYNC].eqn);
 	if (err)
 		mlx4_warn(dev, "MAP_EQ for async EQ %d failed (%d)\n",
-			   priv->eq_table.eq[dev->caps.num_comp_vectors].eqn, err);
+			   priv->eq_table.eq[MLX4_EQ_ASYNC].eqn, err);
 
-	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
-		eq_set_ci(&priv->eq_table.eq[i], 1);
+	/* arm ASYNC eq */
+	eq_set_ci(&priv->eq_table.eq[MLX4_EQ_ASYNC], 1);
 
 	return 0;
 
-err_out_async:
-	mlx4_free_eq(dev, &priv->eq_table.eq[dev->caps.num_comp_vectors]);
-
-err_out_comp:
-	i = dev->caps.num_comp_vectors - 1;
-
 err_out_unmap:
-	while (i >= 0) {
-		mlx4_free_eq(dev, &priv->eq_table.eq[i]);
-		--i;
+	while (i >= 0)
+		mlx4_free_eq(dev, &priv->eq_table.eq[i--]);
+#ifdef CONFIG_RFS_ACCEL
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		if (mlx4_priv(dev)->port[i].rmap) {
+			free_irq_cpu_rmap(mlx4_priv(dev)->port[i].rmap);
+			mlx4_priv(dev)->port[i].rmap = NULL;
+		}
 	}
+#endif
+	mlx4_free_irqs(dev);
+
+err_out_clr_int:
 	if (!mlx4_is_slave(dev))
 		mlx4_unmap_clr_int(dev);
-	mlx4_free_irqs(dev);
 
 err_out_bitmap:
 	mlx4_unmap_uar(dev);
@@ -1300,11 +1320,19 @@ void mlx4_cleanup_eq_table(struct mlx4_dev *dev)
 	int i;
 
 	mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 1,
-		    priv->eq_table.eq[dev->caps.num_comp_vectors].eqn);
+		    priv->eq_table.eq[MLX4_EQ_ASYNC].eqn);
 
+#ifdef CONFIG_RFS_ACCEL
+	for (i = 1; i <= dev->caps.num_ports; i++) {
+		if (mlx4_priv(dev)->port[i].rmap) {
+			free_irq_cpu_rmap(mlx4_priv(dev)->port[i].rmap);
+			mlx4_priv(dev)->port[i].rmap = NULL;
+		}
+	}
+#endif
 	mlx4_free_irqs(dev);
 
-	for (i = 0; i < dev->caps.num_comp_vectors + dev->caps.comp_pool + 1; ++i)
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; ++i)
 		mlx4_free_eq(dev, &priv->eq_table.eq[i]);
 
 	if (!mlx4_is_slave(dev))
@@ -1336,6 +1364,10 @@ int mlx4_test_interrupts(struct mlx4_dev *dev)
 	 * and performing a NOP command
 	 */
 	for(i = 0; !err && (i < dev->caps.num_comp_vectors); ++i) {
+		/* Make sure request_irq was called */
+		if (!priv->eq_table.eq[i].have_irq)
+			continue;
+
 		/* Temporary use polling for command completions */
 		mlx4_cmd_use_polling(dev);
 
@@ -1355,87 +1387,169 @@ int mlx4_test_interrupts(struct mlx4_dev *dev)
 
 	/* Return to default */
 	mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0,
-		    priv->eq_table.eq[dev->caps.num_comp_vectors].eqn);
+		    priv->eq_table.eq[MLX4_EQ_ASYNC].eqn);
 	return err;
 }
 EXPORT_SYMBOL(mlx4_test_interrupts);
 
-int mlx4_assign_eq(struct mlx4_dev *dev, char *name, struct cpu_rmap *rmap,
-		   int *vector)
+bool mlx4_is_eq_vector_valid(struct mlx4_dev *dev, u8 port, int vector)
 {
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	vector = MLX4_CQ_TO_EQ_VECTOR(vector);
+	if (vector < 0 || (vector >= dev->caps.num_comp_vectors + 1) ||
+	    (vector == MLX4_EQ_ASYNC))
+		return false;
+
+	return test_bit(port - 1, priv->eq_table.eq[vector].actv_ports.ports);
+}
+EXPORT_SYMBOL(mlx4_is_eq_vector_valid);
 
+u32 mlx4_get_eqs_per_port(struct mlx4_dev *dev, u8 port)
+{
 	struct mlx4_priv *priv = mlx4_priv(dev);
-	int vec = 0, err = 0, i;
+	unsigned int i;
+	unsigned int sum = 0;
+
+	for (i = 0; i < dev->caps.num_comp_vectors + 1; i++)
+		sum += !!test_bit(port - 1,
+				  priv->eq_table.eq[i].actv_ports.ports);
+
+	return sum;
+}
+EXPORT_SYMBOL(mlx4_get_eqs_per_port);
+
+int mlx4_is_eq_shared(struct mlx4_dev *dev, int vector)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	vector = MLX4_CQ_TO_EQ_VECTOR(vector);
+	if (vector <= 0 || (vector >= dev->caps.num_comp_vectors + 1))
+		return -EINVAL;
+
+	return !!(bitmap_weight(priv->eq_table.eq[vector].actv_ports.ports,
+				dev->caps.num_ports) > 1);
+}
+EXPORT_SYMBOL(mlx4_is_eq_shared);
+
+struct cpu_rmap *mlx4_get_cpu_rmap(struct mlx4_dev *dev, int port)
+{
+	return mlx4_priv(dev)->port[port].rmap;
+}
+EXPORT_SYMBOL(mlx4_get_cpu_rmap);
+
+int mlx4_assign_eq(struct mlx4_dev *dev, u8 port, int *vector)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int err = 0, i = 0;
+	u32 min_ref_count_val = (u32)-1;
+	int requested_vector = MLX4_CQ_TO_EQ_VECTOR(*vector);
+	int *prequested_vector = NULL;
+
 
 	mutex_lock(&priv->msix_ctl.pool_lock);
-	for (i = 0; !vec && i < dev->caps.comp_pool; i++) {
-		if (~priv->msix_ctl.pool_bm & 1ULL << i) {
-			priv->msix_ctl.pool_bm |= 1ULL << i;
-			vec = dev->caps.num_comp_vectors + 1 + i;
-			snprintf(priv->eq_table.irq_names +
-					vec * MLX4_IRQNAME_SIZE,
-					MLX4_IRQNAME_SIZE, "%s", name);
-#ifdef CONFIG_RFS_ACCEL
-			if (rmap) {
-				err = irq_cpu_rmap_add(rmap,
-						       priv->eq_table.eq[vec].irq);
-				if (err)
-					mlx4_warn(dev, "Failed adding irq rmap\n");
+	if (requested_vector < (dev->caps.num_comp_vectors + 1) &&
+	    (requested_vector >= 0) &&
+	    (requested_vector != MLX4_EQ_ASYNC)) {
+		if (test_bit(port - 1,
+			     priv->eq_table.eq[requested_vector].actv_ports.ports)) {
+			prequested_vector = &requested_vector;
+		} else {
+			struct mlx4_eq *eq;
+
+			for (i = 1; i < port;
+			     requested_vector += mlx4_get_eqs_per_port(dev, i++))
+				;
+
+			eq = &priv->eq_table.eq[requested_vector];
+			if (requested_vector < dev->caps.num_comp_vectors + 1 &&
+			    test_bit(port - 1, eq->actv_ports.ports)) {
+				prequested_vector = &requested_vector;
 			}
-#endif
-			err = request_irq(priv->eq_table.eq[vec].irq,
-					  mlx4_msi_x_interrupt, 0,
-					  &priv->eq_table.irq_names[vec<<5],
-					  priv->eq_table.eq + vec);
-			if (err) {
-				/*zero out bit by fliping it*/
-				priv->msix_ctl.pool_bm ^= 1 << i;
-				vec = 0;
-				continue;
-				/*we dont want to break here*/
+		}
+	}
+
+	if  (!prequested_vector) {
+		requested_vector = -1;
+		for (i = 0; min_ref_count_val && i < dev->caps.num_comp_vectors + 1;
+		     i++) {
+			struct mlx4_eq *eq = &priv->eq_table.eq[i];
+
+			if (min_ref_count_val > eq->ref_count &&
+			    test_bit(port - 1, eq->actv_ports.ports)) {
+				min_ref_count_val = eq->ref_count;
+				requested_vector = i;
 			}
+		}
+
+		if (requested_vector < 0) {
+			err = -ENOSPC;
+			goto err_unlock;
+		}
+
+		prequested_vector = &requested_vector;
+	}
+
+	if (!test_bit(*prequested_vector, priv->msix_ctl.pool_bm) &&
+	    dev->flags & MLX4_FLAG_MSI_X) {
+		set_bit(*prequested_vector, priv->msix_ctl.pool_bm);
+		snprintf(priv->eq_table.irq_names +
+			 *prequested_vector * MLX4_IRQNAME_SIZE,
+			 MLX4_IRQNAME_SIZE, "mlx4-%d@%s",
+			 *prequested_vector, dev_name(&dev->persist->pdev->dev));
 
-			eq_set_ci(&priv->eq_table.eq[vec], 1);
+		err = request_irq(priv->eq_table.eq[*prequested_vector].irq,
+				  mlx4_msi_x_interrupt, 0,
+				  &priv->eq_table.irq_names[*prequested_vector << 5],
+				  priv->eq_table.eq + *prequested_vector);
+
+		if (err) {
+			clear_bit(*prequested_vector, priv->msix_ctl.pool_bm);
+			*prequested_vector = -1;
+		} else {
+#if defined(CONFIG_SMP)
+			mlx4_set_eq_affinity_hint(priv, *prequested_vector);
+#endif
+			eq_set_ci(&priv->eq_table.eq[*prequested_vector], 1);
+			priv->eq_table.eq[*prequested_vector].have_irq = 1;
 		}
 	}
+
+	if (!err && *prequested_vector >= 0)
+		priv->eq_table.eq[*prequested_vector].ref_count++;
+
+err_unlock:
 	mutex_unlock(&priv->msix_ctl.pool_lock);
 
-	if (vec) {
-		*vector = vec;
-	} else {
+	if (!err && *prequested_vector >= 0)
+		*vector = MLX4_EQ_TO_CQ_VECTOR(*prequested_vector);
+	else
 		*vector = 0;
-		err = (i == dev->caps.comp_pool) ? -ENOSPC : err;
-	}
+
 	return err;
 }
 EXPORT_SYMBOL(mlx4_assign_eq);
 
-int mlx4_eq_get_irq(struct mlx4_dev *dev, int vec)
+int mlx4_eq_get_irq(struct mlx4_dev *dev, int cq_vec)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
 
-	return priv->eq_table.eq[vec].irq;
+	return priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(cq_vec)].irq;
 }
 EXPORT_SYMBOL(mlx4_eq_get_irq);
 
 void mlx4_release_eq(struct mlx4_dev *dev, int vec)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
-	/*bm index*/
-	int i = vec - dev->caps.num_comp_vectors - 1;
-
-	if (likely(i >= 0)) {
-		/*sanity check , making sure were not trying to free irq's
-		  Belonging to a legacy EQ*/
-		mutex_lock(&priv->msix_ctl.pool_lock);
-		if (priv->msix_ctl.pool_bm & 1ULL << i) {
-			free_irq(priv->eq_table.eq[vec].irq,
-				 &priv->eq_table.eq[vec]);
-			priv->msix_ctl.pool_bm &= ~(1ULL << i);
-		}
-		mutex_unlock(&priv->msix_ctl.pool_lock);
-	}
+	int eq_vec = MLX4_CQ_TO_EQ_VECTOR(vec);
 
+	mutex_lock(&priv->msix_ctl.pool_lock);
+	priv->eq_table.eq[eq_vec].ref_count--;
+
+	/* once we allocated EQ, we don't release it because it might be binded
+	 * to cpu_rmap.
+	 */
+	mutex_unlock(&priv->msix_ctl.pool_lock);
 }
 EXPORT_SYMBOL(mlx4_release_eq);
 
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/fw.c b/kernel/drivers/net/ethernet/mellanox/mlx4/fw.c
index e30bf57ad..90db94e83 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/fw.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/fw.c
@@ -154,6 +154,9 @@ static void dump_dev_cap_flags2(struct mlx4_dev *dev, u64 flags)
 		[26] = "Port ETS Scheduler support",
 		[27] = "Port beacon support",
 		[28] = "RX-ALL support",
+		[29] = "802.1ad offload support",
+		[31] = "Modifying loopback source checks using UPDATE_QP support",
+		[32] = "Loopback source checks support",
 	};
 	int i;
 
@@ -307,6 +310,7 @@ int mlx4_QUERY_FUNC_CAP_wrapper(struct mlx4_dev *dev, int slave,
 
 #define QUERY_FUNC_CAP_FLAGS0_FORCE_PHY_WQE_GID 0x80
 #define QUERY_FUNC_CAP_SUPPORTS_NON_POWER_OF_2_NUM_EQS (1 << 31)
+#define QUERY_FUNC_CAP_PHV_BIT			0x40
 
 	if (vhcr->op_modifier == 1) {
 		struct mlx4_active_ports actv_ports =
@@ -351,6 +355,12 @@ int mlx4_QUERY_FUNC_CAP_wrapper(struct mlx4_dev *dev, int slave,
 		MLX4_PUT(outbox->buf, dev->caps.phys_port_id[vhcr->in_modifier],
 			 QUERY_FUNC_CAP_PHYS_PORT_ID);
 
+		if (dev->caps.phv_bit[port]) {
+			field = QUERY_FUNC_CAP_PHV_BIT;
+			MLX4_PUT(outbox->buf, field,
+				 QUERY_FUNC_CAP_FLAGS0_OFFSET);
+		}
+
 	} else if (vhcr->op_modifier == 0) {
 		struct mlx4_active_ports actv_ports =
 			mlx4_get_active_ports(dev, slave);
@@ -600,6 +610,9 @@ int mlx4_QUERY_FUNC_CAP(struct mlx4_dev *dev, u8 gen_or_port,
 		MLX4_GET(func_cap->phys_port_id, outbox,
 			 QUERY_FUNC_CAP_PHYS_PORT_ID);
 
+	MLX4_GET(field, outbox, QUERY_FUNC_CAP_FLAGS0_OFFSET);
+	func_cap->flags |= (field & QUERY_FUNC_CAP_PHV_BIT);
+
 	/* All other resources are allocated by the master, but we still report
 	 * 'num' and 'reserved' capabilities as follows:
 	 * - num remains the maximum resource index
@@ -700,6 +713,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 #define QUERY_DEV_CAP_D_MPT_ENTRY_SZ_OFFSET	0x92
 #define QUERY_DEV_CAP_BMME_FLAGS_OFFSET		0x94
 #define QUERY_DEV_CAP_CONFIG_DEV_OFFSET		0x94
+#define QUERY_DEV_CAP_PHV_EN_OFFSET		0x96
 #define QUERY_DEV_CAP_RSVD_LKEY_OFFSET		0x98
 #define QUERY_DEV_CAP_MAX_ICM_SZ_OFFSET		0xa0
 #define QUERY_DEV_CAP_ETH_BACKPL_OFFSET		0x9c
@@ -898,6 +912,12 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_CONFIG_DEV;
 	if (field & (1 << 2))
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_IGNORE_FCS;
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_PHV_EN_OFFSET);
+	if (field & 0x80)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_PHV_EN;
+	if (field & 0x40)
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN;
+
 	MLX4_GET(dev_cap->reserved_lkey, outbox,
 		 QUERY_DEV_CAP_RSVD_LKEY_OFFSET);
 	MLX4_GET(field32, outbox, QUERY_DEV_CAP_ETH_BACKPL_OFFSET);
@@ -946,6 +966,10 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	MLX4_GET(field32, outbox, QUERY_DEV_CAP_EXT_2_FLAGS_OFFSET);
 	if (field32 & (1 << 16))
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_UPDATE_QP;
+	if (field32 & (1 << 18))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB;
+	if (field32 & (1 << 19))
+		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_LB_SRC_CHK;
 	if (field32 & (1 << 26))
 		dev_cap->flags2 |= MLX4_DEV_CAP_FLAG2_VLAN_CONTROL;
 	if (field32 & (1 << 20))
@@ -1992,6 +2016,10 @@ int mlx4_QUERY_HCA(struct mlx4_dev *dev,
 	MLX4_GET(param->uar_page_sz, outbox, INIT_HCA_UAR_PAGE_SZ_OFFSET);
 	MLX4_GET(param->log_uar_sz, outbox, INIT_HCA_LOG_UAR_SZ_OFFSET);
 
+	/* phv_check enable */
+	MLX4_GET(byte_field, outbox, INIT_HCA_CACHELINE_SZ_OFFSET);
+	if (byte_field & 0x2)
+		param->phv_check_en = 1;
 out:
 	mlx4_free_cmd_mailbox(dev, mailbox);
 
@@ -2758,3 +2786,79 @@ int mlx4_ACCESS_REG_wrapper(struct mlx4_dev *dev, int slave,
 			    0, MLX4_CMD_ACCESS_REG, MLX4_CMD_TIME_CLASS_C,
 			    MLX4_CMD_NATIVE);
 }
+
+static int mlx4_SET_PORT_phv_bit(struct mlx4_dev *dev, u8 port, u8 phv_bit)
+{
+#define SET_PORT_GEN_PHV_VALID	0x10
+#define SET_PORT_GEN_PHV_EN	0x80
+
+	struct mlx4_cmd_mailbox *mailbox;
+	struct mlx4_set_port_general_context *context;
+	u32 in_mod;
+	int err;
+
+	mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	context = mailbox->buf;
+
+	context->v_ignore_fcs |=  SET_PORT_GEN_PHV_VALID;
+	if (phv_bit)
+		context->phv_en |=  SET_PORT_GEN_PHV_EN;
+
+	in_mod = MLX4_SET_PORT_GENERAL << 8 | port;
+	err = mlx4_cmd(dev, mailbox->dma, in_mod, MLX4_SET_PORT_ETH_OPCODE,
+		       MLX4_CMD_SET_PORT, MLX4_CMD_TIME_CLASS_B,
+		       MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, mailbox);
+	return err;
+}
+
+int get_phv_bit(struct mlx4_dev *dev, u8 port, int *phv)
+{
+	int err;
+	struct mlx4_func_cap func_cap;
+
+	memset(&func_cap, 0, sizeof(func_cap));
+	err = mlx4_QUERY_FUNC_CAP(dev, port, &func_cap);
+	if (!err)
+		*phv = func_cap.flags & QUERY_FUNC_CAP_PHV_BIT;
+	return err;
+}
+EXPORT_SYMBOL(get_phv_bit);
+
+int set_phv_bit(struct mlx4_dev *dev, u8 port, int new_val)
+{
+	int ret;
+
+	if (mlx4_is_slave(dev))
+		return -EPERM;
+
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PHV_EN &&
+	    !(dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_SKIP_OUTER_VLAN)) {
+		ret = mlx4_SET_PORT_phv_bit(dev, port, new_val);
+		if (!ret)
+			dev->caps.phv_bit[port] = new_val;
+		return ret;
+	}
+
+	return -EOPNOTSUPP;
+}
+EXPORT_SYMBOL(set_phv_bit);
+
+void mlx4_replace_zero_macs(struct mlx4_dev *dev)
+{
+	int i;
+	u8 mac_addr[ETH_ALEN];
+
+	dev->port_random_macs = 0;
+	for (i = 1; i <= dev->caps.num_ports; ++i)
+		if (!dev->caps.def_mac[i] &&
+		    dev->caps.port_type[i] == MLX4_PORT_TYPE_ETH) {
+			eth_random_addr(mac_addr);
+			dev->port_random_macs |= 1 << i;
+			dev->caps.def_mac[i] = mlx4_mac_to_u64(mac_addr);
+		}
+}
+EXPORT_SYMBOL_GPL(mlx4_replace_zero_macs);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/fw.h b/kernel/drivers/net/ethernet/mellanox/mlx4/fw.h
index 07cb7c246..08de5555c 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/fw.h
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/fw.h
@@ -204,6 +204,7 @@ struct mlx4_init_hca_param {
 	u16 cqe_size; /* For use only when CQE stride feature enabled */
 	u16 eqe_size; /* For use only when EQE stride feature enabled */
 	u8 rss_ip_frags;
+	u8 phv_check_en; /* for QUERY_HCA */
 };
 
 struct mlx4_init_ib_param {
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/intf.c b/kernel/drivers/net/ethernet/mellanox/mlx4/intf.c
index 0d80aed59..0472941af 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/intf.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/intf.c
@@ -63,8 +63,11 @@ static void mlx4_add_device(struct mlx4_interface *intf, struct mlx4_priv *priv)
 		spin_lock_irq(&priv->ctx_lock);
 		list_add_tail(&dev_ctx->list, &priv->ctx_list);
 		spin_unlock_irq(&priv->ctx_lock);
+		if (intf->activate)
+			intf->activate(&priv->dev, dev_ctx->context);
 	} else
 		kfree(dev_ctx);
+
 }
 
 static void mlx4_remove_device(struct mlx4_interface *intf, struct mlx4_priv *priv)
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/main.c b/kernel/drivers/net/ethernet/mellanox/mlx4/main.c
index ced5ecab5..31c491e02 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -405,6 +405,21 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	dev->caps.max_gso_sz	     = dev_cap->max_gso_sz;
 	dev->caps.max_rss_tbl_sz     = dev_cap->max_rss_tbl_sz;
 
+	if (dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_PHV_EN) {
+		struct mlx4_init_hca_param hca_param;
+
+		memset(&hca_param, 0, sizeof(hca_param));
+		err = mlx4_QUERY_HCA(dev, &hca_param);
+		/* Turn off PHV_EN flag in case phv_check_en is set.
+		 * phv_check_en is a HW check that parse the packet and verify
+		 * phv bit was reported correctly in the wqe. To allow QinQ
+		 * PHV_EN flag should be set and phv_check_en must be cleared
+		 * otherwise QinQ packets will be drop by the HW.
+		 */
+		if (err || hca_param.phv_check_en)
+			dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_PHV_EN;
+	}
+
 	/* Sense port always allowed on supported devices for ConnectX-1 and -2 */
 	if (mlx4_priv(dev)->pci_dev_data & MLX4_PCI_DEV_FORCE_SENSE_PORT)
 		dev->caps.flags |= MLX4_DEV_CAP_FLAG_SENSE_SUPPORT;
@@ -479,7 +494,15 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 		}
 	}
 
-	dev->caps.max_counters = 1 << ilog2(dev_cap->max_counters);
+	if (mlx4_is_master(dev) && (dev->caps.num_ports == 2) &&
+	    (port_type_array[0] == MLX4_PORT_TYPE_IB) &&
+	    (port_type_array[1] == MLX4_PORT_TYPE_ETH)) {
+		mlx4_warn(dev,
+			  "Granular QoS per VF not supported with IB/Eth configuration\n");
+		dev->caps.flags2 &= ~MLX4_DEV_CAP_FLAG2_QOS_VPP;
+	}
+
+	dev->caps.max_counters = dev_cap->max_counters;
 
 	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_FW] = dev_cap->reserved_qps;
 	dev->caps.reserved_qps_cnt[MLX4_QP_REGION_ETH_ADDR] =
@@ -840,6 +863,8 @@ static int mlx4_slave_cap(struct mlx4_dev *dev)
 		return -ENODEV;
 	}
 
+	mlx4_replace_zero_macs(dev);
+
 	dev->caps.qp0_qkey = kcalloc(dev->caps.num_ports, sizeof(u32), GFP_KERNEL);
 	dev->caps.qp0_tunnel = kcalloc(dev->caps.num_ports, sizeof (u32), GFP_KERNEL);
 	dev->caps.qp0_proxy = kcalloc(dev->caps.num_ports, sizeof (u32), GFP_KERNEL);
@@ -867,9 +892,10 @@ static int mlx4_slave_cap(struct mlx4_dev *dev)
 		dev->caps.qp1_proxy[i - 1] = func_cap.qp1_proxy_qpn;
 		dev->caps.port_mask[i] = dev->caps.port_type[i];
 		dev->caps.phys_port_id[i] = func_cap.phys_port_id;
-		if (mlx4_get_slave_pkey_gid_tbl_len(dev, i,
-						    &dev->caps.gid_table_len[i],
-						    &dev->caps.pkey_table_len[i]))
+		err = mlx4_get_slave_pkey_gid_tbl_len(dev, i,
+						      &dev->caps.gid_table_len[i],
+						      &dev->caps.pkey_table_len[i]);
+		if (err)
 			goto err_mem;
 	}
 
@@ -881,6 +907,7 @@ static int mlx4_slave_cap(struct mlx4_dev *dev)
 			 dev->caps.uar_page_size * dev->caps.num_uars,
 			 (unsigned long long)
 			 pci_resource_len(dev->persist->pdev, 2));
+		err = -ENOMEM;
 		goto err_mem;
 	}
 
@@ -1674,6 +1701,25 @@ static int map_internal_clock(struct mlx4_dev *dev)
 	return 0;
 }
 
+int mlx4_get_internal_clock_params(struct mlx4_dev *dev,
+				   struct mlx4_clock_params *params)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	if (mlx4_is_slave(dev))
+		return -ENOTSUPP;
+
+	if (!params)
+		return -EINVAL;
+
+	params->bar = priv->fw.clock_bar;
+	params->offset = priv->fw.clock_offset;
+	params->size = MLX4_CLOCK_SIZE;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_get_internal_clock_params);
+
 static void unmap_internal_clock(struct mlx4_dev *dev)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
@@ -2193,20 +2239,78 @@ err_free_icm:
 static int mlx4_init_counters_table(struct mlx4_dev *dev)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
-	int nent;
+	int nent_pow2;
 
 	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
 		return -ENOENT;
 
-	nent = dev->caps.max_counters;
-	return mlx4_bitmap_init(&priv->counters_bitmap, nent, nent - 1, 0, 0);
+	if (!dev->caps.max_counters)
+		return -ENOSPC;
+
+	nent_pow2 = roundup_pow_of_two(dev->caps.max_counters);
+	/* reserve last counter index for sink counter */
+	return mlx4_bitmap_init(&priv->counters_bitmap, nent_pow2,
+				nent_pow2 - 1, 0,
+				nent_pow2 - dev->caps.max_counters + 1);
 }
 
 static void mlx4_cleanup_counters_table(struct mlx4_dev *dev)
 {
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
+		return;
+
+	if (!dev->caps.max_counters)
+		return;
+
 	mlx4_bitmap_cleanup(&mlx4_priv(dev)->counters_bitmap);
 }
 
+static void mlx4_cleanup_default_counters(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port;
+
+	for (port = 0; port < dev->caps.num_ports; port++)
+		if (priv->def_counter[port] != -1)
+			mlx4_counter_free(dev,  priv->def_counter[port]);
+}
+
+static int mlx4_allocate_default_counters(struct mlx4_dev *dev)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	int port, err = 0;
+	u32 idx;
+
+	for (port = 0; port < dev->caps.num_ports; port++)
+		priv->def_counter[port] = -1;
+
+	for (port = 0; port < dev->caps.num_ports; port++) {
+		err = mlx4_counter_alloc(dev, &idx);
+
+		if (!err || err == -ENOSPC) {
+			priv->def_counter[port] = idx;
+		} else if (err == -ENOENT) {
+			err = 0;
+			continue;
+		} else if (mlx4_is_slave(dev) && err == -EINVAL) {
+			priv->def_counter[port] = MLX4_SINK_COUNTER_INDEX(dev);
+			mlx4_warn(dev, "can't allocate counter from old PF driver, using index %d\n",
+				  MLX4_SINK_COUNTER_INDEX(dev));
+			err = 0;
+		} else {
+			mlx4_err(dev, "%s: failed to allocate default counter port %d err %d\n",
+				 __func__, port + 1, err);
+			mlx4_cleanup_default_counters(dev);
+			return err;
+		}
+
+		mlx4_dbg(dev, "%s: default counter index %d for port %d\n",
+			 __func__, priv->def_counter[port], port + 1);
+	}
+
+	return err;
+}
+
 int __mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
@@ -2215,8 +2319,10 @@ int __mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx)
 		return -ENOENT;
 
 	*idx = mlx4_bitmap_alloc(&priv->counters_bitmap);
-	if (*idx == -1)
-		return -ENOMEM;
+	if (*idx == -1) {
+		*idx = MLX4_SINK_COUNTER_INDEX(dev);
+		return -ENOSPC;
+	}
 
 	return 0;
 }
@@ -2239,8 +2345,35 @@ int mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx)
 }
 EXPORT_SYMBOL_GPL(mlx4_counter_alloc);
 
+static int __mlx4_clear_if_stat(struct mlx4_dev *dev,
+				u8 counter_index)
+{
+	struct mlx4_cmd_mailbox *if_stat_mailbox;
+	int err;
+	u32 if_stat_in_mod = (counter_index & 0xff) | MLX4_QUERY_IF_STAT_RESET;
+
+	if_stat_mailbox = mlx4_alloc_cmd_mailbox(dev);
+	if (IS_ERR(if_stat_mailbox))
+		return PTR_ERR(if_stat_mailbox);
+
+	err = mlx4_cmd_box(dev, 0, if_stat_mailbox->dma, if_stat_in_mod, 0,
+			   MLX4_CMD_QUERY_IF_STAT, MLX4_CMD_TIME_CLASS_C,
+			   MLX4_CMD_NATIVE);
+
+	mlx4_free_cmd_mailbox(dev, if_stat_mailbox);
+	return err;
+}
+
 void __mlx4_counter_free(struct mlx4_dev *dev, u32 idx)
 {
+	if (!(dev->caps.flags & MLX4_DEV_CAP_FLAG_COUNTERS))
+		return;
+
+	if (idx == MLX4_SINK_COUNTER_INDEX(dev))
+		return;
+
+	__mlx4_clear_if_stat(dev, idx);
+
 	mlx4_bitmap_free(&mlx4_priv(dev)->counters_bitmap, idx, MLX4_USE_RR);
 	return;
 }
@@ -2260,6 +2393,14 @@ void mlx4_counter_free(struct mlx4_dev *dev, u32 idx)
 }
 EXPORT_SYMBOL_GPL(mlx4_counter_free);
 
+int mlx4_get_default_counter_index(struct mlx4_dev *dev, int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+
+	return priv->def_counter[port - 1];
+}
+EXPORT_SYMBOL_GPL(mlx4_get_default_counter_index);
+
 void mlx4_set_admin_guid(struct mlx4_dev *dev, __be64 guid, int entry, int port)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
@@ -2364,11 +2505,11 @@ static int mlx4_setup_hca(struct mlx4_dev *dev)
 	if (err) {
 		if (dev->flags & MLX4_FLAG_MSI_X) {
 			mlx4_warn(dev, "NOP command failed to generate MSI-X interrupt IRQ %d)\n",
-				  priv->eq_table.eq[dev->caps.num_comp_vectors].irq);
+				  priv->eq_table.eq[MLX4_EQ_ASYNC].irq);
 			mlx4_warn(dev, "Trying again without MSI-X\n");
 		} else {
 			mlx4_err(dev, "NOP command failed to generate interrupt (IRQ %d), aborting\n",
-				 priv->eq_table.eq[dev->caps.num_comp_vectors].irq);
+				 priv->eq_table.eq[MLX4_EQ_ASYNC].irq);
 			mlx4_err(dev, "BIOS or ACPI interrupt routing problem?\n");
 		}
 
@@ -2395,10 +2536,18 @@ static int mlx4_setup_hca(struct mlx4_dev *dev)
 		goto err_srq_table_free;
 	}
 
-	err = mlx4_init_counters_table(dev);
-	if (err && err != -ENOENT) {
-		mlx4_err(dev, "Failed to initialize counters table, aborting\n");
-		goto err_qp_table_free;
+	if (!mlx4_is_slave(dev)) {
+		err = mlx4_init_counters_table(dev);
+		if (err && err != -ENOENT) {
+			mlx4_err(dev, "Failed to initialize counters table, aborting\n");
+			goto err_qp_table_free;
+		}
+	}
+
+	err = mlx4_allocate_default_counters(dev);
+	if (err) {
+		mlx4_err(dev, "Failed to allocate default counters, aborting\n");
+		goto err_counters_table_free;
 	}
 
 	if (!mlx4_is_slave(dev)) {
@@ -2432,15 +2581,19 @@ static int mlx4_setup_hca(struct mlx4_dev *dev)
 			if (err) {
 				mlx4_err(dev, "Failed to set port %d, aborting\n",
 					 port);
-				goto err_counters_table_free;
+				goto err_default_countes_free;
 			}
 		}
 	}
 
 	return 0;
 
+err_default_countes_free:
+	mlx4_cleanup_default_counters(dev);
+
 err_counters_table_free:
-	mlx4_cleanup_counters_table(dev);
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_counters_table(dev);
 
 err_qp_table_free:
 	mlx4_cleanup_qp_table(dev);
@@ -2481,17 +2634,50 @@ err_uar_table_free:
 	return err;
 }
 
+static int mlx4_init_affinity_hint(struct mlx4_dev *dev, int port, int eqn)
+{
+	int requested_cpu = 0;
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_eq *eq;
+	int off = 0;
+	int i;
+
+	if (eqn > dev->caps.num_comp_vectors)
+		return -EINVAL;
+
+	for (i = 1; i < port; i++)
+		off += mlx4_get_eqs_per_port(dev, i);
+
+	requested_cpu = eqn - off - !!(eqn > MLX4_EQ_ASYNC);
+
+	/* Meaning EQs are shared, and this call comes from the second port */
+	if (requested_cpu < 0)
+		return 0;
+
+	eq = &priv->eq_table.eq[eqn];
+
+	if (!zalloc_cpumask_var(&eq->affinity_mask, GFP_KERNEL))
+		return -ENOMEM;
+
+	cpumask_set_cpu(requested_cpu, eq->affinity_mask);
+
+	return 0;
+}
+
 static void mlx4_enable_msi_x(struct mlx4_dev *dev)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
 	struct msix_entry *entries;
 	int i;
+	int port = 0;
 
 	if (msi_x) {
-		int nreq = dev->caps.num_ports * num_online_cpus() + MSIX_LEGACY_SZ;
+		int nreq = dev->caps.num_ports * num_online_cpus() + 1;
 
 		nreq = min_t(int, dev->caps.num_eqs - dev->caps.reserved_eqs,
 			     nreq);
+		if (nreq > MAX_MSIX)
+			nreq = MAX_MSIX;
 
 		entries = kcalloc(nreq, sizeof *entries, GFP_KERNEL);
 		if (!entries)
@@ -2503,20 +2689,55 @@ static void mlx4_enable_msi_x(struct mlx4_dev *dev)
 		nreq = pci_enable_msix_range(dev->persist->pdev, entries, 2,
 					     nreq);
 
-		if (nreq < 0) {
+		if (nreq < 0 || nreq < MLX4_EQ_ASYNC) {
 			kfree(entries);
 			goto no_msi;
-		} else if (nreq < MSIX_LEGACY_SZ +
-			   dev->caps.num_ports * MIN_MSIX_P_PORT) {
-			/*Working in legacy mode , all EQ's shared*/
-			dev->caps.comp_pool           = 0;
-			dev->caps.num_comp_vectors = nreq - 1;
-		} else {
-			dev->caps.comp_pool           = nreq - MSIX_LEGACY_SZ;
-			dev->caps.num_comp_vectors = MSIX_LEGACY_SZ - 1;
 		}
-		for (i = 0; i < nreq; ++i)
-			priv->eq_table.eq[i].irq = entries[i].vector;
+		/* 1 is reserved for events (asyncrounous EQ) */
+		dev->caps.num_comp_vectors = nreq - 1;
+
+		priv->eq_table.eq[MLX4_EQ_ASYNC].irq = entries[0].vector;
+		bitmap_zero(priv->eq_table.eq[MLX4_EQ_ASYNC].actv_ports.ports,
+			    dev->caps.num_ports);
+
+		for (i = 0; i < dev->caps.num_comp_vectors + 1; i++) {
+			if (i == MLX4_EQ_ASYNC)
+				continue;
+
+			priv->eq_table.eq[i].irq =
+				entries[i + 1 - !!(i > MLX4_EQ_ASYNC)].vector;
+
+			if (MLX4_IS_LEGACY_EQ_MODE(dev->caps)) {
+				bitmap_fill(priv->eq_table.eq[i].actv_ports.ports,
+					    dev->caps.num_ports);
+				/* We don't set affinity hint when there
+				 * aren't enough EQs
+				 */
+			} else {
+				set_bit(port,
+					priv->eq_table.eq[i].actv_ports.ports);
+				if (mlx4_init_affinity_hint(dev, port + 1, i))
+					mlx4_warn(dev, "Couldn't init hint cpumask for EQ %d\n",
+						  i);
+			}
+			/* We divide the Eqs evenly between the two ports.
+			 * (dev->caps.num_comp_vectors / dev->caps.num_ports)
+			 * refers to the number of Eqs per port
+			 * (i.e eqs_per_port). Theoretically, we would like to
+			 * write something like (i + 1) % eqs_per_port == 0.
+			 * However, since there's an asynchronous Eq, we have
+			 * to skip over it by comparing this condition to
+			 * !!((i + 1) > MLX4_EQ_ASYNC).
+			 */
+			if ((dev->caps.num_comp_vectors > dev->caps.num_ports) &&
+			    ((i + 1) %
+			     (dev->caps.num_comp_vectors / dev->caps.num_ports)) ==
+			    !!((i + 1) > MLX4_EQ_ASYNC))
+				/* If dev->caps.num_comp_vectors < dev->caps.num_ports,
+				 * everything is shared anyway.
+				 */
+				port++;
+		}
 
 		dev->flags |= MLX4_FLAG_MSI_X;
 
@@ -2526,10 +2747,15 @@ static void mlx4_enable_msi_x(struct mlx4_dev *dev)
 
 no_msi:
 	dev->caps.num_comp_vectors = 1;
-	dev->caps.comp_pool	   = 0;
 
-	for (i = 0; i < 2; ++i)
+	BUG_ON(MLX4_EQ_ASYNC >= 2);
+	for (i = 0; i < 2; ++i) {
 		priv->eq_table.eq[i].irq = dev->persist->pdev->irq;
+		if (i != MLX4_EQ_ASYNC) {
+			bitmap_fill(priv->eq_table.eq[i].actv_ports.ports,
+				    dev->caps.num_ports);
+		}
+	}
 }
 
 static int mlx4_init_port_info(struct mlx4_dev *dev, int port)
@@ -2594,6 +2820,10 @@ static void mlx4_cleanup_port_info(struct mlx4_port_info *info)
 	device_remove_file(&info->dev->persist->pdev->dev, &info->port_attr);
 	device_remove_file(&info->dev->persist->pdev->dev,
 			   &info->port_mtu_attr);
+#ifdef CONFIG_RFS_ACCEL
+	free_irq_cpu_rmap(info->rmap);
+	info->rmap = NULL;
+#endif
 }
 
 static int mlx4_init_steering(struct mlx4_dev *dev)
@@ -2703,6 +2933,8 @@ static u64 mlx4_enable_sriov(struct mlx4_dev *dev, struct pci_dev *pdev,
 {
 	u64 dev_flags = dev->flags;
 	int err = 0;
+	int fw_enabled_sriov_vfs = min(pci_sriov_get_totalvfs(pdev),
+					MLX4_MAX_NUM_VF);
 
 	if (reset_flow) {
 		dev->dev_vfs = kcalloc(total_vfs, sizeof(*dev->dev_vfs),
@@ -2728,6 +2960,12 @@ static u64 mlx4_enable_sriov(struct mlx4_dev *dev, struct pci_dev *pdev,
 	}
 
 	if (!(dev->flags &  MLX4_FLAG_SRIOV)) {
+		if (total_vfs > fw_enabled_sriov_vfs) {
+			mlx4_err(dev, "requested vfs (%d) > available vfs (%d). Continuing without SR_IOV\n",
+				 total_vfs, fw_enabled_sriov_vfs);
+			err = -ENOMEM;
+			goto disable_sriov;
+		}
 		mlx4_warn(dev, "Enabling SR-IOV with %d VFs\n", total_vfs);
 		err = pci_enable_sriov(pdev, total_vfs);
 	}
@@ -2749,6 +2987,7 @@ disable_sriov:
 free_mem:
 	dev->persist->num_vfs = 0;
 	kfree(dev->dev_vfs);
+        dev->dev_vfs = NULL;
 	return dev_flags & ~MLX4_FLAG_MASTER;
 }
 
@@ -2900,6 +3139,7 @@ slave_start:
 								  existing_vfs,
 								  reset_flow);
 
+				mlx4_close_fw(dev);
 				mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
 				dev->flags = dev_flags;
 				if (!SRIOV_VALID_STATE(dev->flags)) {
@@ -2988,18 +3228,6 @@ slave_start:
 	/* In master functions, the communication channel must be initialized
 	 * after obtaining its address from fw */
 	if (mlx4_is_master(dev)) {
-		int ib_ports = 0;
-
-		mlx4_foreach_port(i, dev, MLX4_PORT_TYPE_IB)
-			ib_ports++;
-
-		if (ib_ports &&
-		    (num_vfs_argc > 1 || probe_vfs_argc > 1)) {
-			mlx4_err(dev,
-				 "Invalid syntax of num_vfs/probe_vfs with IB port - single port VFs syntax is only supported when all ports are configured as ethernet\n");
-			err = -EINVAL;
-			goto err_close;
-		}
 		if (dev->caps.num_ports < 2 &&
 		    num_vfs_argc > 1) {
 			err = -EINVAL;
@@ -3036,7 +3264,7 @@ slave_start:
 	if (err)
 		goto err_master_mfunc;
 
-	priv->msix_ctl.pool_bm = 0;
+	bitmap_zero(priv->msix_ctl.pool_bm, MAX_MSIX);
 	mutex_init(&priv->msix_ctl.pool_lock);
 
 	mlx4_enable_msi_x(dev);
@@ -3058,7 +3286,6 @@ slave_start:
 	    !mlx4_is_mfunc(dev)) {
 		dev->flags &= ~MLX4_FLAG_MSI_X;
 		dev->caps.num_comp_vectors = 1;
-		dev->caps.comp_pool	   = 0;
 		pci_disable_msix(pdev);
 		err = mlx4_setup_hca(dev);
 	}
@@ -3109,7 +3336,9 @@ err_port:
 	for (--port; port >= 1; --port)
 		mlx4_cleanup_port_info(&priv->port[port]);
 
-	mlx4_cleanup_counters_table(dev);
+	mlx4_cleanup_default_counters(dev);
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_counters_table(dev);
 	mlx4_cleanup_qp_table(dev);
 	mlx4_cleanup_srq_table(dev);
 	mlx4_cleanup_cq_table(dev);
@@ -3218,20 +3447,20 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data,
 			goto err_disable_pdev;
 		}
 	}
-	if (total_vfs >= MLX4_MAX_NUM_VF) {
+	if (total_vfs > MLX4_MAX_NUM_VF) {
 		dev_err(&pdev->dev,
-			"Requested more VF's (%d) than allowed (%d)\n",
-			total_vfs, MLX4_MAX_NUM_VF - 1);
+			"Requested more VF's (%d) than allowed by hw (%d)\n",
+			total_vfs, MLX4_MAX_NUM_VF);
 		err = -EINVAL;
 		goto err_disable_pdev;
 	}
 
 	for (i = 0; i < MLX4_MAX_PORTS; i++) {
-		if (nvfs[i] + nvfs[2] >= MLX4_MAX_NUM_VF_P_PORT) {
+		if (nvfs[i] + nvfs[2] > MLX4_MAX_NUM_VF_P_PORT) {
 			dev_err(&pdev->dev,
-				"Requested more VF's (%d) for port (%d) than allowed (%d)\n",
+				"Requested more VF's (%d) for port (%d) than allowed by driver (%d)\n",
 				nvfs[i] + nvfs[2], i + 1,
-				MLX4_MAX_NUM_VF_P_PORT - 1);
+				MLX4_MAX_NUM_VF_P_PORT);
 			err = -EINVAL;
 			goto err_disable_pdev;
 		}
@@ -3407,7 +3636,9 @@ static void mlx4_unload_one(struct pci_dev *pdev)
 		mlx4_free_resource_tracker(dev,
 					   RES_TR_FREE_SLAVES_ONLY);
 
-	mlx4_cleanup_counters_table(dev);
+	mlx4_cleanup_default_counters(dev);
+	if (!mlx4_is_slave(dev))
+		mlx4_cleanup_counters_table(dev);
 	mlx4_cleanup_qp_table(dev);
 	mlx4_cleanup_srq_table(dev);
 	mlx4_cleanup_cq_table(dev);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/mcg.c b/kernel/drivers/net/ethernet/mellanox/mlx4/mcg.c
index bd9ea0d01..1d4e2e054 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/mcg.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/mcg.c
@@ -1184,10 +1184,11 @@ out:
 	if (prot == MLX4_PROT_ETH) {
 		/* manage the steering entry for promisc mode */
 		if (new_entry)
-			new_steering_entry(dev, port, steer, index, qp->qpn);
+			err = new_steering_entry(dev, port, steer,
+						 index, qp->qpn);
 		else
-			existing_steering_entry(dev, port, steer,
-						index, qp->qpn);
+			err = existing_steering_entry(dev, port, steer,
+						      index, qp->qpn);
 	}
 	if (err && link && index != -1) {
 		if (index < dev->caps.num_mgms)
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 502d3dd2c..e1cf9036a 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -65,6 +65,8 @@
 
 #define INIT_HCA_TPT_MW_ENABLE          (1 << 7)
 
+#define MLX4_QUERY_IF_STAT_RESET	BIT(31)
+
 enum {
 	MLX4_HCR_BASE		= 0x80680,
 	MLX4_HCR_SIZE		= 0x0001c,
@@ -287,6 +289,12 @@ struct mlx4_icm_table {
 #define MLX4_CQE_SIZE_MASK_STRIDE	0x3
 #define MLX4_EQE_SIZE_MASK_STRIDE	0x30
 
+#define MLX4_EQ_ASYNC			0
+#define MLX4_EQ_TO_CQ_VECTOR(vector)	((vector) - \
+					 !!((int)(vector) >= MLX4_EQ_ASYNC))
+#define MLX4_CQ_TO_EQ_VECTOR(vector)	((vector) + \
+					 !!((int)(vector) >= MLX4_EQ_ASYNC))
+
 /*
  * Must be packed because mtt_seg is 64 bits but only aligned to 32 bits.
  */
@@ -391,6 +399,9 @@ struct mlx4_eq {
 	struct mlx4_buf_list   *page_list;
 	struct mlx4_mtt		mtt;
 	struct mlx4_eq_tasklet	tasklet_ctx;
+	struct mlx4_active_ports actv_ports;
+	u32			ref_count;
+	cpumask_var_t		affinity_mask;
 };
 
 struct mlx4_slave_eqe {
@@ -776,6 +787,9 @@ struct mlx4_set_port_general_context {
 	u8 pprx;
 	u8 pfcrx;
 	u16 reserved4;
+	u32 reserved5;
+	u8 phv_en;
+	u8 reserved6[3];
 };
 
 struct mlx4_set_port_rqp_calc_context {
@@ -808,6 +822,7 @@ struct mlx4_port_info {
 	struct mlx4_vlan_table	vlan_table;
 	struct mlx4_roce_gid_table gid_table;
 	int			base_qpn;
+	struct cpu_rmap		*rmap;
 };
 
 struct mlx4_sense {
@@ -818,7 +833,7 @@ struct mlx4_sense {
 };
 
 struct mlx4_msix_ctl {
-	u64		pool_bm;
+	DECLARE_BITMAP(pool_bm, MAX_MSIX);
 	struct mutex	pool_lock;
 };
 
@@ -864,6 +879,7 @@ struct mlx4_priv {
 	struct mlx4_qp_table	qp_table;
 	struct mlx4_mcg_table	mcg_table;
 	struct mlx4_bitmap	counters_bitmap;
+	int			def_counter[MLX4_MAX_PORTS];
 
 	struct mlx4_catas_err	catas_err;
 
@@ -997,6 +1013,8 @@ int __mlx4_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
 		     int start_index, int npages, u64 *page_list);
 int __mlx4_counter_alloc(struct mlx4_dev *dev, u32 *idx);
 void __mlx4_counter_free(struct mlx4_dev *dev, u32 idx);
+int mlx4_calc_vf_counters(struct mlx4_dev *dev, int slave, int port,
+			  struct mlx4_counter *data);
 int __mlx4_xrcd_alloc(struct mlx4_dev *dev, u32 *xrcdn);
 void __mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn);
 
@@ -1360,6 +1378,8 @@ void mlx4_vf_immed_vlan_work_handler(struct work_struct *_work);
 
 void mlx4_init_quotas(struct mlx4_dev *dev);
 
+/* for VFs, replace zero MACs with randomly-generated MACs at driver start */
+void mlx4_replace_zero_macs(struct mlx4_dev *dev);
 int mlx4_get_slave_num_gids(struct mlx4_dev *dev, int slave, int port);
 /* Returns the VF index of slave */
 int mlx4_get_vf_indx(struct mlx4_dev *dev, int slave);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
index 909fcf803..c41f15102 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h
@@ -95,6 +95,7 @@
  */
 
 #define MLX4_EN_PRIV_FLAGS_BLUEFLAME 1
+#define MLX4_EN_PRIV_FLAGS_PHV	     2
 
 #define MLX4_EN_WATCHDOG_TIMEOUT	(15 * HZ)
 
@@ -339,7 +340,7 @@ struct mlx4_en_cq {
 	struct napi_struct	napi;
 	int size;
 	int buf_size;
-	unsigned vector;
+	int vector;
 	enum cq_type is_tx;
 	u16 moder_time;
 	u16 moder_cnt;
@@ -567,6 +568,7 @@ struct mlx4_en_priv {
 #endif
 	struct mlx4_en_perf_stats pstats;
 	struct mlx4_en_pkt_stats pkstats;
+	struct mlx4_en_counter_stats pf_stats;
 	struct mlx4_en_flow_stats_rx rx_priority_flowstats[MLX4_NUM_PRIORITIES];
 	struct mlx4_en_flow_stats_tx tx_priority_flowstats[MLX4_NUM_PRIORITIES];
 	struct mlx4_en_flow_stats_rx rx_flowstats;
@@ -582,6 +584,7 @@ struct mlx4_en_priv {
 	struct device *ddev;
 	struct hlist_head mac_hash[MLX4_EN_MAC_HASH_SIZE];
 	struct hwtstamp_config hwtstamp_config;
+	u32 counter_index;
 
 #ifdef CONFIG_MLX4_EN_DCB
 	struct ieee_ets ets;
@@ -795,7 +798,8 @@ void mlx4_en_fill_qp_context(struct mlx4_en_priv *priv, int size, int stride,
 void mlx4_en_sqp_event(struct mlx4_qp *qp, enum mlx4_event event);
 int mlx4_en_map_buffer(struct mlx4_buf *buf);
 void mlx4_en_unmap_buffer(struct mlx4_buf *buf);
-
+int mlx4_en_change_mcast_lb(struct mlx4_en_priv *priv, struct mlx4_qp *qp,
+			    int loopback);
 void mlx4_en_calc_rx_buf(struct net_device *dev);
 int mlx4_en_config_rss_steer(struct mlx4_en_priv *priv);
 void mlx4_en_release_rss_steer(struct mlx4_en_priv *priv);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h b/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h
index 00555832a..7fd466c0b 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/mlx4_stats.h
@@ -23,6 +23,14 @@ struct mlx4_en_pkt_stats {
 #define NUM_PKT_STATS		43
 };
 
+struct mlx4_en_counter_stats {
+	unsigned long rx_packets;
+	unsigned long rx_bytes;
+	unsigned long tx_packets;
+	unsigned long tx_bytes;
+#define NUM_PF_STATS      4
+};
+
 struct mlx4_en_port_stats {
 	unsigned long tso_packets;
 	unsigned long xmit_more;
@@ -99,7 +107,7 @@ enum {
 };
 
 #define NUM_ALL_STATS	(NUM_MAIN_STATS + NUM_PORT_STATS + NUM_PKT_STATS + \
-			 NUM_FLOW_STATS + NUM_PERF_STATS)
+			 NUM_FLOW_STATS + NUM_PERF_STATS + NUM_PF_STATS)
 
 #define MLX4_FIND_NETDEV_STAT(n) (offsetof(struct net_device_stats, n) / \
 				  sizeof(((struct net_device_stats *)0)->n))
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/mr.c b/kernel/drivers/net/ethernet/mellanox/mlx4/mr.c
index 78f51e103..93195191f 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/mr.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/mr.c
@@ -318,7 +318,7 @@ int mlx4_mr_hw_get_mpt(struct mlx4_dev *dev, struct mlx4_mr *mmr,
 				key, NULL);
 	} else {
 		mailbox = mlx4_alloc_cmd_mailbox(dev);
-		if (IS_ERR_OR_NULL(mailbox))
+		if (IS_ERR(mailbox))
 			return PTR_ERR(mailbox);
 
 		err = mlx4_cmd_box(dev, 0, mailbox->dma, key,
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/profile.c b/kernel/drivers/net/ethernet/mellanox/mlx4/profile.c
index 2bf437aaf..bae8b22ed 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/profile.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/profile.c
@@ -82,7 +82,6 @@ u64 mlx4_make_profile(struct mlx4_dev *dev,
 
 	u64 total_size = 0;
 	struct mlx4_resource *profile;
-	struct mlx4_resource tmp;
 	struct sysinfo si;
 	int i, j;
 
@@ -149,11 +148,8 @@ u64 mlx4_make_profile(struct mlx4_dev *dev,
 	 */
 	for (i = MLX4_RES_NUM; i > 0; --i)
 		for (j = 1; j < i; ++j) {
-			if (profile[j].size > profile[j - 1].size) {
-				tmp	       = profile[j];
-				profile[j]     = profile[j - 1];
-				profile[j - 1] = tmp;
-			}
+			if (profile[j].size > profile[j - 1].size)
+				swap(profile[j], profile[j - 1]);
 		}
 
 	for (i = 0; i < MLX4_RES_NUM; ++i) {
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/qp.c b/kernel/drivers/net/ethernet/mellanox/mlx4/qp.c
index b75214a80..168823dde 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/qp.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/qp.c
@@ -422,20 +422,37 @@ int mlx4_update_qp(struct mlx4_dev *dev, u32 qpn,
 	u64 qp_mask = 0;
 	int err = 0;
 
+	if (!attr || (attr & ~MLX4_UPDATE_QP_SUPPORTED_ATTRS))
+		return -EINVAL;
+
 	mailbox = mlx4_alloc_cmd_mailbox(dev);
 	if (IS_ERR(mailbox))
 		return PTR_ERR(mailbox);
 
 	cmd = (struct mlx4_update_qp_context *)mailbox->buf;
 
-	if (!attr || (attr & ~MLX4_UPDATE_QP_SUPPORTED_ATTRS))
-		return -EINVAL;
-
 	if (attr & MLX4_UPDATE_QP_SMAC) {
 		pri_addr_path_mask |= 1ULL << MLX4_UPD_QP_PATH_MASK_MAC_INDEX;
 		cmd->qp_context.pri_path.grh_mylmc = params->smac_index;
 	}
 
+	if (attr & MLX4_UPDATE_QP_ETH_SRC_CHECK_MC_LB) {
+		if (!(dev->caps.flags2
+		      & MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB)) {
+			mlx4_warn(dev,
+				  "Trying to set src check LB, but it isn't supported\n");
+			err = -ENOTSUPP;
+			goto out;
+		}
+		pri_addr_path_mask |=
+			1ULL << MLX4_UPD_QP_PATH_MASK_ETH_SRC_CHECK_MC_LB;
+		if (params->flags &
+		    MLX4_UPDATE_QP_PARAMS_FLAGS_ETH_CHECK_MC_LB) {
+			cmd->qp_context.pri_path.fl |=
+				MLX4_FL_ETH_SRC_CHECK_MC_LB;
+		}
+	}
+
 	if (attr & MLX4_UPDATE_QP_VSD) {
 		qp_mask |= 1ULL << MLX4_UPD_QP_MASK_VSD;
 		if (params->flags & MLX4_UPDATE_QP_PARAMS_FLAGS_VSD_ENABLE)
@@ -458,7 +475,7 @@ int mlx4_update_qp(struct mlx4_dev *dev, u32 qpn,
 	err = mlx4_cmd(dev, mailbox->dma, qpn & 0xffffff, 0,
 		       MLX4_CMD_UPDATE_QP, MLX4_CMD_TIME_CLASS_A,
 		       MLX4_CMD_NATIVE);
-
+out:
 	mlx4_free_cmd_mailbox(dev, mailbox);
 	return err;
 }
@@ -749,7 +766,7 @@ int mlx4_init_qp_table(struct mlx4_dev *dev)
 
 	{
 		int sort[MLX4_NUM_QP_REGION];
-		int i, j, tmp;
+		int i, j;
 		int last_base = dev->caps.num_qps;
 
 		for (i = 1; i < MLX4_NUM_QP_REGION; ++i)
@@ -758,11 +775,8 @@ int mlx4_init_qp_table(struct mlx4_dev *dev)
 		for (i = MLX4_NUM_QP_REGION; i > MLX4_QP_REGION_BOTTOM; --i) {
 			for (j = MLX4_QP_REGION_BOTTOM + 2; j < i; ++j) {
 				if (dev->caps.reserved_qps_cnt[sort[j]] >
-				    dev->caps.reserved_qps_cnt[sort[j - 1]]) {
-					tmp             = sort[j];
-					sort[j]         = sort[j - 1];
-					sort[j - 1]     = tmp;
-				}
+				    dev->caps.reserved_qps_cnt[sort[j - 1]])
+					swap(sort[j], sort[j - 1]);
 			}
 		}
 
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/kernel/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
index bafe2180c..cad6c44df 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c
@@ -46,8 +46,11 @@
 
 #include "mlx4.h"
 #include "fw.h"
+#include "mlx4_stats.h"
 
 #define MLX4_MAC_VALID		(1ull << 63)
+#define MLX4_PF_COUNTERS_PER_PORT	2
+#define MLX4_VF_COUNTERS_PER_PORT	1
 
 struct mac_res {
 	struct list_head list;
@@ -459,11 +462,21 @@ void mlx4_init_quotas(struct mlx4_dev *dev)
 	dev->quotas.mpt =
 		priv->mfunc.master.res_tracker.res_alloc[RES_MPT].quota[pf];
 }
+
+static int get_max_gauranteed_vfs_counter(struct mlx4_dev *dev)
+{
+	/* reduce the sink counter */
+	return (dev->caps.max_counters - 1 -
+		(MLX4_PF_COUNTERS_PER_PORT * MLX4_MAX_PORTS))
+		/ MLX4_MAX_PORTS;
+}
+
 int mlx4_init_resource_tracker(struct mlx4_dev *dev)
 {
 	struct mlx4_priv *priv = mlx4_priv(dev);
 	int i, j;
 	int t;
+	int max_vfs_guarantee_counter = get_max_gauranteed_vfs_counter(dev);
 
 	priv->mfunc.master.res_tracker.slave_list =
 		kzalloc(dev->num_slaves * sizeof(struct slave_list),
@@ -499,6 +512,9 @@ int mlx4_init_resource_tracker(struct mlx4_dev *dev)
 			res_alloc->allocated = kzalloc((dev->persist->
 							num_vfs + 1) *
 						       sizeof(int), GFP_KERNEL);
+		/* Reduce the sink counter */
+		if (i == RES_COUNTER)
+			res_alloc->res_free = dev->caps.max_counters - 1;
 
 		if (!res_alloc->quota || !res_alloc->guaranteed ||
 		    !res_alloc->allocated)
@@ -577,9 +593,17 @@ int mlx4_init_resource_tracker(struct mlx4_dev *dev)
 				break;
 			case RES_COUNTER:
 				res_alloc->quota[t] = dev->caps.max_counters;
-				res_alloc->guaranteed[t] = 0;
 				if (t == mlx4_master_func_num(dev))
-					res_alloc->res_free = res_alloc->quota[t];
+					res_alloc->guaranteed[t] =
+						MLX4_PF_COUNTERS_PER_PORT *
+						MLX4_MAX_PORTS;
+				else if (t <= max_vfs_guarantee_counter)
+					res_alloc->guaranteed[t] =
+						MLX4_VF_COUNTERS_PER_PORT *
+						MLX4_MAX_PORTS;
+				else
+					res_alloc->guaranteed[t] = 0;
+				res_alloc->res_free -= res_alloc->guaranteed[t];
 				break;
 			default:
 				break;
@@ -700,6 +724,9 @@ static void update_gid(struct mlx4_dev *dev, struct mlx4_cmd_mailbox *inbox,
 	}
 }
 
+static int handle_counter(struct mlx4_dev *dev, struct mlx4_qp_context *qpc,
+			  u8 slave, int port);
+
 static int update_vport_qp_param(struct mlx4_dev *dev,
 				 struct mlx4_cmd_mailbox *inbox,
 				 u8 slave, u32 qpn)
@@ -715,6 +742,10 @@ static int update_vport_qp_param(struct mlx4_dev *dev,
 	vp_oper = &priv->mfunc.master.vf_oper[slave].vport[port];
 	qp_type	= (be32_to_cpu(qpc->flags) >> 16) & 0xff;
 
+	err = handle_counter(dev, qpc, slave, port);
+	if (err)
+		goto out;
+
 	if (MLX4_VGT != vp_oper->state.default_vlan) {
 		/* the reserved QPs (special, proxy, tunnel)
 		 * do not operate over vlans
@@ -739,9 +770,12 @@ static int update_vport_qp_param(struct mlx4_dev *dev,
 			}
 		}
 
+		/* preserve IF_COUNTER flag */
+		qpc->pri_path.vlan_control &=
+			MLX4_CTRL_ETH_SRC_CHECK_IF_COUNTER;
 		if (vp_oper->state.link_state == IFLA_VF_LINK_STATE_DISABLE &&
 		    dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_UPDATE_QP) {
-			qpc->pri_path.vlan_control =
+			qpc->pri_path.vlan_control |=
 				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
 				MLX4_VLAN_CTRL_ETH_TX_BLOCK_PRIO_TAGGED |
 				MLX4_VLAN_CTRL_ETH_TX_BLOCK_UNTAGGED |
@@ -749,12 +783,12 @@ static int update_vport_qp_param(struct mlx4_dev *dev,
 				MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED |
 				MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
 		} else if (0 != vp_oper->state.default_vlan) {
-			qpc->pri_path.vlan_control =
+			qpc->pri_path.vlan_control |=
 				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
 				MLX4_VLAN_CTRL_ETH_RX_BLOCK_PRIO_TAGGED |
 				MLX4_VLAN_CTRL_ETH_RX_BLOCK_UNTAGGED;
 		} else { /* priority tagged */
-			qpc->pri_path.vlan_control =
+			qpc->pri_path.vlan_control |=
 				MLX4_VLAN_CTRL_ETH_TX_BLOCK_TAGGED |
 				MLX4_VLAN_CTRL_ETH_RX_BLOCK_TAGGED;
 		}
@@ -859,6 +893,83 @@ static void put_res(struct mlx4_dev *dev, int slave, u64 res_id,
 	spin_unlock_irq(mlx4_tlock(dev));
 }
 
+static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
+			     u64 in_param, u64 *out_param, int port);
+
+static int handle_existing_counter(struct mlx4_dev *dev, u8 slave, int port,
+				   int counter_index)
+{
+	struct res_common *r;
+	struct res_counter *counter;
+	int ret = 0;
+
+	if (counter_index == MLX4_SINK_COUNTER_INDEX(dev))
+		return ret;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	r = find_res(dev, counter_index, RES_COUNTER);
+	if (!r || r->owner != slave)
+		ret = -EINVAL;
+	counter = container_of(r, struct res_counter, com);
+	if (!counter->port)
+		counter->port = port;
+
+	spin_unlock_irq(mlx4_tlock(dev));
+	return ret;
+}
+
+static int handle_unexisting_counter(struct mlx4_dev *dev,
+				     struct mlx4_qp_context *qpc, u8 slave,
+				     int port)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *tmp;
+	struct res_counter *counter;
+	u64 counter_idx = MLX4_SINK_COUNTER_INDEX(dev);
+	int err = 0;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry(tmp,
+			    &tracker->slave_list[slave].res_list[RES_COUNTER],
+			    list) {
+		counter = container_of(tmp, struct res_counter, com);
+		if (port == counter->port) {
+			qpc->pri_path.counter_index  = counter->com.res_id;
+			spin_unlock_irq(mlx4_tlock(dev));
+			return 0;
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+
+	/* No existing counter, need to allocate a new counter */
+	err = counter_alloc_res(dev, slave, RES_OP_RESERVE, 0, 0, &counter_idx,
+				port);
+	if (err == -ENOENT) {
+		err = 0;
+	} else if (err && err != -ENOSPC) {
+		mlx4_err(dev, "%s: failed to create new counter for slave %d err %d\n",
+			 __func__, slave, err);
+	} else {
+		qpc->pri_path.counter_index = counter_idx;
+		mlx4_dbg(dev, "%s: alloc new counter for slave %d index %d\n",
+			 __func__, slave, qpc->pri_path.counter_index);
+		err = 0;
+	}
+
+	return err;
+}
+
+static int handle_counter(struct mlx4_dev *dev, struct mlx4_qp_context *qpc,
+			  u8 slave, int port)
+{
+	if (qpc->pri_path.counter_index != MLX4_SINK_COUNTER_INDEX(dev))
+		return handle_existing_counter(dev, slave, port,
+					       qpc->pri_path.counter_index);
+
+	return handle_unexisting_counter(dev, qpc, slave, port);
+}
+
 static struct res_common *alloc_qp_tr(int id)
 {
 	struct res_qp *ret;
@@ -952,7 +1063,7 @@ static struct res_common *alloc_srq_tr(int id)
 	return &ret->com;
 }
 
-static struct res_common *alloc_counter_tr(int id)
+static struct res_common *alloc_counter_tr(int id, int port)
 {
 	struct res_counter *ret;
 
@@ -962,6 +1073,7 @@ static struct res_common *alloc_counter_tr(int id)
 
 	ret->com.res_id = id;
 	ret->com.state = RES_COUNTER_ALLOCATED;
+	ret->port = port;
 
 	return &ret->com;
 }
@@ -1022,7 +1134,7 @@ static struct res_common *alloc_tr(u64 id, enum mlx4_resource type, int slave,
 		pr_err("implementation missing\n");
 		return NULL;
 	case RES_COUNTER:
-		ret = alloc_counter_tr(id);
+		ret = alloc_counter_tr(id, extra);
 		break;
 	case RES_XRCD:
 		ret = alloc_xrcdn_tr(id);
@@ -1039,6 +1151,53 @@ static struct res_common *alloc_tr(u64 id, enum mlx4_resource type, int slave,
 	return ret;
 }
 
+int mlx4_calc_vf_counters(struct mlx4_dev *dev, int slave, int port,
+			  struct mlx4_counter *data)
+{
+	struct mlx4_priv *priv = mlx4_priv(dev);
+	struct mlx4_resource_tracker *tracker = &priv->mfunc.master.res_tracker;
+	struct res_common *tmp;
+	struct res_counter *counter;
+	int *counters_arr;
+	int i = 0, err = 0;
+
+	memset(data, 0, sizeof(*data));
+
+	counters_arr = kmalloc_array(dev->caps.max_counters,
+				     sizeof(*counters_arr), GFP_KERNEL);
+	if (!counters_arr)
+		return -ENOMEM;
+
+	spin_lock_irq(mlx4_tlock(dev));
+	list_for_each_entry(tmp,
+			    &tracker->slave_list[slave].res_list[RES_COUNTER],
+			    list) {
+		counter = container_of(tmp, struct res_counter, com);
+		if (counter->port == port) {
+			counters_arr[i] = (int)tmp->res_id;
+			i++;
+		}
+	}
+	spin_unlock_irq(mlx4_tlock(dev));
+	counters_arr[i] = -1;
+
+	i = 0;
+
+	while (counters_arr[i] != -1) {
+		err = mlx4_get_counter_stats(dev, counters_arr[i], data,
+					     0);
+		if (err) {
+			memset(data, 0, sizeof(*data));
+			goto table_changed;
+		}
+		i++;
+	}
+
+table_changed:
+	kfree(counters_arr);
+	return 0;
+}
+
 static int add_res_range(struct mlx4_dev *dev, int slave, u64 base, int count,
 			 enum mlx4_resource type, int extra)
 {
@@ -1082,8 +1241,10 @@ static int add_res_range(struct mlx4_dev *dev, int slave, u64 base, int count,
 	return 0;
 
 undo:
-	for (--i; i >= base; --i)
+	for (--i; i >= 0; --i) {
 		rb_erase(&res_arr[i]->node, root);
+		list_del_init(&res_arr[i]->list);
+	}
 
 	spin_unlock_irq(mlx4_tlock(dev));
 
@@ -2001,7 +2162,7 @@ static int vlan_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
 }
 
 static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
-			     u64 in_param, u64 *out_param)
+			     u64 in_param, u64 *out_param, int port)
 {
 	u32 index;
 	int err;
@@ -2019,7 +2180,7 @@ static int counter_alloc_res(struct mlx4_dev *dev, int slave, int op, int cmd,
 		return err;
 	}
 
-	err = add_res_range(dev, slave, index, 1, RES_COUNTER, 0);
+	err = add_res_range(dev, slave, index, 1, RES_COUNTER, port);
 	if (err) {
 		__mlx4_counter_free(dev, index);
 		mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0);
@@ -2101,7 +2262,7 @@ int mlx4_ALLOC_RES_wrapper(struct mlx4_dev *dev, int slave,
 
 	case RES_COUNTER:
 		err = counter_alloc_res(dev, slave, vhcr->op_modifier, alop,
-					vhcr->in_param, &vhcr->out_param);
+					vhcr->in_param, &vhcr->out_param, 0);
 		break;
 
 	case RES_XRCD:
@@ -2335,6 +2496,9 @@ static int counter_free_res(struct mlx4_dev *dev, int slave, int op, int cmd,
 		return -EINVAL;
 
 	index = get_param_l(&in_param);
+	if (index == MLX4_SINK_COUNTER_INDEX(dev))
+		return 0;
+
 	err = rem_res_range(dev, slave, index, 1, RES_COUNTER, 0);
 	if (err)
 		return err;
@@ -2703,6 +2867,10 @@ static void adjust_proxy_tun_qkey(struct mlx4_dev *dev, struct mlx4_vhcr *vhcr,
 	context->qkey = cpu_to_be32(qkey);
 }
 
+static int adjust_qp_sched_queue(struct mlx4_dev *dev, int slave,
+				 struct mlx4_qp_context *qpc,
+				 struct mlx4_cmd_mailbox *inbox);
+
 int mlx4_RST2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
 			     struct mlx4_vhcr *vhcr,
 			     struct mlx4_cmd_mailbox *inbox,
@@ -2725,6 +2893,10 @@ int mlx4_RST2INIT_QP_wrapper(struct mlx4_dev *dev, int slave,
 	struct res_srq *srq;
 	int local_qpn = be32_to_cpu(qpc->local_qpn) & 0xffffff;
 
+	err = adjust_qp_sched_queue(dev, slave, qpc, inbox);
+	if (err)
+		return err;
+
 	err = qp_res_start_move_to(dev, slave, qpn, RES_QP_HW, &qp, 0);
 	if (err)
 		return err;
@@ -3526,8 +3698,8 @@ static int adjust_qp_sched_queue(struct mlx4_dev *dev, int slave,
 	pri_sched_queue = (qpc->pri_path.sched_queue & ~(1 << 6)) |
 			  ((port & 1) << 6);
 
-	if (optpar & MLX4_QP_OPTPAR_PRIMARY_ADDR_PATH ||
-	    mlx4_is_eth(dev, port + 1)) {
+	if (optpar & (MLX4_QP_OPTPAR_PRIMARY_ADDR_PATH | MLX4_QP_OPTPAR_SCHED_QUEUE) ||
+	    qpc->pri_path.sched_queue || mlx4_is_eth(dev, port + 1)) {
 		qpc->pri_path.sched_queue = pri_sched_queue;
 	}
 
@@ -3595,9 +3767,6 @@ int mlx4_INIT2RTR_QP_wrapper(struct mlx4_dev *dev, int slave,
 	update_gid(dev, inbox, (u8)slave);
 	adjust_proxy_tun_qkey(dev, vhcr, qpc);
 	orig_sched_queue = qpc->pri_path.sched_queue;
-	err = update_vport_qp_param(dev, inbox, slave, qpn);
-	if (err)
-		return err;
 
 	err = get_res(dev, slave, qpn, RES_QP, &qp);
 	if (err)
@@ -3607,6 +3776,10 @@ int mlx4_INIT2RTR_QP_wrapper(struct mlx4_dev *dev, int slave,
 		goto out;
 	}
 
+	err = update_vport_qp_param(dev, inbox, slave, qpn);
+	if (err)
+		goto out;
+
 	err = mlx4_DMA_wrapper(dev, slave, vhcr, inbox, outbox, cmd);
 out:
 	/* if no error, save sched queue value passed in by VF. This is
@@ -3965,6 +4138,22 @@ static int validate_eth_header_mac(int slave, struct _rule_hw *eth_header,
 	return 0;
 }
 
+static void handle_eth_header_mcast_prio(struct mlx4_net_trans_rule_hw_ctrl *ctrl,
+					 struct _rule_hw *eth_header)
+{
+	if (is_multicast_ether_addr(eth_header->eth.dst_mac) ||
+	    is_broadcast_ether_addr(eth_header->eth.dst_mac)) {
+		struct mlx4_net_trans_rule_hw_eth *eth =
+			(struct mlx4_net_trans_rule_hw_eth *)eth_header;
+		struct _rule_hw *next_rule = (struct _rule_hw *)(eth + 1);
+		bool last_rule = next_rule->size == 0 && next_rule->id == 0 &&
+			next_rule->rsvd == 0;
+
+		if (last_rule)
+			ctrl->prio = cpu_to_be16(MLX4_DOMAIN_NIC);
+	}
+}
+
 /*
  * In case of missing eth header, append eth header with a MAC address
  * assigned to the VF.
@@ -4025,7 +4214,9 @@ static int add_eth_header(struct mlx4_dev *dev, int slave,
 
 }
 
-#define MLX4_UPD_QP_PATH_MASK_SUPPORTED (1ULL << MLX4_UPD_QP_PATH_MASK_MAC_INDEX)
+#define MLX4_UPD_QP_PATH_MASK_SUPPORTED      (                                \
+	1ULL << MLX4_UPD_QP_PATH_MASK_MAC_INDEX                     |\
+	1ULL << MLX4_UPD_QP_PATH_MASK_ETH_SRC_CHECK_MC_LB)
 int mlx4_UPDATE_QP_wrapper(struct mlx4_dev *dev, int slave,
 			   struct mlx4_vhcr *vhcr,
 			   struct mlx4_cmd_mailbox *inbox,
@@ -4048,6 +4239,16 @@ int mlx4_UPDATE_QP_wrapper(struct mlx4_dev *dev, int slave,
 	    (pri_addr_path_mask & ~MLX4_UPD_QP_PATH_MASK_SUPPORTED))
 		return -EPERM;
 
+	if ((pri_addr_path_mask &
+	     (1ULL << MLX4_UPD_QP_PATH_MASK_ETH_SRC_CHECK_MC_LB)) &&
+		!(dev->caps.flags2 &
+		  MLX4_DEV_CAP_FLAG2_UPDATE_QP_SRC_CHECK_LB)) {
+			mlx4_warn(dev,
+				  "Src check LB for slave %d isn't supported\n",
+				   slave);
+		return -ENOTSUPP;
+	}
+
 	/* Just change the smac for the QP */
 	err = get_res(dev, slave, qpn, RES_QP, &rqp);
 	if (err) {
@@ -4105,9 +4306,10 @@ int mlx4_QP_FLOW_STEERING_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
 		return -EOPNOTSUPP;
 
 	ctrl = (struct mlx4_net_trans_rule_hw_ctrl *)inbox->buf;
-	ctrl->port = mlx4_slave_convert_port(dev, slave, ctrl->port);
-	if (ctrl->port <= 0)
+	err = mlx4_slave_convert_port(dev, slave, ctrl->port);
+	if (err <= 0)
 		return -EINVAL;
+	ctrl->port = err;
 	qpn = be32_to_cpu(ctrl->qpn) & 0xffffff;
 	err = get_res(dev, slave, qpn, RES_QP, &rqp);
 	if (err) {
@@ -4117,6 +4319,12 @@ int mlx4_QP_FLOW_STEERING_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
 	rule_header = (struct _rule_hw *)(ctrl + 1);
 	header_id = map_hw_to_sw_id(be16_to_cpu(rule_header->id));
 
+	if (header_id == MLX4_NET_TRANS_RULE_ID_ETH)
+		handle_eth_header_mcast_prio(ctrl, rule_header);
+
+	if (slave == dev->caps.function)
+		goto execute;
+
 	switch (header_id) {
 	case MLX4_NET_TRANS_RULE_ID_ETH:
 		if (validate_eth_header_mac(slave, rule_header, rlist)) {
@@ -4143,6 +4351,7 @@ int mlx4_QP_FLOW_STEERING_ATTACH_wrapper(struct mlx4_dev *dev, int slave,
 		goto err_put;
 	}
 
+execute:
 	err = mlx4_cmd_imm(dev, inbox->dma, &vhcr->out_param,
 			   vhcr->in_modifier, 0,
 			   MLX4_QP_FLOW_STEERING_ATTACH, MLX4_CMD_TIME_CLASS_A,
@@ -4744,26 +4953,41 @@ static void rem_slave_counters(struct mlx4_dev *dev, int slave)
 	struct res_counter *counter;
 	struct res_counter *tmp;
 	int err;
-	int index;
+	int *counters_arr = NULL;
+	int i, j;
 
 	err = move_all_busy(dev, slave, RES_COUNTER);
 	if (err)
 		mlx4_warn(dev, "rem_slave_counters: Could not move all counters - too busy for slave %d\n",
 			  slave);
 
-	spin_lock_irq(mlx4_tlock(dev));
-	list_for_each_entry_safe(counter, tmp, counter_list, com.list) {
-		if (counter->com.owner == slave) {
-			index = counter->com.res_id;
-			rb_erase(&counter->com.node,
-				 &tracker->res_tree[RES_COUNTER]);
-			list_del(&counter->com.list);
-			kfree(counter);
-			__mlx4_counter_free(dev, index);
+	counters_arr = kmalloc_array(dev->caps.max_counters,
+				     sizeof(*counters_arr), GFP_KERNEL);
+	if (!counters_arr)
+		return;
+
+	do {
+		i = 0;
+		j = 0;
+		spin_lock_irq(mlx4_tlock(dev));
+		list_for_each_entry_safe(counter, tmp, counter_list, com.list) {
+			if (counter->com.owner == slave) {
+				counters_arr[i++] = counter->com.res_id;
+				rb_erase(&counter->com.node,
+					 &tracker->res_tree[RES_COUNTER]);
+				list_del(&counter->com.list);
+				kfree(counter);
+			}
+		}
+		spin_unlock_irq(mlx4_tlock(dev));
+
+		while (j < i) {
+			__mlx4_counter_free(dev, counters_arr[j++]);
 			mlx4_release_resource(dev, slave, RES_COUNTER, 1, 0);
 		}
-	}
-	spin_unlock_irq(mlx4_tlock(dev));
+	} while (i);
+
+	kfree(counters_arr);
 }
 
 static void rem_slave_xrcdns(struct mlx4_dev *dev, int slave)
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/Kconfig b/kernel/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
index 8ff57e8e3..158c88c69 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/Kconfig
@@ -3,6 +3,18 @@
 #
 
 config MLX5_CORE
-	tristate
+	tristate "Mellanox Technologies ConnectX-4 and Connect-IB core driver"
 	depends on PCI
 	default n
+	---help---
+	  Core driver for low level functionality of the ConnectX-4 and
+	  Connect-IB cards by Mellanox Technologies.
+
+config MLX5_CORE_EN
+	bool "Mellanox Technologies ConnectX-4 Ethernet support"
+	depends on NETDEVICES && ETHERNET && PCI && MLX5_CORE
+	default n
+	---help---
+	  Ethernet support in Mellanox Technologies ConnectX-4 NIC.
+	  Ethernet and Infiniband support in ConnectX-4 are currently mutually
+	  exclusive.
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/kernel/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index 105780bb9..26a68b8af 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -2,4 +2,7 @@ obj-$(CONFIG_MLX5_CORE)		+= mlx5_core.o
 
 mlx5_core-y :=	main.o cmd.o debugfs.o fw.o eq.o uar.o pagealloc.o \
 		health.o mcg.o cq.o srq.o alloc.o qp.o port.o mr.o pd.o   \
-		mad.o
+		mad.o transobj.o vport.o
+mlx5_core-$(CONFIG_MLX5_CORE_EN) += wq.o flow_table.o \
+		en_main.o en_flow_table.o en_ethtool.o en_tx.o en_rx.o \
+		en_txrx.o
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/alloc.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
index ac0f7bf4b..6cb383046 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
@@ -42,99 +42,65 @@
 #include "mlx5_core.h"
 
 /* Handling for queue buffers -- we allocate a bunch of memory and
- * register it in a memory region at HCA virtual address 0.  If the
- * requested size is > max_direct, we split the allocation into
- * multiple pages, so we don't require too much contiguous memory.
+ * register it in a memory region at HCA virtual address 0.
  */
 
-int mlx5_buf_alloc(struct mlx5_core_dev *dev, int size, int max_direct,
-		   struct mlx5_buf *buf)
+static void *mlx5_dma_zalloc_coherent_node(struct mlx5_core_dev *dev,
+					   size_t size, dma_addr_t *dma_handle,
+					   int node)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	int original_node;
+	void *cpu_handle;
+
+	mutex_lock(&priv->alloc_mutex);
+	original_node = dev_to_node(&dev->pdev->dev);
+	set_dev_node(&dev->pdev->dev, node);
+	cpu_handle = dma_zalloc_coherent(&dev->pdev->dev, size,
+					 dma_handle, GFP_KERNEL);
+	set_dev_node(&dev->pdev->dev, original_node);
+	mutex_unlock(&priv->alloc_mutex);
+	return cpu_handle;
+}
+
+int mlx5_buf_alloc_node(struct mlx5_core_dev *dev, int size,
+			struct mlx5_buf *buf, int node)
 {
 	dma_addr_t t;
 
 	buf->size = size;
-	if (size <= max_direct) {
-		buf->nbufs        = 1;
-		buf->npages       = 1;
-		buf->page_shift   = (u8)get_order(size) + PAGE_SHIFT;
-		buf->direct.buf   = dma_zalloc_coherent(&dev->pdev->dev,
-							size, &t, GFP_KERNEL);
-		if (!buf->direct.buf)
-			return -ENOMEM;
-
-		buf->direct.map = t;
-
-		while (t & ((1 << buf->page_shift) - 1)) {
-			--buf->page_shift;
-			buf->npages *= 2;
-		}
-	} else {
-		int i;
-
-		buf->direct.buf  = NULL;
-		buf->nbufs       = (size + PAGE_SIZE - 1) / PAGE_SIZE;
-		buf->npages      = buf->nbufs;
-		buf->page_shift  = PAGE_SHIFT;
-		buf->page_list   = kcalloc(buf->nbufs, sizeof(*buf->page_list),
-					   GFP_KERNEL);
-		if (!buf->page_list)
-			return -ENOMEM;
-
-		for (i = 0; i < buf->nbufs; i++) {
-			buf->page_list[i].buf =
-				dma_zalloc_coherent(&dev->pdev->dev, PAGE_SIZE,
-						    &t, GFP_KERNEL);
-			if (!buf->page_list[i].buf)
-				goto err_free;
-
-			buf->page_list[i].map = t;
-		}
-
-		if (BITS_PER_LONG == 64) {
-			struct page **pages;
-			pages = kmalloc(sizeof(*pages) * buf->nbufs, GFP_KERNEL);
-			if (!pages)
-				goto err_free;
-			for (i = 0; i < buf->nbufs; i++)
-				pages[i] = virt_to_page(buf->page_list[i].buf);
-			buf->direct.buf = vmap(pages, buf->nbufs, VM_MAP, PAGE_KERNEL);
-			kfree(pages);
-			if (!buf->direct.buf)
-				goto err_free;
-		}
+	buf->npages       = 1;
+	buf->page_shift   = (u8)get_order(size) + PAGE_SHIFT;
+	buf->direct.buf   = mlx5_dma_zalloc_coherent_node(dev, size,
+							  &t, node);
+	if (!buf->direct.buf)
+		return -ENOMEM;
+
+	buf->direct.map = t;
+
+	while (t & ((1 << buf->page_shift) - 1)) {
+		--buf->page_shift;
+		buf->npages *= 2;
 	}
 
 	return 0;
+}
 
-err_free:
-	mlx5_buf_free(dev, buf);
-
-	return -ENOMEM;
+int mlx5_buf_alloc(struct mlx5_core_dev *dev, int size, struct mlx5_buf *buf)
+{
+	return mlx5_buf_alloc_node(dev, size, buf, dev->priv.numa_node);
 }
 EXPORT_SYMBOL_GPL(mlx5_buf_alloc);
 
 void mlx5_buf_free(struct mlx5_core_dev *dev, struct mlx5_buf *buf)
 {
-	int i;
-
-	if (buf->nbufs == 1)
-		dma_free_coherent(&dev->pdev->dev, buf->size, buf->direct.buf,
-				  buf->direct.map);
-	else {
-		if (BITS_PER_LONG == 64)
-			vunmap(buf->direct.buf);
-
-		for (i = 0; i < buf->nbufs; i++)
-			if (buf->page_list[i].buf)
-				dma_free_coherent(&dev->pdev->dev, PAGE_SIZE,
-						  buf->page_list[i].buf,
-						  buf->page_list[i].map);
-		kfree(buf->page_list);
-	}
+	dma_free_coherent(&dev->pdev->dev, buf->size, buf->direct.buf,
+			  buf->direct.map);
 }
 EXPORT_SYMBOL_GPL(mlx5_buf_free);
 
-static struct mlx5_db_pgdir *mlx5_alloc_db_pgdir(struct device *dma_device)
+static struct mlx5_db_pgdir *mlx5_alloc_db_pgdir(struct mlx5_core_dev *dev,
+						 int node)
 {
 	struct mlx5_db_pgdir *pgdir;
 
@@ -143,8 +109,9 @@ static struct mlx5_db_pgdir *mlx5_alloc_db_pgdir(struct device *dma_device)
 		return NULL;
 
 	bitmap_fill(pgdir->bitmap, MLX5_DB_PER_PAGE);
-	pgdir->db_page = dma_alloc_coherent(dma_device, PAGE_SIZE,
-					    &pgdir->db_dma, GFP_KERNEL);
+
+	pgdir->db_page = mlx5_dma_zalloc_coherent_node(dev, PAGE_SIZE,
+						       &pgdir->db_dma, node);
 	if (!pgdir->db_page) {
 		kfree(pgdir);
 		return NULL;
@@ -177,7 +144,7 @@ static int mlx5_alloc_db_from_pgdir(struct mlx5_db_pgdir *pgdir,
 	return 0;
 }
 
-int mlx5_db_alloc(struct mlx5_core_dev *dev, struct mlx5_db *db)
+int mlx5_db_alloc_node(struct mlx5_core_dev *dev, struct mlx5_db *db, int node)
 {
 	struct mlx5_db_pgdir *pgdir;
 	int ret = 0;
@@ -188,7 +155,7 @@ int mlx5_db_alloc(struct mlx5_core_dev *dev, struct mlx5_db *db)
 		if (!mlx5_alloc_db_from_pgdir(pgdir, db))
 			goto out;
 
-	pgdir = mlx5_alloc_db_pgdir(&(dev->pdev->dev));
+	pgdir = mlx5_alloc_db_pgdir(dev, node);
 	if (!pgdir) {
 		ret = -ENOMEM;
 		goto out;
@@ -204,6 +171,12 @@ out:
 
 	return ret;
 }
+EXPORT_SYMBOL_GPL(mlx5_db_alloc_node);
+
+int mlx5_db_alloc(struct mlx5_core_dev *dev, struct mlx5_db *db)
+{
+	return mlx5_db_alloc_node(dev, db, dev->priv.numa_node);
+}
 EXPORT_SYMBOL_GPL(mlx5_db_alloc);
 
 void mlx5_db_free(struct mlx5_core_dev *dev, struct mlx5_db *db)
@@ -230,10 +203,7 @@ void mlx5_fill_page_array(struct mlx5_buf *buf, __be64 *pas)
 	int i;
 
 	for (i = 0; i < buf->npages; i++) {
-		if (buf->nbufs == 1)
-			addr = buf->direct.map + (i << buf->page_shift);
-		else
-			addr = buf->page_list[i].map;
+		addr = buf->direct.map + (i << buf->page_shift);
 
 		pas[i] = cpu_to_be64(addr);
 	}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/cmd.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
index e3273faf4..037fc4cdf 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -30,7 +30,7 @@
  * SOFTWARE.
  */
 
-#include <asm-generic/kmap_types.h>
+#include <linux/highmem.h>
 #include <linux/module.h>
 #include <linux/errno.h>
 #include <linux/pci.h>
@@ -75,25 +75,6 @@ enum {
 	MLX5_CMD_DELIVERY_STAT_CMD_DESCR_ERR		= 0x10,
 };
 
-enum {
-	MLX5_CMD_STAT_OK			= 0x0,
-	MLX5_CMD_STAT_INT_ERR			= 0x1,
-	MLX5_CMD_STAT_BAD_OP_ERR		= 0x2,
-	MLX5_CMD_STAT_BAD_PARAM_ERR		= 0x3,
-	MLX5_CMD_STAT_BAD_SYS_STATE_ERR		= 0x4,
-	MLX5_CMD_STAT_BAD_RES_ERR		= 0x5,
-	MLX5_CMD_STAT_RES_BUSY			= 0x6,
-	MLX5_CMD_STAT_LIM_ERR			= 0x8,
-	MLX5_CMD_STAT_BAD_RES_STATE_ERR		= 0x9,
-	MLX5_CMD_STAT_IX_ERR			= 0xa,
-	MLX5_CMD_STAT_NO_RES_ERR		= 0xf,
-	MLX5_CMD_STAT_BAD_INP_LEN_ERR		= 0x50,
-	MLX5_CMD_STAT_BAD_OUTP_LEN_ERR		= 0x51,
-	MLX5_CMD_STAT_BAD_QP_STATE_ERR		= 0x10,
-	MLX5_CMD_STAT_BAD_PKT_ERR		= 0x30,
-	MLX5_CMD_STAT_BAD_SIZE_OUTS_CQES_ERR	= 0x40,
-};
-
 static struct mlx5_cmd_work_ent *alloc_cmd(struct mlx5_cmd *cmd,
 					   struct mlx5_cmd_msg *in,
 					   struct mlx5_cmd_msg *out,
@@ -273,6 +254,156 @@ static void dump_buf(void *buf, int size, int data_only, int offset)
 		pr_debug("\n");
 }
 
+enum {
+	MLX5_DRIVER_STATUS_ABORTED = 0xfe,
+	MLX5_DRIVER_SYND = 0xbadd00de,
+};
+
+static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op,
+				       u32 *synd, u8 *status)
+{
+	*synd = 0;
+	*status = 0;
+
+	switch (op) {
+	case MLX5_CMD_OP_TEARDOWN_HCA:
+	case MLX5_CMD_OP_DISABLE_HCA:
+	case MLX5_CMD_OP_MANAGE_PAGES:
+	case MLX5_CMD_OP_DESTROY_MKEY:
+	case MLX5_CMD_OP_DESTROY_EQ:
+	case MLX5_CMD_OP_DESTROY_CQ:
+	case MLX5_CMD_OP_DESTROY_QP:
+	case MLX5_CMD_OP_DESTROY_PSV:
+	case MLX5_CMD_OP_DESTROY_SRQ:
+	case MLX5_CMD_OP_DESTROY_XRC_SRQ:
+	case MLX5_CMD_OP_DESTROY_DCT:
+	case MLX5_CMD_OP_DEALLOC_Q_COUNTER:
+	case MLX5_CMD_OP_DEALLOC_PD:
+	case MLX5_CMD_OP_DEALLOC_UAR:
+	case MLX5_CMD_OP_DETTACH_FROM_MCG:
+	case MLX5_CMD_OP_DEALLOC_XRCD:
+	case MLX5_CMD_OP_DEALLOC_TRANSPORT_DOMAIN:
+	case MLX5_CMD_OP_DELETE_VXLAN_UDP_DPORT:
+	case MLX5_CMD_OP_DELETE_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_DESTROY_TIR:
+	case MLX5_CMD_OP_DESTROY_SQ:
+	case MLX5_CMD_OP_DESTROY_RQ:
+	case MLX5_CMD_OP_DESTROY_RMP:
+	case MLX5_CMD_OP_DESTROY_TIS:
+	case MLX5_CMD_OP_DESTROY_RQT:
+	case MLX5_CMD_OP_DESTROY_FLOW_TABLE:
+	case MLX5_CMD_OP_DESTROY_FLOW_GROUP:
+	case MLX5_CMD_OP_DELETE_FLOW_TABLE_ENTRY:
+		return MLX5_CMD_STAT_OK;
+
+	case MLX5_CMD_OP_QUERY_HCA_CAP:
+	case MLX5_CMD_OP_QUERY_ADAPTER:
+	case MLX5_CMD_OP_INIT_HCA:
+	case MLX5_CMD_OP_ENABLE_HCA:
+	case MLX5_CMD_OP_QUERY_PAGES:
+	case MLX5_CMD_OP_SET_HCA_CAP:
+	case MLX5_CMD_OP_QUERY_ISSI:
+	case MLX5_CMD_OP_SET_ISSI:
+	case MLX5_CMD_OP_CREATE_MKEY:
+	case MLX5_CMD_OP_QUERY_MKEY:
+	case MLX5_CMD_OP_QUERY_SPECIAL_CONTEXTS:
+	case MLX5_CMD_OP_PAGE_FAULT_RESUME:
+	case MLX5_CMD_OP_CREATE_EQ:
+	case MLX5_CMD_OP_QUERY_EQ:
+	case MLX5_CMD_OP_GEN_EQE:
+	case MLX5_CMD_OP_CREATE_CQ:
+	case MLX5_CMD_OP_QUERY_CQ:
+	case MLX5_CMD_OP_MODIFY_CQ:
+	case MLX5_CMD_OP_CREATE_QP:
+	case MLX5_CMD_OP_RST2INIT_QP:
+	case MLX5_CMD_OP_INIT2RTR_QP:
+	case MLX5_CMD_OP_RTR2RTS_QP:
+	case MLX5_CMD_OP_RTS2RTS_QP:
+	case MLX5_CMD_OP_SQERR2RTS_QP:
+	case MLX5_CMD_OP_2ERR_QP:
+	case MLX5_CMD_OP_2RST_QP:
+	case MLX5_CMD_OP_QUERY_QP:
+	case MLX5_CMD_OP_SQD_RTS_QP:
+	case MLX5_CMD_OP_INIT2INIT_QP:
+	case MLX5_CMD_OP_CREATE_PSV:
+	case MLX5_CMD_OP_CREATE_SRQ:
+	case MLX5_CMD_OP_QUERY_SRQ:
+	case MLX5_CMD_OP_ARM_RQ:
+	case MLX5_CMD_OP_CREATE_XRC_SRQ:
+	case MLX5_CMD_OP_QUERY_XRC_SRQ:
+	case MLX5_CMD_OP_ARM_XRC_SRQ:
+	case MLX5_CMD_OP_CREATE_DCT:
+	case MLX5_CMD_OP_DRAIN_DCT:
+	case MLX5_CMD_OP_QUERY_DCT:
+	case MLX5_CMD_OP_ARM_DCT_FOR_KEY_VIOLATION:
+	case MLX5_CMD_OP_QUERY_VPORT_STATE:
+	case MLX5_CMD_OP_MODIFY_VPORT_STATE:
+	case MLX5_CMD_OP_QUERY_ESW_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_ESW_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_NIC_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_ROCE_ADDRESS:
+	case MLX5_CMD_OP_SET_ROCE_ADDRESS:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_CONTEXT:
+	case MLX5_CMD_OP_MODIFY_HCA_VPORT_CONTEXT:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_GID:
+	case MLX5_CMD_OP_QUERY_HCA_VPORT_PKEY:
+	case MLX5_CMD_OP_QUERY_VPORT_COUNTER:
+	case MLX5_CMD_OP_ALLOC_Q_COUNTER:
+	case MLX5_CMD_OP_QUERY_Q_COUNTER:
+	case MLX5_CMD_OP_ALLOC_PD:
+	case MLX5_CMD_OP_ALLOC_UAR:
+	case MLX5_CMD_OP_CONFIG_INT_MODERATION:
+	case MLX5_CMD_OP_ACCESS_REG:
+	case MLX5_CMD_OP_ATTACH_TO_MCG:
+	case MLX5_CMD_OP_GET_DROPPED_PACKET_LOG:
+	case MLX5_CMD_OP_MAD_IFC:
+	case MLX5_CMD_OP_QUERY_MAD_DEMUX:
+	case MLX5_CMD_OP_SET_MAD_DEMUX:
+	case MLX5_CMD_OP_NOP:
+	case MLX5_CMD_OP_ALLOC_XRCD:
+	case MLX5_CMD_OP_ALLOC_TRANSPORT_DOMAIN:
+	case MLX5_CMD_OP_QUERY_CONG_STATUS:
+	case MLX5_CMD_OP_MODIFY_CONG_STATUS:
+	case MLX5_CMD_OP_QUERY_CONG_PARAMS:
+	case MLX5_CMD_OP_MODIFY_CONG_PARAMS:
+	case MLX5_CMD_OP_QUERY_CONG_STATISTICS:
+	case MLX5_CMD_OP_ADD_VXLAN_UDP_DPORT:
+	case MLX5_CMD_OP_SET_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_QUERY_L2_TABLE_ENTRY:
+	case MLX5_CMD_OP_CREATE_TIR:
+	case MLX5_CMD_OP_MODIFY_TIR:
+	case MLX5_CMD_OP_QUERY_TIR:
+	case MLX5_CMD_OP_CREATE_SQ:
+	case MLX5_CMD_OP_MODIFY_SQ:
+	case MLX5_CMD_OP_QUERY_SQ:
+	case MLX5_CMD_OP_CREATE_RQ:
+	case MLX5_CMD_OP_MODIFY_RQ:
+	case MLX5_CMD_OP_QUERY_RQ:
+	case MLX5_CMD_OP_CREATE_RMP:
+	case MLX5_CMD_OP_MODIFY_RMP:
+	case MLX5_CMD_OP_QUERY_RMP:
+	case MLX5_CMD_OP_CREATE_TIS:
+	case MLX5_CMD_OP_MODIFY_TIS:
+	case MLX5_CMD_OP_QUERY_TIS:
+	case MLX5_CMD_OP_CREATE_RQT:
+	case MLX5_CMD_OP_MODIFY_RQT:
+	case MLX5_CMD_OP_QUERY_RQT:
+	case MLX5_CMD_OP_CREATE_FLOW_TABLE:
+	case MLX5_CMD_OP_QUERY_FLOW_TABLE:
+	case MLX5_CMD_OP_CREATE_FLOW_GROUP:
+	case MLX5_CMD_OP_QUERY_FLOW_GROUP:
+	case MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY:
+	case MLX5_CMD_OP_QUERY_FLOW_TABLE_ENTRY:
+		*status = MLX5_DRIVER_STATUS_ABORTED;
+		*synd = MLX5_DRIVER_SYND;
+		return -EIO;
+	default:
+		mlx5_core_err(dev, "Unknown FW command (%d)\n", op);
+		return -EINVAL;
+	}
+}
+
 const char *mlx5_command_str(int command)
 {
 	switch (command) {
@@ -390,8 +521,17 @@ const char *mlx5_command_str(int command)
 	case MLX5_CMD_OP_ARM_RQ:
 		return "ARM_RQ";
 
-	case MLX5_CMD_OP_RESIZE_SRQ:
-		return "RESIZE_SRQ";
+	case MLX5_CMD_OP_CREATE_XRC_SRQ:
+		return "CREATE_XRC_SRQ";
+
+	case MLX5_CMD_OP_DESTROY_XRC_SRQ:
+		return "DESTROY_XRC_SRQ";
+
+	case MLX5_CMD_OP_QUERY_XRC_SRQ:
+		return "QUERY_XRC_SRQ";
+
+	case MLX5_CMD_OP_ARM_XRC_SRQ:
+		return "ARM_XRC_SRQ";
 
 	case MLX5_CMD_OP_ALLOC_PD:
 		return "ALLOC_PD";
@@ -408,8 +548,8 @@ const char *mlx5_command_str(int command)
 	case MLX5_CMD_OP_ATTACH_TO_MCG:
 		return "ATTACH_TO_MCG";
 
-	case MLX5_CMD_OP_DETACH_FROM_MCG:
-		return "DETACH_FROM_MCG";
+	case MLX5_CMD_OP_DETTACH_FROM_MCG:
+		return "DETTACH_FROM_MCG";
 
 	case MLX5_CMD_OP_ALLOC_XRCD:
 		return "ALLOC_XRCD";
@@ -483,6 +623,7 @@ static void cmd_work_handler(struct work_struct *work)
 	struct mlx5_core_dev *dev = container_of(cmd, struct mlx5_core_dev, cmd);
 	struct mlx5_cmd_layout *lay;
 	struct semaphore *sem;
+	unsigned long flags;
 
 	sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem;
 	down(sem);
@@ -495,6 +636,9 @@ static void cmd_work_handler(struct work_struct *work)
 		}
 	} else {
 		ent->idx = cmd->max_reg_cmds;
+		spin_lock_irqsave(&cmd->alloc_lock, flags);
+		clear_bit(ent->idx, &cmd->bitmask);
+		spin_unlock_irqrestore(&cmd->alloc_lock, flags);
 	}
 
 	ent->token = alloc_token(cmd);
@@ -594,6 +738,16 @@ static int wait_func(struct mlx5_core_dev *dev, struct mlx5_cmd_work_ent *ent)
 	return err;
 }
 
+static __be32 *get_synd_ptr(struct mlx5_outbox_hdr *out)
+{
+	return &out->syndrome;
+}
+
+static u8 *get_status_ptr(struct mlx5_outbox_hdr *out)
+{
+	return &out->status;
+}
+
 /*  Notes:
  *    1. Callback functions may not sleep
  *    2. page queue commands do not support asynchrous completion
@@ -1091,7 +1245,7 @@ static void free_msg(struct mlx5_core_dev *dev, struct mlx5_cmd_msg *msg)
 	}
 }
 
-void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
+void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, u64 vec)
 {
 	struct mlx5_cmd *cmd = &dev->cmd;
 	struct mlx5_cmd_work_ent *ent;
@@ -1102,7 +1256,10 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
 	s64 ds;
 	struct mlx5_cmd_stats *stats;
 	unsigned long flags;
+	unsigned long vector;
 
+	/* there can be at most 32 command queues */
+	vector = vec & 0xffffffff;
 	for (i = 0; i < (1 << cmd->log_sz); i++) {
 		if (test_bit(i, &vector)) {
 			struct semaphore *sem;
@@ -1120,11 +1277,16 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
 					ent->ret = verify_signature(ent);
 				else
 					ent->ret = 0;
-				ent->status = ent->lay->status_own >> 1;
+				if (vec & MLX5_TRIGGERED_CMD_COMP)
+					ent->status = MLX5_DRIVER_STATUS_ABORTED;
+				else
+					ent->status = ent->lay->status_own >> 1;
+
 				mlx5_core_dbg(dev, "command completed. ret 0x%x, delivery status %s(0x%x)\n",
 					      ent->ret, deliv_status_to_str(ent->status), ent->status);
 			}
 			free_ent(cmd, ent->idx);
+
 			if (ent->callback) {
 				ds = ent->ts2 - ent->ts1;
 				if (ent->op < ARRAY_SIZE(cmd->stats)) {
@@ -1146,6 +1308,7 @@ void mlx5_cmd_comp_handler(struct mlx5_core_dev *dev, unsigned long vector)
 				mlx5_free_cmd_msg(dev, ent->out);
 				free_msg(dev, ent->in);
 
+				err = err ? err : ent->status;
 				free_cmd(ent);
 				callback(err, context);
 			} else {
@@ -1193,6 +1356,11 @@ static struct mlx5_cmd_msg *alloc_msg(struct mlx5_core_dev *dev, int in_size,
 	return msg;
 }
 
+static u16 opcode_from_in(struct mlx5_inbox_hdr *in)
+{
+	return be16_to_cpu(in->opcode);
+}
+
 static int is_manage_pages(struct mlx5_inbox_hdr *in)
 {
 	return be16_to_cpu(in->opcode) == MLX5_CMD_OP_MANAGE_PAGES;
@@ -1207,6 +1375,15 @@ static int cmd_exec(struct mlx5_core_dev *dev, void *in, int in_size, void *out,
 	gfp_t gfp;
 	int err;
 	u8 status = 0;
+	u32 drv_synd;
+
+	if (pci_channel_offline(dev->pdev) ||
+	    dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		err = mlx5_internal_err_ret_value(dev, opcode_from_in(in), &drv_synd, &status);
+		*get_synd_ptr(out) = cpu_to_be32(drv_synd);
+		*get_status_ptr(out) = status;
+		return err;
+	}
 
 	pages_queue = is_manage_pages(in);
 	gfp = callback ? GFP_ATOMIC : GFP_KERNEL;
@@ -1373,6 +1550,7 @@ int mlx5_cmd_init(struct mlx5_core_dev *dev)
 	int err;
 	int i;
 
+	memset(cmd, 0, sizeof(*cmd));
 	cmd_if_rev = cmdif_rev(dev);
 	if (cmd_if_rev != CMD_IF_REV) {
 		dev_err(&dev->pdev->dev,
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/cq.c
index eb0cf81f5..b51e42d6f 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/cq.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/cq.c
@@ -219,11 +219,30 @@ int mlx5_core_modify_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq,
 }
 EXPORT_SYMBOL(mlx5_core_modify_cq);
 
+int mlx5_core_modify_cq_moderation(struct mlx5_core_dev *dev,
+				   struct mlx5_core_cq *cq,
+				   u16 cq_period,
+				   u16 cq_max_count)
+{
+	struct mlx5_modify_cq_mbox_in in;
+
+	memset(&in, 0, sizeof(in));
+
+	in.cqn              = cpu_to_be32(cq->cqn);
+	in.ctx.cq_period    = cpu_to_be16(cq_period);
+	in.ctx.cq_max_count = cpu_to_be16(cq_max_count);
+	in.field_select     = cpu_to_be32(MLX5_CQ_MODIFY_PERIOD |
+					  MLX5_CQ_MODIFY_COUNT);
+
+	return mlx5_core_modify_cq(dev, cq, &in, sizeof(in));
+}
+
 int mlx5_init_cq_table(struct mlx5_core_dev *dev)
 {
 	struct mlx5_cq_table *table = &dev->priv.cq_table;
 	int err;
 
+	memset(table, 0, sizeof(*table));
 	spin_lock_init(&table->lock);
 	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
 	err = mlx5_cq_debugfs_init(dev);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en.h b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en.h
new file mode 100644
index 000000000..22e72bf1a
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -0,0 +1,633 @@
+/*
+ * Copyright (c) 2015, 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 <linux/if_vlan.h>
+#include <linux/etherdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/qp.h>
+#include <linux/mlx5/cq.h>
+#include <linux/mlx5/vport.h>
+#include "wq.h"
+#include "transobj.h"
+#include "mlx5_core.h"
+
+#define MLX5E_MAX_NUM_TC	8
+
+#define MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE                0x6
+#define MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE                0xa
+#define MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE                0xd
+
+#define MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE                0x1
+#define MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE                0xa
+#define MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE                0xd
+
+#define MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ                 (64 * 1024)
+#define MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC      0x10
+#define MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_PKTS      0x20
+#define MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC      0x10
+#define MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_PKTS      0x20
+#define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES                0x80
+
+#define MLX5E_LOG_INDIR_RQT_SIZE       0x7
+#define MLX5E_INDIR_RQT_SIZE           BIT(MLX5E_LOG_INDIR_RQT_SIZE)
+#define MLX5E_MAX_NUM_CHANNELS         (MLX5E_INDIR_RQT_SIZE >> 1)
+#define MLX5E_TX_CQ_POLL_BUDGET        128
+#define MLX5E_UPDATE_STATS_INTERVAL    200 /* msecs */
+#define MLX5E_SQ_BF_BUDGET             16
+
+static const char vport_strings[][ETH_GSTRING_LEN] = {
+	/* vport statistics */
+	"rx_packets",
+	"rx_bytes",
+	"tx_packets",
+	"tx_bytes",
+	"rx_error_packets",
+	"rx_error_bytes",
+	"tx_error_packets",
+	"tx_error_bytes",
+	"rx_unicast_packets",
+	"rx_unicast_bytes",
+	"tx_unicast_packets",
+	"tx_unicast_bytes",
+	"rx_multicast_packets",
+	"rx_multicast_bytes",
+	"tx_multicast_packets",
+	"tx_multicast_bytes",
+	"rx_broadcast_packets",
+	"rx_broadcast_bytes",
+	"tx_broadcast_packets",
+	"tx_broadcast_bytes",
+
+	/* SW counters */
+	"tso_packets",
+	"tso_bytes",
+	"lro_packets",
+	"lro_bytes",
+	"rx_csum_good",
+	"rx_csum_none",
+	"rx_csum_sw",
+	"tx_csum_offload",
+	"tx_queue_stopped",
+	"tx_queue_wake",
+	"tx_queue_dropped",
+	"rx_wqe_err",
+};
+
+struct mlx5e_vport_stats {
+	/* HW counters */
+	u64 rx_packets;
+	u64 rx_bytes;
+	u64 tx_packets;
+	u64 tx_bytes;
+	u64 rx_error_packets;
+	u64 rx_error_bytes;
+	u64 tx_error_packets;
+	u64 tx_error_bytes;
+	u64 rx_unicast_packets;
+	u64 rx_unicast_bytes;
+	u64 tx_unicast_packets;
+	u64 tx_unicast_bytes;
+	u64 rx_multicast_packets;
+	u64 rx_multicast_bytes;
+	u64 tx_multicast_packets;
+	u64 tx_multicast_bytes;
+	u64 rx_broadcast_packets;
+	u64 rx_broadcast_bytes;
+	u64 tx_broadcast_packets;
+	u64 tx_broadcast_bytes;
+
+	/* SW counters */
+	u64 tso_packets;
+	u64 tso_bytes;
+	u64 lro_packets;
+	u64 lro_bytes;
+	u64 rx_csum_good;
+	u64 rx_csum_none;
+	u64 rx_csum_sw;
+	u64 tx_csum_offload;
+	u64 tx_queue_stopped;
+	u64 tx_queue_wake;
+	u64 tx_queue_dropped;
+	u64 rx_wqe_err;
+
+#define NUM_VPORT_COUNTERS     32
+};
+
+static const char pport_strings[][ETH_GSTRING_LEN] = {
+	/* IEEE802.3 counters */
+	"frames_tx",
+	"frames_rx",
+	"check_seq_err",
+	"alignment_err",
+	"octets_tx",
+	"octets_received",
+	"multicast_xmitted",
+	"broadcast_xmitted",
+	"multicast_rx",
+	"broadcast_rx",
+	"in_range_len_errors",
+	"out_of_range_len",
+	"too_long_errors",
+	"symbol_err",
+	"mac_control_tx",
+	"mac_control_rx",
+	"unsupported_op_rx",
+	"pause_ctrl_rx",
+	"pause_ctrl_tx",
+
+	/* RFC2863 counters */
+	"in_octets",
+	"in_ucast_pkts",
+	"in_discards",
+	"in_errors",
+	"in_unknown_protos",
+	"out_octets",
+	"out_ucast_pkts",
+	"out_discards",
+	"out_errors",
+	"in_multicast_pkts",
+	"in_broadcast_pkts",
+	"out_multicast_pkts",
+	"out_broadcast_pkts",
+
+	/* RFC2819 counters */
+	"drop_events",
+	"octets",
+	"pkts",
+	"broadcast_pkts",
+	"multicast_pkts",
+	"crc_align_errors",
+	"undersize_pkts",
+	"oversize_pkts",
+	"fragments",
+	"jabbers",
+	"collisions",
+	"p64octets",
+	"p65to127octets",
+	"p128to255octets",
+	"p256to511octets",
+	"p512to1023octets",
+	"p1024to1518octets",
+	"p1519to2047octets",
+	"p2048to4095octets",
+	"p4096to8191octets",
+	"p8192to10239octets",
+};
+
+#define NUM_IEEE_802_3_COUNTERS		19
+#define NUM_RFC_2863_COUNTERS		13
+#define NUM_RFC_2819_COUNTERS		21
+#define NUM_PPORT_COUNTERS		(NUM_IEEE_802_3_COUNTERS + \
+					 NUM_RFC_2863_COUNTERS + \
+					 NUM_RFC_2819_COUNTERS)
+
+struct mlx5e_pport_stats {
+	__be64 IEEE_802_3_counters[NUM_IEEE_802_3_COUNTERS];
+	__be64 RFC_2863_counters[NUM_RFC_2863_COUNTERS];
+	__be64 RFC_2819_counters[NUM_RFC_2819_COUNTERS];
+};
+
+static const char rq_stats_strings[][ETH_GSTRING_LEN] = {
+	"packets",
+	"csum_none",
+	"csum_sw",
+	"lro_packets",
+	"lro_bytes",
+	"wqe_err"
+};
+
+struct mlx5e_rq_stats {
+	u64 packets;
+	u64 csum_none;
+	u64 csum_sw;
+	u64 lro_packets;
+	u64 lro_bytes;
+	u64 wqe_err;
+#define NUM_RQ_STATS 6
+};
+
+static const char sq_stats_strings[][ETH_GSTRING_LEN] = {
+	"packets",
+	"tso_packets",
+	"tso_bytes",
+	"csum_offload_none",
+	"stopped",
+	"wake",
+	"dropped",
+	"nop"
+};
+
+struct mlx5e_sq_stats {
+	u64 packets;
+	u64 tso_packets;
+	u64 tso_bytes;
+	u64 csum_offload_none;
+	u64 stopped;
+	u64 wake;
+	u64 dropped;
+	u64 nop;
+#define NUM_SQ_STATS 8
+};
+
+struct mlx5e_stats {
+	struct mlx5e_vport_stats   vport;
+	struct mlx5e_pport_stats   pport;
+};
+
+struct mlx5e_params {
+	u8  log_sq_size;
+	u8  log_rq_size;
+	u16 num_channels;
+	u8  default_vlan_prio;
+	u8  num_tc;
+	u16 rx_cq_moderation_usec;
+	u16 rx_cq_moderation_pkts;
+	u16 tx_cq_moderation_usec;
+	u16 tx_cq_moderation_pkts;
+	u16 min_rx_wqes;
+	bool lro_en;
+	u32 lro_wqe_sz;
+	u16 tx_max_inline;
+	u8  rss_hfunc;
+	u8  toeplitz_hash_key[40];
+	u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE];
+};
+
+enum {
+	MLX5E_RQ_STATE_POST_WQES_ENABLE,
+};
+
+enum cq_flags {
+	MLX5E_CQ_HAS_CQES = 1,
+};
+
+struct mlx5e_cq {
+	/* data path - accessed per cqe */
+	struct mlx5_cqwq           wq;
+	unsigned long              flags;
+
+	/* data path - accessed per napi poll */
+	struct napi_struct        *napi;
+	struct mlx5_core_cq        mcq;
+	struct mlx5e_channel      *channel;
+	struct mlx5e_priv         *priv;
+
+	/* control */
+	struct mlx5_wq_ctrl        wq_ctrl;
+} ____cacheline_aligned_in_smp;
+
+struct mlx5e_rq {
+	/* data path */
+	struct mlx5_wq_ll      wq;
+	u32                    wqe_sz;
+	struct sk_buff       **skb;
+
+	struct device         *pdev;
+	struct net_device     *netdev;
+	struct mlx5e_rq_stats  stats;
+	struct mlx5e_cq        cq;
+
+	unsigned long          state;
+	int                    ix;
+
+	/* control */
+	struct mlx5_wq_ctrl    wq_ctrl;
+	u32                    rqn;
+	struct mlx5e_channel  *channel;
+	struct mlx5e_priv     *priv;
+} ____cacheline_aligned_in_smp;
+
+struct mlx5e_tx_skb_cb {
+	u32 num_bytes;
+	u8  num_wqebbs;
+	u8  num_dma;
+};
+
+#define MLX5E_TX_SKB_CB(__skb) ((struct mlx5e_tx_skb_cb *)__skb->cb)
+
+enum mlx5e_dma_map_type {
+	MLX5E_DMA_MAP_SINGLE,
+	MLX5E_DMA_MAP_PAGE
+};
+
+struct mlx5e_sq_dma {
+	dma_addr_t              addr;
+	u32                     size;
+	enum mlx5e_dma_map_type type;
+};
+
+enum {
+	MLX5E_SQ_STATE_WAKE_TXQ_ENABLE,
+};
+
+struct mlx5e_sq {
+	/* data path */
+
+	/* dirtied @completion */
+	u16                        cc;
+	u32                        dma_fifo_cc;
+
+	/* dirtied @xmit */
+	u16                        pc ____cacheline_aligned_in_smp;
+	u32                        dma_fifo_pc;
+	u16                        bf_offset;
+	u16                        prev_cc;
+	u8                         bf_budget;
+	struct mlx5e_sq_stats      stats;
+
+	struct mlx5e_cq            cq;
+
+	/* pointers to per packet info: write@xmit, read@completion */
+	struct sk_buff           **skb;
+	struct mlx5e_sq_dma       *dma_fifo;
+
+	/* read only */
+	struct mlx5_wq_cyc         wq;
+	u32                        dma_fifo_mask;
+	void __iomem              *uar_map;
+	void __iomem              *uar_bf_map;
+	struct netdev_queue       *txq;
+	u32                        sqn;
+	u16                        bf_buf_size;
+	u16                        max_inline;
+	u16                        edge;
+	struct device             *pdev;
+	__be32                     mkey_be;
+	unsigned long              state;
+
+	/* control path */
+	struct mlx5_wq_ctrl        wq_ctrl;
+	struct mlx5_uar            uar;
+	struct mlx5e_channel      *channel;
+	int                        tc;
+} ____cacheline_aligned_in_smp;
+
+static inline bool mlx5e_sq_has_room_for(struct mlx5e_sq *sq, u16 n)
+{
+	return (((sq->wq.sz_m1 & (sq->cc - sq->pc)) >= n) ||
+		(sq->cc  == sq->pc));
+}
+
+enum channel_flags {
+	MLX5E_CHANNEL_NAPI_SCHED = 1,
+};
+
+struct mlx5e_channel {
+	/* data path */
+	struct mlx5e_rq            rq;
+	struct mlx5e_sq            sq[MLX5E_MAX_NUM_TC];
+	struct napi_struct         napi;
+	struct device             *pdev;
+	struct net_device         *netdev;
+	__be32                     mkey_be;
+	u8                         num_tc;
+	unsigned long              flags;
+
+	/* control */
+	struct mlx5e_priv         *priv;
+	int                        ix;
+	int                        cpu;
+};
+
+enum mlx5e_traffic_types {
+	MLX5E_TT_IPV4_TCP,
+	MLX5E_TT_IPV6_TCP,
+	MLX5E_TT_IPV4_UDP,
+	MLX5E_TT_IPV6_UDP,
+	MLX5E_TT_IPV4_IPSEC_AH,
+	MLX5E_TT_IPV6_IPSEC_AH,
+	MLX5E_TT_IPV4_IPSEC_ESP,
+	MLX5E_TT_IPV6_IPSEC_ESP,
+	MLX5E_TT_IPV4,
+	MLX5E_TT_IPV6,
+	MLX5E_TT_ANY,
+	MLX5E_NUM_TT,
+};
+
+enum mlx5e_rqt_ix {
+	MLX5E_INDIRECTION_RQT,
+	MLX5E_SINGLE_RQ_RQT,
+	MLX5E_NUM_RQT,
+};
+
+struct mlx5e_eth_addr_info {
+	u8  addr[ETH_ALEN + 2];
+	u32 tt_vec;
+	u32 ft_ix[MLX5E_NUM_TT]; /* flow table index per traffic type */
+};
+
+#define MLX5E_ETH_ADDR_HASH_SIZE (1 << BITS_PER_BYTE)
+
+struct mlx5e_eth_addr_db {
+	struct hlist_head          netdev_uc[MLX5E_ETH_ADDR_HASH_SIZE];
+	struct hlist_head          netdev_mc[MLX5E_ETH_ADDR_HASH_SIZE];
+	struct mlx5e_eth_addr_info broadcast;
+	struct mlx5e_eth_addr_info allmulti;
+	struct mlx5e_eth_addr_info promisc;
+	bool                       broadcast_enabled;
+	bool                       allmulti_enabled;
+	bool                       promisc_enabled;
+};
+
+enum {
+	MLX5E_STATE_ASYNC_EVENTS_ENABLE,
+	MLX5E_STATE_OPENED,
+	MLX5E_STATE_DESTROYING,
+};
+
+struct mlx5e_vlan_db {
+	u32           active_vlans_ft_ix[VLAN_N_VID];
+	u32           untagged_rule_ft_ix;
+	u32           any_vlan_rule_ft_ix;
+	bool          filter_disabled;
+};
+
+struct mlx5e_flow_table {
+	void *vlan;
+	void *main;
+};
+
+struct mlx5e_priv {
+	/* priv data path fields - start */
+	int                        default_vlan_prio;
+	struct mlx5e_sq            **txq_to_sq_map;
+	int channeltc_to_txq_map[MLX5E_MAX_NUM_CHANNELS][MLX5E_MAX_NUM_TC];
+	/* priv data path fields - end */
+
+	unsigned long              state;
+	struct mutex               state_lock; /* Protects Interface state */
+	struct mlx5_uar            cq_uar;
+	u32                        pdn;
+	u32                        tdn;
+	struct mlx5_core_mr        mr;
+	struct mlx5e_rq            drop_rq;
+
+	struct mlx5e_channel     **channel;
+	u32                        tisn[MLX5E_MAX_NUM_TC];
+	u32                        rqtn[MLX5E_NUM_RQT];
+	u32                        tirn[MLX5E_NUM_TT];
+
+	struct mlx5e_flow_table    ft;
+	struct mlx5e_eth_addr_db   eth_addr;
+	struct mlx5e_vlan_db       vlan;
+
+	struct mlx5e_params        params;
+	spinlock_t                 async_events_spinlock; /* sync hw events */
+	struct work_struct         update_carrier_work;
+	struct work_struct         set_rx_mode_work;
+	struct delayed_work        update_stats_work;
+
+	struct mlx5_core_dev      *mdev;
+	struct net_device         *netdev;
+	struct mlx5e_stats         stats;
+};
+
+#define MLX5E_NET_IP_ALIGN 2
+
+struct mlx5e_tx_wqe {
+	struct mlx5_wqe_ctrl_seg ctrl;
+	struct mlx5_wqe_eth_seg  eth;
+};
+
+struct mlx5e_rx_wqe {
+	struct mlx5_wqe_srq_next_seg  next;
+	struct mlx5_wqe_data_seg      data;
+};
+
+enum mlx5e_link_mode {
+	MLX5E_1000BASE_CX_SGMII	 = 0,
+	MLX5E_1000BASE_KX	 = 1,
+	MLX5E_10GBASE_CX4	 = 2,
+	MLX5E_10GBASE_KX4	 = 3,
+	MLX5E_10GBASE_KR	 = 4,
+	MLX5E_20GBASE_KR2	 = 5,
+	MLX5E_40GBASE_CR4	 = 6,
+	MLX5E_40GBASE_KR4	 = 7,
+	MLX5E_56GBASE_R4	 = 8,
+	MLX5E_10GBASE_CR	 = 12,
+	MLX5E_10GBASE_SR	 = 13,
+	MLX5E_10GBASE_ER	 = 14,
+	MLX5E_40GBASE_SR4	 = 15,
+	MLX5E_40GBASE_LR4	 = 16,
+	MLX5E_100GBASE_CR4	 = 20,
+	MLX5E_100GBASE_SR4	 = 21,
+	MLX5E_100GBASE_KR4	 = 22,
+	MLX5E_100GBASE_LR4	 = 23,
+	MLX5E_100BASE_TX	 = 24,
+	MLX5E_100BASE_T		 = 25,
+	MLX5E_10GBASE_T		 = 26,
+	MLX5E_25GBASE_CR	 = 27,
+	MLX5E_25GBASE_KR	 = 28,
+	MLX5E_25GBASE_SR	 = 29,
+	MLX5E_50GBASE_CR2	 = 30,
+	MLX5E_50GBASE_KR2	 = 31,
+	MLX5E_LINK_MODES_NUMBER,
+};
+
+#define MLX5E_PROT_MASK(link_mode) (1 << link_mode)
+
+void mlx5e_send_nop(struct mlx5e_sq *sq, bool notify_hw);
+u16 mlx5e_select_queue(struct net_device *dev, struct sk_buff *skb,
+		       void *accel_priv, select_queue_fallback_t fallback);
+netdev_tx_t mlx5e_xmit(struct sk_buff *skb, struct net_device *dev);
+
+void mlx5e_completion_event(struct mlx5_core_cq *mcq);
+void mlx5e_cq_error_event(struct mlx5_core_cq *mcq, enum mlx5_event event);
+int mlx5e_napi_poll(struct napi_struct *napi, int budget);
+bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq);
+bool mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget);
+bool mlx5e_post_rx_wqes(struct mlx5e_rq *rq);
+struct mlx5_cqe64 *mlx5e_get_cqe(struct mlx5e_cq *cq);
+
+void mlx5e_update_stats(struct mlx5e_priv *priv);
+
+int mlx5e_create_flow_tables(struct mlx5e_priv *priv);
+void mlx5e_destroy_flow_tables(struct mlx5e_priv *priv);
+void mlx5e_init_eth_addr(struct mlx5e_priv *priv);
+void mlx5e_set_rx_mode_work(struct work_struct *work);
+
+int mlx5e_vlan_rx_add_vid(struct net_device *dev, __always_unused __be16 proto,
+			  u16 vid);
+int mlx5e_vlan_rx_kill_vid(struct net_device *dev, __always_unused __be16 proto,
+			   u16 vid);
+void mlx5e_enable_vlan_filter(struct mlx5e_priv *priv);
+void mlx5e_disable_vlan_filter(struct mlx5e_priv *priv);
+
+int mlx5e_redirect_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix);
+
+int mlx5e_open_locked(struct net_device *netdev);
+int mlx5e_close_locked(struct net_device *netdev);
+
+static inline void mlx5e_tx_notify_hw(struct mlx5e_sq *sq,
+				      struct mlx5e_tx_wqe *wqe, int bf_sz)
+{
+	u16 ofst = MLX5_BF_OFFSET + sq->bf_offset;
+
+	/* ensure wqe is visible to device before updating doorbell record */
+	dma_wmb();
+
+	*sq->wq.db = cpu_to_be32(sq->pc);
+
+	/* ensure doorbell record is visible to device before ringing the
+	 * doorbell
+	 */
+	wmb();
+
+	if (bf_sz) {
+		__iowrite64_copy(sq->uar_bf_map + ofst, &wqe->ctrl, bf_sz);
+
+		/* flush the write-combining mapped buffer */
+		wmb();
+
+	} else {
+		mlx5_write64((__be32 *)&wqe->ctrl, sq->uar_map + ofst, NULL);
+	}
+
+	sq->bf_offset ^= sq->bf_buf_size;
+}
+
+static inline void mlx5e_cq_arm(struct mlx5e_cq *cq)
+{
+	struct mlx5_core_cq *mcq;
+
+	mcq = &cq->mcq;
+	mlx5_cq_arm(mcq, MLX5_CQ_DB_REQ_NOT, mcq->uar->map, NULL, cq->wq.cc);
+}
+
+static inline int mlx5e_get_max_num_channels(struct mlx5_core_dev *mdev)
+{
+	return min_t(int, mdev->priv.eq_table.num_comp_vectors,
+		     MLX5E_MAX_NUM_CHANNELS);
+}
+
+extern const struct ethtool_ops mlx5e_ethtool_ops;
+u16 mlx5e_get_max_inline_cap(struct mlx5_core_dev *mdev);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
new file mode 100644
index 000000000..2e022e900
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -0,0 +1,881 @@
+/*
+ * Copyright (c) 2015, 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 "en.h"
+
+static void mlx5e_get_drvinfo(struct net_device *dev,
+			      struct ethtool_drvinfo *drvinfo)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	strlcpy(drvinfo->driver, DRIVER_NAME, sizeof(drvinfo->driver));
+	strlcpy(drvinfo->version, DRIVER_VERSION " (" DRIVER_RELDATE ")",
+		sizeof(drvinfo->version));
+	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
+		 "%d.%d.%d",
+		 fw_rev_maj(mdev), fw_rev_min(mdev), fw_rev_sub(mdev));
+	strlcpy(drvinfo->bus_info, pci_name(mdev->pdev),
+		sizeof(drvinfo->bus_info));
+}
+
+static const struct {
+	u32 supported;
+	u32 advertised;
+	u32 speed;
+} ptys2ethtool_table[MLX5E_LINK_MODES_NUMBER] = {
+	[MLX5E_1000BASE_CX_SGMII] = {
+		.supported  = SUPPORTED_1000baseKX_Full,
+		.advertised = ADVERTISED_1000baseKX_Full,
+		.speed      = 1000,
+	},
+	[MLX5E_1000BASE_KX] = {
+		.supported  = SUPPORTED_1000baseKX_Full,
+		.advertised = ADVERTISED_1000baseKX_Full,
+		.speed      = 1000,
+	},
+	[MLX5E_10GBASE_CX4] = {
+		.supported  = SUPPORTED_10000baseKX4_Full,
+		.advertised = ADVERTISED_10000baseKX4_Full,
+		.speed      = 10000,
+	},
+	[MLX5E_10GBASE_KX4] = {
+		.supported  = SUPPORTED_10000baseKX4_Full,
+		.advertised = ADVERTISED_10000baseKX4_Full,
+		.speed      = 10000,
+	},
+	[MLX5E_10GBASE_KR] = {
+		.supported  = SUPPORTED_10000baseKR_Full,
+		.advertised = ADVERTISED_10000baseKR_Full,
+		.speed      = 10000,
+	},
+	[MLX5E_20GBASE_KR2] = {
+		.supported  = SUPPORTED_20000baseKR2_Full,
+		.advertised = ADVERTISED_20000baseKR2_Full,
+		.speed      = 20000,
+	},
+	[MLX5E_40GBASE_CR4] = {
+		.supported  = SUPPORTED_40000baseCR4_Full,
+		.advertised = ADVERTISED_40000baseCR4_Full,
+		.speed      = 40000,
+	},
+	[MLX5E_40GBASE_KR4] = {
+		.supported  = SUPPORTED_40000baseKR4_Full,
+		.advertised = ADVERTISED_40000baseKR4_Full,
+		.speed      = 40000,
+	},
+	[MLX5E_56GBASE_R4] = {
+		.supported  = SUPPORTED_56000baseKR4_Full,
+		.advertised = ADVERTISED_56000baseKR4_Full,
+		.speed      = 56000,
+	},
+	[MLX5E_10GBASE_CR] = {
+		.supported  = SUPPORTED_10000baseKR_Full,
+		.advertised = ADVERTISED_10000baseKR_Full,
+		.speed      = 10000,
+	},
+	[MLX5E_10GBASE_SR] = {
+		.supported  = SUPPORTED_10000baseKR_Full,
+		.advertised = ADVERTISED_10000baseKR_Full,
+		.speed      = 10000,
+	},
+	[MLX5E_10GBASE_ER] = {
+		.supported  = SUPPORTED_10000baseKR_Full,
+		.advertised = ADVERTISED_10000baseKR_Full,
+		.speed      = 10000,
+	},
+	[MLX5E_40GBASE_SR4] = {
+		.supported  = SUPPORTED_40000baseSR4_Full,
+		.advertised = ADVERTISED_40000baseSR4_Full,
+		.speed      = 40000,
+	},
+	[MLX5E_40GBASE_LR4] = {
+		.supported  = SUPPORTED_40000baseLR4_Full,
+		.advertised = ADVERTISED_40000baseLR4_Full,
+		.speed      = 40000,
+	},
+	[MLX5E_100GBASE_CR4] = {
+		.speed      = 100000,
+	},
+	[MLX5E_100GBASE_SR4] = {
+		.speed      = 100000,
+	},
+	[MLX5E_100GBASE_KR4] = {
+		.speed      = 100000,
+	},
+	[MLX5E_100GBASE_LR4] = {
+		.speed      = 100000,
+	},
+	[MLX5E_100BASE_TX]   = {
+		.speed      = 100,
+	},
+	[MLX5E_100BASE_T]    = {
+		.supported  = SUPPORTED_100baseT_Full,
+		.advertised = ADVERTISED_100baseT_Full,
+		.speed      = 100,
+	},
+	[MLX5E_10GBASE_T]    = {
+		.supported  = SUPPORTED_10000baseT_Full,
+		.advertised = ADVERTISED_10000baseT_Full,
+		.speed      = 1000,
+	},
+	[MLX5E_25GBASE_CR]   = {
+		.speed      = 25000,
+	},
+	[MLX5E_25GBASE_KR]   = {
+		.speed      = 25000,
+	},
+	[MLX5E_25GBASE_SR]   = {
+		.speed      = 25000,
+	},
+	[MLX5E_50GBASE_CR2]  = {
+		.speed      = 50000,
+	},
+	[MLX5E_50GBASE_KR2]  = {
+		.speed      = 50000,
+	},
+};
+
+static int mlx5e_get_sset_count(struct net_device *dev, int sset)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	switch (sset) {
+	case ETH_SS_STATS:
+		return NUM_VPORT_COUNTERS + NUM_PPORT_COUNTERS +
+		       priv->params.num_channels * NUM_RQ_STATS +
+		       priv->params.num_channels * priv->params.num_tc *
+						   NUM_SQ_STATS;
+	/* fallthrough */
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static void mlx5e_get_strings(struct net_device *dev,
+			      uint32_t stringset, uint8_t *data)
+{
+	int i, j, tc, idx = 0;
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	switch (stringset) {
+	case ETH_SS_PRIV_FLAGS:
+		break;
+
+	case ETH_SS_TEST:
+		break;
+
+	case ETH_SS_STATS:
+		/* VPORT counters */
+		for (i = 0; i < NUM_VPORT_COUNTERS; i++)
+			strcpy(data + (idx++) * ETH_GSTRING_LEN,
+			       vport_strings[i]);
+
+		/* PPORT counters */
+		for (i = 0; i < NUM_PPORT_COUNTERS; i++)
+			strcpy(data + (idx++) * ETH_GSTRING_LEN,
+			       pport_strings[i]);
+
+		/* per channel counters */
+		for (i = 0; i < priv->params.num_channels; i++)
+			for (j = 0; j < NUM_RQ_STATS; j++)
+				sprintf(data + (idx++) * ETH_GSTRING_LEN,
+					"rx%d_%s", i, rq_stats_strings[j]);
+
+		for (i = 0; i < priv->params.num_channels; i++)
+			for (tc = 0; tc < priv->params.num_tc; tc++)
+				for (j = 0; j < NUM_SQ_STATS; j++)
+					sprintf(data +
+						(idx++) * ETH_GSTRING_LEN,
+						"tx%d_%d_%s", i, tc,
+						sq_stats_strings[j]);
+		break;
+	}
+}
+
+static void mlx5e_get_ethtool_stats(struct net_device *dev,
+				    struct ethtool_stats *stats, u64 *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int i, j, tc, idx = 0;
+
+	if (!data)
+		return;
+
+	mutex_lock(&priv->state_lock);
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+		mlx5e_update_stats(priv);
+	mutex_unlock(&priv->state_lock);
+
+	for (i = 0; i < NUM_VPORT_COUNTERS; i++)
+		data[idx++] = ((u64 *)&priv->stats.vport)[i];
+
+	for (i = 0; i < NUM_PPORT_COUNTERS; i++)
+		data[idx++] = be64_to_cpu(((__be64 *)&priv->stats.pport)[i]);
+
+	/* per channel counters */
+	for (i = 0; i < priv->params.num_channels; i++)
+		for (j = 0; j < NUM_RQ_STATS; j++)
+			data[idx++] = !test_bit(MLX5E_STATE_OPENED,
+						&priv->state) ? 0 :
+				       ((u64 *)&priv->channel[i]->rq.stats)[j];
+
+	for (i = 0; i < priv->params.num_channels; i++)
+		for (tc = 0; tc < priv->params.num_tc; tc++)
+			for (j = 0; j < NUM_SQ_STATS; j++)
+				data[idx++] = !test_bit(MLX5E_STATE_OPENED,
+							&priv->state) ? 0 :
+				((u64 *)&priv->channel[i]->sq[tc].stats)[j];
+}
+
+static void mlx5e_get_ringparam(struct net_device *dev,
+				struct ethtool_ringparam *param)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	param->rx_max_pending = 1 << MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE;
+	param->tx_max_pending = 1 << MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE;
+	param->rx_pending     = 1 << priv->params.log_rq_size;
+	param->tx_pending     = 1 << priv->params.log_sq_size;
+}
+
+static int mlx5e_set_ringparam(struct net_device *dev,
+			       struct ethtool_ringparam *param)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	bool was_opened;
+	u16 min_rx_wqes;
+	u8 log_rq_size;
+	u8 log_sq_size;
+	int err = 0;
+
+	if (param->rx_jumbo_pending) {
+		netdev_info(dev, "%s: rx_jumbo_pending not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+	if (param->rx_mini_pending) {
+		netdev_info(dev, "%s: rx_mini_pending not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+	if (param->rx_pending < (1 << MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE)) {
+		netdev_info(dev, "%s: rx_pending (%d) < min (%d)\n",
+			    __func__, param->rx_pending,
+			    1 << MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE);
+		return -EINVAL;
+	}
+	if (param->rx_pending > (1 << MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE)) {
+		netdev_info(dev, "%s: rx_pending (%d) > max (%d)\n",
+			    __func__, param->rx_pending,
+			    1 << MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE);
+		return -EINVAL;
+	}
+	if (param->tx_pending < (1 << MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE)) {
+		netdev_info(dev, "%s: tx_pending (%d) < min (%d)\n",
+			    __func__, param->tx_pending,
+			    1 << MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE);
+		return -EINVAL;
+	}
+	if (param->tx_pending > (1 << MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE)) {
+		netdev_info(dev, "%s: tx_pending (%d) > max (%d)\n",
+			    __func__, param->tx_pending,
+			    1 << MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE);
+		return -EINVAL;
+	}
+
+	log_rq_size = order_base_2(param->rx_pending);
+	log_sq_size = order_base_2(param->tx_pending);
+	min_rx_wqes = min_t(u16, param->rx_pending - 1,
+			    MLX5E_PARAMS_DEFAULT_MIN_RX_WQES);
+
+	if (log_rq_size == priv->params.log_rq_size &&
+	    log_sq_size == priv->params.log_sq_size &&
+	    min_rx_wqes == priv->params.min_rx_wqes)
+		return 0;
+
+	mutex_lock(&priv->state_lock);
+
+	was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+	if (was_opened)
+		mlx5e_close_locked(dev);
+
+	priv->params.log_rq_size = log_rq_size;
+	priv->params.log_sq_size = log_sq_size;
+	priv->params.min_rx_wqes = min_rx_wqes;
+
+	if (was_opened)
+		err = mlx5e_open_locked(dev);
+
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static void mlx5e_get_channels(struct net_device *dev,
+			       struct ethtool_channels *ch)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	ch->max_combined   = mlx5e_get_max_num_channels(priv->mdev);
+	ch->combined_count = priv->params.num_channels;
+}
+
+static int mlx5e_set_channels(struct net_device *dev,
+			      struct ethtool_channels *ch)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int ncv = mlx5e_get_max_num_channels(priv->mdev);
+	unsigned int count = ch->combined_count;
+	bool was_opened;
+	int err = 0;
+
+	if (!count) {
+		netdev_info(dev, "%s: combined_count=0 not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+	if (ch->rx_count || ch->tx_count) {
+		netdev_info(dev, "%s: separate rx/tx count not supported\n",
+			    __func__);
+		return -EINVAL;
+	}
+	if (count > ncv) {
+		netdev_info(dev, "%s: count (%d) > max (%d)\n",
+			    __func__, count, ncv);
+		return -EINVAL;
+	}
+
+	if (priv->params.num_channels == count)
+		return 0;
+
+	mutex_lock(&priv->state_lock);
+
+	was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+	if (was_opened)
+		mlx5e_close_locked(dev);
+
+	priv->params.num_channels = count;
+
+	if (was_opened)
+		err = mlx5e_open_locked(dev);
+
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_get_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	coal->rx_coalesce_usecs       = priv->params.rx_cq_moderation_usec;
+	coal->rx_max_coalesced_frames = priv->params.rx_cq_moderation_pkts;
+	coal->tx_coalesce_usecs       = priv->params.tx_cq_moderation_usec;
+	coal->tx_max_coalesced_frames = priv->params.tx_cq_moderation_pkts;
+
+	return 0;
+}
+
+static int mlx5e_set_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *coal)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_channel *c;
+	int tc;
+	int i;
+
+	priv->params.tx_cq_moderation_usec = coal->tx_coalesce_usecs;
+	priv->params.tx_cq_moderation_pkts = coal->tx_max_coalesced_frames;
+	priv->params.rx_cq_moderation_usec = coal->rx_coalesce_usecs;
+	priv->params.rx_cq_moderation_pkts = coal->rx_max_coalesced_frames;
+
+	for (i = 0; i < priv->params.num_channels; ++i) {
+		c = priv->channel[i];
+
+		for (tc = 0; tc < c->num_tc; tc++) {
+			mlx5_core_modify_cq_moderation(mdev,
+						&c->sq[tc].cq.mcq,
+						coal->tx_coalesce_usecs,
+						coal->tx_max_coalesced_frames);
+		}
+
+		mlx5_core_modify_cq_moderation(mdev, &c->rq.cq.mcq,
+					       coal->rx_coalesce_usecs,
+					       coal->rx_max_coalesced_frames);
+	}
+
+	return 0;
+}
+
+static u32 ptys2ethtool_supported_link(u32 eth_proto_cap)
+{
+	int i;
+	u32 supported_modes = 0;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (eth_proto_cap & MLX5E_PROT_MASK(i))
+			supported_modes |= ptys2ethtool_table[i].supported;
+	}
+	return supported_modes;
+}
+
+static u32 ptys2ethtool_adver_link(u32 eth_proto_cap)
+{
+	int i;
+	u32 advertising_modes = 0;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (eth_proto_cap & MLX5E_PROT_MASK(i))
+			advertising_modes |= ptys2ethtool_table[i].advertised;
+	}
+	return advertising_modes;
+}
+
+static u32 ptys2ethtool_supported_port(u32 eth_proto_cap)
+{
+	if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_10GBASE_CR)
+			   | MLX5E_PROT_MASK(MLX5E_10GBASE_SR)
+			   | MLX5E_PROT_MASK(MLX5E_40GBASE_CR4)
+			   | MLX5E_PROT_MASK(MLX5E_40GBASE_SR4)
+			   | MLX5E_PROT_MASK(MLX5E_100GBASE_SR4)
+			   | MLX5E_PROT_MASK(MLX5E_1000BASE_CX_SGMII))) {
+		return SUPPORTED_FIBRE;
+	}
+
+	if (eth_proto_cap & (MLX5E_PROT_MASK(MLX5E_100GBASE_KR4)
+			   | MLX5E_PROT_MASK(MLX5E_40GBASE_KR4)
+			   | MLX5E_PROT_MASK(MLX5E_10GBASE_KR)
+			   | MLX5E_PROT_MASK(MLX5E_10GBASE_KX4)
+			   | MLX5E_PROT_MASK(MLX5E_1000BASE_KX))) {
+		return SUPPORTED_Backplane;
+	}
+	return 0;
+}
+
+static void get_speed_duplex(struct net_device *netdev,
+			     u32 eth_proto_oper,
+			     struct ethtool_cmd *cmd)
+{
+	int i;
+	u32 speed = SPEED_UNKNOWN;
+	u8 duplex = DUPLEX_UNKNOWN;
+
+	if (!netif_carrier_ok(netdev))
+		goto out;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (eth_proto_oper & MLX5E_PROT_MASK(i)) {
+			speed = ptys2ethtool_table[i].speed;
+			duplex = DUPLEX_FULL;
+			break;
+		}
+	}
+out:
+	ethtool_cmd_speed_set(cmd, speed);
+	cmd->duplex = duplex;
+}
+
+static void get_supported(u32 eth_proto_cap, u32 *supported)
+{
+	*supported |= ptys2ethtool_supported_port(eth_proto_cap);
+	*supported |= ptys2ethtool_supported_link(eth_proto_cap);
+	*supported |= SUPPORTED_Pause | SUPPORTED_Asym_Pause;
+}
+
+static void get_advertising(u32 eth_proto_cap, u8 tx_pause,
+			    u8 rx_pause, u32 *advertising)
+{
+	*advertising |= ptys2ethtool_adver_link(eth_proto_cap);
+	*advertising |= tx_pause ? ADVERTISED_Pause : 0;
+	*advertising |= (tx_pause ^ rx_pause) ? ADVERTISED_Asym_Pause : 0;
+}
+
+static u8 get_connector_port(u32 eth_proto)
+{
+	if (eth_proto & (MLX5E_PROT_MASK(MLX5E_10GBASE_SR)
+			 | MLX5E_PROT_MASK(MLX5E_40GBASE_SR4)
+			 | MLX5E_PROT_MASK(MLX5E_100GBASE_SR4)
+			 | MLX5E_PROT_MASK(MLX5E_1000BASE_CX_SGMII))) {
+			return PORT_FIBRE;
+	}
+
+	if (eth_proto & (MLX5E_PROT_MASK(MLX5E_40GBASE_CR4)
+			 | MLX5E_PROT_MASK(MLX5E_10GBASE_CR)
+			 | MLX5E_PROT_MASK(MLX5E_100GBASE_CR4))) {
+			return PORT_DA;
+	}
+
+	if (eth_proto & (MLX5E_PROT_MASK(MLX5E_10GBASE_KX4)
+			 | MLX5E_PROT_MASK(MLX5E_10GBASE_KR)
+			 | MLX5E_PROT_MASK(MLX5E_40GBASE_KR4)
+			 | MLX5E_PROT_MASK(MLX5E_100GBASE_KR4))) {
+			return PORT_NONE;
+	}
+
+	return PORT_OTHER;
+}
+
+static void get_lp_advertising(u32 eth_proto_lp, u32 *lp_advertising)
+{
+	*lp_advertising = ptys2ethtool_adver_link(eth_proto_lp);
+}
+
+static int mlx5e_get_settings(struct net_device *netdev,
+			      struct ethtool_cmd *cmd)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	u32 eth_proto_cap;
+	u32 eth_proto_admin;
+	u32 eth_proto_lp;
+	u32 eth_proto_oper;
+	int err;
+
+	err = mlx5_query_port_ptys(mdev, out, sizeof(out), MLX5_PTYS_EN, 1);
+
+	if (err) {
+		netdev_err(netdev, "%s: query port ptys failed: %d\n",
+			   __func__, err);
+		goto err_query_ptys;
+	}
+
+	eth_proto_cap   = MLX5_GET(ptys_reg, out, eth_proto_capability);
+	eth_proto_admin = MLX5_GET(ptys_reg, out, eth_proto_admin);
+	eth_proto_oper  = MLX5_GET(ptys_reg, out, eth_proto_oper);
+	eth_proto_lp    = MLX5_GET(ptys_reg, out, eth_proto_lp_advertise);
+
+	cmd->supported   = 0;
+	cmd->advertising = 0;
+
+	get_supported(eth_proto_cap, &cmd->supported);
+	get_advertising(eth_proto_admin, 0, 0, &cmd->advertising);
+	get_speed_duplex(netdev, eth_proto_oper, cmd);
+
+	eth_proto_oper = eth_proto_oper ? eth_proto_oper : eth_proto_cap;
+
+	cmd->port = get_connector_port(eth_proto_oper);
+	get_lp_advertising(eth_proto_lp, &cmd->lp_advertising);
+
+	cmd->transceiver = XCVR_INTERNAL;
+
+err_query_ptys:
+	return err;
+}
+
+static u32 mlx5e_ethtool2ptys_adver_link(u32 link_modes)
+{
+	u32 i, ptys_modes = 0;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (ptys2ethtool_table[i].advertised & link_modes)
+			ptys_modes |= MLX5E_PROT_MASK(i);
+	}
+
+	return ptys_modes;
+}
+
+static u32 mlx5e_ethtool2ptys_speed_link(u32 speed)
+{
+	u32 i, speed_links = 0;
+
+	for (i = 0; i < MLX5E_LINK_MODES_NUMBER; ++i) {
+		if (ptys2ethtool_table[i].speed == speed)
+			speed_links |= MLX5E_PROT_MASK(i);
+	}
+
+	return speed_links;
+}
+
+static int mlx5e_set_settings(struct net_device *netdev,
+			      struct ethtool_cmd *cmd)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 link_modes;
+	u32 speed;
+	u32 eth_proto_cap, eth_proto_admin;
+	enum mlx5_port_status ps;
+	int err;
+
+	speed = ethtool_cmd_speed(cmd);
+
+	link_modes = cmd->autoneg == AUTONEG_ENABLE ?
+		mlx5e_ethtool2ptys_adver_link(cmd->advertising) :
+		mlx5e_ethtool2ptys_speed_link(speed);
+
+	err = mlx5_query_port_proto_cap(mdev, &eth_proto_cap, MLX5_PTYS_EN);
+	if (err) {
+		netdev_err(netdev, "%s: query port eth proto cap failed: %d\n",
+			   __func__, err);
+		goto out;
+	}
+
+	link_modes = link_modes & eth_proto_cap;
+	if (!link_modes) {
+		netdev_err(netdev, "%s: Not supported link mode(s) requested",
+			   __func__);
+		err = -EINVAL;
+		goto out;
+	}
+
+	err = mlx5_query_port_proto_admin(mdev, &eth_proto_admin, MLX5_PTYS_EN);
+	if (err) {
+		netdev_err(netdev, "%s: query port eth proto admin failed: %d\n",
+			   __func__, err);
+		goto out;
+	}
+
+	if (link_modes == eth_proto_admin)
+		goto out;
+
+	mlx5_query_port_admin_status(mdev, &ps);
+	if (ps == MLX5_PORT_UP)
+		mlx5_set_port_admin_status(mdev, MLX5_PORT_DOWN);
+	mlx5_set_port_proto(mdev, link_modes, MLX5_PTYS_EN);
+	if (ps == MLX5_PORT_UP)
+		mlx5_set_port_admin_status(mdev, MLX5_PORT_UP);
+
+out:
+	return err;
+}
+
+static u32 mlx5e_get_rxfh_key_size(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	return sizeof(priv->params.toeplitz_hash_key);
+}
+
+static u32 mlx5e_get_rxfh_indir_size(struct net_device *netdev)
+{
+	return MLX5E_INDIR_RQT_SIZE;
+}
+
+static int mlx5e_get_rxfh(struct net_device *netdev, u32 *indir, u8 *key,
+			  u8 *hfunc)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	if (indir)
+		memcpy(indir, priv->params.indirection_rqt,
+		       sizeof(priv->params.indirection_rqt));
+
+	if (key)
+		memcpy(key, priv->params.toeplitz_hash_key,
+		       sizeof(priv->params.toeplitz_hash_key));
+
+	if (hfunc)
+		*hfunc = priv->params.rss_hfunc;
+
+	return 0;
+}
+
+static int mlx5e_set_rxfh(struct net_device *dev, const u32 *indir,
+			  const u8 *key, const u8 hfunc)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	bool close_open;
+	int err = 0;
+
+	if ((hfunc != ETH_RSS_HASH_NO_CHANGE) &&
+	    (hfunc != ETH_RSS_HASH_XOR) &&
+	    (hfunc != ETH_RSS_HASH_TOP))
+		return -EINVAL;
+
+	mutex_lock(&priv->state_lock);
+
+	if (indir) {
+		memcpy(priv->params.indirection_rqt, indir,
+		       sizeof(priv->params.indirection_rqt));
+		mlx5e_redirect_rqt(priv, MLX5E_INDIRECTION_RQT);
+	}
+
+	close_open = (key || (hfunc != ETH_RSS_HASH_NO_CHANGE)) &&
+		     test_bit(MLX5E_STATE_OPENED, &priv->state);
+	if (close_open)
+		mlx5e_close_locked(dev);
+
+	if (key)
+		memcpy(priv->params.toeplitz_hash_key, key,
+		       sizeof(priv->params.toeplitz_hash_key));
+
+	if (hfunc != ETH_RSS_HASH_NO_CHANGE)
+		priv->params.rss_hfunc = hfunc;
+
+	if (close_open)
+		err = mlx5e_open_locked(priv->netdev);
+
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_get_rxnfc(struct net_device *netdev,
+			   struct ethtool_rxnfc *info, u32 *rule_locs)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err = 0;
+
+	switch (info->cmd) {
+	case ETHTOOL_GRXRINGS:
+		info->data = priv->params.num_channels;
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+static int mlx5e_get_tunable(struct net_device *dev,
+			     const struct ethtool_tunable *tuna,
+			     void *data)
+{
+	const struct mlx5e_priv *priv = netdev_priv(dev);
+	int err = 0;
+
+	switch (tuna->id) {
+	case ETHTOOL_TX_COPYBREAK:
+		*(u32 *)data = priv->params.tx_max_inline;
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static int mlx5e_set_tunable(struct net_device *dev,
+			     const struct ethtool_tunable *tuna,
+			     const void *data)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	bool was_opened;
+	u32 val;
+	int err = 0;
+
+	switch (tuna->id) {
+	case ETHTOOL_TX_COPYBREAK:
+		val = *(u32 *)data;
+		if (val > mlx5e_get_max_inline_cap(mdev)) {
+			err = -EINVAL;
+			break;
+		}
+
+		mutex_lock(&priv->state_lock);
+
+		was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+		if (was_opened)
+			mlx5e_close_locked(dev);
+
+		priv->params.tx_max_inline = val;
+
+		if (was_opened)
+			err = mlx5e_open_locked(dev);
+
+		mutex_unlock(&priv->state_lock);
+		break;
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	return err;
+}
+
+static void mlx5e_get_pauseparam(struct net_device *netdev,
+				 struct ethtool_pauseparam *pauseparam)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	err = mlx5_query_port_pause(mdev, &pauseparam->rx_pause,
+				    &pauseparam->tx_pause);
+	if (err) {
+		netdev_err(netdev, "%s: mlx5_query_port_pause failed:0x%x\n",
+			   __func__, err);
+	}
+}
+
+static int mlx5e_set_pauseparam(struct net_device *netdev,
+				struct ethtool_pauseparam *pauseparam)
+{
+	struct mlx5e_priv *priv    = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+	int err;
+
+	if (pauseparam->autoneg)
+		return -EINVAL;
+
+	err = mlx5_set_port_pause(mdev,
+				  pauseparam->rx_pause ? 1 : 0,
+				  pauseparam->tx_pause ? 1 : 0);
+	if (err) {
+		netdev_err(netdev, "%s: mlx5_set_port_pause failed:0x%x\n",
+			   __func__, err);
+	}
+
+	return err;
+}
+
+const struct ethtool_ops mlx5e_ethtool_ops = {
+	.get_drvinfo       = mlx5e_get_drvinfo,
+	.get_link          = ethtool_op_get_link,
+	.get_strings       = mlx5e_get_strings,
+	.get_sset_count    = mlx5e_get_sset_count,
+	.get_ethtool_stats = mlx5e_get_ethtool_stats,
+	.get_ringparam     = mlx5e_get_ringparam,
+	.set_ringparam     = mlx5e_set_ringparam,
+	.get_channels      = mlx5e_get_channels,
+	.set_channels      = mlx5e_set_channels,
+	.get_coalesce      = mlx5e_get_coalesce,
+	.set_coalesce      = mlx5e_set_coalesce,
+	.get_settings      = mlx5e_get_settings,
+	.set_settings      = mlx5e_set_settings,
+	.get_rxfh_key_size   = mlx5e_get_rxfh_key_size,
+	.get_rxfh_indir_size = mlx5e_get_rxfh_indir_size,
+	.get_rxfh          = mlx5e_get_rxfh,
+	.set_rxfh          = mlx5e_set_rxfh,
+	.get_rxnfc         = mlx5e_get_rxnfc,
+	.get_tunable       = mlx5e_get_tunable,
+	.set_tunable       = mlx5e_set_tunable,
+	.get_pauseparam    = mlx5e_get_pauseparam,
+	.set_pauseparam    = mlx5e_set_pauseparam,
+};
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_flow_table.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_flow_table.c
new file mode 100644
index 000000000..22d603f78
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_flow_table.c
@@ -0,0 +1,907 @@
+/*
+ * Copyright (c) 2015, 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 <linux/list.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/tcp.h>
+#include <linux/mlx5/flow_table.h>
+#include "en.h"
+
+enum {
+	MLX5E_FULLMATCH = 0,
+	MLX5E_ALLMULTI  = 1,
+	MLX5E_PROMISC   = 2,
+};
+
+enum {
+	MLX5E_UC        = 0,
+	MLX5E_MC_IPV4   = 1,
+	MLX5E_MC_IPV6   = 2,
+	MLX5E_MC_OTHER  = 3,
+};
+
+enum {
+	MLX5E_ACTION_NONE = 0,
+	MLX5E_ACTION_ADD  = 1,
+	MLX5E_ACTION_DEL  = 2,
+};
+
+struct mlx5e_eth_addr_hash_node {
+	struct hlist_node          hlist;
+	u8                         action;
+	struct mlx5e_eth_addr_info ai;
+};
+
+static inline int mlx5e_hash_eth_addr(u8 *addr)
+{
+	return addr[5];
+}
+
+static void mlx5e_add_eth_addr_to_hash(struct hlist_head *hash, u8 *addr)
+{
+	struct mlx5e_eth_addr_hash_node *hn;
+	int ix = mlx5e_hash_eth_addr(addr);
+	int found = 0;
+
+	hlist_for_each_entry(hn, &hash[ix], hlist)
+		if (ether_addr_equal_64bits(hn->ai.addr, addr)) {
+			found = 1;
+			break;
+		}
+
+	if (found) {
+		hn->action = MLX5E_ACTION_NONE;
+		return;
+	}
+
+	hn = kzalloc(sizeof(*hn), GFP_ATOMIC);
+	if (!hn)
+		return;
+
+	ether_addr_copy(hn->ai.addr, addr);
+	hn->action = MLX5E_ACTION_ADD;
+
+	hlist_add_head(&hn->hlist, &hash[ix]);
+}
+
+static void mlx5e_del_eth_addr_from_hash(struct mlx5e_eth_addr_hash_node *hn)
+{
+	hlist_del(&hn->hlist);
+	kfree(hn);
+}
+
+static void mlx5e_del_eth_addr_from_flow_table(struct mlx5e_priv *priv,
+					       struct mlx5e_eth_addr_info *ai)
+{
+	void *ft = priv->ft.main;
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV6_IPSEC_ESP))
+		mlx5_del_flow_table_entry(ft,
+					  ai->ft_ix[MLX5E_TT_IPV6_IPSEC_ESP]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV4_IPSEC_ESP))
+		mlx5_del_flow_table_entry(ft,
+					  ai->ft_ix[MLX5E_TT_IPV4_IPSEC_ESP]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV6_IPSEC_AH))
+		mlx5_del_flow_table_entry(ft,
+					  ai->ft_ix[MLX5E_TT_IPV6_IPSEC_AH]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV4_IPSEC_AH))
+		mlx5_del_flow_table_entry(ft,
+					  ai->ft_ix[MLX5E_TT_IPV4_IPSEC_AH]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV6_TCP))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_IPV6_TCP]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV4_TCP))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_IPV4_TCP]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV6_UDP))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_IPV6_UDP]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV4_UDP))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_IPV4_UDP]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV6))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_IPV6]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_IPV4))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_IPV4]);
+
+	if (ai->tt_vec & BIT(MLX5E_TT_ANY))
+		mlx5_del_flow_table_entry(ft, ai->ft_ix[MLX5E_TT_ANY]);
+}
+
+static int mlx5e_get_eth_addr_type(u8 *addr)
+{
+	if (is_unicast_ether_addr(addr))
+		return MLX5E_UC;
+
+	if ((addr[0] == 0x01) &&
+	    (addr[1] == 0x00) &&
+	    (addr[2] == 0x5e) &&
+	   !(addr[3] &  0x80))
+		return MLX5E_MC_IPV4;
+
+	if ((addr[0] == 0x33) &&
+	    (addr[1] == 0x33))
+		return MLX5E_MC_IPV6;
+
+	return MLX5E_MC_OTHER;
+}
+
+static u32 mlx5e_get_tt_vec(struct mlx5e_eth_addr_info *ai, int type)
+{
+	int eth_addr_type;
+	u32 ret;
+
+	switch (type) {
+	case MLX5E_FULLMATCH:
+		eth_addr_type = mlx5e_get_eth_addr_type(ai->addr);
+		switch (eth_addr_type) {
+		case MLX5E_UC:
+			ret =
+				BIT(MLX5E_TT_IPV4_TCP)       |
+				BIT(MLX5E_TT_IPV6_TCP)       |
+				BIT(MLX5E_TT_IPV4_UDP)       |
+				BIT(MLX5E_TT_IPV6_UDP)       |
+				BIT(MLX5E_TT_IPV4_IPSEC_AH)  |
+				BIT(MLX5E_TT_IPV6_IPSEC_AH)  |
+				BIT(MLX5E_TT_IPV4_IPSEC_ESP) |
+				BIT(MLX5E_TT_IPV6_IPSEC_ESP) |
+				BIT(MLX5E_TT_IPV4)           |
+				BIT(MLX5E_TT_IPV6)           |
+				BIT(MLX5E_TT_ANY)            |
+				0;
+			break;
+
+		case MLX5E_MC_IPV4:
+			ret =
+				BIT(MLX5E_TT_IPV4_UDP)       |
+				BIT(MLX5E_TT_IPV4)           |
+				0;
+			break;
+
+		case MLX5E_MC_IPV6:
+			ret =
+				BIT(MLX5E_TT_IPV6_UDP)       |
+				BIT(MLX5E_TT_IPV6)           |
+				0;
+			break;
+
+		case MLX5E_MC_OTHER:
+			ret =
+				BIT(MLX5E_TT_ANY)            |
+				0;
+			break;
+		}
+
+		break;
+
+	case MLX5E_ALLMULTI:
+		ret =
+			BIT(MLX5E_TT_IPV4_UDP) |
+			BIT(MLX5E_TT_IPV6_UDP) |
+			BIT(MLX5E_TT_IPV4)     |
+			BIT(MLX5E_TT_IPV6)     |
+			BIT(MLX5E_TT_ANY)      |
+			0;
+		break;
+
+	default: /* MLX5E_PROMISC */
+		ret =
+			BIT(MLX5E_TT_IPV4_TCP)       |
+			BIT(MLX5E_TT_IPV6_TCP)       |
+			BIT(MLX5E_TT_IPV4_UDP)       |
+			BIT(MLX5E_TT_IPV6_UDP)       |
+			BIT(MLX5E_TT_IPV4_IPSEC_AH)  |
+			BIT(MLX5E_TT_IPV6_IPSEC_AH)  |
+			BIT(MLX5E_TT_IPV4_IPSEC_ESP) |
+			BIT(MLX5E_TT_IPV6_IPSEC_ESP) |
+			BIT(MLX5E_TT_IPV4)           |
+			BIT(MLX5E_TT_IPV6)           |
+			BIT(MLX5E_TT_ANY)            |
+			0;
+		break;
+	}
+
+	return ret;
+}
+
+static int __mlx5e_add_eth_addr_rule(struct mlx5e_priv *priv,
+				     struct mlx5e_eth_addr_info *ai, int type,
+				     void *flow_context, void *match_criteria)
+{
+	u8 match_criteria_enable = 0;
+	void *match_value;
+	void *dest;
+	u8   *dmac;
+	u8   *match_criteria_dmac;
+	void *ft   = priv->ft.main;
+	u32  *tirn = priv->tirn;
+	u32  *ft_ix;
+	u32  tt_vec;
+	int  err;
+
+	match_value = MLX5_ADDR_OF(flow_context, flow_context, match_value);
+	dmac = MLX5_ADDR_OF(fte_match_param, match_value,
+			    outer_headers.dmac_47_16);
+	match_criteria_dmac = MLX5_ADDR_OF(fte_match_param, match_criteria,
+					   outer_headers.dmac_47_16);
+	dest = MLX5_ADDR_OF(flow_context, flow_context, destination);
+
+	MLX5_SET(flow_context, flow_context, action,
+		 MLX5_FLOW_CONTEXT_ACTION_FWD_DEST);
+	MLX5_SET(flow_context, flow_context, destination_list_size, 1);
+	MLX5_SET(dest_format_struct, dest, destination_type,
+		 MLX5_FLOW_CONTEXT_DEST_TYPE_TIR);
+
+	switch (type) {
+	case MLX5E_FULLMATCH:
+		match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		memset(match_criteria_dmac, 0xff, ETH_ALEN);
+		ether_addr_copy(dmac, ai->addr);
+		break;
+
+	case MLX5E_ALLMULTI:
+		match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+		match_criteria_dmac[0] = 0x01;
+		dmac[0] = 0x01;
+		break;
+
+	case MLX5E_PROMISC:
+		break;
+	}
+
+	tt_vec = mlx5e_get_tt_vec(ai, type);
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_ANY];
+	if (tt_vec & BIT(MLX5E_TT_ANY)) {
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_ANY]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_ANY);
+	}
+
+	match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria,
+			 outer_headers.ethertype);
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV4];
+	if (tt_vec & BIT(MLX5E_TT_IPV4)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IP);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV4]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV4);
+	}
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV6];
+	if (tt_vec & BIT(MLX5E_TT_IPV6)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IPV6);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV6]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV6);
+	}
+
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria,
+			 outer_headers.ip_protocol);
+	MLX5_SET(fte_match_param, match_value, outer_headers.ip_protocol,
+		 IPPROTO_UDP);
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV4_UDP];
+	if (tt_vec & BIT(MLX5E_TT_IPV4_UDP)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IP);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV4_UDP]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV4_UDP);
+	}
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV6_UDP];
+	if (tt_vec & BIT(MLX5E_TT_IPV6_UDP)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IPV6);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV6_UDP]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV6_UDP);
+	}
+
+	MLX5_SET(fte_match_param, match_value, outer_headers.ip_protocol,
+		 IPPROTO_TCP);
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV4_TCP];
+	if (tt_vec & BIT(MLX5E_TT_IPV4_TCP)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IP);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV4_TCP]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV4_TCP);
+	}
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV6_TCP];
+	if (tt_vec & BIT(MLX5E_TT_IPV6_TCP)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IPV6);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV6_TCP]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV6_TCP);
+	}
+
+	MLX5_SET(fte_match_param, match_value, outer_headers.ip_protocol,
+		 IPPROTO_AH);
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV4_IPSEC_AH];
+	if (tt_vec & BIT(MLX5E_TT_IPV4_IPSEC_AH)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IP);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV4_IPSEC_AH]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV4_IPSEC_AH);
+	}
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV6_IPSEC_AH];
+	if (tt_vec & BIT(MLX5E_TT_IPV6_IPSEC_AH)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IPV6);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV6_IPSEC_AH]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV6_IPSEC_AH);
+	}
+
+	MLX5_SET(fte_match_param, match_value, outer_headers.ip_protocol,
+		 IPPROTO_ESP);
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV4_IPSEC_ESP];
+	if (tt_vec & BIT(MLX5E_TT_IPV4_IPSEC_ESP)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IP);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV4_IPSEC_ESP]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV4_IPSEC_ESP);
+	}
+
+	ft_ix = &ai->ft_ix[MLX5E_TT_IPV6_IPSEC_ESP];
+	if (tt_vec & BIT(MLX5E_TT_IPV6_IPSEC_ESP)) {
+		MLX5_SET(fte_match_param, match_value, outer_headers.ethertype,
+			 ETH_P_IPV6);
+		MLX5_SET(dest_format_struct, dest, destination_id,
+			 tirn[MLX5E_TT_IPV6_IPSEC_ESP]);
+		err = mlx5_add_flow_table_entry(ft, match_criteria_enable,
+						match_criteria, flow_context,
+						ft_ix);
+		if (err)
+			goto err_del_ai;
+
+		ai->tt_vec |= BIT(MLX5E_TT_IPV6_IPSEC_ESP);
+	}
+
+	return 0;
+
+err_del_ai:
+	mlx5e_del_eth_addr_from_flow_table(priv, ai);
+
+	return err;
+}
+
+static int mlx5e_add_eth_addr_rule(struct mlx5e_priv *priv,
+				   struct mlx5e_eth_addr_info *ai, int type)
+{
+	u32 *flow_context;
+	u32 *match_criteria;
+	int err;
+
+	flow_context   = mlx5_vzalloc(MLX5_ST_SZ_BYTES(flow_context) +
+				      MLX5_ST_SZ_BYTES(dest_format_struct));
+	match_criteria = mlx5_vzalloc(MLX5_ST_SZ_BYTES(fte_match_param));
+	if (!flow_context || !match_criteria) {
+		netdev_err(priv->netdev, "%s: alloc failed\n", __func__);
+		err = -ENOMEM;
+		goto add_eth_addr_rule_out;
+	}
+
+	err = __mlx5e_add_eth_addr_rule(priv, ai, type, flow_context,
+					match_criteria);
+	if (err)
+		netdev_err(priv->netdev, "%s: failed\n", __func__);
+
+add_eth_addr_rule_out:
+	kvfree(match_criteria);
+	kvfree(flow_context);
+	return err;
+}
+
+enum mlx5e_vlan_rule_type {
+	MLX5E_VLAN_RULE_TYPE_UNTAGGED,
+	MLX5E_VLAN_RULE_TYPE_ANY_VID,
+	MLX5E_VLAN_RULE_TYPE_MATCH_VID,
+};
+
+static int mlx5e_add_vlan_rule(struct mlx5e_priv *priv,
+			       enum mlx5e_vlan_rule_type rule_type, u16 vid)
+{
+	u8 match_criteria_enable = 0;
+	u32 *flow_context;
+	void *match_value;
+	void *dest;
+	u32 *match_criteria;
+	u32 *ft_ix;
+	int err;
+
+	flow_context   = mlx5_vzalloc(MLX5_ST_SZ_BYTES(flow_context) +
+				      MLX5_ST_SZ_BYTES(dest_format_struct));
+	match_criteria = mlx5_vzalloc(MLX5_ST_SZ_BYTES(fte_match_param));
+	if (!flow_context || !match_criteria) {
+		netdev_err(priv->netdev, "%s: alloc failed\n", __func__);
+		err = -ENOMEM;
+		goto add_vlan_rule_out;
+	}
+	match_value = MLX5_ADDR_OF(flow_context, flow_context, match_value);
+	dest = MLX5_ADDR_OF(flow_context, flow_context, destination);
+
+	MLX5_SET(flow_context, flow_context, action,
+		 MLX5_FLOW_CONTEXT_ACTION_FWD_DEST);
+	MLX5_SET(flow_context, flow_context, destination_list_size, 1);
+	MLX5_SET(dest_format_struct, dest, destination_type,
+		 MLX5_FLOW_CONTEXT_DEST_TYPE_FLOW_TABLE);
+	MLX5_SET(dest_format_struct, dest, destination_id,
+		 mlx5_get_flow_table_id(priv->ft.main));
+
+	match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, match_criteria,
+			 outer_headers.vlan_tag);
+
+	switch (rule_type) {
+	case MLX5E_VLAN_RULE_TYPE_UNTAGGED:
+		ft_ix = &priv->vlan.untagged_rule_ft_ix;
+		break;
+	case MLX5E_VLAN_RULE_TYPE_ANY_VID:
+		ft_ix = &priv->vlan.any_vlan_rule_ft_ix;
+		MLX5_SET(fte_match_param, match_value, outer_headers.vlan_tag,
+			 1);
+		break;
+	default: /* MLX5E_VLAN_RULE_TYPE_MATCH_VID */
+		ft_ix = &priv->vlan.active_vlans_ft_ix[vid];
+		MLX5_SET(fte_match_param, match_value, outer_headers.vlan_tag,
+			 1);
+		MLX5_SET_TO_ONES(fte_match_param, match_criteria,
+				 outer_headers.first_vid);
+		MLX5_SET(fte_match_param, match_value, outer_headers.first_vid,
+			 vid);
+		break;
+	}
+
+	err = mlx5_add_flow_table_entry(priv->ft.vlan, match_criteria_enable,
+					match_criteria, flow_context, ft_ix);
+	if (err)
+		netdev_err(priv->netdev, "%s: failed\n", __func__);
+
+add_vlan_rule_out:
+	kvfree(match_criteria);
+	kvfree(flow_context);
+	return err;
+}
+
+static void mlx5e_del_vlan_rule(struct mlx5e_priv *priv,
+				enum mlx5e_vlan_rule_type rule_type, u16 vid)
+{
+	switch (rule_type) {
+	case MLX5E_VLAN_RULE_TYPE_UNTAGGED:
+		mlx5_del_flow_table_entry(priv->ft.vlan,
+					  priv->vlan.untagged_rule_ft_ix);
+		break;
+	case MLX5E_VLAN_RULE_TYPE_ANY_VID:
+		mlx5_del_flow_table_entry(priv->ft.vlan,
+					  priv->vlan.any_vlan_rule_ft_ix);
+		break;
+	case MLX5E_VLAN_RULE_TYPE_MATCH_VID:
+		mlx5_del_flow_table_entry(priv->ft.vlan,
+					  priv->vlan.active_vlans_ft_ix[vid]);
+		break;
+	}
+}
+
+void mlx5e_enable_vlan_filter(struct mlx5e_priv *priv)
+{
+	if (!priv->vlan.filter_disabled)
+		return;
+
+	priv->vlan.filter_disabled = false;
+	if (priv->netdev->flags & IFF_PROMISC)
+		return;
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_VID, 0);
+}
+
+void mlx5e_disable_vlan_filter(struct mlx5e_priv *priv)
+{
+	if (priv->vlan.filter_disabled)
+		return;
+
+	priv->vlan.filter_disabled = true;
+	if (priv->netdev->flags & IFF_PROMISC)
+		return;
+	mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_VID, 0);
+}
+
+int mlx5e_vlan_rx_add_vid(struct net_device *dev, __always_unused __be16 proto,
+			  u16 vid)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	return mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_VID, vid);
+}
+
+int mlx5e_vlan_rx_kill_vid(struct net_device *dev, __always_unused __be16 proto,
+			   u16 vid)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_MATCH_VID, vid);
+
+	return 0;
+}
+
+#define mlx5e_for_each_hash_node(hn, tmp, hash, i) \
+	for (i = 0; i < MLX5E_ETH_ADDR_HASH_SIZE; i++) \
+		hlist_for_each_entry_safe(hn, tmp, &hash[i], hlist)
+
+static void mlx5e_execute_action(struct mlx5e_priv *priv,
+				 struct mlx5e_eth_addr_hash_node *hn)
+{
+	switch (hn->action) {
+	case MLX5E_ACTION_ADD:
+		mlx5e_add_eth_addr_rule(priv, &hn->ai, MLX5E_FULLMATCH);
+		hn->action = MLX5E_ACTION_NONE;
+		break;
+
+	case MLX5E_ACTION_DEL:
+		mlx5e_del_eth_addr_from_flow_table(priv, &hn->ai);
+		mlx5e_del_eth_addr_from_hash(hn);
+		break;
+	}
+}
+
+static void mlx5e_sync_netdev_addr(struct mlx5e_priv *priv)
+{
+	struct net_device *netdev = priv->netdev;
+	struct netdev_hw_addr *ha;
+
+	netif_addr_lock_bh(netdev);
+
+	mlx5e_add_eth_addr_to_hash(priv->eth_addr.netdev_uc,
+				   priv->netdev->dev_addr);
+
+	netdev_for_each_uc_addr(ha, netdev)
+		mlx5e_add_eth_addr_to_hash(priv->eth_addr.netdev_uc, ha->addr);
+
+	netdev_for_each_mc_addr(ha, netdev)
+		mlx5e_add_eth_addr_to_hash(priv->eth_addr.netdev_mc, ha->addr);
+
+	netif_addr_unlock_bh(netdev);
+}
+
+static void mlx5e_apply_netdev_addr(struct mlx5e_priv *priv)
+{
+	struct mlx5e_eth_addr_hash_node *hn;
+	struct hlist_node *tmp;
+	int i;
+
+	mlx5e_for_each_hash_node(hn, tmp, priv->eth_addr.netdev_uc, i)
+		mlx5e_execute_action(priv, hn);
+
+	mlx5e_for_each_hash_node(hn, tmp, priv->eth_addr.netdev_mc, i)
+		mlx5e_execute_action(priv, hn);
+}
+
+static void mlx5e_handle_netdev_addr(struct mlx5e_priv *priv)
+{
+	struct mlx5e_eth_addr_hash_node *hn;
+	struct hlist_node *tmp;
+	int i;
+
+	mlx5e_for_each_hash_node(hn, tmp, priv->eth_addr.netdev_uc, i)
+		hn->action = MLX5E_ACTION_DEL;
+	mlx5e_for_each_hash_node(hn, tmp, priv->eth_addr.netdev_mc, i)
+		hn->action = MLX5E_ACTION_DEL;
+
+	if (!test_bit(MLX5E_STATE_DESTROYING, &priv->state))
+		mlx5e_sync_netdev_addr(priv);
+
+	mlx5e_apply_netdev_addr(priv);
+}
+
+void mlx5e_set_rx_mode_work(struct work_struct *work)
+{
+	struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv,
+					       set_rx_mode_work);
+
+	struct mlx5e_eth_addr_db *ea = &priv->eth_addr;
+	struct net_device *ndev = priv->netdev;
+
+	bool rx_mode_enable   = !test_bit(MLX5E_STATE_DESTROYING, &priv->state);
+	bool promisc_enabled   = rx_mode_enable && (ndev->flags & IFF_PROMISC);
+	bool allmulti_enabled  = rx_mode_enable && (ndev->flags & IFF_ALLMULTI);
+	bool broadcast_enabled = rx_mode_enable;
+
+	bool enable_promisc    = !ea->promisc_enabled   &&  promisc_enabled;
+	bool disable_promisc   =  ea->promisc_enabled   && !promisc_enabled;
+	bool enable_allmulti   = !ea->allmulti_enabled  &&  allmulti_enabled;
+	bool disable_allmulti  =  ea->allmulti_enabled  && !allmulti_enabled;
+	bool enable_broadcast  = !ea->broadcast_enabled &&  broadcast_enabled;
+	bool disable_broadcast =  ea->broadcast_enabled && !broadcast_enabled;
+
+	if (enable_promisc) {
+		mlx5e_add_eth_addr_rule(priv, &ea->promisc, MLX5E_PROMISC);
+		if (!priv->vlan.filter_disabled)
+			mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_VID,
+					    0);
+	}
+	if (enable_allmulti)
+		mlx5e_add_eth_addr_rule(priv, &ea->allmulti, MLX5E_ALLMULTI);
+	if (enable_broadcast)
+		mlx5e_add_eth_addr_rule(priv, &ea->broadcast, MLX5E_FULLMATCH);
+
+	mlx5e_handle_netdev_addr(priv);
+
+	if (disable_broadcast)
+		mlx5e_del_eth_addr_from_flow_table(priv, &ea->broadcast);
+	if (disable_allmulti)
+		mlx5e_del_eth_addr_from_flow_table(priv, &ea->allmulti);
+	if (disable_promisc) {
+		if (!priv->vlan.filter_disabled)
+			mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_ANY_VID,
+					    0);
+		mlx5e_del_eth_addr_from_flow_table(priv, &ea->promisc);
+	}
+
+	ea->promisc_enabled   = promisc_enabled;
+	ea->allmulti_enabled  = allmulti_enabled;
+	ea->broadcast_enabled = broadcast_enabled;
+}
+
+void mlx5e_init_eth_addr(struct mlx5e_priv *priv)
+{
+	ether_addr_copy(priv->eth_addr.broadcast.addr, priv->netdev->broadcast);
+}
+
+static int mlx5e_create_main_flow_table(struct mlx5e_priv *priv)
+{
+	struct mlx5_flow_table_group *g;
+	u8 *dmac;
+
+	g = kcalloc(9, sizeof(*g), GFP_KERNEL);
+	if (!g)
+		return -ENOMEM;
+
+	g[0].log_sz = 3;
+	g[0].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, g[0].match_criteria,
+			 outer_headers.ethertype);
+	MLX5_SET_TO_ONES(fte_match_param, g[0].match_criteria,
+			 outer_headers.ip_protocol);
+
+	g[1].log_sz = 1;
+	g[1].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, g[1].match_criteria,
+			 outer_headers.ethertype);
+
+	g[2].log_sz = 0;
+
+	g[3].log_sz = 14;
+	g[3].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	dmac = MLX5_ADDR_OF(fte_match_param, g[3].match_criteria,
+			    outer_headers.dmac_47_16);
+	memset(dmac, 0xff, ETH_ALEN);
+	MLX5_SET_TO_ONES(fte_match_param, g[3].match_criteria,
+			 outer_headers.ethertype);
+	MLX5_SET_TO_ONES(fte_match_param, g[3].match_criteria,
+			 outer_headers.ip_protocol);
+
+	g[4].log_sz = 13;
+	g[4].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	dmac = MLX5_ADDR_OF(fte_match_param, g[4].match_criteria,
+			    outer_headers.dmac_47_16);
+	memset(dmac, 0xff, ETH_ALEN);
+	MLX5_SET_TO_ONES(fte_match_param, g[4].match_criteria,
+			 outer_headers.ethertype);
+
+	g[5].log_sz = 11;
+	g[5].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	dmac = MLX5_ADDR_OF(fte_match_param, g[5].match_criteria,
+			    outer_headers.dmac_47_16);
+	memset(dmac, 0xff, ETH_ALEN);
+
+	g[6].log_sz = 2;
+	g[6].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	dmac = MLX5_ADDR_OF(fte_match_param, g[6].match_criteria,
+			    outer_headers.dmac_47_16);
+	dmac[0] = 0x01;
+	MLX5_SET_TO_ONES(fte_match_param, g[6].match_criteria,
+			 outer_headers.ethertype);
+	MLX5_SET_TO_ONES(fte_match_param, g[6].match_criteria,
+			 outer_headers.ip_protocol);
+
+	g[7].log_sz = 1;
+	g[7].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	dmac = MLX5_ADDR_OF(fte_match_param, g[7].match_criteria,
+			    outer_headers.dmac_47_16);
+	dmac[0] = 0x01;
+	MLX5_SET_TO_ONES(fte_match_param, g[7].match_criteria,
+			 outer_headers.ethertype);
+
+	g[8].log_sz = 0;
+	g[8].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	dmac = MLX5_ADDR_OF(fte_match_param, g[8].match_criteria,
+			    outer_headers.dmac_47_16);
+	dmac[0] = 0x01;
+	priv->ft.main = mlx5_create_flow_table(priv->mdev, 1,
+					       MLX5_FLOW_TABLE_TYPE_NIC_RCV,
+					       9, g);
+	kfree(g);
+
+	return priv->ft.main ? 0 : -ENOMEM;
+}
+
+static void mlx5e_destroy_main_flow_table(struct mlx5e_priv *priv)
+{
+	mlx5_destroy_flow_table(priv->ft.main);
+}
+
+static int mlx5e_create_vlan_flow_table(struct mlx5e_priv *priv)
+{
+	struct mlx5_flow_table_group *g;
+
+	g = kcalloc(2, sizeof(*g), GFP_KERNEL);
+	if (!g)
+		return -ENOMEM;
+
+	g[0].log_sz = 12;
+	g[0].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, g[0].match_criteria,
+			 outer_headers.vlan_tag);
+	MLX5_SET_TO_ONES(fte_match_param, g[0].match_criteria,
+			 outer_headers.first_vid);
+
+	/* untagged + any vlan id */
+	g[1].log_sz = 1;
+	g[1].match_criteria_enable = MLX5_MATCH_OUTER_HEADERS;
+	MLX5_SET_TO_ONES(fte_match_param, g[1].match_criteria,
+			 outer_headers.vlan_tag);
+
+	priv->ft.vlan = mlx5_create_flow_table(priv->mdev, 0,
+					       MLX5_FLOW_TABLE_TYPE_NIC_RCV,
+					       2, g);
+
+	kfree(g);
+	return priv->ft.vlan ? 0 : -ENOMEM;
+}
+
+static void mlx5e_destroy_vlan_flow_table(struct mlx5e_priv *priv)
+{
+	mlx5_destroy_flow_table(priv->ft.vlan);
+}
+
+int mlx5e_create_flow_tables(struct mlx5e_priv *priv)
+{
+	int err;
+
+	err = mlx5e_create_main_flow_table(priv);
+	if (err)
+		return err;
+
+	err = mlx5e_create_vlan_flow_table(priv);
+	if (err)
+		goto err_destroy_main_flow_table;
+
+	err = mlx5e_add_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_UNTAGGED, 0);
+	if (err)
+		goto err_destroy_vlan_flow_table;
+
+	return 0;
+
+err_destroy_vlan_flow_table:
+	mlx5e_destroy_vlan_flow_table(priv);
+
+err_destroy_main_flow_table:
+	mlx5e_destroy_main_flow_table(priv);
+
+	return err;
+}
+
+void mlx5e_destroy_flow_tables(struct mlx5e_priv *priv)
+{
+	mlx5e_del_vlan_rule(priv, MLX5E_VLAN_RULE_TYPE_UNTAGGED, 0);
+	mlx5e_destroy_vlan_flow_table(priv);
+	mlx5e_destroy_main_flow_table(priv);
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
new file mode 100644
index 000000000..1203d892e
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -0,0 +1,2276 @@
+/*
+ * Copyright (c) 2015, 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 <linux/mlx5/flow_table.h>
+#include "en.h"
+
+struct mlx5e_rq_param {
+	u32                        rqc[MLX5_ST_SZ_DW(rqc)];
+	struct mlx5_wq_param       wq;
+};
+
+struct mlx5e_sq_param {
+	u32                        sqc[MLX5_ST_SZ_DW(sqc)];
+	struct mlx5_wq_param       wq;
+	u16                        max_inline;
+};
+
+struct mlx5e_cq_param {
+	u32                        cqc[MLX5_ST_SZ_DW(cqc)];
+	struct mlx5_wq_param       wq;
+	u16                        eq_ix;
+};
+
+struct mlx5e_channel_param {
+	struct mlx5e_rq_param      rq;
+	struct mlx5e_sq_param      sq;
+	struct mlx5e_cq_param      rx_cq;
+	struct mlx5e_cq_param      tx_cq;
+};
+
+static void mlx5e_update_carrier(struct mlx5e_priv *priv)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u8 port_state;
+
+	port_state = mlx5_query_vport_state(mdev,
+		MLX5_QUERY_VPORT_STATE_IN_OP_MOD_VNIC_VPORT);
+
+	if (port_state == VPORT_STATE_UP)
+		netif_carrier_on(priv->netdev);
+	else
+		netif_carrier_off(priv->netdev);
+}
+
+static void mlx5e_update_carrier_work(struct work_struct *work)
+{
+	struct mlx5e_priv *priv = container_of(work, struct mlx5e_priv,
+					       update_carrier_work);
+
+	mutex_lock(&priv->state_lock);
+	if (test_bit(MLX5E_STATE_OPENED, &priv->state))
+		mlx5e_update_carrier(priv);
+	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)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5e_vport_stats *s = &priv->stats.vport;
+	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;
+	int i, j;
+
+	out = mlx5_vzalloc(outlen);
+	if (!out)
+		return;
+
+	/* Collect firts the SW counters and then HW for consistency */
+	s->tso_packets		= 0;
+	s->tso_bytes		= 0;
+	s->tx_queue_stopped	= 0;
+	s->tx_queue_wake	= 0;
+	s->tx_queue_dropped	= 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;
+	for (i = 0; i < priv->params.num_channels; i++) {
+		rq_stats = &priv->channel[i]->rq.stats;
+
+		s->lro_packets	+= rq_stats->lro_packets;
+		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_wqe_err   += rq_stats->wqe_err;
+
+		for (j = 0; j < priv->params.num_tc; j++) {
+			sq_stats = &priv->channel[i]->sq[j].stats;
+
+			s->tso_packets		+= sq_stats->tso_packets;
+			s->tso_bytes		+= sq_stats->tso_bytes;
+			s->tx_queue_stopped	+= sq_stats->stopped;
+			s->tx_queue_wake	+= sq_stats->wake;
+			s->tx_queue_dropped	+= sq_stats->dropped;
+			tx_offload_none		+= sq_stats->csum_offload_none;
+		}
+	}
+
+	/* HW counters */
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(query_vport_counter_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VPORT_COUNTER);
+	MLX5_SET(query_vport_counter_in, in, op_mod, 0);
+	MLX5_SET(query_vport_counter_in, in, other_vport, 0);
+
+	memset(out, 0, outlen);
+
+	if (mlx5_cmd_exec(mdev, in, sizeof(in), out, outlen))
+		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);
+
+	s->rx_packets =
+		s->rx_unicast_packets +
+		s->rx_multicast_packets +
+		s->rx_broadcast_packets;
+	s->rx_bytes =
+		s->rx_unicast_bytes +
+		s->rx_multicast_bytes +
+		s->rx_broadcast_bytes;
+	s->tx_packets =
+		s->tx_unicast_packets +
+		s->tx_multicast_packets +
+		s->tx_broadcast_packets;
+	s->tx_bytes =
+		s->tx_unicast_bytes +
+		s->tx_multicast_bytes +
+		s->tx_broadcast_bytes;
+
+	/* Update calculated offload counters */
+	s->tx_csum_offload = s->tx_packets - tx_offload_none;
+	s->rx_csum_good    = s->rx_packets - s->rx_csum_none -
+			       s->rx_csum_sw;
+
+	mlx5e_update_pport_counters(priv);
+free_out:
+	kvfree(out);
+}
+
+static void mlx5e_update_stats_work(struct work_struct *work)
+{
+	struct delayed_work *dwork = to_delayed_work(work);
+	struct mlx5e_priv *priv = container_of(dwork, struct mlx5e_priv,
+					       update_stats_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));
+	}
+	mutex_unlock(&priv->state_lock);
+}
+
+static void __mlx5e_async_event(struct mlx5e_priv *priv,
+				enum mlx5_dev_event event)
+{
+	switch (event) {
+	case MLX5_DEV_EVENT_PORT_UP:
+	case MLX5_DEV_EVENT_PORT_DOWN:
+		schedule_work(&priv->update_carrier_work);
+		break;
+
+	default:
+		break;
+	}
+}
+
+static void mlx5e_async_event(struct mlx5_core_dev *mdev, void *vpriv,
+			      enum mlx5_dev_event event, unsigned long param)
+{
+	struct mlx5e_priv *priv = vpriv;
+
+	spin_lock(&priv->async_events_spinlock);
+	if (test_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLE, &priv->state))
+		__mlx5e_async_event(priv, event);
+	spin_unlock(&priv->async_events_spinlock);
+}
+
+static void mlx5e_enable_async_events(struct mlx5e_priv *priv)
+{
+	set_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLE, &priv->state);
+}
+
+static void mlx5e_disable_async_events(struct mlx5e_priv *priv)
+{
+	spin_lock_irq(&priv->async_events_spinlock);
+	clear_bit(MLX5E_STATE_ASYNC_EVENTS_ENABLE, &priv->state);
+	spin_unlock_irq(&priv->async_events_spinlock);
+}
+
+#define MLX5E_HW2SW_MTU(hwmtu) (hwmtu - (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN))
+#define MLX5E_SW2HW_MTU(swmtu) (swmtu + (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN))
+
+static int mlx5e_create_rq(struct mlx5e_channel *c,
+			   struct mlx5e_rq_param *param,
+			   struct mlx5e_rq *rq)
+{
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqc = param->rqc;
+	void *rqc_wq = MLX5_ADDR_OF(rqc, rqc, wq);
+	int wq_sz;
+	int err;
+	int i;
+
+	param->wq.db_numa_node = cpu_to_node(c->cpu);
+
+	err = mlx5_wq_ll_create(mdev, &param->wq, rqc_wq, &rq->wq,
+				&rq->wq_ctrl);
+	if (err)
+		return err;
+
+	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);
+
+	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);
+	}
+
+	rq->pdev    = c->pdev;
+	rq->netdev  = c->netdev;
+	rq->channel = c;
+	rq->ix      = c->ix;
+	rq->priv    = c->priv;
+
+	return 0;
+
+err_rq_wq_destroy:
+	mlx5_wq_destroy(&rq->wq_ctrl);
+
+	return err;
+}
+
+static void mlx5e_destroy_rq(struct mlx5e_rq *rq)
+{
+	kfree(rq->skb);
+	mlx5_wq_destroy(&rq->wq_ctrl);
+}
+
+static int mlx5e_enable_rq(struct mlx5e_rq *rq, struct mlx5e_rq_param *param)
+{
+	struct mlx5e_priv *priv = rq->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *rqc;
+	void *wq;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(create_rq_in) +
+		sizeof(u64) * rq->wq_ctrl.buf.npages;
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	rqc = MLX5_ADDR_OF(create_rq_in, in, ctx);
+	wq  = MLX5_ADDR_OF(rqc, rqc, wq);
+
+	memcpy(rqc, param->rqc, sizeof(param->rqc));
+
+	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(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);
+
+	mlx5_fill_page_array(&rq->wq_ctrl.buf,
+			     (__be64 *)MLX5_ADDR_OF(wq, wq, pas));
+
+	err = mlx5_core_create_rq(mdev, in, inlen, &rq->rqn);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_modify_rq(struct mlx5e_rq *rq, int curr_state, int next_state)
+{
+	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, curr_state);
+	MLX5_SET(rqc, rqc, state, next_state);
+
+	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);
+}
+
+static int mlx5e_wait_for_min_rx_wqes(struct mlx5e_rq *rq)
+{
+	unsigned long exp_time = jiffies + msecs_to_jiffies(20000);
+	struct mlx5e_channel *c = rq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_wq_ll *wq = &rq->wq;
+
+	while (time_before(jiffies, exp_time)) {
+		if (wq->cur_sz >= priv->params.min_rx_wqes)
+			return 0;
+
+		msleep(20);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int mlx5e_open_rq(struct mlx5e_channel *c,
+			 struct mlx5e_rq_param *param,
+			 struct mlx5e_rq *rq)
+{
+	int err;
+
+	err = mlx5e_create_rq(c, param, rq);
+	if (err)
+		return err;
+
+	err = mlx5e_enable_rq(rq, param);
+	if (err)
+		goto err_destroy_rq;
+
+	err = mlx5e_modify_rq(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() */
+
+	return 0;
+
+err_disable_rq:
+	mlx5e_disable_rq(rq);
+err_destroy_rq:
+	mlx5e_destroy_rq(rq);
+
+	return err;
+}
+
+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);
+	while (!mlx5_wq_ll_is_empty(&rq->wq))
+		msleep(20);
+
+	/* avoid destroying rq before mlx5e_poll_rx_cq() is done with it */
+	napi_synchronize(&rq->channel->napi);
+
+	mlx5e_disable_rq(rq);
+	mlx5e_destroy_rq(rq);
+}
+
+static void mlx5e_free_sq_db(struct mlx5e_sq *sq)
+{
+	kfree(sq->dma_fifo);
+	kfree(sq->skb);
+}
+
+static int mlx5e_alloc_sq_db(struct mlx5e_sq *sq, int numa)
+{
+	int wq_sz = mlx5_wq_cyc_get_size(&sq->wq);
+	int df_sz = wq_sz * MLX5_SEND_WQEBB_NUM_DS;
+
+	sq->skb = kzalloc_node(wq_sz * sizeof(*sq->skb), GFP_KERNEL, numa);
+	sq->dma_fifo = kzalloc_node(df_sz * sizeof(*sq->dma_fifo), GFP_KERNEL,
+				    numa);
+
+	if (!sq->skb || !sq->dma_fifo) {
+		mlx5e_free_sq_db(sq);
+		return -ENOMEM;
+	}
+
+	sq->dma_fifo_mask = df_sz - 1;
+
+	return 0;
+}
+
+static int mlx5e_create_sq(struct mlx5e_channel *c,
+			   int tc,
+			   struct mlx5e_sq_param *param,
+			   struct mlx5e_sq *sq)
+{
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	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);
+	if (err)
+		return err;
+
+	param->wq.db_numa_node = cpu_to_node(c->cpu);
+
+	err = mlx5_wq_cyc_create(mdev, &param->wq, sqc_wq, &sq->wq,
+				 &sq->wq_ctrl);
+	if (err)
+		goto err_unmap_free_uar;
+
+	sq->wq.db       = &sq->wq.db[MLX5_SND_DBR];
+	sq->uar_map     = sq->uar.map;
+	sq->uar_bf_map  = sq->uar.bf_map;
+	sq->bf_buf_size = (1 << MLX5_CAP_GEN(mdev, log_bf_reg_size)) / 2;
+	sq->max_inline  = param->max_inline;
+
+	err = mlx5e_alloc_sq_db(sq, cpu_to_node(c->cpu));
+	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);
+
+	sq->pdev      = c->pdev;
+	sq->mkey_be   = c->mkey_be;
+	sq->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_sq_wq_destroy:
+	mlx5_wq_destroy(&sq->wq_ctrl);
+
+err_unmap_free_uar:
+	mlx5_unmap_free_uar(mdev, &sq->uar);
+
+	return err;
+}
+
+static void mlx5e_destroy_sq(struct mlx5e_sq *sq)
+{
+	struct mlx5e_channel *c = sq->channel;
+	struct mlx5e_priv *priv = c->priv;
+
+	mlx5e_free_sq_db(sq);
+	mlx5_wq_destroy(&sq->wq_ctrl);
+	mlx5_unmap_free_uar(priv->mdev, &sq->uar);
+}
+
+static int mlx5e_enable_sq(struct mlx5e_sq *sq, struct mlx5e_sq_param *param)
+{
+	struct mlx5e_channel *c = sq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *sqc;
+	void *wq;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(create_sq_in) +
+		sizeof(u64) * sq->wq_ctrl.buf.npages;
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	sqc = MLX5_ADDR_OF(create_sq_in, in, ctx);
+	wq = MLX5_ADDR_OF(sqc, sqc, wq);
+
+	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, state,		MLX5_SQC_STATE_RST);
+	MLX5_SET(sqc,  sqc, tis_lst_sz,		1);
+	MLX5_SET(sqc,  sqc, flush_in_error_en,	1);
+
+	MLX5_SET(wq,   wq, wq_type,       MLX5_WQ_TYPE_CYCLIC);
+	MLX5_SET(wq,   wq, uar_page,      sq->uar.index);
+	MLX5_SET(wq,   wq, log_wq_pg_sz,  sq->wq_ctrl.buf.page_shift -
+					  MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET64(wq, wq, dbr_addr,      sq->wq_ctrl.db.dma);
+
+	mlx5_fill_page_array(&sq->wq_ctrl.buf,
+			     (__be64 *)MLX5_ADDR_OF(wq, wq, pas));
+
+	err = mlx5_core_create_sq(mdev, in, inlen, &sq->sqn);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_modify_sq(struct mlx5e_sq *sq, int curr_state, int next_state)
+{
+	struct mlx5e_channel *c = sq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *sqc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_sq_in);
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	sqc = MLX5_ADDR_OF(modify_sq_in, in, ctx);
+
+	MLX5_SET(modify_sq_in, in, sq_state, curr_state);
+	MLX5_SET(sqc, sqc, state, next_state);
+
+	err = mlx5_core_modify_sq(mdev, sq->sqn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5e_disable_sq(struct mlx5e_sq *sq)
+{
+	struct mlx5e_channel *c = sq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	mlx5_core_destroy_sq(mdev, sq->sqn);
+}
+
+static int mlx5e_open_sq(struct mlx5e_channel *c,
+			 int tc,
+			 struct mlx5e_sq_param *param,
+			 struct mlx5e_sq *sq)
+{
+	int err;
+
+	err = mlx5e_create_sq(c, tc, param, sq);
+	if (err)
+		return err;
+
+	err = mlx5e_enable_sq(sq, param);
+	if (err)
+		goto err_destroy_sq;
+
+	err = mlx5e_modify_sq(sq, MLX5_SQC_STATE_RST, MLX5_SQC_STATE_RDY);
+	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);
+
+	return 0;
+
+err_disable_sq:
+	mlx5e_disable_sq(sq);
+err_destroy_sq:
+	mlx5e_destroy_sq(sq);
+
+	return err;
+}
+
+static inline void netif_tx_disable_queue(struct netdev_queue *txq)
+{
+	__netif_tx_lock_bh(txq);
+	netif_tx_stop_queue(txq);
+	__netif_tx_unlock_bh(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);
+
+	/* 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);
+	while (sq->cc != sq->pc) /* wait till sq is empty */
+		msleep(20);
+
+	/* avoid destroying sq before mlx5e_poll_tx_cq() is done with it */
+	napi_synchronize(&sq->channel->napi);
+
+	mlx5e_disable_sq(sq);
+	mlx5e_destroy_sq(sq);
+}
+
+static int mlx5e_create_cq(struct mlx5e_channel *c,
+			   struct mlx5e_cq_param *param,
+			   struct mlx5e_cq *cq)
+{
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_core_cq *mcq = &cq->mcq;
+	int eqn_not_used;
+	unsigned int irqn;
+	int err;
+	u32 i;
+
+	param->wq.buf_numa_node = cpu_to_node(c->cpu);
+	param->wq.db_numa_node  = cpu_to_node(c->cpu);
+	param->eq_ix   = c->ix;
+
+	err = mlx5_cqwq_create(mdev, &param->wq, param->cqc, &cq->wq,
+			       &cq->wq_ctrl);
+	if (err)
+		return err;
+
+	mlx5_vector2eqn(mdev, param->eq_ix, &eqn_not_used, &irqn);
+
+	cq->napi        = &c->napi;
+
+	mcq->cqe_sz     = 64;
+	mcq->set_ci_db  = cq->wq_ctrl.db.db;
+	mcq->arm_db     = cq->wq_ctrl.db.db + 1;
+	*mcq->set_ci_db = 0;
+	*mcq->arm_db    = 0;
+	mcq->vector     = param->eq_ix;
+	mcq->comp       = mlx5e_completion_event;
+	mcq->event      = mlx5e_cq_error_event;
+	mcq->irqn       = irqn;
+	mcq->uar        = &priv->cq_uar;
+
+	for (i = 0; i < mlx5_cqwq_get_size(&cq->wq); i++) {
+		struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(&cq->wq, i);
+
+		cqe->op_own = 0xf1;
+	}
+
+	cq->channel = c;
+	cq->priv = priv;
+
+	return 0;
+}
+
+static void mlx5e_destroy_cq(struct mlx5e_cq *cq)
+{
+	mlx5_wq_destroy(&cq->wq_ctrl);
+}
+
+static int mlx5e_enable_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param)
+{
+	struct mlx5e_priv *priv = cq->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_core_cq *mcq = &cq->mcq;
+
+	void *in;
+	void *cqc;
+	int inlen;
+	unsigned int irqn_not_used;
+	int eqn;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(create_cq_in) +
+		sizeof(u64) * cq->wq_ctrl.buf.npages;
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	cqc = MLX5_ADDR_OF(create_cq_in, in, cq_context);
+
+	memcpy(cqc, param->cqc, sizeof(param->cqc));
+
+	mlx5_fill_page_array(&cq->wq_ctrl.buf,
+			     (__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas));
+
+	mlx5_vector2eqn(mdev, param->eq_ix, &eqn, &irqn_not_used);
+
+	MLX5_SET(cqc,   cqc, c_eqn,         eqn);
+	MLX5_SET(cqc,   cqc, uar_page,      mcq->uar->index);
+	MLX5_SET(cqc,   cqc, log_page_size, cq->wq_ctrl.buf.page_shift -
+					    MLX5_ADAPTER_PAGE_SHIFT);
+	MLX5_SET64(cqc, cqc, dbr_addr,      cq->wq_ctrl.db.dma);
+
+	err = mlx5_core_create_cq(mdev, mcq, in, inlen);
+
+	kvfree(in);
+
+	if (err)
+		return err;
+
+	mlx5e_cq_arm(cq);
+
+	return 0;
+}
+
+static void mlx5e_disable_cq(struct mlx5e_cq *cq)
+{
+	struct mlx5e_priv *priv = cq->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	mlx5_core_destroy_cq(mdev, &cq->mcq);
+}
+
+static int mlx5e_open_cq(struct mlx5e_channel *c,
+			 struct mlx5e_cq_param *param,
+			 struct mlx5e_cq *cq,
+			 u16 moderation_usecs,
+			 u16 moderation_frames)
+{
+	int err;
+	struct mlx5e_priv *priv = c->priv;
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	err = mlx5e_create_cq(c, param, cq);
+	if (err)
+		return err;
+
+	err = mlx5e_enable_cq(cq, param);
+	if (err)
+		goto err_destroy_cq;
+
+	err = mlx5_core_modify_cq_moderation(mdev, &cq->mcq,
+					     moderation_usecs,
+					     moderation_frames);
+	if (err)
+		goto err_destroy_cq;
+
+	return 0;
+
+err_destroy_cq:
+	mlx5e_destroy_cq(cq);
+
+	return err;
+}
+
+static void mlx5e_close_cq(struct mlx5e_cq *cq)
+{
+	mlx5e_disable_cq(cq);
+	mlx5e_destroy_cq(cq);
+}
+
+static int mlx5e_get_cpu(struct mlx5e_priv *priv, int ix)
+{
+	return cpumask_first(priv->mdev->priv.irq_info[ix].mask);
+}
+
+static int mlx5e_open_tx_cqs(struct mlx5e_channel *c,
+			     struct mlx5e_channel_param *cparam)
+{
+	struct mlx5e_priv *priv = c->priv;
+	int err;
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++) {
+		err = mlx5e_open_cq(c, &cparam->tx_cq, &c->sq[tc].cq,
+				    priv->params.tx_cq_moderation_usec,
+				    priv->params.tx_cq_moderation_pkts);
+		if (err)
+			goto err_close_tx_cqs;
+	}
+
+	return 0;
+
+err_close_tx_cqs:
+	for (tc--; tc >= 0; tc--)
+		mlx5e_close_cq(&c->sq[tc].cq);
+
+	return err;
+}
+
+static void mlx5e_close_tx_cqs(struct mlx5e_channel *c)
+{
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++)
+		mlx5e_close_cq(&c->sq[tc].cq);
+}
+
+static int mlx5e_open_sqs(struct mlx5e_channel *c,
+			  struct mlx5e_channel_param *cparam)
+{
+	int err;
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++) {
+		err = mlx5e_open_sq(c, tc, &cparam->sq, &c->sq[tc]);
+		if (err)
+			goto err_close_sqs;
+	}
+
+	return 0;
+
+err_close_sqs:
+	for (tc--; tc >= 0; tc--)
+		mlx5e_close_sq(&c->sq[tc]);
+
+	return err;
+}
+
+static void mlx5e_close_sqs(struct mlx5e_channel *c)
+{
+	int tc;
+
+	for (tc = 0; tc < c->num_tc; tc++)
+		mlx5e_close_sq(&c->sq[tc]);
+}
+
+static void mlx5e_build_channeltc_to_txq_map(struct mlx5e_priv *priv, int ix)
+{
+	int i;
+
+	for (i = 0; i < MLX5E_MAX_NUM_TC; i++)
+		priv->channeltc_to_txq_map[ix][i] =
+			ix + i * priv->params.num_channels;
+}
+
+static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
+			      struct mlx5e_channel_param *cparam,
+			      struct mlx5e_channel **cp)
+{
+	struct net_device *netdev = priv->netdev;
+	int cpu = mlx5e_get_cpu(priv, ix);
+	struct mlx5e_channel *c;
+	int err;
+
+	c = kzalloc_node(sizeof(*c), GFP_KERNEL, cpu_to_node(cpu));
+	if (!c)
+		return -ENOMEM;
+
+	c->priv     = priv;
+	c->ix       = ix;
+	c->cpu      = cpu;
+	c->pdev     = &priv->mdev->pdev->dev;
+	c->netdev   = priv->netdev;
+	c->mkey_be  = cpu_to_be32(priv->mr.key);
+	c->num_tc   = priv->params.num_tc;
+
+	mlx5e_build_channeltc_to_txq_map(priv, ix);
+
+	netif_napi_add(netdev, &c->napi, mlx5e_napi_poll, 64);
+
+	err = mlx5e_open_tx_cqs(c, cparam);
+	if (err)
+		goto err_napi_del;
+
+	err = mlx5e_open_cq(c, &cparam->rx_cq, &c->rq.cq,
+			    priv->params.rx_cq_moderation_usec,
+			    priv->params.rx_cq_moderation_pkts);
+	if (err)
+		goto err_close_tx_cqs;
+
+	napi_enable(&c->napi);
+
+	err = mlx5e_open_sqs(c, cparam);
+	if (err)
+		goto err_disable_napi;
+
+	err = mlx5e_open_rq(c, &cparam->rq, &c->rq);
+	if (err)
+		goto err_close_sqs;
+
+	netif_set_xps_queue(netdev, get_cpu_mask(c->cpu), ix);
+	*cp = c;
+
+	return 0;
+
+err_close_sqs:
+	mlx5e_close_sqs(c);
+
+err_disable_napi:
+	napi_disable(&c->napi);
+	mlx5e_close_cq(&c->rq.cq);
+
+err_close_tx_cqs:
+	mlx5e_close_tx_cqs(c);
+
+err_napi_del:
+	netif_napi_del(&c->napi);
+	kfree(c);
+
+	return err;
+}
+
+static void mlx5e_close_channel(struct mlx5e_channel *c)
+{
+	mlx5e_close_rq(&c->rq);
+	mlx5e_close_sqs(c);
+	napi_disable(&c->napi);
+	mlx5e_close_cq(&c->rq.cq);
+	mlx5e_close_tx_cqs(c);
+	netif_napi_del(&c->napi);
+	kfree(c);
+}
+
+static void mlx5e_build_rq_param(struct mlx5e_priv *priv,
+				 struct mlx5e_rq_param *param)
+{
+	void *rqc = param->rqc;
+	void *wq = MLX5_ADDR_OF(rqc, rqc, wq);
+
+	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);
+
+	param->wq.buf_numa_node = dev_to_node(&priv->mdev->pdev->dev);
+	param->wq.linear = 1;
+}
+
+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);
+
+	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);
+	param->max_inline = priv->params.tx_max_inline;
+}
+
+static void mlx5e_build_common_cq_param(struct mlx5e_priv *priv,
+					struct mlx5e_cq_param *param)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, uar_page, priv->cq_uar.index);
+}
+
+static void mlx5e_build_rx_cq_param(struct mlx5e_priv *priv,
+				    struct mlx5e_cq_param *param)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, log_cq_size,  priv->params.log_rq_size);
+
+	mlx5e_build_common_cq_param(priv, param);
+}
+
+static void mlx5e_build_tx_cq_param(struct mlx5e_priv *priv,
+				    struct mlx5e_cq_param *param)
+{
+	void *cqc = param->cqc;
+
+	MLX5_SET(cqc, cqc, log_cq_size,  priv->params.log_sq_size);
+
+	mlx5e_build_common_cq_param(priv, param);
+}
+
+static void mlx5e_build_channel_param(struct mlx5e_priv *priv,
+				      struct mlx5e_channel_param *cparam)
+{
+	memset(cparam, 0, sizeof(*cparam));
+
+	mlx5e_build_rq_param(priv, &cparam->rq);
+	mlx5e_build_sq_param(priv, &cparam->sq);
+	mlx5e_build_rx_cq_param(priv, &cparam->rx_cq);
+	mlx5e_build_tx_cq_param(priv, &cparam->tx_cq);
+}
+
+static int mlx5e_open_channels(struct mlx5e_priv *priv)
+{
+	struct mlx5e_channel_param cparam;
+	int nch = priv->params.num_channels;
+	int err = -ENOMEM;
+	int i;
+	int j;
+
+	priv->channel = kcalloc(nch, sizeof(struct mlx5e_channel *),
+				GFP_KERNEL);
+
+	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)
+		goto err_free_txq_to_sq_map;
+
+	mlx5e_build_channel_param(priv, &cparam);
+	for (i = 0; i < nch; i++) {
+		err = mlx5e_open_channel(priv, i, &cparam, &priv->channel[i]);
+		if (err)
+			goto err_close_channels;
+	}
+
+	for (j = 0; j < nch; j++) {
+		err = mlx5e_wait_for_min_rx_wqes(&priv->channel[j]->rq);
+		if (err)
+			goto err_close_channels;
+	}
+
+	return 0;
+
+err_close_channels:
+	for (i--; i >= 0; i--)
+		mlx5e_close_channel(priv->channel[i]);
+
+err_free_txq_to_sq_map:
+	kfree(priv->txq_to_sq_map);
+	kfree(priv->channel);
+
+	return err;
+}
+
+static void mlx5e_close_channels(struct mlx5e_priv *priv)
+{
+	int i;
+
+	for (i = 0; i < priv->params.num_channels; i++)
+		mlx5e_close_channel(priv->channel[i]);
+
+	kfree(priv->txq_to_sq_map);
+	kfree(priv->channel);
+}
+
+static int mlx5e_rx_hash_fn(int hfunc)
+{
+	return (hfunc == ETH_RSS_HASH_TOP) ?
+	       MLX5_RX_HASH_FN_TOEPLITZ :
+	       MLX5_RX_HASH_FN_INVERTED_XOR8;
+}
+
+static int mlx5e_bits_invert(unsigned long a, int size)
+{
+	int inv = 0;
+	int i;
+
+	for (i = 0; i < size; i++)
+		inv |= (test_bit(size - i - 1, &a) ? 1 : 0) << i;
+
+	return inv;
+}
+
+static void mlx5e_fill_indir_rqt_rqns(struct mlx5e_priv *priv, void *rqtc)
+{
+	int i;
+
+	for (i = 0; i < MLX5E_INDIR_RQT_SIZE; i++) {
+		int ix = i;
+
+		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];
+		ix = ix % priv->params.num_channels;
+		MLX5_SET(rqtc, rqtc, rq_num[i],
+			 test_bit(MLX5E_STATE_OPENED, &priv->state) ?
+			 priv->channel[ix]->rq.rqn :
+			 priv->drop_rq.rqn);
+	}
+}
+
+static void mlx5e_fill_rqt_rqns(struct mlx5e_priv *priv, void *rqtc,
+				enum mlx5e_rqt_ix rqt_ix)
+{
+
+	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;
+	}
+}
+
+static int mlx5e_create_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
+{
+	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;
+
+	inlen = MLX5_ST_SZ_BYTES(create_rqt_in) + sizeof(u32) * sz;
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	rqtc = MLX5_ADDR_OF(create_rqt_in, in, rqt_context);
+
+	MLX5_SET(rqtc, rqtc, rqt_actual_size, sz);
+	MLX5_SET(rqtc, rqtc, rqt_max_size, sz);
+
+	mlx5e_fill_rqt_rqns(priv, rqtc, rqt_ix);
+
+	err = mlx5_core_create_rqt(mdev, in, inlen, &priv->rqtn[rqt_ix]);
+
+	kvfree(in);
+
+	return err;
+}
+
+int mlx5e_redirect_rqt(struct mlx5e_priv *priv, enum mlx5e_rqt_ix rqt_ix)
+{
+	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;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_rqt_in) + sizeof(u32) * sz;
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	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);
+
+	MLX5_SET(modify_rqt_in, in, bitmask.rqn_list, 1);
+
+	err = mlx5_core_modify_rqt(mdev, priv->rqtn[rqt_ix], 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);
+}
+
+static void mlx5e_build_tir_ctx_lro(void *tirc, struct mlx5e_priv *priv)
+{
+	if (!priv->params.lro_en)
+		return;
+
+#define ROUGH_MAX_L2_L3_HDR_SZ 256
+
+	MLX5_SET(tirc, tirc, lro_enable_mask,
+		 MLX5_TIRC_LRO_ENABLE_MASK_IPV4_LRO |
+		 MLX5_TIRC_LRO_ENABLE_MASK_IPV6_LRO);
+	MLX5_SET(tirc, tirc, lro_max_ip_payload_size,
+		 (priv->params.lro_wqe_sz -
+		  ROUGH_MAX_L2_L3_HDR_SZ) >> 8);
+	MLX5_SET(tirc, tirc, lro_timeout_period_usecs,
+		 MLX5_CAP_ETH(priv->mdev,
+			      lro_timer_supported_periods[2]));
+}
+
+static int mlx5e_modify_tir_lro(struct mlx5e_priv *priv, int tt)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	void *in;
+	void *tirc;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_tir_in, in, bitmask.lro, 1);
+	tirc = MLX5_ADDR_OF(modify_tir_in, in, ctx);
+
+	mlx5e_build_tir_ctx_lro(tirc, priv);
+
+	err = mlx5_core_modify_tir(mdev, priv->tirn[tt], in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_refresh_tir_self_loopback_enable(struct mlx5_core_dev *mdev,
+						  u32 tirn)
+{
+	void *in;
+	int inlen;
+	int err;
+
+	inlen = MLX5_ST_SZ_BYTES(modify_tir_in);
+	in = mlx5_vzalloc(inlen);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(modify_tir_in, in, bitmask.self_lb_en, 1);
+
+	err = mlx5_core_modify_tir(mdev, tirn, in, inlen);
+
+	kvfree(in);
+
+	return err;
+}
+
+static int mlx5e_refresh_tirs_self_loopback_enable(struct mlx5e_priv *priv)
+{
+	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;
+	}
+
+	return 0;
+}
+
+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;
+	int err;
+
+	err = mlx5_set_port_mtu(mdev, MLX5E_SW2HW_MTU(netdev->mtu), 1);
+	if (err)
+		return err;
+
+	mlx5_query_port_oper_mtu(mdev, &hw_mtu, 1);
+
+	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);
+	return 0;
+}
+
+int mlx5e_open_locked(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int num_txqs;
+	int err;
+
+	set_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	num_txqs = priv->params.num_channels * priv->params.num_tc;
+	netif_set_real_num_tx_queues(netdev, num_txqs);
+	netif_set_real_num_rx_queues(netdev, priv->params.num_channels);
+
+	err = mlx5e_set_dev_port_mtu(netdev);
+	if (err)
+		goto err_clear_state_opened_flag;
+
+	err = mlx5e_open_channels(priv);
+	if (err) {
+		netdev_err(netdev, "%s: mlx5e_open_channels failed, %d\n",
+			   __func__, err);
+		goto err_clear_state_opened_flag;
+	}
+
+	err = mlx5e_refresh_tirs_self_loopback_enable(priv);
+	if (err) {
+		netdev_err(netdev, "%s: mlx5e_refresh_tirs_self_loopback_enable failed, %d\n",
+			   __func__, err);
+		goto err_close_channels;
+	}
+
+	mlx5e_update_carrier(priv);
+	mlx5e_redirect_rqts(priv);
+
+	schedule_delayed_work(&priv->update_stats_work, 0);
+
+	return 0;
+
+err_close_channels:
+	mlx5e_close_channels(priv);
+err_clear_state_opened_flag:
+	clear_bit(MLX5E_STATE_OPENED, &priv->state);
+	return err;
+}
+
+static int mlx5e_open(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+	err = mlx5e_open_locked(netdev);
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+int mlx5e_close_locked(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	/* May already be CLOSED in case a previous configuration operation
+	 * (e.g RX/TX queue size change) that involves close&open failed.
+	 */
+	if (!test_bit(MLX5E_STATE_OPENED, &priv->state))
+		return 0;
+
+	clear_bit(MLX5E_STATE_OPENED, &priv->state);
+
+	mlx5e_redirect_rqts(priv);
+	netif_carrier_off(priv->netdev);
+	mlx5e_close_channels(priv);
+
+	return 0;
+}
+
+static int mlx5e_close(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err;
+
+	mutex_lock(&priv->state_lock);
+	err = mlx5e_close_locked(netdev);
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static int mlx5e_create_drop_rq(struct mlx5e_priv *priv,
+				struct mlx5e_rq *rq,
+				struct mlx5e_rq_param *param)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	void *rqc = param->rqc;
+	void *rqc_wq = MLX5_ADDR_OF(rqc, rqc, wq);
+	int err;
+
+	param->wq.db_numa_node = param->wq.buf_numa_node;
+
+	err = mlx5_wq_ll_create(mdev, &param->wq, rqc_wq, &rq->wq,
+				&rq->wq_ctrl);
+	if (err)
+		return err;
+
+	rq->priv = priv;
+
+	return 0;
+}
+
+static int mlx5e_create_drop_cq(struct mlx5e_priv *priv,
+				struct mlx5e_cq *cq,
+				struct mlx5e_cq_param *param)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_core_cq *mcq = &cq->mcq;
+	int eqn_not_used;
+	unsigned int irqn;
+	int err;
+
+	err = mlx5_cqwq_create(mdev, &param->wq, param->cqc, &cq->wq,
+			       &cq->wq_ctrl);
+	if (err)
+		return err;
+
+	mlx5_vector2eqn(mdev, param->eq_ix, &eqn_not_used, &irqn);
+
+	mcq->cqe_sz     = 64;
+	mcq->set_ci_db  = cq->wq_ctrl.db.db;
+	mcq->arm_db     = cq->wq_ctrl.db.db + 1;
+	*mcq->set_ci_db = 0;
+	*mcq->arm_db    = 0;
+	mcq->vector     = param->eq_ix;
+	mcq->comp       = mlx5e_completion_event;
+	mcq->event      = mlx5e_cq_error_event;
+	mcq->irqn       = irqn;
+	mcq->uar        = &priv->cq_uar;
+
+	cq->priv = priv;
+
+	return 0;
+}
+
+static int mlx5e_open_drop_rq(struct mlx5e_priv *priv)
+{
+	struct mlx5e_cq_param cq_param;
+	struct mlx5e_rq_param rq_param;
+	struct mlx5e_rq *rq = &priv->drop_rq;
+	struct mlx5e_cq *cq = &priv->drop_rq.cq;
+	int err;
+
+	memset(&cq_param, 0, sizeof(cq_param));
+	memset(&rq_param, 0, sizeof(rq_param));
+	mlx5e_build_rx_cq_param(priv, &cq_param);
+	mlx5e_build_rq_param(priv, &rq_param);
+
+	err = mlx5e_create_drop_cq(priv, cq, &cq_param);
+	if (err)
+		return err;
+
+	err = mlx5e_enable_cq(cq, &cq_param);
+	if (err)
+		goto err_destroy_cq;
+
+	err = mlx5e_create_drop_rq(priv, rq, &rq_param);
+	if (err)
+		goto err_disable_cq;
+
+	err = mlx5e_enable_rq(rq, &rq_param);
+	if (err)
+		goto err_destroy_rq;
+
+	return 0;
+
+err_destroy_rq:
+	mlx5e_destroy_rq(&priv->drop_rq);
+
+err_disable_cq:
+	mlx5e_disable_cq(&priv->drop_rq.cq);
+
+err_destroy_cq:
+	mlx5e_destroy_cq(&priv->drop_rq.cq);
+
+	return err;
+}
+
+static void mlx5e_close_drop_rq(struct mlx5e_priv *priv)
+{
+	mlx5e_disable_rq(&priv->drop_rq);
+	mlx5e_destroy_rq(&priv->drop_rq);
+	mlx5e_disable_cq(&priv->drop_rq.cq);
+	mlx5e_destroy_cq(&priv->drop_rq.cq);
+}
+
+static int mlx5e_create_tis(struct mlx5e_priv *priv, int tc)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 in[MLX5_ST_SZ_DW(create_tis_in)];
+	void *tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(tisc, tisc, prio,  tc);
+	MLX5_SET(tisc, tisc, transport_domain, priv->tdn);
+
+	return mlx5_core_create_tis(mdev, in, sizeof(in), &priv->tisn[tc]);
+}
+
+static void mlx5e_destroy_tis(struct mlx5e_priv *priv, int tc)
+{
+	mlx5_core_destroy_tis(priv->mdev, priv->tisn[tc]);
+}
+
+static int mlx5e_create_tises(struct mlx5e_priv *priv)
+{
+	int err;
+	int tc;
+
+	for (tc = 0; tc < priv->params.num_tc; tc++) {
+		err = mlx5e_create_tis(priv, tc);
+		if (err)
+			goto err_close_tises;
+	}
+
+	return 0;
+
+err_close_tises:
+	for (tc--; tc >= 0; tc--)
+		mlx5e_destroy_tis(priv, tc);
+
+	return err;
+}
+
+static void mlx5e_destroy_tises(struct mlx5e_priv *priv)
+{
+	int tc;
+
+	for (tc = 0; tc < priv->params.num_tc; tc++)
+		mlx5e_destroy_tis(priv, tc);
+}
+
+static void mlx5e_build_tir_ctx(struct mlx5e_priv *priv, u32 *tirc, int tt)
+{
+	void *hfso = MLX5_ADDR_OF(tirc, tirc, rx_hash_field_selector_outer);
+
+	MLX5_SET(tirc, tirc, transport_domain, priv->tdn);
+
+#define MLX5_HASH_IP            (MLX5_HASH_FIELD_SEL_SRC_IP   |\
+				 MLX5_HASH_FIELD_SEL_DST_IP)
+
+#define MLX5_HASH_IP_L4PORTS    (MLX5_HASH_FIELD_SEL_SRC_IP   |\
+				 MLX5_HASH_FIELD_SEL_DST_IP   |\
+				 MLX5_HASH_FIELD_SEL_L4_SPORT |\
+				 MLX5_HASH_FIELD_SEL_L4_DPORT)
+
+#define MLX5_HASH_IP_IPSEC_SPI  (MLX5_HASH_FIELD_SEL_SRC_IP   |\
+				 MLX5_HASH_FIELD_SEL_DST_IP   |\
+				 MLX5_HASH_FIELD_SEL_IPSEC_SPI)
+
+	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]);
+		MLX5_SET(tirc, tirc, rx_hash_fn,
+			 mlx5e_rx_hash_fn(priv->params.rss_hfunc));
+		if (priv->params.rss_hfunc == ETH_RSS_HASH_TOP) {
+			void *rss_key = MLX5_ADDR_OF(tirc, tirc,
+						     rx_hash_toeplitz_key);
+			size_t len = MLX5_FLD_SZ_BYTES(tirc,
+						       rx_hash_toeplitz_key);
+
+			MLX5_SET(tirc, tirc, rx_hash_symmetric, 1);
+			memcpy(rss_key, priv->params.toeplitz_hash_key, len);
+		}
+		break;
+	}
+
+	switch (tt) {
+	case MLX5E_TT_IPV4_TCP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_TCP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV6_TCP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_TCP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV4_UDP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_UDP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV6_UDP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, l4_prot_type,
+			 MLX5_L4_PROT_TYPE_UDP);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_L4PORTS);
+		break;
+
+	case MLX5E_TT_IPV4_IPSEC_AH:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV6_IPSEC_AH:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV4_IPSEC_ESP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV6_IPSEC_ESP:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP_IPSEC_SPI);
+		break;
+
+	case MLX5E_TT_IPV4:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV4);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP);
+		break;
+
+	case MLX5E_TT_IPV6:
+		MLX5_SET(rx_hash_field_select, hfso, l3_prot_type,
+			 MLX5_L3_PROT_TYPE_IPV6);
+		MLX5_SET(rx_hash_field_select, hfso, selected_fields,
+			 MLX5_HASH_IP);
+		break;
+	}
+}
+
+static int mlx5e_create_tir(struct mlx5e_priv *priv, int tt)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	u32 *in;
+	void *tirc;
+	int inlen;
+	int err;
+
+	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);
+
+	mlx5e_build_tir_ctx(priv, tirc, tt);
+
+	err = mlx5_core_create_tir(mdev, in, inlen, &priv->tirn[tt]);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5e_destroy_tir(struct mlx5e_priv *priv, int tt)
+{
+	mlx5_core_destroy_tir(priv->mdev, priv->tirn[tt]);
+}
+
+static int mlx5e_create_tirs(struct mlx5e_priv *priv)
+{
+	int err;
+	int i;
+
+	for (i = 0; i < MLX5E_NUM_TT; i++) {
+		err = mlx5e_create_tir(priv, i);
+		if (err)
+			goto err_destroy_tirs;
+	}
+
+	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 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_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;
+
+	return stats;
+}
+
+static void mlx5e_set_rx_mode(struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
+	schedule_work(&priv->set_rx_mode_work);
+}
+
+static int mlx5e_set_mac(struct net_device *netdev, void *addr)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct sockaddr *saddr = addr;
+
+	if (!is_valid_ether_addr(saddr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	netif_addr_lock_bh(netdev);
+	ether_addr_copy(netdev->dev_addr, saddr->sa_data);
+	netif_addr_unlock_bh(netdev);
+
+	schedule_work(&priv->set_rx_mode_work);
+
+	return 0;
+}
+
+static int mlx5e_set_features(struct net_device *netdev,
+			      netdev_features_t features)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	int err = 0;
+	netdev_features_t changes = features ^ netdev->features;
+
+	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);
+		mlx5e_modify_tir_lro(priv, MLX5E_TT_IPV4_TCP);
+		mlx5e_modify_tir_lro(priv, MLX5E_TT_IPV6_TCP);
+
+		if (was_opened)
+			err = 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 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;
+	int err = 0;
+
+	mlx5_query_port_max_mtu(mdev, &max_mtu, 1);
+
+	max_mtu = MLX5E_HW2SW_MTU(max_mtu);
+
+	if (new_mtu > max_mtu) {
+		netdev_err(netdev,
+			   "%s: Bad MTU (%d) > (%d) Max\n",
+			   __func__, new_mtu, max_mtu);
+		return -EINVAL;
+	}
+
+	mutex_lock(&priv->state_lock);
+
+	was_opened = test_bit(MLX5E_STATE_OPENED, &priv->state);
+	if (was_opened)
+		mlx5e_close_locked(netdev);
+
+	netdev->mtu = new_mtu;
+
+	if (was_opened)
+		err = mlx5e_open_locked(netdev);
+
+	mutex_unlock(&priv->state_lock);
+
+	return err;
+}
+
+static struct net_device_ops mlx5e_netdev_ops = {
+	.ndo_open                = mlx5e_open,
+	.ndo_stop                = mlx5e_close,
+	.ndo_start_xmit          = mlx5e_xmit,
+	.ndo_get_stats64         = mlx5e_get_stats,
+	.ndo_set_rx_mode         = mlx5e_set_rx_mode,
+	.ndo_set_mac_address     = mlx5e_set_mac,
+	.ndo_vlan_rx_add_vid	 = mlx5e_vlan_rx_add_vid,
+	.ndo_vlan_rx_kill_vid	 = mlx5e_vlan_rx_kill_vid,
+	.ndo_set_features        = mlx5e_set_features,
+	.ndo_change_mtu		 = mlx5e_change_mtu,
+};
+
+static int mlx5e_check_required_hca_cap(struct mlx5_core_dev *mdev)
+{
+	if (MLX5_CAP_GEN(mdev, port_type) != MLX5_CAP_PORT_TYPE_ETH)
+		return -ENOTSUPP;
+	if (!MLX5_CAP_GEN(mdev, eth_net_offloads) ||
+	    !MLX5_CAP_GEN(mdev, nic_flow_table) ||
+	    !MLX5_CAP_ETH(mdev, csum_cap) ||
+	    !MLX5_CAP_ETH(mdev, max_lso_cap) ||
+	    !MLX5_CAP_ETH(mdev, vlan_cap) ||
+	    !MLX5_CAP_ETH(mdev, rss_ind_tbl_cap) ||
+	    MLX5_CAP_FLOWTABLE(mdev,
+			       flow_table_properties_nic_receive.max_ft_level)
+			       < 3) {
+		mlx5_core_warn(mdev,
+			       "Not creating net device, some required device capabilities are missing\n");
+		return -ENOTSUPP;
+	}
+	if (!MLX5_CAP_ETH(mdev, self_lb_en_modifiable))
+		mlx5_core_warn(mdev, "Self loop back prevention is not supported\n");
+
+	return 0;
+}
+
+u16 mlx5e_get_max_inline_cap(struct mlx5_core_dev *mdev)
+{
+	int bf_buf_size = (1 << MLX5_CAP_GEN(mdev, log_bf_reg_size)) / 2;
+
+	return bf_buf_size -
+	       sizeof(struct mlx5e_tx_wqe) +
+	       2 /*sizeof(mlx5e_tx_wqe.inline_hdr_start)*/;
+}
+
+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);
+	int i;
+
+	priv->params.log_sq_size           =
+		MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE;
+	priv->params.log_rq_size           =
+		MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE;
+	priv->params.rx_cq_moderation_usec =
+		MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC;
+	priv->params.rx_cq_moderation_pkts =
+		MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_PKTS;
+	priv->params.tx_cq_moderation_usec =
+		MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC;
+	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.default_vlan_prio     = 0;
+	priv->params.rss_hfunc             = ETH_RSS_HASH_XOR;
+
+	netdev_rss_key_fill(priv->params.toeplitz_hash_key,
+			    sizeof(priv->params.toeplitz_hash_key));
+
+	for (i = 0; i < MLX5E_INDIR_RQT_SIZE; i++)
+		priv->params.indirection_rqt[i] = i % num_channels;
+
+	priv->params.lro_wqe_sz            =
+		MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ;
+
+	priv->mdev                         = mdev;
+	priv->netdev                       = netdev;
+	priv->params.num_channels          = num_channels;
+	priv->default_vlan_prio            = priv->params.default_vlan_prio;
+
+	spin_lock_init(&priv->async_events_spinlock);
+	mutex_init(&priv->state_lock);
+
+	INIT_WORK(&priv->update_carrier_work, mlx5e_update_carrier_work);
+	INIT_WORK(&priv->set_rx_mode_work, mlx5e_set_rx_mode_work);
+	INIT_DELAYED_WORK(&priv->update_stats_work, mlx5e_update_stats_work);
+}
+
+static void mlx5e_set_netdev_dev_addr(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+
+	mlx5_query_nic_vport_mac_address(priv->mdev, netdev->dev_addr);
+}
+
+static void mlx5e_build_netdev(struct net_device *netdev)
+{
+	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5_core_dev *mdev = priv->mdev;
+
+	SET_NETDEV_DEV(netdev, &mdev->pdev->dev);
+
+	if (priv->params.num_tc > 1)
+		mlx5e_netdev_ops.ndo_select_queue = mlx5e_select_queue;
+
+	netdev->netdev_ops        = &mlx5e_netdev_ops;
+	netdev->watchdog_timeo    = 15 * HZ;
+
+	netdev->ethtool_ops	  = &mlx5e_ethtool_ops;
+
+	netdev->vlan_features    |= NETIF_F_SG;
+	netdev->vlan_features    |= NETIF_F_IP_CSUM;
+	netdev->vlan_features    |= NETIF_F_IPV6_CSUM;
+	netdev->vlan_features    |= NETIF_F_GRO;
+	netdev->vlan_features    |= NETIF_F_TSO;
+	netdev->vlan_features    |= NETIF_F_TSO6;
+	netdev->vlan_features    |= NETIF_F_RXCSUM;
+	netdev->vlan_features    |= NETIF_F_RXHASH;
+
+	if (!!MLX5_CAP_ETH(mdev, lro_cap))
+		netdev->vlan_features    |= NETIF_F_LRO;
+
+	netdev->hw_features       = netdev->vlan_features;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_TX;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_RX;
+	netdev->hw_features      |= NETIF_F_HW_VLAN_CTAG_FILTER;
+
+	netdev->features          = netdev->hw_features;
+	if (!priv->params.lro_en)
+		netdev->features  &= ~NETIF_F_LRO;
+
+	netdev->features         |= NETIF_F_HIGHDMA;
+
+	netdev->priv_flags       |= IFF_UNICAST_FLT;
+
+	mlx5e_set_netdev_dev_addr(netdev);
+}
+
+static int mlx5e_create_mkey(struct mlx5e_priv *priv, u32 pdn,
+			     struct mlx5_core_mr *mr)
+{
+	struct mlx5_core_dev *mdev = priv->mdev;
+	struct mlx5_create_mkey_mbox_in *in;
+	int err;
+
+	in = mlx5_vzalloc(sizeof(*in));
+	if (!in)
+		return -ENOMEM;
+
+	in->seg.flags = MLX5_PERM_LOCAL_WRITE |
+			MLX5_PERM_LOCAL_READ  |
+			MLX5_ACCESS_MODE_PA;
+	in->seg.flags_pd = cpu_to_be32(pdn | MLX5_MKEY_LEN64);
+	in->seg.qpn_mkey7_0 = cpu_to_be32(0xffffff << 8);
+
+	err = mlx5_core_create_mkey(mdev, mr, in, sizeof(*in), NULL, NULL,
+				    NULL);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void *mlx5e_create_netdev(struct mlx5_core_dev *mdev)
+{
+	struct net_device *netdev;
+	struct mlx5e_priv *priv;
+	int nch = mlx5e_get_max_num_channels(mdev);
+	int err;
+
+	if (mlx5e_check_required_hca_cap(mdev))
+		return NULL;
+
+	netdev = alloc_etherdev_mqs(sizeof(struct mlx5e_priv), nch, nch);
+	if (!netdev) {
+		mlx5_core_err(mdev, "alloc_etherdev_mqs() failed\n");
+		return NULL;
+	}
+
+	mlx5e_build_netdev_priv(mdev, netdev, nch);
+	mlx5e_build_netdev(netdev);
+
+	netif_carrier_off(netdev);
+
+	priv = netdev_priv(netdev);
+
+	err = mlx5_alloc_map_uar(mdev, &priv->cq_uar);
+	if (err) {
+		mlx5_core_err(mdev, "alloc_map uar failed, %d\n", err);
+		goto err_free_netdev;
+	}
+
+	err = mlx5_core_alloc_pd(mdev, &priv->pdn);
+	if (err) {
+		mlx5_core_err(mdev, "alloc pd failed, %d\n", err);
+		goto err_unmap_free_uar;
+	}
+
+	err = mlx5_alloc_transport_domain(mdev, &priv->tdn);
+	if (err) {
+		mlx5_core_err(mdev, "alloc td failed, %d\n", err);
+		goto err_dealloc_pd;
+	}
+
+	err = mlx5e_create_mkey(priv, priv->pdn, &priv->mr);
+	if (err) {
+		mlx5_core_err(mdev, "create mkey failed, %d\n", err);
+		goto err_dealloc_transport_domain;
+	}
+
+	err = mlx5e_create_tises(priv);
+	if (err) {
+		mlx5_core_warn(mdev, "create tises failed, %d\n", err);
+		goto err_destroy_mkey;
+	}
+
+	err = mlx5e_open_drop_rq(priv);
+	if (err) {
+		mlx5_core_err(mdev, "open drop rq failed, %d\n", err);
+		goto err_destroy_tises;
+	}
+
+	err = mlx5e_create_rqt(priv, MLX5E_INDIRECTION_RQT);
+	if (err) {
+		mlx5_core_warn(mdev, "create rqt(INDIR) 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;
+	}
+
+	err = mlx5e_create_flow_tables(priv);
+	if (err) {
+		mlx5_core_warn(mdev, "create flow tables failed, %d\n", err);
+		goto err_destroy_tirs;
+	}
+
+	mlx5e_init_eth_addr(priv);
+
+	err = register_netdev(netdev);
+	if (err) {
+		mlx5_core_err(mdev, "register_netdev failed, %d\n", err);
+		goto err_destroy_flow_tables;
+	}
+
+	mlx5e_enable_async_events(priv);
+	schedule_work(&priv->set_rx_mode_work);
+
+	return priv;
+
+err_destroy_flow_tables:
+	mlx5e_destroy_flow_tables(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_close_drop_rq:
+	mlx5e_close_drop_rq(priv);
+
+err_destroy_tises:
+	mlx5e_destroy_tises(priv);
+
+err_destroy_mkey:
+	mlx5_core_destroy_mkey(mdev, &priv->mr);
+
+err_dealloc_transport_domain:
+	mlx5_dealloc_transport_domain(mdev, priv->tdn);
+
+err_dealloc_pd:
+	mlx5_core_dealloc_pd(mdev, priv->pdn);
+
+err_unmap_free_uar:
+	mlx5_unmap_free_uar(mdev, &priv->cq_uar);
+
+err_free_netdev:
+	free_netdev(netdev);
+
+	return NULL;
+}
+
+static void mlx5e_destroy_netdev(struct mlx5_core_dev *mdev, void *vpriv)
+{
+	struct mlx5e_priv *priv = vpriv;
+	struct net_device *netdev = priv->netdev;
+
+	set_bit(MLX5E_STATE_DESTROYING, &priv->state);
+
+	schedule_work(&priv->set_rx_mode_work);
+	mlx5e_disable_async_events(priv);
+	flush_scheduled_work();
+	unregister_netdev(netdev);
+	mlx5e_destroy_flow_tables(priv);
+	mlx5e_destroy_tirs(priv);
+	mlx5e_destroy_rqt(priv, MLX5E_SINGLE_RQ_RQT);
+	mlx5e_destroy_rqt(priv, MLX5E_INDIRECTION_RQT);
+	mlx5e_close_drop_rq(priv);
+	mlx5e_destroy_tises(priv);
+	mlx5_core_destroy_mkey(priv->mdev, &priv->mr);
+	mlx5_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);
+}
+
+static void *mlx5e_get_netdev(void *vpriv)
+{
+	struct mlx5e_priv *priv = vpriv;
+
+	return priv->netdev;
+}
+
+static struct mlx5_interface mlx5e_interface = {
+	.add       = mlx5e_create_netdev,
+	.remove    = mlx5e_destroy_netdev,
+	.event     = mlx5e_async_event,
+	.protocol  = MLX5_INTERFACE_PROTOCOL_ETH,
+	.get_dev   = mlx5e_get_netdev,
+};
+
+void mlx5e_init(void)
+{
+	mlx5_register_interface(&mlx5e_interface);
+}
+
+void mlx5e_cleanup(void)
+{
+	mlx5_unregister_interface(&mlx5e_interface);
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
new file mode 100644
index 000000000..cf0098596
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -0,0 +1,278 @@
+/*
+ * Copyright (c) 2015, 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 <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/tcp.h>
+#include "en.h"
+
+static inline int mlx5e_alloc_rx_wqe(struct mlx5e_rq *rq,
+				     struct mlx5e_rx_wqe *wqe, u16 ix)
+{
+	struct sk_buff *skb;
+	dma_addr_t dma_addr;
+
+	skb = netdev_alloc_skb(rq->netdev, rq->wqe_sz);
+	if (unlikely(!skb))
+		return -ENOMEM;
+
+	dma_addr = dma_map_single(rq->pdev,
+				  /* hw start padding */
+				  skb->data,
+				  /* hw end padding */
+				  rq->wqe_sz,
+				  DMA_FROM_DEVICE);
+
+	if (unlikely(dma_mapping_error(rq->pdev, dma_addr)))
+		goto err_free_skb;
+
+	skb_reserve(skb, MLX5E_NET_IP_ALIGN);
+
+	*((dma_addr_t *)skb->cb) = dma_addr;
+	wqe->data.addr = cpu_to_be64(dma_addr + MLX5E_NET_IP_ALIGN);
+
+	rq->skb[ix] = skb;
+
+	return 0;
+
+err_free_skb:
+	dev_kfree_skb(skb);
+
+	return -ENOMEM;
+}
+
+bool mlx5e_post_rx_wqes(struct mlx5e_rq *rq)
+{
+	struct mlx5_wq_ll *wq = &rq->wq;
+
+	if (unlikely(!test_bit(MLX5E_RQ_STATE_POST_WQES_ENABLE, &rq->state)))
+		return false;
+
+	while (!mlx5_wq_ll_is_full(wq)) {
+		struct mlx5e_rx_wqe *wqe = mlx5_wq_ll_get_wqe(wq, wq->head);
+
+		if (unlikely(mlx5e_alloc_rx_wqe(rq, wqe, wq->head)))
+			break;
+
+		mlx5_wq_ll_push(wq, be16_to_cpu(wqe->next.next_wqe_index));
+	}
+
+	/* ensure wqes are visible to device before updating doorbell record */
+	dma_wmb();
+
+	mlx5_wq_ll_update_db_record(wq);
+
+	return !mlx5_wq_ll_is_full(wq);
+}
+
+static void mlx5e_lro_update_hdr(struct sk_buff *skb, struct mlx5_cqe64 *cqe)
+{
+	struct ethhdr	*eth	= (struct ethhdr *)(skb->data);
+	struct iphdr	*ipv4	= (struct iphdr *)(skb->data + ETH_HLEN);
+	struct ipv6hdr	*ipv6	= (struct ipv6hdr *)(skb->data + ETH_HLEN);
+	struct tcphdr	*tcp;
+
+	u8 l4_hdr_type = get_cqe_l4_hdr_type(cqe);
+	int tcp_ack = ((CQE_L4_HDR_TYPE_TCP_ACK_NO_DATA  == l4_hdr_type) ||
+		       (CQE_L4_HDR_TYPE_TCP_ACK_AND_DATA == l4_hdr_type));
+
+	u16 tot_len = be32_to_cpu(cqe->byte_cnt) - ETH_HLEN;
+
+	if (eth->h_proto == htons(ETH_P_IP)) {
+		tcp = (struct tcphdr *)(skb->data + ETH_HLEN +
+					sizeof(struct iphdr));
+		ipv6 = NULL;
+		skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
+	} else {
+		tcp = (struct tcphdr *)(skb->data + ETH_HLEN +
+					sizeof(struct ipv6hdr));
+		ipv4 = NULL;
+		skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
+	}
+
+	if (get_cqe_lro_tcppsh(cqe))
+		tcp->psh                = 1;
+
+	if (tcp_ack) {
+		tcp->ack                = 1;
+		tcp->ack_seq            = cqe->lro_ack_seq_num;
+		tcp->window             = cqe->lro_tcp_win;
+	}
+
+	if (ipv4) {
+		ipv4->ttl               = cqe->lro_min_ttl;
+		ipv4->tot_len           = cpu_to_be16(tot_len);
+		ipv4->check             = 0;
+		ipv4->check             = ip_fast_csum((unsigned char *)ipv4,
+						       ipv4->ihl);
+	} else {
+		ipv6->hop_limit         = cqe->lro_min_ttl;
+		ipv6->payload_len       = cpu_to_be16(tot_len -
+						      sizeof(struct ipv6hdr));
+	}
+}
+
+static inline void mlx5e_skb_set_hash(struct mlx5_cqe64 *cqe,
+				      struct sk_buff *skb)
+{
+	u8 cht = cqe->rss_hash_type;
+	int ht = (cht & CQE_RSS_HTYPE_L4) ? PKT_HASH_TYPE_L4 :
+		 (cht & CQE_RSS_HTYPE_IP) ? PKT_HASH_TYPE_L3 :
+					    PKT_HASH_TYPE_NONE;
+	skb_set_hash(skb, be32_to_cpu(cqe->rss_hash_result), ht);
+}
+
+static inline bool is_first_ethertype_ip(struct sk_buff *skb)
+{
+	__be16 ethertype = ((struct ethhdr *)skb->data)->h_proto;
+
+	return (ethertype == htons(ETH_P_IP) || ethertype == htons(ETH_P_IPV6));
+}
+
+static inline void mlx5e_handle_csum(struct net_device *netdev,
+				     struct mlx5_cqe64 *cqe,
+				     struct mlx5e_rq *rq,
+				     struct sk_buff *skb)
+{
+	if (unlikely(!(netdev->features & NETIF_F_RXCSUM)))
+		goto csum_none;
+
+	if (likely(cqe->hds_ip_ext & CQE_L4_OK)) {
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+	} else if (is_first_ethertype_ip(skb)) {
+		skb->ip_summed = CHECKSUM_COMPLETE;
+		skb->csum = csum_unfold((__force __sum16)cqe->check_sum);
+		rq->stats.csum_sw++;
+	} else {
+		goto csum_none;
+	}
+
+	return;
+
+csum_none:
+	skb->ip_summed = CHECKSUM_NONE;
+	rq->stats.csum_none++;
+}
+
+static inline void mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
+				      struct mlx5e_rq *rq,
+				      struct sk_buff *skb)
+{
+	struct net_device *netdev = rq->netdev;
+	u32 cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
+	int lro_num_seg;
+
+	skb_put(skb, cqe_bcnt);
+
+	lro_num_seg = be32_to_cpu(cqe->srqn) >> 24;
+	if (lro_num_seg > 1) {
+		mlx5e_lro_update_hdr(skb, cqe);
+		skb_shinfo(skb)->gso_size = DIV_ROUND_UP(cqe_bcnt, lro_num_seg);
+		rq->stats.lro_packets++;
+		rq->stats.lro_bytes += cqe_bcnt;
+	}
+
+	mlx5e_handle_csum(netdev, cqe, rq, skb);
+
+	skb->protocol = eth_type_trans(skb, netdev);
+
+	skb_record_rx_queue(skb, rq->ix);
+
+	if (likely(netdev->features & NETIF_F_RXHASH))
+		mlx5e_skb_set_hash(cqe, skb);
+
+	if (cqe_has_vlan(cqe))
+		__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
+				       be16_to_cpu(cqe->vlan_info));
+}
+
+bool mlx5e_poll_rx_cq(struct mlx5e_cq *cq, int budget)
+{
+	struct mlx5e_rq *rq = container_of(cq, struct mlx5e_rq, cq);
+	int i;
+
+	/* avoid accessing cq (dma coherent memory) if not needed */
+	if (!test_and_clear_bit(MLX5E_CQ_HAS_CQES, &cq->flags))
+		return false;
+
+	for (i = 0; i < budget; i++) {
+		struct mlx5e_rx_wqe *wqe;
+		struct mlx5_cqe64 *cqe;
+		struct sk_buff *skb;
+		__be16 wqe_counter_be;
+		u16 wqe_counter;
+
+		cqe = mlx5e_get_cqe(cq);
+		if (!cqe)
+			break;
+
+		mlx5_cqwq_pop(&cq->wq);
+
+		wqe_counter_be = cqe->wqe_counter;
+		wqe_counter    = be16_to_cpu(wqe_counter_be);
+		wqe            = mlx5_wq_ll_get_wqe(&rq->wq, wqe_counter);
+		skb            = rq->skb[wqe_counter];
+		prefetch(skb->data);
+		rq->skb[wqe_counter] = NULL;
+
+		dma_unmap_single(rq->pdev,
+				 *((dma_addr_t *)skb->cb),
+				 rq->wqe_sz,
+				 DMA_FROM_DEVICE);
+
+		if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) {
+			rq->stats.wqe_err++;
+			dev_kfree_skb(skb);
+			goto wq_ll_pop;
+		}
+
+		mlx5e_build_rx_skb(cqe, rq, skb);
+		rq->stats.packets++;
+		napi_gro_receive(cq->napi, skb);
+
+wq_ll_pop:
+		mlx5_wq_ll_pop(&rq->wq, wqe_counter_be,
+			       &wqe->next.next_wqe_index);
+	}
+
+	mlx5_cqwq_update_db_record(&cq->wq);
+
+	/* ensure cq space is freed before enabling more cqes */
+	wmb();
+
+	if (i == budget) {
+		set_bit(MLX5E_CQ_HAS_CQES, &cq->flags);
+		return true;
+	}
+
+	return false;
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
new file mode 100644
index 000000000..1341b1d3c
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -0,0 +1,406 @@
+/*
+ * Copyright (c) 2015, 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 <linux/tcp.h>
+#include <linux/if_vlan.h>
+#include "en.h"
+
+#define MLX5E_SQ_NOPS_ROOM  MLX5_SEND_WQE_MAX_WQEBBS
+#define MLX5E_SQ_STOP_ROOM (MLX5_SEND_WQE_MAX_WQEBBS +\
+			    MLX5E_SQ_NOPS_ROOM)
+
+void mlx5e_send_nop(struct mlx5e_sq *sq, bool notify_hw)
+{
+	struct mlx5_wq_cyc                *wq  = &sq->wq;
+
+	u16 pi = sq->pc & wq->sz_m1;
+	struct mlx5e_tx_wqe              *wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
+
+	struct mlx5_wqe_ctrl_seg         *cseg = &wqe->ctrl;
+
+	memset(cseg, 0, sizeof(*cseg));
+
+	cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | MLX5_OPCODE_NOP);
+	cseg->qpn_ds           = cpu_to_be32((sq->sqn << 8) | 0x01);
+
+	sq->skb[pi] = NULL;
+	sq->pc++;
+
+	if (notify_hw) {
+		cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
+		mlx5e_tx_notify_hw(sq, wqe, 0);
+	}
+}
+
+static inline void mlx5e_tx_dma_unmap(struct device *pdev,
+				      struct mlx5e_sq_dma *dma)
+{
+	switch (dma->type) {
+	case MLX5E_DMA_MAP_SINGLE:
+		dma_unmap_single(pdev, dma->addr, dma->size, DMA_TO_DEVICE);
+		break;
+	case MLX5E_DMA_MAP_PAGE:
+		dma_unmap_page(pdev, dma->addr, dma->size, DMA_TO_DEVICE);
+		break;
+	default:
+		WARN_ONCE(true, "mlx5e_tx_dma_unmap unknown DMA type!\n");
+	}
+}
+
+static inline void mlx5e_dma_push(struct mlx5e_sq *sq,
+				  dma_addr_t addr,
+				  u32 size,
+				  enum mlx5e_dma_map_type map_type)
+{
+	sq->dma_fifo[sq->dma_fifo_pc & sq->dma_fifo_mask].addr = addr;
+	sq->dma_fifo[sq->dma_fifo_pc & sq->dma_fifo_mask].size = size;
+	sq->dma_fifo[sq->dma_fifo_pc & sq->dma_fifo_mask].type = map_type;
+	sq->dma_fifo_pc++;
+}
+
+static inline struct mlx5e_sq_dma *mlx5e_dma_get(struct mlx5e_sq *sq, u32 i)
+{
+	return &sq->dma_fifo[i & sq->dma_fifo_mask];
+}
+
+static void mlx5e_dma_unmap_wqe_err(struct mlx5e_sq *sq, struct sk_buff *skb)
+{
+	int i;
+
+	for (i = 0; i < MLX5E_TX_SKB_CB(skb)->num_dma; i++) {
+		struct mlx5e_sq_dma *last_pushed_dma =
+			mlx5e_dma_get(sq, --sq->dma_fifo_pc);
+
+		mlx5e_tx_dma_unmap(sq->pdev, last_pushed_dma);
+	}
+}
+
+u16 mlx5e_select_queue(struct net_device *dev, struct sk_buff *skb,
+		       void *accel_priv, select_queue_fallback_t fallback)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	int channel_ix = fallback(dev, skb);
+	int up = skb_vlan_tag_present(skb)        ?
+		 skb->vlan_tci >> VLAN_PRIO_SHIFT :
+		 priv->default_vlan_prio;
+	int tc = netdev_get_prio_tc_map(dev, up);
+
+	return priv->channeltc_to_txq_map[channel_ix][tc];
+}
+
+static inline u16 mlx5e_get_inline_hdr_size(struct mlx5e_sq *sq,
+					    struct sk_buff *skb, bool bf)
+{
+	/* Some NIC TX decisions, e.g loopback, are based on the packet
+	 * headers and occur before the data gather.
+	 * Therefore these headers must be copied into the WQE
+	 */
+#define MLX5E_MIN_INLINE ETH_HLEN
+
+	if (bf) {
+		u16 ihs = skb_headlen(skb);
+
+		if (skb_vlan_tag_present(skb))
+			ihs += VLAN_HLEN;
+
+		if (ihs <= sq->max_inline)
+			return skb_headlen(skb);
+	}
+
+	return MLX5E_MIN_INLINE;
+}
+
+static inline void mlx5e_insert_vlan(void *start, struct sk_buff *skb, u16 ihs)
+{
+	struct vlan_ethhdr *vhdr = (struct vlan_ethhdr *)start;
+	int cpy1_sz = 2 * ETH_ALEN;
+	int cpy2_sz = ihs - cpy1_sz;
+
+	skb_copy_from_linear_data(skb, vhdr, cpy1_sz);
+	skb_pull_inline(skb, cpy1_sz);
+	vhdr->h_vlan_proto = skb->vlan_proto;
+	vhdr->h_vlan_TCI = cpu_to_be16(skb_vlan_tag_get(skb));
+	skb_copy_from_linear_data(skb, &vhdr->h_vlan_encapsulated_proto,
+				  cpy2_sz);
+	skb_pull_inline(skb, cpy2_sz);
+}
+
+static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
+{
+	struct mlx5_wq_cyc       *wq   = &sq->wq;
+
+	u16 pi = sq->pc & wq->sz_m1;
+	struct mlx5e_tx_wqe      *wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
+
+	struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
+	struct mlx5_wqe_eth_seg  *eseg = &wqe->eth;
+	struct mlx5_wqe_data_seg *dseg;
+
+	u8  opcode = MLX5_OPCODE_SEND;
+	dma_addr_t dma_addr = 0;
+	bool bf = false;
+	u16 headlen;
+	u16 ds_cnt;
+	u16 ihs;
+	int i;
+
+	memset(wqe, 0, sizeof(*wqe));
+
+	if (likely(skb->ip_summed == CHECKSUM_PARTIAL))
+		eseg->cs_flags	= MLX5_ETH_WQE_L3_CSUM | MLX5_ETH_WQE_L4_CSUM;
+	else
+		sq->stats.csum_offload_none++;
+
+	if (sq->cc != sq->prev_cc) {
+		sq->prev_cc = sq->cc;
+		sq->bf_budget = (sq->cc == sq->pc) ? MLX5E_SQ_BF_BUDGET : 0;
+	}
+
+	if (skb_is_gso(skb)) {
+		u32 payload_len;
+
+		eseg->mss    = cpu_to_be16(skb_shinfo(skb)->gso_size);
+		opcode       = MLX5_OPCODE_LSO;
+		ihs          = skb_transport_offset(skb) + tcp_hdrlen(skb);
+		payload_len  = skb->len - ihs;
+		MLX5E_TX_SKB_CB(skb)->num_bytes = skb->len +
+					(skb_shinfo(skb)->gso_segs - 1) * ihs;
+		sq->stats.tso_packets++;
+		sq->stats.tso_bytes += payload_len;
+	} else {
+		bf = sq->bf_budget &&
+		     !skb->xmit_more &&
+		     !skb_shinfo(skb)->nr_frags;
+		ihs = mlx5e_get_inline_hdr_size(sq, skb, bf);
+		MLX5E_TX_SKB_CB(skb)->num_bytes = max_t(unsigned int, skb->len,
+							ETH_ZLEN);
+	}
+
+	if (skb_vlan_tag_present(skb)) {
+		mlx5e_insert_vlan(eseg->inline_hdr_start, skb, ihs);
+		ihs += VLAN_HLEN;
+	} else {
+		skb_copy_from_linear_data(skb, eseg->inline_hdr_start, ihs);
+		skb_pull_inline(skb, ihs);
+	}
+
+	eseg->inline_hdr_sz = cpu_to_be16(ihs);
+
+	ds_cnt  = sizeof(*wqe) / MLX5_SEND_WQE_DS;
+	ds_cnt += DIV_ROUND_UP(ihs - sizeof(eseg->inline_hdr_start),
+			       MLX5_SEND_WQE_DS);
+	dseg    = (struct mlx5_wqe_data_seg *)cseg + ds_cnt;
+
+	MLX5E_TX_SKB_CB(skb)->num_dma = 0;
+
+	headlen = skb_headlen(skb);
+	if (headlen) {
+		dma_addr = dma_map_single(sq->pdev, skb->data, headlen,
+					  DMA_TO_DEVICE);
+		if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
+			goto dma_unmap_wqe_err;
+
+		dseg->addr       = cpu_to_be64(dma_addr);
+		dseg->lkey       = sq->mkey_be;
+		dseg->byte_count = cpu_to_be32(headlen);
+
+		mlx5e_dma_push(sq, dma_addr, headlen, MLX5E_DMA_MAP_SINGLE);
+		MLX5E_TX_SKB_CB(skb)->num_dma++;
+
+		dseg++;
+	}
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+		struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
+		int fsz = skb_frag_size(frag);
+
+		dma_addr = skb_frag_dma_map(sq->pdev, frag, 0, fsz,
+					    DMA_TO_DEVICE);
+		if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
+			goto dma_unmap_wqe_err;
+
+		dseg->addr       = cpu_to_be64(dma_addr);
+		dseg->lkey       = sq->mkey_be;
+		dseg->byte_count = cpu_to_be32(fsz);
+
+		mlx5e_dma_push(sq, dma_addr, fsz, MLX5E_DMA_MAP_PAGE);
+		MLX5E_TX_SKB_CB(skb)->num_dma++;
+
+		dseg++;
+	}
+
+	ds_cnt += MLX5E_TX_SKB_CB(skb)->num_dma;
+
+	cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
+	cseg->qpn_ds           = cpu_to_be32((sq->sqn << 8) | ds_cnt);
+
+	sq->skb[pi] = skb;
+
+	MLX5E_TX_SKB_CB(skb)->num_wqebbs = DIV_ROUND_UP(ds_cnt,
+							MLX5_SEND_WQEBB_NUM_DS);
+	sq->pc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
+
+	netdev_tx_sent_queue(sq->txq, MLX5E_TX_SKB_CB(skb)->num_bytes);
+
+	if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_SQ_STOP_ROOM))) {
+		netif_tx_stop_queue(sq->txq);
+		sq->stats.stopped++;
+	}
+
+	if (!skb->xmit_more || netif_xmit_stopped(sq->txq)) {
+		int bf_sz = 0;
+
+		if (bf && sq->uar_bf_map)
+			bf_sz = MLX5E_TX_SKB_CB(skb)->num_wqebbs << 3;
+
+		cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
+		mlx5e_tx_notify_hw(sq, wqe, bf_sz);
+	}
+
+	/* fill sq edge with nops to avoid wqe wrap around */
+	while ((sq->pc & wq->sz_m1) > sq->edge)
+		mlx5e_send_nop(sq, false);
+
+	sq->bf_budget = bf ? sq->bf_budget - 1 : 0;
+
+	sq->stats.packets++;
+	return NETDEV_TX_OK;
+
+dma_unmap_wqe_err:
+	sq->stats.dropped++;
+	mlx5e_dma_unmap_wqe_err(sq, skb);
+
+	dev_kfree_skb_any(skb);
+
+	return NETDEV_TX_OK;
+}
+
+netdev_tx_t mlx5e_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct mlx5e_sq *sq = priv->txq_to_sq_map[skb_get_queue_mapping(skb)];
+
+	return mlx5e_sq_xmit(sq, skb);
+}
+
+bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq)
+{
+	struct mlx5e_sq *sq;
+	u32 dma_fifo_cc;
+	u32 nbytes;
+	u16 npkts;
+	u16 sqcc;
+	int i;
+
+	/* avoid accessing cq (dma coherent memory) if not needed */
+	if (!test_and_clear_bit(MLX5E_CQ_HAS_CQES, &cq->flags))
+		return false;
+
+	sq = container_of(cq, struct mlx5e_sq, cq);
+
+	npkts = 0;
+	nbytes = 0;
+
+	/* sq->cc must be updated only after mlx5_cqwq_update_db_record(),
+	 * otherwise a cq overrun may occur
+	 */
+	sqcc = sq->cc;
+
+	/* avoid dirtying sq cache line every cqe */
+	dma_fifo_cc = sq->dma_fifo_cc;
+
+	for (i = 0; i < MLX5E_TX_CQ_POLL_BUDGET; i++) {
+		struct mlx5_cqe64 *cqe;
+		u16 wqe_counter;
+		bool last_wqe;
+
+		cqe = mlx5e_get_cqe(cq);
+		if (!cqe)
+			break;
+
+		mlx5_cqwq_pop(&cq->wq);
+
+		wqe_counter = be16_to_cpu(cqe->wqe_counter);
+
+		do {
+			struct sk_buff *skb;
+			u16 ci;
+			int j;
+
+			last_wqe = (sqcc == wqe_counter);
+
+			ci = sqcc & sq->wq.sz_m1;
+			skb = sq->skb[ci];
+
+			if (unlikely(!skb)) { /* nop */
+				sq->stats.nop++;
+				sqcc++;
+				continue;
+			}
+
+			for (j = 0; j < MLX5E_TX_SKB_CB(skb)->num_dma; j++) {
+				struct mlx5e_sq_dma *dma =
+					mlx5e_dma_get(sq, dma_fifo_cc++);
+
+				mlx5e_tx_dma_unmap(sq->pdev, dma);
+			}
+
+			npkts++;
+			nbytes += MLX5E_TX_SKB_CB(skb)->num_bytes;
+			sqcc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
+			dev_kfree_skb(skb);
+		} while (!last_wqe);
+	}
+
+	mlx5_cqwq_update_db_record(&cq->wq);
+
+	/* ensure cq space is freed before enabling more cqes */
+	wmb();
+
+	sq->dma_fifo_cc = dma_fifo_cc;
+	sq->cc = sqcc;
+
+	netdev_tx_completed_queue(sq->txq, npkts, nbytes);
+
+	if (netif_tx_queue_stopped(sq->txq) &&
+	    mlx5e_sq_has_room_for(sq, MLX5E_SQ_STOP_ROOM) &&
+	    likely(test_bit(MLX5E_SQ_STATE_WAKE_TXQ_ENABLE, &sq->state))) {
+				netif_tx_wake_queue(sq->txq);
+				sq->stats.wake++;
+	}
+	if (i == MLX5E_TX_CQ_POLL_BUDGET) {
+		set_bit(MLX5E_CQ_HAS_CQES, &cq->flags);
+		return true;
+	}
+
+	return false;
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
new file mode 100644
index 000000000..2c7cb6755
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -0,0 +1,105 @@
+/*
+ * Copyright (c) 2015, 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 "en.h"
+
+struct mlx5_cqe64 *mlx5e_get_cqe(struct mlx5e_cq *cq)
+{
+	struct mlx5_cqwq *wq = &cq->wq;
+	u32 ci = mlx5_cqwq_get_ci(wq);
+	struct mlx5_cqe64 *cqe = mlx5_cqwq_get_wqe(wq, ci);
+	int cqe_ownership_bit = cqe->op_own & MLX5_CQE_OWNER_MASK;
+	int sw_ownership_val = mlx5_cqwq_get_wrap_cnt(wq) & 1;
+
+	if (cqe_ownership_bit != sw_ownership_val)
+		return NULL;
+
+	/* ensure cqe content is read after cqe ownership bit */
+	rmb();
+
+	return cqe;
+}
+
+int mlx5e_napi_poll(struct napi_struct *napi, int budget)
+{
+	struct mlx5e_channel *c = container_of(napi, struct mlx5e_channel,
+					       napi);
+	bool busy = false;
+	int i;
+
+	clear_bit(MLX5E_CHANNEL_NAPI_SCHED, &c->flags);
+
+	for (i = 0; i < c->num_tc; i++)
+		busy |= mlx5e_poll_tx_cq(&c->sq[i].cq);
+
+	busy |= mlx5e_poll_rx_cq(&c->rq.cq, budget);
+
+	busy |= mlx5e_post_rx_wqes(&c->rq);
+
+	if (busy)
+		return budget;
+
+	napi_complete(napi);
+
+	/* avoid losing completion event during/after polling cqs */
+	if (test_bit(MLX5E_CHANNEL_NAPI_SCHED, &c->flags)) {
+		napi_schedule(napi);
+		return 0;
+	}
+
+	for (i = 0; i < c->num_tc; i++)
+		mlx5e_cq_arm(&c->sq[i].cq);
+	mlx5e_cq_arm(&c->rq.cq);
+
+	return 0;
+}
+
+void mlx5e_completion_event(struct mlx5_core_cq *mcq)
+{
+	struct mlx5e_cq *cq = container_of(mcq, struct mlx5e_cq, mcq);
+
+	set_bit(MLX5E_CQ_HAS_CQES, &cq->flags);
+	set_bit(MLX5E_CHANNEL_NAPI_SCHED, &cq->channel->flags);
+	barrier();
+	napi_schedule(cq->napi);
+}
+
+void mlx5e_cq_error_event(struct mlx5_core_cq *mcq, enum mlx5_event event)
+{
+	struct mlx5e_cq *cq = container_of(mcq, struct mlx5e_cq, mcq);
+	struct mlx5e_channel *c = cq->channel;
+	struct mlx5e_priv *priv = c->priv;
+	struct net_device *netdev = priv->netdev;
+
+	netdev_err(netdev, "%s: cqn=0x%.6x event=0x%.2x\n",
+		   __func__, mcq->cqn, event);
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/eq.c
index 58800e4f3..713ead583 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/eq.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/eq.c
@@ -339,15 +339,15 @@ static void init_eq_buf(struct mlx5_eq *eq)
 int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx,
 		       int nent, u64 mask, const char *name, struct mlx5_uar *uar)
 {
-	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	struct mlx5_priv *priv = &dev->priv;
 	struct mlx5_create_eq_mbox_in *in;
 	struct mlx5_create_eq_mbox_out out;
 	int err;
 	int inlen;
 
 	eq->nent = roundup_pow_of_two(nent + MLX5_NUM_SPARE_EQE);
-	err = mlx5_buf_alloc(dev, eq->nent * MLX5_EQE_SIZE, 2 * PAGE_SIZE,
-			     &eq->buf);
+	eq->cons_index = 0;
+	err = mlx5_buf_alloc(dev, eq->nent * MLX5_EQE_SIZE, &eq->buf);
 	if (err)
 		return err;
 
@@ -378,14 +378,15 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx,
 		goto err_in;
 	}
 
-	snprintf(eq->name, MLX5_MAX_EQ_NAME, "%s@pci:%s",
+	snprintf(priv->irq_info[vecidx].name, MLX5_MAX_IRQ_NAME, "%s@pci:%s",
 		 name, pci_name(dev->pdev));
+
 	eq->eqn = out.eq_number;
-	eq->irqn = vecidx;
+	eq->irqn = priv->msix_arr[vecidx].vector;
 	eq->dev = dev;
 	eq->doorbell = uar->map + MLX5_EQ_DOORBEL_OFFSET;
-	err = request_irq(table->msix_arr[vecidx].vector, mlx5_msix_handler, 0,
-			  eq->name, eq);
+	err = request_irq(eq->irqn, mlx5_msix_handler, 0,
+			  priv->irq_info[vecidx].name, eq);
 	if (err)
 		goto err_eq;
 
@@ -401,7 +402,7 @@ int mlx5_create_map_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq, u8 vecidx,
 	return 0;
 
 err_irq:
-	free_irq(table->msix_arr[vecidx].vector, eq);
+	free_irq(priv->msix_arr[vecidx].vector, eq);
 
 err_eq:
 	mlx5_cmd_destroy_eq(dev, eq->eqn);
@@ -417,16 +418,15 @@ EXPORT_SYMBOL_GPL(mlx5_create_map_eq);
 
 int mlx5_destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
 {
-	struct mlx5_eq_table *table = &dev->priv.eq_table;
 	int err;
 
 	mlx5_debug_eq_remove(dev, eq);
-	free_irq(table->msix_arr[eq->irqn].vector, eq);
+	free_irq(eq->irqn, eq);
 	err = mlx5_cmd_destroy_eq(dev, eq->eqn);
 	if (err)
 		mlx5_core_warn(dev, "failed to destroy a previously created eq: eqn %d\n",
 			       eq->eqn);
-	synchronize_irq(table->msix_arr[eq->irqn].vector);
+	synchronize_irq(eq->irqn);
 	mlx5_buf_free(dev, &eq->buf);
 
 	return err;
@@ -456,7 +456,7 @@ int mlx5_start_eqs(struct mlx5_core_dev *dev)
 	u32 async_event_mask = MLX5_ASYNC_EVENT_MASK;
 	int err;
 
-	if (dev->caps.gen.flags & MLX5_DEV_CAP_FLAG_ON_DMND_PG)
+	if (MLX5_CAP_GEN(dev, pg))
 		async_event_mask |= (1ull << MLX5_EVENT_TYPE_PAGE_FAULT);
 
 	err = mlx5_create_map_eq(dev, &table->cmd_eq, MLX5_EQ_VEC_CMD,
@@ -479,7 +479,7 @@ int mlx5_start_eqs(struct mlx5_core_dev *dev)
 
 	err = mlx5_create_map_eq(dev, &table->pages_eq,
 				 MLX5_EQ_VEC_PAGES,
-				 dev->caps.gen.max_vf + 1,
+				 /* TODO: sriov max_vf + */ 1,
 				 1 << MLX5_EVENT_TYPE_PAGE_REQUEST, "mlx5_pages_eq",
 				 &dev->priv.uuari.uars[0]);
 	if (err) {
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/flow_table.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/flow_table.c
new file mode 100644
index 000000000..ca90b9bc3
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/flow_table.c
@@ -0,0 +1,422 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 <linux/export.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/flow_table.h>
+#include "mlx5_core.h"
+
+struct mlx5_ftg {
+	struct mlx5_flow_table_group    g;
+	u32				id;
+	u32				start_ix;
+};
+
+struct mlx5_flow_table {
+	struct mlx5_core_dev	*dev;
+	u8			level;
+	u8			type;
+	u32			id;
+	struct mutex		mutex; /* sync bitmap alloc */
+	u16			num_groups;
+	struct mlx5_ftg		*group;
+	unsigned long		*bitmap;
+	u32			size;
+};
+
+static int mlx5_set_flow_entry_cmd(struct mlx5_flow_table *ft, u32 group_ix,
+				   u32 flow_index, void *flow_context)
+{
+	u32 out[MLX5_ST_SZ_DW(set_fte_out)];
+	u32 *in;
+	void *in_flow_context;
+	int fcdls =
+		MLX5_GET(flow_context, flow_context, destination_list_size) *
+		MLX5_ST_SZ_BYTES(dest_format_struct);
+	int inlen = MLX5_ST_SZ_BYTES(set_fte_in) + fcdls;
+	int err;
+
+	in = mlx5_vzalloc(inlen);
+	if (!in) {
+		mlx5_core_warn(ft->dev, "failed to allocate inbox\n");
+		return -ENOMEM;
+	}
+
+	MLX5_SET(set_fte_in, in, table_type, ft->type);
+	MLX5_SET(set_fte_in, in, table_id,   ft->id);
+	MLX5_SET(set_fte_in, in, flow_index, flow_index);
+	MLX5_SET(set_fte_in, in, opcode, MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY);
+
+	in_flow_context = MLX5_ADDR_OF(set_fte_in, in, flow_context);
+	memcpy(in_flow_context, flow_context,
+	       MLX5_ST_SZ_BYTES(flow_context) + fcdls);
+
+	MLX5_SET(flow_context, in_flow_context, group_id,
+		 ft->group[group_ix].id);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(ft->dev, in, inlen, out,
+					 sizeof(out));
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5_del_flow_entry_cmd(struct mlx5_flow_table *ft, u32 flow_index)
+{
+	u32 in[MLX5_ST_SZ_DW(delete_fte_in)];
+	u32 out[MLX5_ST_SZ_DW(delete_fte_out)];
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+#define MLX5_SET_DFTEI(p, x, v) MLX5_SET(delete_fte_in, p, x, v)
+	MLX5_SET_DFTEI(in, table_type, ft->type);
+	MLX5_SET_DFTEI(in, table_id,   ft->id);
+	MLX5_SET_DFTEI(in, flow_index, flow_index);
+	MLX5_SET_DFTEI(in, opcode,     MLX5_CMD_OP_DELETE_FLOW_TABLE_ENTRY);
+
+	mlx5_cmd_exec_check_status(ft->dev, in, sizeof(in), out, sizeof(out));
+}
+
+static void mlx5_destroy_flow_group_cmd(struct mlx5_flow_table *ft, int i)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_flow_group_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_flow_group_out)];
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+#define MLX5_SET_DFGI(p, x, v) MLX5_SET(destroy_flow_group_in, p, x, v)
+	MLX5_SET_DFGI(in, table_type, ft->type);
+	MLX5_SET_DFGI(in, table_id,   ft->id);
+	MLX5_SET_DFGI(in, opcode, MLX5_CMD_OP_DESTROY_FLOW_GROUP);
+	MLX5_SET_DFGI(in, group_id, ft->group[i].id);
+	mlx5_cmd_exec_check_status(ft->dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_create_flow_group_cmd(struct mlx5_flow_table *ft, int i)
+{
+	u32 out[MLX5_ST_SZ_DW(create_flow_group_out)];
+	u32 *in;
+	void *in_match_criteria;
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_flow_table_group *g = &ft->group[i].g;
+	u32 start_ix = ft->group[i].start_ix;
+	u32 end_ix = start_ix + (1 << g->log_sz) - 1;
+	int err;
+
+	in = mlx5_vzalloc(inlen);
+	if (!in) {
+		mlx5_core_warn(ft->dev, "failed to allocate inbox\n");
+		return -ENOMEM;
+	}
+	in_match_criteria = MLX5_ADDR_OF(create_flow_group_in, in,
+					 match_criteria);
+
+	memset(out, 0, sizeof(out));
+
+#define MLX5_SET_CFGI(p, x, v) MLX5_SET(create_flow_group_in, p, x, v)
+	MLX5_SET_CFGI(in, table_type,            ft->type);
+	MLX5_SET_CFGI(in, table_id,              ft->id);
+	MLX5_SET_CFGI(in, opcode,                MLX5_CMD_OP_CREATE_FLOW_GROUP);
+	MLX5_SET_CFGI(in, start_flow_index,      start_ix);
+	MLX5_SET_CFGI(in, end_flow_index,        end_ix);
+	MLX5_SET_CFGI(in, match_criteria_enable, g->match_criteria_enable);
+
+	memcpy(in_match_criteria, g->match_criteria,
+	       MLX5_ST_SZ_BYTES(fte_match_param));
+
+	err = mlx5_cmd_exec_check_status(ft->dev, in, inlen, out,
+					 sizeof(out));
+	if (!err)
+		ft->group[i].id = MLX5_GET(create_flow_group_out, out,
+					   group_id);
+
+	kvfree(in);
+
+	return err;
+}
+
+static void mlx5_destroy_flow_table_groups(struct mlx5_flow_table *ft)
+{
+	int i;
+
+	for (i = 0; i < ft->num_groups; i++)
+		mlx5_destroy_flow_group_cmd(ft, i);
+}
+
+static int mlx5_create_flow_table_groups(struct mlx5_flow_table *ft)
+{
+	int err;
+	int i;
+
+	for (i = 0; i < ft->num_groups; i++) {
+		err = mlx5_create_flow_group_cmd(ft, i);
+		if (err)
+			goto err_destroy_flow_table_groups;
+	}
+
+	return 0;
+
+err_destroy_flow_table_groups:
+	for (i--; i >= 0; i--)
+		mlx5_destroy_flow_group_cmd(ft, i);
+
+	return err;
+}
+
+static int mlx5_create_flow_table_cmd(struct mlx5_flow_table *ft)
+{
+	u32 in[MLX5_ST_SZ_DW(create_flow_table_in)];
+	u32 out[MLX5_ST_SZ_DW(create_flow_table_out)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(create_flow_table_in, in, table_type, ft->type);
+	MLX5_SET(create_flow_table_in, in, level,      ft->level);
+	MLX5_SET(create_flow_table_in, in, log_size,   order_base_2(ft->size));
+
+	MLX5_SET(create_flow_table_in, in, opcode,
+		 MLX5_CMD_OP_CREATE_FLOW_TABLE);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(ft->dev, in, sizeof(in), out,
+					 sizeof(out));
+	if (err)
+		return err;
+
+	ft->id = MLX5_GET(create_flow_table_out, out, table_id);
+
+	return 0;
+}
+
+static void mlx5_destroy_flow_table_cmd(struct mlx5_flow_table *ft)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_flow_table_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_flow_table_out)];
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+#define MLX5_SET_DFTI(p, x, v) MLX5_SET(destroy_flow_table_in, p, x, v)
+	MLX5_SET_DFTI(in, table_type, ft->type);
+	MLX5_SET_DFTI(in, table_id,   ft->id);
+	MLX5_SET_DFTI(in, opcode, MLX5_CMD_OP_DESTROY_FLOW_TABLE);
+
+	mlx5_cmd_exec_check_status(ft->dev, in, sizeof(in), out, sizeof(out));
+}
+
+static int mlx5_find_group(struct mlx5_flow_table *ft, u8 match_criteria_enable,
+			   u32 *match_criteria, int *group_ix)
+{
+	void *mc_outer = MLX5_ADDR_OF(fte_match_param, match_criteria,
+				      outer_headers);
+	void *mc_misc  = MLX5_ADDR_OF(fte_match_param, match_criteria,
+				      misc_parameters);
+	void *mc_inner = MLX5_ADDR_OF(fte_match_param, match_criteria,
+				      inner_headers);
+	int mc_outer_sz = MLX5_ST_SZ_BYTES(fte_match_set_lyr_2_4);
+	int mc_misc_sz  = MLX5_ST_SZ_BYTES(fte_match_set_misc);
+	int mc_inner_sz = MLX5_ST_SZ_BYTES(fte_match_set_lyr_2_4);
+	int i;
+
+	for (i = 0; i < ft->num_groups; i++) {
+		struct mlx5_flow_table_group *g = &ft->group[i].g;
+		void *gmc_outer = MLX5_ADDR_OF(fte_match_param,
+					       g->match_criteria,
+					       outer_headers);
+		void *gmc_misc  = MLX5_ADDR_OF(fte_match_param,
+					       g->match_criteria,
+					       misc_parameters);
+		void *gmc_inner = MLX5_ADDR_OF(fte_match_param,
+					       g->match_criteria,
+					       inner_headers);
+
+		if (g->match_criteria_enable != match_criteria_enable)
+			continue;
+
+		if (match_criteria_enable & MLX5_MATCH_OUTER_HEADERS)
+			if (memcmp(mc_outer, gmc_outer, mc_outer_sz))
+				continue;
+
+		if (match_criteria_enable & MLX5_MATCH_MISC_PARAMETERS)
+			if (memcmp(mc_misc, gmc_misc, mc_misc_sz))
+				continue;
+
+		if (match_criteria_enable & MLX5_MATCH_INNER_HEADERS)
+			if (memcmp(mc_inner, gmc_inner, mc_inner_sz))
+				continue;
+
+		*group_ix = i;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static int alloc_flow_index(struct mlx5_flow_table *ft, int group_ix, u32 *ix)
+{
+	struct mlx5_ftg *g = &ft->group[group_ix];
+	int err = 0;
+
+	mutex_lock(&ft->mutex);
+
+	*ix = find_next_zero_bit(ft->bitmap, ft->size, g->start_ix);
+	if (*ix >= (g->start_ix + (1 << g->g.log_sz)))
+		err = -ENOSPC;
+	else
+		__set_bit(*ix, ft->bitmap);
+
+	mutex_unlock(&ft->mutex);
+
+	return err;
+}
+
+static void mlx5_free_flow_index(struct mlx5_flow_table *ft, u32 ix)
+{
+	__clear_bit(ix, ft->bitmap);
+}
+
+int mlx5_add_flow_table_entry(void *flow_table, u8 match_criteria_enable,
+			      void *match_criteria, void *flow_context,
+			      u32 *flow_index)
+{
+	struct mlx5_flow_table *ft = flow_table;
+	int group_ix;
+	int err;
+
+	err = mlx5_find_group(ft, match_criteria_enable, match_criteria,
+			      &group_ix);
+	if (err) {
+		mlx5_core_warn(ft->dev, "mlx5_find_group failed\n");
+		return err;
+	}
+
+	err = alloc_flow_index(ft, group_ix, flow_index);
+	if (err) {
+		mlx5_core_warn(ft->dev, "alloc_flow_index failed\n");
+		return err;
+	}
+
+	return mlx5_set_flow_entry_cmd(ft, group_ix, *flow_index, flow_context);
+}
+EXPORT_SYMBOL(mlx5_add_flow_table_entry);
+
+void mlx5_del_flow_table_entry(void *flow_table, u32 flow_index)
+{
+	struct mlx5_flow_table *ft = flow_table;
+
+	mlx5_del_flow_entry_cmd(ft, flow_index);
+	mlx5_free_flow_index(ft, flow_index);
+}
+EXPORT_SYMBOL(mlx5_del_flow_table_entry);
+
+void *mlx5_create_flow_table(struct mlx5_core_dev *dev, u8 level, u8 table_type,
+			     u16 num_groups,
+			     struct mlx5_flow_table_group *group)
+{
+	struct mlx5_flow_table *ft;
+	u32 start_ix = 0;
+	u32 ft_size = 0;
+	void *gr;
+	void *bm;
+	int err;
+	int i;
+
+	for (i = 0; i < num_groups; i++)
+		ft_size += (1 << group[i].log_sz);
+
+	ft = kzalloc(sizeof(*ft), GFP_KERNEL);
+	gr = kcalloc(num_groups, sizeof(struct mlx5_ftg), GFP_KERNEL);
+	bm = kcalloc(BITS_TO_LONGS(ft_size), sizeof(uintptr_t), GFP_KERNEL);
+	if (!ft || !gr || !bm)
+		goto err_free_ft;
+
+	ft->group	= gr;
+	ft->bitmap	= bm;
+	ft->num_groups	= num_groups;
+	ft->level	= level;
+	ft->type	= table_type;
+	ft->size	= ft_size;
+	ft->dev		= dev;
+	mutex_init(&ft->mutex);
+
+	for (i = 0; i < ft->num_groups; i++) {
+		memcpy(&ft->group[i].g, &group[i], sizeof(*group));
+		ft->group[i].start_ix = start_ix;
+		start_ix += 1 << group[i].log_sz;
+	}
+
+	err = mlx5_create_flow_table_cmd(ft);
+	if (err)
+		goto err_free_ft;
+
+	err = mlx5_create_flow_table_groups(ft);
+	if (err)
+		goto err_destroy_flow_table_cmd;
+
+	return ft;
+
+err_destroy_flow_table_cmd:
+	mlx5_destroy_flow_table_cmd(ft);
+
+err_free_ft:
+	mlx5_core_warn(dev, "failed to alloc flow table\n");
+	kfree(bm);
+	kfree(gr);
+	kfree(ft);
+
+	return NULL;
+}
+EXPORT_SYMBOL(mlx5_create_flow_table);
+
+void mlx5_destroy_flow_table(void *flow_table)
+{
+	struct mlx5_flow_table *ft = flow_table;
+
+	mlx5_destroy_flow_table_groups(ft);
+	mlx5_destroy_flow_table_cmd(ft);
+	kfree(ft->bitmap);
+	kfree(ft->group);
+	kfree(ft);
+}
+EXPORT_SYMBOL(mlx5_destroy_flow_table);
+
+u32 mlx5_get_flow_table_id(void *flow_table)
+{
+	struct mlx5_flow_table *ft = flow_table;
+
+	return ft->id;
+}
+EXPORT_SYMBOL(mlx5_get_flow_table_id);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/fw.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/fw.c
index 4b4cda3bc..9335e5ae1 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/fw.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/fw.c
@@ -35,79 +35,133 @@
 #include <linux/module.h>
 #include "mlx5_core.h"
 
-int mlx5_cmd_query_adapter(struct mlx5_core_dev *dev)
+static int mlx5_cmd_query_adapter(struct mlx5_core_dev *dev, u32 *out,
+				  int outlen)
 {
-	struct mlx5_cmd_query_adapter_mbox_out *out;
-	struct mlx5_cmd_query_adapter_mbox_in in;
+	u32 in[MLX5_ST_SZ_DW(query_adapter_in)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(query_adapter_in, in, opcode, MLX5_CMD_OP_QUERY_ADAPTER);
+
+	return mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, outlen);
+}
+
+int mlx5_query_board_id(struct mlx5_core_dev *dev)
+{
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_adapter_out);
 	int err;
 
-	out = kzalloc(sizeof(*out), GFP_KERNEL);
+	out = kzalloc(outlen, GFP_KERNEL);
 	if (!out)
 		return -ENOMEM;
 
-	memset(&in, 0, sizeof(in));
-	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_QUERY_ADAPTER);
-	err = mlx5_cmd_exec(dev, &in, sizeof(in), out, sizeof(*out));
+	err = mlx5_cmd_query_adapter(dev, out, outlen);
 	if (err)
-		goto out_out;
-
-	if (out->hdr.status) {
-		err = mlx5_cmd_status_to_err(&out->hdr);
-		goto out_out;
-	}
+		goto out;
 
-	memcpy(dev->board_id, out->vsd_psid, sizeof(out->vsd_psid));
+	memcpy(dev->board_id,
+	       MLX5_ADDR_OF(query_adapter_out, out,
+			    query_adapter_struct.vsd_contd_psid),
+	       MLX5_FLD_SZ_BYTES(query_adapter_out,
+				 query_adapter_struct.vsd_contd_psid));
 
-out_out:
+out:
 	kfree(out);
-
 	return err;
 }
 
-int mlx5_cmd_query_hca_cap(struct mlx5_core_dev *dev, struct mlx5_caps *caps)
+int mlx5_core_query_vendor_id(struct mlx5_core_dev *mdev, u32 *vendor_id)
 {
-	return mlx5_core_get_caps(dev, caps, HCA_CAP_OPMOD_GET_CUR);
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_adapter_out);
+	int err;
+
+	out = kzalloc(outlen, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	err = mlx5_cmd_query_adapter(mdev, out, outlen);
+	if (err)
+		goto out;
+
+	*vendor_id = MLX5_GET(query_adapter_out, out,
+			      query_adapter_struct.ieee_vendor_id);
+out:
+	kfree(out);
+	return err;
 }
+EXPORT_SYMBOL(mlx5_core_query_vendor_id);
 
-int mlx5_query_odp_caps(struct mlx5_core_dev *dev, struct mlx5_odp_caps *caps)
+int mlx5_query_hca_caps(struct mlx5_core_dev *dev)
 {
-	u8 in[MLX5_ST_SZ_BYTES(query_hca_cap_in)];
-	int out_sz = MLX5_ST_SZ_BYTES(query_hca_cap_out);
-	void *out;
 	int err;
 
-	if (!(dev->caps.gen.flags & MLX5_DEV_CAP_FLAG_ON_DMND_PG))
-		return -ENOTSUPP;
+	err = mlx5_core_get_caps(dev, MLX5_CAP_GENERAL, HCA_CAP_OPMOD_GET_CUR);
+	if (err)
+		return err;
 
-	memset(in, 0, sizeof(in));
-	out = kzalloc(out_sz, GFP_KERNEL);
-	if (!out)
-		return -ENOMEM;
-	MLX5_SET(query_hca_cap_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_CAP);
-	MLX5_SET(query_hca_cap_in, in, op_mod, HCA_CAP_OPMOD_GET_ODP_CUR);
-	err = mlx5_cmd_exec(dev, in, sizeof(in), out, out_sz);
+	err = mlx5_core_get_caps(dev, MLX5_CAP_GENERAL, HCA_CAP_OPMOD_GET_MAX);
 	if (err)
-		goto out;
+		return err;
 
-	err = mlx5_cmd_status_to_err_v2(out);
-	if (err) {
-		mlx5_core_warn(dev, "query cur hca ODP caps failed, %d\n", err);
-		goto out;
+	if (MLX5_CAP_GEN(dev, eth_net_offloads)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ETHERNET_OFFLOADS,
+					 HCA_CAP_OPMOD_GET_CUR);
+		if (err)
+			return err;
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ETHERNET_OFFLOADS,
+					 HCA_CAP_OPMOD_GET_MAX);
+		if (err)
+			return err;
 	}
 
-	memcpy(caps, MLX5_ADDR_OF(query_hca_cap_out, out, capability_struct),
-	       sizeof(*caps));
+	if (MLX5_CAP_GEN(dev, pg)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ODP,
+					 HCA_CAP_OPMOD_GET_CUR);
+		if (err)
+			return err;
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ODP,
+					 HCA_CAP_OPMOD_GET_MAX);
+		if (err)
+			return err;
+	}
 
-	mlx5_core_dbg(dev, "on-demand paging capabilities:\nrc: %08x\nuc: %08x\nud: %08x\n",
-		be32_to_cpu(caps->per_transport_caps.rc_odp_caps),
-		be32_to_cpu(caps->per_transport_caps.uc_odp_caps),
-		be32_to_cpu(caps->per_transport_caps.ud_odp_caps));
+	if (MLX5_CAP_GEN(dev, atomic)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ATOMIC,
+					 HCA_CAP_OPMOD_GET_CUR);
+		if (err)
+			return err;
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ATOMIC,
+					 HCA_CAP_OPMOD_GET_MAX);
+		if (err)
+			return err;
+	}
 
-out:
-	kfree(out);
-	return err;
+	if (MLX5_CAP_GEN(dev, roce)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ROCE,
+					 HCA_CAP_OPMOD_GET_CUR);
+		if (err)
+			return err;
+		err = mlx5_core_get_caps(dev, MLX5_CAP_ROCE,
+					 HCA_CAP_OPMOD_GET_MAX);
+		if (err)
+			return err;
+	}
+
+	if (MLX5_CAP_GEN(dev, nic_flow_table)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_FLOW_TABLE,
+					 HCA_CAP_OPMOD_GET_CUR);
+		if (err)
+			return err;
+		err = mlx5_core_get_caps(dev, MLX5_CAP_FLOW_TABLE,
+					 HCA_CAP_OPMOD_GET_MAX);
+		if (err)
+			return err;
+	}
+	return 0;
 }
-EXPORT_SYMBOL(mlx5_query_odp_caps);
 
 int mlx5_cmd_init_hca(struct mlx5_core_dev *dev)
 {
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/health.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/health.c
index 292d76f2a..f5deb642d 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/health.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/health.c
@@ -34,6 +34,7 @@
 #include <linux/module.h>
 #include <linux/random.h>
 #include <linux/vmalloc.h>
+#include <linux/hardirq.h>
 #include <linux/mlx5/driver.h>
 #include <linux/mlx5/cmd.h>
 #include "mlx5_core.h"
@@ -46,39 +47,113 @@ enum {
 enum {
 	MLX5_HEALTH_SYNDR_FW_ERR		= 0x1,
 	MLX5_HEALTH_SYNDR_IRISC_ERR		= 0x7,
+	MLX5_HEALTH_SYNDR_HW_UNRECOVERABLE_ERR	= 0x8,
 	MLX5_HEALTH_SYNDR_CRC_ERR		= 0x9,
 	MLX5_HEALTH_SYNDR_FETCH_PCI_ERR		= 0xa,
 	MLX5_HEALTH_SYNDR_HW_FTL_ERR		= 0xb,
 	MLX5_HEALTH_SYNDR_ASYNC_EQ_OVERRUN_ERR	= 0xc,
 	MLX5_HEALTH_SYNDR_EQ_ERR		= 0xd,
+	MLX5_HEALTH_SYNDR_EQ_INV		= 0xe,
 	MLX5_HEALTH_SYNDR_FFSER_ERR		= 0xf,
+	MLX5_HEALTH_SYNDR_HIGH_TEMP		= 0x10
 };
 
-static DEFINE_SPINLOCK(health_lock);
-static LIST_HEAD(health_list);
-static struct work_struct health_work;
+enum {
+	MLX5_NIC_IFC_FULL		= 0,
+	MLX5_NIC_IFC_DISABLED		= 1,
+	MLX5_NIC_IFC_NO_DRAM_NIC	= 2
+};
 
-static void health_care(struct work_struct *work)
+static u8 get_nic_interface(struct mlx5_core_dev *dev)
 {
-	struct mlx5_core_health *health, *n;
-	struct mlx5_core_dev *dev;
-	struct mlx5_priv *priv;
-	LIST_HEAD(tlist);
+	return (ioread32be(&dev->iseg->cmdq_addr_l_sz) >> 8) & 3;
+}
+
+static void trigger_cmd_completions(struct mlx5_core_dev *dev)
+{
+	unsigned long flags;
+	u64 vector;
 
-	spin_lock_irq(&health_lock);
-	list_splice_init(&health_list, &tlist);
+	/* wait for pending handlers to complete */
+	synchronize_irq(dev->priv.msix_arr[MLX5_EQ_VEC_CMD].vector);
+	spin_lock_irqsave(&dev->cmd.alloc_lock, flags);
+	vector = ~dev->cmd.bitmask & ((1ul << (1 << dev->cmd.log_sz)) - 1);
+	if (!vector)
+		goto no_trig;
+
+	vector |= MLX5_TRIGGERED_CMD_COMP;
+	spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags);
+
+	mlx5_core_dbg(dev, "vector 0x%llx\n", vector);
+	mlx5_cmd_comp_handler(dev, vector);
+	return;
+
+no_trig:
+	spin_unlock_irqrestore(&dev->cmd.alloc_lock, flags);
+}
+
+static int in_fatal(struct mlx5_core_dev *dev)
+{
+	struct mlx5_core_health *health = &dev->priv.health;
+	struct health_buffer __iomem *h = health->health;
 
-	spin_unlock_irq(&health_lock);
+	if (get_nic_interface(dev) == MLX5_NIC_IFC_DISABLED)
+		return 1;
 
-	list_for_each_entry_safe(health, n, &tlist, list) {
-		priv = container_of(health, struct mlx5_priv, health);
-		dev = container_of(priv, struct mlx5_core_dev, priv);
-		mlx5_core_warn(dev, "handling bad device here\n");
-		/* nothing yet */
-		spin_lock_irq(&health_lock);
-		list_del_init(&health->list);
-		spin_unlock_irq(&health_lock);
+	if (ioread32be(&h->fw_ver) == 0xffffffff)
+		return 1;
+
+	return 0;
+}
+
+void mlx5_enter_error_state(struct mlx5_core_dev *dev)
+{
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
+		return;
+
+	mlx5_core_err(dev, "start\n");
+	if (pci_channel_offline(dev->pdev) || in_fatal(dev))
+		dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+
+	mlx5_core_event(dev, MLX5_DEV_EVENT_SYS_ERROR, 0);
+	mlx5_core_err(dev, "end\n");
+}
+
+static void mlx5_handle_bad_state(struct mlx5_core_dev *dev)
+{
+	u8 nic_interface = get_nic_interface(dev);
+
+	switch (nic_interface) {
+	case MLX5_NIC_IFC_FULL:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is full driver\n");
+		break;
+
+	case MLX5_NIC_IFC_DISABLED:
+		mlx5_core_warn(dev, "starting teardown\n");
+		break;
+
+	case MLX5_NIC_IFC_NO_DRAM_NIC:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is no dram nic\n");
+		break;
+	default:
+		mlx5_core_warn(dev, "Expected to see disabled NIC but it is has invalid value %d\n",
+			       nic_interface);
 	}
+
+	mlx5_disable_device(dev);
+}
+
+static void health_care(struct work_struct *work)
+{
+	struct mlx5_core_health *health;
+	struct mlx5_core_dev *dev;
+	struct mlx5_priv *priv;
+
+	health = container_of(work, struct mlx5_core_health, work);
+	priv = container_of(health, struct mlx5_priv, health);
+	dev = container_of(priv, struct mlx5_core_dev, priv);
+	mlx5_core_warn(dev, "handling bad device here\n");
+	mlx5_handle_bad_state(dev);
 }
 
 static const char *hsynd_str(u8 synd)
@@ -88,6 +163,8 @@ static const char *hsynd_str(u8 synd)
 		return "firmware internal error";
 	case MLX5_HEALTH_SYNDR_IRISC_ERR:
 		return "irisc not responding";
+	case MLX5_HEALTH_SYNDR_HW_UNRECOVERABLE_ERR:
+		return "unrecoverable hardware error";
 	case MLX5_HEALTH_SYNDR_CRC_ERR:
 		return "firmware CRC error";
 	case MLX5_HEALTH_SYNDR_FETCH_PCI_ERR:
@@ -98,48 +175,81 @@ static const char *hsynd_str(u8 synd)
 		return "async EQ buffer overrun";
 	case MLX5_HEALTH_SYNDR_EQ_ERR:
 		return "EQ error";
+	case MLX5_HEALTH_SYNDR_EQ_INV:
+		return "Invalid EQ refrenced";
 	case MLX5_HEALTH_SYNDR_FFSER_ERR:
 		return "FFSER error";
+	case MLX5_HEALTH_SYNDR_HIGH_TEMP:
+		return "High temprature";
 	default:
 		return "unrecognized error";
 	}
 }
 
-static u16 read_be16(__be16 __iomem *p)
+static u16 get_maj(u32 fw)
 {
-	return swab16(readl((__force u16 __iomem *) p));
+	return fw >> 28;
 }
 
-static u32 read_be32(__be32 __iomem *p)
+static u16 get_min(u32 fw)
 {
-	return swab32(readl((__force u32 __iomem *) p));
+	return fw >> 16 & 0xfff;
+}
+
+static u16 get_sub(u32 fw)
+{
+	return fw & 0xffff;
 }
 
 static void print_health_info(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_health *health = &dev->priv.health;
 	struct health_buffer __iomem *h = health->health;
+	char fw_str[18];
+	u32 fw;
 	int i;
 
+	/* If the syndrom is 0, the device is OK and no need to print buffer */
+	if (!ioread8(&h->synd))
+		return;
+
 	for (i = 0; i < ARRAY_SIZE(h->assert_var); i++)
-		pr_info("assert_var[%d] 0x%08x\n", i, read_be32(h->assert_var + i));
+		dev_err(&dev->pdev->dev, "assert_var[%d] 0x%08x\n", i, ioread32be(h->assert_var + i));
+
+	dev_err(&dev->pdev->dev, "assert_exit_ptr 0x%08x\n", ioread32be(&h->assert_exit_ptr));
+	dev_err(&dev->pdev->dev, "assert_callra 0x%08x\n", ioread32be(&h->assert_callra));
+	fw = ioread32be(&h->fw_ver);
+	sprintf(fw_str, "%d.%d.%d", get_maj(fw), get_min(fw), get_sub(fw));
+	dev_err(&dev->pdev->dev, "fw_ver %s\n", fw_str);
+	dev_err(&dev->pdev->dev, "hw_id 0x%08x\n", ioread32be(&h->hw_id));
+	dev_err(&dev->pdev->dev, "irisc_index %d\n", ioread8(&h->irisc_index));
+	dev_err(&dev->pdev->dev, "synd 0x%x: %s\n", ioread8(&h->synd), hsynd_str(ioread8(&h->synd)));
+	dev_err(&dev->pdev->dev, "ext_synd 0x%04x\n", ioread16be(&h->ext_synd));
+}
+
+static unsigned long get_next_poll_jiffies(void)
+{
+	unsigned long next;
 
-	pr_info("assert_exit_ptr 0x%08x\n", read_be32(&h->assert_exit_ptr));
-	pr_info("assert_callra 0x%08x\n", read_be32(&h->assert_callra));
-	pr_info("fw_ver 0x%08x\n", read_be32(&h->fw_ver));
-	pr_info("hw_id 0x%08x\n", read_be32(&h->hw_id));
-	pr_info("irisc_index %d\n", readb(&h->irisc_index));
-	pr_info("synd 0x%x: %s\n", readb(&h->synd), hsynd_str(readb(&h->synd)));
-	pr_info("ext_sync 0x%04x\n", read_be16(&h->ext_sync));
+	get_random_bytes(&next, sizeof(next));
+	next %= HZ;
+	next += jiffies + MLX5_HEALTH_POLL_INTERVAL;
+
+	return next;
 }
 
 static void poll_health(unsigned long data)
 {
 	struct mlx5_core_dev *dev = (struct mlx5_core_dev *)data;
 	struct mlx5_core_health *health = &dev->priv.health;
-	unsigned long next;
 	u32 count;
 
+	if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+		trigger_cmd_completions(dev);
+		mod_timer(&health->timer, get_next_poll_jiffies());
+		return;
+	}
+
 	count = ioread32be(health->health_counter);
 	if (count == health->prev)
 		++health->miss_counter;
@@ -148,18 +258,16 @@ static void poll_health(unsigned long data)
 
 	health->prev = count;
 	if (health->miss_counter == MAX_MISSES) {
-		mlx5_core_err(dev, "device's health compromised\n");
+		dev_err(&dev->pdev->dev, "device's health compromised - reached miss count\n");
 		print_health_info(dev);
-		spin_lock_irq(&health_lock);
-		list_add_tail(&health->list, &health_list);
-		spin_unlock_irq(&health_lock);
-
-		queue_work(mlx5_core_wq, &health_work);
 	} else {
-		get_random_bytes(&next, sizeof(next));
-		next %= HZ;
-		next += jiffies + MLX5_HEALTH_POLL_INTERVAL;
-		mod_timer(&health->timer, next);
+		mod_timer(&health->timer, get_next_poll_jiffies());
+	}
+
+	if (in_fatal(dev) && !health->sick) {
+		health->sick = true;
+		print_health_info(dev);
+		queue_work(health->wq, &health->work);
 	}
 }
 
@@ -167,7 +275,6 @@ void mlx5_start_health_poll(struct mlx5_core_dev *dev)
 {
 	struct mlx5_core_health *health = &dev->priv.health;
 
-	INIT_LIST_HEAD(&health->list);
 	init_timer(&health->timer);
 	health->health = &dev->iseg->health;
 	health->health_counter = &dev->iseg->health_counter;
@@ -183,18 +290,33 @@ void mlx5_stop_health_poll(struct mlx5_core_dev *dev)
 	struct mlx5_core_health *health = &dev->priv.health;
 
 	del_timer_sync(&health->timer);
-
-	spin_lock_irq(&health_lock);
-	if (!list_empty(&health->list))
-		list_del_init(&health->list);
-	spin_unlock_irq(&health_lock);
 }
 
-void mlx5_health_cleanup(void)
+void mlx5_health_cleanup(struct mlx5_core_dev *dev)
 {
+	struct mlx5_core_health *health = &dev->priv.health;
+
+	destroy_workqueue(health->wq);
 }
 
-void  __init mlx5_health_init(void)
+int mlx5_health_init(struct mlx5_core_dev *dev)
 {
-	INIT_WORK(&health_work, health_care);
+	struct mlx5_core_health *health;
+	char *name;
+
+	health = &dev->priv.health;
+	name = kmalloc(64, GFP_KERNEL);
+	if (!name)
+		return -ENOMEM;
+
+	strcpy(name, "mlx5_health");
+	strcat(name, dev_name(&dev->pdev->dev));
+	health->wq = create_singlethread_workqueue(name);
+	kfree(name);
+	if (!health->wq)
+		return -ENOMEM;
+
+	INIT_WORK(&health->work, health_care);
+
+	return 0;
 }
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mad.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mad.c
index ee1b0b965..1368dac00 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mad.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mad.c
@@ -36,7 +36,7 @@
 #include <linux/mlx5/cmd.h>
 #include "mlx5_core.h"
 
-int mlx5_core_mad_ifc(struct mlx5_core_dev *dev, void *inb, void *outb,
+int mlx5_core_mad_ifc(struct mlx5_core_dev *dev, const void *inb, void *outb,
 		      u16 opmod, u8 port)
 {
 	struct mlx5_mad_ifc_mbox_in *in = NULL;
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/main.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 28425e5ea..6cf6d93d8 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -30,7 +30,7 @@
  * SOFTWARE.
  */
 
-#include <asm-generic/kmap_types.h>
+#include <linux/highmem.h>
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/errno.h>
@@ -38,19 +38,18 @@
 #include <linux/dma-mapping.h>
 #include <linux/slab.h>
 #include <linux/io-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
 #include <linux/mlx5/driver.h>
 #include <linux/mlx5/cq.h>
 #include <linux/mlx5/qp.h>
 #include <linux/mlx5/srq.h>
 #include <linux/debugfs.h>
 #include <linux/kmod.h>
+#include <linux/delay.h>
 #include <linux/mlx5/mlx5_ifc.h>
 #include "mlx5_core.h"
 
-#define DRIVER_NAME "mlx5_core"
-#define DRIVER_VERSION "3.0"
-#define DRIVER_RELDATE  "January 2015"
-
 MODULE_AUTHOR("Eli Cohen <eli@mellanox.com>");
 MODULE_DESCRIPTION("Mellanox Connect-IB, ConnectX-4 core driver");
 MODULE_LICENSE("Dual BSD/GPL");
@@ -65,7 +64,6 @@ static int prof_sel = MLX5_DEFAULT_PROF;
 module_param_named(prof_sel, prof_sel, int, 0444);
 MODULE_PARM_DESC(prof_sel, "profile selector. Valid range 0 - 2");
 
-struct workqueue_struct *mlx5_core_wq;
 static LIST_HEAD(intf_list);
 static LIST_HEAD(dev_list);
 static DEFINE_MUTEX(intf_mutex);
@@ -155,6 +153,25 @@ static struct mlx5_profile profile[] = {
 	},
 };
 
+#define FW_INIT_TIMEOUT_MILI	2000
+#define FW_INIT_WAIT_MS		2
+
+static int wait_fw_init(struct mlx5_core_dev *dev, u32 max_wait_mili)
+{
+	unsigned long end = jiffies + msecs_to_jiffies(max_wait_mili);
+	int err = 0;
+
+	while (fw_initializing(dev)) {
+		if (time_after(jiffies, end)) {
+			err = -EBUSY;
+			break;
+		}
+		msleep(FW_INIT_WAIT_MS);
+	}
+
+	return err;
+}
+
 static int set_dma_caps(struct pci_dev *pdev)
 {
 	int err;
@@ -185,6 +202,34 @@ static int set_dma_caps(struct pci_dev *pdev)
 	return err;
 }
 
+static int mlx5_pci_enable_device(struct mlx5_core_dev *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err = 0;
+
+	mutex_lock(&dev->pci_status_mutex);
+	if (dev->pci_status == MLX5_PCI_STATUS_DISABLED) {
+		err = pci_enable_device(pdev);
+		if (!err)
+			dev->pci_status = MLX5_PCI_STATUS_ENABLED;
+	}
+	mutex_unlock(&dev->pci_status_mutex);
+
+	return err;
+}
+
+static void mlx5_pci_disable_device(struct mlx5_core_dev *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+
+	mutex_lock(&dev->pci_status_mutex);
+	if (dev->pci_status == MLX5_PCI_STATUS_ENABLED) {
+		pci_disable_device(pdev);
+		dev->pci_status = MLX5_PCI_STATUS_DISABLED;
+	}
+	mutex_unlock(&dev->pci_status_mutex);
+}
+
 static int request_bar(struct pci_dev *pdev)
 {
 	int err = 0;
@@ -208,24 +253,28 @@ static void release_bar(struct pci_dev *pdev)
 
 static int mlx5_enable_msix(struct mlx5_core_dev *dev)
 {
-	struct mlx5_eq_table *table = &dev->priv.eq_table;
-	int num_eqs = 1 << dev->caps.gen.log_max_eq;
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_eq_table *table = &priv->eq_table;
+	int num_eqs = 1 << MLX5_CAP_GEN(dev, log_max_eq);
 	int nvec;
 	int i;
 
-	nvec = dev->caps.gen.num_ports * num_online_cpus() + MLX5_EQ_VEC_COMP_BASE;
+	nvec = MLX5_CAP_GEN(dev, num_ports) * num_online_cpus() +
+	       MLX5_EQ_VEC_COMP_BASE;
 	nvec = min_t(int, nvec, num_eqs);
 	if (nvec <= MLX5_EQ_VEC_COMP_BASE)
 		return -ENOMEM;
 
-	table->msix_arr = kzalloc(nvec * sizeof(*table->msix_arr), GFP_KERNEL);
-	if (!table->msix_arr)
-		return -ENOMEM;
+	priv->msix_arr = kcalloc(nvec, sizeof(*priv->msix_arr), GFP_KERNEL);
+
+	priv->irq_info = kcalloc(nvec, sizeof(*priv->irq_info), GFP_KERNEL);
+	if (!priv->msix_arr || !priv->irq_info)
+		goto err_free_msix;
 
 	for (i = 0; i < nvec; i++)
-		table->msix_arr[i].entry = i;
+		priv->msix_arr[i].entry = i;
 
-	nvec = pci_enable_msix_range(dev->pdev, table->msix_arr,
+	nvec = pci_enable_msix_range(dev->pdev, priv->msix_arr,
 				     MLX5_EQ_VEC_COMP_BASE + 1, nvec);
 	if (nvec < 0)
 		return nvec;
@@ -233,14 +282,20 @@ static int mlx5_enable_msix(struct mlx5_core_dev *dev)
 	table->num_comp_vectors = nvec - MLX5_EQ_VEC_COMP_BASE;
 
 	return 0;
+
+err_free_msix:
+	kfree(priv->irq_info);
+	kfree(priv->msix_arr);
+	return -ENOMEM;
 }
 
 static void mlx5_disable_msix(struct mlx5_core_dev *dev)
 {
-	struct mlx5_eq_table *table = &dev->priv.eq_table;
+	struct mlx5_priv *priv = &dev->priv;
 
 	pci_disable_msix(dev->pdev);
-	kfree(table->msix_arr);
+	kfree(priv->irq_info);
+	kfree(priv->msix_arr);
 }
 
 struct mlx5_reg_host_endianess {
@@ -277,98 +332,20 @@ static u16 to_fw_pkey_sz(u32 size)
 	}
 }
 
-/* selectively copy writable fields clearing any reserved area
- */
-static void copy_rw_fields(void *to, struct mlx5_caps *from)
-{
-	__be64 *flags_off = (__be64 *)MLX5_ADDR_OF(cmd_hca_cap, to, reserved_22);
-	u64 v64;
-
-	MLX5_SET(cmd_hca_cap, to, log_max_qp, from->gen.log_max_qp);
-	MLX5_SET(cmd_hca_cap, to, log_max_ra_req_qp, from->gen.log_max_ra_req_qp);
-	MLX5_SET(cmd_hca_cap, to, log_max_ra_res_qp, from->gen.log_max_ra_res_qp);
-	MLX5_SET(cmd_hca_cap, to, pkey_table_size, from->gen.pkey_table_size);
-	MLX5_SET(cmd_hca_cap, to, pkey_table_size, to_fw_pkey_sz(from->gen.pkey_table_size));
-	MLX5_SET(cmd_hca_cap, to, log_uar_page_sz, PAGE_SHIFT - 12);
-	v64 = from->gen.flags & MLX5_CAP_BITS_RW_MASK;
-	*flags_off = cpu_to_be64(v64);
-}
-
-static u16 get_pkey_table_size(int pkey)
-{
-	if (pkey > MLX5_MAX_LOG_PKEY_TABLE)
-		return 0;
-
-	return MLX5_MIN_PKEY_TABLE_SIZE << pkey;
-}
-
-static void fw2drv_caps(struct mlx5_caps *caps, void *out)
-{
-	struct mlx5_general_caps *gen = &caps->gen;
-
-	gen->max_srq_wqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_srq_sz);
-	gen->max_wqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_qp_sz);
-	gen->log_max_qp = MLX5_GET_PR(cmd_hca_cap, out, log_max_qp);
-	gen->log_max_strq = MLX5_GET_PR(cmd_hca_cap, out, log_max_strq_sz);
-	gen->log_max_srq = MLX5_GET_PR(cmd_hca_cap, out, log_max_srqs);
-	gen->max_cqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_cq_sz);
-	gen->log_max_cq = MLX5_GET_PR(cmd_hca_cap, out, log_max_cq);
-	gen->max_eqes = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_max_eq_sz);
-	gen->log_max_mkey = MLX5_GET_PR(cmd_hca_cap, out, log_max_mkey);
-	gen->log_max_eq = MLX5_GET_PR(cmd_hca_cap, out, log_max_eq);
-	gen->max_indirection = MLX5_GET_PR(cmd_hca_cap, out, max_indirection);
-	gen->log_max_mrw_sz = MLX5_GET_PR(cmd_hca_cap, out, log_max_mrw_sz);
-	gen->log_max_bsf_list_size = MLX5_GET_PR(cmd_hca_cap, out, log_max_bsf_list_size);
-	gen->log_max_klm_list_size = MLX5_GET_PR(cmd_hca_cap, out, log_max_klm_list_size);
-	gen->log_max_ra_req_dc = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_req_dc);
-	gen->log_max_ra_res_dc = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_res_dc);
-	gen->log_max_ra_req_qp = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_req_qp);
-	gen->log_max_ra_res_qp = MLX5_GET_PR(cmd_hca_cap, out, log_max_ra_res_qp);
-	gen->max_qp_counters = MLX5_GET_PR(cmd_hca_cap, out, max_qp_cnt);
-	gen->pkey_table_size = get_pkey_table_size(MLX5_GET_PR(cmd_hca_cap, out, pkey_table_size));
-	gen->local_ca_ack_delay = MLX5_GET_PR(cmd_hca_cap, out, local_ca_ack_delay);
-	gen->num_ports = MLX5_GET_PR(cmd_hca_cap, out, num_ports);
-	gen->log_max_msg = MLX5_GET_PR(cmd_hca_cap, out, log_max_msg);
-	gen->stat_rate_support = MLX5_GET_PR(cmd_hca_cap, out, stat_rate_support);
-	gen->flags = be64_to_cpu(*(__be64 *)MLX5_ADDR_OF(cmd_hca_cap, out, reserved_22));
-	pr_debug("flags = 0x%llx\n", gen->flags);
-	gen->uar_sz = MLX5_GET_PR(cmd_hca_cap, out, uar_sz);
-	gen->min_log_pg_sz = MLX5_GET_PR(cmd_hca_cap, out, log_pg_sz);
-	gen->bf_reg_size = MLX5_GET_PR(cmd_hca_cap, out, bf);
-	gen->bf_reg_size = 1 << MLX5_GET_PR(cmd_hca_cap, out, log_bf_reg_size);
-	gen->max_sq_desc_sz = MLX5_GET_PR(cmd_hca_cap, out, max_wqe_sz_sq);
-	gen->max_rq_desc_sz = MLX5_GET_PR(cmd_hca_cap, out, max_wqe_sz_rq);
-	gen->max_dc_sq_desc_sz = MLX5_GET_PR(cmd_hca_cap, out, max_wqe_sz_sq_dc);
-	gen->max_qp_mcg = MLX5_GET_PR(cmd_hca_cap, out, max_qp_mcg);
-	gen->log_max_pd = MLX5_GET_PR(cmd_hca_cap, out, log_max_pd);
-	gen->log_max_xrcd = MLX5_GET_PR(cmd_hca_cap, out, log_max_xrcd);
-	gen->log_uar_page_sz = MLX5_GET_PR(cmd_hca_cap, out, log_uar_page_sz);
-}
-
-static const char *caps_opmod_str(u16 opmod)
-{
-	switch (opmod) {
-	case HCA_CAP_OPMOD_GET_MAX:
-		return "GET_MAX";
-	case HCA_CAP_OPMOD_GET_CUR:
-		return "GET_CUR";
-	default:
-		return "Invalid";
-	}
-}
-
-int mlx5_core_get_caps(struct mlx5_core_dev *dev, struct mlx5_caps *caps,
-		       u16 opmod)
+int mlx5_core_get_caps(struct mlx5_core_dev *dev, enum mlx5_cap_type cap_type,
+		       enum mlx5_cap_mode cap_mode)
 {
 	u8 in[MLX5_ST_SZ_BYTES(query_hca_cap_in)];
 	int out_sz = MLX5_ST_SZ_BYTES(query_hca_cap_out);
-	void *out;
+	void *out, *hca_caps;
+	u16 opmod = (cap_type << 1) | (cap_mode & 0x01);
 	int err;
 
 	memset(in, 0, sizeof(in));
 	out = kzalloc(out_sz, GFP_KERNEL);
 	if (!out)
 		return -ENOMEM;
+
 	MLX5_SET(query_hca_cap_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_CAP);
 	MLX5_SET(query_hca_cap_in, in, op_mod, opmod);
 	err = mlx5_cmd_exec(dev, in, sizeof(in), out, out_sz);
@@ -377,12 +354,30 @@ int mlx5_core_get_caps(struct mlx5_core_dev *dev, struct mlx5_caps *caps,
 
 	err = mlx5_cmd_status_to_err_v2(out);
 	if (err) {
-		mlx5_core_warn(dev, "query max hca cap failed, %d\n", err);
+		mlx5_core_warn(dev,
+			       "QUERY_HCA_CAP : type(%x) opmode(%x) Failed(%d)\n",
+			       cap_type, cap_mode, err);
 		goto query_ex;
 	}
-	mlx5_core_dbg(dev, "%s\n", caps_opmod_str(opmod));
-	fw2drv_caps(caps, MLX5_ADDR_OF(query_hca_cap_out, out, capability_struct));
 
+	hca_caps =  MLX5_ADDR_OF(query_hca_cap_out, out, capability);
+
+	switch (cap_mode) {
+	case HCA_CAP_OPMOD_GET_MAX:
+		memcpy(dev->hca_caps_max[cap_type], hca_caps,
+		       MLX5_UN_SZ_BYTES(hca_cap_union));
+		break;
+	case HCA_CAP_OPMOD_GET_CUR:
+		memcpy(dev->hca_caps_cur[cap_type], hca_caps,
+		       MLX5_UN_SZ_BYTES(hca_cap_union));
+		break;
+	default:
+		mlx5_core_warn(dev,
+			       "Tried to query dev cap type(%x) with wrong opmode(%x)\n",
+			       cap_type, cap_mode);
+		err = -EINVAL;
+		break;
+	}
 query_ex:
 	kfree(out);
 	return err;
@@ -409,49 +404,47 @@ static int handle_hca_cap(struct mlx5_core_dev *dev)
 {
 	void *set_ctx = NULL;
 	struct mlx5_profile *prof = dev->profile;
-	struct mlx5_caps *cur_caps = NULL;
-	struct mlx5_caps *max_caps = NULL;
 	int err = -ENOMEM;
 	int set_sz = MLX5_ST_SZ_BYTES(set_hca_cap_in);
+	void *set_hca_cap;
 
 	set_ctx = kzalloc(set_sz, GFP_KERNEL);
 	if (!set_ctx)
 		goto query_ex;
 
-	max_caps = kzalloc(sizeof(*max_caps), GFP_KERNEL);
-	if (!max_caps)
-		goto query_ex;
-
-	cur_caps = kzalloc(sizeof(*cur_caps), GFP_KERNEL);
-	if (!cur_caps)
-		goto query_ex;
-
-	err = mlx5_core_get_caps(dev, max_caps, HCA_CAP_OPMOD_GET_MAX);
+	err = mlx5_core_get_caps(dev, MLX5_CAP_GENERAL, HCA_CAP_OPMOD_GET_MAX);
 	if (err)
 		goto query_ex;
 
-	err = mlx5_core_get_caps(dev, cur_caps, HCA_CAP_OPMOD_GET_CUR);
+	err = mlx5_core_get_caps(dev, MLX5_CAP_GENERAL, HCA_CAP_OPMOD_GET_CUR);
 	if (err)
 		goto query_ex;
 
+	set_hca_cap = MLX5_ADDR_OF(set_hca_cap_in, set_ctx,
+				   capability);
+	memcpy(set_hca_cap, dev->hca_caps_cur[MLX5_CAP_GENERAL],
+	       MLX5_ST_SZ_BYTES(cmd_hca_cap));
+
+	mlx5_core_dbg(dev, "Current Pkey table size %d Setting new size %d\n",
+		      mlx5_to_sw_pkey_sz(MLX5_CAP_GEN(dev, pkey_table_size)),
+		      128);
 	/* we limit the size of the pkey table to 128 entries for now */
-	cur_caps->gen.pkey_table_size = 128;
+	MLX5_SET(cmd_hca_cap, set_hca_cap, pkey_table_size,
+		 to_fw_pkey_sz(128));
 
 	if (prof->mask & MLX5_PROF_MASK_QP_SIZE)
-		cur_caps->gen.log_max_qp = prof->log_max_qp;
+		MLX5_SET(cmd_hca_cap, set_hca_cap, log_max_qp,
+			 prof->log_max_qp);
 
-	/* disable checksum */
-	cur_caps->gen.flags &= ~MLX5_DEV_CAP_FLAG_CMDIF_CSUM;
+	/* disable cmdif checksum */
+	MLX5_SET(cmd_hca_cap, set_hca_cap, cmdif_checksum, 0);
+
+	MLX5_SET(cmd_hca_cap, set_hca_cap, log_uar_page_sz, PAGE_SHIFT - 12);
 
-	copy_rw_fields(MLX5_ADDR_OF(set_hca_cap_in, set_ctx, hca_capability_struct),
-		       cur_caps);
 	err = set_caps(dev, set_ctx, set_sz);
 
 query_ex:
-	kfree(cur_caps);
-	kfree(max_caps);
 	kfree(set_ctx);
-
 	return err;
 }
 
@@ -507,7 +500,76 @@ static int mlx5_core_disable_hca(struct mlx5_core_dev *dev)
 	return 0;
 }
 
-int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn, int *irqn)
+static int mlx5_irq_set_affinity_hint(struct mlx5_core_dev *mdev, int i)
+{
+	struct mlx5_priv *priv  = &mdev->priv;
+	struct msix_entry *msix = priv->msix_arr;
+	int irq                 = msix[i + MLX5_EQ_VEC_COMP_BASE].vector;
+	int numa_node           = priv->numa_node;
+	int err;
+
+	if (!zalloc_cpumask_var(&priv->irq_info[i].mask, GFP_KERNEL)) {
+		mlx5_core_warn(mdev, "zalloc_cpumask_var failed");
+		return -ENOMEM;
+	}
+
+	cpumask_set_cpu(cpumask_local_spread(i, numa_node),
+			priv->irq_info[i].mask);
+
+	err = irq_set_affinity_hint(irq, priv->irq_info[i].mask);
+	if (err) {
+		mlx5_core_warn(mdev, "irq_set_affinity_hint failed,irq 0x%.4x",
+			       irq);
+		goto err_clear_mask;
+	}
+
+	return 0;
+
+err_clear_mask:
+	free_cpumask_var(priv->irq_info[i].mask);
+	return err;
+}
+
+static void mlx5_irq_clear_affinity_hint(struct mlx5_core_dev *mdev, int i)
+{
+	struct mlx5_priv *priv  = &mdev->priv;
+	struct msix_entry *msix = priv->msix_arr;
+	int irq                 = msix[i + MLX5_EQ_VEC_COMP_BASE].vector;
+
+	irq_set_affinity_hint(irq, NULL);
+	free_cpumask_var(priv->irq_info[i].mask);
+}
+
+static int mlx5_irq_set_affinity_hints(struct mlx5_core_dev *mdev)
+{
+	int err;
+	int i;
+
+	for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++) {
+		err = mlx5_irq_set_affinity_hint(mdev, i);
+		if (err)
+			goto err_out;
+	}
+
+	return 0;
+
+err_out:
+	for (i--; i >= 0; i--)
+		mlx5_irq_clear_affinity_hint(mdev, i);
+
+	return err;
+}
+
+static void mlx5_irq_clear_affinity_hints(struct mlx5_core_dev *mdev)
+{
+	int i;
+
+	for (i = 0; i < mdev->priv.eq_table.num_comp_vectors; i++)
+		mlx5_irq_clear_affinity_hint(mdev, i);
+}
+
+int mlx5_vector2eqn(struct mlx5_core_dev *dev, int vector, int *eqn,
+		    unsigned int *irqn)
 {
 	struct mlx5_eq_table *table = &dev->priv.eq_table;
 	struct mlx5_eq *eq, *n;
@@ -549,7 +611,7 @@ static void free_comp_eqs(struct mlx5_core_dev *dev)
 static int alloc_comp_eqs(struct mlx5_core_dev *dev)
 {
 	struct mlx5_eq_table *table = &dev->priv.eq_table;
-	char name[MLX5_MAX_EQ_NAME];
+	char name[MLX5_MAX_IRQ_NAME];
 	struct mlx5_eq *eq;
 	int ncomp_vec;
 	int nent;
@@ -566,7 +628,7 @@ static int alloc_comp_eqs(struct mlx5_core_dev *dev)
 			goto clean;
 		}
 
-		snprintf(name, MLX5_MAX_EQ_NAME, "mlx5_comp%d", i);
+		snprintf(name, MLX5_MAX_IRQ_NAME, "mlx5_comp%d", i);
 		err = mlx5_create_map_eq(dev, eq,
 					 i + MLX5_EQ_VEC_COMP_BASE, nent, 0,
 					 name, &dev->priv.uuari.uars[0]);
@@ -588,12 +650,197 @@ clean:
 	return err;
 }
 
-static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
+#ifdef CONFIG_MLX5_CORE_EN
+static int mlx5_core_set_issi(struct mlx5_core_dev *dev)
 {
-	struct mlx5_priv *priv = &dev->priv;
+	u32 query_in[MLX5_ST_SZ_DW(query_issi_in)];
+	u32 query_out[MLX5_ST_SZ_DW(query_issi_out)];
+	u32 set_in[MLX5_ST_SZ_DW(set_issi_in)];
+	u32 set_out[MLX5_ST_SZ_DW(set_issi_out)];
 	int err;
+	u32 sup_issi;
+
+	memset(query_in, 0, sizeof(query_in));
+	memset(query_out, 0, sizeof(query_out));
+
+	MLX5_SET(query_issi_in, query_in, opcode, MLX5_CMD_OP_QUERY_ISSI);
+
+	err = mlx5_cmd_exec_check_status(dev, query_in, sizeof(query_in),
+					 query_out, sizeof(query_out));
+	if (err) {
+		if (((struct mlx5_outbox_hdr *)query_out)->status ==
+		    MLX5_CMD_STAT_BAD_OP_ERR) {
+			pr_debug("Only ISSI 0 is supported\n");
+			return 0;
+		}
+
+		pr_err("failed to query ISSI\n");
+		return err;
+	}
+
+	sup_issi = MLX5_GET(query_issi_out, query_out, supported_issi_dw0);
+
+	if (sup_issi & (1 << 1)) {
+		memset(set_in, 0, sizeof(set_in));
+		memset(set_out, 0, sizeof(set_out));
+
+		MLX5_SET(set_issi_in, set_in, opcode, MLX5_CMD_OP_SET_ISSI);
+		MLX5_SET(set_issi_in, set_in, current_issi, 1);
+
+		err = mlx5_cmd_exec_check_status(dev, set_in, sizeof(set_in),
+						 set_out, sizeof(set_out));
+		if (err) {
+			pr_err("failed to set ISSI=1\n");
+			return err;
+		}
+
+		dev->issi = 1;
+
+		return 0;
+	} else if (sup_issi & (1 << 0) || !sup_issi) {
+		return 0;
+	}
+
+	return -ENOTSUPP;
+}
+#endif
+
+static int map_bf_area(struct mlx5_core_dev *dev)
+{
+	resource_size_t bf_start = pci_resource_start(dev->pdev, 0);
+	resource_size_t bf_len = pci_resource_len(dev->pdev, 0);
+
+	dev->priv.bf_mapping = io_mapping_create_wc(bf_start, bf_len);
+
+	return dev->priv.bf_mapping ? 0 : -ENOMEM;
+}
+
+static void unmap_bf_area(struct mlx5_core_dev *dev)
+{
+	if (dev->priv.bf_mapping)
+		io_mapping_free(dev->priv.bf_mapping);
+}
+
+static void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	dev_ctx = kmalloc(sizeof(*dev_ctx), GFP_KERNEL);
+	if (!dev_ctx)
+		return;
+
+	dev_ctx->intf    = intf;
+	dev_ctx->context = intf->add(dev);
+
+	if (dev_ctx->context) {
+		spin_lock_irq(&priv->ctx_lock);
+		list_add_tail(&dev_ctx->list, &priv->ctx_list);
+		spin_unlock_irq(&priv->ctx_lock);
+	} else {
+		kfree(dev_ctx);
+	}
+}
+
+static void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
+{
+	struct mlx5_device_context *dev_ctx;
+	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
+
+	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
+		if (dev_ctx->intf == intf) {
+			spin_lock_irq(&priv->ctx_lock);
+			list_del(&dev_ctx->list);
+			spin_unlock_irq(&priv->ctx_lock);
+
+			intf->remove(dev, dev_ctx->context);
+			kfree(dev_ctx);
+			return;
+		}
+}
+
+static int mlx5_register_device(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_interface *intf;
+
+	mutex_lock(&intf_mutex);
+	list_add_tail(&priv->dev_list, &dev_list);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx5_add_device(intf, priv);
+	mutex_unlock(&intf_mutex);
+
+	return 0;
+}
+
+static void mlx5_unregister_device(struct mlx5_core_dev *dev)
+{
+	struct mlx5_priv *priv = &dev->priv;
+	struct mlx5_interface *intf;
+
+	mutex_lock(&intf_mutex);
+	list_for_each_entry(intf, &intf_list, list)
+		mlx5_remove_device(intf, priv);
+	list_del(&priv->dev_list);
+	mutex_unlock(&intf_mutex);
+}
+
+int mlx5_register_interface(struct mlx5_interface *intf)
+{
+	struct mlx5_priv *priv;
+
+	if (!intf->add || !intf->remove)
+		return -EINVAL;
+
+	mutex_lock(&intf_mutex);
+	list_add_tail(&intf->list, &intf_list);
+	list_for_each_entry(priv, &dev_list, dev_list)
+		mlx5_add_device(intf, priv);
+	mutex_unlock(&intf_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL(mlx5_register_interface);
+
+void mlx5_unregister_interface(struct mlx5_interface *intf)
+{
+	struct mlx5_priv *priv;
+
+	mutex_lock(&intf_mutex);
+	list_for_each_entry(priv, &dev_list, dev_list)
+		mlx5_remove_device(intf, priv);
+	list_del(&intf->list);
+	mutex_unlock(&intf_mutex);
+}
+EXPORT_SYMBOL(mlx5_unregister_interface);
+
+void *mlx5_get_protocol_dev(struct mlx5_core_dev *mdev, int protocol)
+{
+	struct mlx5_priv *priv = &mdev->priv;
+	struct mlx5_device_context *dev_ctx;
+	unsigned long flags;
+	void *result = NULL;
+
+	spin_lock_irqsave(&priv->ctx_lock, flags);
+
+	list_for_each_entry(dev_ctx, &mdev->priv.ctx_list, list)
+		if ((dev_ctx->intf->protocol == protocol) &&
+		    dev_ctx->intf->get_dev) {
+			result = dev_ctx->intf->get_dev(dev_ctx->context);
+			break;
+		}
+
+	spin_unlock_irqrestore(&priv->ctx_lock, flags);
+
+	return result;
+}
+EXPORT_SYMBOL(mlx5_get_protocol_dev);
+
+static int mlx5_pci_init(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err = 0;
 
-	dev->pdev = pdev;
 	pci_set_drvdata(dev->pdev, dev);
 	strncpy(priv->name, dev_name(&pdev->dev), MLX5_MAX_NAME_LEN);
 	priv->name[MLX5_MAX_NAME_LEN - 1] = 0;
@@ -602,11 +849,15 @@ static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 	INIT_LIST_HEAD(&priv->pgdir_list);
 	spin_lock_init(&priv->mkey_lock);
 
+	mutex_init(&priv->alloc_mutex);
+
+	priv->numa_node = dev_to_node(&dev->pdev->dev);
+
 	priv->dbg_root = debugfs_create_dir(dev_name(&pdev->dev), mlx5_debugfs_root);
 	if (!priv->dbg_root)
 		return -ENOMEM;
 
-	err = pci_enable_device(pdev);
+	err = mlx5_pci_enable_device(dev);
 	if (err) {
 		dev_err(&pdev->dev, "Cannot enable PCI device, aborting\n");
 		goto err_dbg;
@@ -633,13 +884,61 @@ static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 		dev_err(&pdev->dev, "Failed mapping initialization segment, aborting\n");
 		goto err_clr_master;
 	}
+
+	return 0;
+
+err_clr_master:
+	pci_clear_master(dev->pdev);
+	release_bar(dev->pdev);
+err_disable:
+	mlx5_pci_disable_device(dev);
+
+err_dbg:
+	debugfs_remove(priv->dbg_root);
+	return err;
+}
+
+static void mlx5_pci_close(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+{
+	iounmap(dev->iseg);
+	pci_clear_master(dev->pdev);
+	release_bar(dev->pdev);
+	mlx5_pci_disable_device(dev);
+	debugfs_remove(priv->dbg_root);
+}
+
+#define MLX5_IB_MOD "mlx5_ib"
+static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
+{
+	struct pci_dev *pdev = dev->pdev;
+	int err;
+
+	mutex_lock(&dev->intf_state_mutex);
+	if (dev->interface_state == MLX5_INTERFACE_STATE_UP) {
+		dev_warn(&dev->pdev->dev, "%s: interface is up, NOP\n",
+			 __func__);
+		goto out;
+	}
+
 	dev_info(&pdev->dev, "firmware version: %d.%d.%d\n", fw_rev_maj(dev),
 		 fw_rev_min(dev), fw_rev_sub(dev));
 
+	/* on load removing any previous indication of internal error, device is
+	 * up
+	 */
+	dev->state = MLX5_DEVICE_STATE_UP;
+
 	err = mlx5_cmd_init(dev);
 	if (err) {
 		dev_err(&pdev->dev, "Failed initializing command interface, aborting\n");
-		goto err_unmap;
+		goto out_err;
+	}
+
+	err = wait_fw_init(dev, FW_INIT_TIMEOUT_MILI);
+	if (err) {
+		dev_err(&dev->pdev->dev, "Firmware over %d MS in initializing state, aborting\n",
+			FW_INIT_TIMEOUT_MILI);
+		goto out_err;
 	}
 
 	mlx5_pagealloc_init(dev);
@@ -650,6 +949,14 @@ static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 		goto err_pagealloc_cleanup;
 	}
 
+#ifdef CONFIG_MLX5_CORE_EN
+	err = mlx5_core_set_issi(dev);
+	if (err) {
+		dev_err(&pdev->dev, "failed to set issi\n");
+		goto err_disable_hca;
+	}
+#endif
+
 	err = mlx5_satisfy_startup_pages(dev, 1);
 	if (err) {
 		dev_err(&pdev->dev, "failed to allocate boot pages\n");
@@ -688,15 +995,15 @@ static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 
 	mlx5_start_health_poll(dev);
 
-	err = mlx5_cmd_query_hca_cap(dev, &dev->caps);
+	err = mlx5_query_hca_caps(dev);
 	if (err) {
 		dev_err(&pdev->dev, "query hca failed\n");
 		goto err_stop_poll;
 	}
 
-	err = mlx5_cmd_query_adapter(dev);
+	err = mlx5_query_board_id(dev);
 	if (err) {
-		dev_err(&pdev->dev, "query adapter failed\n");
+		dev_err(&pdev->dev, "query board id failed\n");
 		goto err_stop_poll;
 	}
 
@@ -730,6 +1037,15 @@ static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 		goto err_stop_eqs;
 	}
 
+	if (map_bf_area(dev))
+		dev_err(&pdev->dev, "Failed to map blue flame area\n");
+
+	err = mlx5_irq_set_affinity_hints(dev);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to alloc affinity hint cpumask\n");
+		goto err_unmap_bf_area;
+	}
+
 	MLX5_INIT_DOORBELL_LOCK(&priv->cq_uar_lock);
 
 	mlx5_init_cq_table(dev);
@@ -737,8 +1053,34 @@ static int mlx5_dev_init(struct mlx5_core_dev *dev, struct pci_dev *pdev)
 	mlx5_init_srq_table(dev);
 	mlx5_init_mr_table(dev);
 
+	err = mlx5_register_device(dev);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_register_device failed %d\n", err);
+		goto err_reg_dev;
+	}
+
+	err = request_module_nowait(MLX5_IB_MOD);
+	if (err)
+		pr_info("failed request module on %s\n", MLX5_IB_MOD);
+
+	dev->interface_state = MLX5_INTERFACE_STATE_UP;
+out:
+	mutex_unlock(&dev->intf_state_mutex);
+
 	return 0;
 
+err_reg_dev:
+	mlx5_cleanup_mr_table(dev);
+	mlx5_cleanup_srq_table(dev);
+	mlx5_cleanup_qp_table(dev);
+	mlx5_cleanup_cq_table(dev);
+	mlx5_irq_clear_affinity_hints(dev);
+
+err_unmap_bf_area:
+	unmap_bf_area(dev);
+
+	free_comp_eqs(dev);
+
 err_stop_eqs:
 	mlx5_stop_eqs(dev);
 
@@ -755,7 +1097,7 @@ err_stop_poll:
 	mlx5_stop_health_poll(dev);
 	if (mlx5_cmd_teardown_hca(dev)) {
 		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
-		return err;
+		goto out_err;
 	}
 
 err_pagealloc_stop:
@@ -771,167 +1113,55 @@ err_pagealloc_cleanup:
 	mlx5_pagealloc_cleanup(dev);
 	mlx5_cmd_cleanup(dev);
 
-err_unmap:
-	iounmap(dev->iseg);
-
-err_clr_master:
-	pci_clear_master(dev->pdev);
-	release_bar(dev->pdev);
-
-err_disable:
-	pci_disable_device(dev->pdev);
+out_err:
+	dev->state = MLX5_DEVICE_STATE_INTERNAL_ERROR;
+	mutex_unlock(&dev->intf_state_mutex);
 
-err_dbg:
-	debugfs_remove(priv->dbg_root);
 	return err;
 }
 
-static void mlx5_dev_cleanup(struct mlx5_core_dev *dev)
+static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv)
 {
-	struct mlx5_priv *priv = &dev->priv;
+	int err = 0;
 
+	mutex_lock(&dev->intf_state_mutex);
+	if (dev->interface_state == MLX5_INTERFACE_STATE_DOWN) {
+		dev_warn(&dev->pdev->dev, "%s: interface is down, NOP\n",
+			 __func__);
+		goto out;
+	}
+	mlx5_unregister_device(dev);
+	mlx5_cleanup_mr_table(dev);
 	mlx5_cleanup_srq_table(dev);
 	mlx5_cleanup_qp_table(dev);
 	mlx5_cleanup_cq_table(dev);
+	mlx5_irq_clear_affinity_hints(dev);
+	unmap_bf_area(dev);
 	free_comp_eqs(dev);
 	mlx5_stop_eqs(dev);
 	mlx5_free_uuars(dev, &priv->uuari);
 	mlx5_eq_cleanup(dev);
 	mlx5_disable_msix(dev);
 	mlx5_stop_health_poll(dev);
-	if (mlx5_cmd_teardown_hca(dev)) {
+	err = mlx5_cmd_teardown_hca(dev);
+	if (err) {
 		dev_err(&dev->pdev->dev, "tear_down_hca failed, skip cleanup\n");
-		return;
+		goto out;
 	}
 	mlx5_pagealloc_stop(dev);
 	mlx5_reclaim_startup_pages(dev);
 	mlx5_core_disable_hca(dev);
 	mlx5_pagealloc_cleanup(dev);
 	mlx5_cmd_cleanup(dev);
-	iounmap(dev->iseg);
-	pci_clear_master(dev->pdev);
-	release_bar(dev->pdev);
-	pci_disable_device(dev->pdev);
-	debugfs_remove(priv->dbg_root);
-}
-
-static void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	dev_ctx = kmalloc(sizeof(*dev_ctx), GFP_KERNEL);
-	if (!dev_ctx) {
-		pr_warn("mlx5_add_device: alloc context failed\n");
-		return;
-	}
-
-	dev_ctx->intf    = intf;
-	dev_ctx->context = intf->add(dev);
-
-	if (dev_ctx->context) {
-		spin_lock_irq(&priv->ctx_lock);
-		list_add_tail(&dev_ctx->list, &priv->ctx_list);
-		spin_unlock_irq(&priv->ctx_lock);
-	} else {
-		kfree(dev_ctx);
-	}
-}
-
-static void mlx5_remove_device(struct mlx5_interface *intf, struct mlx5_priv *priv)
-{
-	struct mlx5_device_context *dev_ctx;
-	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev, priv);
-
-	list_for_each_entry(dev_ctx, &priv->ctx_list, list)
-		if (dev_ctx->intf == intf) {
-			spin_lock_irq(&priv->ctx_lock);
-			list_del(&dev_ctx->list);
-			spin_unlock_irq(&priv->ctx_lock);
-
-			intf->remove(dev, dev_ctx->context);
-			kfree(dev_ctx);
-			return;
-		}
-}
-static int mlx5_register_device(struct mlx5_core_dev *dev)
-{
-	struct mlx5_priv *priv = &dev->priv;
-	struct mlx5_interface *intf;
-
-	mutex_lock(&intf_mutex);
-	list_add_tail(&priv->dev_list, &dev_list);
-	list_for_each_entry(intf, &intf_list, list)
-		mlx5_add_device(intf, priv);
-	mutex_unlock(&intf_mutex);
 
-	return 0;
-}
-static void mlx5_unregister_device(struct mlx5_core_dev *dev)
-{
-	struct mlx5_priv *priv = &dev->priv;
-	struct mlx5_interface *intf;
-
-	mutex_lock(&intf_mutex);
-	list_for_each_entry(intf, &intf_list, list)
-		mlx5_remove_device(intf, priv);
-	list_del(&priv->dev_list);
-	mutex_unlock(&intf_mutex);
-}
-
-int mlx5_register_interface(struct mlx5_interface *intf)
-{
-	struct mlx5_priv *priv;
-
-	if (!intf->add || !intf->remove)
-		return -EINVAL;
-
-	mutex_lock(&intf_mutex);
-	list_add_tail(&intf->list, &intf_list);
-	list_for_each_entry(priv, &dev_list, dev_list)
-		mlx5_add_device(intf, priv);
-	mutex_unlock(&intf_mutex);
-
-	return 0;
-}
-EXPORT_SYMBOL(mlx5_register_interface);
-
-void mlx5_unregister_interface(struct mlx5_interface *intf)
-{
-	struct mlx5_priv *priv;
-
-	mutex_lock(&intf_mutex);
-	list_for_each_entry(priv, &dev_list, dev_list)
-	       mlx5_remove_device(intf, priv);
-	list_del(&intf->list);
-	mutex_unlock(&intf_mutex);
-}
-EXPORT_SYMBOL(mlx5_unregister_interface);
-
-void *mlx5_get_protocol_dev(struct mlx5_core_dev *mdev, int protocol)
-{
-	struct mlx5_priv *priv = &mdev->priv;
-	struct mlx5_device_context *dev_ctx;
-	unsigned long flags;
-	void *result = NULL;
-
-	spin_lock_irqsave(&priv->ctx_lock, flags);
-
-	list_for_each_entry(dev_ctx, &mdev->priv.ctx_list, list)
-		if ((dev_ctx->intf->protocol == protocol) &&
-		    dev_ctx->intf->get_dev) {
-			result = dev_ctx->intf->get_dev(dev_ctx->context);
-			break;
-		}
-
-	spin_unlock_irqrestore(&priv->ctx_lock, flags);
-
-	return result;
+out:
+	dev->interface_state = MLX5_INTERFACE_STATE_DOWN;
+	mutex_unlock(&dev->intf_state_mutex);
+	return err;
 }
-EXPORT_SYMBOL(mlx5_get_protocol_dev);
 
-static void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
-			    unsigned long param)
+void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
+		     unsigned long param)
 {
 	struct mlx5_priv *priv = &dev->priv;
 	struct mlx5_device_context *dev_ctx;
@@ -952,7 +1182,6 @@ struct mlx5_core_event_handler {
 		      void *data);
 };
 
-#define MLX5_IB_MOD "mlx5_ib"
 
 static int init_one(struct pci_dev *pdev,
 		    const struct pci_device_id *id)
@@ -976,43 +1205,166 @@ static int init_one(struct pci_dev *pdev,
 		prof_sel = MLX5_DEFAULT_PROF;
 	}
 	dev->profile = &profile[prof_sel];
+	dev->pdev = pdev;
 	dev->event = mlx5_core_event;
 
 	INIT_LIST_HEAD(&priv->ctx_list);
 	spin_lock_init(&priv->ctx_lock);
-	err = mlx5_dev_init(dev, pdev);
+	mutex_init(&dev->pci_status_mutex);
+	mutex_init(&dev->intf_state_mutex);
+	err = mlx5_pci_init(dev, priv);
 	if (err) {
-		dev_err(&pdev->dev, "mlx5_dev_init failed %d\n", err);
-		goto out;
+		dev_err(&pdev->dev, "mlx5_pci_init failed with error code %d\n", err);
+		goto clean_dev;
 	}
 
-	err = mlx5_register_device(dev);
+	err = mlx5_health_init(dev);
 	if (err) {
-		dev_err(&pdev->dev, "mlx5_register_device failed %d\n", err);
-		goto out_init;
+		dev_err(&pdev->dev, "mlx5_health_init failed with error code %d\n", err);
+		goto close_pci;
 	}
 
-	err = request_module_nowait(MLX5_IB_MOD);
-	if (err)
-		pr_info("failed request module on %s\n", MLX5_IB_MOD);
+	err = mlx5_load_one(dev, priv);
+	if (err) {
+		dev_err(&pdev->dev, "mlx5_load_one failed with error code %d\n", err);
+		goto clean_health;
+	}
 
 	return 0;
 
-out_init:
-	mlx5_dev_cleanup(dev);
-out:
+clean_health:
+	mlx5_health_cleanup(dev);
+close_pci:
+	mlx5_pci_close(dev, priv);
+clean_dev:
+	pci_set_drvdata(pdev, NULL);
 	kfree(dev);
+
 	return err;
 }
+
 static void remove_one(struct pci_dev *pdev)
 {
 	struct mlx5_core_dev *dev  = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
 
-	mlx5_unregister_device(dev);
-	mlx5_dev_cleanup(dev);
+	if (mlx5_unload_one(dev, priv)) {
+		dev_err(&dev->pdev->dev, "mlx5_unload_one failed\n");
+		mlx5_health_cleanup(dev);
+		return;
+	}
+	mlx5_health_cleanup(dev);
+	mlx5_pci_close(dev, priv);
+	pci_set_drvdata(pdev, NULL);
 	kfree(dev);
 }
 
+static pci_ers_result_t mlx5_pci_err_detected(struct pci_dev *pdev,
+					      pci_channel_state_t state)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+	mlx5_enter_error_state(dev);
+	mlx5_unload_one(dev, priv);
+	mlx5_pci_disable_device(dev);
+	return state == pci_channel_io_perm_failure ?
+		PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_NEED_RESET;
+}
+
+static pci_ers_result_t mlx5_pci_slot_reset(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	int err = 0;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+
+	err = mlx5_pci_enable_device(dev);
+	if (err) {
+		dev_err(&pdev->dev, "%s: mlx5_pci_enable_device failed with error code: %d\n"
+			, __func__, err);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+	pci_set_master(pdev);
+	pci_set_power_state(pdev, PCI_D0);
+	pci_restore_state(pdev);
+
+	return err ? PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_RECOVERED;
+}
+
+void mlx5_disable_device(struct mlx5_core_dev *dev)
+{
+	mlx5_pci_err_detected(dev->pdev, 0);
+}
+
+/* wait for the device to show vital signs. For now we check
+ * that we can read the device ID and that the health buffer
+ * shows a non zero value which is different than 0xffffffff
+ */
+static void wait_vital(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_core_health *health = &dev->priv.health;
+	const int niter = 100;
+	u32 count;
+	u16 did;
+	int i;
+
+	/* Wait for firmware to be ready after reset */
+	msleep(1000);
+	for (i = 0; i < niter; i++) {
+		if (pci_read_config_word(pdev, 2, &did)) {
+			dev_warn(&pdev->dev, "failed reading config word\n");
+			break;
+		}
+		if (did == pdev->device) {
+			dev_info(&pdev->dev, "device ID correctly read after %d iterations\n", i);
+			break;
+		}
+		msleep(50);
+	}
+	if (i == niter)
+		dev_warn(&pdev->dev, "%s-%d: could not read device ID\n", __func__, __LINE__);
+
+	for (i = 0; i < niter; i++) {
+		count = ioread32be(health->health_counter);
+		if (count && count != 0xffffffff) {
+			dev_info(&pdev->dev, "Counter value 0x%x after %d iterations\n", count, i);
+			break;
+		}
+		msleep(50);
+	}
+
+	if (i == niter)
+		dev_warn(&pdev->dev, "%s-%d: could not read device ID\n", __func__, __LINE__);
+}
+
+static void mlx5_pci_resume(struct pci_dev *pdev)
+{
+	struct mlx5_core_dev *dev = pci_get_drvdata(pdev);
+	struct mlx5_priv *priv = &dev->priv;
+	int err;
+
+	dev_info(&pdev->dev, "%s was called\n", __func__);
+
+	pci_save_state(pdev);
+	wait_vital(pdev);
+
+	err = mlx5_load_one(dev, priv);
+	if (err)
+		dev_err(&pdev->dev, "%s: mlx5_load_one failed with error code: %d\n"
+			, __func__, err);
+	else
+		dev_info(&pdev->dev, "%s: device recovered\n", __func__);
+}
+
+static const struct pci_error_handlers mlx5_err_handler = {
+	.error_detected = mlx5_pci_err_detected,
+	.slot_reset	= mlx5_pci_slot_reset,
+	.resume		= mlx5_pci_resume
+};
+
 static const struct pci_device_id mlx5_core_pci_table[] = {
 	{ PCI_VDEVICE(MELLANOX, 0x1011) }, /* Connect-IB */
 	{ PCI_VDEVICE(MELLANOX, 0x1012) }, /* Connect-IB VF */
@@ -1029,7 +1381,8 @@ static struct pci_driver mlx5_core_driver = {
 	.name           = DRIVER_NAME,
 	.id_table       = mlx5_core_pci_table,
 	.probe          = init_one,
-	.remove         = remove_one
+	.remove         = remove_one,
+	.err_handler	= &mlx5_err_handler
 };
 
 static int __init init(void)
@@ -1037,22 +1390,17 @@ static int __init init(void)
 	int err;
 
 	mlx5_register_debugfs();
-	mlx5_core_wq = create_singlethread_workqueue("mlx5_core_wq");
-	if (!mlx5_core_wq) {
-		err = -ENOMEM;
-		goto err_debug;
-	}
-	mlx5_health_init();
 
 	err = pci_register_driver(&mlx5_core_driver);
 	if (err)
-		goto err_health;
+		goto err_debug;
+
+#ifdef CONFIG_MLX5_CORE_EN
+	mlx5e_init();
+#endif
 
 	return 0;
 
-err_health:
-	mlx5_health_cleanup();
-	destroy_workqueue(mlx5_core_wq);
 err_debug:
 	mlx5_unregister_debugfs();
 	return err;
@@ -1060,9 +1408,10 @@ err_debug:
 
 static void __exit cleanup(void)
 {
+#ifdef CONFIG_MLX5_CORE_EN
+	mlx5e_cleanup();
+#endif
 	pci_unregister_driver(&mlx5_core_driver);
-	mlx5_health_cleanup();
-	destroy_workqueue(mlx5_core_wq);
 	mlx5_unregister_debugfs();
 }
 
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mcg.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mcg.c
index d79fd85d1..d5a0c2d61 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mcg.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mcg.c
@@ -91,7 +91,7 @@ int mlx5_core_detach_mcg(struct mlx5_core_dev *dev, union ib_gid *mgid, u32 qpn)
 
 	memset(&in, 0, sizeof(in));
 	memset(&out, 0, sizeof(out));
-	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DETACH_FROM_MCG);
+	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DETTACH_FROM_MCG);
 	memcpy(in.gid, mgid, sizeof(*mgid));
 	in.qpn = cpu_to_be32(qpn);
 	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index a051b906a..cee5b7a83 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013-2015, Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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
@@ -37,27 +37,31 @@
 #include <linux/kernel.h>
 #include <linux/sched.h>
 
+#define DRIVER_NAME "mlx5_core"
+#define DRIVER_VERSION "3.0-1"
+#define DRIVER_RELDATE  "January 2015"
+
 extern int mlx5_core_debug_mask;
 
-#define mlx5_core_dbg(dev, format, ...)					\
-	pr_debug("%s:%s:%d:(pid %d): " format,				\
-		 (dev)->priv.name, __func__, __LINE__, current->pid,	\
+#define mlx5_core_dbg(__dev, format, ...)				\
+	dev_dbg(&(__dev)->pdev->dev, "%s:%s:%d:(pid %d): " format,	\
+		 (__dev)->priv.name, __func__, __LINE__, current->pid,	\
 		 ##__VA_ARGS__)
 
-#define mlx5_core_dbg_mask(dev, mask, format, ...)			\
+#define mlx5_core_dbg_mask(__dev, mask, format, ...)			\
 do {									\
 	if ((mask) & mlx5_core_debug_mask)				\
-		mlx5_core_dbg(dev, format, ##__VA_ARGS__);		\
+		mlx5_core_dbg(__dev, format, ##__VA_ARGS__);		\
 } while (0)
 
-#define mlx5_core_err(dev, format, ...)					\
-	pr_err("%s:%s:%d:(pid %d): " format,				\
-	       (dev)->priv.name, __func__, __LINE__, current->pid,	\
+#define mlx5_core_err(__dev, format, ...)				\
+	dev_err(&(__dev)->pdev->dev, "%s:%s:%d:(pid %d): " format,	\
+	       (__dev)->priv.name, __func__, __LINE__, current->pid,	\
 	       ##__VA_ARGS__)
 
-#define mlx5_core_warn(dev, format, ...)				\
-	pr_warn("%s:%s:%d:(pid %d): " format,				\
-		(dev)->priv.name, __func__, __LINE__, current->pid,	\
+#define mlx5_core_warn(__dev, format, ...)				\
+	dev_warn(&(__dev)->pdev->dev, "%s:%s:%d:(pid %d): " format,	\
+		(__dev)->priv.name, __func__, __LINE__, current->pid,	\
 		##__VA_ARGS__)
 
 enum {
@@ -65,11 +69,29 @@ enum {
 	MLX5_CMD_TIME, /* print command execution time */
 };
 
+static inline int mlx5_cmd_exec_check_status(struct mlx5_core_dev *dev, u32 *in,
+					     int in_size, u32 *out,
+					     int out_size)
+{
+	int err;
+
+	err = mlx5_cmd_exec(dev, in, in_size, out, out_size);
+	if (err)
+		return err;
 
-int mlx5_cmd_query_hca_cap(struct mlx5_core_dev *dev,
-			   struct mlx5_caps *caps);
-int mlx5_cmd_query_adapter(struct mlx5_core_dev *dev);
+	return mlx5_cmd_status_to_err((struct mlx5_outbox_hdr *)out);
+}
+
+int mlx5_query_hca_caps(struct mlx5_core_dev *dev);
+int mlx5_query_board_id(struct mlx5_core_dev *dev);
 int mlx5_cmd_init_hca(struct mlx5_core_dev *dev);
 int mlx5_cmd_teardown_hca(struct mlx5_core_dev *dev);
+void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event,
+		     unsigned long param);
+void mlx5_enter_error_state(struct mlx5_core_dev *dev);
+void mlx5_disable_device(struct mlx5_core_dev *dev);
+
+void mlx5e_init(void);
+void mlx5e_cleanup(void);
 
 #endif /* __MLX5_CORE_H__ */
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mr.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mr.c
index 1adb300dd..6fa22b51e 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/mr.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/mr.c
@@ -40,6 +40,7 @@ void mlx5_init_mr_table(struct mlx5_core_dev *dev)
 {
 	struct mlx5_mr_table *table = &dev->priv.mr_table;
 
+	memset(table, 0, sizeof(*table));
 	rwlock_init(&table->lock);
 	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
 }
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
index 8a64542ab..4d3377b12 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
@@ -30,7 +30,7 @@
  * SOFTWARE.
  */
 
-#include <asm-generic/kmap_types.h>
+#include <linux/highmem.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/mlx5/driver.h>
@@ -275,12 +275,36 @@ out_alloc:
 
 	return err;
 }
+
+static void page_notify_fail(struct mlx5_core_dev *dev, u16 func_id)
+{
+	struct mlx5_manage_pages_inbox *in;
+	struct mlx5_manage_pages_outbox out;
+	int err;
+
+	in = kzalloc(sizeof(*in), GFP_KERNEL);
+	if (!in)
+		return;
+
+	memset(&out, 0, sizeof(out));
+	in->hdr.opcode = cpu_to_be16(MLX5_CMD_OP_MANAGE_PAGES);
+	in->hdr.opmod = cpu_to_be16(MLX5_PAGES_CANT_GIVE);
+	in->func_id = cpu_to_be16(func_id);
+	err = mlx5_cmd_exec(dev, in, sizeof(*in), &out, sizeof(out));
+	if (!err)
+		err = mlx5_cmd_status_to_err(&out.hdr);
+
+	if (err)
+		mlx5_core_warn(dev, "page notify failed\n");
+
+	kfree(in);
+}
+
 static int give_pages(struct mlx5_core_dev *dev, u16 func_id, int npages,
 		      int notify_fail)
 {
 	struct mlx5_manage_pages_inbox *in;
 	struct mlx5_manage_pages_outbox out;
-	struct mlx5_manage_pages_inbox *nin;
 	int inlen;
 	u64 addr;
 	int err;
@@ -289,8 +313,9 @@ static int give_pages(struct mlx5_core_dev *dev, u16 func_id, int npages,
 	inlen = sizeof(*in) + npages * sizeof(in->pas[0]);
 	in = mlx5_vzalloc(inlen);
 	if (!in) {
+		err = -ENOMEM;
 		mlx5_core_warn(dev, "vzalloc failed %d\n", inlen);
-		return -ENOMEM;
+		goto out_free;
 	}
 	memset(&out, 0, sizeof(out));
 
@@ -316,43 +341,29 @@ retry:
 	if (err) {
 		mlx5_core_warn(dev, "func_id 0x%x, npages %d, err %d\n",
 			       func_id, npages, err);
-		goto out_alloc;
+		goto out_4k;
 	}
 	dev->priv.fw_pages += npages;
 
-	if (out.hdr.status) {
-		err = mlx5_cmd_status_to_err(&out.hdr);
-		if (err) {
-			mlx5_core_warn(dev, "func_id 0x%x, npages %d, status %d\n",
-				       func_id, npages, out.hdr.status);
-			goto out_alloc;
-		}
+	err = mlx5_cmd_status_to_err(&out.hdr);
+	if (err) {
+		mlx5_core_warn(dev, "func_id 0x%x, npages %d, status %d\n",
+			       func_id, npages, out.hdr.status);
+		goto out_4k;
 	}
 
 	mlx5_core_dbg(dev, "err %d\n", err);
 
-	goto out_free;
-
-out_alloc:
-	if (notify_fail) {
-		nin = kzalloc(sizeof(*nin), GFP_KERNEL);
-		if (!nin) {
-			mlx5_core_warn(dev, "allocation failed\n");
-			goto out_4k;
-		}
-		memset(&out, 0, sizeof(out));
-		nin->hdr.opcode = cpu_to_be16(MLX5_CMD_OP_MANAGE_PAGES);
-		nin->hdr.opmod = cpu_to_be16(MLX5_PAGES_CANT_GIVE);
-		if (mlx5_cmd_exec(dev, nin, sizeof(*nin), &out, sizeof(out)))
-			mlx5_core_warn(dev, "page notify failed\n");
-		kfree(nin);
-	}
+	kvfree(in);
+	return 0;
 
 out_4k:
 	for (i--; i >= 0; i--)
 		free_4k(dev, be64_to_cpu(in->pas[i]));
 out_free:
 	kvfree(in);
+	if (notify_fail)
+		page_notify_fail(dev, func_id);
 	return err;
 }
 
@@ -482,15 +493,20 @@ int mlx5_reclaim_startup_pages(struct mlx5_core_dev *dev)
 	struct fw_page *fwp;
 	struct rb_node *p;
 	int nclaimed = 0;
-	int err;
+	int err = 0;
 
 	do {
 		p = rb_first(&dev->priv.page_root);
 		if (p) {
 			fwp = rb_entry(p, struct fw_page, rb_node);
-			err = reclaim_pages(dev, fwp->func_id,
-					    optimal_reclaimed_pages(),
-					    &nclaimed);
+			if (dev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) {
+				free_4k(dev, fwp->addr);
+				nclaimed = 1;
+			} else {
+				err = reclaim_pages(dev, fwp->func_id,
+						    optimal_reclaimed_pages(),
+						    &nclaimed);
+			}
 			if (err) {
 				mlx5_core_warn(dev, "failed reclaiming pages (%d)\n",
 					       err);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/port.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/port.c
index 49e90f261..a87e773e9 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/port.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/port.c
@@ -90,15 +90,276 @@ int mlx5_set_port_caps(struct mlx5_core_dev *dev, u8 port_num, u32 caps)
 {
 	struct mlx5_reg_pcap in;
 	struct mlx5_reg_pcap out;
-	int err;
 
 	memset(&in, 0, sizeof(in));
 	in.caps_127_96 = cpu_to_be32(caps);
 	in.port_num = port_num;
 
-	err = mlx5_core_access_reg(dev, &in, sizeof(in), &out,
-				   sizeof(out), MLX5_REG_PCAP, 0, 1);
-
-	return err;
+	return mlx5_core_access_reg(dev, &in, sizeof(in), &out,
+				    sizeof(out), MLX5_REG_PCAP, 0, 1);
 }
 EXPORT_SYMBOL_GPL(mlx5_set_port_caps);
+
+int mlx5_query_port_ptys(struct mlx5_core_dev *dev, u32 *ptys,
+			 int ptys_size, int proto_mask, u8 local_port)
+{
+	u32 in[MLX5_ST_SZ_DW(ptys_reg)];
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(ptys_reg, in, local_port, local_port);
+	MLX5_SET(ptys_reg, in, proto_mask, proto_mask);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), ptys,
+				    ptys_size, MLX5_REG_PTYS, 0, 0);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_ptys);
+
+int mlx5_query_port_proto_cap(struct mlx5_core_dev *dev,
+			      u32 *proto_cap, int proto_mask)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), proto_mask, 1);
+	if (err)
+		return err;
+
+	if (proto_mask == MLX5_PTYS_EN)
+		*proto_cap = MLX5_GET(ptys_reg, out, eth_proto_capability);
+	else
+		*proto_cap = MLX5_GET(ptys_reg, out, ib_proto_capability);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_proto_cap);
+
+int mlx5_query_port_proto_admin(struct mlx5_core_dev *dev,
+				u32 *proto_admin, int proto_mask)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), proto_mask, 1);
+	if (err)
+		return err;
+
+	if (proto_mask == MLX5_PTYS_EN)
+		*proto_admin = MLX5_GET(ptys_reg, out, eth_proto_admin);
+	else
+		*proto_admin = MLX5_GET(ptys_reg, out, ib_proto_admin);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_proto_admin);
+
+int mlx5_query_port_link_width_oper(struct mlx5_core_dev *dev,
+				    u8 *link_width_oper, u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), MLX5_PTYS_IB, local_port);
+	if (err)
+		return err;
+
+	*link_width_oper = MLX5_GET(ptys_reg, out, ib_link_width_oper);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_link_width_oper);
+
+int mlx5_query_port_proto_oper(struct mlx5_core_dev *dev,
+			       u8 *proto_oper, int proto_mask,
+			       u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+	int err;
+
+	err = mlx5_query_port_ptys(dev, out, sizeof(out), proto_mask, local_port);
+	if (err)
+		return err;
+
+	if (proto_mask == MLX5_PTYS_EN)
+		*proto_oper = MLX5_GET(ptys_reg, out, eth_proto_oper);
+	else
+		*proto_oper = MLX5_GET(ptys_reg, out, ib_proto_oper);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_proto_oper);
+
+int mlx5_set_port_proto(struct mlx5_core_dev *dev, u32 proto_admin,
+			int proto_mask)
+{
+	u32 in[MLX5_ST_SZ_DW(ptys_reg)];
+	u32 out[MLX5_ST_SZ_DW(ptys_reg)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(ptys_reg, in, local_port, 1);
+	MLX5_SET(ptys_reg, in, proto_mask, proto_mask);
+	if (proto_mask == MLX5_PTYS_EN)
+		MLX5_SET(ptys_reg, in, eth_proto_admin, proto_admin);
+	else
+		MLX5_SET(ptys_reg, in, ib_proto_admin, proto_admin);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PTYS, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_proto);
+
+int mlx5_set_port_admin_status(struct mlx5_core_dev *dev,
+			       enum mlx5_port_status status)
+{
+	u32 in[MLX5_ST_SZ_DW(paos_reg)];
+	u32 out[MLX5_ST_SZ_DW(paos_reg)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(paos_reg, in, local_port, 1);
+	MLX5_SET(paos_reg, in, admin_status, status);
+	MLX5_SET(paos_reg, in, ase, 1);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PAOS, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_admin_status);
+
+int mlx5_query_port_admin_status(struct mlx5_core_dev *dev,
+				 enum mlx5_port_status *status)
+{
+	u32 in[MLX5_ST_SZ_DW(paos_reg)];
+	u32 out[MLX5_ST_SZ_DW(paos_reg)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(paos_reg, in, local_port, 1);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_PAOS, 0, 0);
+	if (err)
+		return err;
+
+	*status = MLX5_GET(paos_reg, out, admin_status);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_admin_status);
+
+static void mlx5_query_port_mtu(struct mlx5_core_dev *dev, int *admin_mtu,
+				int *max_mtu, int *oper_mtu, u8 port)
+{
+	u32 in[MLX5_ST_SZ_DW(pmtu_reg)];
+	u32 out[MLX5_ST_SZ_DW(pmtu_reg)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(pmtu_reg, in, local_port, port);
+
+	mlx5_core_access_reg(dev, in, sizeof(in), out,
+			     sizeof(out), MLX5_REG_PMTU, 0, 0);
+
+	if (max_mtu)
+		*max_mtu  = MLX5_GET(pmtu_reg, out, max_mtu);
+	if (oper_mtu)
+		*oper_mtu = MLX5_GET(pmtu_reg, out, oper_mtu);
+	if (admin_mtu)
+		*admin_mtu = MLX5_GET(pmtu_reg, out, admin_mtu);
+}
+
+int mlx5_set_port_mtu(struct mlx5_core_dev *dev, int mtu, u8 port)
+{
+	u32 in[MLX5_ST_SZ_DW(pmtu_reg)];
+	u32 out[MLX5_ST_SZ_DW(pmtu_reg)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(pmtu_reg, in, admin_mtu, mtu);
+	MLX5_SET(pmtu_reg, in, local_port, port);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_PMTU, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_mtu);
+
+void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, int *max_mtu,
+			     u8 port)
+{
+	mlx5_query_port_mtu(dev, NULL, max_mtu, NULL, port);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_max_mtu);
+
+void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, int *oper_mtu,
+			      u8 port)
+{
+	mlx5_query_port_mtu(dev, NULL, NULL, oper_mtu, port);
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_oper_mtu);
+
+static int mlx5_query_port_pvlc(struct mlx5_core_dev *dev, u32 *pvlc,
+				int pvlc_size,  u8 local_port)
+{
+	u32 in[MLX5_ST_SZ_DW(pvlc_reg)];
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(pvlc_reg, in, local_port, local_port);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), pvlc,
+				    pvlc_size, MLX5_REG_PVLC, 0, 0);
+}
+
+int mlx5_query_port_vl_hw_cap(struct mlx5_core_dev *dev,
+			      u8 *vl_hw_cap, u8 local_port)
+{
+	u32 out[MLX5_ST_SZ_DW(pvlc_reg)];
+	int err;
+
+	err = mlx5_query_port_pvlc(dev, out, sizeof(out), local_port);
+	if (err)
+		return err;
+
+	*vl_hw_cap = MLX5_GET(pvlc_reg, out, vl_hw_cap);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_vl_hw_cap);
+
+int mlx5_set_port_pause(struct mlx5_core_dev *dev, u32 rx_pause, u32 tx_pause)
+{
+	u32 in[MLX5_ST_SZ_DW(pfcc_reg)];
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(pfcc_reg, in, local_port, 1);
+	MLX5_SET(pfcc_reg, in, pptx, tx_pause);
+	MLX5_SET(pfcc_reg, in, pprx, rx_pause);
+
+	return mlx5_core_access_reg(dev, in, sizeof(in), out,
+				    sizeof(out), MLX5_REG_PFCC, 0, 1);
+}
+EXPORT_SYMBOL_GPL(mlx5_set_port_pause);
+
+int mlx5_query_port_pause(struct mlx5_core_dev *dev,
+			  u32 *rx_pause, u32 *tx_pause)
+{
+	u32 in[MLX5_ST_SZ_DW(pfcc_reg)];
+	u32 out[MLX5_ST_SZ_DW(pfcc_reg)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(pfcc_reg, in, local_port, 1);
+
+	err = mlx5_core_access_reg(dev, in, sizeof(in), out,
+				   sizeof(out), MLX5_REG_PFCC, 0, 0);
+	if (err)
+		return err;
+
+	if (rx_pause)
+		*rx_pause = MLX5_GET(pfcc_reg, out, pprx);
+
+	if (tx_pause)
+		*tx_pause = MLX5_GET(pfcc_reg, out, pptx);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_port_pause);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/qp.c
index dc7dbf7e9..30e2ba3f5 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/qp.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/qp.c
@@ -187,10 +187,17 @@ int mlx5_core_create_qp(struct mlx5_core_dev *dev,
 	struct mlx5_destroy_qp_mbox_in din;
 	struct mlx5_destroy_qp_mbox_out dout;
 	int err;
+	void *qpc;
 
 	memset(&out, 0, sizeof(out));
 	in->hdr.opcode = cpu_to_be16(MLX5_CMD_OP_CREATE_QP);
 
+	if (dev->issi) {
+		qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
+		/* 0xffffff means we ask to work with cqe version 0 */
+		MLX5_SET(qpc, qpc, user_index, 0xffffff);
+	}
+
 	err = mlx5_cmd_exec(dev, in, inlen, &out, sizeof(out));
 	if (err) {
 		mlx5_core_warn(dev, "ret %d\n", err);
@@ -338,6 +345,7 @@ void mlx5_init_qp_table(struct mlx5_core_dev *dev)
 {
 	struct mlx5_qp_table *table = &dev->priv.qp_table;
 
+	memset(table, 0, sizeof(*table));
 	spin_lock_init(&table->lock);
 	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
 	mlx5_qp_debugfs_init(dev);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/srq.c
index f9d25dcd0..ffada8019 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/srq.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/srq.c
@@ -37,6 +37,7 @@
 #include <linux/mlx5/srq.h>
 #include <rdma/ib_verbs.h>
 #include "mlx5_core.h"
+#include "transobj.h"
 
 void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type)
 {
@@ -62,6 +63,74 @@ void mlx5_srq_event(struct mlx5_core_dev *dev, u32 srqn, int event_type)
 		complete(&srq->free);
 }
 
+static int get_pas_size(void *srqc)
+{
+	u32 log_page_size = MLX5_GET(srqc, srqc, log_page_size) + 12;
+	u32 log_srq_size  = MLX5_GET(srqc, srqc, log_srq_size);
+	u32 log_rq_stride = MLX5_GET(srqc, srqc, log_rq_stride);
+	u32 page_offset   = MLX5_GET(srqc, srqc, page_offset);
+	u32 po_quanta	  = 1 << (log_page_size - 6);
+	u32 rq_sz	  = 1 << (log_srq_size + 4 + log_rq_stride);
+	u32 page_size	  = 1 << log_page_size;
+	u32 rq_sz_po      = rq_sz + (page_offset * po_quanta);
+	u32 rq_num_pas	  = (rq_sz_po + page_size - 1) / page_size;
+
+	return rq_num_pas * sizeof(u64);
+}
+
+static void rmpc_srqc_reformat(void *srqc, void *rmpc, bool srqc_to_rmpc)
+{
+	void *wq = MLX5_ADDR_OF(rmpc, rmpc, wq);
+
+	if (srqc_to_rmpc) {
+		switch (MLX5_GET(srqc, srqc, state)) {
+		case MLX5_SRQC_STATE_GOOD:
+			MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY);
+			break;
+		case MLX5_SRQC_STATE_ERROR:
+			MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_ERR);
+			break;
+		default:
+			pr_warn("%s: %d: Unknown srq state = 0x%x\n", __func__,
+				__LINE__, MLX5_GET(srqc, srqc, state));
+			MLX5_SET(rmpc, rmpc, state, MLX5_GET(srqc, srqc, state));
+		}
+
+		MLX5_SET(wq,   wq, wq_signature,  MLX5_GET(srqc,  srqc, wq_signature));
+		MLX5_SET(wq,   wq, log_wq_pg_sz,  MLX5_GET(srqc,  srqc, log_page_size));
+		MLX5_SET(wq,   wq, log_wq_stride, MLX5_GET(srqc,  srqc, log_rq_stride) + 4);
+		MLX5_SET(wq,   wq, log_wq_sz,     MLX5_GET(srqc,  srqc, log_srq_size));
+		MLX5_SET(wq,   wq, page_offset,   MLX5_GET(srqc,  srqc, page_offset));
+		MLX5_SET(wq,   wq, lwm,           MLX5_GET(srqc,  srqc, lwm));
+		MLX5_SET(wq,   wq, pd,            MLX5_GET(srqc,  srqc, pd));
+		MLX5_SET64(wq, wq, dbr_addr, MLX5_GET64(srqc,	  srqc, dbr_addr));
+	} else {
+		switch (MLX5_GET(rmpc, rmpc, state)) {
+		case MLX5_RMPC_STATE_RDY:
+			MLX5_SET(srqc, srqc, state, MLX5_SRQC_STATE_GOOD);
+			break;
+		case MLX5_RMPC_STATE_ERR:
+			MLX5_SET(srqc, srqc, state, MLX5_SRQC_STATE_ERROR);
+			break;
+		default:
+			pr_warn("%s: %d: Unknown rmp state = 0x%x\n",
+				__func__, __LINE__,
+				MLX5_GET(rmpc, rmpc, state));
+			MLX5_SET(srqc, srqc, state,
+				 MLX5_GET(rmpc, rmpc, state));
+		}
+
+		MLX5_SET(srqc,   srqc, wq_signature,   MLX5_GET(wq,   wq, wq_signature));
+		MLX5_SET(srqc,   srqc, log_page_size,  MLX5_GET(wq,   wq, log_wq_pg_sz));
+		MLX5_SET(srqc,   srqc, log_rq_stride,  MLX5_GET(wq,   wq, log_wq_stride) - 4);
+		MLX5_SET(srqc,   srqc, log_srq_size,   MLX5_GET(wq,   wq, log_wq_sz));
+		MLX5_SET(srqc,   srqc, page_offset,    MLX5_GET(wq,   wq, page_offset));
+		MLX5_SET(srqc,   srqc, lwm,	       MLX5_GET(wq,   wq, lwm));
+		MLX5_SET(srqc,   srqc, pd,	       MLX5_GET(wq,   wq, pd));
+		MLX5_SET64(srqc, srqc, dbr_addr,       MLX5_GET64(wq, wq, dbr_addr));
+	}
+}
+
 struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn)
 {
 	struct mlx5_srq_table *table = &dev->priv.srq_table;
@@ -79,26 +148,311 @@ struct mlx5_core_srq *mlx5_core_get_srq(struct mlx5_core_dev *dev, u32 srqn)
 }
 EXPORT_SYMBOL(mlx5_core_get_srq);
 
-int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
-			 struct mlx5_create_srq_mbox_in *in, int inlen)
+static int create_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			  struct mlx5_create_srq_mbox_in *in, int inlen)
 {
 	struct mlx5_create_srq_mbox_out out;
-	struct mlx5_srq_table *table = &dev->priv.srq_table;
-	struct mlx5_destroy_srq_mbox_in din;
-	struct mlx5_destroy_srq_mbox_out dout;
 	int err;
 
 	memset(&out, 0, sizeof(out));
+
 	in->hdr.opcode = cpu_to_be16(MLX5_CMD_OP_CREATE_SRQ);
-	err = mlx5_cmd_exec(dev, in, inlen, &out, sizeof(out));
-	if (err)
-		return err;
 
-	if (out.hdr.status)
-		return mlx5_cmd_status_to_err(&out.hdr);
+	err = mlx5_cmd_exec_check_status(dev, (u32 *)in, inlen, (u32 *)(&out),
+					 sizeof(out));
 
 	srq->srqn = be32_to_cpu(out.srqn) & 0xffffff;
 
+	return err;
+}
+
+static int destroy_srq_cmd(struct mlx5_core_dev *dev,
+			   struct mlx5_core_srq *srq)
+{
+	struct mlx5_destroy_srq_mbox_in in;
+	struct mlx5_destroy_srq_mbox_out out;
+
+	memset(&in, 0, sizeof(in));
+	memset(&out, 0, sizeof(out));
+	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DESTROY_SRQ);
+	in.srqn = cpu_to_be32(srq->srqn);
+
+	return mlx5_cmd_exec_check_status(dev, (u32 *)(&in), sizeof(in),
+					  (u32 *)(&out), sizeof(out));
+}
+
+static int arm_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+		       u16 lwm, int is_srq)
+{
+	struct mlx5_arm_srq_mbox_in	in;
+	struct mlx5_arm_srq_mbox_out	out;
+
+	memset(&in, 0, sizeof(in));
+	memset(&out, 0, sizeof(out));
+
+	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_ARM_RQ);
+	in.hdr.opmod = cpu_to_be16(!!is_srq);
+	in.srqn = cpu_to_be32(srq->srqn);
+	in.lwm = cpu_to_be16(lwm);
+
+	return mlx5_cmd_exec_check_status(dev, (u32 *)(&in),
+					  sizeof(in), (u32 *)(&out),
+					  sizeof(out));
+}
+
+static int query_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_query_srq_mbox_out *out)
+{
+	struct mlx5_query_srq_mbox_in in;
+
+	memset(&in, 0, sizeof(in));
+
+	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_QUERY_SRQ);
+	in.srqn = cpu_to_be32(srq->srqn);
+
+	return mlx5_cmd_exec_check_status(dev, (u32 *)(&in), sizeof(in),
+					  (u32 *)out, sizeof(*out));
+}
+
+static int create_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			      struct mlx5_core_srq *srq,
+			      struct mlx5_create_srq_mbox_in *in,
+			      int srq_inlen)
+{
+	u32 create_out[MLX5_ST_SZ_DW(create_xrc_srq_out)];
+	void *create_in;
+	void *srqc;
+	void *xrc_srqc;
+	void *pas;
+	int pas_size;
+	int inlen;
+	int err;
+
+	srqc	  = MLX5_ADDR_OF(create_srq_in, in, srq_context_entry);
+	pas_size  = get_pas_size(srqc);
+	inlen	  = MLX5_ST_SZ_BYTES(create_xrc_srq_in) + pas_size;
+	create_in = mlx5_vzalloc(inlen);
+	if (!create_in)
+		return -ENOMEM;
+
+	xrc_srqc = MLX5_ADDR_OF(create_xrc_srq_in, create_in,
+				xrc_srq_context_entry);
+	pas	 = MLX5_ADDR_OF(create_xrc_srq_in, create_in, pas);
+
+	memcpy(xrc_srqc, srqc, MLX5_ST_SZ_BYTES(srqc));
+	memcpy(pas, in->pas, pas_size);
+	/* 0xffffff means we ask to work with cqe version 0 */
+	MLX5_SET(xrc_srqc,	    xrc_srqc,  user_index, 0xffffff);
+	MLX5_SET(create_xrc_srq_in, create_in, opcode,
+		 MLX5_CMD_OP_CREATE_XRC_SRQ);
+
+	memset(create_out, 0, sizeof(create_out));
+	err = mlx5_cmd_exec_check_status(dev, create_in, inlen, create_out,
+					 sizeof(create_out));
+	if (err)
+		goto out;
+
+	srq->srqn = MLX5_GET(create_xrc_srq_out, create_out, xrc_srqn);
+out:
+	kvfree(create_in);
+	return err;
+}
+
+static int destroy_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			       struct mlx5_core_srq *srq)
+{
+	u32 xrcsrq_in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)];
+	u32 xrcsrq_out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)];
+
+	memset(xrcsrq_in, 0, sizeof(xrcsrq_in));
+	memset(xrcsrq_out, 0, sizeof(xrcsrq_out));
+
+	MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, opcode,
+		 MLX5_CMD_OP_DESTROY_XRC_SRQ);
+	MLX5_SET(destroy_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn);
+
+	return mlx5_cmd_exec_check_status(dev, xrcsrq_in, sizeof(xrcsrq_in),
+					  xrcsrq_out, sizeof(xrcsrq_out));
+}
+
+static int arm_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			   struct mlx5_core_srq *srq, u16 lwm)
+{
+	u32 xrcsrq_in[MLX5_ST_SZ_DW(arm_xrc_srq_in)];
+	u32 xrcsrq_out[MLX5_ST_SZ_DW(arm_xrc_srq_out)];
+
+	memset(xrcsrq_in, 0, sizeof(xrcsrq_in));
+	memset(xrcsrq_out, 0, sizeof(xrcsrq_out));
+
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, opcode,   MLX5_CMD_OP_ARM_XRC_SRQ);
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, op_mod,   MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ);
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn);
+	MLX5_SET(arm_xrc_srq_in, xrcsrq_in, lwm,      lwm);
+
+	return  mlx5_cmd_exec_check_status(dev, xrcsrq_in, sizeof(xrcsrq_in),
+					   xrcsrq_out, sizeof(xrcsrq_out));
+}
+
+static int query_xrc_srq_cmd(struct mlx5_core_dev *dev,
+			     struct mlx5_core_srq *srq,
+			     struct mlx5_query_srq_mbox_out *out)
+{
+	u32 xrcsrq_in[MLX5_ST_SZ_DW(query_xrc_srq_in)];
+	u32 *xrcsrq_out;
+	void *srqc;
+	void *xrc_srqc;
+	int err;
+
+	xrcsrq_out = mlx5_vzalloc(MLX5_ST_SZ_BYTES(query_xrc_srq_out));
+	if (!xrcsrq_out)
+		return -ENOMEM;
+	memset(xrcsrq_in, 0, sizeof(xrcsrq_in));
+
+	MLX5_SET(query_xrc_srq_in, xrcsrq_in, opcode,
+		 MLX5_CMD_OP_QUERY_XRC_SRQ);
+	MLX5_SET(query_xrc_srq_in, xrcsrq_in, xrc_srqn, srq->srqn);
+	err =  mlx5_cmd_exec_check_status(dev, xrcsrq_in, sizeof(xrcsrq_in),
+					  xrcsrq_out,
+					  MLX5_ST_SZ_BYTES(query_xrc_srq_out));
+	if (err)
+		goto out;
+
+	xrc_srqc = MLX5_ADDR_OF(query_xrc_srq_out, xrcsrq_out,
+				xrc_srq_context_entry);
+	srqc = MLX5_ADDR_OF(query_srq_out, out, srq_context_entry);
+	memcpy(srqc, xrc_srqc, MLX5_ST_SZ_BYTES(srqc));
+
+out:
+	kvfree(xrcsrq_out);
+	return err;
+}
+
+static int create_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			  struct mlx5_create_srq_mbox_in *in, int srq_inlen)
+{
+	void *create_in;
+	void *rmpc;
+	void *srqc;
+	int pas_size;
+	int inlen;
+	int err;
+
+	srqc = MLX5_ADDR_OF(create_srq_in, in, srq_context_entry);
+	pas_size = get_pas_size(srqc);
+	inlen = MLX5_ST_SZ_BYTES(create_rmp_in) + pas_size;
+	create_in = mlx5_vzalloc(inlen);
+	if (!create_in)
+		return -ENOMEM;
+
+	rmpc = MLX5_ADDR_OF(create_rmp_in, create_in, ctx);
+
+	memcpy(MLX5_ADDR_OF(rmpc, rmpc, wq.pas), in->pas, pas_size);
+	rmpc_srqc_reformat(srqc, rmpc, true);
+
+	err = mlx5_core_create_rmp(dev, create_in, inlen, &srq->srqn);
+
+	kvfree(create_in);
+	return err;
+}
+
+static int destroy_rmp_cmd(struct mlx5_core_dev *dev,
+			   struct mlx5_core_srq *srq)
+{
+	return mlx5_core_destroy_rmp(dev, srq->srqn);
+}
+
+static int arm_rmp_cmd(struct mlx5_core_dev *dev,
+		       struct mlx5_core_srq *srq,
+		       u16 lwm)
+{
+	void *in;
+	void *rmpc;
+	void *wq;
+	void *bitmask;
+	int err;
+
+	in = mlx5_vzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in));
+	if (!in)
+		return -ENOMEM;
+
+	rmpc =	  MLX5_ADDR_OF(modify_rmp_in,   in,   ctx);
+	bitmask = MLX5_ADDR_OF(modify_rmp_in,   in,   bitmask);
+	wq   =	  MLX5_ADDR_OF(rmpc,	        rmpc, wq);
+
+	MLX5_SET(modify_rmp_in, in,	 rmp_state, MLX5_RMPC_STATE_RDY);
+	MLX5_SET(modify_rmp_in, in,	 rmpn,      srq->srqn);
+	MLX5_SET(wq,		wq,	 lwm,	    lwm);
+	MLX5_SET(rmp_bitmask,	bitmask, lwm,	    1);
+	MLX5_SET(rmpc, rmpc, state, MLX5_RMPC_STATE_RDY);
+
+	err = mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in));
+
+	kvfree(in);
+	return err;
+}
+
+static int query_rmp_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_query_srq_mbox_out *out)
+{
+	u32 *rmp_out;
+	void *rmpc;
+	void *srqc;
+	int err;
+
+	rmp_out =  mlx5_vzalloc(MLX5_ST_SZ_BYTES(query_rmp_out));
+	if (!rmp_out)
+		return -ENOMEM;
+
+	err = mlx5_core_query_rmp(dev, srq->srqn, rmp_out);
+	if (err)
+		goto out;
+
+	srqc = MLX5_ADDR_OF(query_srq_out, out,	    srq_context_entry);
+	rmpc = MLX5_ADDR_OF(query_rmp_out, rmp_out, rmp_context);
+	rmpc_srqc_reformat(srqc, rmpc, false);
+
+out:
+	kvfree(rmp_out);
+	return err;
+}
+
+static int create_srq_split(struct mlx5_core_dev *dev,
+			    struct mlx5_core_srq *srq,
+			    struct mlx5_create_srq_mbox_in *in,
+			    int inlen, int is_xrc)
+{
+	if (!dev->issi)
+		return create_srq_cmd(dev, srq, in, inlen);
+	else if (srq->common.res == MLX5_RES_XSRQ)
+		return create_xrc_srq_cmd(dev, srq, in, inlen);
+	else
+		return create_rmp_cmd(dev, srq, in, inlen);
+}
+
+static int destroy_srq_split(struct mlx5_core_dev *dev,
+			     struct mlx5_core_srq *srq)
+{
+	if (!dev->issi)
+		return destroy_srq_cmd(dev, srq);
+	else if (srq->common.res == MLX5_RES_XSRQ)
+		return destroy_xrc_srq_cmd(dev, srq);
+	else
+		return destroy_rmp_cmd(dev, srq);
+}
+
+int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
+			 struct mlx5_create_srq_mbox_in *in, int inlen,
+			 int is_xrc)
+{
+	int err;
+	struct mlx5_srq_table *table = &dev->priv.srq_table;
+
+	srq->common.res = is_xrc ? MLX5_RES_XSRQ : MLX5_RES_SRQ;
+
+	err = create_srq_split(dev, srq, in, inlen, is_xrc);
+	if (err)
+		return err;
+
 	atomic_set(&srq->refcount, 1);
 	init_completion(&srq->free);
 
@@ -107,25 +461,20 @@ int mlx5_core_create_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
 	spin_unlock_irq(&table->lock);
 	if (err) {
 		mlx5_core_warn(dev, "err %d, srqn 0x%x\n", err, srq->srqn);
-		goto err_cmd;
+		goto err_destroy_srq_split;
 	}
 
 	return 0;
 
-err_cmd:
-	memset(&din, 0, sizeof(din));
-	memset(&dout, 0, sizeof(dout));
-	din.srqn = cpu_to_be32(srq->srqn);
-	din.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DESTROY_SRQ);
-	mlx5_cmd_exec(dev, &din, sizeof(din), &dout, sizeof(dout));
+err_destroy_srq_split:
+	destroy_srq_split(dev, srq);
+
 	return err;
 }
 EXPORT_SYMBOL(mlx5_core_create_srq);
 
 int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq)
 {
-	struct mlx5_destroy_srq_mbox_in in;
-	struct mlx5_destroy_srq_mbox_out out;
 	struct mlx5_srq_table *table = &dev->priv.srq_table;
 	struct mlx5_core_srq *tmp;
 	int err;
@@ -142,17 +491,10 @@ int mlx5_core_destroy_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq)
 		return -EINVAL;
 	}
 
-	memset(&in, 0, sizeof(in));
-	memset(&out, 0, sizeof(out));
-	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_DESTROY_SRQ);
-	in.srqn = cpu_to_be32(srq->srqn);
-	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
+	err = destroy_srq_split(dev, srq);
 	if (err)
 		return err;
 
-	if (out.hdr.status)
-		return mlx5_cmd_status_to_err(&out.hdr);
-
 	if (atomic_dec_and_test(&srq->refcount))
 		complete(&srq->free);
 	wait_for_completion(&srq->free);
@@ -164,48 +506,24 @@ EXPORT_SYMBOL(mlx5_core_destroy_srq);
 int mlx5_core_query_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
 			struct mlx5_query_srq_mbox_out *out)
 {
-	struct mlx5_query_srq_mbox_in in;
-	int err;
-
-	memset(&in, 0, sizeof(in));
-	memset(out, 0, sizeof(*out));
-
-	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_QUERY_SRQ);
-	in.srqn = cpu_to_be32(srq->srqn);
-	err = mlx5_cmd_exec(dev, &in, sizeof(in), out, sizeof(*out));
-	if (err)
-		return err;
-
-	if (out->hdr.status)
-		return mlx5_cmd_status_to_err(&out->hdr);
-
-	return err;
+	if (!dev->issi)
+		return query_srq_cmd(dev, srq, out);
+	else if (srq->common.res == MLX5_RES_XSRQ)
+		return query_xrc_srq_cmd(dev, srq, out);
+	else
+		return query_rmp_cmd(dev, srq, out);
 }
 EXPORT_SYMBOL(mlx5_core_query_srq);
 
 int mlx5_core_arm_srq(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
 		      u16 lwm, int is_srq)
 {
-	struct mlx5_arm_srq_mbox_in	in;
-	struct mlx5_arm_srq_mbox_out	out;
-	int err;
-
-	memset(&in, 0, sizeof(in));
-	memset(&out, 0, sizeof(out));
-
-	in.hdr.opcode = cpu_to_be16(MLX5_CMD_OP_ARM_RQ);
-	in.hdr.opmod = cpu_to_be16(!!is_srq);
-	in.srqn = cpu_to_be32(srq->srqn);
-	in.lwm = cpu_to_be16(lwm);
-
-	err = mlx5_cmd_exec(dev, &in, sizeof(in), &out, sizeof(out));
-	if (err)
-		return err;
-
-	if (out.hdr.status)
-		return mlx5_cmd_status_to_err(&out.hdr);
-
-	return err;
+	if (!dev->issi)
+		return arm_srq_cmd(dev, srq, lwm, is_srq);
+	else if (srq->common.res == MLX5_RES_XSRQ)
+		return arm_xrc_srq_cmd(dev, srq, lwm);
+	else
+		return arm_rmp_cmd(dev, srq, lwm);
 }
 EXPORT_SYMBOL(mlx5_core_arm_srq);
 
@@ -213,6 +531,7 @@ void mlx5_init_srq_table(struct mlx5_core_dev *dev)
 {
 	struct mlx5_srq_table *table = &dev->priv.srq_table;
 
+	memset(table, 0, sizeof(*table));
 	spin_lock_init(&table->lock);
 	INIT_RADIX_TREE(&table->tree, GFP_ATOMIC);
 }
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/transobj.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/transobj.c
new file mode 100644
index 000000000..d7068f54e
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/transobj.c
@@ -0,0 +1,413 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 <linux/mlx5/driver.h>
+#include "mlx5_core.h"
+#include "transobj.h"
+
+int mlx5_alloc_transport_domain(struct mlx5_core_dev *dev, u32 *tdn)
+{
+	u32 in[MLX5_ST_SZ_DW(alloc_transport_domain_in)];
+	u32 out[MLX5_ST_SZ_DW(alloc_transport_domain_out)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+	MLX5_SET(alloc_transport_domain_in, in, opcode,
+		 MLX5_CMD_OP_ALLOC_TRANSPORT_DOMAIN);
+
+	err = mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+	if (!err)
+		*tdn = MLX5_GET(alloc_transport_domain_out, out,
+				transport_domain);
+
+	return err;
+}
+
+void mlx5_dealloc_transport_domain(struct mlx5_core_dev *dev, u32 tdn)
+{
+	u32 in[MLX5_ST_SZ_DW(dealloc_transport_domain_in)];
+	u32 out[MLX5_ST_SZ_DW(dealloc_transport_domain_out)];
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+	MLX5_SET(dealloc_transport_domain_in, in, opcode,
+		 MLX5_CMD_OP_DEALLOC_TRANSPORT_DOMAIN);
+	MLX5_SET(dealloc_transport_domain_in, in, transport_domain, tdn);
+
+	mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_create_rq(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *rqn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_rq_out)];
+	int err;
+
+	MLX5_SET(create_rq_in, in, opcode, MLX5_CMD_OP_CREATE_RQ);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*rqn = MLX5_GET(create_rq_out, out, rqn);
+
+	return err;
+}
+
+int mlx5_core_modify_rq(struct mlx5_core_dev *dev, u32 rqn, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_rq_out)];
+
+	MLX5_SET(modify_rq_in, in, rqn, rqn);
+	MLX5_SET(modify_rq_in, in, opcode, MLX5_CMD_OP_MODIFY_RQ);
+
+	memset(out, 0, sizeof(out));
+	return mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+}
+
+void mlx5_core_destroy_rq(struct mlx5_core_dev *dev, u32 rqn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_rq_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_rq_out)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(destroy_rq_in, in, opcode, MLX5_CMD_OP_DESTROY_RQ);
+	MLX5_SET(destroy_rq_in, in, rqn, rqn);
+
+	mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_create_sq(struct mlx5_core_dev *dev, u32 *in, int inlen, u32 *sqn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_sq_out)];
+	int err;
+
+	MLX5_SET(create_sq_in, in, opcode, MLX5_CMD_OP_CREATE_SQ);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*sqn = MLX5_GET(create_sq_out, out, sqn);
+
+	return err;
+}
+
+int mlx5_core_modify_sq(struct mlx5_core_dev *dev, u32 sqn, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_sq_out)];
+
+	MLX5_SET(modify_sq_in, in, sqn, sqn);
+	MLX5_SET(modify_sq_in, in, opcode, MLX5_CMD_OP_MODIFY_SQ);
+
+	memset(out, 0, sizeof(out));
+	return mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+}
+
+void mlx5_core_destroy_sq(struct mlx5_core_dev *dev, u32 sqn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_sq_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_sq_out)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(destroy_sq_in, in, opcode, MLX5_CMD_OP_DESTROY_SQ);
+	MLX5_SET(destroy_sq_in, in, sqn, sqn);
+
+	mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_create_tir(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *tirn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_tir_out)];
+	int err;
+
+	MLX5_SET(create_tir_in, in, opcode, MLX5_CMD_OP_CREATE_TIR);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*tirn = MLX5_GET(create_tir_out, out, tirn);
+
+	return err;
+}
+
+int mlx5_core_modify_tir(struct mlx5_core_dev *dev, u32 tirn, u32 *in,
+			 int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_tir_out)];
+
+	MLX5_SET(modify_tir_in, in, tirn, tirn);
+	MLX5_SET(modify_tir_in, in, opcode, MLX5_CMD_OP_MODIFY_TIR);
+
+	memset(out, 0, sizeof(out));
+	return mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+}
+
+void mlx5_core_destroy_tir(struct mlx5_core_dev *dev, u32 tirn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_tir_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_tir_out)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(destroy_tir_in, in, opcode, MLX5_CMD_OP_DESTROY_TIR);
+	MLX5_SET(destroy_tir_in, in, tirn, tirn);
+
+	mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_create_tis(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *tisn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_tis_out)];
+	int err;
+
+	MLX5_SET(create_tis_in, in, opcode, MLX5_CMD_OP_CREATE_TIS);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*tisn = MLX5_GET(create_tis_out, out, tisn);
+
+	return err;
+}
+
+void mlx5_core_destroy_tis(struct mlx5_core_dev *dev, u32 tisn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_tis_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_tis_out)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(destroy_tis_in, in, opcode, MLX5_CMD_OP_DESTROY_TIS);
+	MLX5_SET(destroy_tis_in, in, tisn, tisn);
+
+	mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *rmpn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_rmp_out)];
+	int err;
+
+	MLX5_SET(create_rmp_in, in, opcode, MLX5_CMD_OP_CREATE_RMP);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*rmpn = MLX5_GET(create_rmp_out, out, rmpn);
+
+	return err;
+}
+
+int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_rmp_out)];
+
+	MLX5_SET(modify_rmp_in, in, opcode, MLX5_CMD_OP_MODIFY_RMP);
+
+	memset(out, 0, sizeof(out));
+	return mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+}
+
+int mlx5_core_destroy_rmp(struct mlx5_core_dev *dev, u32 rmpn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_rmp_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_rmp_out)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(destroy_rmp_in, in, opcode, MLX5_CMD_OP_DESTROY_RMP);
+	MLX5_SET(destroy_rmp_in, in, rmpn, rmpn);
+
+	return mlx5_cmd_exec_check_status(dev, in, sizeof(in), out,
+					  sizeof(out));
+}
+
+int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out)
+{
+	u32 in[MLX5_ST_SZ_DW(query_rmp_in)];
+	int outlen = MLX5_ST_SZ_BYTES(query_rmp_out);
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(query_rmp_in, in, opcode, MLX5_CMD_OP_QUERY_RMP);
+	MLX5_SET(query_rmp_in, in, rmpn,   rmpn);
+
+	return mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, outlen);
+}
+
+int mlx5_core_arm_rmp(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm)
+{
+	void *in;
+	void *rmpc;
+	void *wq;
+	void *bitmask;
+	int  err;
+
+	in = mlx5_vzalloc(MLX5_ST_SZ_BYTES(modify_rmp_in));
+	if (!in)
+		return -ENOMEM;
+
+	rmpc    = MLX5_ADDR_OF(modify_rmp_in,   in,   ctx);
+	bitmask = MLX5_ADDR_OF(modify_rmp_in,   in,   bitmask);
+	wq      = MLX5_ADDR_OF(rmpc,	        rmpc, wq);
+
+	MLX5_SET(modify_rmp_in, in,	 rmp_state, MLX5_RMPC_STATE_RDY);
+	MLX5_SET(modify_rmp_in, in,	 rmpn,      rmpn);
+	MLX5_SET(wq,		wq,	 lwm,	    lwm);
+	MLX5_SET(rmp_bitmask,	bitmask, lwm,	    1);
+	MLX5_SET(rmpc,		rmpc,	 state,	    MLX5_RMPC_STATE_RDY);
+
+	err =  mlx5_core_modify_rmp(dev, in, MLX5_ST_SZ_BYTES(modify_rmp_in));
+
+	kvfree(in);
+
+	return err;
+}
+
+int mlx5_core_create_xsrq(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			  u32 *xsrqn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_xrc_srq_out)];
+	int err;
+
+	MLX5_SET(create_xrc_srq_in, in, opcode,     MLX5_CMD_OP_CREATE_XRC_SRQ);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*xsrqn = MLX5_GET(create_xrc_srq_out, out, xrc_srqn);
+
+	return err;
+}
+
+int mlx5_core_destroy_xsrq(struct mlx5_core_dev *dev, u32 xsrqn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_xrc_srq_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_xrc_srq_out)];
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+	MLX5_SET(destroy_xrc_srq_in, in, opcode,   MLX5_CMD_OP_DESTROY_XRC_SRQ);
+	MLX5_SET(destroy_xrc_srq_in, in, xrc_srqn, xsrqn);
+
+	return mlx5_cmd_exec_check_status(dev, in, sizeof(in), out,
+					  sizeof(out));
+}
+
+int mlx5_core_query_xsrq(struct mlx5_core_dev *dev, u32 xsrqn, u32 *out)
+{
+	u32 in[MLX5_ST_SZ_DW(query_xrc_srq_in)];
+	void *srqc;
+	void *xrc_srqc;
+	int err;
+
+	memset(in, 0, sizeof(in));
+	MLX5_SET(query_xrc_srq_in, in, opcode,   MLX5_CMD_OP_QUERY_XRC_SRQ);
+	MLX5_SET(query_xrc_srq_in, in, xrc_srqn, xsrqn);
+
+	err =  mlx5_cmd_exec_check_status(dev, in, sizeof(in),
+					  out,
+					  MLX5_ST_SZ_BYTES(query_xrc_srq_out));
+	if (!err) {
+		xrc_srqc = MLX5_ADDR_OF(query_xrc_srq_out, out,
+					xrc_srq_context_entry);
+		srqc = MLX5_ADDR_OF(query_srq_out, out, srq_context_entry);
+		memcpy(srqc, xrc_srqc, MLX5_ST_SZ_BYTES(srqc));
+	}
+
+	return err;
+}
+
+int mlx5_core_arm_xsrq(struct mlx5_core_dev *dev, u32 xsrqn, u16 lwm)
+{
+	u32 in[MLX5_ST_SZ_DW(arm_xrc_srq_in)];
+	u32 out[MLX5_ST_SZ_DW(arm_xrc_srq_out)];
+
+	memset(in, 0, sizeof(in));
+	memset(out, 0, sizeof(out));
+
+	MLX5_SET(arm_xrc_srq_in, in, opcode,   MLX5_CMD_OP_ARM_XRC_SRQ);
+	MLX5_SET(arm_xrc_srq_in, in, xrc_srqn, xsrqn);
+	MLX5_SET(arm_xrc_srq_in, in, lwm,      lwm);
+	MLX5_SET(arm_xrc_srq_in, in, op_mod,
+		 MLX5_ARM_XRC_SRQ_IN_OP_MOD_XRC_SRQ);
+
+	return  mlx5_cmd_exec_check_status(dev, in, sizeof(in), out,
+					   sizeof(out));
+}
+
+int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *rqtn)
+{
+	u32 out[MLX5_ST_SZ_DW(create_rqt_out)];
+	int err;
+
+	MLX5_SET(create_rqt_in, in, opcode, MLX5_CMD_OP_CREATE_RQT);
+
+	memset(out, 0, sizeof(out));
+	err = mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+	if (!err)
+		*rqtn = MLX5_GET(create_rqt_out, out, rqtn);
+
+	return err;
+}
+
+int mlx5_core_modify_rqt(struct mlx5_core_dev *dev, u32 rqtn, u32 *in,
+			 int inlen)
+{
+	u32 out[MLX5_ST_SZ_DW(modify_rqt_out)];
+
+	MLX5_SET(modify_rqt_in, in, rqtn, rqtn);
+	MLX5_SET(modify_rqt_in, in, opcode, MLX5_CMD_OP_MODIFY_RQT);
+
+	memset(out, 0, sizeof(out));
+	return mlx5_cmd_exec_check_status(dev, in, inlen, out, sizeof(out));
+}
+
+void mlx5_core_destroy_rqt(struct mlx5_core_dev *dev, u32 rqtn)
+{
+	u32 in[MLX5_ST_SZ_DW(destroy_rqt_in)];
+	u32 out[MLX5_ST_SZ_DW(destroy_rqt_out)];
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(destroy_rqt_in, in, opcode, MLX5_CMD_OP_DESTROY_RQT);
+	MLX5_SET(destroy_rqt_in, in, rqtn, rqtn);
+
+	mlx5_cmd_exec_check_status(dev, in, sizeof(in), out, sizeof(out));
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/transobj.h b/kernel/drivers/net/ethernet/mellanox/mlx5/core/transobj.h
new file mode 100644
index 000000000..74cae5143
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/transobj.h
@@ -0,0 +1,72 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 __TRANSOBJ_H__
+#define __TRANSOBJ_H__
+
+int mlx5_alloc_transport_domain(struct mlx5_core_dev *dev, u32 *tdn);
+void mlx5_dealloc_transport_domain(struct mlx5_core_dev *dev, u32 tdn);
+int mlx5_core_create_rq(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			u32 *rqn);
+int mlx5_core_modify_rq(struct mlx5_core_dev *dev, u32 rqn, u32 *in, int inlen);
+void mlx5_core_destroy_rq(struct mlx5_core_dev *dev, u32 rqn);
+int mlx5_core_create_sq(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			u32 *sqn);
+int mlx5_core_modify_sq(struct mlx5_core_dev *dev, u32 sqn, u32 *in, int inlen);
+void mlx5_core_destroy_sq(struct mlx5_core_dev *dev, u32 sqn);
+int mlx5_core_create_tir(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *tirn);
+int mlx5_core_modify_tir(struct mlx5_core_dev *dev, u32 tirn, u32 *in,
+			 int inlen);
+void mlx5_core_destroy_tir(struct mlx5_core_dev *dev, u32 tirn);
+int mlx5_core_create_tis(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *tisn);
+void mlx5_core_destroy_tis(struct mlx5_core_dev *dev, u32 tisn);
+int mlx5_core_create_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *rmpn);
+int mlx5_core_modify_rmp(struct mlx5_core_dev *dev, u32 *in, int inlen);
+int mlx5_core_destroy_rmp(struct mlx5_core_dev *dev, u32 rmpn);
+int mlx5_core_query_rmp(struct mlx5_core_dev *dev, u32 rmpn, u32 *out);
+int mlx5_core_arm_rmp(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm);
+int mlx5_core_create_xsrq(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			  u32 *rmpn);
+int mlx5_core_destroy_xsrq(struct mlx5_core_dev *dev, u32 rmpn);
+int mlx5_core_query_xsrq(struct mlx5_core_dev *dev, u32 rmpn, u32 *out);
+int mlx5_core_arm_xsrq(struct mlx5_core_dev *dev, u32 rmpn, u16 lwm);
+
+int mlx5_core_create_rqt(struct mlx5_core_dev *dev, u32 *in, int inlen,
+			 u32 *rqtn);
+int mlx5_core_modify_rqt(struct mlx5_core_dev *dev, u32 rqtn, u32 *in,
+			 int inlen);
+void mlx5_core_destroy_rqt(struct mlx5_core_dev *dev, u32 rqtn);
+
+#endif /* __TRANSOBJ_H__ */
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/uar.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/uar.c
index 5a89bb1d6..eb05c845e 100644
--- a/kernel/drivers/net/ethernet/mellanox/mlx5/core/uar.c
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/uar.c
@@ -32,6 +32,7 @@
 
 #include <linux/kernel.h>
 #include <linux/module.h>
+#include <linux/io-mapping.h>
 #include <linux/mlx5/driver.h>
 #include <linux/mlx5/cmd.h>
 #include "mlx5_core.h"
@@ -175,12 +176,13 @@ int mlx5_alloc_uuars(struct mlx5_core_dev *dev, struct mlx5_uuar_info *uuari)
 	for (i = 0; i < tot_uuars; i++) {
 		bf = &uuari->bfs[i];
 
-		bf->buf_size = dev->caps.gen.bf_reg_size / 2;
+		bf->buf_size = (1 << MLX5_CAP_GEN(dev, log_bf_reg_size)) / 2;
 		bf->uar = &uuari->uars[i / MLX5_BF_REGS_PER_PAGE];
 		bf->regreg = uuari->uars[i / MLX5_BF_REGS_PER_PAGE].map;
 		bf->reg = NULL; /* Add WC support */
-		bf->offset = (i % MLX5_BF_REGS_PER_PAGE) * dev->caps.gen.bf_reg_size +
-			MLX5_BF_OFFSET;
+		bf->offset = (i % MLX5_BF_REGS_PER_PAGE) *
+			     (1 << MLX5_CAP_GEN(dev, log_bf_reg_size)) +
+			     MLX5_BF_OFFSET;
 		bf->need_lock = need_uuar_lock(i);
 		spin_lock_init(&bf->lock);
 		spin_lock_init(&bf->lock32);
@@ -223,3 +225,45 @@ int mlx5_free_uuars(struct mlx5_core_dev *dev, struct mlx5_uuar_info *uuari)
 
 	return 0;
 }
+
+int mlx5_alloc_map_uar(struct mlx5_core_dev *mdev, struct mlx5_uar *uar)
+{
+	phys_addr_t pfn;
+	phys_addr_t uar_bar_start;
+	int err;
+
+	err = mlx5_cmd_alloc_uar(mdev, &uar->index);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_cmd_alloc_uar() failed, %d\n", err);
+		return err;
+	}
+
+	uar_bar_start = pci_resource_start(mdev->pdev, 0);
+	pfn           = (uar_bar_start >> PAGE_SHIFT) + uar->index;
+	uar->map      = ioremap(pfn << PAGE_SHIFT, PAGE_SIZE);
+	if (!uar->map) {
+		mlx5_core_warn(mdev, "ioremap() failed, %d\n", err);
+		err = -ENOMEM;
+		goto err_free_uar;
+	}
+
+	if (mdev->priv.bf_mapping)
+		uar->bf_map = io_mapping_map_wc(mdev->priv.bf_mapping,
+						uar->index << PAGE_SHIFT);
+
+	return 0;
+
+err_free_uar:
+	mlx5_cmd_free_uar(mdev, uar->index);
+
+	return err;
+}
+EXPORT_SYMBOL(mlx5_alloc_map_uar);
+
+void mlx5_unmap_free_uar(struct mlx5_core_dev *mdev, struct mlx5_uar *uar)
+{
+	io_mapping_unmap(uar->bf_map);
+	iounmap(uar->map);
+	mlx5_cmd_free_uar(mdev, uar->index);
+}
+EXPORT_SYMBOL(mlx5_unmap_free_uar);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/vport.c
new file mode 100644
index 000000000..b94177ebc
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/vport.c
@@ -0,0 +1,345 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 <linux/export.h>
+#include <linux/etherdevice.h>
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/vport.h>
+#include "mlx5_core.h"
+
+u8 mlx5_query_vport_state(struct mlx5_core_dev *mdev, u8 opmod)
+{
+	u32 in[MLX5_ST_SZ_DW(query_vport_state_in)];
+	u32 out[MLX5_ST_SZ_DW(query_vport_state_out)];
+	int err;
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(query_vport_state_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_VPORT_STATE);
+	MLX5_SET(query_vport_state_in, in, op_mod, opmod);
+
+	err = mlx5_cmd_exec_check_status(mdev, in, sizeof(in), out,
+					 sizeof(out));
+	if (err)
+		mlx5_core_warn(mdev, "MLX5_CMD_OP_QUERY_VPORT_STATE failed\n");
+
+	return MLX5_GET(query_vport_state_out, out, state);
+}
+EXPORT_SYMBOL(mlx5_query_vport_state);
+
+void mlx5_query_nic_vport_mac_address(struct mlx5_core_dev *mdev, u8 *addr)
+{
+	u32  in[MLX5_ST_SZ_DW(query_nic_vport_context_in)];
+	u32 *out;
+	int outlen = MLX5_ST_SZ_BYTES(query_nic_vport_context_out);
+	u8 *out_addr;
+
+	out = mlx5_vzalloc(outlen);
+	if (!out)
+		return;
+
+	out_addr = MLX5_ADDR_OF(query_nic_vport_context_out, out,
+				nic_vport_context.permanent_address);
+
+	memset(in, 0, sizeof(in));
+
+	MLX5_SET(query_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT);
+
+	memset(out, 0, outlen);
+	mlx5_cmd_exec_check_status(mdev, in, sizeof(in), out, outlen);
+
+	ether_addr_copy(addr, &out_addr[2]);
+
+	kvfree(out);
+}
+EXPORT_SYMBOL(mlx5_query_nic_vport_mac_address);
+
+int mlx5_query_hca_vport_gid(struct mlx5_core_dev *dev, u8 other_vport,
+			     u8 port_num, u16  vf_num, u16 gid_index,
+			     union ib_gid *gid)
+{
+	int in_sz = MLX5_ST_SZ_BYTES(query_hca_vport_gid_in);
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_vport_gid_out);
+	int is_group_manager;
+	void *out = NULL;
+	void *in = NULL;
+	union ib_gid *tmp;
+	int tbsz;
+	int nout;
+	int err;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+	tbsz = mlx5_get_gid_table_len(MLX5_CAP_GEN(dev, gid_table_size));
+	mlx5_core_dbg(dev, "vf_num %d, index %d, gid_table_size %d\n",
+		      vf_num, gid_index, tbsz);
+
+	if (gid_index > tbsz && gid_index != 0xffff)
+		return -EINVAL;
+
+	if (gid_index == 0xffff)
+		nout = tbsz;
+	else
+		nout = 1;
+
+	out_sz += nout * sizeof(*gid);
+
+	in = kzalloc(in_sz, GFP_KERNEL);
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(query_hca_vport_gid_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_VPORT_GID);
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_hca_vport_gid_in, in, vport_number, vf_num);
+			MLX5_SET(query_hca_vport_gid_in, in, other_vport, 1);
+		} else {
+			err = -EPERM;
+			goto out;
+		}
+	}
+	MLX5_SET(query_hca_vport_gid_in, in, gid_index, gid_index);
+
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_hca_vport_gid_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out, out_sz);
+	if (err)
+		goto out;
+
+	err = mlx5_cmd_status_to_err_v2(out);
+	if (err)
+		goto out;
+
+	tmp = out + MLX5_ST_SZ_BYTES(query_hca_vport_gid_out);
+	gid->global.subnet_prefix = tmp->global.subnet_prefix;
+	gid->global.interface_id = tmp->global.interface_id;
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_gid);
+
+int mlx5_query_hca_vport_pkey(struct mlx5_core_dev *dev, u8 other_vport,
+			      u8 port_num, u16 vf_num, u16 pkey_index,
+			      u16 *pkey)
+{
+	int in_sz = MLX5_ST_SZ_BYTES(query_hca_vport_pkey_in);
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_vport_pkey_out);
+	int is_group_manager;
+	void *out = NULL;
+	void *in = NULL;
+	void *pkarr;
+	int nout;
+	int tbsz;
+	int err;
+	int i;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+
+	tbsz = mlx5_to_sw_pkey_sz(MLX5_CAP_GEN(dev, pkey_table_size));
+	if (pkey_index > tbsz && pkey_index != 0xffff)
+		return -EINVAL;
+
+	if (pkey_index == 0xffff)
+		nout = tbsz;
+	else
+		nout = 1;
+
+	out_sz += nout * MLX5_ST_SZ_BYTES(pkey);
+
+	in = kzalloc(in_sz, GFP_KERNEL);
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!in || !out) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	MLX5_SET(query_hca_vport_pkey_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_VPORT_PKEY);
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_hca_vport_pkey_in, in, vport_number, vf_num);
+			MLX5_SET(query_hca_vport_pkey_in, in, other_vport, 1);
+		} else {
+			err = -EPERM;
+			goto out;
+		}
+	}
+	MLX5_SET(query_hca_vport_pkey_in, in, pkey_index, pkey_index);
+
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_hca_vport_pkey_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, in_sz, out, out_sz);
+	if (err)
+		goto out;
+
+	err = mlx5_cmd_status_to_err_v2(out);
+	if (err)
+		goto out;
+
+	pkarr = MLX5_ADDR_OF(query_hca_vport_pkey_out, out, pkey);
+	for (i = 0; i < nout; i++, pkey++, pkarr += MLX5_ST_SZ_BYTES(pkey))
+		*pkey = MLX5_GET_PR(pkey, pkarr, pkey);
+
+out:
+	kfree(in);
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_pkey);
+
+int mlx5_query_hca_vport_context(struct mlx5_core_dev *dev,
+				 u8 other_vport, u8 port_num,
+				 u16 vf_num,
+				 struct mlx5_hca_vport_context *rep)
+{
+	int out_sz = MLX5_ST_SZ_BYTES(query_hca_vport_context_out);
+	int in[MLX5_ST_SZ_DW(query_hca_vport_context_in)];
+	int is_group_manager;
+	void *out;
+	void *ctx;
+	int err;
+
+	is_group_manager = MLX5_CAP_GEN(dev, vport_group_manager);
+
+	memset(in, 0, sizeof(in));
+	out = kzalloc(out_sz, GFP_KERNEL);
+	if (!out)
+		return -ENOMEM;
+
+	MLX5_SET(query_hca_vport_context_in, in, opcode, MLX5_CMD_OP_QUERY_HCA_VPORT_CONTEXT);
+
+	if (other_vport) {
+		if (is_group_manager) {
+			MLX5_SET(query_hca_vport_context_in, in, other_vport, 1);
+			MLX5_SET(query_hca_vport_context_in, in, vport_number, vf_num);
+		} else {
+			err = -EPERM;
+			goto ex;
+		}
+	}
+
+	if (MLX5_CAP_GEN(dev, num_ports) == 2)
+		MLX5_SET(query_hca_vport_context_in, in, port_num, port_num);
+
+	err = mlx5_cmd_exec(dev, in, sizeof(in), out,  out_sz);
+	if (err)
+		goto ex;
+	err = mlx5_cmd_status_to_err_v2(out);
+	if (err)
+		goto ex;
+
+	ctx = MLX5_ADDR_OF(query_hca_vport_context_out, out, hca_vport_context);
+	rep->field_select = MLX5_GET_PR(hca_vport_context, ctx, field_select);
+	rep->sm_virt_aware = MLX5_GET_PR(hca_vport_context, ctx, sm_virt_aware);
+	rep->has_smi = MLX5_GET_PR(hca_vport_context, ctx, has_smi);
+	rep->has_raw = MLX5_GET_PR(hca_vport_context, ctx, has_raw);
+	rep->policy = MLX5_GET_PR(hca_vport_context, ctx, vport_state_policy);
+	rep->phys_state = MLX5_GET_PR(hca_vport_context, ctx,
+				      port_physical_state);
+	rep->vport_state = MLX5_GET_PR(hca_vport_context, ctx, vport_state);
+	rep->port_physical_state = MLX5_GET_PR(hca_vport_context, ctx,
+					       port_physical_state);
+	rep->port_guid = MLX5_GET64_PR(hca_vport_context, ctx, port_guid);
+	rep->node_guid = MLX5_GET64_PR(hca_vport_context, ctx, node_guid);
+	rep->cap_mask1 = MLX5_GET_PR(hca_vport_context, ctx, cap_mask1);
+	rep->cap_mask1_perm = MLX5_GET_PR(hca_vport_context, ctx,
+					  cap_mask1_field_select);
+	rep->cap_mask2 = MLX5_GET_PR(hca_vport_context, ctx, cap_mask2);
+	rep->cap_mask2_perm = MLX5_GET_PR(hca_vport_context, ctx,
+					  cap_mask2_field_select);
+	rep->lid = MLX5_GET_PR(hca_vport_context, ctx, lid);
+	rep->init_type_reply = MLX5_GET_PR(hca_vport_context, ctx,
+					   init_type_reply);
+	rep->lmc = MLX5_GET_PR(hca_vport_context, ctx, lmc);
+	rep->subnet_timeout = MLX5_GET_PR(hca_vport_context, ctx,
+					  subnet_timeout);
+	rep->sm_lid = MLX5_GET_PR(hca_vport_context, ctx, sm_lid);
+	rep->sm_sl = MLX5_GET_PR(hca_vport_context, ctx, sm_sl);
+	rep->qkey_violation_counter = MLX5_GET_PR(hca_vport_context, ctx,
+						  qkey_violation_counter);
+	rep->pkey_violation_counter = MLX5_GET_PR(hca_vport_context, ctx,
+						  pkey_violation_counter);
+	rep->grh_required = MLX5_GET_PR(hca_vport_context, ctx, grh_required);
+	rep->sys_image_guid = MLX5_GET64_PR(hca_vport_context, ctx,
+					    system_image_guid);
+
+ex:
+	kfree(out);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_context);
+
+int mlx5_query_hca_vport_system_image_guid(struct mlx5_core_dev *dev,
+					   u64 *sys_image_guid)
+{
+	struct mlx5_hca_vport_context *rep;
+	int err;
+
+	rep = kzalloc(sizeof(*rep), GFP_KERNEL);
+	if (!rep)
+		return -ENOMEM;
+
+	err = mlx5_query_hca_vport_context(dev, 0, 1, 0, rep);
+	if (!err)
+		*sys_image_guid = rep->sys_image_guid;
+
+	kfree(rep);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_system_image_guid);
+
+int mlx5_query_hca_vport_node_guid(struct mlx5_core_dev *dev,
+				   u64 *node_guid)
+{
+	struct mlx5_hca_vport_context *rep;
+	int err;
+
+	rep = kzalloc(sizeof(*rep), GFP_KERNEL);
+	if (!rep)
+		return -ENOMEM;
+
+	err = mlx5_query_hca_vport_context(dev, 0, 1, 0, rep);
+	if (!err)
+		*node_guid = rep->node_guid;
+
+	kfree(rep);
+	return err;
+}
+EXPORT_SYMBOL_GPL(mlx5_query_hca_vport_node_guid);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/wq.c b/kernel/drivers/net/ethernet/mellanox/mlx5/core/wq.c
new file mode 100644
index 000000000..ce21ee5b2
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/wq.c
@@ -0,0 +1,185 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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 <linux/mlx5/driver.h>
+#include "wq.h"
+#include "mlx5_core.h"
+
+u32 mlx5_wq_cyc_get_size(struct mlx5_wq_cyc *wq)
+{
+	return (u32)wq->sz_m1 + 1;
+}
+
+u32 mlx5_cqwq_get_size(struct mlx5_cqwq *wq)
+{
+	return wq->sz_m1 + 1;
+}
+
+u32 mlx5_wq_ll_get_size(struct mlx5_wq_ll *wq)
+{
+	return (u32)wq->sz_m1 + 1;
+}
+
+static u32 mlx5_wq_cyc_get_byte_size(struct mlx5_wq_cyc *wq)
+{
+	return mlx5_wq_cyc_get_size(wq) << wq->log_stride;
+}
+
+static u32 mlx5_cqwq_get_byte_size(struct mlx5_cqwq *wq)
+{
+	return mlx5_cqwq_get_size(wq) << wq->log_stride;
+}
+
+static u32 mlx5_wq_ll_get_byte_size(struct mlx5_wq_ll *wq)
+{
+	return mlx5_wq_ll_get_size(wq) << wq->log_stride;
+}
+
+int mlx5_wq_cyc_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		       void *wqc, struct mlx5_wq_cyc *wq,
+		       struct mlx5_wq_ctrl *wq_ctrl)
+{
+	int err;
+
+	wq->log_stride = MLX5_GET(wq, wqc, log_wq_stride);
+	wq->sz_m1 = (1 << MLX5_GET(wq, wqc, log_wq_sz)) - 1;
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_buf_alloc_node(mdev, mlx5_wq_cyc_get_byte_size(wq),
+				  &wq_ctrl->buf, param->buf_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_buf_alloc() failed, %d\n", err);
+		goto err_db_free;
+	}
+
+	wq->buf = wq_ctrl->buf.direct.buf;
+	wq->db  = wq_ctrl->db.db;
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		     void *cqc, struct mlx5_cqwq *wq,
+		     struct mlx5_wq_ctrl *wq_ctrl)
+{
+	int err;
+
+	wq->log_stride = 6 + MLX5_GET(cqc, cqc, cqe_sz);
+	wq->log_sz = MLX5_GET(cqc, cqc, log_cq_size);
+	wq->sz_m1 = (1 << wq->log_sz) - 1;
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_buf_alloc_node(mdev, mlx5_cqwq_get_byte_size(wq),
+				  &wq_ctrl->buf, param->buf_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_buf_alloc() failed, %d\n", err);
+		goto err_db_free;
+	}
+
+	wq->buf = wq_ctrl->buf.direct.buf;
+	wq->db  = wq_ctrl->db.db;
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		      void *wqc, struct mlx5_wq_ll *wq,
+		      struct mlx5_wq_ctrl *wq_ctrl)
+{
+	struct mlx5_wqe_srq_next_seg *next_seg;
+	int err;
+	int i;
+
+	wq->log_stride = MLX5_GET(wq, wqc, log_wq_stride);
+	wq->sz_m1 = (1 << MLX5_GET(wq, wqc, log_wq_sz)) - 1;
+
+	err = mlx5_db_alloc_node(mdev, &wq_ctrl->db, param->db_numa_node);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_db_alloc() failed, %d\n", err);
+		return err;
+	}
+
+	err = mlx5_buf_alloc(mdev, mlx5_wq_ll_get_byte_size(wq), &wq_ctrl->buf);
+	if (err) {
+		mlx5_core_warn(mdev, "mlx5_buf_alloc() failed, %d\n", err);
+		goto err_db_free;
+	}
+
+	wq->buf = wq_ctrl->buf.direct.buf;
+	wq->db  = wq_ctrl->db.db;
+
+	for (i = 0; i < wq->sz_m1; i++) {
+		next_seg = mlx5_wq_ll_get_wqe(wq, i);
+		next_seg->next_wqe_index = cpu_to_be16(i + 1);
+	}
+	next_seg = mlx5_wq_ll_get_wqe(wq, i);
+	wq->tail_next = &next_seg->next_wqe_index;
+
+	wq_ctrl->mdev = mdev;
+
+	return 0;
+
+err_db_free:
+	mlx5_db_free(mdev, &wq_ctrl->db);
+
+	return err;
+}
+
+void mlx5_wq_destroy(struct mlx5_wq_ctrl *wq_ctrl)
+{
+	mlx5_buf_free(wq_ctrl->mdev, &wq_ctrl->buf);
+	mlx5_db_free(wq_ctrl->mdev, &wq_ctrl->db);
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlx5/core/wq.h b/kernel/drivers/net/ethernet/mellanox/mlx5/core/wq.h
new file mode 100644
index 000000000..6c2a8f950
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlx5/core/wq.h
@@ -0,0 +1,172 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  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_WQ_H__
+#define __MLX5_WQ_H__
+
+#include <linux/mlx5/mlx5_ifc.h>
+
+struct mlx5_wq_param {
+	int		linear;
+	int		buf_numa_node;
+	int		db_numa_node;
+};
+
+struct mlx5_wq_ctrl {
+	struct mlx5_core_dev	*mdev;
+	struct mlx5_buf		buf;
+	struct mlx5_db		db;
+};
+
+struct mlx5_wq_cyc {
+	void			*buf;
+	__be32			*db;
+	u16			sz_m1;
+	u8			log_stride;
+};
+
+struct mlx5_cqwq {
+	void			*buf;
+	__be32			*db;
+	u32			sz_m1;
+	u32			cc; /* consumer counter */
+	u8			log_sz;
+	u8			log_stride;
+};
+
+struct mlx5_wq_ll {
+	void			*buf;
+	__be32			*db;
+	__be16			*tail_next;
+	u16			sz_m1;
+	u16			head;
+	u16			wqe_ctr;
+	u16			cur_sz;
+	u8			log_stride;
+};
+
+int mlx5_wq_cyc_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		       void *wqc, struct mlx5_wq_cyc *wq,
+		       struct mlx5_wq_ctrl *wq_ctrl);
+u32 mlx5_wq_cyc_get_size(struct mlx5_wq_cyc *wq);
+
+int mlx5_cqwq_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		     void *cqc, struct mlx5_cqwq *wq,
+		     struct mlx5_wq_ctrl *wq_ctrl);
+u32 mlx5_cqwq_get_size(struct mlx5_cqwq *wq);
+
+int mlx5_wq_ll_create(struct mlx5_core_dev *mdev, struct mlx5_wq_param *param,
+		      void *wqc, struct mlx5_wq_ll *wq,
+		      struct mlx5_wq_ctrl *wq_ctrl);
+u32 mlx5_wq_ll_get_size(struct mlx5_wq_ll *wq);
+
+void mlx5_wq_destroy(struct mlx5_wq_ctrl *wq_ctrl);
+
+static inline u16 mlx5_wq_cyc_ctr2ix(struct mlx5_wq_cyc *wq, u16 ctr)
+{
+	return ctr & wq->sz_m1;
+}
+
+static inline void *mlx5_wq_cyc_get_wqe(struct mlx5_wq_cyc *wq, u16 ix)
+{
+	return wq->buf + (ix << wq->log_stride);
+}
+
+static inline int mlx5_wq_cyc_cc_bigger(u16 cc1, u16 cc2)
+{
+	int equal   = (cc1 == cc2);
+	int smaller = 0x8000 & (cc1 - cc2);
+
+	return !equal && !smaller;
+}
+
+static inline u32 mlx5_cqwq_get_ci(struct mlx5_cqwq *wq)
+{
+	return wq->cc & wq->sz_m1;
+}
+
+static inline void *mlx5_cqwq_get_wqe(struct mlx5_cqwq *wq, u32 ix)
+{
+	return wq->buf + (ix << wq->log_stride);
+}
+
+static inline u32 mlx5_cqwq_get_wrap_cnt(struct mlx5_cqwq *wq)
+{
+	return wq->cc >> wq->log_sz;
+}
+
+static inline void mlx5_cqwq_pop(struct mlx5_cqwq *wq)
+{
+	wq->cc++;
+}
+
+static inline void mlx5_cqwq_update_db_record(struct mlx5_cqwq *wq)
+{
+	*wq->db = cpu_to_be32(wq->cc & 0xffffff);
+}
+
+static inline int mlx5_wq_ll_is_full(struct mlx5_wq_ll *wq)
+{
+	return wq->cur_sz == wq->sz_m1;
+}
+
+static inline int mlx5_wq_ll_is_empty(struct mlx5_wq_ll *wq)
+{
+	return !wq->cur_sz;
+}
+
+static inline void *mlx5_wq_ll_get_wqe(struct mlx5_wq_ll *wq, u16 ix)
+{
+	return wq->buf + (ix << wq->log_stride);
+}
+
+static inline void mlx5_wq_ll_push(struct mlx5_wq_ll *wq, u16 head_next)
+{
+	wq->head = head_next;
+	wq->wqe_ctr++;
+	wq->cur_sz++;
+}
+
+static inline void mlx5_wq_ll_pop(struct mlx5_wq_ll *wq, __be16 ix,
+				  __be16 *next_tail_next)
+{
+	*wq->tail_next = ix;
+	wq->tail_next = next_tail_next;
+	wq->cur_sz--;
+}
+
+static inline void mlx5_wq_ll_update_db_record(struct mlx5_wq_ll *wq)
+{
+	*wq->db = cpu_to_be32(wq->wqe_ctr);
+}
+
+#endif /* __MLX5_WQ_H__ */
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/Kconfig b/kernel/drivers/net/ethernet/mellanox/mlxsw/Kconfig
new file mode 100644
index 000000000..e36e12219
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/Kconfig
@@ -0,0 +1,43 @@
+#
+# Mellanox switch drivers configuration
+#
+
+config MLXSW_CORE
+	tristate "Mellanox Technologies Switch ASICs support"
+	---help---
+	  This driver supports Mellanox Technologies Switch ASICs family.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called mlxsw_core.
+
+config MLXSW_PCI
+	tristate "PCI bus implementation for Mellanox Technologies Switch ASICs"
+	depends on PCI && HAS_DMA && HAS_IOMEM && MLXSW_CORE
+	default m
+	---help---
+	  This is PCI bus implementation for Mellanox Technologies Switch ASICs.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called mlxsw_pci.
+
+config MLXSW_SWITCHX2
+	tristate "Mellanox Technologies SwitchX-2 support"
+	depends on MLXSW_CORE && NET_SWITCHDEV
+	default m
+	---help---
+	  This driver supports Mellanox Technologies SwitchX-2 Ethernet
+	  Switch ASICs.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called mlxsw_switchx2.
+
+config MLXSW_SPECTRUM
+	tristate "Mellanox Technologies Spectrum support"
+	depends on MLXSW_CORE && NET_SWITCHDEV
+	default m
+	---help---
+	  This driver supports Mellanox Technologies Spectrum Ethernet
+	  Switch ASICs.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called mlxsw_spectrum.
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/Makefile b/kernel/drivers/net/ethernet/mellanox/mlxsw/Makefile
new file mode 100644
index 000000000..af015818f
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/Makefile
@@ -0,0 +1,9 @@
+obj-$(CONFIG_MLXSW_CORE)	+= mlxsw_core.o
+mlxsw_core-objs			:= core.o
+obj-$(CONFIG_MLXSW_PCI)		+= mlxsw_pci.o
+mlxsw_pci-objs			:= pci.o
+obj-$(CONFIG_MLXSW_SWITCHX2)	+= mlxsw_switchx2.o
+mlxsw_switchx2-objs		:= switchx2.o
+obj-$(CONFIG_MLXSW_SPECTRUM)	+= mlxsw_spectrum.o
+mlxsw_spectrum-objs		:= spectrum.o spectrum_buffers.o \
+				   spectrum_switchdev.o
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/cmd.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/cmd.h
new file mode 100644
index 000000000..cd63b8263
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/cmd.h
@@ -0,0 +1,1115 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/cmd.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_CMD_H
+#define _MLXSW_CMD_H
+
+#include "item.h"
+
+#define MLXSW_CMD_MBOX_SIZE	4096
+
+static inline char *mlxsw_cmd_mbox_alloc(void)
+{
+	return kzalloc(MLXSW_CMD_MBOX_SIZE, GFP_KERNEL);
+}
+
+static inline void mlxsw_cmd_mbox_free(char *mbox)
+{
+	kfree(mbox);
+}
+
+static inline void mlxsw_cmd_mbox_zero(char *mbox)
+{
+	memset(mbox, 0, MLXSW_CMD_MBOX_SIZE);
+}
+
+struct mlxsw_core;
+
+int mlxsw_cmd_exec(struct mlxsw_core *mlxsw_core, u16 opcode, u8 opcode_mod,
+		   u32 in_mod, bool out_mbox_direct,
+		   char *in_mbox, size_t in_mbox_size,
+		   char *out_mbox, size_t out_mbox_size);
+
+static inline int mlxsw_cmd_exec_in(struct mlxsw_core *mlxsw_core, u16 opcode,
+				    u8 opcode_mod, u32 in_mod, char *in_mbox,
+				    size_t in_mbox_size)
+{
+	return mlxsw_cmd_exec(mlxsw_core, opcode, opcode_mod, in_mod, false,
+			      in_mbox, in_mbox_size, NULL, 0);
+}
+
+static inline int mlxsw_cmd_exec_out(struct mlxsw_core *mlxsw_core, u16 opcode,
+				     u8 opcode_mod, u32 in_mod,
+				     bool out_mbox_direct,
+				     char *out_mbox, size_t out_mbox_size)
+{
+	return mlxsw_cmd_exec(mlxsw_core, opcode, opcode_mod, in_mod,
+			      out_mbox_direct, NULL, 0,
+			      out_mbox, out_mbox_size);
+}
+
+static inline int mlxsw_cmd_exec_none(struct mlxsw_core *mlxsw_core, u16 opcode,
+				      u8 opcode_mod, u32 in_mod)
+{
+	return mlxsw_cmd_exec(mlxsw_core, opcode, opcode_mod, in_mod, false,
+			      NULL, 0, NULL, 0);
+}
+
+enum mlxsw_cmd_opcode {
+	MLXSW_CMD_OPCODE_QUERY_FW		= 0x004,
+	MLXSW_CMD_OPCODE_QUERY_BOARDINFO	= 0x006,
+	MLXSW_CMD_OPCODE_QUERY_AQ_CAP		= 0x003,
+	MLXSW_CMD_OPCODE_MAP_FA			= 0xFFF,
+	MLXSW_CMD_OPCODE_UNMAP_FA		= 0xFFE,
+	MLXSW_CMD_OPCODE_CONFIG_PROFILE		= 0x100,
+	MLXSW_CMD_OPCODE_ACCESS_REG		= 0x040,
+	MLXSW_CMD_OPCODE_SW2HW_DQ		= 0x201,
+	MLXSW_CMD_OPCODE_HW2SW_DQ		= 0x202,
+	MLXSW_CMD_OPCODE_2ERR_DQ		= 0x01E,
+	MLXSW_CMD_OPCODE_QUERY_DQ		= 0x022,
+	MLXSW_CMD_OPCODE_SW2HW_CQ		= 0x016,
+	MLXSW_CMD_OPCODE_HW2SW_CQ		= 0x017,
+	MLXSW_CMD_OPCODE_QUERY_CQ		= 0x018,
+	MLXSW_CMD_OPCODE_SW2HW_EQ		= 0x013,
+	MLXSW_CMD_OPCODE_HW2SW_EQ		= 0x014,
+	MLXSW_CMD_OPCODE_QUERY_EQ		= 0x015,
+};
+
+static inline const char *mlxsw_cmd_opcode_str(u16 opcode)
+{
+	switch (opcode) {
+	case MLXSW_CMD_OPCODE_QUERY_FW:
+		return "QUERY_FW";
+	case MLXSW_CMD_OPCODE_QUERY_BOARDINFO:
+		return "QUERY_BOARDINFO";
+	case MLXSW_CMD_OPCODE_QUERY_AQ_CAP:
+		return "QUERY_AQ_CAP";
+	case MLXSW_CMD_OPCODE_MAP_FA:
+		return "MAP_FA";
+	case MLXSW_CMD_OPCODE_UNMAP_FA:
+		return "UNMAP_FA";
+	case MLXSW_CMD_OPCODE_CONFIG_PROFILE:
+		return "CONFIG_PROFILE";
+	case MLXSW_CMD_OPCODE_ACCESS_REG:
+		return "ACCESS_REG";
+	case MLXSW_CMD_OPCODE_SW2HW_DQ:
+		return "SW2HW_DQ";
+	case MLXSW_CMD_OPCODE_HW2SW_DQ:
+		return "HW2SW_DQ";
+	case MLXSW_CMD_OPCODE_2ERR_DQ:
+		return "2ERR_DQ";
+	case MLXSW_CMD_OPCODE_QUERY_DQ:
+		return "QUERY_DQ";
+	case MLXSW_CMD_OPCODE_SW2HW_CQ:
+		return "SW2HW_CQ";
+	case MLXSW_CMD_OPCODE_HW2SW_CQ:
+		return "HW2SW_CQ";
+	case MLXSW_CMD_OPCODE_QUERY_CQ:
+		return "QUERY_CQ";
+	case MLXSW_CMD_OPCODE_SW2HW_EQ:
+		return "SW2HW_EQ";
+	case MLXSW_CMD_OPCODE_HW2SW_EQ:
+		return "HW2SW_EQ";
+	case MLXSW_CMD_OPCODE_QUERY_EQ:
+		return "QUERY_EQ";
+	default:
+		return "*UNKNOWN*";
+	}
+}
+
+enum mlxsw_cmd_status {
+	/* Command execution succeeded. */
+	MLXSW_CMD_STATUS_OK		= 0x00,
+	/* Internal error (e.g. bus error) occurred while processing command. */
+	MLXSW_CMD_STATUS_INTERNAL_ERR	= 0x01,
+	/* Operation/command not supported or opcode modifier not supported. */
+	MLXSW_CMD_STATUS_BAD_OP		= 0x02,
+	/* Parameter not supported, parameter out of range. */
+	MLXSW_CMD_STATUS_BAD_PARAM	= 0x03,
+	/* System was not enabled or bad system state. */
+	MLXSW_CMD_STATUS_BAD_SYS_STATE	= 0x04,
+	/* Attempt to access reserved or unallocated resource, or resource in
+	 * inappropriate ownership.
+	 */
+	MLXSW_CMD_STATUS_BAD_RESOURCE	= 0x05,
+	/* Requested resource is currently executing a command. */
+	MLXSW_CMD_STATUS_RESOURCE_BUSY	= 0x06,
+	/* Required capability exceeds device limits. */
+	MLXSW_CMD_STATUS_EXCEED_LIM	= 0x08,
+	/* Resource is not in the appropriate state or ownership. */
+	MLXSW_CMD_STATUS_BAD_RES_STATE	= 0x09,
+	/* Index out of range (might be beyond table size or attempt to
+	 * access a reserved resource).
+	 */
+	MLXSW_CMD_STATUS_BAD_INDEX	= 0x0A,
+	/* NVMEM checksum/CRC failed. */
+	MLXSW_CMD_STATUS_BAD_NVMEM	= 0x0B,
+	/* Bad management packet (silently discarded). */
+	MLXSW_CMD_STATUS_BAD_PKT	= 0x30,
+};
+
+static inline const char *mlxsw_cmd_status_str(u8 status)
+{
+	switch (status) {
+	case MLXSW_CMD_STATUS_OK:
+		return "OK";
+	case MLXSW_CMD_STATUS_INTERNAL_ERR:
+		return "INTERNAL_ERR";
+	case MLXSW_CMD_STATUS_BAD_OP:
+		return "BAD_OP";
+	case MLXSW_CMD_STATUS_BAD_PARAM:
+		return "BAD_PARAM";
+	case MLXSW_CMD_STATUS_BAD_SYS_STATE:
+		return "BAD_SYS_STATE";
+	case MLXSW_CMD_STATUS_BAD_RESOURCE:
+		return "BAD_RESOURCE";
+	case MLXSW_CMD_STATUS_RESOURCE_BUSY:
+		return "RESOURCE_BUSY";
+	case MLXSW_CMD_STATUS_EXCEED_LIM:
+		return "EXCEED_LIM";
+	case MLXSW_CMD_STATUS_BAD_RES_STATE:
+		return "BAD_RES_STATE";
+	case MLXSW_CMD_STATUS_BAD_INDEX:
+		return "BAD_INDEX";
+	case MLXSW_CMD_STATUS_BAD_NVMEM:
+		return "BAD_NVMEM";
+	case MLXSW_CMD_STATUS_BAD_PKT:
+		return "BAD_PKT";
+	default:
+		return "*UNKNOWN*";
+	}
+}
+
+/* QUERY_FW - Query Firmware
+ * -------------------------
+ * OpMod == 0, INMmod == 0
+ * -----------------------
+ * The QUERY_FW command retrieves information related to firmware, command
+ * interface version and the amount of resources that should be allocated to
+ * the firmware.
+ */
+
+static inline int mlxsw_cmd_query_fw(struct mlxsw_core *mlxsw_core,
+				     char *out_mbox)
+{
+	return mlxsw_cmd_exec_out(mlxsw_core, MLXSW_CMD_OPCODE_QUERY_FW,
+				  0, 0, false, out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_query_fw_fw_pages
+ * Amount of physical memory to be allocatedfor firmware usage in 4KB pages.
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_pages, 0x00, 16, 16);
+
+/* cmd_mbox_query_fw_fw_rev_major
+ * Firmware Revision - Major
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_rev_major, 0x00, 0, 16);
+
+/* cmd_mbox_query_fw_fw_rev_subminor
+ * Firmware Sub-minor version (Patch level)
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_rev_subminor, 0x04, 16, 16);
+
+/* cmd_mbox_query_fw_fw_rev_minor
+ * Firmware Revision - Minor
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_rev_minor, 0x04, 0, 16);
+
+/* cmd_mbox_query_fw_core_clk
+ * Internal Clock Frequency (in MHz)
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, core_clk, 0x08, 16, 16);
+
+/* cmd_mbox_query_fw_cmd_interface_rev
+ * Command Interface Interpreter Revision ID. This number is bumped up
+ * every time a non-backward-compatible change is done for the command
+ * interface. The current cmd_interface_rev is 1.
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, cmd_interface_rev, 0x08, 0, 16);
+
+/* cmd_mbox_query_fw_dt
+ * If set, Debug Trace is supported
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, dt, 0x0C, 31, 1);
+
+/* cmd_mbox_query_fw_api_version
+ * Indicates the version of the API, to enable software querying
+ * for compatibility. The current api_version is 1.
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, api_version, 0x0C, 0, 16);
+
+/* cmd_mbox_query_fw_fw_hour
+ * Firmware timestamp - hour
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_hour, 0x10, 24, 8);
+
+/* cmd_mbox_query_fw_fw_minutes
+ * Firmware timestamp - minutes
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_minutes, 0x10, 16, 8);
+
+/* cmd_mbox_query_fw_fw_seconds
+ * Firmware timestamp - seconds
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_seconds, 0x10, 8, 8);
+
+/* cmd_mbox_query_fw_fw_year
+ * Firmware timestamp - year
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_year, 0x14, 16, 16);
+
+/* cmd_mbox_query_fw_fw_month
+ * Firmware timestamp - month
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_month, 0x14, 8, 8);
+
+/* cmd_mbox_query_fw_fw_day
+ * Firmware timestamp - day
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, fw_day, 0x14, 0, 8);
+
+/* cmd_mbox_query_fw_clr_int_base_offset
+ * Clear Interrupt register's offset from clr_int_bar register
+ * in PCI address space.
+ */
+MLXSW_ITEM64(cmd_mbox, query_fw, clr_int_base_offset, 0x20, 0, 64);
+
+/* cmd_mbox_query_fw_clr_int_bar
+ * PCI base address register (BAR) where clr_int register is located.
+ * 00 - BAR 0-1 (64 bit BAR)
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, clr_int_bar, 0x28, 30, 2);
+
+/* cmd_mbox_query_fw_error_buf_offset
+ * Read Only buffer for internal error reports of offset
+ * from error_buf_bar register in PCI address space).
+ */
+MLXSW_ITEM64(cmd_mbox, query_fw, error_buf_offset, 0x30, 0, 64);
+
+/* cmd_mbox_query_fw_error_buf_size
+ * Internal error buffer size in DWORDs
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, error_buf_size, 0x38, 0, 32);
+
+/* cmd_mbox_query_fw_error_int_bar
+ * PCI base address register (BAR) where error buffer
+ * register is located.
+ * 00 - BAR 0-1 (64 bit BAR)
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, error_int_bar, 0x3C, 30, 2);
+
+/* cmd_mbox_query_fw_doorbell_page_offset
+ * Offset of the doorbell page
+ */
+MLXSW_ITEM64(cmd_mbox, query_fw, doorbell_page_offset, 0x40, 0, 64);
+
+/* cmd_mbox_query_fw_doorbell_page_bar
+ * PCI base address register (BAR) of the doorbell page
+ * 00 - BAR 0-1 (64 bit BAR)
+ */
+MLXSW_ITEM32(cmd_mbox, query_fw, doorbell_page_bar, 0x48, 30, 2);
+
+/* QUERY_BOARDINFO - Query Board Information
+ * -----------------------------------------
+ * OpMod == 0 (N/A), INMmod == 0 (N/A)
+ * -----------------------------------
+ * The QUERY_BOARDINFO command retrieves adapter specific parameters.
+ */
+
+static inline int mlxsw_cmd_boardinfo(struct mlxsw_core *mlxsw_core,
+				      char *out_mbox)
+{
+	return mlxsw_cmd_exec_out(mlxsw_core, MLXSW_CMD_OPCODE_QUERY_BOARDINFO,
+				  0, 0, false, out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_boardinfo_intapin
+ * When PCIe interrupt messages are being used, this value is used for clearing
+ * an interrupt. When using MSI-X, this register is not used.
+ */
+MLXSW_ITEM32(cmd_mbox, boardinfo, intapin, 0x10, 24, 8);
+
+/* cmd_mbox_boardinfo_vsd_vendor_id
+ * PCISIG Vendor ID (www.pcisig.com/membership/vid_search) of the vendor
+ * specifying/formatting the VSD. The vsd_vendor_id identifies the management
+ * domain of the VSD/PSID data. Different vendors may choose different VSD/PSID
+ * format and encoding as long as they use their assigned vsd_vendor_id.
+ */
+MLXSW_ITEM32(cmd_mbox, boardinfo, vsd_vendor_id, 0x1C, 0, 16);
+
+/* cmd_mbox_boardinfo_vsd
+ * Vendor Specific Data. The VSD string that is burnt to the Flash
+ * with the firmware.
+ */
+#define MLXSW_CMD_BOARDINFO_VSD_LEN 208
+MLXSW_ITEM_BUF(cmd_mbox, boardinfo, vsd, 0x20, MLXSW_CMD_BOARDINFO_VSD_LEN);
+
+/* cmd_mbox_boardinfo_psid
+ * The PSID field is a 16-ascii (byte) character string which acts as
+ * the board ID. The PSID format is used in conjunction with
+ * Mellanox vsd_vendor_id (15B3h).
+ */
+#define MLXSW_CMD_BOARDINFO_PSID_LEN 16
+MLXSW_ITEM_BUF(cmd_mbox, boardinfo, psid, 0xF0, MLXSW_CMD_BOARDINFO_PSID_LEN);
+
+/* QUERY_AQ_CAP - Query Asynchronous Queues Capabilities
+ * -----------------------------------------------------
+ * OpMod == 0 (N/A), INMmod == 0 (N/A)
+ * -----------------------------------
+ * The QUERY_AQ_CAP command returns the device asynchronous queues
+ * capabilities supported.
+ */
+
+static inline int mlxsw_cmd_query_aq_cap(struct mlxsw_core *mlxsw_core,
+					 char *out_mbox)
+{
+	return mlxsw_cmd_exec_out(mlxsw_core, MLXSW_CMD_OPCODE_QUERY_AQ_CAP,
+				  0, 0, false, out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_query_aq_cap_log_max_sdq_sz
+ * Log (base 2) of max WQEs allowed on SDQ.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, log_max_sdq_sz, 0x00, 24, 8);
+
+/* cmd_mbox_query_aq_cap_max_num_sdqs
+ * Maximum number of SDQs.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, max_num_sdqs, 0x00, 0, 8);
+
+/* cmd_mbox_query_aq_cap_log_max_rdq_sz
+ * Log (base 2) of max WQEs allowed on RDQ.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, log_max_rdq_sz, 0x04, 24, 8);
+
+/* cmd_mbox_query_aq_cap_max_num_rdqs
+ * Maximum number of RDQs.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, max_num_rdqs, 0x04, 0, 8);
+
+/* cmd_mbox_query_aq_cap_log_max_cq_sz
+ * Log (base 2) of max CQEs allowed on CQ.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, log_max_cq_sz, 0x08, 24, 8);
+
+/* cmd_mbox_query_aq_cap_max_num_cqs
+ * Maximum number of CQs.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, max_num_cqs, 0x08, 0, 8);
+
+/* cmd_mbox_query_aq_cap_log_max_eq_sz
+ * Log (base 2) of max EQEs allowed on EQ.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, log_max_eq_sz, 0x0C, 24, 8);
+
+/* cmd_mbox_query_aq_cap_max_num_eqs
+ * Maximum number of EQs.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, max_num_eqs, 0x0C, 0, 8);
+
+/* cmd_mbox_query_aq_cap_max_sg_sq
+ * The maximum S/G list elements in an DSQ. DSQ must not contain
+ * more S/G entries than indicated here.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, max_sg_sq, 0x10, 8, 8);
+
+/* cmd_mbox_query_aq_cap_
+ * The maximum S/G list elements in an DRQ. DRQ must not contain
+ * more S/G entries than indicated here.
+ */
+MLXSW_ITEM32(cmd_mbox, query_aq_cap, max_sg_rq, 0x10, 0, 8);
+
+/* MAP_FA - Map Firmware Area
+ * --------------------------
+ * OpMod == 0 (N/A), INMmod == Number of VPM entries
+ * -------------------------------------------------
+ * The MAP_FA command passes physical pages to the switch. These pages
+ * are used to store the device firmware. MAP_FA can be executed multiple
+ * times until all the firmware area is mapped (the size that should be
+ * mapped is retrieved through the QUERY_FW command). All required pages
+ * must be mapped to finish the initialization phase. Physical memory
+ * passed in this command must be pinned.
+ */
+
+#define MLXSW_CMD_MAP_FA_VPM_ENTRIES_MAX 32
+
+static inline int mlxsw_cmd_map_fa(struct mlxsw_core *mlxsw_core,
+				   char *in_mbox, u32 vpm_entries_count)
+{
+	return mlxsw_cmd_exec_in(mlxsw_core, MLXSW_CMD_OPCODE_MAP_FA,
+				 0, vpm_entries_count,
+				 in_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_map_fa_pa
+ * Physical Address.
+ */
+MLXSW_ITEM64_INDEXED(cmd_mbox, map_fa, pa, 0x00, 12, 52, 0x08, 0x00, true);
+
+/* cmd_mbox_map_fa_log2size
+ * Log (base 2) of the size in 4KB pages of the physical and contiguous memory
+ * that starts at PA_L/H.
+ */
+MLXSW_ITEM32_INDEXED(cmd_mbox, map_fa, log2size, 0x00, 0, 5, 0x08, 0x04, false);
+
+/* UNMAP_FA - Unmap Firmware Area
+ * ------------------------------
+ * OpMod == 0 (N/A), INMmod == 0 (N/A)
+ * -----------------------------------
+ * The UNMAP_FA command unload the firmware and unmaps all the
+ * firmware area. After this command is completed the device will not access
+ * the pages that were mapped to the firmware area. After executing UNMAP_FA
+ * command, software reset must be done prior to execution of MAP_FW command.
+ */
+
+static inline int mlxsw_cmd_unmap_fa(struct mlxsw_core *mlxsw_core)
+{
+	return mlxsw_cmd_exec_none(mlxsw_core, MLXSW_CMD_OPCODE_UNMAP_FA, 0, 0);
+}
+
+/* CONFIG_PROFILE (Set) - Configure Switch Profile
+ * ------------------------------
+ * OpMod == 1 (Set), INMmod == 0 (N/A)
+ * -----------------------------------
+ * The CONFIG_PROFILE command sets the switch profile. The command can be
+ * executed on the device only once at startup in order to allocate and
+ * configure all switch resources and prepare it for operational mode.
+ * It is not possible to change the device profile after the chip is
+ * in operational mode.
+ * Failure of the CONFIG_PROFILE command leaves the hardware in an indeterminate
+ * state therefore it is required to perform software reset to the device
+ * following an unsuccessful completion of the command. It is required
+ * to perform software reset to the device to change an existing profile.
+ */
+
+static inline int mlxsw_cmd_config_profile_set(struct mlxsw_core *mlxsw_core,
+					       char *in_mbox)
+{
+	return mlxsw_cmd_exec_in(mlxsw_core, MLXSW_CMD_OPCODE_CONFIG_PROFILE,
+				 1, 0, in_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_config_profile_set_max_vepa_channels
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_vepa_channels, 0x0C, 0, 1);
+
+/* cmd_mbox_config_profile_set_max_lag
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_lag, 0x0C, 1, 1);
+
+/* cmd_mbox_config_profile_set_max_port_per_lag
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_port_per_lag, 0x0C, 2, 1);
+
+/* cmd_mbox_config_profile_set_max_mid
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_mid, 0x0C, 3, 1);
+
+/* cmd_mbox_config_profile_set_max_pgt
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_pgt, 0x0C, 4, 1);
+
+/* cmd_mbox_config_profile_set_max_system_port
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_system_port, 0x0C, 5, 1);
+
+/* cmd_mbox_config_profile_set_max_vlan_groups
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_vlan_groups, 0x0C, 6, 1);
+
+/* cmd_mbox_config_profile_set_max_regions
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_regions, 0x0C, 7, 1);
+
+/* cmd_mbox_config_profile_set_flood_mode
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_flood_mode, 0x0C, 8, 1);
+
+/* cmd_mbox_config_profile_set_max_flood_tables
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_flood_tables, 0x0C, 9, 1);
+
+/* cmd_mbox_config_profile_set_max_ib_mc
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_ib_mc, 0x0C, 12, 1);
+
+/* cmd_mbox_config_profile_set_max_pkey
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_max_pkey, 0x0C, 13, 1);
+
+/* cmd_mbox_config_profile_set_adaptive_routing_group_cap
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile,
+	     set_adaptive_routing_group_cap, 0x0C, 14, 1);
+
+/* cmd_mbox_config_profile_set_ar_sec
+ * Capability bit. Setting a bit to 1 configures the profile
+ * according to the mailbox contents.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, set_ar_sec, 0x0C, 15, 1);
+
+/* cmd_mbox_config_profile_max_vepa_channels
+ * Maximum number of VEPA channels per port (0 through 16)
+ * 0 - multi-channel VEPA is disabled
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_vepa_channels, 0x10, 0, 8);
+
+/* cmd_mbox_config_profile_max_lag
+ * Maximum number of LAG IDs requested.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_lag, 0x14, 0, 16);
+
+/* cmd_mbox_config_profile_max_port_per_lag
+ * Maximum number of ports per LAG requested.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_port_per_lag, 0x18, 0, 16);
+
+/* cmd_mbox_config_profile_max_mid
+ * Maximum Multicast IDs.
+ * Multicast IDs are allocated from 0 to max_mid-1
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_mid, 0x1C, 0, 16);
+
+/* cmd_mbox_config_profile_max_pgt
+ * Maximum records in the Port Group Table per Switch Partition.
+ * Port Group Table indexes are from 0 to max_pgt-1
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_pgt, 0x20, 0, 16);
+
+/* cmd_mbox_config_profile_max_system_port
+ * The maximum number of system ports that can be allocated.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_system_port, 0x24, 0, 16);
+
+/* cmd_mbox_config_profile_max_vlan_groups
+ * Maximum number VLAN Groups for VLAN binding.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_vlan_groups, 0x28, 0, 12);
+
+/* cmd_mbox_config_profile_max_regions
+ * Maximum number of TCAM Regions.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_regions, 0x2C, 0, 16);
+
+/* cmd_mbox_config_profile_max_flood_tables
+ * Maximum number of single-entry flooding tables. Different flooding tables
+ * can be associated with different packet types.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_flood_tables, 0x30, 16, 4);
+
+/* cmd_mbox_config_profile_max_vid_flood_tables
+ * Maximum number of per-vid flooding tables. Flooding tables are associated
+ * to the different packet types for the different switch partitions.
+ * Table size is 4K entries covering all VID space.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_vid_flood_tables, 0x30, 8, 4);
+
+/* cmd_mbox_config_profile_flood_mode
+ * Flooding mode to use.
+ * 0-2 - Backward compatible modes for SwitchX devices.
+ * 3 - Mixed mode, where:
+ * max_flood_tables indicates the number of single-entry tables.
+ * max_vid_flood_tables indicates the number of per-VID tables.
+ * max_fid_offset_flood_tables indicates the number of FID-offset tables.
+ * max_fid_flood_tables indicates the number of per-FID tables.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, flood_mode, 0x30, 0, 2);
+
+/* cmd_mbox_config_profile_max_fid_offset_flood_tables
+ * Maximum number of FID-offset flooding tables.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile,
+	     max_fid_offset_flood_tables, 0x34, 24, 4);
+
+/* cmd_mbox_config_profile_fid_offset_flood_table_size
+ * The size (number of entries) of each FID-offset flood table.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile,
+	     fid_offset_flood_table_size, 0x34, 0, 16);
+
+/* cmd_mbox_config_profile_max_fid_flood_tables
+ * Maximum number of per-FID flooding tables.
+ *
+ * Note: This flooding tables cover special FIDs only (vFIDs), starting at
+ * FID value 4K and higher.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_fid_flood_tables, 0x38, 24, 4);
+
+/* cmd_mbox_config_profile_fid_flood_table_size
+ * The size (number of entries) of each per-FID table.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, fid_flood_table_size, 0x38, 0, 16);
+
+/* cmd_mbox_config_profile_max_ib_mc
+ * Maximum number of multicast FDB records for InfiniBand
+ * FDB (in 512 chunks) per InfiniBand switch partition.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_ib_mc, 0x40, 0, 15);
+
+/* cmd_mbox_config_profile_max_pkey
+ * Maximum per port PKEY table size (for PKEY enforcement)
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, max_pkey, 0x44, 0, 15);
+
+/* cmd_mbox_config_profile_ar_sec
+ * Primary/secondary capability
+ * Describes the number of adaptive routing sub-groups
+ * 0 - disable primary/secondary (single group)
+ * 1 - enable primary/secondary (2 sub-groups)
+ * 2 - 3 sub-groups: Not supported in SwitchX, SwitchX-2
+ * 3 - 4 sub-groups: Not supported in SwitchX, SwitchX-2
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, ar_sec, 0x4C, 24, 2);
+
+/* cmd_mbox_config_profile_adaptive_routing_group_cap
+ * Adaptive Routing Group Capability. Indicates the number of AR groups
+ * supported. Note that when Primary/secondary is enabled, each
+ * primary/secondary couple consumes 2 adaptive routing entries.
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, adaptive_routing_group_cap, 0x4C, 0, 16);
+
+/* cmd_mbox_config_profile_arn
+ * Adaptive Routing Notification Enable
+ * Not supported in SwitchX, SwitchX-2
+ */
+MLXSW_ITEM32(cmd_mbox, config_profile, arn, 0x50, 31, 1);
+
+/* cmd_mbox_config_profile_swid_config_mask
+ * Modify Switch Partition Configuration mask. When set, the configu-
+ * ration value for the Switch Partition are taken from the mailbox.
+ * When clear, the current configuration values are used.
+ * Bit 0 - set type
+ * Bit 1 - properties
+ * Other - reserved
+ */
+MLXSW_ITEM32_INDEXED(cmd_mbox, config_profile, swid_config_mask,
+		     0x60, 24, 8, 0x08, 0x00, false);
+
+/* cmd_mbox_config_profile_swid_config_type
+ * Switch Partition type.
+ * 0000 - disabled (Switch Partition does not exist)
+ * 0001 - InfiniBand
+ * 0010 - Ethernet
+ * 1000 - router port (SwitchX-2 only)
+ * Other - reserved
+ */
+MLXSW_ITEM32_INDEXED(cmd_mbox, config_profile, swid_config_type,
+		     0x60, 20, 4, 0x08, 0x00, false);
+
+/* cmd_mbox_config_profile_swid_config_properties
+ * Switch Partition properties.
+ */
+MLXSW_ITEM32_INDEXED(cmd_mbox, config_profile, swid_config_properties,
+		     0x60, 0, 8, 0x08, 0x00, false);
+
+/* ACCESS_REG - Access EMAD Supported Register
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == 0 (N/A)
+ * -------------------------------------
+ * The ACCESS_REG command supports accessing device registers. This access
+ * is mainly used for bootstrapping.
+ */
+
+static inline int mlxsw_cmd_access_reg(struct mlxsw_core *mlxsw_core,
+				       char *in_mbox, char *out_mbox)
+{
+	return mlxsw_cmd_exec(mlxsw_core, MLXSW_CMD_OPCODE_ACCESS_REG,
+			      0, 0, false, in_mbox, MLXSW_CMD_MBOX_SIZE,
+			      out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* SW2HW_DQ - Software to Hardware DQ
+ * ----------------------------------
+ * OpMod == 0 (send DQ) / OpMod == 1 (receive DQ)
+ * INMmod == DQ number
+ * ----------------------------------------------
+ * The SW2HW_DQ command transitions a descriptor queue from software to
+ * hardware ownership. The command enables posting WQEs and ringing DoorBells
+ * on the descriptor queue.
+ */
+
+static inline int __mlxsw_cmd_sw2hw_dq(struct mlxsw_core *mlxsw_core,
+				       char *in_mbox, u32 dq_number,
+				       u8 opcode_mod)
+{
+	return mlxsw_cmd_exec_in(mlxsw_core, MLXSW_CMD_OPCODE_SW2HW_DQ,
+				 opcode_mod, dq_number,
+				 in_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+enum {
+	MLXSW_CMD_OPCODE_MOD_SDQ = 0,
+	MLXSW_CMD_OPCODE_MOD_RDQ = 1,
+};
+
+static inline int mlxsw_cmd_sw2hw_sdq(struct mlxsw_core *mlxsw_core,
+				      char *in_mbox, u32 dq_number)
+{
+	return __mlxsw_cmd_sw2hw_dq(mlxsw_core, in_mbox, dq_number,
+				    MLXSW_CMD_OPCODE_MOD_SDQ);
+}
+
+static inline int mlxsw_cmd_sw2hw_rdq(struct mlxsw_core *mlxsw_core,
+				      char *in_mbox, u32 dq_number)
+{
+	return __mlxsw_cmd_sw2hw_dq(mlxsw_core, in_mbox, dq_number,
+				    MLXSW_CMD_OPCODE_MOD_RDQ);
+}
+
+/* cmd_mbox_sw2hw_dq_cq
+ * Number of the CQ that this Descriptor Queue reports completions to.
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_dq, cq, 0x00, 24, 8);
+
+/* cmd_mbox_sw2hw_dq_sdq_tclass
+ * SDQ: CPU Egress TClass
+ * RDQ: Reserved
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_dq, sdq_tclass, 0x00, 16, 6);
+
+/* cmd_mbox_sw2hw_dq_log2_dq_sz
+ * Log (base 2) of the Descriptor Queue size in 4KB pages.
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_dq, log2_dq_sz, 0x00, 0, 6);
+
+/* cmd_mbox_sw2hw_dq_pa
+ * Physical Address.
+ */
+MLXSW_ITEM64_INDEXED(cmd_mbox, sw2hw_dq, pa, 0x10, 12, 52, 0x08, 0x00, true);
+
+/* HW2SW_DQ - Hardware to Software DQ
+ * ----------------------------------
+ * OpMod == 0 (send DQ) / OpMod == 1 (receive DQ)
+ * INMmod == DQ number
+ * ----------------------------------------------
+ * The HW2SW_DQ command transitions a descriptor queue from hardware to
+ * software ownership. Incoming packets on the DQ are silently discarded,
+ * SW should not post descriptors on nonoperational DQs.
+ */
+
+static inline int __mlxsw_cmd_hw2sw_dq(struct mlxsw_core *mlxsw_core,
+				       u32 dq_number, u8 opcode_mod)
+{
+	return mlxsw_cmd_exec_none(mlxsw_core, MLXSW_CMD_OPCODE_HW2SW_DQ,
+				   opcode_mod, dq_number);
+}
+
+static inline int mlxsw_cmd_hw2sw_sdq(struct mlxsw_core *mlxsw_core,
+				      u32 dq_number)
+{
+	return __mlxsw_cmd_hw2sw_dq(mlxsw_core, dq_number,
+				    MLXSW_CMD_OPCODE_MOD_SDQ);
+}
+
+static inline int mlxsw_cmd_hw2sw_rdq(struct mlxsw_core *mlxsw_core,
+				      u32 dq_number)
+{
+	return __mlxsw_cmd_hw2sw_dq(mlxsw_core, dq_number,
+				    MLXSW_CMD_OPCODE_MOD_RDQ);
+}
+
+/* 2ERR_DQ - To Error DQ
+ * ---------------------
+ * OpMod == 0 (send DQ) / OpMod == 1 (receive DQ)
+ * INMmod == DQ number
+ * ----------------------------------------------
+ * The 2ERR_DQ command transitions the DQ into the error state from the state
+ * in which it has been. While the command is executed, some in-process
+ * descriptors may complete. Once the DQ transitions into the error state,
+ * if there are posted descriptors on the RDQ/SDQ, the hardware writes
+ * a completion with error (flushed) for all descriptors posted in the RDQ/SDQ.
+ * When the command is completed successfully, the DQ is already in
+ * the error state.
+ */
+
+static inline int __mlxsw_cmd_2err_dq(struct mlxsw_core *mlxsw_core,
+				      u32 dq_number, u8 opcode_mod)
+{
+	return mlxsw_cmd_exec_none(mlxsw_core, MLXSW_CMD_OPCODE_2ERR_DQ,
+				   opcode_mod, dq_number);
+}
+
+static inline int mlxsw_cmd_2err_sdq(struct mlxsw_core *mlxsw_core,
+				     u32 dq_number)
+{
+	return __mlxsw_cmd_2err_dq(mlxsw_core, dq_number,
+				   MLXSW_CMD_OPCODE_MOD_SDQ);
+}
+
+static inline int mlxsw_cmd_2err_rdq(struct mlxsw_core *mlxsw_core,
+				     u32 dq_number)
+{
+	return __mlxsw_cmd_2err_dq(mlxsw_core, dq_number,
+				   MLXSW_CMD_OPCODE_MOD_RDQ);
+}
+
+/* QUERY_DQ - Query DQ
+ * ---------------------
+ * OpMod == 0 (send DQ) / OpMod == 1 (receive DQ)
+ * INMmod == DQ number
+ * ----------------------------------------------
+ * The QUERY_DQ command retrieves a snapshot of DQ parameters from the hardware.
+ *
+ * Note: Output mailbox has the same format as SW2HW_DQ.
+ */
+
+static inline int __mlxsw_cmd_query_dq(struct mlxsw_core *mlxsw_core,
+				       char *out_mbox, u32 dq_number,
+				       u8 opcode_mod)
+{
+	return mlxsw_cmd_exec_out(mlxsw_core, MLXSW_CMD_OPCODE_2ERR_DQ,
+				  opcode_mod, dq_number, false,
+				  out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+static inline int mlxsw_cmd_query_sdq(struct mlxsw_core *mlxsw_core,
+				      char *out_mbox, u32 dq_number)
+{
+	return __mlxsw_cmd_query_dq(mlxsw_core, out_mbox, dq_number,
+				    MLXSW_CMD_OPCODE_MOD_SDQ);
+}
+
+static inline int mlxsw_cmd_query_rdq(struct mlxsw_core *mlxsw_core,
+				      char *out_mbox, u32 dq_number)
+{
+	return __mlxsw_cmd_query_dq(mlxsw_core, out_mbox, dq_number,
+				    MLXSW_CMD_OPCODE_MOD_RDQ);
+}
+
+/* SW2HW_CQ - Software to Hardware CQ
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == CQ number
+ * -------------------------------------
+ * The SW2HW_CQ command transfers ownership of a CQ context entry from software
+ * to hardware. The command takes the CQ context entry from the input mailbox
+ * and stores it in the CQC in the ownership of the hardware. The command fails
+ * if the requested CQC entry is already in the ownership of the hardware.
+ */
+
+static inline int mlxsw_cmd_sw2hw_cq(struct mlxsw_core *mlxsw_core,
+				     char *in_mbox, u32 cq_number)
+{
+	return mlxsw_cmd_exec_in(mlxsw_core, MLXSW_CMD_OPCODE_SW2HW_CQ,
+				 0, cq_number, in_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_sw2hw_cq_cv
+ * CQE Version.
+ * 0 - CQE Version 0, 1 - CQE Version 1
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_cq, cv, 0x00, 28, 4);
+
+/* cmd_mbox_sw2hw_cq_c_eqn
+ * Event Queue this CQ reports completion events to.
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_cq, c_eqn, 0x00, 24, 1);
+
+/* cmd_mbox_sw2hw_cq_oi
+ * When set, overrun ignore is enabled. When set, updates of
+ * CQ consumer counter (poll for completion) or Request completion
+ * notifications (Arm CQ) DoorBells should not be rung on that CQ.
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_cq, oi, 0x00, 12, 1);
+
+/* cmd_mbox_sw2hw_cq_st
+ * Event delivery state machine
+ * 0x0 - FIRED
+ * 0x1 - ARMED (Request for Notification)
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_cq, st, 0x00, 8, 1);
+
+/* cmd_mbox_sw2hw_cq_log_cq_size
+ * Log (base 2) of the CQ size (in entries).
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_cq, log_cq_size, 0x00, 0, 4);
+
+/* cmd_mbox_sw2hw_cq_producer_counter
+ * Producer Counter. The counter is incremented for each CQE that is
+ * written by the HW to the CQ.
+ * Maintained by HW (valid for the QUERY_CQ command only)
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_cq, producer_counter, 0x04, 0, 16);
+
+/* cmd_mbox_sw2hw_cq_pa
+ * Physical Address.
+ */
+MLXSW_ITEM64_INDEXED(cmd_mbox, sw2hw_cq, pa, 0x10, 11, 53, 0x08, 0x00, true);
+
+/* HW2SW_CQ - Hardware to Software CQ
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == CQ number
+ * -------------------------------------
+ * The HW2SW_CQ command transfers ownership of a CQ context entry from hardware
+ * to software. The CQC entry is invalidated as a result of this command.
+ */
+
+static inline int mlxsw_cmd_hw2sw_cq(struct mlxsw_core *mlxsw_core,
+				     u32 cq_number)
+{
+	return mlxsw_cmd_exec_none(mlxsw_core, MLXSW_CMD_OPCODE_HW2SW_CQ,
+				   0, cq_number);
+}
+
+/* QUERY_CQ - Query CQ
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == CQ number
+ * -------------------------------------
+ * The QUERY_CQ command retrieves a snapshot of the current CQ context entry.
+ * The command stores the snapshot in the output mailbox in the software format.
+ * Note that the CQ context state and values are not affected by the QUERY_CQ
+ * command. The QUERY_CQ command is for debug purposes only.
+ *
+ * Note: Output mailbox has the same format as SW2HW_CQ.
+ */
+
+static inline int mlxsw_cmd_query_cq(struct mlxsw_core *mlxsw_core,
+				     char *out_mbox, u32 cq_number)
+{
+	return mlxsw_cmd_exec_out(mlxsw_core, MLXSW_CMD_OPCODE_QUERY_CQ,
+				  0, cq_number, false,
+				  out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* SW2HW_EQ - Software to Hardware EQ
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == EQ number
+ * -------------------------------------
+ * The SW2HW_EQ command transfers ownership of an EQ context entry from software
+ * to hardware. The command takes the EQ context entry from the input mailbox
+ * and stores it in the EQC in the ownership of the hardware. The command fails
+ * if the requested EQC entry is already in the ownership of the hardware.
+ */
+
+static inline int mlxsw_cmd_sw2hw_eq(struct mlxsw_core *mlxsw_core,
+				     char *in_mbox, u32 eq_number)
+{
+	return mlxsw_cmd_exec_in(mlxsw_core, MLXSW_CMD_OPCODE_SW2HW_EQ,
+				 0, eq_number, in_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+/* cmd_mbox_sw2hw_eq_int_msix
+ * When set, MSI-X cycles will be generated by this EQ.
+ * When cleared, an interrupt will be generated by this EQ.
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_eq, int_msix, 0x00, 24, 1);
+
+/* cmd_mbox_sw2hw_eq_int_oi
+ * When set, overrun ignore is enabled.
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_eq, oi, 0x00, 12, 1);
+
+/* cmd_mbox_sw2hw_eq_int_st
+ * Event delivery state machine
+ * 0x0 - FIRED
+ * 0x1 - ARMED (Request for Notification)
+ * 0x11 - Always ARMED
+ * other - reserved
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_eq, st, 0x00, 8, 2);
+
+/* cmd_mbox_sw2hw_eq_int_log_eq_size
+ * Log (base 2) of the EQ size (in entries).
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_eq, log_eq_size, 0x00, 0, 4);
+
+/* cmd_mbox_sw2hw_eq_int_producer_counter
+ * Producer Counter. The counter is incremented for each EQE that is written
+ * by the HW to the EQ.
+ * Maintained by HW (valid for the QUERY_EQ command only)
+ */
+MLXSW_ITEM32(cmd_mbox, sw2hw_eq, producer_counter, 0x04, 0, 16);
+
+/* cmd_mbox_sw2hw_eq_int_pa
+ * Physical Address.
+ */
+MLXSW_ITEM64_INDEXED(cmd_mbox, sw2hw_eq, pa, 0x10, 11, 53, 0x08, 0x00, true);
+
+/* HW2SW_EQ - Hardware to Software EQ
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == EQ number
+ * -------------------------------------
+ */
+
+static inline int mlxsw_cmd_hw2sw_eq(struct mlxsw_core *mlxsw_core,
+				     u32 eq_number)
+{
+	return mlxsw_cmd_exec_none(mlxsw_core, MLXSW_CMD_OPCODE_HW2SW_EQ,
+				   0, eq_number);
+}
+
+/* QUERY_EQ - Query EQ
+ * ----------------------------------
+ * OpMod == 0 (N/A), INMmod == EQ number
+ * -------------------------------------
+ *
+ * Note: Output mailbox has the same format as SW2HW_EQ.
+ */
+
+static inline int mlxsw_cmd_query_eq(struct mlxsw_core *mlxsw_core,
+				     char *out_mbox, u32 eq_number)
+{
+	return mlxsw_cmd_exec_out(mlxsw_core, MLXSW_CMD_OPCODE_QUERY_EQ,
+				  0, eq_number, false,
+				  out_mbox, MLXSW_CMD_MBOX_SIZE);
+}
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/core.c b/kernel/drivers/net/ethernet/mellanox/mlxsw/core.c
new file mode 100644
index 000000000..97f0d93ca
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/core.c
@@ -0,0 +1,1299 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/core.c
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/export.h>
+#include <linux/err.h>
+#include <linux/if_link.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include <linux/u64_stats_sync.h>
+#include <linux/netdevice.h>
+#include <linux/wait.h>
+#include <linux/skbuff.h>
+#include <linux/etherdevice.h>
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/gfp.h>
+#include <linux/random.h>
+#include <linux/jiffies.h>
+#include <linux/mutex.h>
+#include <linux/rcupdate.h>
+#include <linux/slab.h>
+#include <asm/byteorder.h>
+
+#include "core.h"
+#include "item.h"
+#include "cmd.h"
+#include "port.h"
+#include "trap.h"
+#include "emad.h"
+#include "reg.h"
+
+static LIST_HEAD(mlxsw_core_driver_list);
+static DEFINE_SPINLOCK(mlxsw_core_driver_list_lock);
+
+static const char mlxsw_core_driver_name[] = "mlxsw_core";
+
+static struct dentry *mlxsw_core_dbg_root;
+
+struct mlxsw_core_pcpu_stats {
+	u64			trap_rx_packets[MLXSW_TRAP_ID_MAX];
+	u64			trap_rx_bytes[MLXSW_TRAP_ID_MAX];
+	u64			port_rx_packets[MLXSW_PORT_MAX_PORTS];
+	u64			port_rx_bytes[MLXSW_PORT_MAX_PORTS];
+	struct u64_stats_sync	syncp;
+	u32			trap_rx_dropped[MLXSW_TRAP_ID_MAX];
+	u32			port_rx_dropped[MLXSW_PORT_MAX_PORTS];
+	u32			trap_rx_invalid;
+	u32			port_rx_invalid;
+};
+
+struct mlxsw_core {
+	struct mlxsw_driver *driver;
+	const struct mlxsw_bus *bus;
+	void *bus_priv;
+	const struct mlxsw_bus_info *bus_info;
+	struct list_head rx_listener_list;
+	struct list_head event_listener_list;
+	struct {
+		struct sk_buff *resp_skb;
+		u64 tid;
+		wait_queue_head_t wait;
+		bool trans_active;
+		struct mutex lock; /* One EMAD transaction at a time. */
+		bool use_emad;
+	} emad;
+	struct mlxsw_core_pcpu_stats __percpu *pcpu_stats;
+	struct dentry *dbg_dir;
+	struct {
+		struct debugfs_blob_wrapper vsd_blob;
+		struct debugfs_blob_wrapper psid_blob;
+	} dbg;
+	unsigned long driver_priv[0];
+	/* driver_priv has to be always the last item */
+};
+
+struct mlxsw_rx_listener_item {
+	struct list_head list;
+	struct mlxsw_rx_listener rxl;
+	void *priv;
+};
+
+struct mlxsw_event_listener_item {
+	struct list_head list;
+	struct mlxsw_event_listener el;
+	void *priv;
+};
+
+/******************
+ * EMAD processing
+ ******************/
+
+/* emad_eth_hdr_dmac
+ * Destination MAC in EMAD's Ethernet header.
+ * Must be set to 01:02:c9:00:00:01
+ */
+MLXSW_ITEM_BUF(emad, eth_hdr, dmac, 0x00, 6);
+
+/* emad_eth_hdr_smac
+ * Source MAC in EMAD's Ethernet header.
+ * Must be set to 00:02:c9:01:02:03
+ */
+MLXSW_ITEM_BUF(emad, eth_hdr, smac, 0x06, 6);
+
+/* emad_eth_hdr_ethertype
+ * Ethertype in EMAD's Ethernet header.
+ * Must be set to 0x8932
+ */
+MLXSW_ITEM32(emad, eth_hdr, ethertype, 0x0C, 16, 16);
+
+/* emad_eth_hdr_mlx_proto
+ * Mellanox protocol.
+ * Must be set to 0x0.
+ */
+MLXSW_ITEM32(emad, eth_hdr, mlx_proto, 0x0C, 8, 8);
+
+/* emad_eth_hdr_ver
+ * Mellanox protocol version.
+ * Must be set to 0x0.
+ */
+MLXSW_ITEM32(emad, eth_hdr, ver, 0x0C, 4, 4);
+
+/* emad_op_tlv_type
+ * Type of the TLV.
+ * Must be set to 0x1 (operation TLV).
+ */
+MLXSW_ITEM32(emad, op_tlv, type, 0x00, 27, 5);
+
+/* emad_op_tlv_len
+ * Length of the operation TLV in u32.
+ * Must be set to 0x4.
+ */
+MLXSW_ITEM32(emad, op_tlv, len, 0x00, 16, 11);
+
+/* emad_op_tlv_dr
+ * Direct route bit. Setting to 1 indicates the EMAD is a direct route
+ * EMAD. DR TLV must follow.
+ *
+ * Note: Currently not supported and must not be set.
+ */
+MLXSW_ITEM32(emad, op_tlv, dr, 0x00, 15, 1);
+
+/* emad_op_tlv_status
+ * Returned status in case of EMAD response. Must be set to 0 in case
+ * of EMAD request.
+ * 0x0 - success
+ * 0x1 - device is busy. Requester should retry
+ * 0x2 - Mellanox protocol version not supported
+ * 0x3 - unknown TLV
+ * 0x4 - register not supported
+ * 0x5 - operation class not supported
+ * 0x6 - EMAD method not supported
+ * 0x7 - bad parameter (e.g. port out of range)
+ * 0x8 - resource not available
+ * 0x9 - message receipt acknowledgment. Requester should retry
+ * 0x70 - internal error
+ */
+MLXSW_ITEM32(emad, op_tlv, status, 0x00, 8, 7);
+
+/* emad_op_tlv_register_id
+ * Register ID of register within register TLV.
+ */
+MLXSW_ITEM32(emad, op_tlv, register_id, 0x04, 16, 16);
+
+/* emad_op_tlv_r
+ * Response bit. Setting to 1 indicates Response, otherwise request.
+ */
+MLXSW_ITEM32(emad, op_tlv, r, 0x04, 15, 1);
+
+/* emad_op_tlv_method
+ * EMAD method type.
+ * 0x1 - query
+ * 0x2 - write
+ * 0x3 - send (currently not supported)
+ * 0x4 - event
+ */
+MLXSW_ITEM32(emad, op_tlv, method, 0x04, 8, 7);
+
+/* emad_op_tlv_class
+ * EMAD operation class. Must be set to 0x1 (REG_ACCESS).
+ */
+MLXSW_ITEM32(emad, op_tlv, class, 0x04, 0, 8);
+
+/* emad_op_tlv_tid
+ * EMAD transaction ID. Used for pairing request and response EMADs.
+ */
+MLXSW_ITEM64(emad, op_tlv, tid, 0x08, 0, 64);
+
+/* emad_reg_tlv_type
+ * Type of the TLV.
+ * Must be set to 0x3 (register TLV).
+ */
+MLXSW_ITEM32(emad, reg_tlv, type, 0x00, 27, 5);
+
+/* emad_reg_tlv_len
+ * Length of the operation TLV in u32.
+ */
+MLXSW_ITEM32(emad, reg_tlv, len, 0x00, 16, 11);
+
+/* emad_end_tlv_type
+ * Type of the TLV.
+ * Must be set to 0x0 (end TLV).
+ */
+MLXSW_ITEM32(emad, end_tlv, type, 0x00, 27, 5);
+
+/* emad_end_tlv_len
+ * Length of the end TLV in u32.
+ * Must be set to 1.
+ */
+MLXSW_ITEM32(emad, end_tlv, len, 0x00, 16, 11);
+
+enum mlxsw_core_reg_access_type {
+	MLXSW_CORE_REG_ACCESS_TYPE_QUERY,
+	MLXSW_CORE_REG_ACCESS_TYPE_WRITE,
+};
+
+static inline const char *
+mlxsw_core_reg_access_type_str(enum mlxsw_core_reg_access_type type)
+{
+	switch (type) {
+	case MLXSW_CORE_REG_ACCESS_TYPE_QUERY:
+		return "query";
+	case MLXSW_CORE_REG_ACCESS_TYPE_WRITE:
+		return "write";
+	}
+	BUG();
+}
+
+static void mlxsw_emad_pack_end_tlv(char *end_tlv)
+{
+	mlxsw_emad_end_tlv_type_set(end_tlv, MLXSW_EMAD_TLV_TYPE_END);
+	mlxsw_emad_end_tlv_len_set(end_tlv, MLXSW_EMAD_END_TLV_LEN);
+}
+
+static void mlxsw_emad_pack_reg_tlv(char *reg_tlv,
+				    const struct mlxsw_reg_info *reg,
+				    char *payload)
+{
+	mlxsw_emad_reg_tlv_type_set(reg_tlv, MLXSW_EMAD_TLV_TYPE_REG);
+	mlxsw_emad_reg_tlv_len_set(reg_tlv, reg->len / sizeof(u32) + 1);
+	memcpy(reg_tlv + sizeof(u32), payload, reg->len);
+}
+
+static void mlxsw_emad_pack_op_tlv(char *op_tlv,
+				   const struct mlxsw_reg_info *reg,
+				   enum mlxsw_core_reg_access_type type,
+				   struct mlxsw_core *mlxsw_core)
+{
+	mlxsw_emad_op_tlv_type_set(op_tlv, MLXSW_EMAD_TLV_TYPE_OP);
+	mlxsw_emad_op_tlv_len_set(op_tlv, MLXSW_EMAD_OP_TLV_LEN);
+	mlxsw_emad_op_tlv_dr_set(op_tlv, 0);
+	mlxsw_emad_op_tlv_status_set(op_tlv, 0);
+	mlxsw_emad_op_tlv_register_id_set(op_tlv, reg->id);
+	mlxsw_emad_op_tlv_r_set(op_tlv, MLXSW_EMAD_OP_TLV_REQUEST);
+	if (type == MLXSW_CORE_REG_ACCESS_TYPE_QUERY)
+		mlxsw_emad_op_tlv_method_set(op_tlv,
+					     MLXSW_EMAD_OP_TLV_METHOD_QUERY);
+	else
+		mlxsw_emad_op_tlv_method_set(op_tlv,
+					     MLXSW_EMAD_OP_TLV_METHOD_WRITE);
+	mlxsw_emad_op_tlv_class_set(op_tlv,
+				    MLXSW_EMAD_OP_TLV_CLASS_REG_ACCESS);
+	mlxsw_emad_op_tlv_tid_set(op_tlv, mlxsw_core->emad.tid);
+}
+
+static int mlxsw_emad_construct_eth_hdr(struct sk_buff *skb)
+{
+	char *eth_hdr = skb_push(skb, MLXSW_EMAD_ETH_HDR_LEN);
+
+	mlxsw_emad_eth_hdr_dmac_memcpy_to(eth_hdr, MLXSW_EMAD_EH_DMAC);
+	mlxsw_emad_eth_hdr_smac_memcpy_to(eth_hdr, MLXSW_EMAD_EH_SMAC);
+	mlxsw_emad_eth_hdr_ethertype_set(eth_hdr, MLXSW_EMAD_EH_ETHERTYPE);
+	mlxsw_emad_eth_hdr_mlx_proto_set(eth_hdr, MLXSW_EMAD_EH_MLX_PROTO);
+	mlxsw_emad_eth_hdr_ver_set(eth_hdr, MLXSW_EMAD_EH_PROTO_VERSION);
+
+	skb_reset_mac_header(skb);
+
+	return 0;
+}
+
+static void mlxsw_emad_construct(struct sk_buff *skb,
+				 const struct mlxsw_reg_info *reg,
+				 char *payload,
+				 enum mlxsw_core_reg_access_type type,
+				 struct mlxsw_core *mlxsw_core)
+{
+	char *buf;
+
+	buf = skb_push(skb, MLXSW_EMAD_END_TLV_LEN * sizeof(u32));
+	mlxsw_emad_pack_end_tlv(buf);
+
+	buf = skb_push(skb, reg->len + sizeof(u32));
+	mlxsw_emad_pack_reg_tlv(buf, reg, payload);
+
+	buf = skb_push(skb, MLXSW_EMAD_OP_TLV_LEN * sizeof(u32));
+	mlxsw_emad_pack_op_tlv(buf, reg, type, mlxsw_core);
+
+	mlxsw_emad_construct_eth_hdr(skb);
+}
+
+static char *mlxsw_emad_op_tlv(const struct sk_buff *skb)
+{
+	return ((char *) (skb->data + MLXSW_EMAD_ETH_HDR_LEN));
+}
+
+static char *mlxsw_emad_reg_tlv(const struct sk_buff *skb)
+{
+	return ((char *) (skb->data + MLXSW_EMAD_ETH_HDR_LEN +
+				      MLXSW_EMAD_OP_TLV_LEN * sizeof(u32)));
+}
+
+static char *mlxsw_emad_reg_payload(const char *op_tlv)
+{
+	return ((char *) (op_tlv + (MLXSW_EMAD_OP_TLV_LEN + 1) * sizeof(u32)));
+}
+
+static u64 mlxsw_emad_get_tid(const struct sk_buff *skb)
+{
+	char *op_tlv;
+
+	op_tlv = mlxsw_emad_op_tlv(skb);
+	return mlxsw_emad_op_tlv_tid_get(op_tlv);
+}
+
+static bool mlxsw_emad_is_resp(const struct sk_buff *skb)
+{
+	char *op_tlv;
+
+	op_tlv = mlxsw_emad_op_tlv(skb);
+	return (mlxsw_emad_op_tlv_r_get(op_tlv) == MLXSW_EMAD_OP_TLV_RESPONSE);
+}
+
+#define MLXSW_EMAD_TIMEOUT_MS 200
+
+static int __mlxsw_emad_transmit(struct mlxsw_core *mlxsw_core,
+				 struct sk_buff *skb,
+				 const struct mlxsw_tx_info *tx_info)
+{
+	int err;
+	int ret;
+
+	mlxsw_core->emad.trans_active = true;
+
+	err = mlxsw_core_skb_transmit(mlxsw_core->driver_priv, skb, tx_info);
+	if (err) {
+		dev_err(mlxsw_core->bus_info->dev, "Failed to transmit EMAD (tid=%llx)\n",
+			mlxsw_core->emad.tid);
+		dev_kfree_skb(skb);
+		goto trans_inactive_out;
+	}
+
+	ret = wait_event_timeout(mlxsw_core->emad.wait,
+				 !(mlxsw_core->emad.trans_active),
+				 msecs_to_jiffies(MLXSW_EMAD_TIMEOUT_MS));
+	if (!ret) {
+		dev_warn(mlxsw_core->bus_info->dev, "EMAD timed-out (tid=%llx)\n",
+			 mlxsw_core->emad.tid);
+		err = -EIO;
+		goto trans_inactive_out;
+	}
+
+	return 0;
+
+trans_inactive_out:
+	mlxsw_core->emad.trans_active = false;
+	return err;
+}
+
+static int mlxsw_emad_process_status(struct mlxsw_core *mlxsw_core,
+				     char *op_tlv)
+{
+	enum mlxsw_emad_op_tlv_status status;
+	u64 tid;
+
+	status = mlxsw_emad_op_tlv_status_get(op_tlv);
+	tid = mlxsw_emad_op_tlv_tid_get(op_tlv);
+
+	switch (status) {
+	case MLXSW_EMAD_OP_TLV_STATUS_SUCCESS:
+		return 0;
+	case MLXSW_EMAD_OP_TLV_STATUS_BUSY:
+	case MLXSW_EMAD_OP_TLV_STATUS_MESSAGE_RECEIPT_ACK:
+		dev_warn(mlxsw_core->bus_info->dev, "Reg access status again (tid=%llx,status=%x(%s))\n",
+			 tid, status, mlxsw_emad_op_tlv_status_str(status));
+		return -EAGAIN;
+	case MLXSW_EMAD_OP_TLV_STATUS_VERSION_NOT_SUPPORTED:
+	case MLXSW_EMAD_OP_TLV_STATUS_UNKNOWN_TLV:
+	case MLXSW_EMAD_OP_TLV_STATUS_REGISTER_NOT_SUPPORTED:
+	case MLXSW_EMAD_OP_TLV_STATUS_CLASS_NOT_SUPPORTED:
+	case MLXSW_EMAD_OP_TLV_STATUS_METHOD_NOT_SUPPORTED:
+	case MLXSW_EMAD_OP_TLV_STATUS_BAD_PARAMETER:
+	case MLXSW_EMAD_OP_TLV_STATUS_RESOURCE_NOT_AVAILABLE:
+	case MLXSW_EMAD_OP_TLV_STATUS_INTERNAL_ERROR:
+	default:
+		dev_err(mlxsw_core->bus_info->dev, "Reg access status failed (tid=%llx,status=%x(%s))\n",
+			tid, status, mlxsw_emad_op_tlv_status_str(status));
+		return -EIO;
+	}
+}
+
+static int mlxsw_emad_process_status_skb(struct mlxsw_core *mlxsw_core,
+					 struct sk_buff *skb)
+{
+	return mlxsw_emad_process_status(mlxsw_core, mlxsw_emad_op_tlv(skb));
+}
+
+static int mlxsw_emad_transmit(struct mlxsw_core *mlxsw_core,
+			       struct sk_buff *skb,
+			       const struct mlxsw_tx_info *tx_info)
+{
+	struct sk_buff *trans_skb;
+	int n_retry;
+	int err;
+
+	n_retry = 0;
+retry:
+	/* We copy the EMAD to a new skb, since we might need
+	 * to retransmit it in case of failure.
+	 */
+	trans_skb = skb_copy(skb, GFP_KERNEL);
+	if (!trans_skb) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	err = __mlxsw_emad_transmit(mlxsw_core, trans_skb, tx_info);
+	if (!err) {
+		struct sk_buff *resp_skb = mlxsw_core->emad.resp_skb;
+
+		err = mlxsw_emad_process_status_skb(mlxsw_core, resp_skb);
+		if (err)
+			dev_kfree_skb(resp_skb);
+		if (!err || err != -EAGAIN)
+			goto out;
+	}
+	if (n_retry++ < MLXSW_EMAD_MAX_RETRY)
+		goto retry;
+
+out:
+	dev_kfree_skb(skb);
+	mlxsw_core->emad.tid++;
+	return err;
+}
+
+static void mlxsw_emad_rx_listener_func(struct sk_buff *skb, u8 local_port,
+					void *priv)
+{
+	struct mlxsw_core *mlxsw_core = priv;
+
+	if (mlxsw_emad_is_resp(skb) &&
+	    mlxsw_core->emad.trans_active &&
+	    mlxsw_emad_get_tid(skb) == mlxsw_core->emad.tid) {
+		mlxsw_core->emad.resp_skb = skb;
+		mlxsw_core->emad.trans_active = false;
+		wake_up(&mlxsw_core->emad.wait);
+	} else {
+		dev_kfree_skb(skb);
+	}
+}
+
+static const struct mlxsw_rx_listener mlxsw_emad_rx_listener = {
+	.func = mlxsw_emad_rx_listener_func,
+	.local_port = MLXSW_PORT_DONT_CARE,
+	.trap_id = MLXSW_TRAP_ID_ETHEMAD,
+};
+
+static int mlxsw_emad_traps_set(struct mlxsw_core *mlxsw_core)
+{
+	char htgt_pl[MLXSW_REG_HTGT_LEN];
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int err;
+
+	mlxsw_reg_htgt_pack(htgt_pl, MLXSW_REG_HTGT_TRAP_GROUP_EMAD);
+	err = mlxsw_reg_write(mlxsw_core, MLXSW_REG(htgt), htgt_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_TRAP_TO_CPU,
+			    MLXSW_TRAP_ID_ETHEMAD);
+	return mlxsw_reg_write(mlxsw_core, MLXSW_REG(hpkt), hpkt_pl);
+}
+
+static int mlxsw_emad_init(struct mlxsw_core *mlxsw_core)
+{
+	int err;
+
+	/* Set the upper 32 bits of the transaction ID field to a random
+	 * number. This allows us to discard EMADs addressed to other
+	 * devices.
+	 */
+	get_random_bytes(&mlxsw_core->emad.tid, 4);
+	mlxsw_core->emad.tid = mlxsw_core->emad.tid << 32;
+
+	init_waitqueue_head(&mlxsw_core->emad.wait);
+	mlxsw_core->emad.trans_active = false;
+	mutex_init(&mlxsw_core->emad.lock);
+
+	err = mlxsw_core_rx_listener_register(mlxsw_core,
+					      &mlxsw_emad_rx_listener,
+					      mlxsw_core);
+	if (err)
+		return err;
+
+	err = mlxsw_emad_traps_set(mlxsw_core);
+	if (err)
+		goto err_emad_trap_set;
+
+	mlxsw_core->emad.use_emad = true;
+
+	return 0;
+
+err_emad_trap_set:
+	mlxsw_core_rx_listener_unregister(mlxsw_core,
+					  &mlxsw_emad_rx_listener,
+					  mlxsw_core);
+	return err;
+}
+
+static void mlxsw_emad_fini(struct mlxsw_core *mlxsw_core)
+{
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+
+	mlxsw_core->emad.use_emad = false;
+	mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_DISCARD,
+			    MLXSW_TRAP_ID_ETHEMAD);
+	mlxsw_reg_write(mlxsw_core, MLXSW_REG(hpkt), hpkt_pl);
+
+	mlxsw_core_rx_listener_unregister(mlxsw_core,
+					  &mlxsw_emad_rx_listener,
+					  mlxsw_core);
+}
+
+static struct sk_buff *mlxsw_emad_alloc(const struct mlxsw_core *mlxsw_core,
+					u16 reg_len)
+{
+	struct sk_buff *skb;
+	u16 emad_len;
+
+	emad_len = (reg_len + sizeof(u32) + MLXSW_EMAD_ETH_HDR_LEN +
+		    (MLXSW_EMAD_OP_TLV_LEN + MLXSW_EMAD_END_TLV_LEN) *
+		    sizeof(u32) + mlxsw_core->driver->txhdr_len);
+	if (emad_len > MLXSW_EMAD_MAX_FRAME_LEN)
+		return NULL;
+
+	skb = netdev_alloc_skb(NULL, emad_len);
+	if (!skb)
+		return NULL;
+	memset(skb->data, 0, emad_len);
+	skb_reserve(skb, emad_len);
+
+	return skb;
+}
+
+/*****************
+ * Core functions
+ *****************/
+
+static int mlxsw_core_rx_stats_dbg_read(struct seq_file *file, void *data)
+{
+	struct mlxsw_core *mlxsw_core = file->private;
+	struct mlxsw_core_pcpu_stats *p;
+	u64 rx_packets, rx_bytes;
+	u64 tmp_rx_packets, tmp_rx_bytes;
+	u32 rx_dropped, rx_invalid;
+	unsigned int start;
+	int i;
+	int j;
+	static const char hdr[] =
+		"     NUM   RX_PACKETS     RX_BYTES RX_DROPPED\n";
+
+	seq_printf(file, hdr);
+	for (i = 0; i < MLXSW_TRAP_ID_MAX; i++) {
+		rx_packets = 0;
+		rx_bytes = 0;
+		rx_dropped = 0;
+		for_each_possible_cpu(j) {
+			p = per_cpu_ptr(mlxsw_core->pcpu_stats, j);
+			do {
+				start = u64_stats_fetch_begin(&p->syncp);
+				tmp_rx_packets = p->trap_rx_packets[i];
+				tmp_rx_bytes = p->trap_rx_bytes[i];
+			} while (u64_stats_fetch_retry(&p->syncp, start));
+
+			rx_packets += tmp_rx_packets;
+			rx_bytes += tmp_rx_bytes;
+			rx_dropped += p->trap_rx_dropped[i];
+		}
+		seq_printf(file, "trap %3d %12llu %12llu %10u\n",
+			   i, rx_packets, rx_bytes, rx_dropped);
+	}
+	rx_invalid = 0;
+	for_each_possible_cpu(j) {
+		p = per_cpu_ptr(mlxsw_core->pcpu_stats, j);
+		rx_invalid += p->trap_rx_invalid;
+	}
+	seq_printf(file, "trap INV                           %10u\n",
+		   rx_invalid);
+
+	for (i = 0; i < MLXSW_PORT_MAX_PORTS; i++) {
+		rx_packets = 0;
+		rx_bytes = 0;
+		rx_dropped = 0;
+		for_each_possible_cpu(j) {
+			p = per_cpu_ptr(mlxsw_core->pcpu_stats, j);
+			do {
+				start = u64_stats_fetch_begin(&p->syncp);
+				tmp_rx_packets = p->port_rx_packets[i];
+				tmp_rx_bytes = p->port_rx_bytes[i];
+			} while (u64_stats_fetch_retry(&p->syncp, start));
+
+			rx_packets += tmp_rx_packets;
+			rx_bytes += tmp_rx_bytes;
+			rx_dropped += p->port_rx_dropped[i];
+		}
+		seq_printf(file, "port %3d %12llu %12llu %10u\n",
+			   i, rx_packets, rx_bytes, rx_dropped);
+	}
+	rx_invalid = 0;
+	for_each_possible_cpu(j) {
+		p = per_cpu_ptr(mlxsw_core->pcpu_stats, j);
+		rx_invalid += p->port_rx_invalid;
+	}
+	seq_printf(file, "port INV                           %10u\n",
+		   rx_invalid);
+	return 0;
+}
+
+static int mlxsw_core_rx_stats_dbg_open(struct inode *inode, struct file *f)
+{
+	struct mlxsw_core *mlxsw_core = inode->i_private;
+
+	return single_open(f, mlxsw_core_rx_stats_dbg_read, mlxsw_core);
+}
+
+static const struct file_operations mlxsw_core_rx_stats_dbg_ops = {
+	.owner = THIS_MODULE,
+	.open = mlxsw_core_rx_stats_dbg_open,
+	.release = single_release,
+	.read = seq_read,
+	.llseek = seq_lseek
+};
+
+static void mlxsw_core_buf_dump_dbg(struct mlxsw_core *mlxsw_core,
+				    const char *buf, size_t size)
+{
+	__be32 *m = (__be32 *) buf;
+	int i;
+	int count = size / sizeof(__be32);
+
+	for (i = count - 1; i >= 0; i--)
+		if (m[i])
+			break;
+	i++;
+	count = i ? i : 1;
+	for (i = 0; i < count; i += 4)
+		dev_dbg(mlxsw_core->bus_info->dev, "%04x - %08x %08x %08x %08x\n",
+			i * 4, be32_to_cpu(m[i]), be32_to_cpu(m[i + 1]),
+			be32_to_cpu(m[i + 2]), be32_to_cpu(m[i + 3]));
+}
+
+int mlxsw_core_driver_register(struct mlxsw_driver *mlxsw_driver)
+{
+	spin_lock(&mlxsw_core_driver_list_lock);
+	list_add_tail(&mlxsw_driver->list, &mlxsw_core_driver_list);
+	spin_unlock(&mlxsw_core_driver_list_lock);
+	return 0;
+}
+EXPORT_SYMBOL(mlxsw_core_driver_register);
+
+void mlxsw_core_driver_unregister(struct mlxsw_driver *mlxsw_driver)
+{
+	spin_lock(&mlxsw_core_driver_list_lock);
+	list_del(&mlxsw_driver->list);
+	spin_unlock(&mlxsw_core_driver_list_lock);
+}
+EXPORT_SYMBOL(mlxsw_core_driver_unregister);
+
+static struct mlxsw_driver *__driver_find(const char *kind)
+{
+	struct mlxsw_driver *mlxsw_driver;
+
+	list_for_each_entry(mlxsw_driver, &mlxsw_core_driver_list, list) {
+		if (strcmp(mlxsw_driver->kind, kind) == 0)
+			return mlxsw_driver;
+	}
+	return NULL;
+}
+
+static struct mlxsw_driver *mlxsw_core_driver_get(const char *kind)
+{
+	struct mlxsw_driver *mlxsw_driver;
+
+	spin_lock(&mlxsw_core_driver_list_lock);
+	mlxsw_driver = __driver_find(kind);
+	if (!mlxsw_driver) {
+		spin_unlock(&mlxsw_core_driver_list_lock);
+		request_module(MLXSW_MODULE_ALIAS_PREFIX "%s", kind);
+		spin_lock(&mlxsw_core_driver_list_lock);
+		mlxsw_driver = __driver_find(kind);
+	}
+	if (mlxsw_driver) {
+		if (!try_module_get(mlxsw_driver->owner))
+			mlxsw_driver = NULL;
+	}
+
+	spin_unlock(&mlxsw_core_driver_list_lock);
+	return mlxsw_driver;
+}
+
+static void mlxsw_core_driver_put(const char *kind)
+{
+	struct mlxsw_driver *mlxsw_driver;
+
+	spin_lock(&mlxsw_core_driver_list_lock);
+	mlxsw_driver = __driver_find(kind);
+	spin_unlock(&mlxsw_core_driver_list_lock);
+	if (!mlxsw_driver)
+		return;
+	module_put(mlxsw_driver->owner);
+}
+
+static int mlxsw_core_debugfs_init(struct mlxsw_core *mlxsw_core)
+{
+	const struct mlxsw_bus_info *bus_info = mlxsw_core->bus_info;
+
+	mlxsw_core->dbg_dir = debugfs_create_dir(bus_info->device_name,
+						 mlxsw_core_dbg_root);
+	if (!mlxsw_core->dbg_dir)
+		return -ENOMEM;
+	debugfs_create_file("rx_stats", S_IRUGO, mlxsw_core->dbg_dir,
+			    mlxsw_core, &mlxsw_core_rx_stats_dbg_ops);
+	mlxsw_core->dbg.vsd_blob.data = (void *) &bus_info->vsd;
+	mlxsw_core->dbg.vsd_blob.size = sizeof(bus_info->vsd);
+	debugfs_create_blob("vsd", S_IRUGO, mlxsw_core->dbg_dir,
+			    &mlxsw_core->dbg.vsd_blob);
+	mlxsw_core->dbg.psid_blob.data = (void *) &bus_info->psid;
+	mlxsw_core->dbg.psid_blob.size = sizeof(bus_info->psid);
+	debugfs_create_blob("psid", S_IRUGO, mlxsw_core->dbg_dir,
+			    &mlxsw_core->dbg.psid_blob);
+	return 0;
+}
+
+static void mlxsw_core_debugfs_fini(struct mlxsw_core *mlxsw_core)
+{
+	debugfs_remove_recursive(mlxsw_core->dbg_dir);
+}
+
+int mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
+				   const struct mlxsw_bus *mlxsw_bus,
+				   void *bus_priv)
+{
+	const char *device_kind = mlxsw_bus_info->device_kind;
+	struct mlxsw_core *mlxsw_core;
+	struct mlxsw_driver *mlxsw_driver;
+	size_t alloc_size;
+	int err;
+
+	mlxsw_driver = mlxsw_core_driver_get(device_kind);
+	if (!mlxsw_driver)
+		return -EINVAL;
+	alloc_size = sizeof(*mlxsw_core) + mlxsw_driver->priv_size;
+	mlxsw_core = kzalloc(alloc_size, GFP_KERNEL);
+	if (!mlxsw_core) {
+		err = -ENOMEM;
+		goto err_core_alloc;
+	}
+
+	INIT_LIST_HEAD(&mlxsw_core->rx_listener_list);
+	INIT_LIST_HEAD(&mlxsw_core->event_listener_list);
+	mlxsw_core->driver = mlxsw_driver;
+	mlxsw_core->bus = mlxsw_bus;
+	mlxsw_core->bus_priv = bus_priv;
+	mlxsw_core->bus_info = mlxsw_bus_info;
+
+	mlxsw_core->pcpu_stats =
+		netdev_alloc_pcpu_stats(struct mlxsw_core_pcpu_stats);
+	if (!mlxsw_core->pcpu_stats) {
+		err = -ENOMEM;
+		goto err_alloc_stats;
+	}
+
+	err = mlxsw_bus->init(bus_priv, mlxsw_core, mlxsw_driver->profile);
+	if (err)
+		goto err_bus_init;
+
+	err = mlxsw_emad_init(mlxsw_core);
+	if (err)
+		goto err_emad_init;
+
+	err = mlxsw_driver->init(mlxsw_core->driver_priv, mlxsw_core,
+				 mlxsw_bus_info);
+	if (err)
+		goto err_driver_init;
+
+	err = mlxsw_core_debugfs_init(mlxsw_core);
+	if (err)
+		goto err_debugfs_init;
+
+	return 0;
+
+err_debugfs_init:
+	mlxsw_core->driver->fini(mlxsw_core->driver_priv);
+err_driver_init:
+	mlxsw_emad_fini(mlxsw_core);
+err_emad_init:
+	mlxsw_bus->fini(bus_priv);
+err_bus_init:
+	free_percpu(mlxsw_core->pcpu_stats);
+err_alloc_stats:
+	kfree(mlxsw_core);
+err_core_alloc:
+	mlxsw_core_driver_put(device_kind);
+	return err;
+}
+EXPORT_SYMBOL(mlxsw_core_bus_device_register);
+
+void mlxsw_core_bus_device_unregister(struct mlxsw_core *mlxsw_core)
+{
+	const char *device_kind = mlxsw_core->bus_info->device_kind;
+
+	mlxsw_core_debugfs_fini(mlxsw_core);
+	mlxsw_core->driver->fini(mlxsw_core->driver_priv);
+	mlxsw_emad_fini(mlxsw_core);
+	mlxsw_core->bus->fini(mlxsw_core->bus_priv);
+	free_percpu(mlxsw_core->pcpu_stats);
+	kfree(mlxsw_core);
+	mlxsw_core_driver_put(device_kind);
+}
+EXPORT_SYMBOL(mlxsw_core_bus_device_unregister);
+
+static struct mlxsw_core *__mlxsw_core_get(void *driver_priv)
+{
+	return container_of(driver_priv, struct mlxsw_core, driver_priv);
+}
+
+bool mlxsw_core_skb_transmit_busy(void *driver_priv,
+				  const struct mlxsw_tx_info *tx_info)
+{
+	struct mlxsw_core *mlxsw_core = __mlxsw_core_get(driver_priv);
+
+	return mlxsw_core->bus->skb_transmit_busy(mlxsw_core->bus_priv,
+						  tx_info);
+}
+EXPORT_SYMBOL(mlxsw_core_skb_transmit_busy);
+
+int mlxsw_core_skb_transmit(void *driver_priv, struct sk_buff *skb,
+			    const struct mlxsw_tx_info *tx_info)
+{
+	struct mlxsw_core *mlxsw_core = __mlxsw_core_get(driver_priv);
+
+	return mlxsw_core->bus->skb_transmit(mlxsw_core->bus_priv, skb,
+					     tx_info);
+}
+EXPORT_SYMBOL(mlxsw_core_skb_transmit);
+
+static bool __is_rx_listener_equal(const struct mlxsw_rx_listener *rxl_a,
+				   const struct mlxsw_rx_listener *rxl_b)
+{
+	return (rxl_a->func == rxl_b->func &&
+		rxl_a->local_port == rxl_b->local_port &&
+		rxl_a->trap_id == rxl_b->trap_id);
+}
+
+static struct mlxsw_rx_listener_item *
+__find_rx_listener_item(struct mlxsw_core *mlxsw_core,
+			const struct mlxsw_rx_listener *rxl,
+			void *priv)
+{
+	struct mlxsw_rx_listener_item *rxl_item;
+
+	list_for_each_entry(rxl_item, &mlxsw_core->rx_listener_list, list) {
+		if (__is_rx_listener_equal(&rxl_item->rxl, rxl) &&
+		    rxl_item->priv == priv)
+			return rxl_item;
+	}
+	return NULL;
+}
+
+int mlxsw_core_rx_listener_register(struct mlxsw_core *mlxsw_core,
+				    const struct mlxsw_rx_listener *rxl,
+				    void *priv)
+{
+	struct mlxsw_rx_listener_item *rxl_item;
+
+	rxl_item = __find_rx_listener_item(mlxsw_core, rxl, priv);
+	if (rxl_item)
+		return -EEXIST;
+	rxl_item = kmalloc(sizeof(*rxl_item), GFP_KERNEL);
+	if (!rxl_item)
+		return -ENOMEM;
+	rxl_item->rxl = *rxl;
+	rxl_item->priv = priv;
+
+	list_add_rcu(&rxl_item->list, &mlxsw_core->rx_listener_list);
+	return 0;
+}
+EXPORT_SYMBOL(mlxsw_core_rx_listener_register);
+
+void mlxsw_core_rx_listener_unregister(struct mlxsw_core *mlxsw_core,
+				       const struct mlxsw_rx_listener *rxl,
+				       void *priv)
+{
+	struct mlxsw_rx_listener_item *rxl_item;
+
+	rxl_item = __find_rx_listener_item(mlxsw_core, rxl, priv);
+	if (!rxl_item)
+		return;
+	list_del_rcu(&rxl_item->list);
+	synchronize_rcu();
+	kfree(rxl_item);
+}
+EXPORT_SYMBOL(mlxsw_core_rx_listener_unregister);
+
+static void mlxsw_core_event_listener_func(struct sk_buff *skb, u8 local_port,
+					   void *priv)
+{
+	struct mlxsw_event_listener_item *event_listener_item = priv;
+	struct mlxsw_reg_info reg;
+	char *payload;
+	char *op_tlv = mlxsw_emad_op_tlv(skb);
+	char *reg_tlv = mlxsw_emad_reg_tlv(skb);
+
+	reg.id = mlxsw_emad_op_tlv_register_id_get(op_tlv);
+	reg.len = (mlxsw_emad_reg_tlv_len_get(reg_tlv) - 1) * sizeof(u32);
+	payload = mlxsw_emad_reg_payload(op_tlv);
+	event_listener_item->el.func(&reg, payload, event_listener_item->priv);
+	dev_kfree_skb(skb);
+}
+
+static bool __is_event_listener_equal(const struct mlxsw_event_listener *el_a,
+				      const struct mlxsw_event_listener *el_b)
+{
+	return (el_a->func == el_b->func &&
+		el_a->trap_id == el_b->trap_id);
+}
+
+static struct mlxsw_event_listener_item *
+__find_event_listener_item(struct mlxsw_core *mlxsw_core,
+			   const struct mlxsw_event_listener *el,
+			   void *priv)
+{
+	struct mlxsw_event_listener_item *el_item;
+
+	list_for_each_entry(el_item, &mlxsw_core->event_listener_list, list) {
+		if (__is_event_listener_equal(&el_item->el, el) &&
+		    el_item->priv == priv)
+			return el_item;
+	}
+	return NULL;
+}
+
+int mlxsw_core_event_listener_register(struct mlxsw_core *mlxsw_core,
+				       const struct mlxsw_event_listener *el,
+				       void *priv)
+{
+	int err;
+	struct mlxsw_event_listener_item *el_item;
+	const struct mlxsw_rx_listener rxl = {
+		.func = mlxsw_core_event_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = el->trap_id,
+	};
+
+	el_item = __find_event_listener_item(mlxsw_core, el, priv);
+	if (el_item)
+		return -EEXIST;
+	el_item = kmalloc(sizeof(*el_item), GFP_KERNEL);
+	if (!el_item)
+		return -ENOMEM;
+	el_item->el = *el;
+	el_item->priv = priv;
+
+	err = mlxsw_core_rx_listener_register(mlxsw_core, &rxl, el_item);
+	if (err)
+		goto err_rx_listener_register;
+
+	/* No reason to save item if we did not manage to register an RX
+	 * listener for it.
+	 */
+	list_add_rcu(&el_item->list, &mlxsw_core->event_listener_list);
+
+	return 0;
+
+err_rx_listener_register:
+	kfree(el_item);
+	return err;
+}
+EXPORT_SYMBOL(mlxsw_core_event_listener_register);
+
+void mlxsw_core_event_listener_unregister(struct mlxsw_core *mlxsw_core,
+					  const struct mlxsw_event_listener *el,
+					  void *priv)
+{
+	struct mlxsw_event_listener_item *el_item;
+	const struct mlxsw_rx_listener rxl = {
+		.func = mlxsw_core_event_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = el->trap_id,
+	};
+
+	el_item = __find_event_listener_item(mlxsw_core, el, priv);
+	if (!el_item)
+		return;
+	mlxsw_core_rx_listener_unregister(mlxsw_core, &rxl, el_item);
+	list_del(&el_item->list);
+	kfree(el_item);
+}
+EXPORT_SYMBOL(mlxsw_core_event_listener_unregister);
+
+static int mlxsw_core_reg_access_emad(struct mlxsw_core *mlxsw_core,
+				      const struct mlxsw_reg_info *reg,
+				      char *payload,
+				      enum mlxsw_core_reg_access_type type)
+{
+	int err;
+	char *op_tlv;
+	struct sk_buff *skb;
+	struct mlxsw_tx_info tx_info = {
+		.local_port = MLXSW_PORT_CPU_PORT,
+		.is_emad = true,
+	};
+
+	skb = mlxsw_emad_alloc(mlxsw_core, reg->len);
+	if (!skb)
+		return -ENOMEM;
+
+	mlxsw_emad_construct(skb, reg, payload, type, mlxsw_core);
+	mlxsw_core->driver->txhdr_construct(skb, &tx_info);
+
+	dev_dbg(mlxsw_core->bus_info->dev, "EMAD send (tid=%llx)\n",
+		mlxsw_core->emad.tid);
+	mlxsw_core_buf_dump_dbg(mlxsw_core, skb->data, skb->len);
+
+	err = mlxsw_emad_transmit(mlxsw_core, skb, &tx_info);
+	if (!err) {
+		op_tlv = mlxsw_emad_op_tlv(mlxsw_core->emad.resp_skb);
+		memcpy(payload, mlxsw_emad_reg_payload(op_tlv),
+		       reg->len);
+
+		dev_dbg(mlxsw_core->bus_info->dev, "EMAD recv (tid=%llx)\n",
+			mlxsw_core->emad.tid - 1);
+		mlxsw_core_buf_dump_dbg(mlxsw_core,
+					mlxsw_core->emad.resp_skb->data,
+					mlxsw_core->emad.resp_skb->len);
+
+		dev_kfree_skb(mlxsw_core->emad.resp_skb);
+	}
+
+	return err;
+}
+
+static int mlxsw_core_reg_access_cmd(struct mlxsw_core *mlxsw_core,
+				     const struct mlxsw_reg_info *reg,
+				     char *payload,
+				     enum mlxsw_core_reg_access_type type)
+{
+	int err, n_retry;
+	char *in_mbox, *out_mbox, *tmp;
+
+	in_mbox = mlxsw_cmd_mbox_alloc();
+	if (!in_mbox)
+		return -ENOMEM;
+
+	out_mbox = mlxsw_cmd_mbox_alloc();
+	if (!out_mbox) {
+		err = -ENOMEM;
+		goto free_in_mbox;
+	}
+
+	mlxsw_emad_pack_op_tlv(in_mbox, reg, type, mlxsw_core);
+	tmp = in_mbox + MLXSW_EMAD_OP_TLV_LEN * sizeof(u32);
+	mlxsw_emad_pack_reg_tlv(tmp, reg, payload);
+
+	n_retry = 0;
+retry:
+	err = mlxsw_cmd_access_reg(mlxsw_core, in_mbox, out_mbox);
+	if (!err) {
+		err = mlxsw_emad_process_status(mlxsw_core, out_mbox);
+		if (err == -EAGAIN && n_retry++ < MLXSW_EMAD_MAX_RETRY)
+			goto retry;
+	}
+
+	if (!err)
+		memcpy(payload, mlxsw_emad_reg_payload(out_mbox),
+		       reg->len);
+
+	mlxsw_core->emad.tid++;
+	mlxsw_cmd_mbox_free(out_mbox);
+free_in_mbox:
+	mlxsw_cmd_mbox_free(in_mbox);
+	return err;
+}
+
+static int mlxsw_core_reg_access(struct mlxsw_core *mlxsw_core,
+				 const struct mlxsw_reg_info *reg,
+				 char *payload,
+				 enum mlxsw_core_reg_access_type type)
+{
+	u64 cur_tid;
+	int err;
+
+	if (mutex_lock_interruptible(&mlxsw_core->emad.lock)) {
+		dev_err(mlxsw_core->bus_info->dev, "Reg access interrupted (reg_id=%x(%s),type=%s)\n",
+			reg->id, mlxsw_reg_id_str(reg->id),
+			mlxsw_core_reg_access_type_str(type));
+		return -EINTR;
+	}
+
+	cur_tid = mlxsw_core->emad.tid;
+	dev_dbg(mlxsw_core->bus_info->dev, "Reg access (tid=%llx,reg_id=%x(%s),type=%s)\n",
+		cur_tid, reg->id, mlxsw_reg_id_str(reg->id),
+		mlxsw_core_reg_access_type_str(type));
+
+	/* During initialization EMAD interface is not available to us,
+	 * so we default to command interface. We switch to EMAD interface
+	 * after setting the appropriate traps.
+	 */
+	if (!mlxsw_core->emad.use_emad)
+		err = mlxsw_core_reg_access_cmd(mlxsw_core, reg,
+						payload, type);
+	else
+		err = mlxsw_core_reg_access_emad(mlxsw_core, reg,
+						 payload, type);
+
+	if (err)
+		dev_err(mlxsw_core->bus_info->dev, "Reg access failed (tid=%llx,reg_id=%x(%s),type=%s)\n",
+			cur_tid, reg->id, mlxsw_reg_id_str(reg->id),
+			mlxsw_core_reg_access_type_str(type));
+
+	mutex_unlock(&mlxsw_core->emad.lock);
+	return err;
+}
+
+int mlxsw_reg_query(struct mlxsw_core *mlxsw_core,
+		    const struct mlxsw_reg_info *reg, char *payload)
+{
+	return mlxsw_core_reg_access(mlxsw_core, reg, payload,
+				     MLXSW_CORE_REG_ACCESS_TYPE_QUERY);
+}
+EXPORT_SYMBOL(mlxsw_reg_query);
+
+int mlxsw_reg_write(struct mlxsw_core *mlxsw_core,
+		    const struct mlxsw_reg_info *reg, char *payload)
+{
+	return mlxsw_core_reg_access(mlxsw_core, reg, payload,
+				     MLXSW_CORE_REG_ACCESS_TYPE_WRITE);
+}
+EXPORT_SYMBOL(mlxsw_reg_write);
+
+void mlxsw_core_skb_receive(struct mlxsw_core *mlxsw_core, struct sk_buff *skb,
+			    struct mlxsw_rx_info *rx_info)
+{
+	struct mlxsw_rx_listener_item *rxl_item;
+	const struct mlxsw_rx_listener *rxl;
+	struct mlxsw_core_pcpu_stats *pcpu_stats;
+	u8 local_port = rx_info->sys_port;
+	bool found = false;
+
+	dev_dbg_ratelimited(mlxsw_core->bus_info->dev, "%s: sys_port = %d, trap_id = 0x%x\n",
+			    __func__, rx_info->sys_port, rx_info->trap_id);
+
+	if ((rx_info->trap_id >= MLXSW_TRAP_ID_MAX) ||
+	    (local_port >= MLXSW_PORT_MAX_PORTS))
+		goto drop;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(rxl_item, &mlxsw_core->rx_listener_list, list) {
+		rxl = &rxl_item->rxl;
+		if ((rxl->local_port == MLXSW_PORT_DONT_CARE ||
+		     rxl->local_port == local_port) &&
+		    rxl->trap_id == rx_info->trap_id) {
+			found = true;
+			break;
+		}
+	}
+	rcu_read_unlock();
+	if (!found)
+		goto drop;
+
+	pcpu_stats = this_cpu_ptr(mlxsw_core->pcpu_stats);
+	u64_stats_update_begin(&pcpu_stats->syncp);
+	pcpu_stats->port_rx_packets[local_port]++;
+	pcpu_stats->port_rx_bytes[local_port] += skb->len;
+	pcpu_stats->trap_rx_packets[rx_info->trap_id]++;
+	pcpu_stats->trap_rx_bytes[rx_info->trap_id] += skb->len;
+	u64_stats_update_end(&pcpu_stats->syncp);
+
+	rxl->func(skb, local_port, rxl_item->priv);
+	return;
+
+drop:
+	if (rx_info->trap_id >= MLXSW_TRAP_ID_MAX)
+		this_cpu_inc(mlxsw_core->pcpu_stats->trap_rx_invalid);
+	else
+		this_cpu_inc(mlxsw_core->pcpu_stats->trap_rx_dropped[rx_info->trap_id]);
+	if (local_port >= MLXSW_PORT_MAX_PORTS)
+		this_cpu_inc(mlxsw_core->pcpu_stats->port_rx_invalid);
+	else
+		this_cpu_inc(mlxsw_core->pcpu_stats->port_rx_dropped[local_port]);
+	dev_kfree_skb(skb);
+}
+EXPORT_SYMBOL(mlxsw_core_skb_receive);
+
+int mlxsw_cmd_exec(struct mlxsw_core *mlxsw_core, u16 opcode, u8 opcode_mod,
+		   u32 in_mod, bool out_mbox_direct,
+		   char *in_mbox, size_t in_mbox_size,
+		   char *out_mbox, size_t out_mbox_size)
+{
+	u8 status;
+	int err;
+
+	BUG_ON(in_mbox_size % sizeof(u32) || out_mbox_size % sizeof(u32));
+	if (!mlxsw_core->bus->cmd_exec)
+		return -EOPNOTSUPP;
+
+	dev_dbg(mlxsw_core->bus_info->dev, "Cmd exec (opcode=%x(%s),opcode_mod=%x,in_mod=%x)\n",
+		opcode, mlxsw_cmd_opcode_str(opcode), opcode_mod, in_mod);
+	if (in_mbox) {
+		dev_dbg(mlxsw_core->bus_info->dev, "Input mailbox:\n");
+		mlxsw_core_buf_dump_dbg(mlxsw_core, in_mbox, in_mbox_size);
+	}
+
+	err = mlxsw_core->bus->cmd_exec(mlxsw_core->bus_priv, opcode,
+					opcode_mod, in_mod, out_mbox_direct,
+					in_mbox, in_mbox_size,
+					out_mbox, out_mbox_size, &status);
+
+	if (err == -EIO && status != MLXSW_CMD_STATUS_OK) {
+		dev_err(mlxsw_core->bus_info->dev, "Cmd exec failed (opcode=%x(%s),opcode_mod=%x,in_mod=%x,status=%x(%s))\n",
+			opcode, mlxsw_cmd_opcode_str(opcode), opcode_mod,
+			in_mod, status, mlxsw_cmd_status_str(status));
+	} else if (err == -ETIMEDOUT) {
+		dev_err(mlxsw_core->bus_info->dev, "Cmd exec timed-out (opcode=%x(%s),opcode_mod=%x,in_mod=%x)\n",
+			opcode, mlxsw_cmd_opcode_str(opcode), opcode_mod,
+			in_mod);
+	}
+
+	if (!err && out_mbox) {
+		dev_dbg(mlxsw_core->bus_info->dev, "Output mailbox:\n");
+		mlxsw_core_buf_dump_dbg(mlxsw_core, out_mbox, out_mbox_size);
+	}
+	return err;
+}
+EXPORT_SYMBOL(mlxsw_cmd_exec);
+
+static int __init mlxsw_core_module_init(void)
+{
+	mlxsw_core_dbg_root = debugfs_create_dir(mlxsw_core_driver_name, NULL);
+	if (!mlxsw_core_dbg_root)
+		return -ENOMEM;
+	return 0;
+}
+
+static void __exit mlxsw_core_module_exit(void)
+{
+	debugfs_remove_recursive(mlxsw_core_dbg_root);
+}
+
+module_init(mlxsw_core_module_init);
+module_exit(mlxsw_core_module_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Jiri Pirko <jiri@mellanox.com>");
+MODULE_DESCRIPTION("Mellanox switch device core driver");
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/core.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/core.h
new file mode 100644
index 000000000..807827350
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/core.h
@@ -0,0 +1,212 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/core.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_CORE_H
+#define _MLXSW_CORE_H
+
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/slab.h>
+#include <linux/gfp.h>
+#include <linux/types.h>
+#include <linux/skbuff.h>
+
+#include "trap.h"
+#include "reg.h"
+
+#include "cmd.h"
+
+#define MLXSW_MODULE_ALIAS_PREFIX "mlxsw-driver-"
+#define MODULE_MLXSW_DRIVER_ALIAS(kind)	\
+	MODULE_ALIAS(MLXSW_MODULE_ALIAS_PREFIX kind)
+
+#define MLXSW_DEVICE_KIND_SWITCHX2 "switchx2"
+#define MLXSW_DEVICE_KIND_SPECTRUM "spectrum"
+
+struct mlxsw_core;
+struct mlxsw_driver;
+struct mlxsw_bus;
+struct mlxsw_bus_info;
+
+int mlxsw_core_driver_register(struct mlxsw_driver *mlxsw_driver);
+void mlxsw_core_driver_unregister(struct mlxsw_driver *mlxsw_driver);
+
+int mlxsw_core_bus_device_register(const struct mlxsw_bus_info *mlxsw_bus_info,
+				   const struct mlxsw_bus *mlxsw_bus,
+				   void *bus_priv);
+void mlxsw_core_bus_device_unregister(struct mlxsw_core *mlxsw_core);
+
+struct mlxsw_tx_info {
+	u8 local_port;
+	bool is_emad;
+};
+
+bool mlxsw_core_skb_transmit_busy(void *driver_priv,
+				  const struct mlxsw_tx_info *tx_info);
+
+int mlxsw_core_skb_transmit(void *driver_priv, struct sk_buff *skb,
+			    const struct mlxsw_tx_info *tx_info);
+
+struct mlxsw_rx_listener {
+	void (*func)(struct sk_buff *skb, u8 local_port, void *priv);
+	u8 local_port;
+	u16 trap_id;
+};
+
+struct mlxsw_event_listener {
+	void (*func)(const struct mlxsw_reg_info *reg,
+		     char *payload, void *priv);
+	enum mlxsw_event_trap_id trap_id;
+};
+
+int mlxsw_core_rx_listener_register(struct mlxsw_core *mlxsw_core,
+				    const struct mlxsw_rx_listener *rxl,
+				    void *priv);
+void mlxsw_core_rx_listener_unregister(struct mlxsw_core *mlxsw_core,
+				       const struct mlxsw_rx_listener *rxl,
+				       void *priv);
+
+int mlxsw_core_event_listener_register(struct mlxsw_core *mlxsw_core,
+				       const struct mlxsw_event_listener *el,
+				       void *priv);
+void mlxsw_core_event_listener_unregister(struct mlxsw_core *mlxsw_core,
+					  const struct mlxsw_event_listener *el,
+					  void *priv);
+
+int mlxsw_reg_query(struct mlxsw_core *mlxsw_core,
+		    const struct mlxsw_reg_info *reg, char *payload);
+int mlxsw_reg_write(struct mlxsw_core *mlxsw_core,
+		    const struct mlxsw_reg_info *reg, char *payload);
+
+struct mlxsw_rx_info {
+	u16 sys_port;
+	int trap_id;
+};
+
+void mlxsw_core_skb_receive(struct mlxsw_core *mlxsw_core, struct sk_buff *skb,
+			    struct mlxsw_rx_info *rx_info);
+
+#define MLXSW_CONFIG_PROFILE_SWID_COUNT 8
+
+struct mlxsw_swid_config {
+	u8	used_type:1,
+		used_properties:1;
+	u8	type;
+	u8	properties;
+};
+
+struct mlxsw_config_profile {
+	u16	used_max_vepa_channels:1,
+		used_max_lag:1,
+		used_max_port_per_lag:1,
+		used_max_mid:1,
+		used_max_pgt:1,
+		used_max_system_port:1,
+		used_max_vlan_groups:1,
+		used_max_regions:1,
+		used_flood_tables:1,
+		used_flood_mode:1,
+		used_max_ib_mc:1,
+		used_max_pkey:1,
+		used_ar_sec:1,
+		used_adaptive_routing_group_cap:1;
+	u8	max_vepa_channels;
+	u16	max_lag;
+	u16	max_port_per_lag;
+	u16	max_mid;
+	u16	max_pgt;
+	u16	max_system_port;
+	u16	max_vlan_groups;
+	u16	max_regions;
+	u8	max_flood_tables;
+	u8	max_vid_flood_tables;
+	u8	flood_mode;
+	u8	max_fid_offset_flood_tables;
+	u16	fid_offset_flood_table_size;
+	u8	max_fid_flood_tables;
+	u16	fid_flood_table_size;
+	u16	max_ib_mc;
+	u16	max_pkey;
+	u8	ar_sec;
+	u16	adaptive_routing_group_cap;
+	u8	arn;
+	struct mlxsw_swid_config swid_config[MLXSW_CONFIG_PROFILE_SWID_COUNT];
+};
+
+struct mlxsw_driver {
+	struct list_head list;
+	const char *kind;
+	struct module *owner;
+	size_t priv_size;
+	int (*init)(void *driver_priv, struct mlxsw_core *mlxsw_core,
+		    const struct mlxsw_bus_info *mlxsw_bus_info);
+	void (*fini)(void *driver_priv);
+	void (*txhdr_construct)(struct sk_buff *skb,
+				const struct mlxsw_tx_info *tx_info);
+	u8 txhdr_len;
+	const struct mlxsw_config_profile *profile;
+};
+
+struct mlxsw_bus {
+	const char *kind;
+	int (*init)(void *bus_priv, struct mlxsw_core *mlxsw_core,
+		    const struct mlxsw_config_profile *profile);
+	void (*fini)(void *bus_priv);
+	bool (*skb_transmit_busy)(void *bus_priv,
+				  const struct mlxsw_tx_info *tx_info);
+	int (*skb_transmit)(void *bus_priv, struct sk_buff *skb,
+			    const struct mlxsw_tx_info *tx_info);
+	int (*cmd_exec)(void *bus_priv, u16 opcode, u8 opcode_mod,
+			u32 in_mod, bool out_mbox_direct,
+			char *in_mbox, size_t in_mbox_size,
+			char *out_mbox, size_t out_mbox_size,
+			u8 *p_status);
+};
+
+struct mlxsw_bus_info {
+	const char *device_kind;
+	const char *device_name;
+	struct device *dev;
+	struct {
+		u16 major;
+		u16 minor;
+		u16 subminor;
+	} fw_rev;
+	u8 vsd[MLXSW_CMD_BOARDINFO_VSD_LEN];
+	u8 psid[MLXSW_CMD_BOARDINFO_PSID_LEN];
+};
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/emad.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/emad.h
new file mode 100644
index 000000000..97b6bb5d9
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/emad.h
@@ -0,0 +1,127 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/emad.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_EMAD_H
+#define _MLXSW_EMAD_H
+
+#define MLXSW_EMAD_MAX_FRAME_LEN 1518	/* Length in u8 */
+#define MLXSW_EMAD_MAX_RETRY 5
+
+/* EMAD Ethernet header */
+#define MLXSW_EMAD_ETH_HDR_LEN 0x10	/* Length in u8 */
+#define MLXSW_EMAD_EH_DMAC "\x01\x02\xc9\x00\x00\x01"
+#define MLXSW_EMAD_EH_SMAC "\x00\x02\xc9\x01\x02\x03"
+#define MLXSW_EMAD_EH_ETHERTYPE 0x8932
+#define MLXSW_EMAD_EH_MLX_PROTO 0
+#define MLXSW_EMAD_EH_PROTO_VERSION 0
+
+/* EMAD TLV Types */
+enum {
+	MLXSW_EMAD_TLV_TYPE_END,
+	MLXSW_EMAD_TLV_TYPE_OP,
+	MLXSW_EMAD_TLV_TYPE_DR,
+	MLXSW_EMAD_TLV_TYPE_REG,
+	MLXSW_EMAD_TLV_TYPE_USERDATA,
+	MLXSW_EMAD_TLV_TYPE_OOBETH,
+};
+
+/* OP TLV */
+#define MLXSW_EMAD_OP_TLV_LEN 4		/* Length in u32 */
+
+enum {
+	MLXSW_EMAD_OP_TLV_CLASS_REG_ACCESS = 1,
+	MLXSW_EMAD_OP_TLV_CLASS_IPC = 2,
+};
+
+enum mlxsw_emad_op_tlv_status {
+	MLXSW_EMAD_OP_TLV_STATUS_SUCCESS,
+	MLXSW_EMAD_OP_TLV_STATUS_BUSY,
+	MLXSW_EMAD_OP_TLV_STATUS_VERSION_NOT_SUPPORTED,
+	MLXSW_EMAD_OP_TLV_STATUS_UNKNOWN_TLV,
+	MLXSW_EMAD_OP_TLV_STATUS_REGISTER_NOT_SUPPORTED,
+	MLXSW_EMAD_OP_TLV_STATUS_CLASS_NOT_SUPPORTED,
+	MLXSW_EMAD_OP_TLV_STATUS_METHOD_NOT_SUPPORTED,
+	MLXSW_EMAD_OP_TLV_STATUS_BAD_PARAMETER,
+	MLXSW_EMAD_OP_TLV_STATUS_RESOURCE_NOT_AVAILABLE,
+	MLXSW_EMAD_OP_TLV_STATUS_MESSAGE_RECEIPT_ACK,
+	MLXSW_EMAD_OP_TLV_STATUS_INTERNAL_ERROR = 0x70,
+};
+
+static inline char *mlxsw_emad_op_tlv_status_str(u8 status)
+{
+	switch (status) {
+	case MLXSW_EMAD_OP_TLV_STATUS_SUCCESS:
+		return "operation performed";
+	case MLXSW_EMAD_OP_TLV_STATUS_BUSY:
+		return "device is busy";
+	case MLXSW_EMAD_OP_TLV_STATUS_VERSION_NOT_SUPPORTED:
+		return "version not supported";
+	case MLXSW_EMAD_OP_TLV_STATUS_UNKNOWN_TLV:
+		return "unknown TLV";
+	case MLXSW_EMAD_OP_TLV_STATUS_REGISTER_NOT_SUPPORTED:
+		return "register not supported";
+	case MLXSW_EMAD_OP_TLV_STATUS_CLASS_NOT_SUPPORTED:
+		return "class not supported";
+	case MLXSW_EMAD_OP_TLV_STATUS_METHOD_NOT_SUPPORTED:
+		return "method not supported";
+	case MLXSW_EMAD_OP_TLV_STATUS_BAD_PARAMETER:
+		return "bad parameter";
+	case MLXSW_EMAD_OP_TLV_STATUS_RESOURCE_NOT_AVAILABLE:
+		return "resource not available";
+	case MLXSW_EMAD_OP_TLV_STATUS_MESSAGE_RECEIPT_ACK:
+		return "acknowledged. retransmit";
+	case MLXSW_EMAD_OP_TLV_STATUS_INTERNAL_ERROR:
+		return "internal error";
+	default:
+		return "*UNKNOWN*";
+	}
+}
+
+enum {
+	MLXSW_EMAD_OP_TLV_REQUEST,
+	MLXSW_EMAD_OP_TLV_RESPONSE
+};
+
+enum {
+	MLXSW_EMAD_OP_TLV_METHOD_QUERY = 1,
+	MLXSW_EMAD_OP_TLV_METHOD_WRITE = 2,
+	MLXSW_EMAD_OP_TLV_METHOD_SEND = 3,
+	MLXSW_EMAD_OP_TLV_METHOD_EVENT = 5,
+};
+
+/* END TLV */
+#define MLXSW_EMAD_END_TLV_LEN 1	/* Length in u32 */
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/item.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/item.h
new file mode 100644
index 000000000..a94dbda65
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/item.h
@@ -0,0 +1,441 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/item.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_ITEM_H
+#define _MLXSW_ITEM_H
+
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/bitops.h>
+
+struct mlxsw_item {
+	unsigned short	offset;		/* bytes in container */
+	unsigned short	step;		/* step in bytes for indexed items */
+	unsigned short	in_step_offset; /* offset within one step */
+	unsigned char	shift;		/* shift in bits */
+	unsigned char	element_size;	/* size of element in bit array */
+	bool		no_real_shift;
+	union {
+		unsigned char	bits;
+		unsigned short	bytes;
+	} size;
+	const char	*name;
+};
+
+static inline unsigned int
+__mlxsw_item_offset(struct mlxsw_item *item, unsigned short index,
+		    size_t typesize)
+{
+	BUG_ON(index && !item->step);
+	if (item->offset % typesize != 0 ||
+	    item->step % typesize != 0 ||
+	    item->in_step_offset % typesize != 0) {
+		pr_err("mlxsw: item bug (name=%s,offset=%x,step=%x,in_step_offset=%x,typesize=%zx)\n",
+		       item->name, item->offset, item->step,
+		       item->in_step_offset, typesize);
+		BUG();
+	}
+
+	return ((item->offset + item->step * index + item->in_step_offset) /
+		typesize);
+}
+
+static inline u16 __mlxsw_item_get16(char *buf, struct mlxsw_item *item,
+				     unsigned short index)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index, sizeof(u16));
+	__be16 *b = (__be16 *) buf;
+	u16 tmp;
+
+	tmp = be16_to_cpu(b[offset]);
+	tmp >>= item->shift;
+	tmp &= GENMASK(item->size.bits - 1, 0);
+	if (item->no_real_shift)
+		tmp <<= item->shift;
+	return tmp;
+}
+
+static inline void __mlxsw_item_set16(char *buf, struct mlxsw_item *item,
+				      unsigned short index, u16 val)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index,
+						  sizeof(u16));
+	__be16 *b = (__be16 *) buf;
+	u16 mask = GENMASK(item->size.bits - 1, 0) << item->shift;
+	u16 tmp;
+
+	if (!item->no_real_shift)
+		val <<= item->shift;
+	val &= mask;
+	tmp = be16_to_cpu(b[offset]);
+	tmp &= ~mask;
+	tmp |= val;
+	b[offset] = cpu_to_be16(tmp);
+}
+
+static inline u32 __mlxsw_item_get32(char *buf, struct mlxsw_item *item,
+				     unsigned short index)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index, sizeof(u32));
+	__be32 *b = (__be32 *) buf;
+	u32 tmp;
+
+	tmp = be32_to_cpu(b[offset]);
+	tmp >>= item->shift;
+	tmp &= GENMASK(item->size.bits - 1, 0);
+	if (item->no_real_shift)
+		tmp <<= item->shift;
+	return tmp;
+}
+
+static inline void __mlxsw_item_set32(char *buf, struct mlxsw_item *item,
+				      unsigned short index, u32 val)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index,
+						  sizeof(u32));
+	__be32 *b = (__be32 *) buf;
+	u32 mask = GENMASK(item->size.bits - 1, 0) << item->shift;
+	u32 tmp;
+
+	if (!item->no_real_shift)
+		val <<= item->shift;
+	val &= mask;
+	tmp = be32_to_cpu(b[offset]);
+	tmp &= ~mask;
+	tmp |= val;
+	b[offset] = cpu_to_be32(tmp);
+}
+
+static inline u64 __mlxsw_item_get64(char *buf, struct mlxsw_item *item,
+				     unsigned short index)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index, sizeof(u64));
+	__be64 *b = (__be64 *) buf;
+	u64 tmp;
+
+	tmp = be64_to_cpu(b[offset]);
+	tmp >>= item->shift;
+	tmp &= GENMASK_ULL(item->size.bits - 1, 0);
+	if (item->no_real_shift)
+		tmp <<= item->shift;
+	return tmp;
+}
+
+static inline void __mlxsw_item_set64(char *buf, struct mlxsw_item *item,
+				      unsigned short index, u64 val)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index, sizeof(u64));
+	__be64 *b = (__be64 *) buf;
+	u64 mask = GENMASK_ULL(item->size.bits - 1, 0) << item->shift;
+	u64 tmp;
+
+	if (!item->no_real_shift)
+		val <<= item->shift;
+	val &= mask;
+	tmp = be64_to_cpu(b[offset]);
+	tmp &= ~mask;
+	tmp |= val;
+	b[offset] = cpu_to_be64(tmp);
+}
+
+static inline void __mlxsw_item_memcpy_from(char *buf, char *dst,
+					    struct mlxsw_item *item,
+					    unsigned short index)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index, sizeof(char));
+
+	memcpy(dst, &buf[offset], item->size.bytes);
+}
+
+static inline void __mlxsw_item_memcpy_to(char *buf, const char *src,
+					  struct mlxsw_item *item,
+					  unsigned short index)
+{
+	unsigned int offset = __mlxsw_item_offset(item, index, sizeof(char));
+
+	memcpy(&buf[offset], src, item->size.bytes);
+}
+
+static inline u16
+__mlxsw_item_bit_array_offset(struct mlxsw_item *item, u16 index, u8 *shift)
+{
+	u16 max_index, be_index;
+	u16 offset;		/* byte offset inside the array */
+	u8 in_byte_index;
+
+	BUG_ON(index && !item->element_size);
+	if (item->offset % sizeof(u32) != 0 ||
+	    BITS_PER_BYTE % item->element_size != 0) {
+		pr_err("mlxsw: item bug (name=%s,offset=%x,element_size=%x)\n",
+		       item->name, item->offset, item->element_size);
+		BUG();
+	}
+
+	max_index = (item->size.bytes << 3) / item->element_size - 1;
+	be_index = max_index - index;
+	offset = be_index * item->element_size >> 3;
+	in_byte_index  = index % (BITS_PER_BYTE / item->element_size);
+	*shift = in_byte_index * item->element_size;
+
+	return item->offset + offset;
+}
+
+static inline u8 __mlxsw_item_bit_array_get(char *buf, struct mlxsw_item *item,
+					    u16 index)
+{
+	u8 shift, tmp;
+	u16 offset = __mlxsw_item_bit_array_offset(item, index, &shift);
+
+	tmp = buf[offset];
+	tmp >>= shift;
+	tmp &= GENMASK(item->element_size - 1, 0);
+	return tmp;
+}
+
+static inline void __mlxsw_item_bit_array_set(char *buf, struct mlxsw_item *item,
+					      u16 index, u8 val)
+{
+	u8 shift, tmp;
+	u16 offset = __mlxsw_item_bit_array_offset(item, index, &shift);
+	u8 mask = GENMASK(item->element_size - 1, 0) << shift;
+
+	val <<= shift;
+	val &= mask;
+	tmp = buf[offset];
+	tmp &= ~mask;
+	tmp |= val;
+	buf[offset] = tmp;
+}
+
+#define __ITEM_NAME(_type, _cname, _iname)					\
+	mlxsw_##_type##_##_cname##_##_iname##_item
+
+/* _type: cmd_mbox, reg, etc.
+ * _cname: containter name (e.g. command name, register name)
+ * _iname: item name within the container
+ */
+
+#define MLXSW_ITEM16(_type, _cname, _iname, _offset, _shift, _sizebits)		\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.shift = _shift,							\
+	.size = {.bits = _sizebits,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u16 mlxsw_##_type##_##_cname##_##_iname##_get(char *buf)		\
+{										\
+	return __mlxsw_item_get16(buf, &__ITEM_NAME(_type, _cname, _iname), 0);	\
+}										\
+static inline void mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, u16 val)\
+{										\
+	__mlxsw_item_set16(buf, &__ITEM_NAME(_type, _cname, _iname), 0, val);	\
+}
+
+#define MLXSW_ITEM16_INDEXED(_type, _cname, _iname, _offset, _shift, _sizebits,	\
+			     _step, _instepoffset, _norealshift)		\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.step = _step,								\
+	.in_step_offset = _instepoffset,					\
+	.shift = _shift,							\
+	.no_real_shift = _norealshift,						\
+	.size = {.bits = _sizebits,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u16								\
+mlxsw_##_type##_##_cname##_##_iname##_get(char *buf, unsigned short index)	\
+{										\
+	return __mlxsw_item_get16(buf, &__ITEM_NAME(_type, _cname, _iname),	\
+				  index);					\
+}										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, unsigned short index,	\
+					  u16 val)				\
+{										\
+	__mlxsw_item_set16(buf, &__ITEM_NAME(_type, _cname, _iname),		\
+			   index, val);						\
+}
+
+#define MLXSW_ITEM32(_type, _cname, _iname, _offset, _shift, _sizebits)		\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.shift = _shift,							\
+	.size = {.bits = _sizebits,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u32 mlxsw_##_type##_##_cname##_##_iname##_get(char *buf)		\
+{										\
+	return __mlxsw_item_get32(buf, &__ITEM_NAME(_type, _cname, _iname), 0);	\
+}										\
+static inline void mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, u32 val)\
+{										\
+	__mlxsw_item_set32(buf, &__ITEM_NAME(_type, _cname, _iname), 0, val);	\
+}
+
+#define MLXSW_ITEM32_INDEXED(_type, _cname, _iname, _offset, _shift, _sizebits,	\
+			     _step, _instepoffset, _norealshift)		\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.step = _step,								\
+	.in_step_offset = _instepoffset,					\
+	.shift = _shift,							\
+	.no_real_shift = _norealshift,						\
+	.size = {.bits = _sizebits,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u32								\
+mlxsw_##_type##_##_cname##_##_iname##_get(char *buf, unsigned short index)	\
+{										\
+	return __mlxsw_item_get32(buf, &__ITEM_NAME(_type, _cname, _iname),	\
+				  index);					\
+}										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, unsigned short index,	\
+					  u32 val)				\
+{										\
+	__mlxsw_item_set32(buf, &__ITEM_NAME(_type, _cname, _iname),		\
+			   index, val);						\
+}
+
+#define MLXSW_ITEM64(_type, _cname, _iname, _offset, _shift, _sizebits)		\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.shift = _shift,							\
+	.size = {.bits = _sizebits,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u64 mlxsw_##_type##_##_cname##_##_iname##_get(char *buf)		\
+{										\
+	return __mlxsw_item_get64(buf, &__ITEM_NAME(_type, _cname, _iname), 0);	\
+}										\
+static inline void mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, u64 val)\
+{										\
+	__mlxsw_item_set64(buf, &__ITEM_NAME(_type, _cname, _iname), 0,	val);	\
+}
+
+#define MLXSW_ITEM64_INDEXED(_type, _cname, _iname, _offset, _shift,		\
+			     _sizebits, _step, _instepoffset, _norealshift)	\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.step = _step,								\
+	.in_step_offset = _instepoffset,					\
+	.shift = _shift,							\
+	.no_real_shift = _norealshift,						\
+	.size = {.bits = _sizebits,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u64								\
+mlxsw_##_type##_##_cname##_##_iname##_get(char *buf, unsigned short index)	\
+{										\
+	return __mlxsw_item_get64(buf, &__ITEM_NAME(_type, _cname, _iname),	\
+				  index);					\
+}										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, unsigned short index,	\
+					  u64 val)				\
+{										\
+	__mlxsw_item_set64(buf, &__ITEM_NAME(_type, _cname, _iname),		\
+			   index, val);						\
+}
+
+#define MLXSW_ITEM_BUF(_type, _cname, _iname, _offset, _sizebytes)		\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.size = {.bytes = _sizebytes,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_memcpy_from(char *buf, char *dst)		\
+{										\
+	__mlxsw_item_memcpy_from(buf, dst,					\
+				 &__ITEM_NAME(_type, _cname, _iname), 0);	\
+}										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_memcpy_to(char *buf, const char *src)	\
+{										\
+	__mlxsw_item_memcpy_to(buf, src,					\
+			       &__ITEM_NAME(_type, _cname, _iname), 0);		\
+}
+
+#define MLXSW_ITEM_BUF_INDEXED(_type, _cname, _iname, _offset, _sizebytes,	\
+			       _step, _instepoffset)				\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.step = _step,								\
+	.in_step_offset = _instepoffset,					\
+	.size = {.bytes = _sizebytes,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_memcpy_from(char *buf,			\
+						  unsigned short index,		\
+						  char *dst)			\
+{										\
+	__mlxsw_item_memcpy_from(buf, dst,					\
+				 &__ITEM_NAME(_type, _cname, _iname), index);	\
+}										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_memcpy_to(char *buf,			\
+						unsigned short index,		\
+						const char *src)		\
+{										\
+	__mlxsw_item_memcpy_to(buf, src,					\
+			       &__ITEM_NAME(_type, _cname, _iname), index);	\
+}
+
+#define MLXSW_ITEM_BIT_ARRAY(_type, _cname, _iname, _offset, _sizebytes,	\
+			     _element_size)					\
+static struct mlxsw_item __ITEM_NAME(_type, _cname, _iname) = {			\
+	.offset = _offset,							\
+	.element_size = _element_size,						\
+	.size = {.bytes = _sizebytes,},						\
+	.name = #_type "_" #_cname "_" #_iname,					\
+};										\
+static inline u8								\
+mlxsw_##_type##_##_cname##_##_iname##_get(char *buf, u16 index)			\
+{										\
+	return __mlxsw_item_bit_array_get(buf,					\
+					  &__ITEM_NAME(_type, _cname, _iname),	\
+					  index);				\
+}										\
+static inline void								\
+mlxsw_##_type##_##_cname##_##_iname##_set(char *buf, u16 index, u8 val)		\
+{										\
+	return __mlxsw_item_bit_array_set(buf,					\
+					  &__ITEM_NAME(_type, _cname, _iname),	\
+					  index, val);				\
+}										\
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/pci.c b/kernel/drivers/net/ethernet/mellanox/mlxsw/pci.c
new file mode 100644
index 000000000..de69e719d
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/pci.c
@@ -0,0 +1,1847 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/pci.c
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/export.h>
+#include <linux/err.h>
+#include <linux/device.h>
+#include <linux/pci.h>
+#include <linux/interrupt.h>
+#include <linux/wait.h>
+#include <linux/types.h>
+#include <linux/skbuff.h>
+#include <linux/if_vlan.h>
+#include <linux/log2.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include <linux/string.h>
+
+#include "pci.h"
+#include "core.h"
+#include "cmd.h"
+#include "port.h"
+
+static const char mlxsw_pci_driver_name[] = "mlxsw_pci";
+
+static const struct pci_device_id mlxsw_pci_id_table[] = {
+	{PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_SWITCHX2), 0},
+	{PCI_VDEVICE(MELLANOX, PCI_DEVICE_ID_MELLANOX_SPECTRUM), 0},
+	{0, }
+};
+
+static struct dentry *mlxsw_pci_dbg_root;
+
+static const char *mlxsw_pci_device_kind_get(const struct pci_device_id *id)
+{
+	switch (id->device) {
+	case PCI_DEVICE_ID_MELLANOX_SWITCHX2:
+		return MLXSW_DEVICE_KIND_SWITCHX2;
+	case PCI_DEVICE_ID_MELLANOX_SPECTRUM:
+		return MLXSW_DEVICE_KIND_SPECTRUM;
+	default:
+		BUG();
+	}
+}
+
+#define mlxsw_pci_write32(mlxsw_pci, reg, val) \
+	iowrite32be(val, (mlxsw_pci)->hw_addr + (MLXSW_PCI_ ## reg))
+#define mlxsw_pci_read32(mlxsw_pci, reg) \
+	ioread32be((mlxsw_pci)->hw_addr + (MLXSW_PCI_ ## reg))
+
+enum mlxsw_pci_queue_type {
+	MLXSW_PCI_QUEUE_TYPE_SDQ,
+	MLXSW_PCI_QUEUE_TYPE_RDQ,
+	MLXSW_PCI_QUEUE_TYPE_CQ,
+	MLXSW_PCI_QUEUE_TYPE_EQ,
+};
+
+static const char *mlxsw_pci_queue_type_str(enum mlxsw_pci_queue_type q_type)
+{
+	switch (q_type) {
+	case MLXSW_PCI_QUEUE_TYPE_SDQ:
+		return "sdq";
+	case MLXSW_PCI_QUEUE_TYPE_RDQ:
+		return "rdq";
+	case MLXSW_PCI_QUEUE_TYPE_CQ:
+		return "cq";
+	case MLXSW_PCI_QUEUE_TYPE_EQ:
+		return "eq";
+	}
+	BUG();
+}
+
+#define MLXSW_PCI_QUEUE_TYPE_COUNT	4
+
+static const u16 mlxsw_pci_doorbell_type_offset[] = {
+	MLXSW_PCI_DOORBELL_SDQ_OFFSET,	/* for type MLXSW_PCI_QUEUE_TYPE_SDQ */
+	MLXSW_PCI_DOORBELL_RDQ_OFFSET,	/* for type MLXSW_PCI_QUEUE_TYPE_RDQ */
+	MLXSW_PCI_DOORBELL_CQ_OFFSET,	/* for type MLXSW_PCI_QUEUE_TYPE_CQ */
+	MLXSW_PCI_DOORBELL_EQ_OFFSET,	/* for type MLXSW_PCI_QUEUE_TYPE_EQ */
+};
+
+static const u16 mlxsw_pci_doorbell_arm_type_offset[] = {
+	0, /* unused */
+	0, /* unused */
+	MLXSW_PCI_DOORBELL_ARM_CQ_OFFSET, /* for type MLXSW_PCI_QUEUE_TYPE_CQ */
+	MLXSW_PCI_DOORBELL_ARM_EQ_OFFSET, /* for type MLXSW_PCI_QUEUE_TYPE_EQ */
+};
+
+struct mlxsw_pci_mem_item {
+	char *buf;
+	dma_addr_t mapaddr;
+	size_t size;
+};
+
+struct mlxsw_pci_queue_elem_info {
+	char *elem; /* pointer to actual dma mapped element mem chunk */
+	union {
+		struct {
+			struct sk_buff *skb;
+		} sdq;
+		struct {
+			struct sk_buff *skb;
+		} rdq;
+	} u;
+};
+
+struct mlxsw_pci_queue {
+	spinlock_t lock; /* for queue accesses */
+	struct mlxsw_pci_mem_item mem_item;
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	u16 producer_counter;
+	u16 consumer_counter;
+	u16 count; /* number of elements in queue */
+	u8 num; /* queue number */
+	u8 elem_size; /* size of one element */
+	enum mlxsw_pci_queue_type type;
+	struct tasklet_struct tasklet; /* queue processing tasklet */
+	struct mlxsw_pci *pci;
+	union {
+		struct {
+			u32 comp_sdq_count;
+			u32 comp_rdq_count;
+		} cq;
+		struct {
+			u32 ev_cmd_count;
+			u32 ev_comp_count;
+			u32 ev_other_count;
+		} eq;
+	} u;
+};
+
+struct mlxsw_pci_queue_type_group {
+	struct mlxsw_pci_queue *q;
+	u8 count; /* number of queues in group */
+};
+
+struct mlxsw_pci {
+	struct pci_dev *pdev;
+	u8 __iomem *hw_addr;
+	struct mlxsw_pci_queue_type_group queues[MLXSW_PCI_QUEUE_TYPE_COUNT];
+	u32 doorbell_offset;
+	struct msix_entry msix_entry;
+	struct mlxsw_core *core;
+	struct {
+		struct mlxsw_pci_mem_item *items;
+		unsigned int count;
+	} fw_area;
+	struct {
+		struct mlxsw_pci_mem_item out_mbox;
+		struct mlxsw_pci_mem_item in_mbox;
+		struct mutex lock; /* Lock access to command registers */
+		bool nopoll;
+		wait_queue_head_t wait;
+		bool wait_done;
+		struct {
+			u8 status;
+			u64 out_param;
+		} comp;
+	} cmd;
+	struct mlxsw_bus_info bus_info;
+	struct dentry *dbg_dir;
+};
+
+static void mlxsw_pci_queue_tasklet_schedule(struct mlxsw_pci_queue *q)
+{
+	tasklet_schedule(&q->tasklet);
+}
+
+static char *__mlxsw_pci_queue_elem_get(struct mlxsw_pci_queue *q,
+					size_t elem_size, int elem_index)
+{
+	return q->mem_item.buf + (elem_size * elem_index);
+}
+
+static struct mlxsw_pci_queue_elem_info *
+mlxsw_pci_queue_elem_info_get(struct mlxsw_pci_queue *q, int elem_index)
+{
+	return &q->elem_info[elem_index];
+}
+
+static struct mlxsw_pci_queue_elem_info *
+mlxsw_pci_queue_elem_info_producer_get(struct mlxsw_pci_queue *q)
+{
+	int index = q->producer_counter & (q->count - 1);
+
+	if ((q->producer_counter - q->consumer_counter) == q->count)
+		return NULL;
+	return mlxsw_pci_queue_elem_info_get(q, index);
+}
+
+static struct mlxsw_pci_queue_elem_info *
+mlxsw_pci_queue_elem_info_consumer_get(struct mlxsw_pci_queue *q)
+{
+	int index = q->consumer_counter & (q->count - 1);
+
+	return mlxsw_pci_queue_elem_info_get(q, index);
+}
+
+static char *mlxsw_pci_queue_elem_get(struct mlxsw_pci_queue *q, int elem_index)
+{
+	return mlxsw_pci_queue_elem_info_get(q, elem_index)->elem;
+}
+
+static bool mlxsw_pci_elem_hw_owned(struct mlxsw_pci_queue *q, bool owner_bit)
+{
+	return owner_bit != !!(q->consumer_counter & q->count);
+}
+
+static char *mlxsw_pci_queue_sw_elem_get(struct mlxsw_pci_queue *q,
+					 u32 (*get_elem_owner_func)(char *))
+{
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	char *elem;
+	bool owner_bit;
+
+	elem_info = mlxsw_pci_queue_elem_info_consumer_get(q);
+	elem = elem_info->elem;
+	owner_bit = get_elem_owner_func(elem);
+	if (mlxsw_pci_elem_hw_owned(q, owner_bit))
+		return NULL;
+	q->consumer_counter++;
+	rmb(); /* make sure we read owned bit before the rest of elem */
+	return elem;
+}
+
+static struct mlxsw_pci_queue_type_group *
+mlxsw_pci_queue_type_group_get(struct mlxsw_pci *mlxsw_pci,
+			       enum mlxsw_pci_queue_type q_type)
+{
+	return &mlxsw_pci->queues[q_type];
+}
+
+static u8 __mlxsw_pci_queue_count(struct mlxsw_pci *mlxsw_pci,
+				  enum mlxsw_pci_queue_type q_type)
+{
+	struct mlxsw_pci_queue_type_group *queue_group;
+
+	queue_group = mlxsw_pci_queue_type_group_get(mlxsw_pci, q_type);
+	return queue_group->count;
+}
+
+static u8 mlxsw_pci_sdq_count(struct mlxsw_pci *mlxsw_pci)
+{
+	return __mlxsw_pci_queue_count(mlxsw_pci, MLXSW_PCI_QUEUE_TYPE_SDQ);
+}
+
+static u8 mlxsw_pci_rdq_count(struct mlxsw_pci *mlxsw_pci)
+{
+	return __mlxsw_pci_queue_count(mlxsw_pci, MLXSW_PCI_QUEUE_TYPE_RDQ);
+}
+
+static u8 mlxsw_pci_cq_count(struct mlxsw_pci *mlxsw_pci)
+{
+	return __mlxsw_pci_queue_count(mlxsw_pci, MLXSW_PCI_QUEUE_TYPE_CQ);
+}
+
+static u8 mlxsw_pci_eq_count(struct mlxsw_pci *mlxsw_pci)
+{
+	return __mlxsw_pci_queue_count(mlxsw_pci, MLXSW_PCI_QUEUE_TYPE_EQ);
+}
+
+static struct mlxsw_pci_queue *
+__mlxsw_pci_queue_get(struct mlxsw_pci *mlxsw_pci,
+		      enum mlxsw_pci_queue_type q_type, u8 q_num)
+{
+	return &mlxsw_pci->queues[q_type].q[q_num];
+}
+
+static struct mlxsw_pci_queue *mlxsw_pci_sdq_get(struct mlxsw_pci *mlxsw_pci,
+						 u8 q_num)
+{
+	return __mlxsw_pci_queue_get(mlxsw_pci,
+				     MLXSW_PCI_QUEUE_TYPE_SDQ, q_num);
+}
+
+static struct mlxsw_pci_queue *mlxsw_pci_rdq_get(struct mlxsw_pci *mlxsw_pci,
+						 u8 q_num)
+{
+	return __mlxsw_pci_queue_get(mlxsw_pci,
+				     MLXSW_PCI_QUEUE_TYPE_RDQ, q_num);
+}
+
+static struct mlxsw_pci_queue *mlxsw_pci_cq_get(struct mlxsw_pci *mlxsw_pci,
+						u8 q_num)
+{
+	return __mlxsw_pci_queue_get(mlxsw_pci, MLXSW_PCI_QUEUE_TYPE_CQ, q_num);
+}
+
+static struct mlxsw_pci_queue *mlxsw_pci_eq_get(struct mlxsw_pci *mlxsw_pci,
+						u8 q_num)
+{
+	return __mlxsw_pci_queue_get(mlxsw_pci, MLXSW_PCI_QUEUE_TYPE_EQ, q_num);
+}
+
+static void __mlxsw_pci_queue_doorbell_set(struct mlxsw_pci *mlxsw_pci,
+					   struct mlxsw_pci_queue *q,
+					   u16 val)
+{
+	mlxsw_pci_write32(mlxsw_pci,
+			  DOORBELL(mlxsw_pci->doorbell_offset,
+				   mlxsw_pci_doorbell_type_offset[q->type],
+				   q->num), val);
+}
+
+static void __mlxsw_pci_queue_doorbell_arm_set(struct mlxsw_pci *mlxsw_pci,
+					       struct mlxsw_pci_queue *q,
+					       u16 val)
+{
+	mlxsw_pci_write32(mlxsw_pci,
+			  DOORBELL(mlxsw_pci->doorbell_offset,
+				   mlxsw_pci_doorbell_arm_type_offset[q->type],
+				   q->num), val);
+}
+
+static void mlxsw_pci_queue_doorbell_producer_ring(struct mlxsw_pci *mlxsw_pci,
+						   struct mlxsw_pci_queue *q)
+{
+	wmb(); /* ensure all writes are done before we ring a bell */
+	__mlxsw_pci_queue_doorbell_set(mlxsw_pci, q, q->producer_counter);
+}
+
+static void mlxsw_pci_queue_doorbell_consumer_ring(struct mlxsw_pci *mlxsw_pci,
+						   struct mlxsw_pci_queue *q)
+{
+	wmb(); /* ensure all writes are done before we ring a bell */
+	__mlxsw_pci_queue_doorbell_set(mlxsw_pci, q,
+				       q->consumer_counter + q->count);
+}
+
+static void
+mlxsw_pci_queue_doorbell_arm_consumer_ring(struct mlxsw_pci *mlxsw_pci,
+					   struct mlxsw_pci_queue *q)
+{
+	wmb(); /* ensure all writes are done before we ring a bell */
+	__mlxsw_pci_queue_doorbell_arm_set(mlxsw_pci, q, q->consumer_counter);
+}
+
+static dma_addr_t __mlxsw_pci_queue_page_get(struct mlxsw_pci_queue *q,
+					     int page_index)
+{
+	return q->mem_item.mapaddr + MLXSW_PCI_PAGE_SIZE * page_index;
+}
+
+static int mlxsw_pci_sdq_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+			      struct mlxsw_pci_queue *q)
+{
+	int i;
+	int err;
+
+	q->producer_counter = 0;
+	q->consumer_counter = 0;
+
+	/* Set CQ of same number of this SDQ. */
+	mlxsw_cmd_mbox_sw2hw_dq_cq_set(mbox, q->num);
+	mlxsw_cmd_mbox_sw2hw_dq_sdq_tclass_set(mbox, 7);
+	mlxsw_cmd_mbox_sw2hw_dq_log2_dq_sz_set(mbox, 3); /* 8 pages */
+	for (i = 0; i < MLXSW_PCI_AQ_PAGES; i++) {
+		dma_addr_t mapaddr = __mlxsw_pci_queue_page_get(q, i);
+
+		mlxsw_cmd_mbox_sw2hw_dq_pa_set(mbox, i, mapaddr);
+	}
+
+	err = mlxsw_cmd_sw2hw_sdq(mlxsw_pci->core, mbox, q->num);
+	if (err)
+		return err;
+	mlxsw_pci_queue_doorbell_producer_ring(mlxsw_pci, q);
+	return 0;
+}
+
+static void mlxsw_pci_sdq_fini(struct mlxsw_pci *mlxsw_pci,
+			       struct mlxsw_pci_queue *q)
+{
+	mlxsw_cmd_hw2sw_sdq(mlxsw_pci->core, q->num);
+}
+
+static int mlxsw_pci_sdq_dbg_read(struct seq_file *file, void *data)
+{
+	struct mlxsw_pci *mlxsw_pci = dev_get_drvdata(file->private);
+	struct mlxsw_pci_queue *q;
+	int i;
+	static const char hdr[] =
+		"NUM PROD_COUNT CONS_COUNT COUNT\n";
+
+	seq_printf(file, hdr);
+	for (i = 0; i < mlxsw_pci_sdq_count(mlxsw_pci); i++) {
+		q = mlxsw_pci_sdq_get(mlxsw_pci, i);
+		spin_lock_bh(&q->lock);
+		seq_printf(file, "%3d %10d %10d %5d\n",
+			   i, q->producer_counter, q->consumer_counter,
+			   q->count);
+		spin_unlock_bh(&q->lock);
+	}
+	return 0;
+}
+
+static int mlxsw_pci_wqe_frag_map(struct mlxsw_pci *mlxsw_pci, char *wqe,
+				  int index, char *frag_data, size_t frag_len,
+				  int direction)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	dma_addr_t mapaddr;
+
+	mapaddr = pci_map_single(pdev, frag_data, frag_len, direction);
+	if (unlikely(pci_dma_mapping_error(pdev, mapaddr))) {
+		dev_err_ratelimited(&pdev->dev, "failed to dma map tx frag\n");
+		return -EIO;
+	}
+	mlxsw_pci_wqe_address_set(wqe, index, mapaddr);
+	mlxsw_pci_wqe_byte_count_set(wqe, index, frag_len);
+	return 0;
+}
+
+static void mlxsw_pci_wqe_frag_unmap(struct mlxsw_pci *mlxsw_pci, char *wqe,
+				     int index, int direction)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	size_t frag_len = mlxsw_pci_wqe_byte_count_get(wqe, index);
+	dma_addr_t mapaddr = mlxsw_pci_wqe_address_get(wqe, index);
+
+	if (!frag_len)
+		return;
+	pci_unmap_single(pdev, mapaddr, frag_len, direction);
+}
+
+static int mlxsw_pci_rdq_skb_alloc(struct mlxsw_pci *mlxsw_pci,
+				   struct mlxsw_pci_queue_elem_info *elem_info)
+{
+	size_t buf_len = MLXSW_PORT_MAX_MTU;
+	char *wqe = elem_info->elem;
+	struct sk_buff *skb;
+	int err;
+
+	elem_info->u.rdq.skb = NULL;
+	skb = netdev_alloc_skb_ip_align(NULL, buf_len);
+	if (!skb)
+		return -ENOMEM;
+
+	/* Assume that wqe was previously zeroed. */
+
+	err = mlxsw_pci_wqe_frag_map(mlxsw_pci, wqe, 0, skb->data,
+				     buf_len, DMA_FROM_DEVICE);
+	if (err)
+		goto err_frag_map;
+
+	elem_info->u.rdq.skb = skb;
+	return 0;
+
+err_frag_map:
+	dev_kfree_skb_any(skb);
+	return err;
+}
+
+static void mlxsw_pci_rdq_skb_free(struct mlxsw_pci *mlxsw_pci,
+				   struct mlxsw_pci_queue_elem_info *elem_info)
+{
+	struct sk_buff *skb;
+	char *wqe;
+
+	skb = elem_info->u.rdq.skb;
+	wqe = elem_info->elem;
+
+	mlxsw_pci_wqe_frag_unmap(mlxsw_pci, wqe, 0, DMA_FROM_DEVICE);
+	dev_kfree_skb_any(skb);
+}
+
+static int mlxsw_pci_rdq_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+			      struct mlxsw_pci_queue *q)
+{
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	u8 sdq_count = mlxsw_pci_sdq_count(mlxsw_pci);
+	int i;
+	int err;
+
+	q->producer_counter = 0;
+	q->consumer_counter = 0;
+
+	/* Set CQ of same number of this RDQ with base
+	 * above SDQ count as the lower ones are assigned to SDQs.
+	 */
+	mlxsw_cmd_mbox_sw2hw_dq_cq_set(mbox, sdq_count + q->num);
+	mlxsw_cmd_mbox_sw2hw_dq_log2_dq_sz_set(mbox, 3); /* 8 pages */
+	for (i = 0; i < MLXSW_PCI_AQ_PAGES; i++) {
+		dma_addr_t mapaddr = __mlxsw_pci_queue_page_get(q, i);
+
+		mlxsw_cmd_mbox_sw2hw_dq_pa_set(mbox, i, mapaddr);
+	}
+
+	err = mlxsw_cmd_sw2hw_rdq(mlxsw_pci->core, mbox, q->num);
+	if (err)
+		return err;
+
+	mlxsw_pci_queue_doorbell_producer_ring(mlxsw_pci, q);
+
+	for (i = 0; i < q->count; i++) {
+		elem_info = mlxsw_pci_queue_elem_info_producer_get(q);
+		BUG_ON(!elem_info);
+		err = mlxsw_pci_rdq_skb_alloc(mlxsw_pci, elem_info);
+		if (err)
+			goto rollback;
+		/* Everything is set up, ring doorbell to pass elem to HW */
+		q->producer_counter++;
+		mlxsw_pci_queue_doorbell_producer_ring(mlxsw_pci, q);
+	}
+
+	return 0;
+
+rollback:
+	for (i--; i >= 0; i--) {
+		elem_info = mlxsw_pci_queue_elem_info_get(q, i);
+		mlxsw_pci_rdq_skb_free(mlxsw_pci, elem_info);
+	}
+	mlxsw_cmd_hw2sw_rdq(mlxsw_pci->core, q->num);
+
+	return err;
+}
+
+static void mlxsw_pci_rdq_fini(struct mlxsw_pci *mlxsw_pci,
+			       struct mlxsw_pci_queue *q)
+{
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	int i;
+
+	mlxsw_cmd_hw2sw_rdq(mlxsw_pci->core, q->num);
+	for (i = 0; i < q->count; i++) {
+		elem_info = mlxsw_pci_queue_elem_info_get(q, i);
+		mlxsw_pci_rdq_skb_free(mlxsw_pci, elem_info);
+	}
+}
+
+static int mlxsw_pci_rdq_dbg_read(struct seq_file *file, void *data)
+{
+	struct mlxsw_pci *mlxsw_pci = dev_get_drvdata(file->private);
+	struct mlxsw_pci_queue *q;
+	int i;
+	static const char hdr[] =
+		"NUM PROD_COUNT CONS_COUNT COUNT\n";
+
+	seq_printf(file, hdr);
+	for (i = 0; i < mlxsw_pci_rdq_count(mlxsw_pci); i++) {
+		q = mlxsw_pci_rdq_get(mlxsw_pci, i);
+		spin_lock_bh(&q->lock);
+		seq_printf(file, "%3d %10d %10d %5d\n",
+			   i, q->producer_counter, q->consumer_counter,
+			   q->count);
+		spin_unlock_bh(&q->lock);
+	}
+	return 0;
+}
+
+static int mlxsw_pci_cq_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+			     struct mlxsw_pci_queue *q)
+{
+	int i;
+	int err;
+
+	q->consumer_counter = 0;
+
+	for (i = 0; i < q->count; i++) {
+		char *elem = mlxsw_pci_queue_elem_get(q, i);
+
+		mlxsw_pci_cqe_owner_set(elem, 1);
+	}
+
+	mlxsw_cmd_mbox_sw2hw_cq_cv_set(mbox, 0); /* CQE ver 0 */
+	mlxsw_cmd_mbox_sw2hw_cq_c_eqn_set(mbox, MLXSW_PCI_EQ_COMP_NUM);
+	mlxsw_cmd_mbox_sw2hw_cq_oi_set(mbox, 0);
+	mlxsw_cmd_mbox_sw2hw_cq_st_set(mbox, 0);
+	mlxsw_cmd_mbox_sw2hw_cq_log_cq_size_set(mbox, ilog2(q->count));
+	for (i = 0; i < MLXSW_PCI_AQ_PAGES; i++) {
+		dma_addr_t mapaddr = __mlxsw_pci_queue_page_get(q, i);
+
+		mlxsw_cmd_mbox_sw2hw_cq_pa_set(mbox, i, mapaddr);
+	}
+	err = mlxsw_cmd_sw2hw_cq(mlxsw_pci->core, mbox, q->num);
+	if (err)
+		return err;
+	mlxsw_pci_queue_doorbell_consumer_ring(mlxsw_pci, q);
+	mlxsw_pci_queue_doorbell_arm_consumer_ring(mlxsw_pci, q);
+	return 0;
+}
+
+static void mlxsw_pci_cq_fini(struct mlxsw_pci *mlxsw_pci,
+			      struct mlxsw_pci_queue *q)
+{
+	mlxsw_cmd_hw2sw_cq(mlxsw_pci->core, q->num);
+}
+
+static int mlxsw_pci_cq_dbg_read(struct seq_file *file, void *data)
+{
+	struct mlxsw_pci *mlxsw_pci = dev_get_drvdata(file->private);
+
+	struct mlxsw_pci_queue *q;
+	int i;
+	static const char hdr[] =
+		"NUM CONS_INDEX  SDQ_COUNT  RDQ_COUNT COUNT\n";
+
+	seq_printf(file, hdr);
+	for (i = 0; i < mlxsw_pci_cq_count(mlxsw_pci); i++) {
+		q = mlxsw_pci_cq_get(mlxsw_pci, i);
+		spin_lock_bh(&q->lock);
+		seq_printf(file, "%3d %10d %10d %10d %5d\n",
+			   i, q->consumer_counter, q->u.cq.comp_sdq_count,
+			   q->u.cq.comp_rdq_count, q->count);
+		spin_unlock_bh(&q->lock);
+	}
+	return 0;
+}
+
+static void mlxsw_pci_cqe_sdq_handle(struct mlxsw_pci *mlxsw_pci,
+				     struct mlxsw_pci_queue *q,
+				     u16 consumer_counter_limit,
+				     char *cqe)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	char *wqe;
+	struct sk_buff *skb;
+	int i;
+
+	spin_lock(&q->lock);
+	elem_info = mlxsw_pci_queue_elem_info_consumer_get(q);
+	skb = elem_info->u.sdq.skb;
+	wqe = elem_info->elem;
+	for (i = 0; i < MLXSW_PCI_WQE_SG_ENTRIES; i++)
+		mlxsw_pci_wqe_frag_unmap(mlxsw_pci, wqe, i, DMA_TO_DEVICE);
+	dev_kfree_skb_any(skb);
+	elem_info->u.sdq.skb = NULL;
+
+	if (q->consumer_counter++ != consumer_counter_limit)
+		dev_dbg_ratelimited(&pdev->dev, "Consumer counter does not match limit in SDQ\n");
+	spin_unlock(&q->lock);
+}
+
+static void mlxsw_pci_cqe_rdq_handle(struct mlxsw_pci *mlxsw_pci,
+				     struct mlxsw_pci_queue *q,
+				     u16 consumer_counter_limit,
+				     char *cqe)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	char *wqe;
+	struct sk_buff *skb;
+	struct mlxsw_rx_info rx_info;
+	u16 byte_count;
+	int err;
+
+	elem_info = mlxsw_pci_queue_elem_info_consumer_get(q);
+	skb = elem_info->u.sdq.skb;
+	if (!skb)
+		return;
+	wqe = elem_info->elem;
+	mlxsw_pci_wqe_frag_unmap(mlxsw_pci, wqe, 0, DMA_FROM_DEVICE);
+
+	if (q->consumer_counter++ != consumer_counter_limit)
+		dev_dbg_ratelimited(&pdev->dev, "Consumer counter does not match limit in RDQ\n");
+
+	/* We do not support lag now */
+	if (mlxsw_pci_cqe_lag_get(cqe))
+		goto drop;
+
+	rx_info.sys_port = mlxsw_pci_cqe_system_port_get(cqe);
+	rx_info.trap_id = mlxsw_pci_cqe_trap_id_get(cqe);
+
+	byte_count = mlxsw_pci_cqe_byte_count_get(cqe);
+	if (mlxsw_pci_cqe_crc_get(cqe))
+		byte_count -= ETH_FCS_LEN;
+	skb_put(skb, byte_count);
+	mlxsw_core_skb_receive(mlxsw_pci->core, skb, &rx_info);
+
+put_new_skb:
+	memset(wqe, 0, q->elem_size);
+	err = mlxsw_pci_rdq_skb_alloc(mlxsw_pci, elem_info);
+	if (err)
+		dev_dbg_ratelimited(&pdev->dev, "Failed to alloc skb for RDQ\n");
+	/* Everything is set up, ring doorbell to pass elem to HW */
+	q->producer_counter++;
+	mlxsw_pci_queue_doorbell_producer_ring(mlxsw_pci, q);
+	return;
+
+drop:
+	dev_kfree_skb_any(skb);
+	goto put_new_skb;
+}
+
+static char *mlxsw_pci_cq_sw_cqe_get(struct mlxsw_pci_queue *q)
+{
+	return mlxsw_pci_queue_sw_elem_get(q, mlxsw_pci_cqe_owner_get);
+}
+
+static void mlxsw_pci_cq_tasklet(unsigned long data)
+{
+	struct mlxsw_pci_queue *q = (struct mlxsw_pci_queue *) data;
+	struct mlxsw_pci *mlxsw_pci = q->pci;
+	char *cqe;
+	int items = 0;
+	int credits = q->count >> 1;
+
+	while ((cqe = mlxsw_pci_cq_sw_cqe_get(q))) {
+		u16 wqe_counter = mlxsw_pci_cqe_wqe_counter_get(cqe);
+		u8 sendq = mlxsw_pci_cqe_sr_get(cqe);
+		u8 dqn = mlxsw_pci_cqe_dqn_get(cqe);
+
+		if (sendq) {
+			struct mlxsw_pci_queue *sdq;
+
+			sdq = mlxsw_pci_sdq_get(mlxsw_pci, dqn);
+			mlxsw_pci_cqe_sdq_handle(mlxsw_pci, sdq,
+						 wqe_counter, cqe);
+			q->u.cq.comp_sdq_count++;
+		} else {
+			struct mlxsw_pci_queue *rdq;
+
+			rdq = mlxsw_pci_rdq_get(mlxsw_pci, dqn);
+			mlxsw_pci_cqe_rdq_handle(mlxsw_pci, rdq,
+						 wqe_counter, cqe);
+			q->u.cq.comp_rdq_count++;
+		}
+		if (++items == credits)
+			break;
+	}
+	if (items) {
+		mlxsw_pci_queue_doorbell_consumer_ring(mlxsw_pci, q);
+		mlxsw_pci_queue_doorbell_arm_consumer_ring(mlxsw_pci, q);
+	}
+}
+
+static int mlxsw_pci_eq_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+			     struct mlxsw_pci_queue *q)
+{
+	int i;
+	int err;
+
+	q->consumer_counter = 0;
+
+	for (i = 0; i < q->count; i++) {
+		char *elem = mlxsw_pci_queue_elem_get(q, i);
+
+		mlxsw_pci_eqe_owner_set(elem, 1);
+	}
+
+	mlxsw_cmd_mbox_sw2hw_eq_int_msix_set(mbox, 1); /* MSI-X used */
+	mlxsw_cmd_mbox_sw2hw_eq_oi_set(mbox, 0);
+	mlxsw_cmd_mbox_sw2hw_eq_st_set(mbox, 1); /* armed */
+	mlxsw_cmd_mbox_sw2hw_eq_log_eq_size_set(mbox, ilog2(q->count));
+	for (i = 0; i < MLXSW_PCI_AQ_PAGES; i++) {
+		dma_addr_t mapaddr = __mlxsw_pci_queue_page_get(q, i);
+
+		mlxsw_cmd_mbox_sw2hw_eq_pa_set(mbox, i, mapaddr);
+	}
+	err = mlxsw_cmd_sw2hw_eq(mlxsw_pci->core, mbox, q->num);
+	if (err)
+		return err;
+	mlxsw_pci_queue_doorbell_consumer_ring(mlxsw_pci, q);
+	mlxsw_pci_queue_doorbell_arm_consumer_ring(mlxsw_pci, q);
+	return 0;
+}
+
+static void mlxsw_pci_eq_fini(struct mlxsw_pci *mlxsw_pci,
+			      struct mlxsw_pci_queue *q)
+{
+	mlxsw_cmd_hw2sw_eq(mlxsw_pci->core, q->num);
+}
+
+static int mlxsw_pci_eq_dbg_read(struct seq_file *file, void *data)
+{
+	struct mlxsw_pci *mlxsw_pci = dev_get_drvdata(file->private);
+	struct mlxsw_pci_queue *q;
+	int i;
+	static const char hdr[] =
+		"NUM CONS_COUNT     EV_CMD    EV_COMP   EV_OTHER COUNT\n";
+
+	seq_printf(file, hdr);
+	for (i = 0; i < mlxsw_pci_eq_count(mlxsw_pci); i++) {
+		q = mlxsw_pci_eq_get(mlxsw_pci, i);
+		spin_lock_bh(&q->lock);
+		seq_printf(file, "%3d %10d %10d %10d %10d %5d\n",
+			   i, q->consumer_counter, q->u.eq.ev_cmd_count,
+			   q->u.eq.ev_comp_count, q->u.eq.ev_other_count,
+			   q->count);
+		spin_unlock_bh(&q->lock);
+	}
+	return 0;
+}
+
+static void mlxsw_pci_eq_cmd_event(struct mlxsw_pci *mlxsw_pci, char *eqe)
+{
+	mlxsw_pci->cmd.comp.status = mlxsw_pci_eqe_cmd_status_get(eqe);
+	mlxsw_pci->cmd.comp.out_param =
+		((u64) mlxsw_pci_eqe_cmd_out_param_h_get(eqe)) << 32 |
+		mlxsw_pci_eqe_cmd_out_param_l_get(eqe);
+	mlxsw_pci->cmd.wait_done = true;
+	wake_up(&mlxsw_pci->cmd.wait);
+}
+
+static char *mlxsw_pci_eq_sw_eqe_get(struct mlxsw_pci_queue *q)
+{
+	return mlxsw_pci_queue_sw_elem_get(q, mlxsw_pci_eqe_owner_get);
+}
+
+static void mlxsw_pci_eq_tasklet(unsigned long data)
+{
+	struct mlxsw_pci_queue *q = (struct mlxsw_pci_queue *) data;
+	struct mlxsw_pci *mlxsw_pci = q->pci;
+	u8 cq_count = mlxsw_pci_cq_count(mlxsw_pci);
+	unsigned long active_cqns[BITS_TO_LONGS(MLXSW_PCI_CQS_MAX)];
+	char *eqe;
+	u8 cqn;
+	bool cq_handle = false;
+	int items = 0;
+	int credits = q->count >> 1;
+
+	memset(&active_cqns, 0, sizeof(active_cqns));
+
+	while ((eqe = mlxsw_pci_eq_sw_eqe_get(q))) {
+		u8 event_type = mlxsw_pci_eqe_event_type_get(eqe);
+
+		switch (event_type) {
+		case MLXSW_PCI_EQE_EVENT_TYPE_CMD:
+			mlxsw_pci_eq_cmd_event(mlxsw_pci, eqe);
+			q->u.eq.ev_cmd_count++;
+			break;
+		case MLXSW_PCI_EQE_EVENT_TYPE_COMP:
+			cqn = mlxsw_pci_eqe_cqn_get(eqe);
+			set_bit(cqn, active_cqns);
+			cq_handle = true;
+			q->u.eq.ev_comp_count++;
+			break;
+		default:
+			q->u.eq.ev_other_count++;
+		}
+		if (++items == credits)
+			break;
+	}
+	if (items) {
+		mlxsw_pci_queue_doorbell_consumer_ring(mlxsw_pci, q);
+		mlxsw_pci_queue_doorbell_arm_consumer_ring(mlxsw_pci, q);
+	}
+
+	if (!cq_handle)
+		return;
+	for_each_set_bit(cqn, active_cqns, cq_count) {
+		q = mlxsw_pci_cq_get(mlxsw_pci, cqn);
+		mlxsw_pci_queue_tasklet_schedule(q);
+	}
+}
+
+struct mlxsw_pci_queue_ops {
+	const char *name;
+	enum mlxsw_pci_queue_type type;
+	int (*init)(struct mlxsw_pci *mlxsw_pci, char *mbox,
+		    struct mlxsw_pci_queue *q);
+	void (*fini)(struct mlxsw_pci *mlxsw_pci,
+		     struct mlxsw_pci_queue *q);
+	void (*tasklet)(unsigned long data);
+	int (*dbg_read)(struct seq_file *s, void *data);
+	u16 elem_count;
+	u8 elem_size;
+};
+
+static const struct mlxsw_pci_queue_ops mlxsw_pci_sdq_ops = {
+	.type		= MLXSW_PCI_QUEUE_TYPE_SDQ,
+	.init		= mlxsw_pci_sdq_init,
+	.fini		= mlxsw_pci_sdq_fini,
+	.dbg_read	= mlxsw_pci_sdq_dbg_read,
+	.elem_count	= MLXSW_PCI_WQE_COUNT,
+	.elem_size	= MLXSW_PCI_WQE_SIZE,
+};
+
+static const struct mlxsw_pci_queue_ops mlxsw_pci_rdq_ops = {
+	.type		= MLXSW_PCI_QUEUE_TYPE_RDQ,
+	.init		= mlxsw_pci_rdq_init,
+	.fini		= mlxsw_pci_rdq_fini,
+	.dbg_read	= mlxsw_pci_rdq_dbg_read,
+	.elem_count	= MLXSW_PCI_WQE_COUNT,
+	.elem_size	= MLXSW_PCI_WQE_SIZE
+};
+
+static const struct mlxsw_pci_queue_ops mlxsw_pci_cq_ops = {
+	.type		= MLXSW_PCI_QUEUE_TYPE_CQ,
+	.init		= mlxsw_pci_cq_init,
+	.fini		= mlxsw_pci_cq_fini,
+	.tasklet	= mlxsw_pci_cq_tasklet,
+	.dbg_read	= mlxsw_pci_cq_dbg_read,
+	.elem_count	= MLXSW_PCI_CQE_COUNT,
+	.elem_size	= MLXSW_PCI_CQE_SIZE
+};
+
+static const struct mlxsw_pci_queue_ops mlxsw_pci_eq_ops = {
+	.type		= MLXSW_PCI_QUEUE_TYPE_EQ,
+	.init		= mlxsw_pci_eq_init,
+	.fini		= mlxsw_pci_eq_fini,
+	.tasklet	= mlxsw_pci_eq_tasklet,
+	.dbg_read	= mlxsw_pci_eq_dbg_read,
+	.elem_count	= MLXSW_PCI_EQE_COUNT,
+	.elem_size	= MLXSW_PCI_EQE_SIZE
+};
+
+static int mlxsw_pci_queue_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+				const struct mlxsw_pci_queue_ops *q_ops,
+				struct mlxsw_pci_queue *q, u8 q_num)
+{
+	struct mlxsw_pci_mem_item *mem_item = &q->mem_item;
+	int i;
+	int err;
+
+	spin_lock_init(&q->lock);
+	q->num = q_num;
+	q->count = q_ops->elem_count;
+	q->elem_size = q_ops->elem_size;
+	q->type = q_ops->type;
+	q->pci = mlxsw_pci;
+
+	if (q_ops->tasklet)
+		tasklet_init(&q->tasklet, q_ops->tasklet, (unsigned long) q);
+
+	mem_item->size = MLXSW_PCI_AQ_SIZE;
+	mem_item->buf = pci_alloc_consistent(mlxsw_pci->pdev,
+					     mem_item->size,
+					     &mem_item->mapaddr);
+	if (!mem_item->buf)
+		return -ENOMEM;
+	memset(mem_item->buf, 0, mem_item->size);
+
+	q->elem_info = kcalloc(q->count, sizeof(*q->elem_info), GFP_KERNEL);
+	if (!q->elem_info) {
+		err = -ENOMEM;
+		goto err_elem_info_alloc;
+	}
+
+	/* Initialize dma mapped elements info elem_info for
+	 * future easy access.
+	 */
+	for (i = 0; i < q->count; i++) {
+		struct mlxsw_pci_queue_elem_info *elem_info;
+
+		elem_info = mlxsw_pci_queue_elem_info_get(q, i);
+		elem_info->elem =
+			__mlxsw_pci_queue_elem_get(q, q_ops->elem_size, i);
+	}
+
+	mlxsw_cmd_mbox_zero(mbox);
+	err = q_ops->init(mlxsw_pci, mbox, q);
+	if (err)
+		goto err_q_ops_init;
+	return 0;
+
+err_q_ops_init:
+	kfree(q->elem_info);
+err_elem_info_alloc:
+	pci_free_consistent(mlxsw_pci->pdev, mem_item->size,
+			    mem_item->buf, mem_item->mapaddr);
+	return err;
+}
+
+static void mlxsw_pci_queue_fini(struct mlxsw_pci *mlxsw_pci,
+				 const struct mlxsw_pci_queue_ops *q_ops,
+				 struct mlxsw_pci_queue *q)
+{
+	struct mlxsw_pci_mem_item *mem_item = &q->mem_item;
+
+	q_ops->fini(mlxsw_pci, q);
+	kfree(q->elem_info);
+	pci_free_consistent(mlxsw_pci->pdev, mem_item->size,
+			    mem_item->buf, mem_item->mapaddr);
+}
+
+static int mlxsw_pci_queue_group_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+				      const struct mlxsw_pci_queue_ops *q_ops,
+				      u8 num_qs)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	struct mlxsw_pci_queue_type_group *queue_group;
+	char tmp[16];
+	int i;
+	int err;
+
+	queue_group = mlxsw_pci_queue_type_group_get(mlxsw_pci, q_ops->type);
+	queue_group->q = kcalloc(num_qs, sizeof(*queue_group->q), GFP_KERNEL);
+	if (!queue_group->q)
+		return -ENOMEM;
+
+	for (i = 0; i < num_qs; i++) {
+		err = mlxsw_pci_queue_init(mlxsw_pci, mbox, q_ops,
+					   &queue_group->q[i], i);
+		if (err)
+			goto err_queue_init;
+	}
+	queue_group->count = num_qs;
+
+	sprintf(tmp, "%s_stats", mlxsw_pci_queue_type_str(q_ops->type));
+	debugfs_create_devm_seqfile(&pdev->dev, tmp, mlxsw_pci->dbg_dir,
+				    q_ops->dbg_read);
+
+	return 0;
+
+err_queue_init:
+	for (i--; i >= 0; i--)
+		mlxsw_pci_queue_fini(mlxsw_pci, q_ops, &queue_group->q[i]);
+	kfree(queue_group->q);
+	return err;
+}
+
+static void mlxsw_pci_queue_group_fini(struct mlxsw_pci *mlxsw_pci,
+				       const struct mlxsw_pci_queue_ops *q_ops)
+{
+	struct mlxsw_pci_queue_type_group *queue_group;
+	int i;
+
+	queue_group = mlxsw_pci_queue_type_group_get(mlxsw_pci, q_ops->type);
+	for (i = 0; i < queue_group->count; i++)
+		mlxsw_pci_queue_fini(mlxsw_pci, q_ops, &queue_group->q[i]);
+	kfree(queue_group->q);
+}
+
+static int mlxsw_pci_aqs_init(struct mlxsw_pci *mlxsw_pci, char *mbox)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	u8 num_sdqs;
+	u8 sdq_log2sz;
+	u8 num_rdqs;
+	u8 rdq_log2sz;
+	u8 num_cqs;
+	u8 cq_log2sz;
+	u8 num_eqs;
+	u8 eq_log2sz;
+	int err;
+
+	mlxsw_cmd_mbox_zero(mbox);
+	err = mlxsw_cmd_query_aq_cap(mlxsw_pci->core, mbox);
+	if (err)
+		return err;
+
+	num_sdqs = mlxsw_cmd_mbox_query_aq_cap_max_num_sdqs_get(mbox);
+	sdq_log2sz = mlxsw_cmd_mbox_query_aq_cap_log_max_sdq_sz_get(mbox);
+	num_rdqs = mlxsw_cmd_mbox_query_aq_cap_max_num_rdqs_get(mbox);
+	rdq_log2sz = mlxsw_cmd_mbox_query_aq_cap_log_max_rdq_sz_get(mbox);
+	num_cqs = mlxsw_cmd_mbox_query_aq_cap_max_num_cqs_get(mbox);
+	cq_log2sz = mlxsw_cmd_mbox_query_aq_cap_log_max_cq_sz_get(mbox);
+	num_eqs = mlxsw_cmd_mbox_query_aq_cap_max_num_eqs_get(mbox);
+	eq_log2sz = mlxsw_cmd_mbox_query_aq_cap_log_max_eq_sz_get(mbox);
+
+	if (num_sdqs + num_rdqs > num_cqs ||
+	    num_cqs > MLXSW_PCI_CQS_MAX || num_eqs != MLXSW_PCI_EQS_COUNT) {
+		dev_err(&pdev->dev, "Unsupported number of queues\n");
+		return -EINVAL;
+	}
+
+	if ((1 << sdq_log2sz != MLXSW_PCI_WQE_COUNT) ||
+	    (1 << rdq_log2sz != MLXSW_PCI_WQE_COUNT) ||
+	    (1 << cq_log2sz != MLXSW_PCI_CQE_COUNT) ||
+	    (1 << eq_log2sz != MLXSW_PCI_EQE_COUNT)) {
+		dev_err(&pdev->dev, "Unsupported number of async queue descriptors\n");
+		return -EINVAL;
+	}
+
+	err = mlxsw_pci_queue_group_init(mlxsw_pci, mbox, &mlxsw_pci_eq_ops,
+					 num_eqs);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to initialize event queues\n");
+		return err;
+	}
+
+	err = mlxsw_pci_queue_group_init(mlxsw_pci, mbox, &mlxsw_pci_cq_ops,
+					 num_cqs);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to initialize completion queues\n");
+		goto err_cqs_init;
+	}
+
+	err = mlxsw_pci_queue_group_init(mlxsw_pci, mbox, &mlxsw_pci_sdq_ops,
+					 num_sdqs);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to initialize send descriptor queues\n");
+		goto err_sdqs_init;
+	}
+
+	err = mlxsw_pci_queue_group_init(mlxsw_pci, mbox, &mlxsw_pci_rdq_ops,
+					 num_rdqs);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to initialize receive descriptor queues\n");
+		goto err_rdqs_init;
+	}
+
+	/* We have to poll in command interface until queues are initialized */
+	mlxsw_pci->cmd.nopoll = true;
+	return 0;
+
+err_rdqs_init:
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_sdq_ops);
+err_sdqs_init:
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_cq_ops);
+err_cqs_init:
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_eq_ops);
+	return err;
+}
+
+static void mlxsw_pci_aqs_fini(struct mlxsw_pci *mlxsw_pci)
+{
+	mlxsw_pci->cmd.nopoll = false;
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_rdq_ops);
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_sdq_ops);
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_cq_ops);
+	mlxsw_pci_queue_group_fini(mlxsw_pci, &mlxsw_pci_eq_ops);
+}
+
+static void
+mlxsw_pci_config_profile_swid_config(struct mlxsw_pci *mlxsw_pci,
+				     char *mbox, int index,
+				     const struct mlxsw_swid_config *swid)
+{
+	u8 mask = 0;
+
+	if (swid->used_type) {
+		mlxsw_cmd_mbox_config_profile_swid_config_type_set(
+			mbox, index, swid->type);
+		mask |= 1;
+	}
+	if (swid->used_properties) {
+		mlxsw_cmd_mbox_config_profile_swid_config_properties_set(
+			mbox, index, swid->properties);
+		mask |= 2;
+	}
+	mlxsw_cmd_mbox_config_profile_swid_config_mask_set(mbox, index, mask);
+}
+
+static int mlxsw_pci_config_profile(struct mlxsw_pci *mlxsw_pci, char *mbox,
+				    const struct mlxsw_config_profile *profile)
+{
+	int i;
+
+	mlxsw_cmd_mbox_zero(mbox);
+
+	if (profile->used_max_vepa_channels) {
+		mlxsw_cmd_mbox_config_profile_set_max_vepa_channels_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_vepa_channels_set(
+			mbox, profile->max_vepa_channels);
+	}
+	if (profile->used_max_lag) {
+		mlxsw_cmd_mbox_config_profile_set_max_lag_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_lag_set(
+			mbox, profile->max_lag);
+	}
+	if (profile->used_max_port_per_lag) {
+		mlxsw_cmd_mbox_config_profile_set_max_port_per_lag_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_port_per_lag_set(
+			mbox, profile->max_port_per_lag);
+	}
+	if (profile->used_max_mid) {
+		mlxsw_cmd_mbox_config_profile_set_max_mid_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_mid_set(
+			mbox, profile->max_mid);
+	}
+	if (profile->used_max_pgt) {
+		mlxsw_cmd_mbox_config_profile_set_max_pgt_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_pgt_set(
+			mbox, profile->max_pgt);
+	}
+	if (profile->used_max_system_port) {
+		mlxsw_cmd_mbox_config_profile_set_max_system_port_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_system_port_set(
+			mbox, profile->max_system_port);
+	}
+	if (profile->used_max_vlan_groups) {
+		mlxsw_cmd_mbox_config_profile_set_max_vlan_groups_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_vlan_groups_set(
+			mbox, profile->max_vlan_groups);
+	}
+	if (profile->used_max_regions) {
+		mlxsw_cmd_mbox_config_profile_set_max_regions_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_regions_set(
+			mbox, profile->max_regions);
+	}
+	if (profile->used_flood_tables) {
+		mlxsw_cmd_mbox_config_profile_set_flood_tables_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_flood_tables_set(
+			mbox, profile->max_flood_tables);
+		mlxsw_cmd_mbox_config_profile_max_vid_flood_tables_set(
+			mbox, profile->max_vid_flood_tables);
+		mlxsw_cmd_mbox_config_profile_max_fid_offset_flood_tables_set(
+			mbox, profile->max_fid_offset_flood_tables);
+		mlxsw_cmd_mbox_config_profile_fid_offset_flood_table_size_set(
+			mbox, profile->fid_offset_flood_table_size);
+		mlxsw_cmd_mbox_config_profile_max_fid_flood_tables_set(
+			mbox, profile->max_fid_flood_tables);
+		mlxsw_cmd_mbox_config_profile_fid_flood_table_size_set(
+			mbox, profile->fid_flood_table_size);
+	}
+	if (profile->used_flood_mode) {
+		mlxsw_cmd_mbox_config_profile_set_flood_mode_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_flood_mode_set(
+			mbox, profile->flood_mode);
+	}
+	if (profile->used_max_ib_mc) {
+		mlxsw_cmd_mbox_config_profile_set_max_ib_mc_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_ib_mc_set(
+			mbox, profile->max_ib_mc);
+	}
+	if (profile->used_max_pkey) {
+		mlxsw_cmd_mbox_config_profile_set_max_pkey_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_max_pkey_set(
+			mbox, profile->max_pkey);
+	}
+	if (profile->used_ar_sec) {
+		mlxsw_cmd_mbox_config_profile_set_ar_sec_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_ar_sec_set(
+			mbox, profile->ar_sec);
+	}
+	if (profile->used_adaptive_routing_group_cap) {
+		mlxsw_cmd_mbox_config_profile_set_adaptive_routing_group_cap_set(
+			mbox, 1);
+		mlxsw_cmd_mbox_config_profile_adaptive_routing_group_cap_set(
+			mbox, profile->adaptive_routing_group_cap);
+	}
+
+	for (i = 0; i < MLXSW_CONFIG_PROFILE_SWID_COUNT; i++)
+		mlxsw_pci_config_profile_swid_config(mlxsw_pci, mbox, i,
+						     &profile->swid_config[i]);
+
+	return mlxsw_cmd_config_profile_set(mlxsw_pci->core, mbox);
+}
+
+static int mlxsw_pci_boardinfo(struct mlxsw_pci *mlxsw_pci, char *mbox)
+{
+	struct mlxsw_bus_info *bus_info = &mlxsw_pci->bus_info;
+	int err;
+
+	mlxsw_cmd_mbox_zero(mbox);
+	err = mlxsw_cmd_boardinfo(mlxsw_pci->core, mbox);
+	if (err)
+		return err;
+	mlxsw_cmd_mbox_boardinfo_vsd_memcpy_from(mbox, bus_info->vsd);
+	mlxsw_cmd_mbox_boardinfo_psid_memcpy_from(mbox, bus_info->psid);
+	return 0;
+}
+
+static int mlxsw_pci_fw_area_init(struct mlxsw_pci *mlxsw_pci, char *mbox,
+				  u16 num_pages)
+{
+	struct mlxsw_pci_mem_item *mem_item;
+	int nent = 0;
+	int i;
+	int err;
+
+	mlxsw_pci->fw_area.items = kcalloc(num_pages, sizeof(*mem_item),
+					   GFP_KERNEL);
+	if (!mlxsw_pci->fw_area.items)
+		return -ENOMEM;
+	mlxsw_pci->fw_area.count = num_pages;
+
+	mlxsw_cmd_mbox_zero(mbox);
+	for (i = 0; i < num_pages; i++) {
+		mem_item = &mlxsw_pci->fw_area.items[i];
+
+		mem_item->size = MLXSW_PCI_PAGE_SIZE;
+		mem_item->buf = pci_alloc_consistent(mlxsw_pci->pdev,
+						     mem_item->size,
+						     &mem_item->mapaddr);
+		if (!mem_item->buf) {
+			err = -ENOMEM;
+			goto err_alloc;
+		}
+		mlxsw_cmd_mbox_map_fa_pa_set(mbox, nent, mem_item->mapaddr);
+		mlxsw_cmd_mbox_map_fa_log2size_set(mbox, nent, 0); /* 1 page */
+		if (++nent == MLXSW_CMD_MAP_FA_VPM_ENTRIES_MAX) {
+			err = mlxsw_cmd_map_fa(mlxsw_pci->core, mbox, nent);
+			if (err)
+				goto err_cmd_map_fa;
+			nent = 0;
+			mlxsw_cmd_mbox_zero(mbox);
+		}
+	}
+
+	if (nent) {
+		err = mlxsw_cmd_map_fa(mlxsw_pci->core, mbox, nent);
+		if (err)
+			goto err_cmd_map_fa;
+	}
+
+	return 0;
+
+err_cmd_map_fa:
+err_alloc:
+	for (i--; i >= 0; i--) {
+		mem_item = &mlxsw_pci->fw_area.items[i];
+
+		pci_free_consistent(mlxsw_pci->pdev, mem_item->size,
+				    mem_item->buf, mem_item->mapaddr);
+	}
+	kfree(mlxsw_pci->fw_area.items);
+	return err;
+}
+
+static void mlxsw_pci_fw_area_fini(struct mlxsw_pci *mlxsw_pci)
+{
+	struct mlxsw_pci_mem_item *mem_item;
+	int i;
+
+	mlxsw_cmd_unmap_fa(mlxsw_pci->core);
+
+	for (i = 0; i < mlxsw_pci->fw_area.count; i++) {
+		mem_item = &mlxsw_pci->fw_area.items[i];
+
+		pci_free_consistent(mlxsw_pci->pdev, mem_item->size,
+				    mem_item->buf, mem_item->mapaddr);
+	}
+	kfree(mlxsw_pci->fw_area.items);
+}
+
+static irqreturn_t mlxsw_pci_eq_irq_handler(int irq, void *dev_id)
+{
+	struct mlxsw_pci *mlxsw_pci = dev_id;
+	struct mlxsw_pci_queue *q;
+	int i;
+
+	for (i = 0; i < MLXSW_PCI_EQS_COUNT; i++) {
+		q = mlxsw_pci_eq_get(mlxsw_pci, i);
+		mlxsw_pci_queue_tasklet_schedule(q);
+	}
+	return IRQ_HANDLED;
+}
+
+static int mlxsw_pci_mbox_alloc(struct mlxsw_pci *mlxsw_pci,
+				struct mlxsw_pci_mem_item *mbox)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	int err = 0;
+
+	mbox->size = MLXSW_CMD_MBOX_SIZE;
+	mbox->buf = pci_alloc_consistent(pdev, MLXSW_CMD_MBOX_SIZE,
+					 &mbox->mapaddr);
+	if (!mbox->buf) {
+		dev_err(&pdev->dev, "Failed allocating memory for mailbox\n");
+		err = -ENOMEM;
+	}
+
+	return err;
+}
+
+static void mlxsw_pci_mbox_free(struct mlxsw_pci *mlxsw_pci,
+				struct mlxsw_pci_mem_item *mbox)
+{
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+
+	pci_free_consistent(pdev, MLXSW_CMD_MBOX_SIZE, mbox->buf,
+			    mbox->mapaddr);
+}
+
+static int mlxsw_pci_init(void *bus_priv, struct mlxsw_core *mlxsw_core,
+			  const struct mlxsw_config_profile *profile)
+{
+	struct mlxsw_pci *mlxsw_pci = bus_priv;
+	struct pci_dev *pdev = mlxsw_pci->pdev;
+	char *mbox;
+	u16 num_pages;
+	int err;
+
+	mutex_init(&mlxsw_pci->cmd.lock);
+	init_waitqueue_head(&mlxsw_pci->cmd.wait);
+
+	mlxsw_pci->core = mlxsw_core;
+
+	mbox = mlxsw_cmd_mbox_alloc();
+	if (!mbox)
+		return -ENOMEM;
+
+	err = mlxsw_pci_mbox_alloc(mlxsw_pci, &mlxsw_pci->cmd.in_mbox);
+	if (err)
+		goto mbox_put;
+
+	err = mlxsw_pci_mbox_alloc(mlxsw_pci, &mlxsw_pci->cmd.out_mbox);
+	if (err)
+		goto err_out_mbox_alloc;
+
+	err = mlxsw_cmd_query_fw(mlxsw_core, mbox);
+	if (err)
+		goto err_query_fw;
+
+	mlxsw_pci->bus_info.fw_rev.major =
+		mlxsw_cmd_mbox_query_fw_fw_rev_major_get(mbox);
+	mlxsw_pci->bus_info.fw_rev.minor =
+		mlxsw_cmd_mbox_query_fw_fw_rev_minor_get(mbox);
+	mlxsw_pci->bus_info.fw_rev.subminor =
+		mlxsw_cmd_mbox_query_fw_fw_rev_subminor_get(mbox);
+
+	if (mlxsw_cmd_mbox_query_fw_cmd_interface_rev_get(mbox) != 1) {
+		dev_err(&pdev->dev, "Unsupported cmd interface revision ID queried from hw\n");
+		err = -EINVAL;
+		goto err_iface_rev;
+	}
+	if (mlxsw_cmd_mbox_query_fw_doorbell_page_bar_get(mbox) != 0) {
+		dev_err(&pdev->dev, "Unsupported doorbell page bar queried from hw\n");
+		err = -EINVAL;
+		goto err_doorbell_page_bar;
+	}
+
+	mlxsw_pci->doorbell_offset =
+		mlxsw_cmd_mbox_query_fw_doorbell_page_offset_get(mbox);
+
+	num_pages = mlxsw_cmd_mbox_query_fw_fw_pages_get(mbox);
+	err = mlxsw_pci_fw_area_init(mlxsw_pci, mbox, num_pages);
+	if (err)
+		goto err_fw_area_init;
+
+	err = mlxsw_pci_boardinfo(mlxsw_pci, mbox);
+	if (err)
+		goto err_boardinfo;
+
+	err = mlxsw_pci_config_profile(mlxsw_pci, mbox, profile);
+	if (err)
+		goto err_config_profile;
+
+	err = mlxsw_pci_aqs_init(mlxsw_pci, mbox);
+	if (err)
+		goto err_aqs_init;
+
+	err = request_irq(mlxsw_pci->msix_entry.vector,
+			  mlxsw_pci_eq_irq_handler, 0,
+			  mlxsw_pci_driver_name, mlxsw_pci);
+	if (err) {
+		dev_err(&pdev->dev, "IRQ request failed\n");
+		goto err_request_eq_irq;
+	}
+
+	goto mbox_put;
+
+err_request_eq_irq:
+	mlxsw_pci_aqs_fini(mlxsw_pci);
+err_aqs_init:
+err_config_profile:
+err_boardinfo:
+	mlxsw_pci_fw_area_fini(mlxsw_pci);
+err_fw_area_init:
+err_doorbell_page_bar:
+err_iface_rev:
+err_query_fw:
+	mlxsw_pci_mbox_free(mlxsw_pci, &mlxsw_pci->cmd.out_mbox);
+err_out_mbox_alloc:
+	mlxsw_pci_mbox_free(mlxsw_pci, &mlxsw_pci->cmd.in_mbox);
+mbox_put:
+	mlxsw_cmd_mbox_free(mbox);
+	return err;
+}
+
+static void mlxsw_pci_fini(void *bus_priv)
+{
+	struct mlxsw_pci *mlxsw_pci = bus_priv;
+
+	free_irq(mlxsw_pci->msix_entry.vector, mlxsw_pci);
+	mlxsw_pci_aqs_fini(mlxsw_pci);
+	mlxsw_pci_fw_area_fini(mlxsw_pci);
+	mlxsw_pci_mbox_free(mlxsw_pci, &mlxsw_pci->cmd.out_mbox);
+	mlxsw_pci_mbox_free(mlxsw_pci, &mlxsw_pci->cmd.in_mbox);
+}
+
+static struct mlxsw_pci_queue *
+mlxsw_pci_sdq_pick(struct mlxsw_pci *mlxsw_pci,
+		   const struct mlxsw_tx_info *tx_info)
+{
+	u8 sdqn = tx_info->local_port % mlxsw_pci_sdq_count(mlxsw_pci);
+
+	return mlxsw_pci_sdq_get(mlxsw_pci, sdqn);
+}
+
+static bool mlxsw_pci_skb_transmit_busy(void *bus_priv,
+					const struct mlxsw_tx_info *tx_info)
+{
+	struct mlxsw_pci *mlxsw_pci = bus_priv;
+	struct mlxsw_pci_queue *q = mlxsw_pci_sdq_pick(mlxsw_pci, tx_info);
+
+	return !mlxsw_pci_queue_elem_info_producer_get(q);
+}
+
+static int mlxsw_pci_skb_transmit(void *bus_priv, struct sk_buff *skb,
+				  const struct mlxsw_tx_info *tx_info)
+{
+	struct mlxsw_pci *mlxsw_pci = bus_priv;
+	struct mlxsw_pci_queue *q;
+	struct mlxsw_pci_queue_elem_info *elem_info;
+	char *wqe;
+	int i;
+	int err;
+
+	if (skb_shinfo(skb)->nr_frags > MLXSW_PCI_WQE_SG_ENTRIES - 1) {
+		err = skb_linearize(skb);
+		if (err)
+			return err;
+	}
+
+	q = mlxsw_pci_sdq_pick(mlxsw_pci, tx_info);
+	spin_lock_bh(&q->lock);
+	elem_info = mlxsw_pci_queue_elem_info_producer_get(q);
+	if (!elem_info) {
+		/* queue is full */
+		err = -EAGAIN;
+		goto unlock;
+	}
+	elem_info->u.sdq.skb = skb;
+
+	wqe = elem_info->elem;
+	mlxsw_pci_wqe_c_set(wqe, 1); /* always report completion */
+	mlxsw_pci_wqe_lp_set(wqe, !!tx_info->is_emad);
+	mlxsw_pci_wqe_type_set(wqe, MLXSW_PCI_WQE_TYPE_ETHERNET);
+
+	err = mlxsw_pci_wqe_frag_map(mlxsw_pci, wqe, 0, skb->data,
+				     skb_headlen(skb), DMA_TO_DEVICE);
+	if (err)
+		goto unlock;
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+		const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+
+		err = mlxsw_pci_wqe_frag_map(mlxsw_pci, wqe, i + 1,
+					     skb_frag_address(frag),
+					     skb_frag_size(frag),
+					     DMA_TO_DEVICE);
+		if (err)
+			goto unmap_frags;
+	}
+
+	/* Set unused sq entries byte count to zero. */
+	for (i++; i < MLXSW_PCI_WQE_SG_ENTRIES; i++)
+		mlxsw_pci_wqe_byte_count_set(wqe, i, 0);
+
+	/* Everything is set up, ring producer doorbell to get HW going */
+	q->producer_counter++;
+	mlxsw_pci_queue_doorbell_producer_ring(mlxsw_pci, q);
+
+	goto unlock;
+
+unmap_frags:
+	for (; i >= 0; i--)
+		mlxsw_pci_wqe_frag_unmap(mlxsw_pci, wqe, i, DMA_TO_DEVICE);
+unlock:
+	spin_unlock_bh(&q->lock);
+	return err;
+}
+
+static int mlxsw_pci_cmd_exec(void *bus_priv, u16 opcode, u8 opcode_mod,
+			      u32 in_mod, bool out_mbox_direct,
+			      char *in_mbox, size_t in_mbox_size,
+			      char *out_mbox, size_t out_mbox_size,
+			      u8 *p_status)
+{
+	struct mlxsw_pci *mlxsw_pci = bus_priv;
+	dma_addr_t in_mapaddr = mlxsw_pci->cmd.in_mbox.mapaddr;
+	dma_addr_t out_mapaddr = mlxsw_pci->cmd.out_mbox.mapaddr;
+	bool evreq = mlxsw_pci->cmd.nopoll;
+	unsigned long timeout = msecs_to_jiffies(MLXSW_PCI_CIR_TIMEOUT_MSECS);
+	bool *p_wait_done = &mlxsw_pci->cmd.wait_done;
+	int err;
+
+	*p_status = MLXSW_CMD_STATUS_OK;
+
+	err = mutex_lock_interruptible(&mlxsw_pci->cmd.lock);
+	if (err)
+		return err;
+
+	if (in_mbox)
+		memcpy(mlxsw_pci->cmd.in_mbox.buf, in_mbox, in_mbox_size);
+	mlxsw_pci_write32(mlxsw_pci, CIR_IN_PARAM_HI, upper_32_bits(in_mapaddr));
+	mlxsw_pci_write32(mlxsw_pci, CIR_IN_PARAM_LO, lower_32_bits(in_mapaddr));
+
+	mlxsw_pci_write32(mlxsw_pci, CIR_OUT_PARAM_HI, upper_32_bits(out_mapaddr));
+	mlxsw_pci_write32(mlxsw_pci, CIR_OUT_PARAM_LO, lower_32_bits(out_mapaddr));
+
+	mlxsw_pci_write32(mlxsw_pci, CIR_IN_MODIFIER, in_mod);
+	mlxsw_pci_write32(mlxsw_pci, CIR_TOKEN, 0);
+
+	*p_wait_done = false;
+
+	wmb(); /* all needs to be written before we write control register */
+	mlxsw_pci_write32(mlxsw_pci, CIR_CTRL,
+			  MLXSW_PCI_CIR_CTRL_GO_BIT |
+			  (evreq ? MLXSW_PCI_CIR_CTRL_EVREQ_BIT : 0) |
+			  (opcode_mod << MLXSW_PCI_CIR_CTRL_OPCODE_MOD_SHIFT) |
+			  opcode);
+
+	if (!evreq) {
+		unsigned long end;
+
+		end = jiffies + timeout;
+		do {
+			u32 ctrl = mlxsw_pci_read32(mlxsw_pci, CIR_CTRL);
+
+			if (!(ctrl & MLXSW_PCI_CIR_CTRL_GO_BIT)) {
+				*p_wait_done = true;
+				*p_status = ctrl >> MLXSW_PCI_CIR_CTRL_STATUS_SHIFT;
+				break;
+			}
+			cond_resched();
+		} while (time_before(jiffies, end));
+	} else {
+		wait_event_timeout(mlxsw_pci->cmd.wait, *p_wait_done, timeout);
+		*p_status = mlxsw_pci->cmd.comp.status;
+	}
+
+	err = 0;
+	if (*p_wait_done) {
+		if (*p_status)
+			err = -EIO;
+	} else {
+		err = -ETIMEDOUT;
+	}
+
+	if (!err && out_mbox && out_mbox_direct) {
+		/* Some commands don't use output param as address to mailbox
+		 * but they store output directly into registers. In that case,
+		 * copy registers into mbox buffer.
+		 */
+		__be32 tmp;
+
+		if (!evreq) {
+			tmp = cpu_to_be32(mlxsw_pci_read32(mlxsw_pci,
+							   CIR_OUT_PARAM_HI));
+			memcpy(out_mbox, &tmp, sizeof(tmp));
+			tmp = cpu_to_be32(mlxsw_pci_read32(mlxsw_pci,
+							   CIR_OUT_PARAM_LO));
+			memcpy(out_mbox + sizeof(tmp), &tmp, sizeof(tmp));
+		}
+	} else if (!err && out_mbox) {
+		memcpy(out_mbox, mlxsw_pci->cmd.out_mbox.buf, out_mbox_size);
+	}
+
+	mutex_unlock(&mlxsw_pci->cmd.lock);
+
+	return err;
+}
+
+static const struct mlxsw_bus mlxsw_pci_bus = {
+	.kind			= "pci",
+	.init			= mlxsw_pci_init,
+	.fini			= mlxsw_pci_fini,
+	.skb_transmit_busy	= mlxsw_pci_skb_transmit_busy,
+	.skb_transmit		= mlxsw_pci_skb_transmit,
+	.cmd_exec		= mlxsw_pci_cmd_exec,
+};
+
+static int mlxsw_pci_sw_reset(struct mlxsw_pci *mlxsw_pci)
+{
+	mlxsw_pci_write32(mlxsw_pci, SW_RESET, MLXSW_PCI_SW_RESET_RST_BIT);
+	/* Current firware does not let us know when the reset is done.
+	 * So we just wait here for constant time and hope for the best.
+	 */
+	msleep(MLXSW_PCI_SW_RESET_TIMEOUT_MSECS);
+	return 0;
+}
+
+static int mlxsw_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id)
+{
+	struct mlxsw_pci *mlxsw_pci;
+	int err;
+
+	mlxsw_pci = kzalloc(sizeof(*mlxsw_pci), GFP_KERNEL);
+	if (!mlxsw_pci)
+		return -ENOMEM;
+
+	err = pci_enable_device(pdev);
+	if (err) {
+		dev_err(&pdev->dev, "pci_enable_device failed\n");
+		goto err_pci_enable_device;
+	}
+
+	err = pci_request_regions(pdev, mlxsw_pci_driver_name);
+	if (err) {
+		dev_err(&pdev->dev, "pci_request_regions failed\n");
+		goto err_pci_request_regions;
+	}
+
+	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
+	if (!err) {
+		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+		if (err) {
+			dev_err(&pdev->dev, "pci_set_consistent_dma_mask failed\n");
+			goto err_pci_set_dma_mask;
+		}
+	} else {
+		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+		if (err) {
+			dev_err(&pdev->dev, "pci_set_dma_mask failed\n");
+			goto err_pci_set_dma_mask;
+		}
+	}
+
+	if (pci_resource_len(pdev, 0) < MLXSW_PCI_BAR0_SIZE) {
+		dev_err(&pdev->dev, "invalid PCI region size\n");
+		err = -EINVAL;
+		goto err_pci_resource_len_check;
+	}
+
+	mlxsw_pci->hw_addr = ioremap(pci_resource_start(pdev, 0),
+				     pci_resource_len(pdev, 0));
+	if (!mlxsw_pci->hw_addr) {
+		dev_err(&pdev->dev, "ioremap failed\n");
+		err = -EIO;
+		goto err_ioremap;
+	}
+	pci_set_master(pdev);
+
+	mlxsw_pci->pdev = pdev;
+	pci_set_drvdata(pdev, mlxsw_pci);
+
+	err = mlxsw_pci_sw_reset(mlxsw_pci);
+	if (err) {
+		dev_err(&pdev->dev, "Software reset failed\n");
+		goto err_sw_reset;
+	}
+
+	err = pci_enable_msix_exact(pdev, &mlxsw_pci->msix_entry, 1);
+	if (err) {
+		dev_err(&pdev->dev, "MSI-X init failed\n");
+		goto err_msix_init;
+	}
+
+	mlxsw_pci->bus_info.device_kind = mlxsw_pci_device_kind_get(id);
+	mlxsw_pci->bus_info.device_name = pci_name(mlxsw_pci->pdev);
+	mlxsw_pci->bus_info.dev = &pdev->dev;
+
+	mlxsw_pci->dbg_dir = debugfs_create_dir(mlxsw_pci->bus_info.device_name,
+						mlxsw_pci_dbg_root);
+	if (!mlxsw_pci->dbg_dir) {
+		dev_err(&pdev->dev, "Failed to create debugfs dir\n");
+		err = -ENOMEM;
+		goto err_dbg_create_dir;
+	}
+
+	err = mlxsw_core_bus_device_register(&mlxsw_pci->bus_info,
+					     &mlxsw_pci_bus, mlxsw_pci);
+	if (err) {
+		dev_err(&pdev->dev, "cannot register bus device\n");
+		goto err_bus_device_register;
+	}
+
+	return 0;
+
+err_bus_device_register:
+	debugfs_remove_recursive(mlxsw_pci->dbg_dir);
+err_dbg_create_dir:
+	pci_disable_msix(mlxsw_pci->pdev);
+err_msix_init:
+err_sw_reset:
+	iounmap(mlxsw_pci->hw_addr);
+err_ioremap:
+err_pci_resource_len_check:
+err_pci_set_dma_mask:
+	pci_release_regions(pdev);
+err_pci_request_regions:
+	pci_disable_device(pdev);
+err_pci_enable_device:
+	kfree(mlxsw_pci);
+	return err;
+}
+
+static void mlxsw_pci_remove(struct pci_dev *pdev)
+{
+	struct mlxsw_pci *mlxsw_pci = pci_get_drvdata(pdev);
+
+	mlxsw_core_bus_device_unregister(mlxsw_pci->core);
+	debugfs_remove_recursive(mlxsw_pci->dbg_dir);
+	pci_disable_msix(mlxsw_pci->pdev);
+	iounmap(mlxsw_pci->hw_addr);
+	pci_release_regions(mlxsw_pci->pdev);
+	pci_disable_device(mlxsw_pci->pdev);
+	kfree(mlxsw_pci);
+}
+
+static struct pci_driver mlxsw_pci_driver = {
+	.name		= mlxsw_pci_driver_name,
+	.id_table	= mlxsw_pci_id_table,
+	.probe		= mlxsw_pci_probe,
+	.remove		= mlxsw_pci_remove,
+};
+
+static int __init mlxsw_pci_module_init(void)
+{
+	int err;
+
+	mlxsw_pci_dbg_root = debugfs_create_dir(mlxsw_pci_driver_name, NULL);
+	if (!mlxsw_pci_dbg_root)
+		return -ENOMEM;
+	err = pci_register_driver(&mlxsw_pci_driver);
+	if (err)
+		goto err_register_driver;
+	return 0;
+
+err_register_driver:
+	debugfs_remove_recursive(mlxsw_pci_dbg_root);
+	return err;
+}
+
+static void __exit mlxsw_pci_module_exit(void)
+{
+	pci_unregister_driver(&mlxsw_pci_driver);
+	debugfs_remove_recursive(mlxsw_pci_dbg_root);
+}
+
+module_init(mlxsw_pci_module_init);
+module_exit(mlxsw_pci_module_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Jiri Pirko <jiri@mellanox.com>");
+MODULE_DESCRIPTION("Mellanox switch PCI interface driver");
+MODULE_DEVICE_TABLE(pci, mlxsw_pci_id_table);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/pci.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/pci.h
new file mode 100644
index 000000000..142f33d97
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/pci.h
@@ -0,0 +1,226 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/pci.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_PCI_H
+#define _MLXSW_PCI_H
+
+#include <linux/bitops.h>
+
+#include "item.h"
+
+#define PCI_DEVICE_ID_MELLANOX_SWITCHX2	0xc738
+#define PCI_DEVICE_ID_MELLANOX_SPECTRUM	0xcb84
+#define MLXSW_PCI_BAR0_SIZE		(1024 * 1024) /* 1MB */
+#define MLXSW_PCI_PAGE_SIZE		4096
+
+#define MLXSW_PCI_CIR_BASE			0x71000
+#define MLXSW_PCI_CIR_IN_PARAM_HI		MLXSW_PCI_CIR_BASE
+#define MLXSW_PCI_CIR_IN_PARAM_LO		(MLXSW_PCI_CIR_BASE + 0x04)
+#define MLXSW_PCI_CIR_IN_MODIFIER		(MLXSW_PCI_CIR_BASE + 0x08)
+#define MLXSW_PCI_CIR_OUT_PARAM_HI		(MLXSW_PCI_CIR_BASE + 0x0C)
+#define MLXSW_PCI_CIR_OUT_PARAM_LO		(MLXSW_PCI_CIR_BASE + 0x10)
+#define MLXSW_PCI_CIR_TOKEN			(MLXSW_PCI_CIR_BASE + 0x14)
+#define MLXSW_PCI_CIR_CTRL			(MLXSW_PCI_CIR_BASE + 0x18)
+#define MLXSW_PCI_CIR_CTRL_GO_BIT		BIT(23)
+#define MLXSW_PCI_CIR_CTRL_EVREQ_BIT		BIT(22)
+#define MLXSW_PCI_CIR_CTRL_OPCODE_MOD_SHIFT	12
+#define MLXSW_PCI_CIR_CTRL_STATUS_SHIFT		24
+#define MLXSW_PCI_CIR_TIMEOUT_MSECS		1000
+
+#define MLXSW_PCI_SW_RESET			0xF0010
+#define MLXSW_PCI_SW_RESET_RST_BIT		BIT(0)
+#define MLXSW_PCI_SW_RESET_TIMEOUT_MSECS	5000
+
+#define MLXSW_PCI_DOORBELL_SDQ_OFFSET		0x000
+#define MLXSW_PCI_DOORBELL_RDQ_OFFSET		0x200
+#define MLXSW_PCI_DOORBELL_CQ_OFFSET		0x400
+#define MLXSW_PCI_DOORBELL_EQ_OFFSET		0x600
+#define MLXSW_PCI_DOORBELL_ARM_CQ_OFFSET	0x800
+#define MLXSW_PCI_DOORBELL_ARM_EQ_OFFSET	0xA00
+
+#define MLXSW_PCI_DOORBELL(offset, type_offset, num)	\
+	((offset) + (type_offset) + (num) * 4)
+
+#define MLXSW_PCI_CQS_MAX	96
+#define MLXSW_PCI_EQS_COUNT	2
+#define MLXSW_PCI_EQ_ASYNC_NUM	0
+#define MLXSW_PCI_EQ_COMP_NUM	1
+
+#define MLXSW_PCI_AQ_PAGES	8
+#define MLXSW_PCI_AQ_SIZE	(MLXSW_PCI_PAGE_SIZE * MLXSW_PCI_AQ_PAGES)
+#define MLXSW_PCI_WQE_SIZE	32 /* 32 bytes per element */
+#define MLXSW_PCI_CQE_SIZE	16 /* 16 bytes per element */
+#define MLXSW_PCI_EQE_SIZE	16 /* 16 bytes per element */
+#define MLXSW_PCI_WQE_COUNT	(MLXSW_PCI_AQ_SIZE / MLXSW_PCI_WQE_SIZE)
+#define MLXSW_PCI_CQE_COUNT	(MLXSW_PCI_AQ_SIZE / MLXSW_PCI_CQE_SIZE)
+#define MLXSW_PCI_EQE_COUNT	(MLXSW_PCI_AQ_SIZE / MLXSW_PCI_EQE_SIZE)
+#define MLXSW_PCI_EQE_UPDATE_COUNT	0x80
+
+#define MLXSW_PCI_WQE_SG_ENTRIES	3
+#define MLXSW_PCI_WQE_TYPE_ETHERNET	0xA
+
+/* pci_wqe_c
+ * If set it indicates that a completion should be reported upon
+ * execution of this descriptor.
+ */
+MLXSW_ITEM32(pci, wqe, c, 0x00, 31, 1);
+
+/* pci_wqe_lp
+ * Local Processing, set if packet should be processed by the local
+ * switch hardware:
+ * For Ethernet EMAD (Direct Route and non Direct Route) -
+ * must be set if packet destination is local device
+ * For InfiniBand CTL - must be set if packet destination is local device
+ * Otherwise it must be clear
+ * Local Process packets must not exceed the size of 2K (including payload
+ * and headers).
+ */
+MLXSW_ITEM32(pci, wqe, lp, 0x00, 30, 1);
+
+/* pci_wqe_type
+ * Packet type.
+ */
+MLXSW_ITEM32(pci, wqe, type, 0x00, 23, 4);
+
+/* pci_wqe_byte_count
+ * Size of i-th scatter/gather entry, 0 if entry is unused.
+ */
+MLXSW_ITEM16_INDEXED(pci, wqe, byte_count, 0x02, 0, 14, 0x02, 0x00, false);
+
+/* pci_wqe_address
+ * Physical address of i-th scatter/gather entry.
+ * Gather Entries must be 2Byte aligned.
+ */
+MLXSW_ITEM64_INDEXED(pci, wqe, address, 0x08, 0, 64, 0x8, 0x0, false);
+
+/* pci_cqe_lag
+ * Packet arrives from a port which is a LAG
+ */
+MLXSW_ITEM32(pci, cqe, lag, 0x00, 23, 1);
+
+/* pci_cqe_system_port
+ * When lag=0: System port on which the packet was received
+ * When lag=1:
+ * bits [15:4] LAG ID on which the packet was received
+ * bits [3:0] sub_port on which the packet was received
+ */
+MLXSW_ITEM32(pci, cqe, system_port, 0x00, 0, 16);
+
+/* pci_cqe_wqe_counter
+ * WQE count of the WQEs completed on the associated dqn
+ */
+MLXSW_ITEM32(pci, cqe, wqe_counter, 0x04, 16, 16);
+
+/* pci_cqe_byte_count
+ * Byte count of received packets including additional two
+ * Reserved Bytes that are append to the end of the frame.
+ * Reserved for Send CQE.
+ */
+MLXSW_ITEM32(pci, cqe, byte_count, 0x04, 0, 14);
+
+/* pci_cqe_trap_id
+ * Trap ID that captured the packet.
+ */
+MLXSW_ITEM32(pci, cqe, trap_id, 0x08, 0, 8);
+
+/* pci_cqe_crc
+ * Length include CRC. Indicates the length field includes
+ * the packet's CRC.
+ */
+MLXSW_ITEM32(pci, cqe, crc, 0x0C, 8, 1);
+
+/* pci_cqe_e
+ * CQE with Error.
+ */
+MLXSW_ITEM32(pci, cqe, e, 0x0C, 7, 1);
+
+/* pci_cqe_sr
+ * 1 - Send Queue
+ * 0 - Receive Queue
+ */
+MLXSW_ITEM32(pci, cqe, sr, 0x0C, 6, 1);
+
+/* pci_cqe_dqn
+ * Descriptor Queue (DQ) Number.
+ */
+MLXSW_ITEM32(pci, cqe, dqn, 0x0C, 1, 5);
+
+/* pci_cqe_owner
+ * Ownership bit.
+ */
+MLXSW_ITEM32(pci, cqe, owner, 0x0C, 0, 1);
+
+/* pci_eqe_event_type
+ * Event type.
+ */
+MLXSW_ITEM32(pci, eqe, event_type, 0x0C, 24, 8);
+#define MLXSW_PCI_EQE_EVENT_TYPE_COMP	0x00
+#define MLXSW_PCI_EQE_EVENT_TYPE_CMD	0x0A
+
+/* pci_eqe_event_sub_type
+ * Event type.
+ */
+MLXSW_ITEM32(pci, eqe, event_sub_type, 0x0C, 16, 8);
+
+/* pci_eqe_cqn
+ * Completion Queue that triggeret this EQE.
+ */
+MLXSW_ITEM32(pci, eqe, cqn, 0x0C, 8, 7);
+
+/* pci_eqe_owner
+ * Ownership bit.
+ */
+MLXSW_ITEM32(pci, eqe, owner, 0x0C, 0, 1);
+
+/* pci_eqe_cmd_token
+ * Command completion event - token
+ */
+MLXSW_ITEM32(pci, eqe, cmd_token, 0x08, 16, 16);
+
+/* pci_eqe_cmd_status
+ * Command completion event - status
+ */
+MLXSW_ITEM32(pci, eqe, cmd_status, 0x08, 0, 8);
+
+/* pci_eqe_cmd_out_param_h
+ * Command completion event - output parameter - higher part
+ */
+MLXSW_ITEM32(pci, eqe, cmd_out_param_h, 0x0C, 0, 32);
+
+/* pci_eqe_cmd_out_param_l
+ * Command completion event - output parameter - lower part
+ */
+MLXSW_ITEM32(pci, eqe, cmd_out_param_l, 0x10, 0, 32);
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/port.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/port.h
new file mode 100644
index 000000000..726f5435b
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/port.h
@@ -0,0 +1,75 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/port.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#ifndef _MLXSW_PORT_H
+#define _MLXSW_PORT_H
+
+#include <linux/types.h>
+
+#define MLXSW_PORT_MAX_MTU		10000
+
+#define MLXSW_PORT_DEFAULT_VID		1
+
+#define MLXSW_PORT_SWID_DISABLED_PORT	255
+#define MLXSW_PORT_SWID_ALL_SWIDS	254
+#define MLXSW_PORT_SWID_TYPE_ETH	2
+
+#define MLXSW_PORT_MID			0xd000
+
+#define MLXSW_PORT_MAX_PHY_PORTS	0x40
+#define MLXSW_PORT_MAX_PORTS		MLXSW_PORT_MAX_PHY_PORTS
+
+#define MLXSW_PORT_DEVID_BITS_OFFSET	10
+#define MLXSW_PORT_PHY_BITS_OFFSET	4
+#define MLXSW_PORT_PHY_BITS_MASK	(MLXSW_PORT_MAX_PHY_PORTS - 1)
+
+#define MLXSW_PORT_CPU_PORT		0x0
+
+#define MLXSW_PORT_DONT_CARE		(MLXSW_PORT_MAX_PORTS)
+
+enum mlxsw_port_admin_status {
+	MLXSW_PORT_ADMIN_STATUS_UP = 1,
+	MLXSW_PORT_ADMIN_STATUS_DOWN = 2,
+	MLXSW_PORT_ADMIN_STATUS_UP_ONCE = 3,
+	MLXSW_PORT_ADMIN_STATUS_DISABLED = 4,
+};
+
+enum mlxsw_reg_pude_oper_status {
+	MLXSW_PORT_OPER_STATUS_UP = 1,
+	MLXSW_PORT_OPER_STATUS_DOWN = 2,
+	MLXSW_PORT_OPER_STATUS_FAILURE = 4,	/* Can be set to up again. */
+};
+
+#endif /* _MLXSW_PORT_H */
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/reg.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/reg.h
new file mode 100644
index 000000000..236fb5d2a
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -0,0 +1,2460 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/reg.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_REG_H
+#define _MLXSW_REG_H
+
+#include <linux/string.h>
+#include <linux/bitops.h>
+#include <linux/if_vlan.h>
+
+#include "item.h"
+#include "port.h"
+
+struct mlxsw_reg_info {
+	u16 id;
+	u16 len; /* In u8 */
+};
+
+#define MLXSW_REG(type) (&mlxsw_reg_##type)
+#define MLXSW_REG_LEN(type) MLXSW_REG(type)->len
+#define MLXSW_REG_ZERO(type, payload) memset(payload, 0, MLXSW_REG(type)->len)
+
+/* SGCR - Switch General Configuration Register
+ * --------------------------------------------
+ * This register is used for configuration of the switch capabilities.
+ */
+#define MLXSW_REG_SGCR_ID 0x2000
+#define MLXSW_REG_SGCR_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_sgcr = {
+	.id = MLXSW_REG_SGCR_ID,
+	.len = MLXSW_REG_SGCR_LEN,
+};
+
+/* reg_sgcr_llb
+ * Link Local Broadcast (Default=0)
+ * When set, all Link Local packets (224.0.0.X) will be treated as broadcast
+ * packets and ignore the IGMP snooping entries.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sgcr, llb, 0x04, 0, 1);
+
+static inline void mlxsw_reg_sgcr_pack(char *payload, bool llb)
+{
+	MLXSW_REG_ZERO(sgcr, payload);
+	mlxsw_reg_sgcr_llb_set(payload, !!llb);
+}
+
+/* SPAD - Switch Physical Address Register
+ * ---------------------------------------
+ * The SPAD register configures the switch physical MAC address.
+ */
+#define MLXSW_REG_SPAD_ID 0x2002
+#define MLXSW_REG_SPAD_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_spad = {
+	.id = MLXSW_REG_SPAD_ID,
+	.len = MLXSW_REG_SPAD_LEN,
+};
+
+/* reg_spad_base_mac
+ * Base MAC address for the switch partitions.
+ * Per switch partition MAC address is equal to:
+ * base_mac + swid
+ * Access: RW
+ */
+MLXSW_ITEM_BUF(reg, spad, base_mac, 0x02, 6);
+
+/* SSPR - Switch System Port Record Register
+ * -----------------------------------------
+ * Configures the system port to local port mapping.
+ */
+#define MLXSW_REG_SSPR_ID 0x2008
+#define MLXSW_REG_SSPR_LEN 0x8
+
+static const struct mlxsw_reg_info mlxsw_reg_sspr = {
+	.id = MLXSW_REG_SSPR_ID,
+	.len = MLXSW_REG_SSPR_LEN,
+};
+
+/* reg_sspr_m
+ * Master - if set, then the record describes the master system port.
+ * This is needed in case a local port is mapped into several system ports
+ * (for multipathing). That number will be reported as the source system
+ * port when packets are forwarded to the CPU. Only one master port is allowed
+ * per local port.
+ *
+ * Note: Must be set for Spectrum.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sspr, m, 0x00, 31, 1);
+
+/* reg_sspr_local_port
+ * Local port number.
+ *
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sspr, local_port, 0x00, 16, 8);
+
+/* reg_sspr_sub_port
+ * Virtual port within the physical port.
+ * Should be set to 0 when virtual ports are not enabled on the port.
+ *
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sspr, sub_port, 0x00, 8, 8);
+
+/* reg_sspr_system_port
+ * Unique identifier within the stacking domain that represents all the ports
+ * that are available in the system (external ports).
+ *
+ * Currently, only single-ASIC configurations are supported, so we default to
+ * 1:1 mapping between system ports and local ports.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sspr, system_port, 0x04, 0, 16);
+
+static inline void mlxsw_reg_sspr_pack(char *payload, u8 local_port)
+{
+	MLXSW_REG_ZERO(sspr, payload);
+	mlxsw_reg_sspr_m_set(payload, 1);
+	mlxsw_reg_sspr_local_port_set(payload, local_port);
+	mlxsw_reg_sspr_sub_port_set(payload, 0);
+	mlxsw_reg_sspr_system_port_set(payload, local_port);
+}
+
+/* SFDAT - Switch Filtering Database Aging Time
+ * --------------------------------------------
+ * Controls the Switch aging time. Aging time is able to be set per Switch
+ * Partition.
+ */
+#define MLXSW_REG_SFDAT_ID 0x2009
+#define MLXSW_REG_SFDAT_LEN 0x8
+
+static const struct mlxsw_reg_info mlxsw_reg_sfdat = {
+	.id = MLXSW_REG_SFDAT_ID,
+	.len = MLXSW_REG_SFDAT_LEN,
+};
+
+/* reg_sfdat_swid
+ * Switch partition ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sfdat, swid, 0x00, 24, 8);
+
+/* reg_sfdat_age_time
+ * Aging time in seconds
+ * Min - 10 seconds
+ * Max - 1,000,000 seconds
+ * Default is 300 seconds.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfdat, age_time, 0x04, 0, 20);
+
+static inline void mlxsw_reg_sfdat_pack(char *payload, u32 age_time)
+{
+	MLXSW_REG_ZERO(sfdat, payload);
+	mlxsw_reg_sfdat_swid_set(payload, 0);
+	mlxsw_reg_sfdat_age_time_set(payload, age_time);
+}
+
+/* SFD - Switch Filtering Database
+ * -------------------------------
+ * The following register defines the access to the filtering database.
+ * The register supports querying, adding, removing and modifying the database.
+ * The access is optimized for bulk updates in which case more than one
+ * FDB record is present in the same command.
+ */
+#define MLXSW_REG_SFD_ID 0x200A
+#define MLXSW_REG_SFD_BASE_LEN 0x10 /* base length, without records */
+#define MLXSW_REG_SFD_REC_LEN 0x10 /* record length */
+#define MLXSW_REG_SFD_REC_MAX_COUNT 64
+#define MLXSW_REG_SFD_LEN (MLXSW_REG_SFD_BASE_LEN +	\
+			   MLXSW_REG_SFD_REC_LEN * MLXSW_REG_SFD_REC_MAX_COUNT)
+
+static const struct mlxsw_reg_info mlxsw_reg_sfd = {
+	.id = MLXSW_REG_SFD_ID,
+	.len = MLXSW_REG_SFD_LEN,
+};
+
+/* reg_sfd_swid
+ * Switch partition ID for queries. Reserved on Write.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sfd, swid, 0x00, 24, 8);
+
+enum mlxsw_reg_sfd_op {
+	/* Dump entire FDB a (process according to record_locator) */
+	MLXSW_REG_SFD_OP_QUERY_DUMP = 0,
+	/* Query records by {MAC, VID/FID} value */
+	MLXSW_REG_SFD_OP_QUERY_QUERY = 1,
+	/* Query and clear activity. Query records by {MAC, VID/FID} value */
+	MLXSW_REG_SFD_OP_QUERY_QUERY_AND_CLEAR_ACTIVITY = 2,
+	/* Test. Response indicates if each of the records could be
+	 * added to the FDB.
+	 */
+	MLXSW_REG_SFD_OP_WRITE_TEST = 0,
+	/* Add/modify. Aged-out records cannot be added. This command removes
+	 * the learning notification of the {MAC, VID/FID}. Response includes
+	 * the entries that were added to the FDB.
+	 */
+	MLXSW_REG_SFD_OP_WRITE_EDIT = 1,
+	/* Remove record by {MAC, VID/FID}. This command also removes
+	 * the learning notification and aged-out notifications
+	 * of the {MAC, VID/FID}. The response provides current (pre-removal)
+	 * entries as non-aged-out.
+	 */
+	MLXSW_REG_SFD_OP_WRITE_REMOVE = 2,
+	/* Remove learned notification by {MAC, VID/FID}. The response provides
+	 * the removed learning notification.
+	 */
+	MLXSW_REG_SFD_OP_WRITE_REMOVE_NOTIFICATION = 2,
+};
+
+/* reg_sfd_op
+ * Operation.
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, sfd, op, 0x04, 30, 2);
+
+/* reg_sfd_record_locator
+ * Used for querying the FDB. Use record_locator=0 to initiate the
+ * query. When a record is returned, a new record_locator is
+ * returned to be used in the subsequent query.
+ * Reserved for database update.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sfd, record_locator, 0x04, 0, 30);
+
+/* reg_sfd_num_rec
+ * Request: Number of records to read/add/modify/remove
+ * Response: Number of records read/added/replaced/removed
+ * See above description for more details.
+ * Ranges 0..64
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfd, num_rec, 0x08, 0, 8);
+
+static inline void mlxsw_reg_sfd_pack(char *payload, enum mlxsw_reg_sfd_op op,
+				      u32 record_locator)
+{
+	MLXSW_REG_ZERO(sfd, payload);
+	mlxsw_reg_sfd_op_set(payload, op);
+	mlxsw_reg_sfd_record_locator_set(payload, record_locator);
+}
+
+/* reg_sfd_rec_swid
+ * Switch partition ID.
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, rec_swid, MLXSW_REG_SFD_BASE_LEN, 24, 8,
+		     MLXSW_REG_SFD_REC_LEN, 0x00, false);
+
+enum mlxsw_reg_sfd_rec_type {
+	MLXSW_REG_SFD_REC_TYPE_UNICAST = 0x0,
+};
+
+/* reg_sfd_rec_type
+ * FDB record type.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, rec_type, MLXSW_REG_SFD_BASE_LEN, 20, 4,
+		     MLXSW_REG_SFD_REC_LEN, 0x00, false);
+
+enum mlxsw_reg_sfd_rec_policy {
+	/* Replacement disabled, aging disabled. */
+	MLXSW_REG_SFD_REC_POLICY_STATIC_ENTRY = 0,
+	/* (mlag remote): Replacement enabled, aging disabled,
+	 * learning notification enabled on this port.
+	 */
+	MLXSW_REG_SFD_REC_POLICY_DYNAMIC_ENTRY_MLAG = 1,
+	/* (ingress device): Replacement enabled, aging enabled. */
+	MLXSW_REG_SFD_REC_POLICY_DYNAMIC_ENTRY_INGRESS = 3,
+};
+
+/* reg_sfd_rec_policy
+ * Policy.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, rec_policy, MLXSW_REG_SFD_BASE_LEN, 18, 2,
+		     MLXSW_REG_SFD_REC_LEN, 0x00, false);
+
+/* reg_sfd_rec_a
+ * Activity. Set for new static entries. Set for static entries if a frame SMAC
+ * lookup hits on the entry.
+ * To clear the a bit, use "query and clear activity" op.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, rec_a, MLXSW_REG_SFD_BASE_LEN, 16, 1,
+		     MLXSW_REG_SFD_REC_LEN, 0x00, false);
+
+/* reg_sfd_rec_mac
+ * MAC address.
+ * Access: Index
+ */
+MLXSW_ITEM_BUF_INDEXED(reg, sfd, rec_mac, MLXSW_REG_SFD_BASE_LEN, 6,
+		       MLXSW_REG_SFD_REC_LEN, 0x02);
+
+enum mlxsw_reg_sfd_rec_action {
+	/* forward */
+	MLXSW_REG_SFD_REC_ACTION_NOP = 0,
+	/* forward and trap, trap_id is FDB_TRAP */
+	MLXSW_REG_SFD_REC_ACTION_MIRROR_TO_CPU = 1,
+	/* trap and do not forward, trap_id is FDB_TRAP */
+	MLXSW_REG_SFD_REC_ACTION_TRAP = 3,
+	MLXSW_REG_SFD_REC_ACTION_DISCARD_ERROR = 15,
+};
+
+/* reg_sfd_rec_action
+ * Action to apply on the packet.
+ * Note: Dynamic entries can only be configured with NOP action.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, rec_action, MLXSW_REG_SFD_BASE_LEN, 28, 4,
+		     MLXSW_REG_SFD_REC_LEN, 0x0C, false);
+
+/* reg_sfd_uc_sub_port
+ * VEPA channel on local port.
+ * Valid only if local port is a non-stacking port. Must be 0 if multichannel
+ * VEPA is not enabled.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, uc_sub_port, MLXSW_REG_SFD_BASE_LEN, 16, 8,
+		     MLXSW_REG_SFD_REC_LEN, 0x08, false);
+
+/* reg_sfd_uc_fid_vid
+ * Filtering ID or VLAN ID
+ * For SwitchX and SwitchX-2:
+ * - Dynamic entries (policy 2,3) use FID
+ * - Static entries (policy 0) use VID
+ * - When independent learning is configured, VID=FID
+ * For Spectrum: use FID for both Dynamic and Static entries.
+ * VID should not be used.
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, uc_fid_vid, MLXSW_REG_SFD_BASE_LEN, 0, 16,
+		     MLXSW_REG_SFD_REC_LEN, 0x08, false);
+
+/* reg_sfd_uc_system_port
+ * Unique port identifier for the final destination of the packet.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, sfd, uc_system_port, MLXSW_REG_SFD_BASE_LEN, 0, 16,
+		     MLXSW_REG_SFD_REC_LEN, 0x0C, false);
+
+static inline void mlxsw_reg_sfd_uc_pack(char *payload, int rec_index,
+					 enum mlxsw_reg_sfd_rec_policy policy,
+					 const char *mac, u16 vid,
+					 enum mlxsw_reg_sfd_rec_action action,
+					 u8 local_port)
+{
+	u8 num_rec = mlxsw_reg_sfd_num_rec_get(payload);
+
+	if (rec_index >= num_rec)
+		mlxsw_reg_sfd_num_rec_set(payload, rec_index + 1);
+	mlxsw_reg_sfd_rec_swid_set(payload, rec_index, 0);
+	mlxsw_reg_sfd_rec_type_set(payload, rec_index,
+				   MLXSW_REG_SFD_REC_TYPE_UNICAST);
+	mlxsw_reg_sfd_rec_policy_set(payload, rec_index, policy);
+	mlxsw_reg_sfd_rec_mac_memcpy_to(payload, rec_index, mac);
+	mlxsw_reg_sfd_uc_sub_port_set(payload, rec_index, 0);
+	mlxsw_reg_sfd_uc_fid_vid_set(payload, rec_index, vid);
+	mlxsw_reg_sfd_rec_action_set(payload, rec_index, action);
+	mlxsw_reg_sfd_uc_system_port_set(payload, rec_index, local_port);
+}
+
+static inline void mlxsw_reg_sfd_uc_unpack(char *payload, int rec_index,
+					   char *mac, u16 *p_vid,
+					   u8 *p_local_port)
+{
+	mlxsw_reg_sfd_rec_mac_memcpy_from(payload, rec_index, mac);
+	*p_vid = mlxsw_reg_sfd_uc_fid_vid_get(payload, rec_index);
+	*p_local_port = mlxsw_reg_sfd_uc_system_port_get(payload, rec_index);
+}
+
+/* SFN - Switch FDB Notification Register
+ * -------------------------------------------
+ * The switch provides notifications on newly learned FDB entries and
+ * aged out entries. The notifications can be polled by software.
+ */
+#define MLXSW_REG_SFN_ID 0x200B
+#define MLXSW_REG_SFN_BASE_LEN 0x10 /* base length, without records */
+#define MLXSW_REG_SFN_REC_LEN 0x10 /* record length */
+#define MLXSW_REG_SFN_REC_MAX_COUNT 64
+#define MLXSW_REG_SFN_LEN (MLXSW_REG_SFN_BASE_LEN +	\
+			   MLXSW_REG_SFN_REC_LEN * MLXSW_REG_SFN_REC_MAX_COUNT)
+
+static const struct mlxsw_reg_info mlxsw_reg_sfn = {
+	.id = MLXSW_REG_SFN_ID,
+	.len = MLXSW_REG_SFN_LEN,
+};
+
+/* reg_sfn_swid
+ * Switch partition ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sfn, swid, 0x00, 24, 8);
+
+/* reg_sfn_num_rec
+ * Request: Number of learned notifications and aged-out notification
+ * records requested.
+ * Response: Number of notification records returned (must be smaller
+ * than or equal to the value requested)
+ * Ranges 0..64
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, sfn, num_rec, 0x04, 0, 8);
+
+static inline void mlxsw_reg_sfn_pack(char *payload)
+{
+	MLXSW_REG_ZERO(sfn, payload);
+	mlxsw_reg_sfn_swid_set(payload, 0);
+	mlxsw_reg_sfn_num_rec_set(payload, MLXSW_REG_SFN_REC_MAX_COUNT);
+}
+
+/* reg_sfn_rec_swid
+ * Switch partition ID.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, rec_swid, MLXSW_REG_SFN_BASE_LEN, 24, 8,
+		     MLXSW_REG_SFN_REC_LEN, 0x00, false);
+
+enum mlxsw_reg_sfn_rec_type {
+	/* MAC addresses learned on a regular port. */
+	MLXSW_REG_SFN_REC_TYPE_LEARNED_MAC = 0x5,
+	/* Aged-out MAC address on a regular port */
+	MLXSW_REG_SFN_REC_TYPE_AGED_OUT_MAC = 0x7,
+};
+
+/* reg_sfn_rec_type
+ * Notification record type.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, rec_type, MLXSW_REG_SFN_BASE_LEN, 20, 4,
+		     MLXSW_REG_SFN_REC_LEN, 0x00, false);
+
+/* reg_sfn_rec_mac
+ * MAC address.
+ * Access: RO
+ */
+MLXSW_ITEM_BUF_INDEXED(reg, sfn, rec_mac, MLXSW_REG_SFN_BASE_LEN, 6,
+		       MLXSW_REG_SFN_REC_LEN, 0x02);
+
+/* reg_sfn_mac_sub_port
+ * VEPA channel on the local port.
+ * 0 if multichannel VEPA is not enabled.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, mac_sub_port, MLXSW_REG_SFN_BASE_LEN, 16, 8,
+		     MLXSW_REG_SFN_REC_LEN, 0x08, false);
+
+/* reg_sfn_mac_fid
+ * Filtering identifier.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, mac_fid, MLXSW_REG_SFN_BASE_LEN, 0, 16,
+		     MLXSW_REG_SFN_REC_LEN, 0x08, false);
+
+/* reg_sfn_mac_system_port
+ * Unique port identifier for the final destination of the packet.
+ * Access: RO
+ */
+MLXSW_ITEM32_INDEXED(reg, sfn, mac_system_port, MLXSW_REG_SFN_BASE_LEN, 0, 16,
+		     MLXSW_REG_SFN_REC_LEN, 0x0C, false);
+
+static inline void mlxsw_reg_sfn_mac_unpack(char *payload, int rec_index,
+					    char *mac, u16 *p_vid,
+					    u8 *p_local_port)
+{
+	mlxsw_reg_sfn_rec_mac_memcpy_from(payload, rec_index, mac);
+	*p_vid = mlxsw_reg_sfn_mac_fid_get(payload, rec_index);
+	*p_local_port = mlxsw_reg_sfn_mac_system_port_get(payload, rec_index);
+}
+
+/* SPMS - Switch Port MSTP/RSTP State Register
+ * -------------------------------------------
+ * Configures the spanning tree state of a physical port.
+ */
+#define MLXSW_REG_SPMS_ID 0x200D
+#define MLXSW_REG_SPMS_LEN 0x404
+
+static const struct mlxsw_reg_info mlxsw_reg_spms = {
+	.id = MLXSW_REG_SPMS_ID,
+	.len = MLXSW_REG_SPMS_LEN,
+};
+
+/* reg_spms_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spms, local_port, 0x00, 16, 8);
+
+enum mlxsw_reg_spms_state {
+	MLXSW_REG_SPMS_STATE_NO_CHANGE,
+	MLXSW_REG_SPMS_STATE_DISCARDING,
+	MLXSW_REG_SPMS_STATE_LEARNING,
+	MLXSW_REG_SPMS_STATE_FORWARDING,
+};
+
+/* reg_spms_state
+ * Spanning tree state of each VLAN ID (VID) of the local port.
+ * 0 - Do not change spanning tree state (used only when writing).
+ * 1 - Discarding. No learning or forwarding to/from this port (default).
+ * 2 - Learning. Port is learning, but not forwarding.
+ * 3 - Forwarding. Port is learning and forwarding.
+ * Access: RW
+ */
+MLXSW_ITEM_BIT_ARRAY(reg, spms, state, 0x04, 0x400, 2);
+
+static inline void mlxsw_reg_spms_pack(char *payload, u8 local_port)
+{
+	MLXSW_REG_ZERO(spms, payload);
+	mlxsw_reg_spms_local_port_set(payload, local_port);
+}
+
+static inline void mlxsw_reg_spms_vid_pack(char *payload, u16 vid,
+					   enum mlxsw_reg_spms_state state)
+{
+	mlxsw_reg_spms_state_set(payload, vid, state);
+}
+
+/* SPVID - Switch Port VID
+ * -----------------------
+ * The switch port VID configures the default VID for a port.
+ */
+#define MLXSW_REG_SPVID_ID 0x200E
+#define MLXSW_REG_SPVID_LEN 0x08
+
+static const struct mlxsw_reg_info mlxsw_reg_spvid = {
+	.id = MLXSW_REG_SPVID_ID,
+	.len = MLXSW_REG_SPVID_LEN,
+};
+
+/* reg_spvid_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spvid, local_port, 0x00, 16, 8);
+
+/* reg_spvid_sub_port
+ * Virtual port within the physical port.
+ * Should be set to 0 when virtual ports are not enabled on the port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spvid, sub_port, 0x00, 8, 8);
+
+/* reg_spvid_pvid
+ * Port default VID
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvid, pvid, 0x04, 0, 12);
+
+static inline void mlxsw_reg_spvid_pack(char *payload, u8 local_port, u16 pvid)
+{
+	MLXSW_REG_ZERO(spvid, payload);
+	mlxsw_reg_spvid_local_port_set(payload, local_port);
+	mlxsw_reg_spvid_pvid_set(payload, pvid);
+}
+
+/* SPVM - Switch Port VLAN Membership
+ * ----------------------------------
+ * The Switch Port VLAN Membership register configures the VLAN membership
+ * of a port in a VLAN denoted by VID. VLAN membership is managed per
+ * virtual port. The register can be used to add and remove VID(s) from a port.
+ */
+#define MLXSW_REG_SPVM_ID 0x200F
+#define MLXSW_REG_SPVM_BASE_LEN 0x04 /* base length, without records */
+#define MLXSW_REG_SPVM_REC_LEN 0x04 /* record length */
+#define MLXSW_REG_SPVM_REC_MAX_COUNT 256
+#define MLXSW_REG_SPVM_LEN (MLXSW_REG_SPVM_BASE_LEN +	\
+		    MLXSW_REG_SPVM_REC_LEN * MLXSW_REG_SPVM_REC_MAX_COUNT)
+
+static const struct mlxsw_reg_info mlxsw_reg_spvm = {
+	.id = MLXSW_REG_SPVM_ID,
+	.len = MLXSW_REG_SPVM_LEN,
+};
+
+/* reg_spvm_pt
+ * Priority tagged. If this bit is set, packets forwarded to the port with
+ * untagged VLAN membership (u bit is set) will be tagged with priority tag
+ * (VID=0)
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, spvm, pt, 0x00, 31, 1);
+
+/* reg_spvm_pte
+ * Priority Tagged Update Enable. On Write operations, if this bit is cleared,
+ * the pt bit will NOT be updated. To update the pt bit, pte must be set.
+ * Access: WO
+ */
+MLXSW_ITEM32(reg, spvm, pte, 0x00, 30, 1);
+
+/* reg_spvm_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spvm, local_port, 0x00, 16, 8);
+
+/* reg_spvm_sub_port
+ * Virtual port within the physical port.
+ * Should be set to 0 when virtual ports are not enabled on the port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spvm, sub_port, 0x00, 8, 8);
+
+/* reg_spvm_num_rec
+ * Number of records to update. Each record contains: i, e, u, vid.
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, spvm, num_rec, 0x00, 0, 8);
+
+/* reg_spvm_rec_i
+ * Ingress membership in VLAN ID.
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, spvm, rec_i,
+		     MLXSW_REG_SPVM_BASE_LEN, 14, 1,
+		     MLXSW_REG_SPVM_REC_LEN, 0, false);
+
+/* reg_spvm_rec_e
+ * Egress membership in VLAN ID.
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, spvm, rec_e,
+		     MLXSW_REG_SPVM_BASE_LEN, 13, 1,
+		     MLXSW_REG_SPVM_REC_LEN, 0, false);
+
+/* reg_spvm_rec_u
+ * Untagged - port is an untagged member - egress transmission uses untagged
+ * frames on VID<n>
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, spvm, rec_u,
+		     MLXSW_REG_SPVM_BASE_LEN, 12, 1,
+		     MLXSW_REG_SPVM_REC_LEN, 0, false);
+
+/* reg_spvm_rec_vid
+ * Egress membership in VLAN ID.
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, spvm, rec_vid,
+		     MLXSW_REG_SPVM_BASE_LEN, 0, 12,
+		     MLXSW_REG_SPVM_REC_LEN, 0, false);
+
+static inline void mlxsw_reg_spvm_pack(char *payload, u8 local_port,
+				       u16 vid_begin, u16 vid_end,
+				       bool is_member, bool untagged)
+{
+	int size = vid_end - vid_begin + 1;
+	int i;
+
+	MLXSW_REG_ZERO(spvm, payload);
+	mlxsw_reg_spvm_local_port_set(payload, local_port);
+	mlxsw_reg_spvm_num_rec_set(payload, size);
+
+	for (i = 0; i < size; i++) {
+		mlxsw_reg_spvm_rec_i_set(payload, i, is_member);
+		mlxsw_reg_spvm_rec_e_set(payload, i, is_member);
+		mlxsw_reg_spvm_rec_u_set(payload, i, untagged);
+		mlxsw_reg_spvm_rec_vid_set(payload, i, vid_begin + i);
+	}
+}
+
+/* SFGC - Switch Flooding Group Configuration
+ * ------------------------------------------
+ * The following register controls the association of flooding tables and MIDs
+ * to packet types used for flooding.
+ */
+#define MLXSW_REG_SFGC_ID 0x2011
+#define MLXSW_REG_SFGC_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_sfgc = {
+	.id = MLXSW_REG_SFGC_ID,
+	.len = MLXSW_REG_SFGC_LEN,
+};
+
+enum mlxsw_reg_sfgc_type {
+	MLXSW_REG_SFGC_TYPE_BROADCAST,
+	MLXSW_REG_SFGC_TYPE_UNKNOWN_UNICAST,
+	MLXSW_REG_SFGC_TYPE_UNREGISTERED_MULTICAST_IPV4,
+	MLXSW_REG_SFGC_TYPE_UNREGISTERED_MULTICAST_IPV6,
+	MLXSW_REG_SFGC_TYPE_RESERVED,
+	MLXSW_REG_SFGC_TYPE_UNREGISTERED_MULTICAST_NON_IP,
+	MLXSW_REG_SFGC_TYPE_IPV4_LINK_LOCAL,
+	MLXSW_REG_SFGC_TYPE_IPV6_ALL_HOST,
+	MLXSW_REG_SFGC_TYPE_MAX,
+};
+
+/* reg_sfgc_type
+ * The traffic type to reach the flooding table.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sfgc, type, 0x00, 0, 4);
+
+enum mlxsw_reg_sfgc_bridge_type {
+	MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID = 0,
+	MLXSW_REG_SFGC_BRIDGE_TYPE_VFID = 1,
+};
+
+/* reg_sfgc_bridge_type
+ * Access: Index
+ *
+ * Note: SwitchX-2 only supports 802.1Q mode.
+ */
+MLXSW_ITEM32(reg, sfgc, bridge_type, 0x04, 24, 3);
+
+enum mlxsw_flood_table_type {
+	MLXSW_REG_SFGC_TABLE_TYPE_VID = 1,
+	MLXSW_REG_SFGC_TABLE_TYPE_SINGLE = 2,
+	MLXSW_REG_SFGC_TABLE_TYPE_ANY = 0,
+	MLXSW_REG_SFGC_TABLE_TYPE_FID_OFFEST = 3,
+	MLXSW_REG_SFGC_TABLE_TYPE_FID = 4,
+};
+
+/* reg_sfgc_table_type
+ * See mlxsw_flood_table_type
+ * Access: RW
+ *
+ * Note: FID offset and FID types are not supported in SwitchX-2.
+ */
+MLXSW_ITEM32(reg, sfgc, table_type, 0x04, 16, 3);
+
+/* reg_sfgc_flood_table
+ * Flooding table index to associate with the specific type on the specific
+ * switch partition.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfgc, flood_table, 0x04, 0, 6);
+
+/* reg_sfgc_mid
+ * The multicast ID for the swid. Not supported for Spectrum
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfgc, mid, 0x08, 0, 16);
+
+/* reg_sfgc_counter_set_type
+ * Counter Set Type for flow counters.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfgc, counter_set_type, 0x0C, 24, 8);
+
+/* reg_sfgc_counter_index
+ * Counter Index for flow counters.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfgc, counter_index, 0x0C, 0, 24);
+
+static inline void
+mlxsw_reg_sfgc_pack(char *payload, enum mlxsw_reg_sfgc_type type,
+		    enum mlxsw_reg_sfgc_bridge_type bridge_type,
+		    enum mlxsw_flood_table_type table_type,
+		    unsigned int flood_table)
+{
+	MLXSW_REG_ZERO(sfgc, payload);
+	mlxsw_reg_sfgc_type_set(payload, type);
+	mlxsw_reg_sfgc_bridge_type_set(payload, bridge_type);
+	mlxsw_reg_sfgc_table_type_set(payload, table_type);
+	mlxsw_reg_sfgc_flood_table_set(payload, flood_table);
+	mlxsw_reg_sfgc_mid_set(payload, MLXSW_PORT_MID);
+}
+
+/* SFTR - Switch Flooding Table Register
+ * -------------------------------------
+ * The switch flooding table is used for flooding packet replication. The table
+ * defines a bit mask of ports for packet replication.
+ */
+#define MLXSW_REG_SFTR_ID 0x2012
+#define MLXSW_REG_SFTR_LEN 0x420
+
+static const struct mlxsw_reg_info mlxsw_reg_sftr = {
+	.id = MLXSW_REG_SFTR_ID,
+	.len = MLXSW_REG_SFTR_LEN,
+};
+
+/* reg_sftr_swid
+ * Switch partition ID with which to associate the port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sftr, swid, 0x00, 24, 8);
+
+/* reg_sftr_flood_table
+ * Flooding table index to associate with the specific type on the specific
+ * switch partition.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sftr, flood_table, 0x00, 16, 6);
+
+/* reg_sftr_index
+ * Index. Used as an index into the Flooding Table in case the table is
+ * configured to use VID / FID or FID Offset.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sftr, index, 0x00, 0, 16);
+
+/* reg_sftr_table_type
+ * See mlxsw_flood_table_type
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sftr, table_type, 0x04, 16, 3);
+
+/* reg_sftr_range
+ * Range of entries to update
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sftr, range, 0x04, 0, 16);
+
+/* reg_sftr_port
+ * Local port membership (1 bit per port).
+ * Access: RW
+ */
+MLXSW_ITEM_BIT_ARRAY(reg, sftr, port, 0x20, 0x20, 1);
+
+/* reg_sftr_cpu_port_mask
+ * CPU port mask (1 bit per port).
+ * Access: W
+ */
+MLXSW_ITEM_BIT_ARRAY(reg, sftr, port_mask, 0x220, 0x20, 1);
+
+static inline void mlxsw_reg_sftr_pack(char *payload,
+				       unsigned int flood_table,
+				       unsigned int index,
+				       enum mlxsw_flood_table_type table_type,
+				       unsigned int range, u8 port, bool set)
+{
+	MLXSW_REG_ZERO(sftr, payload);
+	mlxsw_reg_sftr_swid_set(payload, 0);
+	mlxsw_reg_sftr_flood_table_set(payload, flood_table);
+	mlxsw_reg_sftr_index_set(payload, index);
+	mlxsw_reg_sftr_table_type_set(payload, table_type);
+	mlxsw_reg_sftr_range_set(payload, range);
+	mlxsw_reg_sftr_port_set(payload, port, set);
+	mlxsw_reg_sftr_port_mask_set(payload, port, 1);
+}
+
+/* SPMLR - Switch Port MAC Learning Register
+ * -----------------------------------------
+ * Controls the Switch MAC learning policy per port.
+ */
+#define MLXSW_REG_SPMLR_ID 0x2018
+#define MLXSW_REG_SPMLR_LEN 0x8
+
+static const struct mlxsw_reg_info mlxsw_reg_spmlr = {
+	.id = MLXSW_REG_SPMLR_ID,
+	.len = MLXSW_REG_SPMLR_LEN,
+};
+
+/* reg_spmlr_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spmlr, local_port, 0x00, 16, 8);
+
+/* reg_spmlr_sub_port
+ * Virtual port within the physical port.
+ * Should be set to 0 when virtual ports are not enabled on the port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, spmlr, sub_port, 0x00, 8, 8);
+
+enum mlxsw_reg_spmlr_learn_mode {
+	MLXSW_REG_SPMLR_LEARN_MODE_DISABLE = 0,
+	MLXSW_REG_SPMLR_LEARN_MODE_ENABLE = 2,
+	MLXSW_REG_SPMLR_LEARN_MODE_SEC = 3,
+};
+
+/* reg_spmlr_learn_mode
+ * Learning mode on the port.
+ * 0 - Learning disabled.
+ * 2 - Learning enabled.
+ * 3 - Security mode.
+ *
+ * In security mode the switch does not learn MACs on the port, but uses the
+ * SMAC to see if it exists on another ingress port. If so, the packet is
+ * classified as a bad packet and is discarded unless the software registers
+ * to receive port security error packets usign HPKT.
+ */
+MLXSW_ITEM32(reg, spmlr, learn_mode, 0x04, 30, 2);
+
+static inline void mlxsw_reg_spmlr_pack(char *payload, u8 local_port,
+					enum mlxsw_reg_spmlr_learn_mode mode)
+{
+	MLXSW_REG_ZERO(spmlr, payload);
+	mlxsw_reg_spmlr_local_port_set(payload, local_port);
+	mlxsw_reg_spmlr_sub_port_set(payload, 0);
+	mlxsw_reg_spmlr_learn_mode_set(payload, mode);
+}
+
+/* SVFA - Switch VID to FID Allocation Register
+ * --------------------------------------------
+ * Controls the VID to FID mapping and {Port, VID} to FID mapping for
+ * virtualized ports.
+ */
+#define MLXSW_REG_SVFA_ID 0x201C
+#define MLXSW_REG_SVFA_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_svfa = {
+	.id = MLXSW_REG_SVFA_ID,
+	.len = MLXSW_REG_SVFA_LEN,
+};
+
+/* reg_svfa_swid
+ * Switch partition ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, svfa, swid, 0x00, 24, 8);
+
+/* reg_svfa_local_port
+ * Local port number.
+ * Access: Index
+ *
+ * Note: Reserved for 802.1Q FIDs.
+ */
+MLXSW_ITEM32(reg, svfa, local_port, 0x00, 16, 8);
+
+enum mlxsw_reg_svfa_mt {
+	MLXSW_REG_SVFA_MT_VID_TO_FID,
+	MLXSW_REG_SVFA_MT_PORT_VID_TO_FID,
+};
+
+/* reg_svfa_mapping_table
+ * Mapping table:
+ * 0 - VID to FID
+ * 1 - {Port, VID} to FID
+ * Access: Index
+ *
+ * Note: Reserved for SwitchX-2.
+ */
+MLXSW_ITEM32(reg, svfa, mapping_table, 0x00, 8, 3);
+
+/* reg_svfa_v
+ * Valid.
+ * Valid if set.
+ * Access: RW
+ *
+ * Note: Reserved for SwitchX-2.
+ */
+MLXSW_ITEM32(reg, svfa, v, 0x00, 0, 1);
+
+/* reg_svfa_fid
+ * Filtering ID.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, svfa, fid, 0x04, 16, 16);
+
+/* reg_svfa_vid
+ * VLAN ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, svfa, vid, 0x04, 0, 12);
+
+/* reg_svfa_counter_set_type
+ * Counter set type for flow counters.
+ * Access: RW
+ *
+ * Note: Reserved for SwitchX-2.
+ */
+MLXSW_ITEM32(reg, svfa, counter_set_type, 0x08, 24, 8);
+
+/* reg_svfa_counter_index
+ * Counter index for flow counters.
+ * Access: RW
+ *
+ * Note: Reserved for SwitchX-2.
+ */
+MLXSW_ITEM32(reg, svfa, counter_index, 0x08, 0, 24);
+
+static inline void mlxsw_reg_svfa_pack(char *payload, u8 local_port,
+				       enum mlxsw_reg_svfa_mt mt, bool valid,
+				       u16 fid, u16 vid)
+{
+	MLXSW_REG_ZERO(svfa, payload);
+	local_port = mt == MLXSW_REG_SVFA_MT_VID_TO_FID ? 0 : local_port;
+	mlxsw_reg_svfa_swid_set(payload, 0);
+	mlxsw_reg_svfa_local_port_set(payload, local_port);
+	mlxsw_reg_svfa_mapping_table_set(payload, mt);
+	mlxsw_reg_svfa_v_set(payload, valid);
+	mlxsw_reg_svfa_fid_set(payload, fid);
+	mlxsw_reg_svfa_vid_set(payload, vid);
+}
+
+/* SVPE - Switch Virtual-Port Enabling Register
+ * --------------------------------------------
+ * Enables port virtualization.
+ */
+#define MLXSW_REG_SVPE_ID 0x201E
+#define MLXSW_REG_SVPE_LEN 0x4
+
+static const struct mlxsw_reg_info mlxsw_reg_svpe = {
+	.id = MLXSW_REG_SVPE_ID,
+	.len = MLXSW_REG_SVPE_LEN,
+};
+
+/* reg_svpe_local_port
+ * Local port number
+ * Access: Index
+ *
+ * Note: CPU port is not supported (uses VLAN mode only).
+ */
+MLXSW_ITEM32(reg, svpe, local_port, 0x00, 16, 8);
+
+/* reg_svpe_vp_en
+ * Virtual port enable.
+ * 0 - Disable, VLAN mode (VID to FID).
+ * 1 - Enable, Virtual port mode ({Port, VID} to FID).
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, svpe, vp_en, 0x00, 8, 1);
+
+static inline void mlxsw_reg_svpe_pack(char *payload, u8 local_port,
+				       bool enable)
+{
+	MLXSW_REG_ZERO(svpe, payload);
+	mlxsw_reg_svpe_local_port_set(payload, local_port);
+	mlxsw_reg_svpe_vp_en_set(payload, enable);
+}
+
+/* SFMR - Switch FID Management Register
+ * -------------------------------------
+ * Creates and configures FIDs.
+ */
+#define MLXSW_REG_SFMR_ID 0x201F
+#define MLXSW_REG_SFMR_LEN 0x18
+
+static const struct mlxsw_reg_info mlxsw_reg_sfmr = {
+	.id = MLXSW_REG_SFMR_ID,
+	.len = MLXSW_REG_SFMR_LEN,
+};
+
+enum mlxsw_reg_sfmr_op {
+	MLXSW_REG_SFMR_OP_CREATE_FID,
+	MLXSW_REG_SFMR_OP_DESTROY_FID,
+};
+
+/* reg_sfmr_op
+ * Operation.
+ * 0 - Create or edit FID.
+ * 1 - Destroy FID.
+ * Access: WO
+ */
+MLXSW_ITEM32(reg, sfmr, op, 0x00, 24, 4);
+
+/* reg_sfmr_fid
+ * Filtering ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sfmr, fid, 0x00, 0, 16);
+
+/* reg_sfmr_fid_offset
+ * FID offset.
+ * Used to point into the flooding table selected by SFGC register if
+ * the table is of type FID-Offset. Otherwise, this field is reserved.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfmr, fid_offset, 0x08, 0, 16);
+
+/* reg_sfmr_vtfp
+ * Valid Tunnel Flood Pointer.
+ * If not set, then nve_tunnel_flood_ptr is reserved and considered NULL.
+ * Access: RW
+ *
+ * Note: Reserved for 802.1Q FIDs.
+ */
+MLXSW_ITEM32(reg, sfmr, vtfp, 0x0C, 31, 1);
+
+/* reg_sfmr_nve_tunnel_flood_ptr
+ * Underlay Flooding and BC Pointer.
+ * Used as a pointer to the first entry of the group based link lists of
+ * flooding or BC entries (for NVE tunnels).
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sfmr, nve_tunnel_flood_ptr, 0x0C, 0, 24);
+
+/* reg_sfmr_vv
+ * VNI Valid.
+ * If not set, then vni is reserved.
+ * Access: RW
+ *
+ * Note: Reserved for 802.1Q FIDs.
+ */
+MLXSW_ITEM32(reg, sfmr, vv, 0x10, 31, 1);
+
+/* reg_sfmr_vni
+ * Virtual Network Identifier.
+ * Access: RW
+ *
+ * Note: A given VNI can only be assigned to one FID.
+ */
+MLXSW_ITEM32(reg, sfmr, vni, 0x10, 0, 24);
+
+static inline void mlxsw_reg_sfmr_pack(char *payload,
+				       enum mlxsw_reg_sfmr_op op, u16 fid,
+				       u16 fid_offset)
+{
+	MLXSW_REG_ZERO(sfmr, payload);
+	mlxsw_reg_sfmr_op_set(payload, op);
+	mlxsw_reg_sfmr_fid_set(payload, fid);
+	mlxsw_reg_sfmr_fid_offset_set(payload, fid_offset);
+	mlxsw_reg_sfmr_vtfp_set(payload, false);
+	mlxsw_reg_sfmr_vv_set(payload, false);
+}
+
+/* SPVMLR - Switch Port VLAN MAC Learning Register
+ * -----------------------------------------------
+ * Controls the switch MAC learning policy per {Port, VID}.
+ */
+#define MLXSW_REG_SPVMLR_ID 0x2020
+#define MLXSW_REG_SPVMLR_BASE_LEN 0x04 /* base length, without records */
+#define MLXSW_REG_SPVMLR_REC_LEN 0x04 /* record length */
+#define MLXSW_REG_SPVMLR_REC_MAX_COUNT 256
+#define MLXSW_REG_SPVMLR_LEN (MLXSW_REG_SPVMLR_BASE_LEN + \
+			      MLXSW_REG_SPVMLR_REC_LEN * \
+			      MLXSW_REG_SPVMLR_REC_MAX_COUNT)
+
+static const struct mlxsw_reg_info mlxsw_reg_spvmlr = {
+	.id = MLXSW_REG_SPVMLR_ID,
+	.len = MLXSW_REG_SPVMLR_LEN,
+};
+
+/* reg_spvmlr_local_port
+ * Local ingress port.
+ * Access: Index
+ *
+ * Note: CPU port is not supported.
+ */
+MLXSW_ITEM32(reg, spvmlr, local_port, 0x00, 16, 8);
+
+/* reg_spvmlr_num_rec
+ * Number of records to update.
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, spvmlr, num_rec, 0x00, 0, 8);
+
+/* reg_spvmlr_rec_learn_enable
+ * 0 - Disable learning for {Port, VID}.
+ * 1 - Enable learning for {Port, VID}.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, spvmlr, rec_learn_enable, MLXSW_REG_SPVMLR_BASE_LEN,
+		     31, 1, MLXSW_REG_SPVMLR_REC_LEN, 0x00, false);
+
+/* reg_spvmlr_rec_vid
+ * VLAN ID to be added/removed from port or for querying.
+ * Access: Index
+ */
+MLXSW_ITEM32_INDEXED(reg, spvmlr, rec_vid, MLXSW_REG_SPVMLR_BASE_LEN, 0, 12,
+		     MLXSW_REG_SPVMLR_REC_LEN, 0x00, false);
+
+static inline void mlxsw_reg_spvmlr_pack(char *payload, u8 local_port,
+					 u16 vid_begin, u16 vid_end,
+					 bool learn_enable)
+{
+	int num_rec = vid_end - vid_begin + 1;
+	int i;
+
+	WARN_ON(num_rec < 1 || num_rec > MLXSW_REG_SPVMLR_REC_MAX_COUNT);
+
+	MLXSW_REG_ZERO(spvmlr, payload);
+	mlxsw_reg_spvmlr_local_port_set(payload, local_port);
+	mlxsw_reg_spvmlr_num_rec_set(payload, num_rec);
+
+	for (i = 0; i < num_rec; i++) {
+		mlxsw_reg_spvmlr_rec_learn_enable_set(payload, i, learn_enable);
+		mlxsw_reg_spvmlr_rec_vid_set(payload, i, vid_begin + i);
+	}
+}
+
+/* PMLP - Ports Module to Local Port Register
+ * ------------------------------------------
+ * Configures the assignment of modules to local ports.
+ */
+#define MLXSW_REG_PMLP_ID 0x5002
+#define MLXSW_REG_PMLP_LEN 0x40
+
+static const struct mlxsw_reg_info mlxsw_reg_pmlp = {
+	.id = MLXSW_REG_PMLP_ID,
+	.len = MLXSW_REG_PMLP_LEN,
+};
+
+/* reg_pmlp_rxtx
+ * 0 - Tx value is used for both Tx and Rx.
+ * 1 - Rx value is taken from a separte field.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, pmlp, rxtx, 0x00, 31, 1);
+
+/* reg_pmlp_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pmlp, local_port, 0x00, 16, 8);
+
+/* reg_pmlp_width
+ * 0 - Unmap local port.
+ * 1 - Lane 0 is used.
+ * 2 - Lanes 0 and 1 are used.
+ * 4 - Lanes 0, 1, 2 and 3 are used.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, pmlp, width, 0x00, 0, 8);
+
+/* reg_pmlp_module
+ * Module number.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, pmlp, module, 0x04, 0, 8, 0x04, 0, false);
+
+/* reg_pmlp_tx_lane
+ * Tx Lane. When rxtx field is cleared, this field is used for Rx as well.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, pmlp, tx_lane, 0x04, 16, 2, 0x04, 16, false);
+
+/* reg_pmlp_rx_lane
+ * Rx Lane. When rxtx field is cleared, this field is ignored and Rx lane is
+ * equal to Tx lane.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, pmlp, rx_lane, 0x04, 24, 2, 0x04, 24, false);
+
+static inline void mlxsw_reg_pmlp_pack(char *payload, u8 local_port)
+{
+	MLXSW_REG_ZERO(pmlp, payload);
+	mlxsw_reg_pmlp_local_port_set(payload, local_port);
+}
+
+/* PMTU - Port MTU Register
+ * ------------------------
+ * Configures and reports the port MTU.
+ */
+#define MLXSW_REG_PMTU_ID 0x5003
+#define MLXSW_REG_PMTU_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_pmtu = {
+	.id = MLXSW_REG_PMTU_ID,
+	.len = MLXSW_REG_PMTU_LEN,
+};
+
+/* reg_pmtu_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pmtu, local_port, 0x00, 16, 8);
+
+/* reg_pmtu_max_mtu
+ * Maximum MTU.
+ * When port type (e.g. Ethernet) is configured, the relevant MTU is
+ * reported, otherwise the minimum between the max_mtu of the different
+ * types is reported.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, pmtu, max_mtu, 0x04, 16, 16);
+
+/* reg_pmtu_admin_mtu
+ * MTU value to set port to. Must be smaller or equal to max_mtu.
+ * Note: If port type is Infiniband, then port must be disabled, when its
+ * MTU is set.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, pmtu, admin_mtu, 0x08, 16, 16);
+
+/* reg_pmtu_oper_mtu
+ * The actual MTU configured on the port. Packets exceeding this size
+ * will be dropped.
+ * Note: In Ethernet and FC oper_mtu == admin_mtu, however, in Infiniband
+ * oper_mtu might be smaller than admin_mtu.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, pmtu, oper_mtu, 0x0C, 16, 16);
+
+static inline void mlxsw_reg_pmtu_pack(char *payload, u8 local_port,
+				       u16 new_mtu)
+{
+	MLXSW_REG_ZERO(pmtu, payload);
+	mlxsw_reg_pmtu_local_port_set(payload, local_port);
+	mlxsw_reg_pmtu_max_mtu_set(payload, 0);
+	mlxsw_reg_pmtu_admin_mtu_set(payload, new_mtu);
+	mlxsw_reg_pmtu_oper_mtu_set(payload, 0);
+}
+
+/* PTYS - Port Type and Speed Register
+ * -----------------------------------
+ * Configures and reports the port speed type.
+ *
+ * Note: When set while the link is up, the changes will not take effect
+ * until the port transitions from down to up state.
+ */
+#define MLXSW_REG_PTYS_ID 0x5004
+#define MLXSW_REG_PTYS_LEN 0x40
+
+static const struct mlxsw_reg_info mlxsw_reg_ptys = {
+	.id = MLXSW_REG_PTYS_ID,
+	.len = MLXSW_REG_PTYS_LEN,
+};
+
+/* reg_ptys_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ptys, local_port, 0x00, 16, 8);
+
+#define MLXSW_REG_PTYS_PROTO_MASK_ETH	BIT(2)
+
+/* reg_ptys_proto_mask
+ * Protocol mask. Indicates which protocol is used.
+ * 0 - Infiniband.
+ * 1 - Fibre Channel.
+ * 2 - Ethernet.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ptys, proto_mask, 0x00, 0, 3);
+
+#define MLXSW_REG_PTYS_ETH_SPEED_SGMII			BIT(0)
+#define MLXSW_REG_PTYS_ETH_SPEED_1000BASE_KX		BIT(1)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CX4		BIT(2)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4		BIT(3)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR		BIT(4)
+#define MLXSW_REG_PTYS_ETH_SPEED_20GBASE_KR2		BIT(5)
+#define MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4		BIT(6)
+#define MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4		BIT(7)
+#define MLXSW_REG_PTYS_ETH_SPEED_56GBASE_R4		BIT(8)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR		BIT(12)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR		BIT(13)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_ER_LR		BIT(14)
+#define MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4		BIT(15)
+#define MLXSW_REG_PTYS_ETH_SPEED_40GBASE_LR4_ER4	BIT(16)
+#define MLXSW_REG_PTYS_ETH_SPEED_50GBASE_KR4		BIT(19)
+#define MLXSW_REG_PTYS_ETH_SPEED_100GBASE_CR4		BIT(20)
+#define MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4		BIT(21)
+#define MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4		BIT(22)
+#define MLXSW_REG_PTYS_ETH_SPEED_100GBASE_LR4_ER4	BIT(23)
+#define MLXSW_REG_PTYS_ETH_SPEED_100BASE_TX		BIT(24)
+#define MLXSW_REG_PTYS_ETH_SPEED_100BASE_T		BIT(25)
+#define MLXSW_REG_PTYS_ETH_SPEED_10GBASE_T		BIT(26)
+#define MLXSW_REG_PTYS_ETH_SPEED_25GBASE_CR		BIT(27)
+#define MLXSW_REG_PTYS_ETH_SPEED_25GBASE_KR		BIT(28)
+#define MLXSW_REG_PTYS_ETH_SPEED_25GBASE_SR		BIT(29)
+#define MLXSW_REG_PTYS_ETH_SPEED_50GBASE_CR2		BIT(30)
+#define MLXSW_REG_PTYS_ETH_SPEED_50GBASE_KR2		BIT(31)
+
+/* reg_ptys_eth_proto_cap
+ * Ethernet port supported speeds and protocols.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, ptys, eth_proto_cap, 0x0C, 0, 32);
+
+/* reg_ptys_eth_proto_admin
+ * Speed and protocol to set port to.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ptys, eth_proto_admin, 0x18, 0, 32);
+
+/* reg_ptys_eth_proto_oper
+ * The current speed and protocol configured for the port.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, ptys, eth_proto_oper, 0x24, 0, 32);
+
+static inline void mlxsw_reg_ptys_pack(char *payload, u8 local_port,
+				       u32 proto_admin)
+{
+	MLXSW_REG_ZERO(ptys, payload);
+	mlxsw_reg_ptys_local_port_set(payload, local_port);
+	mlxsw_reg_ptys_proto_mask_set(payload, MLXSW_REG_PTYS_PROTO_MASK_ETH);
+	mlxsw_reg_ptys_eth_proto_admin_set(payload, proto_admin);
+}
+
+static inline void mlxsw_reg_ptys_unpack(char *payload, u32 *p_eth_proto_cap,
+					 u32 *p_eth_proto_adm,
+					 u32 *p_eth_proto_oper)
+{
+	if (p_eth_proto_cap)
+		*p_eth_proto_cap = mlxsw_reg_ptys_eth_proto_cap_get(payload);
+	if (p_eth_proto_adm)
+		*p_eth_proto_adm = mlxsw_reg_ptys_eth_proto_admin_get(payload);
+	if (p_eth_proto_oper)
+		*p_eth_proto_oper = mlxsw_reg_ptys_eth_proto_oper_get(payload);
+}
+
+/* PPAD - Port Physical Address Register
+ * -------------------------------------
+ * The PPAD register configures the per port physical MAC address.
+ */
+#define MLXSW_REG_PPAD_ID 0x5005
+#define MLXSW_REG_PPAD_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_ppad = {
+	.id = MLXSW_REG_PPAD_ID,
+	.len = MLXSW_REG_PPAD_LEN,
+};
+
+/* reg_ppad_single_base_mac
+ * 0: base_mac, local port should be 0 and mac[7:0] is
+ * reserved. HW will set incremental
+ * 1: single_mac - mac of the local_port
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ppad, single_base_mac, 0x00, 28, 1);
+
+/* reg_ppad_local_port
+ * port number, if single_base_mac = 0 then local_port is reserved
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, ppad, local_port, 0x00, 16, 8);
+
+/* reg_ppad_mac
+ * If single_base_mac = 0 - base MAC address, mac[7:0] is reserved.
+ * If single_base_mac = 1 - the per port MAC address
+ * Access: RW
+ */
+MLXSW_ITEM_BUF(reg, ppad, mac, 0x02, 6);
+
+static inline void mlxsw_reg_ppad_pack(char *payload, bool single_base_mac,
+				       u8 local_port)
+{
+	MLXSW_REG_ZERO(ppad, payload);
+	mlxsw_reg_ppad_single_base_mac_set(payload, !!single_base_mac);
+	mlxsw_reg_ppad_local_port_set(payload, local_port);
+}
+
+/* PAOS - Ports Administrative and Operational Status Register
+ * -----------------------------------------------------------
+ * Configures and retrieves per port administrative and operational status.
+ */
+#define MLXSW_REG_PAOS_ID 0x5006
+#define MLXSW_REG_PAOS_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_paos = {
+	.id = MLXSW_REG_PAOS_ID,
+	.len = MLXSW_REG_PAOS_LEN,
+};
+
+/* reg_paos_swid
+ * Switch partition ID with which to associate the port.
+ * Note: while external ports uses unique local port numbers (and thus swid is
+ * redundant), router ports use the same local port number where swid is the
+ * only indication for the relevant port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, paos, swid, 0x00, 24, 8);
+
+/* reg_paos_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, paos, local_port, 0x00, 16, 8);
+
+/* reg_paos_admin_status
+ * Port administrative state (the desired state of the port):
+ * 1 - Up.
+ * 2 - Down.
+ * 3 - Up once. This means that in case of link failure, the port won't go
+ *     into polling mode, but will wait to be re-enabled by software.
+ * 4 - Disabled by system. Can only be set by hardware.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, paos, admin_status, 0x00, 8, 4);
+
+/* reg_paos_oper_status
+ * Port operational state (the current state):
+ * 1 - Up.
+ * 2 - Down.
+ * 3 - Down by port failure. This means that the device will not let the
+ *     port up again until explicitly specified by software.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, paos, oper_status, 0x00, 0, 4);
+
+/* reg_paos_ase
+ * Admin state update enabled.
+ * Access: WO
+ */
+MLXSW_ITEM32(reg, paos, ase, 0x04, 31, 1);
+
+/* reg_paos_ee
+ * Event update enable. If this bit is set, event generation will be
+ * updated based on the e field.
+ * Access: WO
+ */
+MLXSW_ITEM32(reg, paos, ee, 0x04, 30, 1);
+
+/* reg_paos_e
+ * Event generation on operational state change:
+ * 0 - Do not generate event.
+ * 1 - Generate Event.
+ * 2 - Generate Single Event.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, paos, e, 0x04, 0, 2);
+
+static inline void mlxsw_reg_paos_pack(char *payload, u8 local_port,
+				       enum mlxsw_port_admin_status status)
+{
+	MLXSW_REG_ZERO(paos, payload);
+	mlxsw_reg_paos_swid_set(payload, 0);
+	mlxsw_reg_paos_local_port_set(payload, local_port);
+	mlxsw_reg_paos_admin_status_set(payload, status);
+	mlxsw_reg_paos_oper_status_set(payload, 0);
+	mlxsw_reg_paos_ase_set(payload, 1);
+	mlxsw_reg_paos_ee_set(payload, 1);
+	mlxsw_reg_paos_e_set(payload, 1);
+}
+
+/* PPCNT - Ports Performance Counters Register
+ * -------------------------------------------
+ * The PPCNT register retrieves per port performance counters.
+ */
+#define MLXSW_REG_PPCNT_ID 0x5008
+#define MLXSW_REG_PPCNT_LEN 0x100
+
+static const struct mlxsw_reg_info mlxsw_reg_ppcnt = {
+	.id = MLXSW_REG_PPCNT_ID,
+	.len = MLXSW_REG_PPCNT_LEN,
+};
+
+/* reg_ppcnt_swid
+ * For HCA: must be always 0.
+ * Switch partition ID to associate port with.
+ * Switch partitions are numbered from 0 to 7 inclusively.
+ * Switch partition 254 indicates stacking ports.
+ * Switch partition 255 indicates all switch partitions.
+ * Only valid on Set() operation with local_port=255.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ppcnt, swid, 0x00, 24, 8);
+
+/* reg_ppcnt_local_port
+ * Local port number.
+ * 255 indicates all ports on the device, and is only allowed
+ * for Set() operation.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ppcnt, local_port, 0x00, 16, 8);
+
+/* reg_ppcnt_pnat
+ * Port number access type:
+ * 0 - Local port number
+ * 1 - IB port number
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ppcnt, pnat, 0x00, 14, 2);
+
+/* reg_ppcnt_grp
+ * Performance counter group.
+ * Group 63 indicates all groups. Only valid on Set() operation with
+ * clr bit set.
+ * 0x0: IEEE 802.3 Counters
+ * 0x1: RFC 2863 Counters
+ * 0x2: RFC 2819 Counters
+ * 0x3: RFC 3635 Counters
+ * 0x5: Ethernet Extended Counters
+ * 0x8: Link Level Retransmission Counters
+ * 0x10: Per Priority Counters
+ * 0x11: Per Traffic Class Counters
+ * 0x12: Physical Layer Counters
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ppcnt, grp, 0x00, 0, 6);
+
+/* reg_ppcnt_clr
+ * Clear counters. Setting the clr bit will reset the counter value
+ * for all counters in the counter group. This bit can be set
+ * for both Set() and Get() operation.
+ * Access: OP
+ */
+MLXSW_ITEM32(reg, ppcnt, clr, 0x04, 31, 1);
+
+/* reg_ppcnt_prio_tc
+ * Priority for counter set that support per priority, valid values: 0-7.
+ * Traffic class for counter set that support per traffic class,
+ * valid values: 0- cap_max_tclass-1 .
+ * For HCA: cap_max_tclass is always 8.
+ * Otherwise must be 0.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, ppcnt, prio_tc, 0x04, 0, 5);
+
+/* reg_ppcnt_a_frames_transmitted_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_frames_transmitted_ok,
+	     0x08 + 0x00, 0, 64);
+
+/* reg_ppcnt_a_frames_received_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_frames_received_ok,
+	     0x08 + 0x08, 0, 64);
+
+/* reg_ppcnt_a_frame_check_sequence_errors
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_frame_check_sequence_errors,
+	     0x08 + 0x10, 0, 64);
+
+/* reg_ppcnt_a_alignment_errors
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_alignment_errors,
+	     0x08 + 0x18, 0, 64);
+
+/* reg_ppcnt_a_octets_transmitted_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_octets_transmitted_ok,
+	     0x08 + 0x20, 0, 64);
+
+/* reg_ppcnt_a_octets_received_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_octets_received_ok,
+	     0x08 + 0x28, 0, 64);
+
+/* reg_ppcnt_a_multicast_frames_xmitted_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_multicast_frames_xmitted_ok,
+	     0x08 + 0x30, 0, 64);
+
+/* reg_ppcnt_a_broadcast_frames_xmitted_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_broadcast_frames_xmitted_ok,
+	     0x08 + 0x38, 0, 64);
+
+/* reg_ppcnt_a_multicast_frames_received_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_multicast_frames_received_ok,
+	     0x08 + 0x40, 0, 64);
+
+/* reg_ppcnt_a_broadcast_frames_received_ok
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_broadcast_frames_received_ok,
+	     0x08 + 0x48, 0, 64);
+
+/* reg_ppcnt_a_in_range_length_errors
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_in_range_length_errors,
+	     0x08 + 0x50, 0, 64);
+
+/* reg_ppcnt_a_out_of_range_length_field
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_out_of_range_length_field,
+	     0x08 + 0x58, 0, 64);
+
+/* reg_ppcnt_a_frame_too_long_errors
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_frame_too_long_errors,
+	     0x08 + 0x60, 0, 64);
+
+/* reg_ppcnt_a_symbol_error_during_carrier
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_symbol_error_during_carrier,
+	     0x08 + 0x68, 0, 64);
+
+/* reg_ppcnt_a_mac_control_frames_transmitted
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_mac_control_frames_transmitted,
+	     0x08 + 0x70, 0, 64);
+
+/* reg_ppcnt_a_mac_control_frames_received
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_mac_control_frames_received,
+	     0x08 + 0x78, 0, 64);
+
+/* reg_ppcnt_a_unsupported_opcodes_received
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_unsupported_opcodes_received,
+	     0x08 + 0x80, 0, 64);
+
+/* reg_ppcnt_a_pause_mac_ctrl_frames_received
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_pause_mac_ctrl_frames_received,
+	     0x08 + 0x88, 0, 64);
+
+/* reg_ppcnt_a_pause_mac_ctrl_frames_transmitted
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, a_pause_mac_ctrl_frames_transmitted,
+	     0x08 + 0x90, 0, 64);
+
+static inline void mlxsw_reg_ppcnt_pack(char *payload, u8 local_port)
+{
+	MLXSW_REG_ZERO(ppcnt, payload);
+	mlxsw_reg_ppcnt_swid_set(payload, 0);
+	mlxsw_reg_ppcnt_local_port_set(payload, local_port);
+	mlxsw_reg_ppcnt_pnat_set(payload, 0);
+	mlxsw_reg_ppcnt_grp_set(payload, 0);
+	mlxsw_reg_ppcnt_clr_set(payload, 0);
+	mlxsw_reg_ppcnt_prio_tc_set(payload, 0);
+}
+
+/* PBMC - Port Buffer Management Control Register
+ * ----------------------------------------------
+ * The PBMC register configures and retrieves the port packet buffer
+ * allocation for different Prios, and the Pause threshold management.
+ */
+#define MLXSW_REG_PBMC_ID 0x500C
+#define MLXSW_REG_PBMC_LEN 0x68
+
+static const struct mlxsw_reg_info mlxsw_reg_pbmc = {
+	.id = MLXSW_REG_PBMC_ID,
+	.len = MLXSW_REG_PBMC_LEN,
+};
+
+/* reg_pbmc_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pbmc, local_port, 0x00, 16, 8);
+
+/* reg_pbmc_xoff_timer_value
+ * When device generates a pause frame, it uses this value as the pause
+ * timer (time for the peer port to pause in quota-512 bit time).
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, pbmc, xoff_timer_value, 0x04, 16, 16);
+
+/* reg_pbmc_xoff_refresh
+ * The time before a new pause frame should be sent to refresh the pause RW
+ * state. Using the same units as xoff_timer_value above (in quota-512 bit
+ * time).
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, pbmc, xoff_refresh, 0x04, 0, 16);
+
+/* reg_pbmc_buf_lossy
+ * The field indicates if the buffer is lossy.
+ * 0 - Lossless
+ * 1 - Lossy
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, pbmc, buf_lossy, 0x0C, 25, 1, 0x08, 0x00, false);
+
+/* reg_pbmc_buf_epsb
+ * Eligible for Port Shared buffer.
+ * If epsb is set, packets assigned to buffer are allowed to insert the port
+ * shared buffer.
+ * When buf_lossy is MLXSW_REG_PBMC_LOSSY_LOSSY this field is reserved.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, pbmc, buf_epsb, 0x0C, 24, 1, 0x08, 0x00, false);
+
+/* reg_pbmc_buf_size
+ * The part of the packet buffer array is allocated for the specific buffer.
+ * Units are represented in cells.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, pbmc, buf_size, 0x0C, 0, 16, 0x08, 0x00, false);
+
+static inline void mlxsw_reg_pbmc_pack(char *payload, u8 local_port,
+				       u16 xoff_timer_value, u16 xoff_refresh)
+{
+	MLXSW_REG_ZERO(pbmc, payload);
+	mlxsw_reg_pbmc_local_port_set(payload, local_port);
+	mlxsw_reg_pbmc_xoff_timer_value_set(payload, xoff_timer_value);
+	mlxsw_reg_pbmc_xoff_refresh_set(payload, xoff_refresh);
+}
+
+static inline void mlxsw_reg_pbmc_lossy_buffer_pack(char *payload,
+						    int buf_index,
+						    u16 size)
+{
+	mlxsw_reg_pbmc_buf_lossy_set(payload, buf_index, 1);
+	mlxsw_reg_pbmc_buf_epsb_set(payload, buf_index, 0);
+	mlxsw_reg_pbmc_buf_size_set(payload, buf_index, size);
+}
+
+/* PSPA - Port Switch Partition Allocation
+ * ---------------------------------------
+ * Controls the association of a port with a switch partition and enables
+ * configuring ports as stacking ports.
+ */
+#define MLXSW_REG_PSPA_ID 0x500D
+#define MLXSW_REG_PSPA_LEN 0x8
+
+static const struct mlxsw_reg_info mlxsw_reg_pspa = {
+	.id = MLXSW_REG_PSPA_ID,
+	.len = MLXSW_REG_PSPA_LEN,
+};
+
+/* reg_pspa_swid
+ * Switch partition ID.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, pspa, swid, 0x00, 24, 8);
+
+/* reg_pspa_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pspa, local_port, 0x00, 16, 8);
+
+/* reg_pspa_sub_port
+ * Virtual port within the local port. Set to 0 when virtual ports are
+ * disabled on the local port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pspa, sub_port, 0x00, 8, 8);
+
+static inline void mlxsw_reg_pspa_pack(char *payload, u8 swid, u8 local_port)
+{
+	MLXSW_REG_ZERO(pspa, payload);
+	mlxsw_reg_pspa_swid_set(payload, swid);
+	mlxsw_reg_pspa_local_port_set(payload, local_port);
+	mlxsw_reg_pspa_sub_port_set(payload, 0);
+}
+
+/* HTGT - Host Trap Group Table
+ * ----------------------------
+ * Configures the properties for forwarding to CPU.
+ */
+#define MLXSW_REG_HTGT_ID 0x7002
+#define MLXSW_REG_HTGT_LEN 0x100
+
+static const struct mlxsw_reg_info mlxsw_reg_htgt = {
+	.id = MLXSW_REG_HTGT_ID,
+	.len = MLXSW_REG_HTGT_LEN,
+};
+
+/* reg_htgt_swid
+ * Switch partition ID.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, htgt, swid, 0x00, 24, 8);
+
+#define MLXSW_REG_HTGT_PATH_TYPE_LOCAL 0x0	/* For locally attached CPU */
+
+/* reg_htgt_type
+ * CPU path type.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, htgt, type, 0x00, 8, 4);
+
+enum mlxsw_reg_htgt_trap_group {
+	MLXSW_REG_HTGT_TRAP_GROUP_EMAD,
+	MLXSW_REG_HTGT_TRAP_GROUP_RX,
+	MLXSW_REG_HTGT_TRAP_GROUP_CTRL,
+};
+
+/* reg_htgt_trap_group
+ * Trap group number. User defined number specifying which trap groups
+ * should be forwarded to the CPU. The mapping between trap IDs and trap
+ * groups is configured using HPKT register.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, htgt, trap_group, 0x00, 0, 8);
+
+enum {
+	MLXSW_REG_HTGT_POLICER_DISABLE,
+	MLXSW_REG_HTGT_POLICER_ENABLE,
+};
+
+/* reg_htgt_pide
+ * Enable policer ID specified using 'pid' field.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, htgt, pide, 0x04, 15, 1);
+
+/* reg_htgt_pid
+ * Policer ID for the trap group.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, htgt, pid, 0x04, 0, 8);
+
+#define MLXSW_REG_HTGT_TRAP_TO_CPU 0x0
+
+/* reg_htgt_mirror_action
+ * Mirror action to use.
+ * 0 - Trap to CPU.
+ * 1 - Trap to CPU and mirror to a mirroring agent.
+ * 2 - Mirror to a mirroring agent and do not trap to CPU.
+ * Access: RW
+ *
+ * Note: Mirroring to a mirroring agent is only supported in Spectrum.
+ */
+MLXSW_ITEM32(reg, htgt, mirror_action, 0x08, 8, 2);
+
+/* reg_htgt_mirroring_agent
+ * Mirroring agent.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, htgt, mirroring_agent, 0x08, 0, 3);
+
+/* reg_htgt_priority
+ * Trap group priority.
+ * In case a packet matches multiple classification rules, the packet will
+ * only be trapped once, based on the trap ID associated with the group (via
+ * register HPKT) with the highest priority.
+ * Supported values are 0-7, with 7 represnting the highest priority.
+ * Access: RW
+ *
+ * Note: In SwitchX-2 this field is ignored and the priority value is replaced
+ * by the 'trap_group' field.
+ */
+MLXSW_ITEM32(reg, htgt, priority, 0x0C, 0, 4);
+
+/* reg_htgt_local_path_cpu_tclass
+ * CPU ingress traffic class for the trap group.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, htgt, local_path_cpu_tclass, 0x10, 16, 6);
+
+#define MLXSW_REG_HTGT_LOCAL_PATH_RDQ_EMAD	0x15
+#define MLXSW_REG_HTGT_LOCAL_PATH_RDQ_RX	0x14
+#define MLXSW_REG_HTGT_LOCAL_PATH_RDQ_CTRL	0x13
+
+/* reg_htgt_local_path_rdq
+ * Receive descriptor queue (RDQ) to use for the trap group.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, htgt, local_path_rdq, 0x10, 0, 6);
+
+static inline void mlxsw_reg_htgt_pack(char *payload,
+				       enum mlxsw_reg_htgt_trap_group group)
+{
+	u8 swid, rdq;
+
+	MLXSW_REG_ZERO(htgt, payload);
+	switch (group) {
+	case MLXSW_REG_HTGT_TRAP_GROUP_EMAD:
+		swid = MLXSW_PORT_SWID_ALL_SWIDS;
+		rdq = MLXSW_REG_HTGT_LOCAL_PATH_RDQ_EMAD;
+		break;
+	case MLXSW_REG_HTGT_TRAP_GROUP_RX:
+		swid = 0;
+		rdq = MLXSW_REG_HTGT_LOCAL_PATH_RDQ_RX;
+		break;
+	case MLXSW_REG_HTGT_TRAP_GROUP_CTRL:
+		swid = 0;
+		rdq = MLXSW_REG_HTGT_LOCAL_PATH_RDQ_CTRL;
+		break;
+	}
+	mlxsw_reg_htgt_swid_set(payload, swid);
+	mlxsw_reg_htgt_type_set(payload, MLXSW_REG_HTGT_PATH_TYPE_LOCAL);
+	mlxsw_reg_htgt_trap_group_set(payload, group);
+	mlxsw_reg_htgt_pide_set(payload, MLXSW_REG_HTGT_POLICER_DISABLE);
+	mlxsw_reg_htgt_pid_set(payload, 0);
+	mlxsw_reg_htgt_mirror_action_set(payload, MLXSW_REG_HTGT_TRAP_TO_CPU);
+	mlxsw_reg_htgt_mirroring_agent_set(payload, 0);
+	mlxsw_reg_htgt_priority_set(payload, 0);
+	mlxsw_reg_htgt_local_path_cpu_tclass_set(payload, 7);
+	mlxsw_reg_htgt_local_path_rdq_set(payload, rdq);
+}
+
+/* HPKT - Host Packet Trap
+ * -----------------------
+ * Configures trap IDs inside trap groups.
+ */
+#define MLXSW_REG_HPKT_ID 0x7003
+#define MLXSW_REG_HPKT_LEN 0x10
+
+static const struct mlxsw_reg_info mlxsw_reg_hpkt = {
+	.id = MLXSW_REG_HPKT_ID,
+	.len = MLXSW_REG_HPKT_LEN,
+};
+
+enum {
+	MLXSW_REG_HPKT_ACK_NOT_REQUIRED,
+	MLXSW_REG_HPKT_ACK_REQUIRED,
+};
+
+/* reg_hpkt_ack
+ * Require acknowledgements from the host for events.
+ * If set, then the device will wait for the event it sent to be acknowledged
+ * by the host. This option is only relevant for event trap IDs.
+ * Access: RW
+ *
+ * Note: Currently not supported by firmware.
+ */
+MLXSW_ITEM32(reg, hpkt, ack, 0x00, 24, 1);
+
+enum mlxsw_reg_hpkt_action {
+	MLXSW_REG_HPKT_ACTION_FORWARD,
+	MLXSW_REG_HPKT_ACTION_TRAP_TO_CPU,
+	MLXSW_REG_HPKT_ACTION_MIRROR_TO_CPU,
+	MLXSW_REG_HPKT_ACTION_DISCARD,
+	MLXSW_REG_HPKT_ACTION_SOFT_DISCARD,
+	MLXSW_REG_HPKT_ACTION_TRAP_AND_SOFT_DISCARD,
+};
+
+/* reg_hpkt_action
+ * Action to perform on packet when trapped.
+ * 0 - No action. Forward to CPU based on switching rules.
+ * 1 - Trap to CPU (CPU receives sole copy).
+ * 2 - Mirror to CPU (CPU receives a replica of the packet).
+ * 3 - Discard.
+ * 4 - Soft discard (allow other traps to act on the packet).
+ * 5 - Trap and soft discard (allow other traps to overwrite this trap).
+ * Access: RW
+ *
+ * Note: Must be set to 0 (forward) for event trap IDs, as they are already
+ * addressed to the CPU.
+ */
+MLXSW_ITEM32(reg, hpkt, action, 0x00, 20, 3);
+
+/* reg_hpkt_trap_group
+ * Trap group to associate the trap with.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, hpkt, trap_group, 0x00, 12, 6);
+
+/* reg_hpkt_trap_id
+ * Trap ID.
+ * Access: Index
+ *
+ * Note: A trap ID can only be associated with a single trap group. The device
+ * will associate the trap ID with the last trap group configured.
+ */
+MLXSW_ITEM32(reg, hpkt, trap_id, 0x00, 0, 9);
+
+enum {
+	MLXSW_REG_HPKT_CTRL_PACKET_DEFAULT,
+	MLXSW_REG_HPKT_CTRL_PACKET_NO_BUFFER,
+	MLXSW_REG_HPKT_CTRL_PACKET_USE_BUFFER,
+};
+
+/* reg_hpkt_ctrl
+ * Configure dedicated buffer resources for control packets.
+ * 0 - Keep factory defaults.
+ * 1 - Do not use control buffer for this trap ID.
+ * 2 - Use control buffer for this trap ID.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, hpkt, ctrl, 0x04, 16, 2);
+
+static inline void mlxsw_reg_hpkt_pack(char *payload, u8 action, u16 trap_id)
+{
+	enum mlxsw_reg_htgt_trap_group trap_group;
+
+	MLXSW_REG_ZERO(hpkt, payload);
+	mlxsw_reg_hpkt_ack_set(payload, MLXSW_REG_HPKT_ACK_NOT_REQUIRED);
+	mlxsw_reg_hpkt_action_set(payload, action);
+	switch (trap_id) {
+	case MLXSW_TRAP_ID_ETHEMAD:
+	case MLXSW_TRAP_ID_PUDE:
+		trap_group = MLXSW_REG_HTGT_TRAP_GROUP_EMAD;
+		break;
+	default:
+		trap_group = MLXSW_REG_HTGT_TRAP_GROUP_RX;
+		break;
+	}
+	mlxsw_reg_hpkt_trap_group_set(payload, trap_group);
+	mlxsw_reg_hpkt_trap_id_set(payload, trap_id);
+	mlxsw_reg_hpkt_ctrl_set(payload, MLXSW_REG_HPKT_CTRL_PACKET_DEFAULT);
+}
+
+/* SBPR - Shared Buffer Pools Register
+ * -----------------------------------
+ * The SBPR configures and retrieves the shared buffer pools and configuration.
+ */
+#define MLXSW_REG_SBPR_ID 0xB001
+#define MLXSW_REG_SBPR_LEN 0x14
+
+static const struct mlxsw_reg_info mlxsw_reg_sbpr = {
+	.id = MLXSW_REG_SBPR_ID,
+	.len = MLXSW_REG_SBPR_LEN,
+};
+
+enum mlxsw_reg_sbpr_dir {
+	MLXSW_REG_SBPR_DIR_INGRESS,
+	MLXSW_REG_SBPR_DIR_EGRESS,
+};
+
+/* reg_sbpr_dir
+ * Direction.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbpr, dir, 0x00, 24, 2);
+
+/* reg_sbpr_pool
+ * Pool index.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbpr, pool, 0x00, 0, 4);
+
+/* reg_sbpr_size
+ * Pool size in buffer cells.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbpr, size, 0x04, 0, 24);
+
+enum mlxsw_reg_sbpr_mode {
+	MLXSW_REG_SBPR_MODE_STATIC,
+	MLXSW_REG_SBPR_MODE_DYNAMIC,
+};
+
+/* reg_sbpr_mode
+ * Pool quota calculation mode.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbpr, mode, 0x08, 0, 4);
+
+static inline void mlxsw_reg_sbpr_pack(char *payload, u8 pool,
+				       enum mlxsw_reg_sbpr_dir dir,
+				       enum mlxsw_reg_sbpr_mode mode, u32 size)
+{
+	MLXSW_REG_ZERO(sbpr, payload);
+	mlxsw_reg_sbpr_pool_set(payload, pool);
+	mlxsw_reg_sbpr_dir_set(payload, dir);
+	mlxsw_reg_sbpr_mode_set(payload, mode);
+	mlxsw_reg_sbpr_size_set(payload, size);
+}
+
+/* SBCM - Shared Buffer Class Management Register
+ * ----------------------------------------------
+ * The SBCM register configures and retrieves the shared buffer allocation
+ * and configuration according to Port-PG, including the binding to pool
+ * and definition of the associated quota.
+ */
+#define MLXSW_REG_SBCM_ID 0xB002
+#define MLXSW_REG_SBCM_LEN 0x28
+
+static const struct mlxsw_reg_info mlxsw_reg_sbcm = {
+	.id = MLXSW_REG_SBCM_ID,
+	.len = MLXSW_REG_SBCM_LEN,
+};
+
+/* reg_sbcm_local_port
+ * Local port number.
+ * For Ingress: excludes CPU port and Router port
+ * For Egress: excludes IP Router
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbcm, local_port, 0x00, 16, 8);
+
+/* reg_sbcm_pg_buff
+ * PG buffer - Port PG (dir=ingress) / traffic class (dir=egress)
+ * For PG buffer: range is 0..cap_max_pg_buffers - 1
+ * For traffic class: range is 0..cap_max_tclass - 1
+ * Note that when traffic class is in MC aware mode then the traffic
+ * classes which are MC aware cannot be configured.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbcm, pg_buff, 0x00, 8, 6);
+
+enum mlxsw_reg_sbcm_dir {
+	MLXSW_REG_SBCM_DIR_INGRESS,
+	MLXSW_REG_SBCM_DIR_EGRESS,
+};
+
+/* reg_sbcm_dir
+ * Direction.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbcm, dir, 0x00, 0, 2);
+
+/* reg_sbcm_min_buff
+ * Minimum buffer size for the limiter, in cells.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbcm, min_buff, 0x18, 0, 24);
+
+/* reg_sbcm_max_buff
+ * When the pool associated to the port-pg/tclass is configured to
+ * static, Maximum buffer size for the limiter configured in cells.
+ * When the pool associated to the port-pg/tclass is configured to
+ * dynamic, the max_buff holds the "alpha" parameter, supporting
+ * the following values:
+ * 0: 0
+ * i: (1/128)*2^(i-1), for i=1..14
+ * 0xFF: Infinity
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbcm, max_buff, 0x1C, 0, 24);
+
+/* reg_sbcm_pool
+ * Association of the port-priority to a pool.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbcm, pool, 0x24, 0, 4);
+
+static inline void mlxsw_reg_sbcm_pack(char *payload, u8 local_port, u8 pg_buff,
+				       enum mlxsw_reg_sbcm_dir dir,
+				       u32 min_buff, u32 max_buff, u8 pool)
+{
+	MLXSW_REG_ZERO(sbcm, payload);
+	mlxsw_reg_sbcm_local_port_set(payload, local_port);
+	mlxsw_reg_sbcm_pg_buff_set(payload, pg_buff);
+	mlxsw_reg_sbcm_dir_set(payload, dir);
+	mlxsw_reg_sbcm_min_buff_set(payload, min_buff);
+	mlxsw_reg_sbcm_max_buff_set(payload, max_buff);
+	mlxsw_reg_sbcm_pool_set(payload, pool);
+}
+
+/* SBPM - Shared Buffer Class Management Register
+ * ----------------------------------------------
+ * The SBPM register configures and retrieves the shared buffer allocation
+ * and configuration according to Port-Pool, including the definition
+ * of the associated quota.
+ */
+#define MLXSW_REG_SBPM_ID 0xB003
+#define MLXSW_REG_SBPM_LEN 0x28
+
+static const struct mlxsw_reg_info mlxsw_reg_sbpm = {
+	.id = MLXSW_REG_SBPM_ID,
+	.len = MLXSW_REG_SBPM_LEN,
+};
+
+/* reg_sbpm_local_port
+ * Local port number.
+ * For Ingress: excludes CPU port and Router port
+ * For Egress: excludes IP Router
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbpm, local_port, 0x00, 16, 8);
+
+/* reg_sbpm_pool
+ * The pool associated to quota counting on the local_port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbpm, pool, 0x00, 8, 4);
+
+enum mlxsw_reg_sbpm_dir {
+	MLXSW_REG_SBPM_DIR_INGRESS,
+	MLXSW_REG_SBPM_DIR_EGRESS,
+};
+
+/* reg_sbpm_dir
+ * Direction.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbpm, dir, 0x00, 0, 2);
+
+/* reg_sbpm_min_buff
+ * Minimum buffer size for the limiter, in cells.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbpm, min_buff, 0x18, 0, 24);
+
+/* reg_sbpm_max_buff
+ * When the pool associated to the port-pg/tclass is configured to
+ * static, Maximum buffer size for the limiter configured in cells.
+ * When the pool associated to the port-pg/tclass is configured to
+ * dynamic, the max_buff holds the "alpha" parameter, supporting
+ * the following values:
+ * 0: 0
+ * i: (1/128)*2^(i-1), for i=1..14
+ * 0xFF: Infinity
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbpm, max_buff, 0x1C, 0, 24);
+
+static inline void mlxsw_reg_sbpm_pack(char *payload, u8 local_port, u8 pool,
+				       enum mlxsw_reg_sbpm_dir dir,
+				       u32 min_buff, u32 max_buff)
+{
+	MLXSW_REG_ZERO(sbpm, payload);
+	mlxsw_reg_sbpm_local_port_set(payload, local_port);
+	mlxsw_reg_sbpm_pool_set(payload, pool);
+	mlxsw_reg_sbpm_dir_set(payload, dir);
+	mlxsw_reg_sbpm_min_buff_set(payload, min_buff);
+	mlxsw_reg_sbpm_max_buff_set(payload, max_buff);
+}
+
+/* SBMM - Shared Buffer Multicast Management Register
+ * --------------------------------------------------
+ * The SBMM register configures and retrieves the shared buffer allocation
+ * and configuration for MC packets according to Switch-Priority, including
+ * the binding to pool and definition of the associated quota.
+ */
+#define MLXSW_REG_SBMM_ID 0xB004
+#define MLXSW_REG_SBMM_LEN 0x28
+
+static const struct mlxsw_reg_info mlxsw_reg_sbmm = {
+	.id = MLXSW_REG_SBMM_ID,
+	.len = MLXSW_REG_SBMM_LEN,
+};
+
+/* reg_sbmm_prio
+ * Switch Priority.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, sbmm, prio, 0x00, 8, 4);
+
+/* reg_sbmm_min_buff
+ * Minimum buffer size for the limiter, in cells.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbmm, min_buff, 0x18, 0, 24);
+
+/* reg_sbmm_max_buff
+ * When the pool associated to the port-pg/tclass is configured to
+ * static, Maximum buffer size for the limiter configured in cells.
+ * When the pool associated to the port-pg/tclass is configured to
+ * dynamic, the max_buff holds the "alpha" parameter, supporting
+ * the following values:
+ * 0: 0
+ * i: (1/128)*2^(i-1), for i=1..14
+ * 0xFF: Infinity
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbmm, max_buff, 0x1C, 0, 24);
+
+/* reg_sbmm_pool
+ * Association of the port-priority to a pool.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, sbmm, pool, 0x24, 0, 4);
+
+static inline void mlxsw_reg_sbmm_pack(char *payload, u8 prio, u32 min_buff,
+				       u32 max_buff, u8 pool)
+{
+	MLXSW_REG_ZERO(sbmm, payload);
+	mlxsw_reg_sbmm_prio_set(payload, prio);
+	mlxsw_reg_sbmm_min_buff_set(payload, min_buff);
+	mlxsw_reg_sbmm_max_buff_set(payload, max_buff);
+	mlxsw_reg_sbmm_pool_set(payload, pool);
+}
+
+static inline const char *mlxsw_reg_id_str(u16 reg_id)
+{
+	switch (reg_id) {
+	case MLXSW_REG_SGCR_ID:
+		return "SGCR";
+	case MLXSW_REG_SPAD_ID:
+		return "SPAD";
+	case MLXSW_REG_SSPR_ID:
+		return "SSPR";
+	case MLXSW_REG_SFDAT_ID:
+		return "SFDAT";
+	case MLXSW_REG_SFD_ID:
+		return "SFD";
+	case MLXSW_REG_SFN_ID:
+		return "SFN";
+	case MLXSW_REG_SPMS_ID:
+		return "SPMS";
+	case MLXSW_REG_SPVID_ID:
+		return "SPVID";
+	case MLXSW_REG_SPVM_ID:
+		return "SPVM";
+	case MLXSW_REG_SFGC_ID:
+		return "SFGC";
+	case MLXSW_REG_SFTR_ID:
+		return "SFTR";
+	case MLXSW_REG_SPMLR_ID:
+		return "SPMLR";
+	case MLXSW_REG_SVFA_ID:
+		return "SVFA";
+	case MLXSW_REG_SVPE_ID:
+		return "SVPE";
+	case MLXSW_REG_SFMR_ID:
+		return "SFMR";
+	case MLXSW_REG_SPVMLR_ID:
+		return "SPVMLR";
+	case MLXSW_REG_PMLP_ID:
+		return "PMLP";
+	case MLXSW_REG_PMTU_ID:
+		return "PMTU";
+	case MLXSW_REG_PTYS_ID:
+		return "PTYS";
+	case MLXSW_REG_PPAD_ID:
+		return "PPAD";
+	case MLXSW_REG_PAOS_ID:
+		return "PAOS";
+	case MLXSW_REG_PPCNT_ID:
+		return "PPCNT";
+	case MLXSW_REG_PBMC_ID:
+		return "PBMC";
+	case MLXSW_REG_PSPA_ID:
+		return "PSPA";
+	case MLXSW_REG_HTGT_ID:
+		return "HTGT";
+	case MLXSW_REG_HPKT_ID:
+		return "HPKT";
+	case MLXSW_REG_SBPR_ID:
+		return "SBPR";
+	case MLXSW_REG_SBCM_ID:
+		return "SBCM";
+	case MLXSW_REG_SBPM_ID:
+		return "SBPM";
+	case MLXSW_REG_SBMM_ID:
+		return "SBMM";
+	default:
+		return "*UNKNOWN*";
+	}
+}
+
+/* PUDE - Port Up / Down Event
+ * ---------------------------
+ * Reports the operational state change of a port.
+ */
+#define MLXSW_REG_PUDE_LEN 0x10
+
+/* reg_pude_swid
+ * Switch partition ID with which to associate the port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pude, swid, 0x00, 24, 8);
+
+/* reg_pude_local_port
+ * Local port number.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, pude, local_port, 0x00, 16, 8);
+
+/* reg_pude_admin_status
+ * Port administrative state (the desired state).
+ * 1 - Up.
+ * 2 - Down.
+ * 3 - Up once. This means that in case of link failure, the port won't go
+ *     into polling mode, but will wait to be re-enabled by software.
+ * 4 - Disabled by system. Can only be set by hardware.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, pude, admin_status, 0x00, 8, 4);
+
+/* reg_pude_oper_status
+ * Port operatioanl state.
+ * 1 - Up.
+ * 2 - Down.
+ * 3 - Down by port failure. This means that the device will not let the
+ *     port up again until explicitly specified by software.
+ * Access: RO
+ */
+MLXSW_ITEM32(reg, pude, oper_status, 0x00, 0, 4);
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
new file mode 100644
index 000000000..3be4a2355
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -0,0 +1,1949 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/ethtool.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/skbuff.h>
+#include <linux/if_vlan.h>
+#include <linux/if_bridge.h>
+#include <linux/workqueue.h>
+#include <linux/jiffies.h>
+#include <linux/bitops.h>
+#include <net/switchdev.h>
+#include <generated/utsrelease.h>
+
+#include "spectrum.h"
+#include "core.h"
+#include "reg.h"
+#include "port.h"
+#include "trap.h"
+#include "txheader.h"
+
+static const char mlxsw_sp_driver_name[] = "mlxsw_spectrum";
+static const char mlxsw_sp_driver_version[] = "1.0";
+
+/* tx_hdr_version
+ * Tx header version.
+ * Must be set to 1.
+ */
+MLXSW_ITEM32(tx, hdr, version, 0x00, 28, 4);
+
+/* tx_hdr_ctl
+ * Packet control type.
+ * 0 - Ethernet control (e.g. EMADs, LACP)
+ * 1 - Ethernet data
+ */
+MLXSW_ITEM32(tx, hdr, ctl, 0x00, 26, 2);
+
+/* tx_hdr_proto
+ * Packet protocol type. Must be set to 1 (Ethernet).
+ */
+MLXSW_ITEM32(tx, hdr, proto, 0x00, 21, 3);
+
+/* tx_hdr_rx_is_router
+ * Packet is sent from the router. Valid for data packets only.
+ */
+MLXSW_ITEM32(tx, hdr, rx_is_router, 0x00, 19, 1);
+
+/* tx_hdr_fid_valid
+ * Indicates if the 'fid' field is valid and should be used for
+ * forwarding lookup. Valid for data packets only.
+ */
+MLXSW_ITEM32(tx, hdr, fid_valid, 0x00, 16, 1);
+
+/* tx_hdr_swid
+ * Switch partition ID. Must be set to 0.
+ */
+MLXSW_ITEM32(tx, hdr, swid, 0x00, 12, 3);
+
+/* tx_hdr_control_tclass
+ * Indicates if the packet should use the control TClass and not one
+ * of the data TClasses.
+ */
+MLXSW_ITEM32(tx, hdr, control_tclass, 0x00, 6, 1);
+
+/* tx_hdr_etclass
+ * Egress TClass to be used on the egress device on the egress port.
+ */
+MLXSW_ITEM32(tx, hdr, etclass, 0x00, 0, 4);
+
+/* tx_hdr_port_mid
+ * Destination local port for unicast packets.
+ * Destination multicast ID for multicast packets.
+ *
+ * Control packets are directed to a specific egress port, while data
+ * packets are transmitted through the CPU port (0) into the switch partition,
+ * where forwarding rules are applied.
+ */
+MLXSW_ITEM32(tx, hdr, port_mid, 0x04, 16, 16);
+
+/* tx_hdr_fid
+ * Forwarding ID used for L2 forwarding lookup. Valid only if 'fid_valid' is
+ * set, otherwise calculated based on the packet's VID using VID to FID mapping.
+ * Valid for data packets only.
+ */
+MLXSW_ITEM32(tx, hdr, fid, 0x08, 0, 16);
+
+/* tx_hdr_type
+ * 0 - Data packets
+ * 6 - Control packets
+ */
+MLXSW_ITEM32(tx, hdr, type, 0x0C, 0, 4);
+
+static void mlxsw_sp_txhdr_construct(struct sk_buff *skb,
+				     const struct mlxsw_tx_info *tx_info)
+{
+	char *txhdr = skb_push(skb, MLXSW_TXHDR_LEN);
+
+	memset(txhdr, 0, MLXSW_TXHDR_LEN);
+
+	mlxsw_tx_hdr_version_set(txhdr, MLXSW_TXHDR_VERSION_1);
+	mlxsw_tx_hdr_ctl_set(txhdr, MLXSW_TXHDR_ETH_CTL);
+	mlxsw_tx_hdr_proto_set(txhdr, MLXSW_TXHDR_PROTO_ETH);
+	mlxsw_tx_hdr_swid_set(txhdr, 0);
+	mlxsw_tx_hdr_control_tclass_set(txhdr, 1);
+	mlxsw_tx_hdr_port_mid_set(txhdr, tx_info->local_port);
+	mlxsw_tx_hdr_type_set(txhdr, MLXSW_TXHDR_TYPE_CONTROL);
+}
+
+static int mlxsw_sp_base_mac_get(struct mlxsw_sp *mlxsw_sp)
+{
+	char spad_pl[MLXSW_REG_SPAD_LEN];
+	int err;
+
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(spad), spad_pl);
+	if (err)
+		return err;
+	mlxsw_reg_spad_base_mac_memcpy_from(spad_pl, mlxsw_sp->base_mac);
+	return 0;
+}
+
+static int mlxsw_sp_port_admin_status_set(struct mlxsw_sp_port *mlxsw_sp_port,
+					  bool is_up)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char paos_pl[MLXSW_REG_PAOS_LEN];
+
+	mlxsw_reg_paos_pack(paos_pl, mlxsw_sp_port->local_port,
+			    is_up ? MLXSW_PORT_ADMIN_STATUS_UP :
+			    MLXSW_PORT_ADMIN_STATUS_DOWN);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(paos), paos_pl);
+}
+
+static int mlxsw_sp_port_oper_status_get(struct mlxsw_sp_port *mlxsw_sp_port,
+					 bool *p_is_up)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char paos_pl[MLXSW_REG_PAOS_LEN];
+	u8 oper_status;
+	int err;
+
+	mlxsw_reg_paos_pack(paos_pl, mlxsw_sp_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(paos), paos_pl);
+	if (err)
+		return err;
+	oper_status = mlxsw_reg_paos_oper_status_get(paos_pl);
+	*p_is_up = oper_status == MLXSW_PORT_ADMIN_STATUS_UP ? true : false;
+	return 0;
+}
+
+static int mlxsw_sp_vfid_create(struct mlxsw_sp *mlxsw_sp, u16 vfid)
+{
+	char sfmr_pl[MLXSW_REG_SFMR_LEN];
+	int err;
+
+	mlxsw_reg_sfmr_pack(sfmr_pl, MLXSW_REG_SFMR_OP_CREATE_FID,
+			    MLXSW_SP_VFID_BASE + vfid, 0);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sfmr), sfmr_pl);
+
+	if (err)
+		return err;
+
+	set_bit(vfid, mlxsw_sp->active_vfids);
+	return 0;
+}
+
+static void mlxsw_sp_vfid_destroy(struct mlxsw_sp *mlxsw_sp, u16 vfid)
+{
+	char sfmr_pl[MLXSW_REG_SFMR_LEN];
+
+	clear_bit(vfid, mlxsw_sp->active_vfids);
+
+	mlxsw_reg_sfmr_pack(sfmr_pl, MLXSW_REG_SFMR_OP_DESTROY_FID,
+			    MLXSW_SP_VFID_BASE + vfid, 0);
+	mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sfmr), sfmr_pl);
+}
+
+static int mlxsw_sp_port_dev_addr_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				      unsigned char *addr)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char ppad_pl[MLXSW_REG_PPAD_LEN];
+
+	mlxsw_reg_ppad_pack(ppad_pl, true, mlxsw_sp_port->local_port);
+	mlxsw_reg_ppad_mac_memcpy_to(ppad_pl, addr);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ppad), ppad_pl);
+}
+
+static int mlxsw_sp_port_dev_addr_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	unsigned char *addr = mlxsw_sp_port->dev->dev_addr;
+
+	ether_addr_copy(addr, mlxsw_sp->base_mac);
+	addr[ETH_ALEN - 1] += mlxsw_sp_port->local_port;
+	return mlxsw_sp_port_dev_addr_set(mlxsw_sp_port, addr);
+}
+
+static int mlxsw_sp_port_stp_state_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				       u16 vid, enum mlxsw_reg_spms_state state)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char *spms_pl;
+	int err;
+
+	spms_pl = kmalloc(MLXSW_REG_SPMS_LEN, GFP_KERNEL);
+	if (!spms_pl)
+		return -ENOMEM;
+	mlxsw_reg_spms_pack(spms_pl, mlxsw_sp_port->local_port);
+	mlxsw_reg_spms_vid_pack(spms_pl, vid, state);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spms), spms_pl);
+	kfree(spms_pl);
+	return err;
+}
+
+static int mlxsw_sp_port_mtu_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 mtu)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char pmtu_pl[MLXSW_REG_PMTU_LEN];
+	int max_mtu;
+	int err;
+
+	mtu += MLXSW_TXHDR_LEN + ETH_HLEN;
+	mlxsw_reg_pmtu_pack(pmtu_pl, mlxsw_sp_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(pmtu), pmtu_pl);
+	if (err)
+		return err;
+	max_mtu = mlxsw_reg_pmtu_max_mtu_get(pmtu_pl);
+
+	if (mtu > max_mtu)
+		return -EINVAL;
+
+	mlxsw_reg_pmtu_pack(pmtu_pl, mlxsw_sp_port->local_port, mtu);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pmtu), pmtu_pl);
+}
+
+static int mlxsw_sp_port_swid_set(struct mlxsw_sp_port *mlxsw_sp_port, u8 swid)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char pspa_pl[MLXSW_REG_PSPA_LEN];
+
+	mlxsw_reg_pspa_pack(pspa_pl, swid, mlxsw_sp_port->local_port);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(pspa), pspa_pl);
+}
+
+static int mlxsw_sp_port_vp_mode_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				     bool enable)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char svpe_pl[MLXSW_REG_SVPE_LEN];
+
+	mlxsw_reg_svpe_pack(svpe_pl, mlxsw_sp_port->local_port, enable);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(svpe), svpe_pl);
+}
+
+int mlxsw_sp_port_vid_to_fid_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				 enum mlxsw_reg_svfa_mt mt, bool valid, u16 fid,
+				 u16 vid)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char svfa_pl[MLXSW_REG_SVFA_LEN];
+
+	mlxsw_reg_svfa_pack(svfa_pl, mlxsw_sp_port->local_port, mt, valid,
+			    fid, vid);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(svfa), svfa_pl);
+}
+
+static int mlxsw_sp_port_vid_learning_set(struct mlxsw_sp_port *mlxsw_sp_port,
+					  u16 vid, bool learn_enable)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char *spvmlr_pl;
+	int err;
+
+	spvmlr_pl = kmalloc(MLXSW_REG_SPVMLR_LEN, GFP_KERNEL);
+	if (!spvmlr_pl)
+		return -ENOMEM;
+	mlxsw_reg_spvmlr_pack(spvmlr_pl, mlxsw_sp_port->local_port, vid, vid,
+			      learn_enable);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spvmlr), spvmlr_pl);
+	kfree(spvmlr_pl);
+	return err;
+}
+
+static int
+mlxsw_sp_port_system_port_mapping_set(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char sspr_pl[MLXSW_REG_SSPR_LEN];
+
+	mlxsw_reg_sspr_pack(sspr_pl, mlxsw_sp_port->local_port);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sspr), sspr_pl);
+}
+
+static int mlxsw_sp_port_module_check(struct mlxsw_sp_port *mlxsw_sp_port,
+				      bool *p_usable)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char pmlp_pl[MLXSW_REG_PMLP_LEN];
+	int err;
+
+	mlxsw_reg_pmlp_pack(pmlp_pl, mlxsw_sp_port->local_port);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(pmlp), pmlp_pl);
+	if (err)
+		return err;
+	*p_usable = mlxsw_reg_pmlp_width_get(pmlp_pl) ? true : false;
+	return 0;
+}
+
+static int mlxsw_sp_port_open(struct net_device *dev)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err;
+
+	err = mlxsw_sp_port_admin_status_set(mlxsw_sp_port, true);
+	if (err)
+		return err;
+	netif_start_queue(dev);
+	return 0;
+}
+
+static int mlxsw_sp_port_stop(struct net_device *dev)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+
+	netif_stop_queue(dev);
+	return mlxsw_sp_port_admin_status_set(mlxsw_sp_port, false);
+}
+
+static netdev_tx_t mlxsw_sp_port_xmit(struct sk_buff *skb,
+				      struct net_device *dev)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	struct mlxsw_sp_port_pcpu_stats *pcpu_stats;
+	const struct mlxsw_tx_info tx_info = {
+		.local_port = mlxsw_sp_port->local_port,
+		.is_emad = false,
+	};
+	u64 len;
+	int err;
+
+	if (mlxsw_core_skb_transmit_busy(mlxsw_sp, &tx_info))
+		return NETDEV_TX_BUSY;
+
+	if (unlikely(skb_headroom(skb) < MLXSW_TXHDR_LEN)) {
+		struct sk_buff *skb_orig = skb;
+
+		skb = skb_realloc_headroom(skb, MLXSW_TXHDR_LEN);
+		if (!skb) {
+			this_cpu_inc(mlxsw_sp_port->pcpu_stats->tx_dropped);
+			dev_kfree_skb_any(skb_orig);
+			return NETDEV_TX_OK;
+		}
+	}
+
+	if (eth_skb_pad(skb)) {
+		this_cpu_inc(mlxsw_sp_port->pcpu_stats->tx_dropped);
+		return NETDEV_TX_OK;
+	}
+
+	mlxsw_sp_txhdr_construct(skb, &tx_info);
+	len = skb->len;
+	/* Due to a race we might fail here because of a full queue. In that
+	 * unlikely case we simply drop the packet.
+	 */
+	err = mlxsw_core_skb_transmit(mlxsw_sp, skb, &tx_info);
+
+	if (!err) {
+		pcpu_stats = this_cpu_ptr(mlxsw_sp_port->pcpu_stats);
+		u64_stats_update_begin(&pcpu_stats->syncp);
+		pcpu_stats->tx_packets++;
+		pcpu_stats->tx_bytes += len;
+		u64_stats_update_end(&pcpu_stats->syncp);
+	} else {
+		this_cpu_inc(mlxsw_sp_port->pcpu_stats->tx_dropped);
+		dev_kfree_skb_any(skb);
+	}
+	return NETDEV_TX_OK;
+}
+
+static int mlxsw_sp_port_set_mac_address(struct net_device *dev, void *p)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct sockaddr *addr = p;
+	int err;
+
+	if (!is_valid_ether_addr(addr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	err = mlxsw_sp_port_dev_addr_set(mlxsw_sp_port, addr->sa_data);
+	if (err)
+		return err;
+	memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
+	return 0;
+}
+
+static int mlxsw_sp_port_change_mtu(struct net_device *dev, int mtu)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err;
+
+	err = mlxsw_sp_port_mtu_set(mlxsw_sp_port, mtu);
+	if (err)
+		return err;
+	dev->mtu = mtu;
+	return 0;
+}
+
+static struct rtnl_link_stats64 *
+mlxsw_sp_port_get_stats64(struct net_device *dev,
+			  struct rtnl_link_stats64 *stats)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp_port_pcpu_stats *p;
+	u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
+	u32 tx_dropped = 0;
+	unsigned int start;
+	int i;
+
+	for_each_possible_cpu(i) {
+		p = per_cpu_ptr(mlxsw_sp_port->pcpu_stats, i);
+		do {
+			start = u64_stats_fetch_begin_irq(&p->syncp);
+			rx_packets	= p->rx_packets;
+			rx_bytes	= p->rx_bytes;
+			tx_packets	= p->tx_packets;
+			tx_bytes	= p->tx_bytes;
+		} while (u64_stats_fetch_retry_irq(&p->syncp, start));
+
+		stats->rx_packets	+= rx_packets;
+		stats->rx_bytes		+= rx_bytes;
+		stats->tx_packets	+= tx_packets;
+		stats->tx_bytes		+= tx_bytes;
+		/* tx_dropped is u32, updated without syncp protection. */
+		tx_dropped	+= p->tx_dropped;
+	}
+	stats->tx_dropped	= tx_dropped;
+	return stats;
+}
+
+int mlxsw_sp_port_vlan_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid_begin,
+			   u16 vid_end, bool is_member, bool untagged)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char *spvm_pl;
+	int err;
+
+	spvm_pl = kmalloc(MLXSW_REG_SPVM_LEN, GFP_KERNEL);
+	if (!spvm_pl)
+		return -ENOMEM;
+
+	mlxsw_reg_spvm_pack(spvm_pl, mlxsw_sp_port->local_port,	vid_begin,
+			    vid_end, is_member, untagged);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spvm), spvm_pl);
+	kfree(spvm_pl);
+	return err;
+}
+
+static int mlxsw_sp_port_vp_mode_trans(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	enum mlxsw_reg_svfa_mt mt = MLXSW_REG_SVFA_MT_PORT_VID_TO_FID;
+	u16 vid, last_visited_vid;
+	int err;
+
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, VLAN_N_VID) {
+		err = mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port, mt, true, vid,
+						   vid);
+		if (err) {
+			last_visited_vid = vid;
+			goto err_port_vid_to_fid_set;
+		}
+	}
+
+	err = mlxsw_sp_port_vp_mode_set(mlxsw_sp_port, true);
+	if (err) {
+		last_visited_vid = VLAN_N_VID;
+		goto err_port_vid_to_fid_set;
+	}
+
+	return 0;
+
+err_port_vid_to_fid_set:
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, last_visited_vid)
+		mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port, mt, false, vid,
+					     vid);
+	return err;
+}
+
+static int mlxsw_sp_port_vlan_mode_trans(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	enum mlxsw_reg_svfa_mt mt = MLXSW_REG_SVFA_MT_PORT_VID_TO_FID;
+	u16 vid;
+	int err;
+
+	err = mlxsw_sp_port_vp_mode_set(mlxsw_sp_port, false);
+	if (err)
+		return err;
+
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, VLAN_N_VID) {
+		err = mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port, mt, false,
+						   vid, vid);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int mlxsw_sp_port_add_vid(struct net_device *dev, __be16 __always_unused proto,
+			  u16 vid)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char *sftr_pl;
+	int err;
+
+	/* VLAN 0 is added to HW filter when device goes up, but it is
+	 * reserved in our case, so simply return.
+	 */
+	if (!vid)
+		return 0;
+
+	if (test_bit(vid, mlxsw_sp_port->active_vfids)) {
+		netdev_warn(dev, "VID=%d already configured\n", vid);
+		return 0;
+	}
+
+	if (!test_bit(vid, mlxsw_sp->active_vfids)) {
+		err = mlxsw_sp_vfid_create(mlxsw_sp, vid);
+		if (err) {
+			netdev_err(dev, "Failed to create vFID=%d\n",
+				   MLXSW_SP_VFID_BASE + vid);
+			return err;
+		}
+
+		sftr_pl = kmalloc(MLXSW_REG_SFTR_LEN, GFP_KERNEL);
+		if (!sftr_pl) {
+			err = -ENOMEM;
+			goto err_flood_table_alloc;
+		}
+		mlxsw_reg_sftr_pack(sftr_pl, 0, vid,
+				    MLXSW_REG_SFGC_TABLE_TYPE_FID, 0,
+				    MLXSW_PORT_CPU_PORT, true);
+		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sftr), sftr_pl);
+		kfree(sftr_pl);
+		if (err) {
+			netdev_err(dev, "Failed to configure flood table\n");
+			goto err_flood_table_config;
+		}
+	}
+
+	/* In case we fail in the following steps, we intentionally do not
+	 * destroy the associated vFID.
+	 */
+
+	/* When adding the first VLAN interface on a bridged port we need to
+	 * transition all the active 802.1Q bridge VLANs to use explicit
+	 * {Port, VID} to FID mappings and set the port's mode to Virtual mode.
+	 */
+	if (!mlxsw_sp_port->nr_vfids) {
+		err = mlxsw_sp_port_vp_mode_trans(mlxsw_sp_port);
+		if (err) {
+			netdev_err(dev, "Failed to set to Virtual mode\n");
+			return err;
+		}
+	}
+
+	err = mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port,
+					   MLXSW_REG_SVFA_MT_PORT_VID_TO_FID,
+					   true, MLXSW_SP_VFID_BASE + vid, vid);
+	if (err) {
+		netdev_err(dev, "Failed to map {Port, VID=%d} to vFID=%d\n",
+			   vid, MLXSW_SP_VFID_BASE + vid);
+		goto err_port_vid_to_fid_set;
+	}
+
+	err = mlxsw_sp_port_vid_learning_set(mlxsw_sp_port, vid, false);
+	if (err) {
+		netdev_err(dev, "Failed to disable learning for VID=%d\n", vid);
+		goto err_port_vid_learning_set;
+	}
+
+	err = mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid, true, false);
+	if (err) {
+		netdev_err(dev, "Failed to set VLAN membership for VID=%d\n",
+			   vid);
+		goto err_port_add_vid;
+	}
+
+	err = mlxsw_sp_port_stp_state_set(mlxsw_sp_port, vid,
+					  MLXSW_REG_SPMS_STATE_FORWARDING);
+	if (err) {
+		netdev_err(dev, "Failed to set STP state for VID=%d\n", vid);
+		goto err_port_stp_state_set;
+	}
+
+	mlxsw_sp_port->nr_vfids++;
+	set_bit(vid, mlxsw_sp_port->active_vfids);
+
+	return 0;
+
+err_flood_table_config:
+err_flood_table_alloc:
+	mlxsw_sp_vfid_destroy(mlxsw_sp, vid);
+	return err;
+
+err_port_stp_state_set:
+	mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid, false, false);
+err_port_add_vid:
+	mlxsw_sp_port_vid_learning_set(mlxsw_sp_port, vid, true);
+err_port_vid_learning_set:
+	mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port,
+				     MLXSW_REG_SVFA_MT_PORT_VID_TO_FID, false,
+				     MLXSW_SP_VFID_BASE + vid, vid);
+err_port_vid_to_fid_set:
+	mlxsw_sp_port_vlan_mode_trans(mlxsw_sp_port);
+	return err;
+}
+
+int mlxsw_sp_port_kill_vid(struct net_device *dev,
+			   __be16 __always_unused proto, u16 vid)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err;
+
+	/* VLAN 0 is removed from HW filter when device goes down, but
+	 * it is reserved in our case, so simply return.
+	 */
+	if (!vid)
+		return 0;
+
+	if (!test_bit(vid, mlxsw_sp_port->active_vfids)) {
+		netdev_warn(dev, "VID=%d does not exist\n", vid);
+		return 0;
+	}
+
+	err = mlxsw_sp_port_stp_state_set(mlxsw_sp_port, vid,
+					  MLXSW_REG_SPMS_STATE_DISCARDING);
+	if (err) {
+		netdev_err(dev, "Failed to set STP state for VID=%d\n", vid);
+		return err;
+	}
+
+	err = mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid, false, false);
+	if (err) {
+		netdev_err(dev, "Failed to set VLAN membership for VID=%d\n",
+			   vid);
+		return err;
+	}
+
+	err = mlxsw_sp_port_vid_learning_set(mlxsw_sp_port, vid, true);
+	if (err) {
+		netdev_err(dev, "Failed to enable learning for VID=%d\n", vid);
+		return err;
+	}
+
+	err = mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port,
+					   MLXSW_REG_SVFA_MT_PORT_VID_TO_FID,
+					   false, MLXSW_SP_VFID_BASE + vid,
+					   vid);
+	if (err) {
+		netdev_err(dev, "Failed to invalidate {Port, VID=%d} to vFID=%d mapping\n",
+			   vid, MLXSW_SP_VFID_BASE + vid);
+		return err;
+	}
+
+	/* When removing the last VLAN interface on a bridged port we need to
+	 * transition all active 802.1Q bridge VLANs to use VID to FID
+	 * mappings and set port's mode to VLAN mode.
+	 */
+	if (mlxsw_sp_port->nr_vfids == 1) {
+		err = mlxsw_sp_port_vlan_mode_trans(mlxsw_sp_port);
+		if (err) {
+			netdev_err(dev, "Failed to set to VLAN mode\n");
+			return err;
+		}
+	}
+
+	mlxsw_sp_port->nr_vfids--;
+	clear_bit(vid, mlxsw_sp_port->active_vfids);
+
+	return 0;
+}
+
+static const struct net_device_ops mlxsw_sp_port_netdev_ops = {
+	.ndo_open		= mlxsw_sp_port_open,
+	.ndo_stop		= mlxsw_sp_port_stop,
+	.ndo_start_xmit		= mlxsw_sp_port_xmit,
+	.ndo_set_mac_address	= mlxsw_sp_port_set_mac_address,
+	.ndo_change_mtu		= mlxsw_sp_port_change_mtu,
+	.ndo_get_stats64	= mlxsw_sp_port_get_stats64,
+	.ndo_vlan_rx_add_vid	= mlxsw_sp_port_add_vid,
+	.ndo_vlan_rx_kill_vid	= mlxsw_sp_port_kill_vid,
+	.ndo_fdb_add		= switchdev_port_fdb_add,
+	.ndo_fdb_del		= switchdev_port_fdb_del,
+	.ndo_fdb_dump		= switchdev_port_fdb_dump,
+	.ndo_bridge_setlink	= switchdev_port_bridge_setlink,
+	.ndo_bridge_getlink	= switchdev_port_bridge_getlink,
+	.ndo_bridge_dellink	= switchdev_port_bridge_dellink,
+};
+
+static void mlxsw_sp_port_get_drvinfo(struct net_device *dev,
+				      struct ethtool_drvinfo *drvinfo)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+
+	strlcpy(drvinfo->driver, mlxsw_sp_driver_name, sizeof(drvinfo->driver));
+	strlcpy(drvinfo->version, mlxsw_sp_driver_version,
+		sizeof(drvinfo->version));
+	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
+		 "%d.%d.%d",
+		 mlxsw_sp->bus_info->fw_rev.major,
+		 mlxsw_sp->bus_info->fw_rev.minor,
+		 mlxsw_sp->bus_info->fw_rev.subminor);
+	strlcpy(drvinfo->bus_info, mlxsw_sp->bus_info->device_name,
+		sizeof(drvinfo->bus_info));
+}
+
+struct mlxsw_sp_port_hw_stats {
+	char str[ETH_GSTRING_LEN];
+	u64 (*getter)(char *payload);
+};
+
+static const struct mlxsw_sp_port_hw_stats mlxsw_sp_port_hw_stats[] = {
+	{
+		.str = "a_frames_transmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_frames_transmitted_ok_get,
+	},
+	{
+		.str = "a_frames_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_frames_received_ok_get,
+	},
+	{
+		.str = "a_frame_check_sequence_errors",
+		.getter = mlxsw_reg_ppcnt_a_frame_check_sequence_errors_get,
+	},
+	{
+		.str = "a_alignment_errors",
+		.getter = mlxsw_reg_ppcnt_a_alignment_errors_get,
+	},
+	{
+		.str = "a_octets_transmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_octets_transmitted_ok_get,
+	},
+	{
+		.str = "a_octets_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_octets_received_ok_get,
+	},
+	{
+		.str = "a_multicast_frames_xmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_multicast_frames_xmitted_ok_get,
+	},
+	{
+		.str = "a_broadcast_frames_xmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_broadcast_frames_xmitted_ok_get,
+	},
+	{
+		.str = "a_multicast_frames_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_multicast_frames_received_ok_get,
+	},
+	{
+		.str = "a_broadcast_frames_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_broadcast_frames_received_ok_get,
+	},
+	{
+		.str = "a_in_range_length_errors",
+		.getter = mlxsw_reg_ppcnt_a_in_range_length_errors_get,
+	},
+	{
+		.str = "a_out_of_range_length_field",
+		.getter = mlxsw_reg_ppcnt_a_out_of_range_length_field_get,
+	},
+	{
+		.str = "a_frame_too_long_errors",
+		.getter = mlxsw_reg_ppcnt_a_frame_too_long_errors_get,
+	},
+	{
+		.str = "a_symbol_error_during_carrier",
+		.getter = mlxsw_reg_ppcnt_a_symbol_error_during_carrier_get,
+	},
+	{
+		.str = "a_mac_control_frames_transmitted",
+		.getter = mlxsw_reg_ppcnt_a_mac_control_frames_transmitted_get,
+	},
+	{
+		.str = "a_mac_control_frames_received",
+		.getter = mlxsw_reg_ppcnt_a_mac_control_frames_received_get,
+	},
+	{
+		.str = "a_unsupported_opcodes_received",
+		.getter = mlxsw_reg_ppcnt_a_unsupported_opcodes_received_get,
+	},
+	{
+		.str = "a_pause_mac_ctrl_frames_received",
+		.getter = mlxsw_reg_ppcnt_a_pause_mac_ctrl_frames_received_get,
+	},
+	{
+		.str = "a_pause_mac_ctrl_frames_xmitted",
+		.getter = mlxsw_reg_ppcnt_a_pause_mac_ctrl_frames_transmitted_get,
+	},
+};
+
+#define MLXSW_SP_PORT_HW_STATS_LEN ARRAY_SIZE(mlxsw_sp_port_hw_stats)
+
+static void mlxsw_sp_port_get_strings(struct net_device *dev,
+				      u32 stringset, u8 *data)
+{
+	u8 *p = data;
+	int i;
+
+	switch (stringset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < MLXSW_SP_PORT_HW_STATS_LEN; i++) {
+			memcpy(p, mlxsw_sp_port_hw_stats[i].str,
+			       ETH_GSTRING_LEN);
+			p += ETH_GSTRING_LEN;
+		}
+		break;
+	}
+}
+
+static void mlxsw_sp_port_get_stats(struct net_device *dev,
+				    struct ethtool_stats *stats, u64 *data)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char ppcnt_pl[MLXSW_REG_PPCNT_LEN];
+	int i;
+	int err;
+
+	mlxsw_reg_ppcnt_pack(ppcnt_pl, mlxsw_sp_port->local_port);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ppcnt), ppcnt_pl);
+	for (i = 0; i < MLXSW_SP_PORT_HW_STATS_LEN; i++)
+		data[i] = !err ? mlxsw_sp_port_hw_stats[i].getter(ppcnt_pl) : 0;
+}
+
+static int mlxsw_sp_port_get_sset_count(struct net_device *dev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return MLXSW_SP_PORT_HW_STATS_LEN;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+struct mlxsw_sp_port_link_mode {
+	u32 mask;
+	u32 supported;
+	u32 advertised;
+	u32 speed;
+};
+
+static const struct mlxsw_sp_port_link_mode mlxsw_sp_port_link_mode[] = {
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_100BASE_T,
+		.supported	= SUPPORTED_100baseT_Full,
+		.advertised	= ADVERTISED_100baseT_Full,
+		.speed		= 100,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_100BASE_TX,
+		.speed		= 100,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_SGMII |
+				  MLXSW_REG_PTYS_ETH_SPEED_1000BASE_KX,
+		.supported	= SUPPORTED_1000baseKX_Full,
+		.advertised	= ADVERTISED_1000baseKX_Full,
+		.speed		= 1000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_10GBASE_T,
+		.supported	= SUPPORTED_10000baseT_Full,
+		.advertised	= ADVERTISED_10000baseT_Full,
+		.speed		= 10000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CX4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4,
+		.supported	= SUPPORTED_10000baseKX4_Full,
+		.advertised	= ADVERTISED_10000baseKX4_Full,
+		.speed		= 10000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_ER_LR,
+		.supported	= SUPPORTED_10000baseKR_Full,
+		.advertised	= ADVERTISED_10000baseKR_Full,
+		.speed		= 10000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_20GBASE_KR2,
+		.supported	= SUPPORTED_20000baseKR2_Full,
+		.advertised	= ADVERTISED_20000baseKR2_Full,
+		.speed		= 20000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4,
+		.supported	= SUPPORTED_40000baseCR4_Full,
+		.advertised	= ADVERTISED_40000baseCR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4,
+		.supported	= SUPPORTED_40000baseKR4_Full,
+		.advertised	= ADVERTISED_40000baseKR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4,
+		.supported	= SUPPORTED_40000baseSR4_Full,
+		.advertised	= ADVERTISED_40000baseSR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_LR4_ER4,
+		.supported	= SUPPORTED_40000baseLR4_Full,
+		.advertised	= ADVERTISED_40000baseLR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_25GBASE_CR |
+				  MLXSW_REG_PTYS_ETH_SPEED_25GBASE_KR |
+				  MLXSW_REG_PTYS_ETH_SPEED_25GBASE_SR,
+		.speed		= 25000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_50GBASE_KR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_50GBASE_CR2 |
+				  MLXSW_REG_PTYS_ETH_SPEED_50GBASE_KR2,
+		.speed		= 50000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_56GBASE_R4,
+		.supported	= SUPPORTED_56000baseKR4_Full,
+		.advertised	= ADVERTISED_56000baseKR4_Full,
+		.speed		= 56000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_100GBASE_CR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_100GBASE_LR4_ER4,
+		.speed		= 100000,
+	},
+};
+
+#define MLXSW_SP_PORT_LINK_MODE_LEN ARRAY_SIZE(mlxsw_sp_port_link_mode)
+
+static u32 mlxsw_sp_from_ptys_supported_port(u32 ptys_eth_proto)
+{
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR |
+			      MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_SGMII))
+		return SUPPORTED_FIBRE;
+
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR |
+			      MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_1000BASE_KX))
+		return SUPPORTED_Backplane;
+	return 0;
+}
+
+static u32 mlxsw_sp_from_ptys_supported_link(u32 ptys_eth_proto)
+{
+	u32 modes = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SP_PORT_LINK_MODE_LEN; i++) {
+		if (ptys_eth_proto & mlxsw_sp_port_link_mode[i].mask)
+			modes |= mlxsw_sp_port_link_mode[i].supported;
+	}
+	return modes;
+}
+
+static u32 mlxsw_sp_from_ptys_advert_link(u32 ptys_eth_proto)
+{
+	u32 modes = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SP_PORT_LINK_MODE_LEN; i++) {
+		if (ptys_eth_proto & mlxsw_sp_port_link_mode[i].mask)
+			modes |= mlxsw_sp_port_link_mode[i].advertised;
+	}
+	return modes;
+}
+
+static void mlxsw_sp_from_ptys_speed_duplex(bool carrier_ok, u32 ptys_eth_proto,
+					    struct ethtool_cmd *cmd)
+{
+	u32 speed = SPEED_UNKNOWN;
+	u8 duplex = DUPLEX_UNKNOWN;
+	int i;
+
+	if (!carrier_ok)
+		goto out;
+
+	for (i = 0; i < MLXSW_SP_PORT_LINK_MODE_LEN; i++) {
+		if (ptys_eth_proto & mlxsw_sp_port_link_mode[i].mask) {
+			speed = mlxsw_sp_port_link_mode[i].speed;
+			duplex = DUPLEX_FULL;
+			break;
+		}
+	}
+out:
+	ethtool_cmd_speed_set(cmd, speed);
+	cmd->duplex = duplex;
+}
+
+static u8 mlxsw_sp_port_connector_port(u32 ptys_eth_proto)
+{
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_SGMII))
+		return PORT_FIBRE;
+
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_CR4))
+		return PORT_DA;
+
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR |
+			      MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4))
+		return PORT_NONE;
+
+	return PORT_OTHER;
+}
+
+static int mlxsw_sp_port_get_settings(struct net_device *dev,
+				      struct ethtool_cmd *cmd)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char ptys_pl[MLXSW_REG_PTYS_LEN];
+	u32 eth_proto_cap;
+	u32 eth_proto_admin;
+	u32 eth_proto_oper;
+	int err;
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sp_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ptys), ptys_pl);
+	if (err) {
+		netdev_err(dev, "Failed to get proto");
+		return err;
+	}
+	mlxsw_reg_ptys_unpack(ptys_pl, &eth_proto_cap,
+			      &eth_proto_admin, &eth_proto_oper);
+
+	cmd->supported = mlxsw_sp_from_ptys_supported_port(eth_proto_cap) |
+			 mlxsw_sp_from_ptys_supported_link(eth_proto_cap) |
+			 SUPPORTED_Pause | SUPPORTED_Asym_Pause;
+	cmd->advertising = mlxsw_sp_from_ptys_advert_link(eth_proto_admin);
+	mlxsw_sp_from_ptys_speed_duplex(netif_carrier_ok(dev),
+					eth_proto_oper, cmd);
+
+	eth_proto_oper = eth_proto_oper ? eth_proto_oper : eth_proto_cap;
+	cmd->port = mlxsw_sp_port_connector_port(eth_proto_oper);
+	cmd->lp_advertising = mlxsw_sp_from_ptys_advert_link(eth_proto_oper);
+
+	cmd->transceiver = XCVR_INTERNAL;
+	return 0;
+}
+
+static u32 mlxsw_sp_to_ptys_advert_link(u32 advertising)
+{
+	u32 ptys_proto = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SP_PORT_LINK_MODE_LEN; i++) {
+		if (advertising & mlxsw_sp_port_link_mode[i].advertised)
+			ptys_proto |= mlxsw_sp_port_link_mode[i].mask;
+	}
+	return ptys_proto;
+}
+
+static u32 mlxsw_sp_to_ptys_speed(u32 speed)
+{
+	u32 ptys_proto = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SP_PORT_LINK_MODE_LEN; i++) {
+		if (speed == mlxsw_sp_port_link_mode[i].speed)
+			ptys_proto |= mlxsw_sp_port_link_mode[i].mask;
+	}
+	return ptys_proto;
+}
+
+static int mlxsw_sp_port_set_settings(struct net_device *dev,
+				      struct ethtool_cmd *cmd)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char ptys_pl[MLXSW_REG_PTYS_LEN];
+	u32 speed;
+	u32 eth_proto_new;
+	u32 eth_proto_cap;
+	u32 eth_proto_admin;
+	bool is_up;
+	int err;
+
+	speed = ethtool_cmd_speed(cmd);
+
+	eth_proto_new = cmd->autoneg == AUTONEG_ENABLE ?
+		mlxsw_sp_to_ptys_advert_link(cmd->advertising) :
+		mlxsw_sp_to_ptys_speed(speed);
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sp_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ptys), ptys_pl);
+	if (err) {
+		netdev_err(dev, "Failed to get proto");
+		return err;
+	}
+	mlxsw_reg_ptys_unpack(ptys_pl, &eth_proto_cap, &eth_proto_admin, NULL);
+
+	eth_proto_new = eth_proto_new & eth_proto_cap;
+	if (!eth_proto_new) {
+		netdev_err(dev, "Not supported proto admin requested");
+		return -EINVAL;
+	}
+	if (eth_proto_new == eth_proto_admin)
+		return 0;
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sp_port->local_port, eth_proto_new);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ptys), ptys_pl);
+	if (err) {
+		netdev_err(dev, "Failed to set proto admin");
+		return err;
+	}
+
+	err = mlxsw_sp_port_oper_status_get(mlxsw_sp_port, &is_up);
+	if (err) {
+		netdev_err(dev, "Failed to get oper status");
+		return err;
+	}
+	if (!is_up)
+		return 0;
+
+	err = mlxsw_sp_port_admin_status_set(mlxsw_sp_port, false);
+	if (err) {
+		netdev_err(dev, "Failed to set admin status");
+		return err;
+	}
+
+	err = mlxsw_sp_port_admin_status_set(mlxsw_sp_port, true);
+	if (err) {
+		netdev_err(dev, "Failed to set admin status");
+		return err;
+	}
+
+	return 0;
+}
+
+static const struct ethtool_ops mlxsw_sp_port_ethtool_ops = {
+	.get_drvinfo		= mlxsw_sp_port_get_drvinfo,
+	.get_link		= ethtool_op_get_link,
+	.get_strings		= mlxsw_sp_port_get_strings,
+	.get_ethtool_stats	= mlxsw_sp_port_get_stats,
+	.get_sset_count		= mlxsw_sp_port_get_sset_count,
+	.get_settings		= mlxsw_sp_port_get_settings,
+	.set_settings		= mlxsw_sp_port_set_settings,
+};
+
+static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port;
+	struct net_device *dev;
+	bool usable;
+	int err;
+
+	dev = alloc_etherdev(sizeof(struct mlxsw_sp_port));
+	if (!dev)
+		return -ENOMEM;
+	mlxsw_sp_port = netdev_priv(dev);
+	mlxsw_sp_port->dev = dev;
+	mlxsw_sp_port->mlxsw_sp = mlxsw_sp;
+	mlxsw_sp_port->local_port = local_port;
+	mlxsw_sp_port->learning = 1;
+	mlxsw_sp_port->learning_sync = 1;
+	mlxsw_sp_port->uc_flood = 1;
+	mlxsw_sp_port->pvid = 1;
+
+	mlxsw_sp_port->pcpu_stats =
+		netdev_alloc_pcpu_stats(struct mlxsw_sp_port_pcpu_stats);
+	if (!mlxsw_sp_port->pcpu_stats) {
+		err = -ENOMEM;
+		goto err_alloc_stats;
+	}
+
+	dev->netdev_ops = &mlxsw_sp_port_netdev_ops;
+	dev->ethtool_ops = &mlxsw_sp_port_ethtool_ops;
+
+	err = mlxsw_sp_port_dev_addr_init(mlxsw_sp_port);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Unable to init port mac address\n",
+			mlxsw_sp_port->local_port);
+		goto err_dev_addr_init;
+	}
+
+	netif_carrier_off(dev);
+
+	dev->features |= NETIF_F_NETNS_LOCAL | NETIF_F_LLTX | NETIF_F_SG |
+			 NETIF_F_HW_VLAN_CTAG_FILTER;
+
+	/* Each packet needs to have a Tx header (metadata) on top all other
+	 * headers.
+	 */
+	dev->hard_header_len += MLXSW_TXHDR_LEN;
+
+	err = mlxsw_sp_port_module_check(mlxsw_sp_port, &usable);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to check module\n",
+			mlxsw_sp_port->local_port);
+		goto err_port_module_check;
+	}
+
+	if (!usable) {
+		dev_dbg(mlxsw_sp->bus_info->dev, "Port %d: Not usable, skipping initialization\n",
+			mlxsw_sp_port->local_port);
+		goto port_not_usable;
+	}
+
+	err = mlxsw_sp_port_system_port_mapping_set(mlxsw_sp_port);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to set system port mapping\n",
+			mlxsw_sp_port->local_port);
+		goto err_port_system_port_mapping_set;
+	}
+
+	err = mlxsw_sp_port_swid_set(mlxsw_sp_port, 0);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to set SWID\n",
+			mlxsw_sp_port->local_port);
+		goto err_port_swid_set;
+	}
+
+	err = mlxsw_sp_port_mtu_set(mlxsw_sp_port, ETH_DATA_LEN);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to set MTU\n",
+			mlxsw_sp_port->local_port);
+		goto err_port_mtu_set;
+	}
+
+	err = mlxsw_sp_port_admin_status_set(mlxsw_sp_port, false);
+	if (err)
+		goto err_port_admin_status_set;
+
+	err = mlxsw_sp_port_buffers_init(mlxsw_sp_port);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to initialize buffers\n",
+			mlxsw_sp_port->local_port);
+		goto err_port_buffers_init;
+	}
+
+	mlxsw_sp_port_switchdev_init(mlxsw_sp_port);
+	err = register_netdev(dev);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Port %d: Failed to register netdev\n",
+			mlxsw_sp_port->local_port);
+		goto err_register_netdev;
+	}
+
+	err = mlxsw_sp_port_vlan_init(mlxsw_sp_port);
+	if (err)
+		goto err_port_vlan_init;
+
+	mlxsw_sp->ports[local_port] = mlxsw_sp_port;
+	return 0;
+
+err_port_vlan_init:
+	unregister_netdev(dev);
+err_register_netdev:
+err_port_buffers_init:
+err_port_admin_status_set:
+err_port_mtu_set:
+err_port_swid_set:
+err_port_system_port_mapping_set:
+port_not_usable:
+err_port_module_check:
+err_dev_addr_init:
+	free_percpu(mlxsw_sp_port->pcpu_stats);
+err_alloc_stats:
+	free_netdev(dev);
+	return err;
+}
+
+static void mlxsw_sp_vfids_fini(struct mlxsw_sp *mlxsw_sp)
+{
+	u16 vfid;
+
+	for_each_set_bit(vfid, mlxsw_sp->active_vfids, VLAN_N_VID)
+		mlxsw_sp_vfid_destroy(mlxsw_sp, vfid);
+}
+
+static void mlxsw_sp_port_remove(struct mlxsw_sp *mlxsw_sp, u8 local_port)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = mlxsw_sp->ports[local_port];
+
+	if (!mlxsw_sp_port)
+		return;
+	mlxsw_sp_port_kill_vid(mlxsw_sp_port->dev, 0, 1);
+	unregister_netdev(mlxsw_sp_port->dev); /* This calls ndo_stop */
+	mlxsw_sp_port_switchdev_fini(mlxsw_sp_port);
+	free_percpu(mlxsw_sp_port->pcpu_stats);
+	free_netdev(mlxsw_sp_port->dev);
+}
+
+static void mlxsw_sp_ports_remove(struct mlxsw_sp *mlxsw_sp)
+{
+	int i;
+
+	for (i = 1; i < MLXSW_PORT_MAX_PORTS; i++)
+		mlxsw_sp_port_remove(mlxsw_sp, i);
+	kfree(mlxsw_sp->ports);
+}
+
+static int mlxsw_sp_ports_create(struct mlxsw_sp *mlxsw_sp)
+{
+	size_t alloc_size;
+	int i;
+	int err;
+
+	alloc_size = sizeof(struct mlxsw_sp_port *) * MLXSW_PORT_MAX_PORTS;
+	mlxsw_sp->ports = kzalloc(alloc_size, GFP_KERNEL);
+	if (!mlxsw_sp->ports)
+		return -ENOMEM;
+
+	for (i = 1; i < MLXSW_PORT_MAX_PORTS; i++) {
+		err = mlxsw_sp_port_create(mlxsw_sp, i);
+		if (err)
+			goto err_port_create;
+	}
+	return 0;
+
+err_port_create:
+	for (i--; i >= 1; i--)
+		mlxsw_sp_port_remove(mlxsw_sp, i);
+	kfree(mlxsw_sp->ports);
+	return err;
+}
+
+static void mlxsw_sp_pude_event_func(const struct mlxsw_reg_info *reg,
+				     char *pude_pl, void *priv)
+{
+	struct mlxsw_sp *mlxsw_sp = priv;
+	struct mlxsw_sp_port *mlxsw_sp_port;
+	enum mlxsw_reg_pude_oper_status status;
+	u8 local_port;
+
+	local_port = mlxsw_reg_pude_local_port_get(pude_pl);
+	mlxsw_sp_port = mlxsw_sp->ports[local_port];
+	if (!mlxsw_sp_port) {
+		dev_warn(mlxsw_sp->bus_info->dev, "Port %d: Link event received for non-existent port\n",
+			 local_port);
+		return;
+	}
+
+	status = mlxsw_reg_pude_oper_status_get(pude_pl);
+	if (status == MLXSW_PORT_OPER_STATUS_UP) {
+		netdev_info(mlxsw_sp_port->dev, "link up\n");
+		netif_carrier_on(mlxsw_sp_port->dev);
+	} else {
+		netdev_info(mlxsw_sp_port->dev, "link down\n");
+		netif_carrier_off(mlxsw_sp_port->dev);
+	}
+}
+
+static struct mlxsw_event_listener mlxsw_sp_pude_event = {
+	.func = mlxsw_sp_pude_event_func,
+	.trap_id = MLXSW_TRAP_ID_PUDE,
+};
+
+static int mlxsw_sp_event_register(struct mlxsw_sp *mlxsw_sp,
+				   enum mlxsw_event_trap_id trap_id)
+{
+	struct mlxsw_event_listener *el;
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int err;
+
+	switch (trap_id) {
+	case MLXSW_TRAP_ID_PUDE:
+		el = &mlxsw_sp_pude_event;
+		break;
+	}
+	err = mlxsw_core_event_listener_register(mlxsw_sp->core, el, mlxsw_sp);
+	if (err)
+		return err;
+
+	mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_FORWARD, trap_id);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(hpkt), hpkt_pl);
+	if (err)
+		goto err_event_trap_set;
+
+	return 0;
+
+err_event_trap_set:
+	mlxsw_core_event_listener_unregister(mlxsw_sp->core, el, mlxsw_sp);
+	return err;
+}
+
+static void mlxsw_sp_event_unregister(struct mlxsw_sp *mlxsw_sp,
+				      enum mlxsw_event_trap_id trap_id)
+{
+	struct mlxsw_event_listener *el;
+
+	switch (trap_id) {
+	case MLXSW_TRAP_ID_PUDE:
+		el = &mlxsw_sp_pude_event;
+		break;
+	}
+	mlxsw_core_event_listener_unregister(mlxsw_sp->core, el, mlxsw_sp);
+}
+
+static void mlxsw_sp_rx_listener_func(struct sk_buff *skb, u8 local_port,
+				      void *priv)
+{
+	struct mlxsw_sp *mlxsw_sp = priv;
+	struct mlxsw_sp_port *mlxsw_sp_port = mlxsw_sp->ports[local_port];
+	struct mlxsw_sp_port_pcpu_stats *pcpu_stats;
+
+	if (unlikely(!mlxsw_sp_port)) {
+		dev_warn_ratelimited(mlxsw_sp->bus_info->dev, "Port %d: skb received for non-existent port\n",
+				     local_port);
+		return;
+	}
+
+	skb->dev = mlxsw_sp_port->dev;
+
+	pcpu_stats = this_cpu_ptr(mlxsw_sp_port->pcpu_stats);
+	u64_stats_update_begin(&pcpu_stats->syncp);
+	pcpu_stats->rx_packets++;
+	pcpu_stats->rx_bytes += skb->len;
+	u64_stats_update_end(&pcpu_stats->syncp);
+
+	skb->protocol = eth_type_trans(skb, skb->dev);
+	netif_receive_skb(skb);
+}
+
+static const struct mlxsw_rx_listener mlxsw_sp_rx_listener[] = {
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_FDB_MC,
+	},
+	/* Traps for specific L2 packet types, not trapped as FDB MC */
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_STP,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_LACP,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_EAPOL,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_LLDP,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_MMRP,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_MVRP,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_RPVST,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_DHCP,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_QUERY,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V1_REPORT,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V2_REPORT,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V2_LEAVE,
+	},
+	{
+		.func = mlxsw_sp_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V3_REPORT,
+	},
+};
+
+static int mlxsw_sp_traps_init(struct mlxsw_sp *mlxsw_sp)
+{
+	char htgt_pl[MLXSW_REG_HTGT_LEN];
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int i;
+	int err;
+
+	mlxsw_reg_htgt_pack(htgt_pl, MLXSW_REG_HTGT_TRAP_GROUP_RX);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(htgt), htgt_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_htgt_pack(htgt_pl, MLXSW_REG_HTGT_TRAP_GROUP_CTRL);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(htgt), htgt_pl);
+	if (err)
+		return err;
+
+	for (i = 0; i < ARRAY_SIZE(mlxsw_sp_rx_listener); i++) {
+		err = mlxsw_core_rx_listener_register(mlxsw_sp->core,
+						      &mlxsw_sp_rx_listener[i],
+						      mlxsw_sp);
+		if (err)
+			goto err_rx_listener_register;
+
+		mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_TRAP_TO_CPU,
+				    mlxsw_sp_rx_listener[i].trap_id);
+		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(hpkt), hpkt_pl);
+		if (err)
+			goto err_rx_trap_set;
+	}
+	return 0;
+
+err_rx_trap_set:
+	mlxsw_core_rx_listener_unregister(mlxsw_sp->core,
+					  &mlxsw_sp_rx_listener[i],
+					  mlxsw_sp);
+err_rx_listener_register:
+	for (i--; i >= 0; i--) {
+		mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_FORWARD,
+				    mlxsw_sp_rx_listener[i].trap_id);
+		mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(hpkt), hpkt_pl);
+
+		mlxsw_core_rx_listener_unregister(mlxsw_sp->core,
+						  &mlxsw_sp_rx_listener[i],
+						  mlxsw_sp);
+	}
+	return err;
+}
+
+static void mlxsw_sp_traps_fini(struct mlxsw_sp *mlxsw_sp)
+{
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(mlxsw_sp_rx_listener); i++) {
+		mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_FORWARD,
+				    mlxsw_sp_rx_listener[i].trap_id);
+		mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(hpkt), hpkt_pl);
+
+		mlxsw_core_rx_listener_unregister(mlxsw_sp->core,
+						  &mlxsw_sp_rx_listener[i],
+						  mlxsw_sp);
+	}
+}
+
+static int __mlxsw_sp_flood_init(struct mlxsw_core *mlxsw_core,
+				 enum mlxsw_reg_sfgc_type type,
+				 enum mlxsw_reg_sfgc_bridge_type bridge_type)
+{
+	enum mlxsw_flood_table_type table_type;
+	enum mlxsw_sp_flood_table flood_table;
+	char sfgc_pl[MLXSW_REG_SFGC_LEN];
+
+	if (bridge_type == MLXSW_REG_SFGC_BRIDGE_TYPE_VFID) {
+		table_type = MLXSW_REG_SFGC_TABLE_TYPE_FID;
+		flood_table = 0;
+	} else {
+		table_type = MLXSW_REG_SFGC_TABLE_TYPE_FID_OFFEST;
+		if (type == MLXSW_REG_SFGC_TYPE_UNKNOWN_UNICAST)
+			flood_table = MLXSW_SP_FLOOD_TABLE_UC;
+		else
+			flood_table = MLXSW_SP_FLOOD_TABLE_BM;
+	}
+
+	mlxsw_reg_sfgc_pack(sfgc_pl, type, bridge_type, table_type,
+			    flood_table);
+	return mlxsw_reg_write(mlxsw_core, MLXSW_REG(sfgc), sfgc_pl);
+}
+
+static int mlxsw_sp_flood_init(struct mlxsw_sp *mlxsw_sp)
+{
+	int type, err;
+
+	/* For non-offloaded netdevs, flood all traffic types to CPU
+	 * port.
+	 */
+	for (type = 0; type < MLXSW_REG_SFGC_TYPE_MAX; type++) {
+		if (type == MLXSW_REG_SFGC_TYPE_RESERVED)
+			continue;
+
+		err = __mlxsw_sp_flood_init(mlxsw_sp->core, type,
+					    MLXSW_REG_SFGC_BRIDGE_TYPE_VFID);
+		if (err)
+			return err;
+	}
+
+	/* For bridged ports, use one flooding table for unknown unicast
+	 * traffic and a second table for unregistered multicast and
+	 * broadcast.
+	 */
+	for (type = 0; type < MLXSW_REG_SFGC_TYPE_MAX; type++) {
+		if (type == MLXSW_REG_SFGC_TYPE_RESERVED)
+			continue;
+
+		err = __mlxsw_sp_flood_init(mlxsw_sp->core, type,
+					    MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int mlxsw_sp_init(void *priv, struct mlxsw_core *mlxsw_core,
+			 const struct mlxsw_bus_info *mlxsw_bus_info)
+{
+	struct mlxsw_sp *mlxsw_sp = priv;
+	int err;
+
+	mlxsw_sp->core = mlxsw_core;
+	mlxsw_sp->bus_info = mlxsw_bus_info;
+
+	err = mlxsw_sp_base_mac_get(mlxsw_sp);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to get base mac\n");
+		return err;
+	}
+
+	err = mlxsw_sp_ports_create(mlxsw_sp);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to create ports\n");
+		goto err_ports_create;
+	}
+
+	err = mlxsw_sp_event_register(mlxsw_sp, MLXSW_TRAP_ID_PUDE);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to register for PUDE events\n");
+		goto err_event_register;
+	}
+
+	err = mlxsw_sp_traps_init(mlxsw_sp);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to set traps for RX\n");
+		goto err_rx_listener_register;
+	}
+
+	err = mlxsw_sp_flood_init(mlxsw_sp);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to initialize flood tables\n");
+		goto err_flood_init;
+	}
+
+	err = mlxsw_sp_buffers_init(mlxsw_sp);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to initialize buffers\n");
+		goto err_buffers_init;
+	}
+
+	err = mlxsw_sp_switchdev_init(mlxsw_sp);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to initialize switchdev\n");
+		goto err_switchdev_init;
+	}
+
+	return 0;
+
+err_switchdev_init:
+err_buffers_init:
+err_flood_init:
+	mlxsw_sp_traps_fini(mlxsw_sp);
+err_rx_listener_register:
+	mlxsw_sp_event_unregister(mlxsw_sp, MLXSW_TRAP_ID_PUDE);
+err_event_register:
+	mlxsw_sp_ports_remove(mlxsw_sp);
+err_ports_create:
+	mlxsw_sp_vfids_fini(mlxsw_sp);
+	return err;
+}
+
+static void mlxsw_sp_fini(void *priv)
+{
+	struct mlxsw_sp *mlxsw_sp = priv;
+
+	mlxsw_sp_switchdev_fini(mlxsw_sp);
+	mlxsw_sp_traps_fini(mlxsw_sp);
+	mlxsw_sp_event_unregister(mlxsw_sp, MLXSW_TRAP_ID_PUDE);
+	mlxsw_sp_ports_remove(mlxsw_sp);
+	mlxsw_sp_vfids_fini(mlxsw_sp);
+}
+
+static struct mlxsw_config_profile mlxsw_sp_config_profile = {
+	.used_max_vepa_channels		= 1,
+	.max_vepa_channels		= 0,
+	.used_max_lag			= 1,
+	.max_lag			= 64,
+	.used_max_port_per_lag		= 1,
+	.max_port_per_lag		= 16,
+	.used_max_mid			= 1,
+	.max_mid			= 7000,
+	.used_max_pgt			= 1,
+	.max_pgt			= 0,
+	.used_max_system_port		= 1,
+	.max_system_port		= 64,
+	.used_max_vlan_groups		= 1,
+	.max_vlan_groups		= 127,
+	.used_max_regions		= 1,
+	.max_regions			= 400,
+	.used_flood_tables		= 1,
+	.used_flood_mode		= 1,
+	.flood_mode			= 3,
+	.max_fid_offset_flood_tables	= 2,
+	.fid_offset_flood_table_size	= VLAN_N_VID - 1,
+	.max_fid_flood_tables		= 1,
+	.fid_flood_table_size		= VLAN_N_VID,
+	.used_max_ib_mc			= 1,
+	.max_ib_mc			= 0,
+	.used_max_pkey			= 1,
+	.max_pkey			= 0,
+	.swid_config			= {
+		{
+			.used_type	= 1,
+			.type		= MLXSW_PORT_SWID_TYPE_ETH,
+		}
+	},
+};
+
+static struct mlxsw_driver mlxsw_sp_driver = {
+	.kind			= MLXSW_DEVICE_KIND_SPECTRUM,
+	.owner			= THIS_MODULE,
+	.priv_size		= sizeof(struct mlxsw_sp),
+	.init			= mlxsw_sp_init,
+	.fini			= mlxsw_sp_fini,
+	.txhdr_construct	= mlxsw_sp_txhdr_construct,
+	.txhdr_len		= MLXSW_TXHDR_LEN,
+	.profile		= &mlxsw_sp_config_profile,
+};
+
+static bool mlxsw_sp_port_dev_check(const struct net_device *dev)
+{
+	return dev->netdev_ops == &mlxsw_sp_port_netdev_ops;
+}
+
+static int mlxsw_sp_port_bridge_join(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	struct net_device *dev = mlxsw_sp_port->dev;
+	int err;
+
+	/* When port is not bridged untagged packets are tagged with
+	 * PVID=VID=1, thereby creating an implicit VLAN interface in
+	 * the device. Remove it and let bridge code take care of its
+	 * own VLANs.
+	 */
+	err = mlxsw_sp_port_kill_vid(dev, 0, 1);
+	if (err)
+		netdev_err(dev, "Failed to remove VID 1\n");
+
+	return err;
+}
+
+static int mlxsw_sp_port_bridge_leave(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	struct net_device *dev = mlxsw_sp_port->dev;
+	int err;
+
+	/* Add implicit VLAN interface in the device, so that untagged
+	 * packets will be classified to the default vFID.
+	 */
+	err = mlxsw_sp_port_add_vid(dev, 0, 1);
+	if (err)
+		netdev_err(dev, "Failed to add VID 1\n");
+
+	return err;
+}
+
+static bool mlxsw_sp_master_bridge_check(struct mlxsw_sp *mlxsw_sp,
+					 struct net_device *br_dev)
+{
+	return !mlxsw_sp->master_bridge.dev ||
+	       mlxsw_sp->master_bridge.dev == br_dev;
+}
+
+static void mlxsw_sp_master_bridge_inc(struct mlxsw_sp *mlxsw_sp,
+				       struct net_device *br_dev)
+{
+	mlxsw_sp->master_bridge.dev = br_dev;
+	mlxsw_sp->master_bridge.ref_count++;
+}
+
+static void mlxsw_sp_master_bridge_dec(struct mlxsw_sp *mlxsw_sp,
+				       struct net_device *br_dev)
+{
+	if (--mlxsw_sp->master_bridge.ref_count == 0)
+		mlxsw_sp->master_bridge.dev = NULL;
+}
+
+static int mlxsw_sp_netdevice_event(struct notifier_block *unused,
+				    unsigned long event, void *ptr)
+{
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
+	struct netdev_notifier_changeupper_info *info;
+	struct mlxsw_sp_port *mlxsw_sp_port;
+	struct net_device *upper_dev;
+	struct mlxsw_sp *mlxsw_sp;
+	int err;
+
+	if (!mlxsw_sp_port_dev_check(dev))
+		return NOTIFY_DONE;
+
+	mlxsw_sp_port = netdev_priv(dev);
+	mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	info = ptr;
+
+	switch (event) {
+	case NETDEV_PRECHANGEUPPER:
+		upper_dev = info->upper_dev;
+		/* HW limitation forbids to put ports to multiple bridges. */
+		if (info->master && info->linking &&
+		    netif_is_bridge_master(upper_dev) &&
+		    !mlxsw_sp_master_bridge_check(mlxsw_sp, upper_dev))
+			return NOTIFY_BAD;
+		break;
+	case NETDEV_CHANGEUPPER:
+		upper_dev = info->upper_dev;
+		if (info->master &&
+		    netif_is_bridge_master(upper_dev)) {
+			if (info->linking) {
+				err = mlxsw_sp_port_bridge_join(mlxsw_sp_port);
+				if (err)
+					netdev_err(dev, "Failed to join bridge\n");
+				mlxsw_sp_master_bridge_inc(mlxsw_sp, upper_dev);
+				mlxsw_sp_port->bridged = 1;
+			} else {
+				err = mlxsw_sp_port_bridge_leave(mlxsw_sp_port);
+				if (err)
+					netdev_err(dev, "Failed to leave bridge\n");
+				mlxsw_sp_port->bridged = 0;
+				mlxsw_sp_master_bridge_dec(mlxsw_sp, upper_dev);
+			}
+		}
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block mlxsw_sp_netdevice_nb __read_mostly = {
+	.notifier_call = mlxsw_sp_netdevice_event,
+};
+
+static int __init mlxsw_sp_module_init(void)
+{
+	int err;
+
+	register_netdevice_notifier(&mlxsw_sp_netdevice_nb);
+	err = mlxsw_core_driver_register(&mlxsw_sp_driver);
+	if (err)
+		goto err_core_driver_register;
+	return 0;
+
+err_core_driver_register:
+	unregister_netdevice_notifier(&mlxsw_sp_netdevice_nb);
+	return err;
+}
+
+static void __exit mlxsw_sp_module_exit(void)
+{
+	mlxsw_core_driver_unregister(&mlxsw_sp_driver);
+	unregister_netdevice_notifier(&mlxsw_sp_netdevice_nb);
+}
+
+module_init(mlxsw_sp_module_init);
+module_exit(mlxsw_sp_module_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Jiri Pirko <jiri@mellanox.com>");
+MODULE_DESCRIPTION("Mellanox Spectrum driver");
+MODULE_MLXSW_DRIVER_ALIAS(MLXSW_DEVICE_KIND_SPECTRUM);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
new file mode 100644
index 000000000..4365c8bcc
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -0,0 +1,122 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_SPECTRUM_H
+#define _MLXSW_SPECTRUM_H
+
+#include <linux/types.h>
+#include <linux/netdevice.h>
+#include <linux/bitops.h>
+#include <linux/if_vlan.h>
+#include <net/switchdev.h>
+
+#include "core.h"
+
+#define MLXSW_SP_VFID_BASE VLAN_N_VID
+
+struct mlxsw_sp_port;
+
+struct mlxsw_sp {
+	unsigned long active_vfids[BITS_TO_LONGS(VLAN_N_VID)];
+	unsigned long active_fids[BITS_TO_LONGS(VLAN_N_VID)];
+	struct mlxsw_sp_port **ports;
+	struct mlxsw_core *core;
+	const struct mlxsw_bus_info *bus_info;
+	unsigned char base_mac[ETH_ALEN];
+	struct {
+		struct delayed_work dw;
+#define MLXSW_SP_DEFAULT_LEARNING_INTERVAL 100
+		unsigned int interval; /* ms */
+	} fdb_notify;
+#define MLXSW_SP_DEFAULT_AGEING_TIME 300
+	u32 ageing_time;
+	struct {
+		struct net_device *dev;
+		unsigned int ref_count;
+	} master_bridge;
+};
+
+struct mlxsw_sp_port_pcpu_stats {
+	u64			rx_packets;
+	u64			rx_bytes;
+	u64			tx_packets;
+	u64			tx_bytes;
+	struct u64_stats_sync	syncp;
+	u32			tx_dropped;
+};
+
+struct mlxsw_sp_port {
+	struct net_device *dev;
+	struct mlxsw_sp_port_pcpu_stats __percpu *pcpu_stats;
+	struct mlxsw_sp *mlxsw_sp;
+	u8 local_port;
+	u8 stp_state;
+	u8 learning:1,
+	   learning_sync:1,
+	   uc_flood:1,
+	   bridged:1;
+	u16 pvid;
+	/* 802.1Q bridge VLANs */
+	unsigned long active_vlans[BITS_TO_LONGS(VLAN_N_VID)];
+	/* VLAN interfaces */
+	unsigned long active_vfids[BITS_TO_LONGS(VLAN_N_VID)];
+	u16 nr_vfids;
+};
+
+enum mlxsw_sp_flood_table {
+	MLXSW_SP_FLOOD_TABLE_UC,
+	MLXSW_SP_FLOOD_TABLE_BM,
+};
+
+int mlxsw_sp_buffers_init(struct mlxsw_sp *mlxsw_sp);
+int mlxsw_sp_port_buffers_init(struct mlxsw_sp_port *mlxsw_sp_port);
+
+int mlxsw_sp_switchdev_init(struct mlxsw_sp *mlxsw_sp);
+void mlxsw_sp_switchdev_fini(struct mlxsw_sp *mlxsw_sp);
+int mlxsw_sp_port_vlan_init(struct mlxsw_sp_port *mlxsw_sp_port);
+void mlxsw_sp_port_switchdev_init(struct mlxsw_sp_port *mlxsw_sp_port);
+void mlxsw_sp_port_switchdev_fini(struct mlxsw_sp_port *mlxsw_sp_port);
+int mlxsw_sp_port_vid_to_fid_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				 enum mlxsw_reg_svfa_mt mt, bool valid, u16 fid,
+				 u16 vid);
+int mlxsw_sp_port_vlan_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid_begin,
+			   u16 vid_end, bool is_member, bool untagged);
+int mlxsw_sp_port_add_vid(struct net_device *dev, __be16 __always_unused proto,
+			  u16 vid);
+int mlxsw_sp_port_kill_vid(struct net_device *dev,
+			   __be16 __always_unused proto, u16 vid);
+
+#endif
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum_buffers.c b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum_buffers.c
new file mode 100644
index 000000000..d59195e3f
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum_buffers.c
@@ -0,0 +1,422 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/spectrum_buffers.c
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+
+#include "spectrum.h"
+#include "core.h"
+#include "port.h"
+#include "reg.h"
+
+struct mlxsw_sp_pb {
+	u8 index;
+	u16 size;
+};
+
+#define MLXSW_SP_PB(_index, _size)	\
+	{				\
+		.index = _index,	\
+		.size = _size,		\
+	}
+
+static const struct mlxsw_sp_pb mlxsw_sp_pbs[] = {
+	MLXSW_SP_PB(0, 208),
+	MLXSW_SP_PB(1, 208),
+	MLXSW_SP_PB(2, 208),
+	MLXSW_SP_PB(3, 208),
+	MLXSW_SP_PB(4, 208),
+	MLXSW_SP_PB(5, 208),
+	MLXSW_SP_PB(6, 208),
+	MLXSW_SP_PB(7, 208),
+	MLXSW_SP_PB(9, 208),
+};
+
+#define MLXSW_SP_PBS_LEN ARRAY_SIZE(mlxsw_sp_pbs)
+
+static int mlxsw_sp_port_pb_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	char pbmc_pl[MLXSW_REG_PBMC_LEN];
+	int i;
+
+	mlxsw_reg_pbmc_pack(pbmc_pl, mlxsw_sp_port->local_port,
+			    0xffff, 0xffff / 2);
+	for (i = 0; i < MLXSW_SP_PBS_LEN; i++) {
+		const struct mlxsw_sp_pb *pb;
+
+		pb = &mlxsw_sp_pbs[i];
+		mlxsw_reg_pbmc_lossy_buffer_pack(pbmc_pl, pb->index, pb->size);
+	}
+	return mlxsw_reg_write(mlxsw_sp_port->mlxsw_sp->core,
+			       MLXSW_REG(pbmc), pbmc_pl);
+}
+
+#define MLXSW_SP_SB_BYTES_PER_CELL 96
+
+struct mlxsw_sp_sb_pool {
+	u8 pool;
+	enum mlxsw_reg_sbpr_dir dir;
+	enum mlxsw_reg_sbpr_mode mode;
+	u32 size;
+};
+
+#define MLXSW_SP_SB_POOL_INGRESS_SIZE				\
+	((15000000 - (2 * 20000 * MLXSW_PORT_MAX_PORTS)) /	\
+	 MLXSW_SP_SB_BYTES_PER_CELL)
+#define MLXSW_SP_SB_POOL_EGRESS_SIZE				\
+	((14000000 - (8 * 1500 * MLXSW_PORT_MAX_PORTS)) /	\
+	 MLXSW_SP_SB_BYTES_PER_CELL)
+
+#define MLXSW_SP_SB_POOL(_pool, _dir, _mode, _size)		\
+	{							\
+		.pool = _pool,					\
+		.dir = _dir,					\
+		.mode = _mode,					\
+		.size = _size,					\
+	}
+
+#define MLXSW_SP_SB_POOL_INGRESS(_pool, _size)			\
+	MLXSW_SP_SB_POOL(_pool, MLXSW_REG_SBPR_DIR_INGRESS,	\
+			 MLXSW_REG_SBPR_MODE_DYNAMIC, _size)
+
+#define MLXSW_SP_SB_POOL_EGRESS(_pool, _size)			\
+	MLXSW_SP_SB_POOL(_pool, MLXSW_REG_SBPR_DIR_EGRESS,	\
+			 MLXSW_REG_SBPR_MODE_DYNAMIC, _size)
+
+static const struct mlxsw_sp_sb_pool mlxsw_sp_sb_pools[] = {
+	MLXSW_SP_SB_POOL_INGRESS(0, MLXSW_SP_SB_POOL_INGRESS_SIZE),
+	MLXSW_SP_SB_POOL_INGRESS(1, 0),
+	MLXSW_SP_SB_POOL_INGRESS(2, 0),
+	MLXSW_SP_SB_POOL_INGRESS(3, 0),
+	MLXSW_SP_SB_POOL_EGRESS(0, MLXSW_SP_SB_POOL_EGRESS_SIZE),
+	MLXSW_SP_SB_POOL_EGRESS(1, 0),
+	MLXSW_SP_SB_POOL_EGRESS(2, 0),
+	MLXSW_SP_SB_POOL_EGRESS(2, MLXSW_SP_SB_POOL_EGRESS_SIZE),
+};
+
+#define MLXSW_SP_SB_POOLS_LEN ARRAY_SIZE(mlxsw_sp_sb_pools)
+
+static int mlxsw_sp_sb_pools_init(struct mlxsw_sp *mlxsw_sp)
+{
+	char sbpr_pl[MLXSW_REG_SBPR_LEN];
+	int i;
+	int err;
+
+	for (i = 0; i < MLXSW_SP_SB_POOLS_LEN; i++) {
+		const struct mlxsw_sp_sb_pool *pool;
+
+		pool = &mlxsw_sp_sb_pools[i];
+		mlxsw_reg_sbpr_pack(sbpr_pl, pool->pool, pool->dir,
+				    pool->mode, pool->size);
+		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sbpr), sbpr_pl);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+struct mlxsw_sp_sb_cm {
+	union {
+		u8 pg;
+		u8 tc;
+	} u;
+	enum mlxsw_reg_sbcm_dir dir;
+	u32 min_buff;
+	u32 max_buff;
+	u8 pool;
+};
+
+#define MLXSW_SP_SB_CM(_pg_tc, _dir, _min_buff, _max_buff, _pool)	\
+	{								\
+		.u.pg = _pg_tc,						\
+		.dir = _dir,						\
+		.min_buff = _min_buff,					\
+		.max_buff = _max_buff,					\
+		.pool = _pool,						\
+	}
+
+#define MLXSW_SP_SB_CM_INGRESS(_pg, _min_buff, _max_buff)		\
+	MLXSW_SP_SB_CM(_pg, MLXSW_REG_SBCM_DIR_INGRESS,			\
+		       _min_buff, _max_buff, 0)
+
+#define MLXSW_SP_SB_CM_EGRESS(_tc, _min_buff, _max_buff)		\
+	MLXSW_SP_SB_CM(_tc, MLXSW_REG_SBCM_DIR_EGRESS,			\
+		       _min_buff, _max_buff, 0)
+
+#define MLXSW_SP_CPU_PORT_SB_CM_EGRESS(_tc)				\
+	MLXSW_SP_SB_CM(_tc, MLXSW_REG_SBCM_DIR_EGRESS, 104, 2, 3)
+
+static const struct mlxsw_sp_sb_cm mlxsw_sp_sb_cms[] = {
+	MLXSW_SP_SB_CM_INGRESS(0, 10000 / MLXSW_SP_SB_BYTES_PER_CELL, 8),
+	MLXSW_SP_SB_CM_INGRESS(1, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(2, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(3, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(4, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(5, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(6, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(7, 0, 0),
+	MLXSW_SP_SB_CM_INGRESS(9, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff),
+	MLXSW_SP_SB_CM_EGRESS(0, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(1, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(2, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(3, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(4, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(5, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(6, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(7, 1500 / MLXSW_SP_SB_BYTES_PER_CELL, 9),
+	MLXSW_SP_SB_CM_EGRESS(8, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(9, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(10, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(11, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(12, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(13, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(14, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(15, 0, 0),
+	MLXSW_SP_SB_CM_EGRESS(16, 1, 0xff),
+};
+
+#define MLXSW_SP_SB_CMS_LEN ARRAY_SIZE(mlxsw_sp_sb_cms)
+
+static const struct mlxsw_sp_sb_cm mlxsw_sp_cpu_port_sb_cms[] = {
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(0),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(1),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(2),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(3),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(4),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(5),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(6),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(7),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(8),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(9),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(10),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(11),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(12),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(13),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(14),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(15),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(16),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(17),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(18),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(19),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(20),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(21),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(22),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(23),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(24),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(25),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(26),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(27),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(28),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(29),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(30),
+	MLXSW_SP_CPU_PORT_SB_CM_EGRESS(31),
+};
+
+#define MLXSW_SP_CPU_PORT_SB_MCS_LEN \
+	ARRAY_SIZE(mlxsw_sp_cpu_port_sb_cms)
+
+static int mlxsw_sp_sb_cms_init(struct mlxsw_sp *mlxsw_sp, u8 local_port,
+				const struct mlxsw_sp_sb_cm *cms,
+				size_t cms_len)
+{
+	char sbcm_pl[MLXSW_REG_SBCM_LEN];
+	int i;
+	int err;
+
+	for (i = 0; i < cms_len; i++) {
+		const struct mlxsw_sp_sb_cm *cm;
+
+		cm = &cms[i];
+		mlxsw_reg_sbcm_pack(sbcm_pl, local_port, cm->u.pg, cm->dir,
+				    cm->min_buff, cm->max_buff, cm->pool);
+		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sbcm), sbcm_pl);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+static int mlxsw_sp_port_sb_cms_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	return mlxsw_sp_sb_cms_init(mlxsw_sp_port->mlxsw_sp,
+				    mlxsw_sp_port->local_port, mlxsw_sp_sb_cms,
+				    MLXSW_SP_SB_CMS_LEN);
+}
+
+static int mlxsw_sp_cpu_port_sb_cms_init(struct mlxsw_sp *mlxsw_sp)
+{
+	return mlxsw_sp_sb_cms_init(mlxsw_sp, 0, mlxsw_sp_cpu_port_sb_cms,
+				    MLXSW_SP_CPU_PORT_SB_MCS_LEN);
+}
+
+struct mlxsw_sp_sb_pm {
+	u8 pool;
+	enum mlxsw_reg_sbpm_dir dir;
+	u32 min_buff;
+	u32 max_buff;
+};
+
+#define MLXSW_SP_SB_PM(_pool, _dir, _min_buff, _max_buff)	\
+	{							\
+		.pool = _pool,					\
+		.dir = _dir,					\
+		.min_buff = _min_buff,				\
+		.max_buff = _max_buff,				\
+	}
+
+#define MLXSW_SP_SB_PM_INGRESS(_pool, _min_buff, _max_buff)	\
+	MLXSW_SP_SB_PM(_pool, MLXSW_REG_SBPM_DIR_INGRESS,	\
+		       _min_buff, _max_buff)
+
+#define MLXSW_SP_SB_PM_EGRESS(_pool, _min_buff, _max_buff)	\
+	MLXSW_SP_SB_PM(_pool, MLXSW_REG_SBPM_DIR_EGRESS,	\
+		       _min_buff, _max_buff)
+
+static const struct mlxsw_sp_sb_pm mlxsw_sp_sb_pms[] = {
+	MLXSW_SP_SB_PM_INGRESS(0, 0, 0xff),
+	MLXSW_SP_SB_PM_INGRESS(1, 0, 0),
+	MLXSW_SP_SB_PM_INGRESS(2, 0, 0),
+	MLXSW_SP_SB_PM_INGRESS(3, 0, 0),
+	MLXSW_SP_SB_PM_EGRESS(0, 0, 7),
+	MLXSW_SP_SB_PM_EGRESS(1, 0, 0),
+	MLXSW_SP_SB_PM_EGRESS(2, 0, 0),
+	MLXSW_SP_SB_PM_EGRESS(3, 0, 0),
+};
+
+#define MLXSW_SP_SB_PMS_LEN ARRAY_SIZE(mlxsw_sp_sb_pms)
+
+static int mlxsw_sp_port_sb_pms_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	char sbpm_pl[MLXSW_REG_SBPM_LEN];
+	int i;
+	int err;
+
+	for (i = 0; i < MLXSW_SP_SB_PMS_LEN; i++) {
+		const struct mlxsw_sp_sb_pm *pm;
+
+		pm = &mlxsw_sp_sb_pms[i];
+		mlxsw_reg_sbpm_pack(sbpm_pl, mlxsw_sp_port->local_port,
+				    pm->pool, pm->dir,
+				    pm->min_buff, pm->max_buff);
+		err = mlxsw_reg_write(mlxsw_sp_port->mlxsw_sp->core,
+				      MLXSW_REG(sbpm), sbpm_pl);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+struct mlxsw_sp_sb_mm {
+	u8 prio;
+	u32 min_buff;
+	u32 max_buff;
+	u8 pool;
+};
+
+#define MLXSW_SP_SB_MM(_prio, _min_buff, _max_buff, _pool)	\
+	{							\
+		.prio = _prio,					\
+		.min_buff = _min_buff,				\
+		.max_buff = _max_buff,				\
+		.pool = _pool,					\
+	}
+
+static const struct mlxsw_sp_sb_mm mlxsw_sp_sb_mms[] = {
+	MLXSW_SP_SB_MM(0, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(1, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(2, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(3, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(4, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(5, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(6, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(7, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(8, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(9, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(10, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(11, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(12, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(13, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+	MLXSW_SP_SB_MM(14, 20000 / MLXSW_SP_SB_BYTES_PER_CELL, 0xff, 0),
+};
+
+#define MLXSW_SP_SB_MMS_LEN ARRAY_SIZE(mlxsw_sp_sb_mms)
+
+static int mlxsw_sp_sb_mms_init(struct mlxsw_sp *mlxsw_sp)
+{
+	char sbmm_pl[MLXSW_REG_SBMM_LEN];
+	int i;
+	int err;
+
+	for (i = 0; i < MLXSW_SP_SB_MMS_LEN; i++) {
+		const struct mlxsw_sp_sb_mm *mc;
+
+		mc = &mlxsw_sp_sb_mms[i];
+		mlxsw_reg_sbmm_pack(sbmm_pl, mc->prio, mc->min_buff,
+				    mc->max_buff, mc->pool);
+		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sbmm), sbmm_pl);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+int mlxsw_sp_buffers_init(struct mlxsw_sp *mlxsw_sp)
+{
+	int err;
+
+	err = mlxsw_sp_sb_pools_init(mlxsw_sp);
+	if (err)
+		return err;
+	err = mlxsw_sp_cpu_port_sb_cms_init(mlxsw_sp);
+	if (err)
+		return err;
+	err = mlxsw_sp_sb_mms_init(mlxsw_sp);
+
+	return err;
+}
+
+int mlxsw_sp_port_buffers_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	int err;
+
+	err = mlxsw_sp_port_pb_init(mlxsw_sp_port);
+	if (err)
+		return err;
+	err = mlxsw_sp_port_sb_cms_init(mlxsw_sp_port);
+	if (err)
+		return err;
+	err = mlxsw_sp_port_sb_pms_init(mlxsw_sp_port);
+
+	return err;
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
new file mode 100644
index 000000000..7dbeafa65
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -0,0 +1,906 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/skbuff.h>
+#include <linux/if_vlan.h>
+#include <linux/if_bridge.h>
+#include <linux/workqueue.h>
+#include <linux/jiffies.h>
+#include <linux/rtnetlink.h>
+#include <net/switchdev.h>
+
+#include "spectrum.h"
+#include "core.h"
+#include "reg.h"
+
+static int mlxsw_sp_port_attr_get(struct net_device *dev,
+				  struct switchdev_attr *attr)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+
+	switch (attr->id) {
+	case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
+		attr->u.ppid.id_len = sizeof(mlxsw_sp->base_mac);
+		memcpy(&attr->u.ppid.id, &mlxsw_sp->base_mac,
+		       attr->u.ppid.id_len);
+		break;
+	case SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS:
+		attr->u.brport_flags =
+			(mlxsw_sp_port->learning ? BR_LEARNING : 0) |
+			(mlxsw_sp_port->learning_sync ? BR_LEARNING_SYNC : 0) |
+			(mlxsw_sp_port->uc_flood ? BR_FLOOD : 0);
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+static int mlxsw_sp_port_stp_state_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				       u8 state)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	enum mlxsw_reg_spms_state spms_state;
+	char *spms_pl;
+	u16 vid;
+	int err;
+
+	switch (state) {
+	case BR_STATE_DISABLED: /* fall-through */
+	case BR_STATE_FORWARDING:
+		spms_state = MLXSW_REG_SPMS_STATE_FORWARDING;
+		break;
+	case BR_STATE_LISTENING: /* fall-through */
+	case BR_STATE_LEARNING:
+		spms_state = MLXSW_REG_SPMS_STATE_LEARNING;
+		break;
+	case BR_STATE_BLOCKING:
+		spms_state = MLXSW_REG_SPMS_STATE_DISCARDING;
+		break;
+	default:
+		BUG();
+	}
+
+	spms_pl = kmalloc(MLXSW_REG_SPMS_LEN, GFP_KERNEL);
+	if (!spms_pl)
+		return -ENOMEM;
+	mlxsw_reg_spms_pack(spms_pl, mlxsw_sp_port->local_port);
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, VLAN_N_VID)
+		mlxsw_reg_spms_vid_pack(spms_pl, vid, spms_state);
+
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spms), spms_pl);
+	kfree(spms_pl);
+	return err;
+}
+
+static int mlxsw_sp_port_attr_stp_state_set(struct mlxsw_sp_port *mlxsw_sp_port,
+					    struct switchdev_trans *trans,
+					    u8 state)
+{
+	if (switchdev_trans_ph_prepare(trans))
+		return 0;
+
+	mlxsw_sp_port->stp_state = state;
+	return mlxsw_sp_port_stp_state_set(mlxsw_sp_port, state);
+}
+
+static int __mlxsw_sp_port_flood_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				     u16 fid_begin, u16 fid_end, bool set,
+				     bool only_uc)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	u16 range = fid_end - fid_begin + 1;
+	char *sftr_pl;
+	int err;
+
+	sftr_pl = kmalloc(MLXSW_REG_SFTR_LEN, GFP_KERNEL);
+	if (!sftr_pl)
+		return -ENOMEM;
+
+	mlxsw_reg_sftr_pack(sftr_pl, MLXSW_SP_FLOOD_TABLE_UC, fid_begin,
+			    MLXSW_REG_SFGC_TABLE_TYPE_FID_OFFEST, range,
+			    mlxsw_sp_port->local_port, set);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sftr), sftr_pl);
+	if (err)
+		goto buffer_out;
+
+	/* Flooding control allows one to decide whether a given port will
+	 * flood unicast traffic for which there is no FDB entry.
+	 */
+	if (only_uc)
+		goto buffer_out;
+
+	mlxsw_reg_sftr_pack(sftr_pl, MLXSW_SP_FLOOD_TABLE_BM, fid_begin,
+			    MLXSW_REG_SFGC_TABLE_TYPE_FID_OFFEST, range,
+			    mlxsw_sp_port->local_port, set);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sftr), sftr_pl);
+
+buffer_out:
+	kfree(sftr_pl);
+	return err;
+}
+
+static int mlxsw_sp_port_uc_flood_set(struct mlxsw_sp_port *mlxsw_sp_port,
+				      bool set)
+{
+	struct net_device *dev = mlxsw_sp_port->dev;
+	u16 vid, last_visited_vid;
+	int err;
+
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, VLAN_N_VID) {
+		err = __mlxsw_sp_port_flood_set(mlxsw_sp_port, vid, vid, set,
+						true);
+		if (err) {
+			last_visited_vid = vid;
+			goto err_port_flood_set;
+		}
+	}
+
+	return 0;
+
+err_port_flood_set:
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, last_visited_vid)
+		__mlxsw_sp_port_flood_set(mlxsw_sp_port, vid, vid, !set, true);
+	netdev_err(dev, "Failed to configure unicast flooding\n");
+	return err;
+}
+
+static int mlxsw_sp_port_attr_br_flags_set(struct mlxsw_sp_port *mlxsw_sp_port,
+					   struct switchdev_trans *trans,
+					   unsigned long brport_flags)
+{
+	unsigned long uc_flood = mlxsw_sp_port->uc_flood ? BR_FLOOD : 0;
+	bool set;
+	int err;
+
+	if (switchdev_trans_ph_prepare(trans))
+		return 0;
+
+	if ((uc_flood ^ brport_flags) & BR_FLOOD) {
+		set = mlxsw_sp_port->uc_flood ? false : true;
+		err = mlxsw_sp_port_uc_flood_set(mlxsw_sp_port, set);
+		if (err)
+			return err;
+	}
+
+	mlxsw_sp_port->uc_flood = brport_flags & BR_FLOOD ? 1 : 0;
+	mlxsw_sp_port->learning = brport_flags & BR_LEARNING ? 1 : 0;
+	mlxsw_sp_port->learning_sync = brport_flags & BR_LEARNING_SYNC ? 1 : 0;
+
+	return 0;
+}
+
+static int mlxsw_sp_ageing_set(struct mlxsw_sp *mlxsw_sp, u32 ageing_time)
+{
+	char sfdat_pl[MLXSW_REG_SFDAT_LEN];
+	int err;
+
+	mlxsw_reg_sfdat_pack(sfdat_pl, ageing_time);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sfdat), sfdat_pl);
+	if (err)
+		return err;
+	mlxsw_sp->ageing_time = ageing_time;
+	return 0;
+}
+
+static int mlxsw_sp_port_attr_br_ageing_set(struct mlxsw_sp_port *mlxsw_sp_port,
+					    struct switchdev_trans *trans,
+					    unsigned long ageing_clock_t)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	unsigned long ageing_jiffies = clock_t_to_jiffies(ageing_clock_t);
+	u32 ageing_time = jiffies_to_msecs(ageing_jiffies) / 1000;
+
+	if (switchdev_trans_ph_prepare(trans))
+		return 0;
+
+	return mlxsw_sp_ageing_set(mlxsw_sp, ageing_time);
+}
+
+static int mlxsw_sp_port_attr_set(struct net_device *dev,
+				  const struct switchdev_attr *attr,
+				  struct switchdev_trans *trans)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err = 0;
+
+	switch (attr->id) {
+	case SWITCHDEV_ATTR_ID_PORT_STP_STATE:
+		err = mlxsw_sp_port_attr_stp_state_set(mlxsw_sp_port, trans,
+						       attr->u.stp_state);
+		break;
+	case SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS:
+		err = mlxsw_sp_port_attr_br_flags_set(mlxsw_sp_port, trans,
+						      attr->u.brport_flags);
+		break;
+	case SWITCHDEV_ATTR_ID_BRIDGE_AGEING_TIME:
+		err = mlxsw_sp_port_attr_br_ageing_set(mlxsw_sp_port, trans,
+						       attr->u.ageing_time);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+static int mlxsw_sp_port_pvid_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char spvid_pl[MLXSW_REG_SPVID_LEN];
+
+	mlxsw_reg_spvid_pack(spvid_pl, mlxsw_sp_port->local_port, vid);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(spvid), spvid_pl);
+}
+
+static int mlxsw_sp_fid_create(struct mlxsw_sp *mlxsw_sp, u16 fid)
+{
+	char sfmr_pl[MLXSW_REG_SFMR_LEN];
+	int err;
+
+	mlxsw_reg_sfmr_pack(sfmr_pl, MLXSW_REG_SFMR_OP_CREATE_FID, fid, fid);
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sfmr), sfmr_pl);
+
+	if (err)
+		return err;
+
+	set_bit(fid, mlxsw_sp->active_fids);
+	return 0;
+}
+
+static void mlxsw_sp_fid_destroy(struct mlxsw_sp *mlxsw_sp, u16 fid)
+{
+	char sfmr_pl[MLXSW_REG_SFMR_LEN];
+
+	clear_bit(fid, mlxsw_sp->active_fids);
+
+	mlxsw_reg_sfmr_pack(sfmr_pl, MLXSW_REG_SFMR_OP_DESTROY_FID,
+			    fid, fid);
+	mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sfmr), sfmr_pl);
+}
+
+static int mlxsw_sp_port_fid_map(struct mlxsw_sp_port *mlxsw_sp_port, u16 fid)
+{
+	enum mlxsw_reg_svfa_mt mt;
+
+	if (mlxsw_sp_port->nr_vfids)
+		mt = MLXSW_REG_SVFA_MT_PORT_VID_TO_FID;
+	else
+		mt = MLXSW_REG_SVFA_MT_VID_TO_FID;
+
+	return mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port, mt, true, fid, fid);
+}
+
+static int mlxsw_sp_port_fid_unmap(struct mlxsw_sp_port *mlxsw_sp_port, u16 fid)
+{
+	enum mlxsw_reg_svfa_mt mt;
+
+	if (!mlxsw_sp_port->nr_vfids)
+		return 0;
+
+	mt = MLXSW_REG_SVFA_MT_PORT_VID_TO_FID;
+	return mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port, mt, false, fid, fid);
+}
+
+static int mlxsw_sp_port_add_vids(struct net_device *dev, u16 vid_begin,
+				  u16 vid_end)
+{
+	u16 vid;
+	int err;
+
+	for (vid = vid_begin; vid <= vid_end; vid++) {
+		err = mlxsw_sp_port_add_vid(dev, 0, vid);
+		if (err)
+			goto err_port_add_vid;
+	}
+	return 0;
+
+err_port_add_vid:
+	for (vid--; vid >= vid_begin; vid--)
+		mlxsw_sp_port_kill_vid(dev, 0, vid);
+	return err;
+}
+
+static int __mlxsw_sp_port_vlans_add(struct mlxsw_sp_port *mlxsw_sp_port,
+				     u16 vid_begin, u16 vid_end,
+				     bool flag_untagged, bool flag_pvid)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	struct net_device *dev = mlxsw_sp_port->dev;
+	enum mlxsw_reg_svfa_mt mt;
+	u16 vid, vid_e;
+	int err;
+
+	/* In case this is invoked with BRIDGE_FLAGS_SELF and port is
+	 * not bridged, then packets ingressing through the port with
+	 * the specified VIDs will be directed to CPU.
+	 */
+	if (!mlxsw_sp_port->bridged)
+		return mlxsw_sp_port_add_vids(dev, vid_begin, vid_end);
+
+	for (vid = vid_begin; vid <= vid_end; vid++) {
+		if (!test_bit(vid, mlxsw_sp->active_fids)) {
+			err = mlxsw_sp_fid_create(mlxsw_sp, vid);
+			if (err) {
+				netdev_err(dev, "Failed to create FID=%d\n",
+					   vid);
+				return err;
+			}
+
+			/* When creating a FID, we set a VID to FID mapping
+			 * regardless of the port's mode.
+			 */
+			mt = MLXSW_REG_SVFA_MT_VID_TO_FID;
+			err = mlxsw_sp_port_vid_to_fid_set(mlxsw_sp_port, mt,
+							   true, vid, vid);
+			if (err) {
+				netdev_err(dev, "Failed to create FID=VID=%d mapping\n",
+					   vid);
+				return err;
+			}
+		}
+
+		/* Set FID mapping according to port's mode */
+		err = mlxsw_sp_port_fid_map(mlxsw_sp_port, vid);
+		if (err) {
+			netdev_err(dev, "Failed to map FID=%d", vid);
+			return err;
+		}
+	}
+
+	err = __mlxsw_sp_port_flood_set(mlxsw_sp_port, vid_begin, vid_end,
+					true, false);
+	if (err) {
+		netdev_err(dev, "Failed to configure flooding\n");
+		return err;
+	}
+
+	for (vid = vid_begin; vid <= vid_end;
+	     vid += MLXSW_REG_SPVM_REC_MAX_COUNT) {
+		vid_e = min((u16) (vid + MLXSW_REG_SPVM_REC_MAX_COUNT - 1),
+			    vid_end);
+
+		err = mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid_e, true,
+					     flag_untagged);
+		if (err) {
+			netdev_err(mlxsw_sp_port->dev, "Unable to add VIDs %d-%d\n",
+				   vid, vid_e);
+			return err;
+		}
+	}
+
+	vid = vid_begin;
+	if (flag_pvid && mlxsw_sp_port->pvid != vid) {
+		err = mlxsw_sp_port_pvid_set(mlxsw_sp_port, vid);
+		if (err) {
+			netdev_err(mlxsw_sp_port->dev, "Unable to add PVID %d\n",
+				   vid);
+			return err;
+		}
+		mlxsw_sp_port->pvid = vid;
+	}
+
+	/* Changing activity bits only if HW operation succeded */
+	for (vid = vid_begin; vid <= vid_end; vid++)
+		set_bit(vid, mlxsw_sp_port->active_vlans);
+
+	return mlxsw_sp_port_stp_state_set(mlxsw_sp_port,
+					   mlxsw_sp_port->stp_state);
+}
+
+static int mlxsw_sp_port_vlans_add(struct mlxsw_sp_port *mlxsw_sp_port,
+				   const struct switchdev_obj_port_vlan *vlan,
+				   struct switchdev_trans *trans)
+{
+	bool untagged_flag = vlan->flags & BRIDGE_VLAN_INFO_UNTAGGED;
+	bool pvid_flag = vlan->flags & BRIDGE_VLAN_INFO_PVID;
+
+	if (switchdev_trans_ph_prepare(trans))
+		return 0;
+
+	return __mlxsw_sp_port_vlans_add(mlxsw_sp_port,
+					 vlan->vid_begin, vlan->vid_end,
+					 untagged_flag, pvid_flag);
+}
+
+static int mlxsw_sp_port_fdb_op(struct mlxsw_sp_port *mlxsw_sp_port,
+				const char *mac, u16 vid, bool adding,
+				bool dynamic)
+{
+	enum mlxsw_reg_sfd_rec_policy policy;
+	enum mlxsw_reg_sfd_op op;
+	char *sfd_pl;
+	int err;
+
+	if (!vid)
+		vid = mlxsw_sp_port->pvid;
+
+	sfd_pl = kmalloc(MLXSW_REG_SFD_LEN, GFP_KERNEL);
+	if (!sfd_pl)
+		return -ENOMEM;
+
+	policy = dynamic ? MLXSW_REG_SFD_REC_POLICY_DYNAMIC_ENTRY_INGRESS :
+			   MLXSW_REG_SFD_REC_POLICY_STATIC_ENTRY;
+	op = adding ? MLXSW_REG_SFD_OP_WRITE_EDIT :
+		      MLXSW_REG_SFD_OP_WRITE_REMOVE;
+	mlxsw_reg_sfd_pack(sfd_pl, op, 0);
+	mlxsw_reg_sfd_uc_pack(sfd_pl, 0, policy,
+			      mac, vid, MLXSW_REG_SFD_REC_ACTION_NOP,
+			      mlxsw_sp_port->local_port);
+	err = mlxsw_reg_write(mlxsw_sp_port->mlxsw_sp->core, MLXSW_REG(sfd),
+			      sfd_pl);
+	kfree(sfd_pl);
+
+	return err;
+}
+
+static int
+mlxsw_sp_port_fdb_static_add(struct mlxsw_sp_port *mlxsw_sp_port,
+			     const struct switchdev_obj_port_fdb *fdb,
+			     struct switchdev_trans *trans)
+{
+	if (switchdev_trans_ph_prepare(trans))
+		return 0;
+
+	return mlxsw_sp_port_fdb_op(mlxsw_sp_port, fdb->addr, fdb->vid,
+				    true, false);
+}
+
+static int mlxsw_sp_port_obj_add(struct net_device *dev,
+				 const struct switchdev_obj *obj,
+				 struct switchdev_trans *trans)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err = 0;
+
+	switch (obj->id) {
+	case SWITCHDEV_OBJ_ID_PORT_VLAN:
+		err = mlxsw_sp_port_vlans_add(mlxsw_sp_port,
+					      SWITCHDEV_OBJ_PORT_VLAN(obj),
+					      trans);
+		break;
+	case SWITCHDEV_OBJ_ID_PORT_FDB:
+		err = mlxsw_sp_port_fdb_static_add(mlxsw_sp_port,
+						   SWITCHDEV_OBJ_PORT_FDB(obj),
+						   trans);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+static int mlxsw_sp_port_kill_vids(struct net_device *dev, u16 vid_begin,
+				   u16 vid_end)
+{
+	u16 vid;
+	int err;
+
+	for (vid = vid_begin; vid <= vid_end; vid++) {
+		err = mlxsw_sp_port_kill_vid(dev, 0, vid);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __mlxsw_sp_port_vlans_del(struct mlxsw_sp_port *mlxsw_sp_port,
+				     u16 vid_begin, u16 vid_end, bool init)
+{
+	struct net_device *dev = mlxsw_sp_port->dev;
+	u16 vid, vid_e;
+	int err;
+
+	/* In case this is invoked with BRIDGE_FLAGS_SELF and port is
+	 * not bridged, then prevent packets ingressing through the
+	 * port with the specified VIDs from being trapped to CPU.
+	 */
+	if (!init && !mlxsw_sp_port->bridged)
+		return mlxsw_sp_port_kill_vids(dev, vid_begin, vid_end);
+
+	for (vid = vid_begin; vid <= vid_end;
+	     vid += MLXSW_REG_SPVM_REC_MAX_COUNT) {
+		vid_e = min((u16) (vid + MLXSW_REG_SPVM_REC_MAX_COUNT - 1),
+			    vid_end);
+		err = mlxsw_sp_port_vlan_set(mlxsw_sp_port, vid, vid_e, false,
+					     false);
+		if (err) {
+			netdev_err(mlxsw_sp_port->dev, "Unable to del VIDs %d-%d\n",
+				   vid, vid_e);
+			return err;
+		}
+	}
+
+	if ((mlxsw_sp_port->pvid >= vid_begin) &&
+	    (mlxsw_sp_port->pvid <= vid_end)) {
+		/* Default VLAN is always 1 */
+		mlxsw_sp_port->pvid = 1;
+		err = mlxsw_sp_port_pvid_set(mlxsw_sp_port,
+					     mlxsw_sp_port->pvid);
+		if (err) {
+			netdev_err(mlxsw_sp_port->dev, "Unable to del PVID %d\n",
+				   vid);
+			return err;
+		}
+	}
+
+	if (init)
+		goto out;
+
+	err = __mlxsw_sp_port_flood_set(mlxsw_sp_port, vid_begin, vid_end,
+					false, false);
+	if (err) {
+		netdev_err(dev, "Failed to clear flooding\n");
+		return err;
+	}
+
+	for (vid = vid_begin; vid <= vid_end; vid++) {
+		/* Remove FID mapping in case of Virtual mode */
+		err = mlxsw_sp_port_fid_unmap(mlxsw_sp_port, vid);
+		if (err) {
+			netdev_err(dev, "Failed to unmap FID=%d", vid);
+			return err;
+		}
+	}
+
+out:
+	/* Changing activity bits only if HW operation succeded */
+	for (vid = vid_begin; vid <= vid_end; vid++)
+		clear_bit(vid, mlxsw_sp_port->active_vlans);
+
+	return 0;
+}
+
+static int mlxsw_sp_port_vlans_del(struct mlxsw_sp_port *mlxsw_sp_port,
+				   const struct switchdev_obj_port_vlan *vlan)
+{
+	return __mlxsw_sp_port_vlans_del(mlxsw_sp_port,
+					 vlan->vid_begin, vlan->vid_end, false);
+}
+
+static int
+mlxsw_sp_port_fdb_static_del(struct mlxsw_sp_port *mlxsw_sp_port,
+			     const struct switchdev_obj_port_fdb *fdb)
+{
+	return mlxsw_sp_port_fdb_op(mlxsw_sp_port, fdb->addr, fdb->vid,
+				    false, false);
+}
+
+static int mlxsw_sp_port_obj_del(struct net_device *dev,
+				 const struct switchdev_obj *obj)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err = 0;
+
+	switch (obj->id) {
+	case SWITCHDEV_OBJ_ID_PORT_VLAN:
+		err = mlxsw_sp_port_vlans_del(mlxsw_sp_port,
+					      SWITCHDEV_OBJ_PORT_VLAN(obj));
+		break;
+	case SWITCHDEV_OBJ_ID_PORT_FDB:
+		err = mlxsw_sp_port_fdb_static_del(mlxsw_sp_port,
+						   SWITCHDEV_OBJ_PORT_FDB(obj));
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+static int mlxsw_sp_port_fdb_dump(struct mlxsw_sp_port *mlxsw_sp_port,
+				  struct switchdev_obj_port_fdb *fdb,
+				  switchdev_obj_dump_cb_t *cb)
+{
+	char *sfd_pl;
+	char mac[ETH_ALEN];
+	u16 vid;
+	u8 local_port;
+	u8 num_rec;
+	int stored_err = 0;
+	int i;
+	int err;
+
+	sfd_pl = kmalloc(MLXSW_REG_SFD_LEN, GFP_KERNEL);
+	if (!sfd_pl)
+		return -ENOMEM;
+
+	mlxsw_reg_sfd_pack(sfd_pl, MLXSW_REG_SFD_OP_QUERY_DUMP, 0);
+	do {
+		mlxsw_reg_sfd_num_rec_set(sfd_pl, MLXSW_REG_SFD_REC_MAX_COUNT);
+		err = mlxsw_reg_query(mlxsw_sp_port->mlxsw_sp->core,
+				      MLXSW_REG(sfd), sfd_pl);
+		if (err)
+			goto out;
+
+		num_rec = mlxsw_reg_sfd_num_rec_get(sfd_pl);
+
+		/* Even in case of error, we have to run the dump to the end
+		 * so the session in firmware is finished.
+		 */
+		if (stored_err)
+			continue;
+
+		for (i = 0; i < num_rec; i++) {
+			switch (mlxsw_reg_sfd_rec_type_get(sfd_pl, i)) {
+			case MLXSW_REG_SFD_REC_TYPE_UNICAST:
+				mlxsw_reg_sfd_uc_unpack(sfd_pl, i, mac, &vid,
+							&local_port);
+				if (local_port == mlxsw_sp_port->local_port) {
+					ether_addr_copy(fdb->addr, mac);
+					fdb->ndm_state = NUD_REACHABLE;
+					fdb->vid = vid;
+					err = cb(&fdb->obj);
+					if (err)
+						stored_err = err;
+				}
+			}
+		}
+	} while (num_rec == MLXSW_REG_SFD_REC_MAX_COUNT);
+
+out:
+	kfree(sfd_pl);
+	return stored_err ? stored_err : err;
+}
+
+static int mlxsw_sp_port_vlan_dump(struct mlxsw_sp_port *mlxsw_sp_port,
+				   struct switchdev_obj_port_vlan *vlan,
+				   switchdev_obj_dump_cb_t *cb)
+{
+	u16 vid;
+	int err = 0;
+
+	for_each_set_bit(vid, mlxsw_sp_port->active_vlans, VLAN_N_VID) {
+		vlan->flags = 0;
+		if (vid == mlxsw_sp_port->pvid)
+			vlan->flags |= BRIDGE_VLAN_INFO_PVID;
+		vlan->vid_begin = vid;
+		vlan->vid_end = vid;
+		err = cb(&vlan->obj);
+		if (err)
+			break;
+	}
+	return err;
+}
+
+static int mlxsw_sp_port_obj_dump(struct net_device *dev,
+				  struct switchdev_obj *obj,
+				  switchdev_obj_dump_cb_t *cb)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port = netdev_priv(dev);
+	int err = 0;
+
+	switch (obj->id) {
+	case SWITCHDEV_OBJ_ID_PORT_VLAN:
+		err = mlxsw_sp_port_vlan_dump(mlxsw_sp_port,
+					      SWITCHDEV_OBJ_PORT_VLAN(obj), cb);
+		break;
+	case SWITCHDEV_OBJ_ID_PORT_FDB:
+		err = mlxsw_sp_port_fdb_dump(mlxsw_sp_port,
+					     SWITCHDEV_OBJ_PORT_FDB(obj), cb);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	return err;
+}
+
+static const struct switchdev_ops mlxsw_sp_port_switchdev_ops = {
+	.switchdev_port_attr_get	= mlxsw_sp_port_attr_get,
+	.switchdev_port_attr_set	= mlxsw_sp_port_attr_set,
+	.switchdev_port_obj_add		= mlxsw_sp_port_obj_add,
+	.switchdev_port_obj_del		= mlxsw_sp_port_obj_del,
+	.switchdev_port_obj_dump	= mlxsw_sp_port_obj_dump,
+};
+
+static void mlxsw_sp_fdb_notify_mac_process(struct mlxsw_sp *mlxsw_sp,
+					    char *sfn_pl, int rec_index,
+					    bool adding)
+{
+	struct mlxsw_sp_port *mlxsw_sp_port;
+	char mac[ETH_ALEN];
+	u8 local_port;
+	u16 vid;
+	int err;
+
+	mlxsw_reg_sfn_mac_unpack(sfn_pl, rec_index, mac, &vid, &local_port);
+	mlxsw_sp_port = mlxsw_sp->ports[local_port];
+	if (!mlxsw_sp_port) {
+		dev_err_ratelimited(mlxsw_sp->bus_info->dev, "Incorrect local port in FDB notification\n");
+		return;
+	}
+
+	err = mlxsw_sp_port_fdb_op(mlxsw_sp_port, mac, vid,
+				   adding && mlxsw_sp_port->learning, true);
+	if (err) {
+		if (net_ratelimit())
+			netdev_err(mlxsw_sp_port->dev, "Failed to set FDB entry\n");
+		return;
+	}
+
+	if (mlxsw_sp_port->learning && mlxsw_sp_port->learning_sync) {
+		struct switchdev_notifier_fdb_info info;
+		unsigned long notifier_type;
+
+		info.addr = mac;
+		info.vid = vid;
+		notifier_type = adding ? SWITCHDEV_FDB_ADD : SWITCHDEV_FDB_DEL;
+		call_switchdev_notifiers(notifier_type, mlxsw_sp_port->dev,
+					 &info.info);
+	}
+}
+
+static void mlxsw_sp_fdb_notify_rec_process(struct mlxsw_sp *mlxsw_sp,
+					    char *sfn_pl, int rec_index)
+{
+	switch (mlxsw_reg_sfn_rec_type_get(sfn_pl, rec_index)) {
+	case MLXSW_REG_SFN_REC_TYPE_LEARNED_MAC:
+		mlxsw_sp_fdb_notify_mac_process(mlxsw_sp, sfn_pl,
+						rec_index, true);
+		break;
+	case MLXSW_REG_SFN_REC_TYPE_AGED_OUT_MAC:
+		mlxsw_sp_fdb_notify_mac_process(mlxsw_sp, sfn_pl,
+						rec_index, false);
+		break;
+	}
+}
+
+static void mlxsw_sp_fdb_notify_work_schedule(struct mlxsw_sp *mlxsw_sp)
+{
+	schedule_delayed_work(&mlxsw_sp->fdb_notify.dw,
+			      msecs_to_jiffies(mlxsw_sp->fdb_notify.interval));
+}
+
+static void mlxsw_sp_fdb_notify_work(struct work_struct *work)
+{
+	struct mlxsw_sp *mlxsw_sp;
+	char *sfn_pl;
+	u8 num_rec;
+	int i;
+	int err;
+
+	sfn_pl = kmalloc(MLXSW_REG_SFN_LEN, GFP_KERNEL);
+	if (!sfn_pl)
+		return;
+
+	mlxsw_sp = container_of(work, struct mlxsw_sp, fdb_notify.dw.work);
+
+	rtnl_lock();
+	do {
+		mlxsw_reg_sfn_pack(sfn_pl);
+		err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(sfn), sfn_pl);
+		if (err) {
+			dev_err_ratelimited(mlxsw_sp->bus_info->dev, "Failed to get FDB notifications\n");
+			break;
+		}
+		num_rec = mlxsw_reg_sfn_num_rec_get(sfn_pl);
+		for (i = 0; i < num_rec; i++)
+			mlxsw_sp_fdb_notify_rec_process(mlxsw_sp, sfn_pl, i);
+
+	} while (num_rec);
+	rtnl_unlock();
+
+	kfree(sfn_pl);
+	mlxsw_sp_fdb_notify_work_schedule(mlxsw_sp);
+}
+
+static int mlxsw_sp_fdb_init(struct mlxsw_sp *mlxsw_sp)
+{
+	int err;
+
+	err = mlxsw_sp_ageing_set(mlxsw_sp, MLXSW_SP_DEFAULT_AGEING_TIME);
+	if (err) {
+		dev_err(mlxsw_sp->bus_info->dev, "Failed to set default ageing time\n");
+		return err;
+	}
+	INIT_DELAYED_WORK(&mlxsw_sp->fdb_notify.dw, mlxsw_sp_fdb_notify_work);
+	mlxsw_sp->fdb_notify.interval = MLXSW_SP_DEFAULT_LEARNING_INTERVAL;
+	mlxsw_sp_fdb_notify_work_schedule(mlxsw_sp);
+	return 0;
+}
+
+static void mlxsw_sp_fdb_fini(struct mlxsw_sp *mlxsw_sp)
+{
+	cancel_delayed_work_sync(&mlxsw_sp->fdb_notify.dw);
+}
+
+static void mlxsw_sp_fids_fini(struct mlxsw_sp *mlxsw_sp)
+{
+	u16 fid;
+
+	for_each_set_bit(fid, mlxsw_sp->active_fids, VLAN_N_VID)
+		mlxsw_sp_fid_destroy(mlxsw_sp, fid);
+}
+
+int mlxsw_sp_switchdev_init(struct mlxsw_sp *mlxsw_sp)
+{
+	return mlxsw_sp_fdb_init(mlxsw_sp);
+}
+
+void mlxsw_sp_switchdev_fini(struct mlxsw_sp *mlxsw_sp)
+{
+	mlxsw_sp_fdb_fini(mlxsw_sp);
+	mlxsw_sp_fids_fini(mlxsw_sp);
+}
+
+int mlxsw_sp_port_vlan_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	struct net_device *dev = mlxsw_sp_port->dev;
+	int err;
+
+	/* Allow only untagged packets to ingress and tag them internally
+	 * with VID 1.
+	 */
+	mlxsw_sp_port->pvid = 1;
+	err = __mlxsw_sp_port_vlans_del(mlxsw_sp_port, 0, VLAN_N_VID, true);
+	if (err) {
+		netdev_err(dev, "Unable to init VLANs\n");
+		return err;
+	}
+
+	/* Add implicit VLAN interface in the device, so that untagged
+	 * packets will be classified to the default vFID.
+	 */
+	err = mlxsw_sp_port_add_vid(dev, 0, 1);
+	if (err)
+		netdev_err(dev, "Failed to configure default vFID\n");
+
+	return err;
+}
+
+void mlxsw_sp_port_switchdev_init(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+	mlxsw_sp_port->dev->switchdev_ops = &mlxsw_sp_port_switchdev_ops;
+}
+
+void mlxsw_sp_port_switchdev_fini(struct mlxsw_sp_port *mlxsw_sp_port)
+{
+}
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/switchx2.c b/kernel/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
new file mode 100644
index 000000000..d85960cfb
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/switchx2.c
@@ -0,0 +1,1557 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/switchx2.c
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/skbuff.h>
+#include <linux/if_vlan.h>
+#include <net/switchdev.h>
+#include <generated/utsrelease.h>
+
+#include "core.h"
+#include "reg.h"
+#include "port.h"
+#include "trap.h"
+#include "txheader.h"
+
+static const char mlxsw_sx_driver_name[] = "mlxsw_switchx2";
+static const char mlxsw_sx_driver_version[] = "1.0";
+
+struct mlxsw_sx_port;
+
+struct mlxsw_sx {
+	struct mlxsw_sx_port **ports;
+	struct mlxsw_core *core;
+	const struct mlxsw_bus_info *bus_info;
+	u8 hw_id[ETH_ALEN];
+};
+
+struct mlxsw_sx_port_pcpu_stats {
+	u64			rx_packets;
+	u64			rx_bytes;
+	u64			tx_packets;
+	u64			tx_bytes;
+	struct u64_stats_sync	syncp;
+	u32			tx_dropped;
+};
+
+struct mlxsw_sx_port {
+	struct net_device *dev;
+	struct mlxsw_sx_port_pcpu_stats __percpu *pcpu_stats;
+	struct mlxsw_sx *mlxsw_sx;
+	u8 local_port;
+};
+
+/* tx_hdr_version
+ * Tx header version.
+ * Must be set to 0.
+ */
+MLXSW_ITEM32(tx, hdr, version, 0x00, 28, 4);
+
+/* tx_hdr_ctl
+ * Packet control type.
+ * 0 - Ethernet control (e.g. EMADs, LACP)
+ * 1 - Ethernet data
+ */
+MLXSW_ITEM32(tx, hdr, ctl, 0x00, 26, 2);
+
+/* tx_hdr_proto
+ * Packet protocol type. Must be set to 1 (Ethernet).
+ */
+MLXSW_ITEM32(tx, hdr, proto, 0x00, 21, 3);
+
+/* tx_hdr_etclass
+ * Egress TClass to be used on the egress device on the egress port.
+ * The MSB is specified in the 'ctclass3' field.
+ * Range is 0-15, where 15 is the highest priority.
+ */
+MLXSW_ITEM32(tx, hdr, etclass, 0x00, 18, 3);
+
+/* tx_hdr_swid
+ * Switch partition ID.
+ */
+MLXSW_ITEM32(tx, hdr, swid, 0x00, 12, 3);
+
+/* tx_hdr_port_mid
+ * Destination local port for unicast packets.
+ * Destination multicast ID for multicast packets.
+ *
+ * Control packets are directed to a specific egress port, while data
+ * packets are transmitted through the CPU port (0) into the switch partition,
+ * where forwarding rules are applied.
+ */
+MLXSW_ITEM32(tx, hdr, port_mid, 0x04, 16, 16);
+
+/* tx_hdr_ctclass3
+ * See field 'etclass'.
+ */
+MLXSW_ITEM32(tx, hdr, ctclass3, 0x04, 14, 1);
+
+/* tx_hdr_rdq
+ * RDQ for control packets sent to remote CPU.
+ * Must be set to 0x1F for EMADs, otherwise 0.
+ */
+MLXSW_ITEM32(tx, hdr, rdq, 0x04, 9, 5);
+
+/* tx_hdr_cpu_sig
+ * Signature control for packets going to CPU. Must be set to 0.
+ */
+MLXSW_ITEM32(tx, hdr, cpu_sig, 0x04, 0, 9);
+
+/* tx_hdr_sig
+ * Stacking protocl signature. Must be set to 0xE0E0.
+ */
+MLXSW_ITEM32(tx, hdr, sig, 0x0C, 16, 16);
+
+/* tx_hdr_stclass
+ * Stacking TClass.
+ */
+MLXSW_ITEM32(tx, hdr, stclass, 0x0C, 13, 3);
+
+/* tx_hdr_emad
+ * EMAD bit. Must be set for EMADs.
+ */
+MLXSW_ITEM32(tx, hdr, emad, 0x0C, 5, 1);
+
+/* tx_hdr_type
+ * 0 - Data packets
+ * 6 - Control packets
+ */
+MLXSW_ITEM32(tx, hdr, type, 0x0C, 0, 4);
+
+static void mlxsw_sx_txhdr_construct(struct sk_buff *skb,
+				     const struct mlxsw_tx_info *tx_info)
+{
+	char *txhdr = skb_push(skb, MLXSW_TXHDR_LEN);
+	bool is_emad = tx_info->is_emad;
+
+	memset(txhdr, 0, MLXSW_TXHDR_LEN);
+
+	/* We currently set default values for the egress tclass (QoS). */
+	mlxsw_tx_hdr_version_set(txhdr, MLXSW_TXHDR_VERSION_0);
+	mlxsw_tx_hdr_ctl_set(txhdr, MLXSW_TXHDR_ETH_CTL);
+	mlxsw_tx_hdr_proto_set(txhdr, MLXSW_TXHDR_PROTO_ETH);
+	mlxsw_tx_hdr_etclass_set(txhdr, is_emad ? MLXSW_TXHDR_ETCLASS_6 :
+						  MLXSW_TXHDR_ETCLASS_5);
+	mlxsw_tx_hdr_swid_set(txhdr, 0);
+	mlxsw_tx_hdr_port_mid_set(txhdr, tx_info->local_port);
+	mlxsw_tx_hdr_ctclass3_set(txhdr, MLXSW_TXHDR_CTCLASS3);
+	mlxsw_tx_hdr_rdq_set(txhdr, is_emad ? MLXSW_TXHDR_RDQ_EMAD :
+					      MLXSW_TXHDR_RDQ_OTHER);
+	mlxsw_tx_hdr_cpu_sig_set(txhdr, MLXSW_TXHDR_CPU_SIG);
+	mlxsw_tx_hdr_sig_set(txhdr, MLXSW_TXHDR_SIG);
+	mlxsw_tx_hdr_stclass_set(txhdr, MLXSW_TXHDR_STCLASS_NONE);
+	mlxsw_tx_hdr_emad_set(txhdr, is_emad ? MLXSW_TXHDR_EMAD :
+					       MLXSW_TXHDR_NOT_EMAD);
+	mlxsw_tx_hdr_type_set(txhdr, MLXSW_TXHDR_TYPE_CONTROL);
+}
+
+static int mlxsw_sx_port_admin_status_set(struct mlxsw_sx_port *mlxsw_sx_port,
+					  bool is_up)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char paos_pl[MLXSW_REG_PAOS_LEN];
+
+	mlxsw_reg_paos_pack(paos_pl, mlxsw_sx_port->local_port,
+			    is_up ? MLXSW_PORT_ADMIN_STATUS_UP :
+			    MLXSW_PORT_ADMIN_STATUS_DOWN);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(paos), paos_pl);
+}
+
+static int mlxsw_sx_port_oper_status_get(struct mlxsw_sx_port *mlxsw_sx_port,
+					 bool *p_is_up)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char paos_pl[MLXSW_REG_PAOS_LEN];
+	u8 oper_status;
+	int err;
+
+	mlxsw_reg_paos_pack(paos_pl, mlxsw_sx_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(paos), paos_pl);
+	if (err)
+		return err;
+	oper_status = mlxsw_reg_paos_oper_status_get(paos_pl);
+	*p_is_up = oper_status == MLXSW_PORT_ADMIN_STATUS_UP ? true : false;
+	return 0;
+}
+
+static int mlxsw_sx_port_mtu_set(struct mlxsw_sx_port *mlxsw_sx_port, u16 mtu)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char pmtu_pl[MLXSW_REG_PMTU_LEN];
+	int max_mtu;
+	int err;
+
+	mtu += MLXSW_TXHDR_LEN + ETH_HLEN;
+	mlxsw_reg_pmtu_pack(pmtu_pl, mlxsw_sx_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(pmtu), pmtu_pl);
+	if (err)
+		return err;
+	max_mtu = mlxsw_reg_pmtu_max_mtu_get(pmtu_pl);
+
+	if (mtu > max_mtu)
+		return -EINVAL;
+
+	mlxsw_reg_pmtu_pack(pmtu_pl, mlxsw_sx_port->local_port, mtu);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(pmtu), pmtu_pl);
+}
+
+static int mlxsw_sx_port_swid_set(struct mlxsw_sx_port *mlxsw_sx_port, u8 swid)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char pspa_pl[MLXSW_REG_PSPA_LEN];
+
+	mlxsw_reg_pspa_pack(pspa_pl, swid, mlxsw_sx_port->local_port);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(pspa), pspa_pl);
+}
+
+static int
+mlxsw_sx_port_system_port_mapping_set(struct mlxsw_sx_port *mlxsw_sx_port)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char sspr_pl[MLXSW_REG_SSPR_LEN];
+
+	mlxsw_reg_sspr_pack(sspr_pl, mlxsw_sx_port->local_port);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sspr), sspr_pl);
+}
+
+static int mlxsw_sx_port_module_check(struct mlxsw_sx_port *mlxsw_sx_port,
+				      bool *p_usable)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char pmlp_pl[MLXSW_REG_PMLP_LEN];
+	int err;
+
+	mlxsw_reg_pmlp_pack(pmlp_pl, mlxsw_sx_port->local_port);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(pmlp), pmlp_pl);
+	if (err)
+		return err;
+	*p_usable = mlxsw_reg_pmlp_width_get(pmlp_pl) ? true : false;
+	return 0;
+}
+
+static int mlxsw_sx_port_open(struct net_device *dev)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	int err;
+
+	err = mlxsw_sx_port_admin_status_set(mlxsw_sx_port, true);
+	if (err)
+		return err;
+	netif_start_queue(dev);
+	return 0;
+}
+
+static int mlxsw_sx_port_stop(struct net_device *dev)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+
+	netif_stop_queue(dev);
+	return mlxsw_sx_port_admin_status_set(mlxsw_sx_port, false);
+}
+
+static netdev_tx_t mlxsw_sx_port_xmit(struct sk_buff *skb,
+				      struct net_device *dev)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	struct mlxsw_sx_port_pcpu_stats *pcpu_stats;
+	const struct mlxsw_tx_info tx_info = {
+		.local_port = mlxsw_sx_port->local_port,
+		.is_emad = false,
+	};
+	u64 len;
+	int err;
+
+	if (mlxsw_core_skb_transmit_busy(mlxsw_sx, &tx_info))
+		return NETDEV_TX_BUSY;
+
+	if (unlikely(skb_headroom(skb) < MLXSW_TXHDR_LEN)) {
+		struct sk_buff *skb_orig = skb;
+
+		skb = skb_realloc_headroom(skb, MLXSW_TXHDR_LEN);
+		if (!skb) {
+			this_cpu_inc(mlxsw_sx_port->pcpu_stats->tx_dropped);
+			dev_kfree_skb_any(skb_orig);
+			return NETDEV_TX_OK;
+		}
+	}
+	mlxsw_sx_txhdr_construct(skb, &tx_info);
+	len = skb->len;
+	/* Due to a race we might fail here because of a full queue. In that
+	 * unlikely case we simply drop the packet.
+	 */
+	err = mlxsw_core_skb_transmit(mlxsw_sx, skb, &tx_info);
+
+	if (!err) {
+		pcpu_stats = this_cpu_ptr(mlxsw_sx_port->pcpu_stats);
+		u64_stats_update_begin(&pcpu_stats->syncp);
+		pcpu_stats->tx_packets++;
+		pcpu_stats->tx_bytes += len;
+		u64_stats_update_end(&pcpu_stats->syncp);
+	} else {
+		this_cpu_inc(mlxsw_sx_port->pcpu_stats->tx_dropped);
+		dev_kfree_skb_any(skb);
+	}
+	return NETDEV_TX_OK;
+}
+
+static int mlxsw_sx_port_change_mtu(struct net_device *dev, int mtu)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	int err;
+
+	err = mlxsw_sx_port_mtu_set(mlxsw_sx_port, mtu);
+	if (err)
+		return err;
+	dev->mtu = mtu;
+	return 0;
+}
+
+static struct rtnl_link_stats64 *
+mlxsw_sx_port_get_stats64(struct net_device *dev,
+			  struct rtnl_link_stats64 *stats)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx_port_pcpu_stats *p;
+	u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
+	u32 tx_dropped = 0;
+	unsigned int start;
+	int i;
+
+	for_each_possible_cpu(i) {
+		p = per_cpu_ptr(mlxsw_sx_port->pcpu_stats, i);
+		do {
+			start = u64_stats_fetch_begin_irq(&p->syncp);
+			rx_packets	= p->rx_packets;
+			rx_bytes	= p->rx_bytes;
+			tx_packets	= p->tx_packets;
+			tx_bytes	= p->tx_bytes;
+		} while (u64_stats_fetch_retry_irq(&p->syncp, start));
+
+		stats->rx_packets	+= rx_packets;
+		stats->rx_bytes		+= rx_bytes;
+		stats->tx_packets	+= tx_packets;
+		stats->tx_bytes		+= tx_bytes;
+		/* tx_dropped is u32, updated without syncp protection. */
+		tx_dropped	+= p->tx_dropped;
+	}
+	stats->tx_dropped	= tx_dropped;
+	return stats;
+}
+
+static const struct net_device_ops mlxsw_sx_port_netdev_ops = {
+	.ndo_open		= mlxsw_sx_port_open,
+	.ndo_stop		= mlxsw_sx_port_stop,
+	.ndo_start_xmit		= mlxsw_sx_port_xmit,
+	.ndo_change_mtu		= mlxsw_sx_port_change_mtu,
+	.ndo_get_stats64	= mlxsw_sx_port_get_stats64,
+};
+
+static void mlxsw_sx_port_get_drvinfo(struct net_device *dev,
+				      struct ethtool_drvinfo *drvinfo)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+
+	strlcpy(drvinfo->driver, mlxsw_sx_driver_name, sizeof(drvinfo->driver));
+	strlcpy(drvinfo->version, mlxsw_sx_driver_version,
+		sizeof(drvinfo->version));
+	snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
+		 "%d.%d.%d",
+		 mlxsw_sx->bus_info->fw_rev.major,
+		 mlxsw_sx->bus_info->fw_rev.minor,
+		 mlxsw_sx->bus_info->fw_rev.subminor);
+	strlcpy(drvinfo->bus_info, mlxsw_sx->bus_info->device_name,
+		sizeof(drvinfo->bus_info));
+}
+
+struct mlxsw_sx_port_hw_stats {
+	char str[ETH_GSTRING_LEN];
+	u64 (*getter)(char *payload);
+};
+
+static const struct mlxsw_sx_port_hw_stats mlxsw_sx_port_hw_stats[] = {
+	{
+		.str = "a_frames_transmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_frames_transmitted_ok_get,
+	},
+	{
+		.str = "a_frames_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_frames_received_ok_get,
+	},
+	{
+		.str = "a_frame_check_sequence_errors",
+		.getter = mlxsw_reg_ppcnt_a_frame_check_sequence_errors_get,
+	},
+	{
+		.str = "a_alignment_errors",
+		.getter = mlxsw_reg_ppcnt_a_alignment_errors_get,
+	},
+	{
+		.str = "a_octets_transmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_octets_transmitted_ok_get,
+	},
+	{
+		.str = "a_octets_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_octets_received_ok_get,
+	},
+	{
+		.str = "a_multicast_frames_xmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_multicast_frames_xmitted_ok_get,
+	},
+	{
+		.str = "a_broadcast_frames_xmitted_ok",
+		.getter = mlxsw_reg_ppcnt_a_broadcast_frames_xmitted_ok_get,
+	},
+	{
+		.str = "a_multicast_frames_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_multicast_frames_received_ok_get,
+	},
+	{
+		.str = "a_broadcast_frames_received_ok",
+		.getter = mlxsw_reg_ppcnt_a_broadcast_frames_received_ok_get,
+	},
+	{
+		.str = "a_in_range_length_errors",
+		.getter = mlxsw_reg_ppcnt_a_in_range_length_errors_get,
+	},
+	{
+		.str = "a_out_of_range_length_field",
+		.getter = mlxsw_reg_ppcnt_a_out_of_range_length_field_get,
+	},
+	{
+		.str = "a_frame_too_long_errors",
+		.getter = mlxsw_reg_ppcnt_a_frame_too_long_errors_get,
+	},
+	{
+		.str = "a_symbol_error_during_carrier",
+		.getter = mlxsw_reg_ppcnt_a_symbol_error_during_carrier_get,
+	},
+	{
+		.str = "a_mac_control_frames_transmitted",
+		.getter = mlxsw_reg_ppcnt_a_mac_control_frames_transmitted_get,
+	},
+	{
+		.str = "a_mac_control_frames_received",
+		.getter = mlxsw_reg_ppcnt_a_mac_control_frames_received_get,
+	},
+	{
+		.str = "a_unsupported_opcodes_received",
+		.getter = mlxsw_reg_ppcnt_a_unsupported_opcodes_received_get,
+	},
+	{
+		.str = "a_pause_mac_ctrl_frames_received",
+		.getter = mlxsw_reg_ppcnt_a_pause_mac_ctrl_frames_received_get,
+	},
+	{
+		.str = "a_pause_mac_ctrl_frames_xmitted",
+		.getter = mlxsw_reg_ppcnt_a_pause_mac_ctrl_frames_transmitted_get,
+	},
+};
+
+#define MLXSW_SX_PORT_HW_STATS_LEN ARRAY_SIZE(mlxsw_sx_port_hw_stats)
+
+static void mlxsw_sx_port_get_strings(struct net_device *dev,
+				      u32 stringset, u8 *data)
+{
+	u8 *p = data;
+	int i;
+
+	switch (stringset) {
+	case ETH_SS_STATS:
+		for (i = 0; i < MLXSW_SX_PORT_HW_STATS_LEN; i++) {
+			memcpy(p, mlxsw_sx_port_hw_stats[i].str,
+			       ETH_GSTRING_LEN);
+			p += ETH_GSTRING_LEN;
+		}
+		break;
+	}
+}
+
+static void mlxsw_sx_port_get_stats(struct net_device *dev,
+				    struct ethtool_stats *stats, u64 *data)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char ppcnt_pl[MLXSW_REG_PPCNT_LEN];
+	int i;
+	int err;
+
+	mlxsw_reg_ppcnt_pack(ppcnt_pl, mlxsw_sx_port->local_port);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(ppcnt), ppcnt_pl);
+	for (i = 0; i < MLXSW_SX_PORT_HW_STATS_LEN; i++)
+		data[i] = !err ? mlxsw_sx_port_hw_stats[i].getter(ppcnt_pl) : 0;
+}
+
+static int mlxsw_sx_port_get_sset_count(struct net_device *dev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return MLXSW_SX_PORT_HW_STATS_LEN;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+struct mlxsw_sx_port_link_mode {
+	u32 mask;
+	u32 supported;
+	u32 advertised;
+	u32 speed;
+};
+
+static const struct mlxsw_sx_port_link_mode mlxsw_sx_port_link_mode[] = {
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_100BASE_T,
+		.supported	= SUPPORTED_100baseT_Full,
+		.advertised	= ADVERTISED_100baseT_Full,
+		.speed		= 100,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_100BASE_TX,
+		.speed		= 100,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_SGMII |
+				  MLXSW_REG_PTYS_ETH_SPEED_1000BASE_KX,
+		.supported	= SUPPORTED_1000baseKX_Full,
+		.advertised	= ADVERTISED_1000baseKX_Full,
+		.speed		= 1000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_10GBASE_T,
+		.supported	= SUPPORTED_10000baseT_Full,
+		.advertised	= ADVERTISED_10000baseT_Full,
+		.speed		= 10000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CX4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4,
+		.supported	= SUPPORTED_10000baseKX4_Full,
+		.advertised	= ADVERTISED_10000baseKX4_Full,
+		.speed		= 10000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR |
+				  MLXSW_REG_PTYS_ETH_SPEED_10GBASE_ER_LR,
+		.supported	= SUPPORTED_10000baseKR_Full,
+		.advertised	= ADVERTISED_10000baseKR_Full,
+		.speed		= 10000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_20GBASE_KR2,
+		.supported	= SUPPORTED_20000baseKR2_Full,
+		.advertised	= ADVERTISED_20000baseKR2_Full,
+		.speed		= 20000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4,
+		.supported	= SUPPORTED_40000baseCR4_Full,
+		.advertised	= ADVERTISED_40000baseCR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4,
+		.supported	= SUPPORTED_40000baseKR4_Full,
+		.advertised	= ADVERTISED_40000baseKR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4,
+		.supported	= SUPPORTED_40000baseSR4_Full,
+		.advertised	= ADVERTISED_40000baseSR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_40GBASE_LR4_ER4,
+		.supported	= SUPPORTED_40000baseLR4_Full,
+		.advertised	= ADVERTISED_40000baseLR4_Full,
+		.speed		= 40000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_25GBASE_CR |
+				  MLXSW_REG_PTYS_ETH_SPEED_25GBASE_KR |
+				  MLXSW_REG_PTYS_ETH_SPEED_25GBASE_SR,
+		.speed		= 25000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_50GBASE_KR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_50GBASE_CR2 |
+				  MLXSW_REG_PTYS_ETH_SPEED_50GBASE_KR2,
+		.speed		= 50000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_56GBASE_R4,
+		.supported	= SUPPORTED_56000baseKR4_Full,
+		.advertised	= ADVERTISED_56000baseKR4_Full,
+		.speed		= 56000,
+	},
+	{
+		.mask		= MLXSW_REG_PTYS_ETH_SPEED_100GBASE_CR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4 |
+				  MLXSW_REG_PTYS_ETH_SPEED_100GBASE_LR4_ER4,
+		.speed		= 100000,
+	},
+};
+
+#define MLXSW_SX_PORT_LINK_MODE_LEN ARRAY_SIZE(mlxsw_sx_port_link_mode)
+
+static u32 mlxsw_sx_from_ptys_supported_port(u32 ptys_eth_proto)
+{
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR |
+			      MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_SGMII))
+		return SUPPORTED_FIBRE;
+
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR |
+			      MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_1000BASE_KX))
+		return SUPPORTED_Backplane;
+	return 0;
+}
+
+static u32 mlxsw_sx_from_ptys_supported_link(u32 ptys_eth_proto)
+{
+	u32 modes = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SX_PORT_LINK_MODE_LEN; i++) {
+		if (ptys_eth_proto & mlxsw_sx_port_link_mode[i].mask)
+			modes |= mlxsw_sx_port_link_mode[i].supported;
+	}
+	return modes;
+}
+
+static u32 mlxsw_sx_from_ptys_advert_link(u32 ptys_eth_proto)
+{
+	u32 modes = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SX_PORT_LINK_MODE_LEN; i++) {
+		if (ptys_eth_proto & mlxsw_sx_port_link_mode[i].mask)
+			modes |= mlxsw_sx_port_link_mode[i].advertised;
+	}
+	return modes;
+}
+
+static void mlxsw_sx_from_ptys_speed_duplex(bool carrier_ok, u32 ptys_eth_proto,
+					    struct ethtool_cmd *cmd)
+{
+	u32 speed = SPEED_UNKNOWN;
+	u8 duplex = DUPLEX_UNKNOWN;
+	int i;
+
+	if (!carrier_ok)
+		goto out;
+
+	for (i = 0; i < MLXSW_SX_PORT_LINK_MODE_LEN; i++) {
+		if (ptys_eth_proto & mlxsw_sx_port_link_mode[i].mask) {
+			speed = mlxsw_sx_port_link_mode[i].speed;
+			duplex = DUPLEX_FULL;
+			break;
+		}
+	}
+out:
+	ethtool_cmd_speed_set(cmd, speed);
+	cmd->duplex = duplex;
+}
+
+static u8 mlxsw_sx_port_connector_port(u32 ptys_eth_proto)
+{
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_SR |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_SR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_SGMII))
+		return PORT_FIBRE;
+
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_CR |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_CR4))
+		return PORT_DA;
+
+	if (ptys_eth_proto & (MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KR |
+			      MLXSW_REG_PTYS_ETH_SPEED_10GBASE_KX4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_KR4 |
+			      MLXSW_REG_PTYS_ETH_SPEED_100GBASE_KR4))
+		return PORT_NONE;
+
+	return PORT_OTHER;
+}
+
+static int mlxsw_sx_port_get_settings(struct net_device *dev,
+				      struct ethtool_cmd *cmd)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char ptys_pl[MLXSW_REG_PTYS_LEN];
+	u32 eth_proto_cap;
+	u32 eth_proto_admin;
+	u32 eth_proto_oper;
+	int err;
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sx_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(ptys), ptys_pl);
+	if (err) {
+		netdev_err(dev, "Failed to get proto");
+		return err;
+	}
+	mlxsw_reg_ptys_unpack(ptys_pl, &eth_proto_cap,
+			      &eth_proto_admin, &eth_proto_oper);
+
+	cmd->supported = mlxsw_sx_from_ptys_supported_port(eth_proto_cap) |
+			 mlxsw_sx_from_ptys_supported_link(eth_proto_cap) |
+			 SUPPORTED_Pause | SUPPORTED_Asym_Pause;
+	cmd->advertising = mlxsw_sx_from_ptys_advert_link(eth_proto_admin);
+	mlxsw_sx_from_ptys_speed_duplex(netif_carrier_ok(dev),
+					eth_proto_oper, cmd);
+
+	eth_proto_oper = eth_proto_oper ? eth_proto_oper : eth_proto_cap;
+	cmd->port = mlxsw_sx_port_connector_port(eth_proto_oper);
+	cmd->lp_advertising = mlxsw_sx_from_ptys_advert_link(eth_proto_oper);
+
+	cmd->transceiver = XCVR_INTERNAL;
+	return 0;
+}
+
+static u32 mlxsw_sx_to_ptys_advert_link(u32 advertising)
+{
+	u32 ptys_proto = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SX_PORT_LINK_MODE_LEN; i++) {
+		if (advertising & mlxsw_sx_port_link_mode[i].advertised)
+			ptys_proto |= mlxsw_sx_port_link_mode[i].mask;
+	}
+	return ptys_proto;
+}
+
+static u32 mlxsw_sx_to_ptys_speed(u32 speed)
+{
+	u32 ptys_proto = 0;
+	int i;
+
+	for (i = 0; i < MLXSW_SX_PORT_LINK_MODE_LEN; i++) {
+		if (speed == mlxsw_sx_port_link_mode[i].speed)
+			ptys_proto |= mlxsw_sx_port_link_mode[i].mask;
+	}
+	return ptys_proto;
+}
+
+static int mlxsw_sx_port_set_settings(struct net_device *dev,
+				      struct ethtool_cmd *cmd)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char ptys_pl[MLXSW_REG_PTYS_LEN];
+	u32 speed;
+	u32 eth_proto_new;
+	u32 eth_proto_cap;
+	u32 eth_proto_admin;
+	bool is_up;
+	int err;
+
+	speed = ethtool_cmd_speed(cmd);
+
+	eth_proto_new = cmd->autoneg == AUTONEG_ENABLE ?
+		mlxsw_sx_to_ptys_advert_link(cmd->advertising) :
+		mlxsw_sx_to_ptys_speed(speed);
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sx_port->local_port, 0);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(ptys), ptys_pl);
+	if (err) {
+		netdev_err(dev, "Failed to get proto");
+		return err;
+	}
+	mlxsw_reg_ptys_unpack(ptys_pl, &eth_proto_cap, &eth_proto_admin, NULL);
+
+	eth_proto_new = eth_proto_new & eth_proto_cap;
+	if (!eth_proto_new) {
+		netdev_err(dev, "Not supported proto admin requested");
+		return -EINVAL;
+	}
+	if (eth_proto_new == eth_proto_admin)
+		return 0;
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sx_port->local_port, eth_proto_new);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(ptys), ptys_pl);
+	if (err) {
+		netdev_err(dev, "Failed to set proto admin");
+		return err;
+	}
+
+	err = mlxsw_sx_port_oper_status_get(mlxsw_sx_port, &is_up);
+	if (err) {
+		netdev_err(dev, "Failed to get oper status");
+		return err;
+	}
+	if (!is_up)
+		return 0;
+
+	err = mlxsw_sx_port_admin_status_set(mlxsw_sx_port, false);
+	if (err) {
+		netdev_err(dev, "Failed to set admin status");
+		return err;
+	}
+
+	err = mlxsw_sx_port_admin_status_set(mlxsw_sx_port, true);
+	if (err) {
+		netdev_err(dev, "Failed to set admin status");
+		return err;
+	}
+
+	return 0;
+}
+
+static const struct ethtool_ops mlxsw_sx_port_ethtool_ops = {
+	.get_drvinfo		= mlxsw_sx_port_get_drvinfo,
+	.get_link		= ethtool_op_get_link,
+	.get_strings		= mlxsw_sx_port_get_strings,
+	.get_ethtool_stats	= mlxsw_sx_port_get_stats,
+	.get_sset_count		= mlxsw_sx_port_get_sset_count,
+	.get_settings		= mlxsw_sx_port_get_settings,
+	.set_settings		= mlxsw_sx_port_set_settings,
+};
+
+static int mlxsw_sx_port_attr_get(struct net_device *dev,
+				  struct switchdev_attr *attr)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = netdev_priv(dev);
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+
+	switch (attr->id) {
+	case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
+		attr->u.ppid.id_len = sizeof(mlxsw_sx->hw_id);
+		memcpy(&attr->u.ppid.id, &mlxsw_sx->hw_id, attr->u.ppid.id_len);
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+static const struct switchdev_ops mlxsw_sx_port_switchdev_ops = {
+	.switchdev_port_attr_get	= mlxsw_sx_port_attr_get,
+};
+
+static int mlxsw_sx_hw_id_get(struct mlxsw_sx *mlxsw_sx)
+{
+	char spad_pl[MLXSW_REG_SPAD_LEN];
+	int err;
+
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(spad), spad_pl);
+	if (err)
+		return err;
+	mlxsw_reg_spad_base_mac_memcpy_from(spad_pl, mlxsw_sx->hw_id);
+	return 0;
+}
+
+static int mlxsw_sx_port_dev_addr_get(struct mlxsw_sx_port *mlxsw_sx_port)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	struct net_device *dev = mlxsw_sx_port->dev;
+	char ppad_pl[MLXSW_REG_PPAD_LEN];
+	int err;
+
+	mlxsw_reg_ppad_pack(ppad_pl, false, 0);
+	err = mlxsw_reg_query(mlxsw_sx->core, MLXSW_REG(ppad), ppad_pl);
+	if (err)
+		return err;
+	mlxsw_reg_ppad_mac_memcpy_from(ppad_pl, dev->dev_addr);
+	/* The last byte value in base mac address is guaranteed
+	 * to be such it does not overflow when adding local_port
+	 * value.
+	 */
+	dev->dev_addr[ETH_ALEN - 1] += mlxsw_sx_port->local_port;
+	return 0;
+}
+
+static int mlxsw_sx_port_stp_state_set(struct mlxsw_sx_port *mlxsw_sx_port,
+				       u16 vid, enum mlxsw_reg_spms_state state)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char *spms_pl;
+	int err;
+
+	spms_pl = kmalloc(MLXSW_REG_SPMS_LEN, GFP_KERNEL);
+	if (!spms_pl)
+		return -ENOMEM;
+	mlxsw_reg_spms_pack(spms_pl, mlxsw_sx_port->local_port);
+	mlxsw_reg_spms_vid_pack(spms_pl, vid, state);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(spms), spms_pl);
+	kfree(spms_pl);
+	return err;
+}
+
+static int mlxsw_sx_port_speed_set(struct mlxsw_sx_port *mlxsw_sx_port,
+				   u32 speed)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char ptys_pl[MLXSW_REG_PTYS_LEN];
+
+	mlxsw_reg_ptys_pack(ptys_pl, mlxsw_sx_port->local_port, speed);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(ptys), ptys_pl);
+}
+
+static int
+mlxsw_sx_port_mac_learning_mode_set(struct mlxsw_sx_port *mlxsw_sx_port,
+				    enum mlxsw_reg_spmlr_learn_mode mode)
+{
+	struct mlxsw_sx *mlxsw_sx = mlxsw_sx_port->mlxsw_sx;
+	char spmlr_pl[MLXSW_REG_SPMLR_LEN];
+
+	mlxsw_reg_spmlr_pack(spmlr_pl, mlxsw_sx_port->local_port, mode);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(spmlr), spmlr_pl);
+}
+
+static int mlxsw_sx_port_create(struct mlxsw_sx *mlxsw_sx, u8 local_port)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port;
+	struct net_device *dev;
+	bool usable;
+	int err;
+
+	dev = alloc_etherdev(sizeof(struct mlxsw_sx_port));
+	if (!dev)
+		return -ENOMEM;
+	mlxsw_sx_port = netdev_priv(dev);
+	mlxsw_sx_port->dev = dev;
+	mlxsw_sx_port->mlxsw_sx = mlxsw_sx;
+	mlxsw_sx_port->local_port = local_port;
+
+	mlxsw_sx_port->pcpu_stats =
+		netdev_alloc_pcpu_stats(struct mlxsw_sx_port_pcpu_stats);
+	if (!mlxsw_sx_port->pcpu_stats) {
+		err = -ENOMEM;
+		goto err_alloc_stats;
+	}
+
+	dev->netdev_ops = &mlxsw_sx_port_netdev_ops;
+	dev->ethtool_ops = &mlxsw_sx_port_ethtool_ops;
+	dev->switchdev_ops = &mlxsw_sx_port_switchdev_ops;
+
+	err = mlxsw_sx_port_dev_addr_get(mlxsw_sx_port);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Unable to get port mac address\n",
+			mlxsw_sx_port->local_port);
+		goto err_dev_addr_get;
+	}
+
+	netif_carrier_off(dev);
+
+	dev->features |= NETIF_F_NETNS_LOCAL | NETIF_F_LLTX | NETIF_F_SG |
+			 NETIF_F_VLAN_CHALLENGED;
+
+	/* Each packet needs to have a Tx header (metadata) on top all other
+	 * headers.
+	 */
+	dev->hard_header_len += MLXSW_TXHDR_LEN;
+
+	err = mlxsw_sx_port_module_check(mlxsw_sx_port, &usable);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to check module\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_module_check;
+	}
+
+	if (!usable) {
+		dev_dbg(mlxsw_sx->bus_info->dev, "Port %d: Not usable, skipping initialization\n",
+			mlxsw_sx_port->local_port);
+		goto port_not_usable;
+	}
+
+	err = mlxsw_sx_port_system_port_mapping_set(mlxsw_sx_port);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to set system port mapping\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_system_port_mapping_set;
+	}
+
+	err = mlxsw_sx_port_swid_set(mlxsw_sx_port, 0);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to set SWID\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_swid_set;
+	}
+
+	err = mlxsw_sx_port_speed_set(mlxsw_sx_port,
+				      MLXSW_REG_PTYS_ETH_SPEED_40GBASE_CR4);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to set speed\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_speed_set;
+	}
+
+	err = mlxsw_sx_port_mtu_set(mlxsw_sx_port, ETH_DATA_LEN);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to set MTU\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_mtu_set;
+	}
+
+	err = mlxsw_sx_port_admin_status_set(mlxsw_sx_port, false);
+	if (err)
+		goto err_port_admin_status_set;
+
+	err = mlxsw_sx_port_stp_state_set(mlxsw_sx_port,
+					  MLXSW_PORT_DEFAULT_VID,
+					  MLXSW_REG_SPMS_STATE_FORWARDING);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to set STP state\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_stp_state_set;
+	}
+
+	err = mlxsw_sx_port_mac_learning_mode_set(mlxsw_sx_port,
+						  MLXSW_REG_SPMLR_LEARN_MODE_DISABLE);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to set MAC learning mode\n",
+			mlxsw_sx_port->local_port);
+		goto err_port_mac_learning_mode_set;
+	}
+
+	err = register_netdev(dev);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Port %d: Failed to register netdev\n",
+			mlxsw_sx_port->local_port);
+		goto err_register_netdev;
+	}
+
+	mlxsw_sx->ports[local_port] = mlxsw_sx_port;
+	return 0;
+
+err_register_netdev:
+err_port_mac_learning_mode_set:
+err_port_stp_state_set:
+err_port_admin_status_set:
+err_port_mtu_set:
+err_port_speed_set:
+err_port_swid_set:
+err_port_system_port_mapping_set:
+port_not_usable:
+err_port_module_check:
+err_dev_addr_get:
+	free_percpu(mlxsw_sx_port->pcpu_stats);
+err_alloc_stats:
+	free_netdev(dev);
+	return err;
+}
+
+static void mlxsw_sx_port_remove(struct mlxsw_sx *mlxsw_sx, u8 local_port)
+{
+	struct mlxsw_sx_port *mlxsw_sx_port = mlxsw_sx->ports[local_port];
+
+	if (!mlxsw_sx_port)
+		return;
+	unregister_netdev(mlxsw_sx_port->dev); /* This calls ndo_stop */
+	mlxsw_sx_port_swid_set(mlxsw_sx_port, MLXSW_PORT_SWID_DISABLED_PORT);
+	free_percpu(mlxsw_sx_port->pcpu_stats);
+	free_netdev(mlxsw_sx_port->dev);
+}
+
+static void mlxsw_sx_ports_remove(struct mlxsw_sx *mlxsw_sx)
+{
+	int i;
+
+	for (i = 1; i < MLXSW_PORT_MAX_PORTS; i++)
+		mlxsw_sx_port_remove(mlxsw_sx, i);
+	kfree(mlxsw_sx->ports);
+}
+
+static int mlxsw_sx_ports_create(struct mlxsw_sx *mlxsw_sx)
+{
+	size_t alloc_size;
+	int i;
+	int err;
+
+	alloc_size = sizeof(struct mlxsw_sx_port *) * MLXSW_PORT_MAX_PORTS;
+	mlxsw_sx->ports = kzalloc(alloc_size, GFP_KERNEL);
+	if (!mlxsw_sx->ports)
+		return -ENOMEM;
+
+	for (i = 1; i < MLXSW_PORT_MAX_PORTS; i++) {
+		err = mlxsw_sx_port_create(mlxsw_sx, i);
+		if (err)
+			goto err_port_create;
+	}
+	return 0;
+
+err_port_create:
+	for (i--; i >= 1; i--)
+		mlxsw_sx_port_remove(mlxsw_sx, i);
+	kfree(mlxsw_sx->ports);
+	return err;
+}
+
+static void mlxsw_sx_pude_event_func(const struct mlxsw_reg_info *reg,
+				     char *pude_pl, void *priv)
+{
+	struct mlxsw_sx *mlxsw_sx = priv;
+	struct mlxsw_sx_port *mlxsw_sx_port;
+	enum mlxsw_reg_pude_oper_status status;
+	u8 local_port;
+
+	local_port = mlxsw_reg_pude_local_port_get(pude_pl);
+	mlxsw_sx_port = mlxsw_sx->ports[local_port];
+	if (!mlxsw_sx_port) {
+		dev_warn(mlxsw_sx->bus_info->dev, "Port %d: Link event received for non-existent port\n",
+			 local_port);
+		return;
+	}
+
+	status = mlxsw_reg_pude_oper_status_get(pude_pl);
+	if (status == MLXSW_PORT_OPER_STATUS_UP) {
+		netdev_info(mlxsw_sx_port->dev, "link up\n");
+		netif_carrier_on(mlxsw_sx_port->dev);
+	} else {
+		netdev_info(mlxsw_sx_port->dev, "link down\n");
+		netif_carrier_off(mlxsw_sx_port->dev);
+	}
+}
+
+static struct mlxsw_event_listener mlxsw_sx_pude_event = {
+	.func = mlxsw_sx_pude_event_func,
+	.trap_id = MLXSW_TRAP_ID_PUDE,
+};
+
+static int mlxsw_sx_event_register(struct mlxsw_sx *mlxsw_sx,
+				   enum mlxsw_event_trap_id trap_id)
+{
+	struct mlxsw_event_listener *el;
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int err;
+
+	switch (trap_id) {
+	case MLXSW_TRAP_ID_PUDE:
+		el = &mlxsw_sx_pude_event;
+		break;
+	}
+	err = mlxsw_core_event_listener_register(mlxsw_sx->core, el, mlxsw_sx);
+	if (err)
+		return err;
+
+	mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_FORWARD, trap_id);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(hpkt), hpkt_pl);
+	if (err)
+		goto err_event_trap_set;
+
+	return 0;
+
+err_event_trap_set:
+	mlxsw_core_event_listener_unregister(mlxsw_sx->core, el, mlxsw_sx);
+	return err;
+}
+
+static void mlxsw_sx_event_unregister(struct mlxsw_sx *mlxsw_sx,
+				      enum mlxsw_event_trap_id trap_id)
+{
+	struct mlxsw_event_listener *el;
+
+	switch (trap_id) {
+	case MLXSW_TRAP_ID_PUDE:
+		el = &mlxsw_sx_pude_event;
+		break;
+	}
+	mlxsw_core_event_listener_unregister(mlxsw_sx->core, el, mlxsw_sx);
+}
+
+static void mlxsw_sx_rx_listener_func(struct sk_buff *skb, u8 local_port,
+				      void *priv)
+{
+	struct mlxsw_sx *mlxsw_sx = priv;
+	struct mlxsw_sx_port *mlxsw_sx_port = mlxsw_sx->ports[local_port];
+	struct mlxsw_sx_port_pcpu_stats *pcpu_stats;
+
+	if (unlikely(!mlxsw_sx_port)) {
+		dev_warn_ratelimited(mlxsw_sx->bus_info->dev, "Port %d: skb received for non-existent port\n",
+				     local_port);
+		return;
+	}
+
+	skb->dev = mlxsw_sx_port->dev;
+
+	pcpu_stats = this_cpu_ptr(mlxsw_sx_port->pcpu_stats);
+	u64_stats_update_begin(&pcpu_stats->syncp);
+	pcpu_stats->rx_packets++;
+	pcpu_stats->rx_bytes += skb->len;
+	u64_stats_update_end(&pcpu_stats->syncp);
+
+	skb->protocol = eth_type_trans(skb, skb->dev);
+	netif_receive_skb(skb);
+}
+
+static const struct mlxsw_rx_listener mlxsw_sx_rx_listener[] = {
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_FDB_MC,
+	},
+	/* Traps for specific L2 packet types, not trapped as FDB MC */
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_STP,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_LACP,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_EAPOL,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_LLDP,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_MMRP,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_MVRP,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_RPVST,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_DHCP,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_QUERY,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V1_REPORT,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V2_REPORT,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V2_LEAVE,
+	},
+	{
+		.func = mlxsw_sx_rx_listener_func,
+		.local_port = MLXSW_PORT_DONT_CARE,
+		.trap_id = MLXSW_TRAP_ID_IGMP_V3_REPORT,
+	},
+};
+
+static int mlxsw_sx_traps_init(struct mlxsw_sx *mlxsw_sx)
+{
+	char htgt_pl[MLXSW_REG_HTGT_LEN];
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int i;
+	int err;
+
+	mlxsw_reg_htgt_pack(htgt_pl, MLXSW_REG_HTGT_TRAP_GROUP_RX);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(htgt), htgt_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_htgt_pack(htgt_pl, MLXSW_REG_HTGT_TRAP_GROUP_CTRL);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(htgt), htgt_pl);
+	if (err)
+		return err;
+
+	for (i = 0; i < ARRAY_SIZE(mlxsw_sx_rx_listener); i++) {
+		err = mlxsw_core_rx_listener_register(mlxsw_sx->core,
+						      &mlxsw_sx_rx_listener[i],
+						      mlxsw_sx);
+		if (err)
+			goto err_rx_listener_register;
+
+		mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_TRAP_TO_CPU,
+				    mlxsw_sx_rx_listener[i].trap_id);
+		err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(hpkt), hpkt_pl);
+		if (err)
+			goto err_rx_trap_set;
+	}
+	return 0;
+
+err_rx_trap_set:
+	mlxsw_core_rx_listener_unregister(mlxsw_sx->core,
+					  &mlxsw_sx_rx_listener[i],
+					  mlxsw_sx);
+err_rx_listener_register:
+	for (i--; i >= 0; i--) {
+		mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_FORWARD,
+				    mlxsw_sx_rx_listener[i].trap_id);
+		mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(hpkt), hpkt_pl);
+
+		mlxsw_core_rx_listener_unregister(mlxsw_sx->core,
+						  &mlxsw_sx_rx_listener[i],
+						  mlxsw_sx);
+	}
+	return err;
+}
+
+static void mlxsw_sx_traps_fini(struct mlxsw_sx *mlxsw_sx)
+{
+	char hpkt_pl[MLXSW_REG_HPKT_LEN];
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(mlxsw_sx_rx_listener); i++) {
+		mlxsw_reg_hpkt_pack(hpkt_pl, MLXSW_REG_HPKT_ACTION_FORWARD,
+				    mlxsw_sx_rx_listener[i].trap_id);
+		mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(hpkt), hpkt_pl);
+
+		mlxsw_core_rx_listener_unregister(mlxsw_sx->core,
+						  &mlxsw_sx_rx_listener[i],
+						  mlxsw_sx);
+	}
+}
+
+static int mlxsw_sx_flood_init(struct mlxsw_sx *mlxsw_sx)
+{
+	char sfgc_pl[MLXSW_REG_SFGC_LEN];
+	char sgcr_pl[MLXSW_REG_SGCR_LEN];
+	char *sftr_pl;
+	int err;
+
+	/* Configure a flooding table, which includes only CPU port. */
+	sftr_pl = kmalloc(MLXSW_REG_SFTR_LEN, GFP_KERNEL);
+	if (!sftr_pl)
+		return -ENOMEM;
+	mlxsw_reg_sftr_pack(sftr_pl, 0, 0, MLXSW_REG_SFGC_TABLE_TYPE_SINGLE, 0,
+			    MLXSW_PORT_CPU_PORT, true);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sftr), sftr_pl);
+	kfree(sftr_pl);
+	if (err)
+		return err;
+
+	/* Flood different packet types using the flooding table. */
+	mlxsw_reg_sfgc_pack(sfgc_pl,
+			    MLXSW_REG_SFGC_TYPE_UNKNOWN_UNICAST,
+			    MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID,
+			    MLXSW_REG_SFGC_TABLE_TYPE_SINGLE,
+			    0);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sfgc), sfgc_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_sfgc_pack(sfgc_pl,
+			    MLXSW_REG_SFGC_TYPE_BROADCAST,
+			    MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID,
+			    MLXSW_REG_SFGC_TABLE_TYPE_SINGLE,
+			    0);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sfgc), sfgc_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_sfgc_pack(sfgc_pl,
+			    MLXSW_REG_SFGC_TYPE_UNREGISTERED_MULTICAST_NON_IP,
+			    MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID,
+			    MLXSW_REG_SFGC_TABLE_TYPE_SINGLE,
+			    0);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sfgc), sfgc_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_sfgc_pack(sfgc_pl,
+			    MLXSW_REG_SFGC_TYPE_UNREGISTERED_MULTICAST_IPV6,
+			    MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID,
+			    MLXSW_REG_SFGC_TABLE_TYPE_SINGLE,
+			    0);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sfgc), sfgc_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_sfgc_pack(sfgc_pl,
+			    MLXSW_REG_SFGC_TYPE_UNREGISTERED_MULTICAST_IPV4,
+			    MLXSW_REG_SFGC_BRIDGE_TYPE_1Q_FID,
+			    MLXSW_REG_SFGC_TABLE_TYPE_SINGLE,
+			    0);
+	err = mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sfgc), sfgc_pl);
+	if (err)
+		return err;
+
+	mlxsw_reg_sgcr_pack(sgcr_pl, true);
+	return mlxsw_reg_write(mlxsw_sx->core, MLXSW_REG(sgcr), sgcr_pl);
+}
+
+static int mlxsw_sx_init(void *priv, struct mlxsw_core *mlxsw_core,
+			 const struct mlxsw_bus_info *mlxsw_bus_info)
+{
+	struct mlxsw_sx *mlxsw_sx = priv;
+	int err;
+
+	mlxsw_sx->core = mlxsw_core;
+	mlxsw_sx->bus_info = mlxsw_bus_info;
+
+	err = mlxsw_sx_hw_id_get(mlxsw_sx);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Failed to get switch HW ID\n");
+		return err;
+	}
+
+	err = mlxsw_sx_ports_create(mlxsw_sx);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Failed to create ports\n");
+		return err;
+	}
+
+	err = mlxsw_sx_event_register(mlxsw_sx, MLXSW_TRAP_ID_PUDE);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Failed to register for PUDE events\n");
+		goto err_event_register;
+	}
+
+	err = mlxsw_sx_traps_init(mlxsw_sx);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Failed to set traps for RX\n");
+		goto err_rx_listener_register;
+	}
+
+	err = mlxsw_sx_flood_init(mlxsw_sx);
+	if (err) {
+		dev_err(mlxsw_sx->bus_info->dev, "Failed to initialize flood tables\n");
+		goto err_flood_init;
+	}
+
+	return 0;
+
+err_flood_init:
+	mlxsw_sx_traps_fini(mlxsw_sx);
+err_rx_listener_register:
+	mlxsw_sx_event_unregister(mlxsw_sx, MLXSW_TRAP_ID_PUDE);
+err_event_register:
+	mlxsw_sx_ports_remove(mlxsw_sx);
+	return err;
+}
+
+static void mlxsw_sx_fini(void *priv)
+{
+	struct mlxsw_sx *mlxsw_sx = priv;
+
+	mlxsw_sx_traps_fini(mlxsw_sx);
+	mlxsw_sx_event_unregister(mlxsw_sx, MLXSW_TRAP_ID_PUDE);
+	mlxsw_sx_ports_remove(mlxsw_sx);
+}
+
+static struct mlxsw_config_profile mlxsw_sx_config_profile = {
+	.used_max_vepa_channels		= 1,
+	.max_vepa_channels		= 0,
+	.used_max_lag			= 1,
+	.max_lag			= 64,
+	.used_max_port_per_lag		= 1,
+	.max_port_per_lag		= 16,
+	.used_max_mid			= 1,
+	.max_mid			= 7000,
+	.used_max_pgt			= 1,
+	.max_pgt			= 0,
+	.used_max_system_port		= 1,
+	.max_system_port		= 48000,
+	.used_max_vlan_groups		= 1,
+	.max_vlan_groups		= 127,
+	.used_max_regions		= 1,
+	.max_regions			= 400,
+	.used_flood_tables		= 1,
+	.max_flood_tables		= 2,
+	.max_vid_flood_tables		= 1,
+	.used_flood_mode		= 1,
+	.flood_mode			= 3,
+	.used_max_ib_mc			= 1,
+	.max_ib_mc			= 0,
+	.used_max_pkey			= 1,
+	.max_pkey			= 0,
+	.swid_config			= {
+		{
+			.used_type	= 1,
+			.type		= MLXSW_PORT_SWID_TYPE_ETH,
+		}
+	},
+};
+
+static struct mlxsw_driver mlxsw_sx_driver = {
+	.kind			= MLXSW_DEVICE_KIND_SWITCHX2,
+	.owner			= THIS_MODULE,
+	.priv_size		= sizeof(struct mlxsw_sx),
+	.init			= mlxsw_sx_init,
+	.fini			= mlxsw_sx_fini,
+	.txhdr_construct	= mlxsw_sx_txhdr_construct,
+	.txhdr_len		= MLXSW_TXHDR_LEN,
+	.profile		= &mlxsw_sx_config_profile,
+};
+
+static int __init mlxsw_sx_module_init(void)
+{
+	return mlxsw_core_driver_register(&mlxsw_sx_driver);
+}
+
+static void __exit mlxsw_sx_module_exit(void)
+{
+	mlxsw_core_driver_unregister(&mlxsw_sx_driver);
+}
+
+module_init(mlxsw_sx_module_init);
+module_exit(mlxsw_sx_module_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
+MODULE_AUTHOR("Jiri Pirko <jiri@mellanox.com>");
+MODULE_DESCRIPTION("Mellanox SwitchX-2 driver");
+MODULE_MLXSW_DRIVER_ALIAS(MLXSW_DEVICE_KIND_SWITCHX2);
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/trap.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/trap.h
new file mode 100644
index 000000000..53a9550be
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/trap.h
@@ -0,0 +1,66 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/trap.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Elad Raz <eladr@mellanox.com>
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#ifndef _MLXSW_TRAP_H
+#define _MLXSW_TRAP_H
+
+enum {
+	/* Ethernet EMAD and FDB miss */
+	MLXSW_TRAP_ID_FDB_MC = 0x01,
+	MLXSW_TRAP_ID_ETHEMAD = 0x05,
+	/* L2 traps for specific packet types */
+	MLXSW_TRAP_ID_STP = 0x10,
+	MLXSW_TRAP_ID_LACP = 0x11,
+	MLXSW_TRAP_ID_EAPOL = 0x12,
+	MLXSW_TRAP_ID_LLDP = 0x13,
+	MLXSW_TRAP_ID_MMRP = 0x14,
+	MLXSW_TRAP_ID_MVRP = 0x15,
+	MLXSW_TRAP_ID_RPVST = 0x16,
+	MLXSW_TRAP_ID_DHCP = 0x19,
+	MLXSW_TRAP_ID_IGMP_QUERY = 0x30,
+	MLXSW_TRAP_ID_IGMP_V1_REPORT = 0x31,
+	MLXSW_TRAP_ID_IGMP_V2_REPORT = 0x32,
+	MLXSW_TRAP_ID_IGMP_V2_LEAVE = 0x33,
+	MLXSW_TRAP_ID_IGMP_V3_REPORT = 0x34,
+
+	MLXSW_TRAP_ID_MAX = 0x1FF
+};
+
+enum mlxsw_event_trap_id {
+	/* Port Up/Down event generated by hardware */
+	MLXSW_TRAP_ID_PUDE = 0x8,
+};
+
+#endif /* _MLXSW_TRAP_H */
diff --git a/kernel/drivers/net/ethernet/mellanox/mlxsw/txheader.h b/kernel/drivers/net/ethernet/mellanox/mlxsw/txheader.h
new file mode 100644
index 000000000..fdf94720c
--- /dev/null
+++ b/kernel/drivers/net/ethernet/mellanox/mlxsw/txheader.h
@@ -0,0 +1,81 @@
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/txheader.h
+ * Copyright (c) 2015 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2015 Ido Schimmel <idosch@mellanox.com>
+ * Copyright (c) 2015 Jiri Pirko <jiri@mellanox.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. 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.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _MLXSW_TXHEADER_H
+#define _MLXSW_TXHEADER_H
+
+#define MLXSW_TXHDR_LEN 0x10
+#define MLXSW_TXHDR_VERSION_0 0
+#define MLXSW_TXHDR_VERSION_1 1
+
+enum {
+	MLXSW_TXHDR_ETH_CTL,
+	MLXSW_TXHDR_ETH_DATA,
+};
+
+#define MLXSW_TXHDR_PROTO_ETH 1
+
+enum {
+	MLXSW_TXHDR_ETCLASS_0,
+	MLXSW_TXHDR_ETCLASS_1,
+	MLXSW_TXHDR_ETCLASS_2,
+	MLXSW_TXHDR_ETCLASS_3,
+	MLXSW_TXHDR_ETCLASS_4,
+	MLXSW_TXHDR_ETCLASS_5,
+	MLXSW_TXHDR_ETCLASS_6,
+	MLXSW_TXHDR_ETCLASS_7,
+};
+
+enum {
+	MLXSW_TXHDR_RDQ_OTHER,
+	MLXSW_TXHDR_RDQ_EMAD = 0x1f,
+};
+
+#define MLXSW_TXHDR_CTCLASS3 0
+#define MLXSW_TXHDR_CPU_SIG 0
+#define MLXSW_TXHDR_SIG 0xE0E0
+#define MLXSW_TXHDR_STCLASS_NONE 0
+
+enum {
+	MLXSW_TXHDR_NOT_EMAD,
+	MLXSW_TXHDR_EMAD,
+};
+
+enum {
+	MLXSW_TXHDR_TYPE_DATA,
+	MLXSW_TXHDR_TYPE_CONTROL = 6,
+};
+
+#endif