Merge tag 'mlx5-updates-2020-05-09' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2020-05-09

This series includes updates to mlx5 netdev driver and bonding updates
to support getting the next active tx slave.

1) merge commit with mlx5-next that includes bonding updates from Maor
   Bonding: Add support to get xmit slave
2) Maxim makes some general code improvements to TX data path
3) Tariq makes some general code improvements to kTLS and mlx5 accel layer
in preparation for mlx5 TLS RX.
====================

Signed-off-by: Jakub Kicinski <kuba@kernel.org>

25 files changed, 498 insertions, 284 deletions

diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
index 3a598d04b156..095ea51d1853 100644
--- a/drivers/net/bonding/bond_alb.c
+++ b/drivers/net/bonding/bond_alb.c
@@ -1331,11 +1331,11 @@ static netdev_tx_t bond_do_alb_xmit(struct sk_buff *skb, struct bonding *bond,
 	return bond_tx_drop(bond->dev, skb);
 }
 
-netdev_tx_t bond_tlb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
+struct slave *bond_xmit_tlb_slave_get(struct bonding *bond,
+				      struct sk_buff *skb)
 {
-	struct bonding *bond = netdev_priv(bond_dev);
-	struct ethhdr *eth_data;
 	struct slave *tx_slave = NULL;
+	struct ethhdr *eth_data;
 	u32 hash_index;
 
 	skb_reset_mac_header(skb);
@@ -1357,7 +1357,7 @@ netdev_tx_t bond_tlb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
 				struct bond_up_slave *slaves;
 				unsigned int count;
 
-				slaves = rcu_dereference(bond->slave_arr);
+				slaves = rcu_dereference(bond->usable_slaves);
 				count = slaves ? READ_ONCE(slaves->count) : 0;
 				if (likely(count))
 					tx_slave = slaves->arr[hash_index %
@@ -1366,20 +1366,29 @@ netdev_tx_t bond_tlb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
 			break;
 		}
 	}
-	return bond_do_alb_xmit(skb, bond, tx_slave);
+	return tx_slave;
 }
 
-netdev_tx_t bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
+netdev_tx_t bond_tlb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
-	struct ethhdr *eth_data;
+	struct slave *tx_slave;
+
+	tx_slave = bond_xmit_tlb_slave_get(bond, skb);
+	return bond_do_alb_xmit(skb, bond, tx_slave);
+}
+
+struct slave *bond_xmit_alb_slave_get(struct bonding *bond,
+				      struct sk_buff *skb)
+{
 	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
-	struct slave *tx_slave = NULL;
 	static const __be32 ip_bcast = htonl(0xffffffff);
-	int hash_size = 0;
+	struct slave *tx_slave = NULL;
+	const u8 *hash_start = NULL;
 	bool do_tx_balance = true;
+	struct ethhdr *eth_data;
 	u32 hash_index = 0;
-	const u8 *hash_start = NULL;
+	int hash_size = 0;
 
 	skb_reset_mac_header(skb);
 	eth_data = eth_hdr(skb);
@@ -1491,14 +1500,22 @@ netdev_tx_t bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
 			struct bond_up_slave *slaves;
 			unsigned int count;
 
-			slaves = rcu_dereference(bond->slave_arr);
+			slaves = rcu_dereference(bond->usable_slaves);
 			count = slaves ? READ_ONCE(slaves->count) : 0;
 			if (likely(count))
 				tx_slave = slaves->arr[bond_xmit_hash(bond, skb) %
 						       count];
 		}
 	}
+	return tx_slave;
+}
+
+netdev_tx_t bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev)
+{
+	struct bonding *bond = netdev_priv(bond_dev);
+	struct slave *tx_slave = NULL;
 
+	tx_slave = bond_xmit_alb_slave_get(bond, skb);
 	return bond_do_alb_xmit(skb, bond, tx_slave);
 }
 
diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
index 4f9e7c421f57..a25c65d4af71 100644
--- a/drivers/net/bonding/bond_main.c
+++ b/drivers/net/bonding/bond_main.c
@@ -3923,16 +3923,15 @@ unwind:
 }
 
 /**
- * bond_xmit_slave_id - transmit skb through slave with slave_id
+ * bond_get_slave_by_id - get xmit slave with slave_id
  * @bond: bonding device that is transmitting
- * @skb: buffer to transmit
  * @slave_id: slave id up to slave_cnt-1 through which to transmit
  *
- * This function tries to transmit through slave with slave_id but in case
+ * This function tries to get slave with slave_id but in case
  * it fails, it tries to find the first available slave for transmission.
- * The skb is consumed in all cases, thus the function is void.
  */
-static netdev_tx_t bond_xmit_slave_id(struct bonding *bond, struct sk_buff *skb, int slave_id)
+static struct slave *bond_get_slave_by_id(struct bonding *bond,
+					  int slave_id)
 {
 	struct list_head *iter;
 	struct slave *slave;
@@ -3942,7 +3941,7 @@ static netdev_tx_t bond_xmit_slave_id(struct bonding *bond, struct sk_buff *skb,
 	bond_for_each_slave_rcu(bond, slave, iter) {
 		if (--i < 0) {
 			if (bond_slave_can_tx(slave))
-				return bond_dev_queue_xmit(bond, skb, slave->dev);
+				return slave;
 		}
 	}
 
@@ -3952,10 +3951,10 @@ static netdev_tx_t bond_xmit_slave_id(struct bonding *bond, struct sk_buff *skb,
 		if (--i < 0)
 			break;
 		if (bond_slave_can_tx(slave))
-			return bond_dev_queue_xmit(bond, skb, slave->dev);
+			return slave;
 	}
 	/* no slave that can tx has been found */
-	return bond_tx_drop(bond->dev, skb);
+	return NULL;
 }
 
 /**
@@ -3991,10 +3990,9 @@ static u32 bond_rr_gen_slave_id(struct bonding *bond)
 	return slave_id;
 }
 
-static netdev_tx_t bond_xmit_roundrobin(struct sk_buff *skb,
-					struct net_device *bond_dev)
+static struct slave *bond_xmit_roundrobin_slave_get(struct bonding *bond,
+						    struct sk_buff *skb)
 {
-	struct bonding *bond = netdev_priv(bond_dev);
 	struct slave *slave;
 	int slave_cnt;
 	u32 slave_id;
@@ -4016,20 +4014,39 @@ static netdev_tx_t bond_xmit_roundrobin(struct sk_buff *skb,
 		if (iph->protocol == IPPROTO_IGMP) {
 			slave = rcu_dereference(bond->curr_active_slave);
 			if (slave)
-				return bond_dev_queue_xmit(bond, skb, slave->dev);
-			return bond_xmit_slave_id(bond, skb, 0);
+				return slave;
+			return bond_get_slave_by_id(bond, 0);
 		}
 	}
 
 non_igmp:
 	slave_cnt = READ_ONCE(bond->slave_cnt);
 	if (likely(slave_cnt)) {
-		slave_id = bond_rr_gen_slave_id(bond);
-		return bond_xmit_slave_id(bond, skb, slave_id % slave_cnt);
+		slave_id = bond_rr_gen_slave_id(bond) % slave_cnt;
+		return bond_get_slave_by_id(bond, slave_id);
 	}
+	return NULL;
+}
+
+static netdev_tx_t bond_xmit_roundrobin(struct sk_buff *skb,
+					struct net_device *bond_dev)
+{
+	struct bonding *bond = netdev_priv(bond_dev);
+	struct slave *slave;
+
+	slave = bond_xmit_roundrobin_slave_get(bond, skb);
+	if (likely(slave))
+		return bond_dev_queue_xmit(bond, skb, slave->dev);
+
 	return bond_tx_drop(bond_dev, skb);
 }
 
+static struct slave *bond_xmit_activebackup_slave_get(struct bonding *bond,
+						      struct sk_buff *skb)
+{
+	return rcu_dereference(bond->curr_active_slave);
+}
+
 /* In active-backup mode, we know that bond->curr_active_slave is always valid if
  * the bond has a usable interface.
  */
@@ -4039,7 +4056,7 @@ static netdev_tx_t bond_xmit_activebackup(struct sk_buff *skb,
 	struct bonding *bond = netdev_priv(bond_dev);
 	struct slave *slave;
 
-	slave = rcu_dereference(bond->curr_active_slave);
+	slave = bond_xmit_activebackup_slave_get(bond, skb);
 	if (slave)
 		return bond_dev_queue_xmit(bond, skb, slave->dev);
 
@@ -4077,6 +4094,61 @@ err:
 	bond_slave_arr_work_rearm(bond, 1);
 }
 
+static void bond_skip_slave(struct bond_up_slave *slaves,
+			    struct slave *skipslave)
+{
+	int idx;
+
+	/* Rare situation where caller has asked to skip a specific
+	 * slave but allocation failed (most likely!). BTW this is
+	 * only possible when the call is initiated from
+	 * __bond_release_one(). In this situation; overwrite the
+	 * skipslave entry in the array with the last entry from the
+	 * array to avoid a situation where the xmit path may choose
+	 * this to-be-skipped slave to send a packet out.
+	 */
+	for (idx = 0; slaves && idx < slaves->count; idx++) {
+		if (skipslave == slaves->arr[idx]) {
+			slaves->arr[idx] =
+				slaves->arr[slaves->count - 1];
+			slaves->count--;
+			break;
+		}
+	}
+}
+
+static void bond_set_slave_arr(struct bonding *bond,
+			       struct bond_up_slave *usable_slaves,
+			       struct bond_up_slave *all_slaves)
+{
+	struct bond_up_slave *usable, *all;
+
+	usable = rtnl_dereference(bond->usable_slaves);
+	rcu_assign_pointer(bond->usable_slaves, usable_slaves);
+	kfree_rcu(usable, rcu);
+
+	all = rtnl_dereference(bond->all_slaves);
+	rcu_assign_pointer(bond->all_slaves, all_slaves);
+	kfree_rcu(all, rcu);
+}
+
+static void bond_reset_slave_arr(struct bonding *bond)
+{
+	struct bond_up_slave *usable, *all;
+
+	usable = rtnl_dereference(bond->usable_slaves);
+	if (usable) {
+		RCU_INIT_POINTER(bond->usable_slaves, NULL);
+		kfree_rcu(usable, rcu);
+	}
+
+	all = rtnl_dereference(bond->all_slaves);
+	if (all) {
+		RCU_INIT_POINTER(bond->all_slaves, NULL);
+		kfree_rcu(all, rcu);
+	}
+}
+
 /* Build the usable slaves array in control path for modes that use xmit-hash
  * to determine the slave interface -
  * (a) BOND_MODE_8023AD
@@ -4087,9 +4159,9 @@ err:
  */
 int bond_update_slave_arr(struct bonding *bond, struct slave *skipslave)
 {
+	struct bond_up_slave *usable_slaves = NULL, *all_slaves = NULL;
 	struct slave *slave;
 	struct list_head *iter;
-	struct bond_up_slave *new_arr, *old_arr;
 	int agg_id = 0;
 	int ret = 0;
 
@@ -4097,11 +4169,12 @@ int bond_update_slave_arr(struct bonding *bond, struct slave *skipslave)
 	WARN_ON(lockdep_is_held(&bond->mode_lock));
 #endif
 
-	new_arr = kzalloc(offsetof(struct bond_up_slave, arr[bond->slave_cnt]),
-			  GFP_KERNEL);
-	if (!new_arr) {
+	usable_slaves = kzalloc(struct_size(usable_slaves, arr,
+					    bond->slave_cnt), GFP_KERNEL);
+	all_slaves = kzalloc(struct_size(all_slaves, arr,
+					 bond->slave_cnt), GFP_KERNEL);
+	if (!usable_slaves || !all_slaves) {
 		ret = -ENOMEM;
-		pr_err("Failed to build slave-array.\n");
 		goto out;
 	}
 	if (BOND_MODE(bond) == BOND_MODE_8023AD) {
@@ -4109,20 +4182,19 @@ int bond_update_slave_arr(struct bonding *bond, struct slave *skipslave)
 
 		if (bond_3ad_get_active_agg_info(bond, &ad_info)) {
 			pr_debug("bond_3ad_get_active_agg_info failed\n");
-			kfree_rcu(new_arr, rcu);
 			/* No active aggragator means it's not safe to use
 			 * the previous array.
 			 */
-			old_arr = rtnl_dereference(bond->slave_arr);
-			if (old_arr) {
-				RCU_INIT_POINTER(bond->slave_arr, NULL);
-				kfree_rcu(old_arr, rcu);
-			}
+			bond_reset_slave_arr(bond);
 			goto out;
 		}
 		agg_id = ad_info.aggregator_id;
 	}
 	bond_for_each_slave(bond, slave, iter) {
+		if (skipslave == slave)
+			continue;
+
+		all_slaves->arr[all_slaves->count++] = slave;
 		if (BOND_MODE(bond) == BOND_MODE_8023AD) {
 			struct aggregator *agg;
 
@@ -4132,44 +4204,45 @@ int bond_update_slave_arr(struct bonding *bond, struct slave *skipslave)
 		}
 		if (!bond_slave_can_tx(slave))
 			continue;
-		if (skipslave == slave)
-			continue;
 
 		slave_dbg(bond->dev, slave->dev, "Adding slave to tx hash array[%d]\n",
-			  new_arr->count);
+			  usable_slaves->count);
 
-		new_arr->arr[new_arr->count++] = slave;
+		usable_slaves->arr[usable_slaves->count++] = slave;
 	}
 
-	old_arr = rtnl_dereference(bond->slave_arr);
-	rcu_assign_pointer(bond->slave_arr, new_arr);
-	if (old_arr)
-		kfree_rcu(old_arr, rcu);
+	bond_set_slave_arr(bond, usable_slaves, all_slaves);
+	return ret;
 out:
 	if (ret != 0 && skipslave) {
-		int idx;
-
-		/* Rare situation where caller has asked to skip a specific
-		 * slave but allocation failed (most likely!). BTW this is
-		 * only possible when the call is initiated from
-		 * __bond_release_one(). In this situation; overwrite the
-		 * skipslave entry in the array with the last entry from the
-		 * array to avoid a situation where the xmit path may choose
-		 * this to-be-skipped slave to send a packet out.
-		 */
-		old_arr = rtnl_dereference(bond->slave_arr);
-		for (idx = 0; old_arr != NULL && idx < old_arr->count; idx++) {
-			if (skipslave == old_arr->arr[idx]) {
-				old_arr->arr[idx] =
-				    old_arr->arr[old_arr->count-1];
-				old_arr->count--;
-				break;
-			}
-		}
+		bond_skip_slave(rtnl_dereference(bond->all_slaves),
+				skipslave);
+		bond_skip_slave(rtnl_dereference(bond->usable_slaves),
+				skipslave);
 	}
+	kfree_rcu(all_slaves, rcu);
+	kfree_rcu(usable_slaves, rcu);
+
 	return ret;
 }
 
+static struct slave *bond_xmit_3ad_xor_slave_get(struct bonding *bond,
+						 struct sk_buff *skb,
+						 struct bond_up_slave *slaves)
+{
+	struct slave *slave;
+	unsigned int count;
+	u32 hash;
+
+	hash = bond_xmit_hash(bond, skb);
+	count = slaves ? READ_ONCE(slaves->count) : 0;
+	if (unlikely(!count))
+		return NULL;
+
+	slave = slaves->arr[hash % count];
+	return slave;
+}
+
 /* Use this Xmit function for 3AD as well as XOR modes. The current
  * usable slave array is formed in the control path. The xmit function
  * just calculates hash and sends the packet out.
@@ -4178,16 +4251,14 @@ static netdev_tx_t bond_3ad_xor_xmit(struct sk_buff *skb,
 				     struct net_device *dev)
 {
 	struct bonding *bond = netdev_priv(dev);
-	struct slave *slave;
 	struct bond_up_slave *slaves;
-	unsigned int count;
+	struct slave *slave;
 
-	slaves = rcu_dereference(bond->slave_arr);
-	count = slaves ? READ_ONCE(slaves->count) : 0;
-	if (likely(count)) {
-		slave = slaves->arr[bond_xmit_hash(bond, skb) % count];
+	slaves = rcu_dereference(bond->usable_slaves);
+	slave = bond_xmit_3ad_xor_slave_get(bond, skb, slaves);
+	if (likely(slave))
 		return bond_dev_queue_xmit(bond, skb, slave->dev);
-	}
+
 	return bond_tx_drop(dev, skb);
 }
 
@@ -4269,6 +4340,48 @@ static u16 bond_select_queue(struct net_device *dev, struct sk_buff *skb,
 	return txq;
 }
 
+static struct net_device *bond_xmit_get_slave(struct net_device *master_dev,
+					      struct sk_buff *skb,
+					      bool all_slaves)
+{
+	struct bonding *bond = netdev_priv(master_dev);
+	struct bond_up_slave *slaves;
+	struct slave *slave = NULL;
+
+	switch (BOND_MODE(bond)) {
+	case BOND_MODE_ROUNDROBIN:
+		slave = bond_xmit_roundrobin_slave_get(bond, skb);
+		break;
+	case BOND_MODE_ACTIVEBACKUP:
+		slave = bond_xmit_activebackup_slave_get(bond, skb);
+		break;
+	case BOND_MODE_8023AD:
+	case BOND_MODE_XOR:
+		if (all_slaves)
+			slaves = rcu_dereference(bond->all_slaves);
+		else
+			slaves = rcu_dereference(bond->usable_slaves);
+		slave = bond_xmit_3ad_xor_slave_get(bond, skb, slaves);
+		break;
+	case BOND_MODE_BROADCAST:
+		break;
+	case BOND_MODE_ALB:
+		slave = bond_xmit_alb_slave_get(bond, skb);
+		break;
+	case BOND_MODE_TLB:
+		slave = bond_xmit_tlb_slave_get(bond, skb);
+		break;
+	default:
+		/* Should never happen, mode already checked */
+		WARN_ONCE(true, "Unknown bonding mode");
+		break;
+	}
+
+	if (slave)
+		return slave->dev;
+	return NULL;
+}
+
 static netdev_tx_t __bond_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct bonding *bond = netdev_priv(dev);
@@ -4389,6 +4502,7 @@ static const struct net_device_ops bond_netdev_ops = {
 	.ndo_del_slave		= bond_release,
 	.ndo_fix_features	= bond_fix_features,
 	.ndo_features_check	= passthru_features_check,
+	.ndo_get_xmit_slave	= bond_xmit_get_slave,
 };
 
 static const struct device_type bond_type = {
@@ -4456,9 +4570,9 @@ void bond_setup(struct net_device *bond_dev)
 static void bond_uninit(struct net_device *bond_dev)
 {
 	struct bonding *bond = netdev_priv(bond_dev);
+	struct bond_up_slave *usable, *all;
 	struct list_head *iter;
 	struct slave *slave;
-	struct bond_up_slave *arr;
 
 	bond_netpoll_cleanup(bond_dev);
 
@@ -4467,10 +4581,16 @@ static void bond_uninit(struct net_device *bond_dev)
 		__bond_release_one(bond_dev, slave->dev, true, true);
 	netdev_info(bond_dev, "Released all slaves\n");
 
-	arr = rtnl_dereference(bond->slave_arr);
-	if (arr) {
-		RCU_INIT_POINTER(bond->slave_arr, NULL);
-		kfree_rcu(arr, rcu);
+	usable = rtnl_dereference(bond->usable_slaves);
+	if (usable) {
+		RCU_INIT_POINTER(bond->usable_slaves, NULL);
+		kfree_rcu(usable, rcu);
+	}
+
+	all = rtnl_dereference(bond->all_slaves);
+	if (all) {
+		RCU_INIT_POINTER(bond->all_slaves, NULL);
+		kfree_rcu(all, rcu);
 	}
 
 	list_del(&bond->bond_list);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h b/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h
index c13260467750..82b185121edb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/accel.h
@@ -5,7 +5,6 @@
 
 #include <linux/skbuff.h>
 #include <linux/netdevice.h>
-#include "en.h"
 
 static inline bool is_metadata_hdr_valid(struct sk_buff *skb)
 {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 0864b76ca2c0..3bd64c63865b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -339,16 +339,6 @@ struct mlx5e_cq_decomp {
 	u16                        wqe_counter;
 } ____cacheline_aligned_in_smp;
 
-struct mlx5e_tx_wqe_info {
-	struct sk_buff *skb;
-	u32 num_bytes;
-	u8  num_wqebbs;
-	u8  num_dma;
-#ifdef CONFIG_MLX5_EN_TLS
-	struct page *resync_dump_frag_page;
-#endif
-};
-
 enum mlx5e_dma_map_type {
 	MLX5E_DMA_MAP_SINGLE,
 	MLX5E_DMA_MAP_PAGE
@@ -370,18 +360,6 @@ enum {
 	MLX5E_SQ_STATE_PENDING_XSK_TX,
 };
 
-struct mlx5e_icosq_wqe_info {
-	u8  opcode;
-	u8 num_wqebbs;
-
-	/* Auxiliary data for different opcodes. */
-	union {
-		struct {
-			struct mlx5e_rq *rq;
-		} umr;
-	};
-};
-
 struct mlx5e_txqsq {
 	/* data path */
 
@@ -484,11 +462,6 @@ struct mlx5e_xdp_info_fifo {
 	u32 mask;
 };
 
-struct mlx5e_xdp_wqe_info {
-	u8 num_wqebbs;
-	u8 num_pkts;
-};
-
 struct mlx5e_xdp_mpwqe {
 	/* Current MPWQE session */
 	struct mlx5e_tx_wqe *wqe;
@@ -919,8 +892,8 @@ void mlx5e_build_ptys2ethtool_map(void);
 u16 mlx5e_select_queue(struct net_device *dev, struct sk_buff *skb,
 		       struct net_device *sb_dev);
 netdev_tx_t mlx5e_xmit(struct sk_buff *skb, struct net_device *dev);
-netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
-			  struct mlx5e_tx_wqe *wqe, u16 pi, bool xmit_more);
+void mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+		   struct mlx5e_tx_wqe *wqe, u16 pi, bool xmit_more);
 
 void mlx5e_trigger_irq(struct mlx5e_icosq *sq);
 void mlx5e_completion_event(struct mlx5_core_cq *mcq, struct mlx5_eqe *eqe);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h b/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h
index 89fe65593c16..dce2bbbf9109 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h
@@ -27,6 +27,11 @@
 
 #define INL_HDR_START_SZ (sizeof(((struct mlx5_wqe_eth_seg *)NULL)->inline_hdr.start))
 
+enum mlx5e_icosq_wqe_type {
+	MLX5E_ICOSQ_WQE_NOP,
+	MLX5E_ICOSQ_WQE_UMR_RX,
+};
+
 static inline bool
 mlx5e_wqc_has_room_for(struct mlx5_wq_cyc *wq, u16 cc, u16 pc, u16 n)
 {
@@ -81,6 +86,16 @@ mlx5e_post_nop_fence(struct mlx5_wq_cyc *wq, u32 sqn, u16 *pc)
 	return wqe;
 }
 
+struct mlx5e_tx_wqe_info {
+	struct sk_buff *skb;
+	u32 num_bytes;
+	u8 num_wqebbs;
+	u8 num_dma;
+#ifdef CONFIG_MLX5_EN_TLS
+	struct page *resync_dump_frag_page;
+#endif
+};
+
 static inline u16 mlx5e_txqsq_get_next_pi(struct mlx5e_txqsq *sq, u16 size)
 {
 	struct mlx5_wq_cyc *wq = &sq->wq;
@@ -109,6 +124,18 @@ static inline u16 mlx5e_txqsq_get_next_pi(struct mlx5e_txqsq *sq, u16 size)
 	return pi;
 }
 
+struct mlx5e_icosq_wqe_info {
+	u8 wqe_type;
+	u8 num_wqebbs;
+
+	/* Auxiliary data for different wqe types. */
+	union {
+		struct {
+			struct mlx5e_rq *rq;
+		} umr;
+	};
+};
+
 static inline u16 mlx5e_icosq_get_next_pi(struct mlx5e_icosq *sq, u16 size)
 {
 	struct mlx5_wq_cyc *wq = &sq->wq;
@@ -125,7 +152,7 @@ static inline u16 mlx5e_icosq_get_next_pi(struct mlx5e_icosq *sq, u16 size)
 		/* Fill SQ frag edge with NOPs to avoid WQE wrapping two pages. */
 		for (; wi < edge_wi; wi++) {
 			*wi = (struct mlx5e_icosq_wqe_info) {
-				.opcode = MLX5_OPCODE_NOP,
+				.wqe_type   = MLX5E_ICOSQ_WQE_NOP,
 				.num_wqebbs = 1,
 			};
 			mlx5e_post_nop(wq, sq->sqn, &sq->pc);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
index ed6f045febeb..e2e01f064c1e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h
@@ -137,6 +137,11 @@ mlx5e_xdp_no_room_for_inline_pkt(struct mlx5e_xdp_mpwqe *session)
 	       session->ds_count + MLX5E_XDP_INLINE_WQE_MAX_DS_CNT > MLX5E_XDP_MPW_MAX_NUM_DS;
 }
 
+struct mlx5e_xdp_wqe_info {
+	u8 num_wqebbs;
+	u8 num_pkts;
+};
+
 static inline void
 mlx5e_xdp_mpwqe_add_dseg(struct mlx5e_xdpsq *sq,
 			 struct mlx5e_xdp_xmit_data *xdptxd,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
index a6f65d4b2f36..fac145dcf2ce 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h
@@ -102,33 +102,49 @@ mlx5e_udp_gso_handle_tx_skb(struct sk_buff *skb)
 	udp_hdr(skb)->len = htons(payload_len);
 }
 
-static inline struct sk_buff *
-mlx5e_accel_handle_tx(struct sk_buff *skb,
-		      struct mlx5e_txqsq *sq,
-		      struct net_device *dev,
-		      struct mlx5e_tx_wqe **wqe,
-		      u16 *pi)
+struct mlx5e_accel_tx_state {
+#ifdef CONFIG_MLX5_EN_TLS
+	struct mlx5e_accel_tx_tls_state tls;
+#endif
+};
+
+static inline bool mlx5e_accel_tx_begin(struct net_device *dev,
+					struct mlx5e_txqsq *sq,
+					struct sk_buff *skb,
+					struct mlx5e_accel_tx_state *state)
 {
+	if (skb_is_gso(skb) && skb_shinfo(skb)->gso_type & SKB_GSO_UDP_L4)
+		mlx5e_udp_gso_handle_tx_skb(skb);
+
 #ifdef CONFIG_MLX5_EN_TLS
 	if (test_bit(MLX5E_SQ_STATE_TLS, &sq->state)) {
-		skb = mlx5e_tls_handle_tx_skb(dev, sq, skb, wqe, pi);
-		if (unlikely(!skb))
-			return NULL;
+		/* May send SKBs and WQEs. */
+		if (unlikely(!mlx5e_tls_handle_tx_skb(dev, sq, skb, &state->tls)))
+			return false;
 	}
 #endif
 
+	return true;
+}
+
+static inline bool mlx5e_accel_tx_finish(struct mlx5e_priv *priv,
+					 struct mlx5e_txqsq *sq,
+					 struct sk_buff *skb,
+					 struct mlx5e_tx_wqe *wqe,
+					 struct mlx5e_accel_tx_state *state)
+{
+#ifdef CONFIG_MLX5_EN_TLS
+	mlx5e_tls_handle_tx_wqe(sq, &wqe->ctrl, &state->tls);
+#endif
+
 #ifdef CONFIG_MLX5_EN_IPSEC
 	if (test_bit(MLX5E_SQ_STATE_IPSEC, &sq->state)) {
-		skb = mlx5e_ipsec_handle_tx_skb(dev, *wqe, skb);
-		if (unlikely(!skb))
-			return NULL;
+		if (unlikely(!mlx5e_ipsec_handle_tx_skb(priv, &wqe->eth, skb)))
+			return false;
 	}
 #endif
 
-	if (skb_is_gso(skb) && skb_shinfo(skb)->gso_type & SKB_GSO_UDP_L4)
-		mlx5e_udp_gso_handle_tx_skb(skb);
-
-	return skb;
+	return true;
 }
 
 #endif /* __MLX5E_EN_ACCEL_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
index 0dd17514caae..824b87ac8f9e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c
@@ -233,11 +233,10 @@ static void mlx5e_ipsec_set_metadata(struct sk_buff *skb,
 		   ntohs(mdata->content.tx.seq));
 }
 
-struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev,
-					  struct mlx5e_tx_wqe *wqe,
-					  struct sk_buff *skb)
+bool mlx5e_ipsec_handle_tx_skb(struct mlx5e_priv *priv,
+			       struct mlx5_wqe_eth_seg *eseg,
+			       struct sk_buff *skb)
 {
-	struct mlx5e_priv *priv = netdev_priv(netdev);
 	struct xfrm_offload *xo = xfrm_offload(skb);
 	struct mlx5e_ipsec_metadata *mdata;
 	struct mlx5e_ipsec_sa_entry *sa_entry;
@@ -245,7 +244,7 @@ struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev,
 	struct sec_path *sp;
 
 	if (!xo)
-		return skb;
+		return true;
 
 	sp = skb_sec_path(skb);
 	if (unlikely(sp->len != 1)) {
@@ -276,16 +275,16 @@ struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev,
 		atomic64_inc(&priv->ipsec->sw_stats.ipsec_tx_drop_metadata);
 		goto drop;
 	}
-	mlx5e_ipsec_set_swp(skb, &wqe->eth, x->props.mode, xo);
+	mlx5e_ipsec_set_swp(skb, eseg, x->props.mode, xo);
 	sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle;
 	sa_entry->set_iv_op(skb, x, xo);
 	mlx5e_ipsec_set_metadata(skb, mdata, xo);
 
-	return skb;
+	return true;
 
 drop:
 	kfree_skb(skb);
-	return NULL;
+	return false;
 }
 
 static inline struct xfrm_state *
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
index db84500b024f..ba02643586a5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h
@@ -52,9 +52,9 @@ void mlx5e_ipsec_set_iv_esn(struct sk_buff *skb, struct xfrm_state *x,
 			    struct xfrm_offload *xo);
 void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_state *x,
 			struct xfrm_offload *xo);
-struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev,
-					  struct mlx5e_tx_wqe *wqe,
-					  struct sk_buff *skb);
+bool mlx5e_ipsec_handle_tx_skb(struct mlx5e_priv *priv,
+			       struct mlx5_wqe_eth_seg *eseg,
+			       struct sk_buff *skb);
 
 #endif /* CONFIG_MLX5_EN_IPSEC */
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h
index 9daaec244385..dabbc5f226ce 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.h
@@ -9,6 +9,7 @@
 #ifdef CONFIG_MLX5_EN_TLS
 #include <net/tls.h>
 #include "accel/tls.h"
+#include "en_accel/tls_rxtx.h"
 
 #define MLX5E_KTLS_STATIC_UMR_WQE_SZ \
 	(offsetof(struct mlx5e_umr_wqe, tls_static_params_ctx) + \
@@ -95,10 +96,9 @@ mlx5e_get_ktls_tx_priv_ctx(struct tls_context *tls_ctx)
 void mlx5e_ktls_build_netdev(struct mlx5e_priv *priv);
 void mlx5e_ktls_tx_offload_set_pending(struct mlx5e_ktls_offload_context_tx *priv_tx);
 
-struct sk_buff *mlx5e_ktls_handle_tx_skb(struct net_device *netdev,
-					 struct mlx5e_txqsq *sq,
-					 struct sk_buff *skb,
-					 struct mlx5e_tx_wqe **wqe, u16 *pi);
+bool mlx5e_ktls_handle_tx_skb(struct tls_context *tls_ctx, struct mlx5e_txqsq *sq,
+			      struct sk_buff *skb, int datalen,
+			      struct mlx5e_accel_tx_tls_state *state);
 void mlx5e_ktls_tx_handle_resync_dump_comp(struct mlx5e_txqsq *sq,
 					   struct mlx5e_tx_wqe_info *wi,
 					   u32 *dma_fifo_cc);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
index ba973937f0b5..3cd78d9503c1 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c
@@ -108,10 +108,11 @@ static void tx_fill_wi(struct mlx5e_txqsq *sq,
 {
 	struct mlx5e_tx_wqe_info *wi = &sq->db.wqe_info[pi];
 
-	memset(wi, 0, sizeof(*wi));
-	wi->num_wqebbs = num_wqebbs;
-	wi->num_bytes  = num_bytes;
-	wi->resync_dump_frag_page = page;
+	*wi = (struct mlx5e_tx_wqe_info) {
+		.num_wqebbs = num_wqebbs,
+		.num_bytes  = num_bytes,
+		.resync_dump_frag_page = page,
+	};
 }
 
 void mlx5e_ktls_tx_offload_set_pending(struct mlx5e_ktls_offload_context_tx *priv_tx)
@@ -134,14 +135,14 @@ post_static_params(struct mlx5e_txqsq *sq,
 		   struct mlx5e_ktls_offload_context_tx *priv_tx,
 		   bool fence)
 {
+	u16 pi, num_wqebbs = MLX5E_KTLS_STATIC_WQEBBS;
 	struct mlx5e_umr_wqe *umr_wqe;
-	u16 pi;
 
-	pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
+	pi = mlx5e_txqsq_get_next_pi(sq, num_wqebbs);
 	umr_wqe = MLX5E_TLS_FETCH_UMR_WQE(sq, pi);
 	build_static_params(umr_wqe, sq->pc, sq->sqn, priv_tx, fence);
-	tx_fill_wi(sq, pi, MLX5E_KTLS_STATIC_WQEBBS, 0, NULL);
-	sq->pc += MLX5E_KTLS_STATIC_WQEBBS;
+	tx_fill_wi(sq, pi, num_wqebbs, 0, NULL);
+	sq->pc += num_wqebbs;
 }
 
 static void
@@ -149,14 +150,14 @@ post_progress_params(struct mlx5e_txqsq *sq,
 		     struct mlx5e_ktls_offload_context_tx *priv_tx,
 		     bool fence)
 {
+	u16 pi, num_wqebbs = MLX5E_KTLS_PROGRESS_WQEBBS;
 	struct mlx5e_tx_wqe *wqe;
-	u16 pi;
 
-	pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
+	pi = mlx5e_txqsq_get_next_pi(sq, num_wqebbs);
 	wqe = MLX5E_TLS_FETCH_PROGRESS_WQE(sq, pi);
 	build_progress_params(wqe, sq->pc, sq->sqn, priv_tx, fence);
-	tx_fill_wi(sq, pi, MLX5E_KTLS_PROGRESS_WQEBBS, 0, NULL);
-	sq->pc += MLX5E_KTLS_PROGRESS_WQEBBS;
+	tx_fill_wi(sq, pi, num_wqebbs, 0, NULL);
+	sq->pc += num_wqebbs;
 }
 
 static void
@@ -166,8 +167,6 @@ mlx5e_ktls_tx_post_param_wqes(struct mlx5e_txqsq *sq,
 {
 	bool progress_fence = skip_static_post || !fence_first_post;
 
-	mlx5e_txqsq_get_next_pi(sq, MLX5E_KTLS_STATIC_WQEBBS + MLX5E_KTLS_PROGRESS_WQEBBS);
-
 	if (!skip_static_post)
 		post_static_params(sq, priv_tx, fence_first_post);
 
@@ -274,6 +273,7 @@ tx_post_resync_dump(struct mlx5e_txqsq *sq, skb_frag_t *frag, u32 tisn, bool fir
 	int fsz;
 	u16 pi;
 
+	BUILD_BUG_ON(MLX5E_KTLS_DUMP_WQEBBS != 1);
 	pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
 	wqe = MLX5E_TLS_FETCH_DUMP_WQE(sq, pi);
 
@@ -342,7 +342,6 @@ mlx5e_ktls_tx_handle_ooo(struct mlx5e_ktls_offload_context_tx *priv_tx,
 	struct mlx5e_sq_stats *stats = sq->stats;
 	enum mlx5e_ktls_sync_retval ret;
 	struct tx_sync_info info = {};
-	u8 num_wqebbs;
 	int i = 0;
 
 	ret = tx_sync_info_get(priv_tx, seq, datalen, &info);
@@ -371,9 +370,6 @@ mlx5e_ktls_tx_handle_ooo(struct mlx5e_ktls_offload_context_tx *priv_tx,
 		return MLX5E_KTLS_SYNC_DONE;
 	}
 
-	num_wqebbs = mlx5e_ktls_dumps_num_wqebbs(sq, info.nr_frags, info.sync_len);
-	mlx5e_txqsq_get_next_pi(sq, num_wqebbs);
-
 	for (; i < info.nr_frags; i++) {
 		unsigned int orig_fsz, frag_offset = 0, n = 0;
 		skb_frag_t *f = &info.frags[i];
@@ -413,35 +409,18 @@ err_out:
 	return MLX5E_KTLS_SYNC_FAIL;
 }
 
-struct sk_buff *mlx5e_ktls_handle_tx_skb(struct net_device *netdev,
-					 struct mlx5e_txqsq *sq,
-					 struct sk_buff *skb,
-					 struct mlx5e_tx_wqe **wqe, u16 *pi)
+bool mlx5e_ktls_handle_tx_skb(struct tls_context *tls_ctx, struct mlx5e_txqsq *sq,
+			      struct sk_buff *skb, int datalen,
+			      struct mlx5e_accel_tx_tls_state *state)
 {
 	struct mlx5e_ktls_offload_context_tx *priv_tx;
 	struct mlx5e_sq_stats *stats = sq->stats;
-	struct mlx5_wqe_ctrl_seg *cseg;
-	struct tls_context *tls_ctx;
-	int datalen;
 	u32 seq;
 
-	if (!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk))
-		goto out;
-
-	datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb));
-	if (!datalen)
-		goto out;
-
-	tls_ctx = tls_get_ctx(skb->sk);
-	if (WARN_ON_ONCE(tls_ctx->netdev != netdev))
-		goto err_out;
-
 	priv_tx = mlx5e_get_ktls_tx_priv_ctx(tls_ctx);
 
 	if (unlikely(mlx5e_ktls_tx_offload_test_and_clear_pending(priv_tx))) {
 		mlx5e_ktls_tx_post_param_wqes(sq, priv_tx, false, false);
-		*pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
-		*wqe = MLX5E_TX_FETCH_WQE(sq, *pi);
 		stats->tls_ctx++;
 	}
 
@@ -452,31 +431,28 @@ struct sk_buff *mlx5e_ktls_handle_tx_skb(struct net_device *netdev,
 
 		switch (ret) {
 		case MLX5E_KTLS_SYNC_DONE:
-			*pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
-			*wqe = MLX5E_TX_FETCH_WQE(sq, *pi);
 			break;
 		case MLX5E_KTLS_SYNC_SKIP_NO_DATA:
 			if (likely(!skb->decrypted))
 				goto out;
 			WARN_ON_ONCE(1);
 			/* fall-through */
-		default: /* MLX5E_KTLS_SYNC_FAIL */
+		case MLX5E_KTLS_SYNC_FAIL:
 			goto err_out;
 		}
 	}
 
 	priv_tx->expected_seq = seq + datalen;
 
-	cseg = &(*wqe)->ctrl;
-	cseg->tisn = cpu_to_be32(priv_tx->tisn << 8);
+	state->tls_tisn = priv_tx->tisn;
 
 	stats->tls_encrypted_packets += skb_is_gso(skb) ? skb_shinfo(skb)->gso_segs : 1;
 	stats->tls_encrypted_bytes   += datalen;
 
 out:
-	return skb;
+	return true;
 
 err_out:
 	dev_kfree_skb_any(skb);
-	return NULL;
+	return false;
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
index 1d7ddeb7a46b..05454a843b28 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.c
@@ -184,18 +184,17 @@ static void mlx5e_tls_complete_sync_skb(struct sk_buff *skb,
 	nskb->queue_mapping = skb->queue_mapping;
 }
 
-static struct sk_buff *
-mlx5e_tls_handle_ooo(struct mlx5e_tls_offload_context_tx *context,
-		     struct mlx5e_txqsq *sq, struct sk_buff *skb,
-		     struct mlx5e_tx_wqe **wqe,
-		     u16 *pi,
-		     struct mlx5e_tls *tls)
+static bool mlx5e_tls_handle_ooo(struct mlx5e_tls_offload_context_tx *context,
+				 struct mlx5e_txqsq *sq, struct sk_buff *skb,
+				 struct mlx5e_tls *tls)
 {
 	u32 tcp_seq = ntohl(tcp_hdr(skb)->seq);
+	struct mlx5e_tx_wqe *wqe;
 	struct sync_info info;
 	struct sk_buff *nskb;
 	int linear_len = 0;
 	int headln;
+	u16 pi;
 	int i;
 
 	sq->stats->tls_ooo++;
@@ -217,7 +216,7 @@ mlx5e_tls_handle_ooo(struct mlx5e_tls_offload_context_tx *context,
 		if (likely(payload <= -info.sync_len))
 			/* SKB payload doesn't require offload
 			 */
-			return skb;
+			return true;
 
 		atomic64_inc(&tls->sw_stats.tx_tls_drop_bypass_required);
 		goto err_out;
@@ -247,21 +246,19 @@ mlx5e_tls_handle_ooo(struct mlx5e_tls_offload_context_tx *context,
 	sq->stats->tls_resync_bytes += nskb->len;
 	mlx5e_tls_complete_sync_skb(skb, nskb, tcp_seq, headln,
 				    cpu_to_be64(info.rcd_sn));
-	mlx5e_sq_xmit(sq, nskb, *wqe, *pi, true);
-	*pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
-	*wqe = MLX5E_TX_FETCH_WQE(sq, *pi);
-	return skb;
+	pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
+	wqe = MLX5E_TX_FETCH_WQE(sq, pi);
+	mlx5e_sq_xmit(sq, nskb, wqe, pi, true);
+
+	return true;
 
 err_out:
 	dev_kfree_skb_any(skb);
-	return NULL;
+	return false;
 }
 
-struct sk_buff *mlx5e_tls_handle_tx_skb(struct net_device *netdev,
-					struct mlx5e_txqsq *sq,
-					struct sk_buff *skb,
-					struct mlx5e_tx_wqe **wqe,
-					u16 *pi)
+bool mlx5e_tls_handle_tx_skb(struct net_device *netdev, struct mlx5e_txqsq *sq,
+			     struct sk_buff *skb, struct mlx5e_accel_tx_tls_state *state)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
 	struct mlx5e_tls_offload_context_tx *context;
@@ -270,41 +267,45 @@ struct sk_buff *mlx5e_tls_handle_tx_skb(struct net_device *netdev,
 	int datalen;
 	u32 skb_seq;
 
-	if (MLX5_CAP_GEN(sq->channel->mdev, tls_tx)) {
-		skb = mlx5e_ktls_handle_tx_skb(netdev, sq, skb, wqe, pi);
-		goto out;
-	}
-
 	if (!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk))
-		goto out;
+		return true;
 
 	datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb));
 	if (!datalen)
-		goto out;
+		return true;
 
 	tls_ctx = tls_get_ctx(skb->sk);
-	if (unlikely(tls_ctx->netdev != netdev))
-		goto out;
+	if (WARN_ON_ONCE(tls_ctx->netdev != netdev))
+		goto err_out;
+
+	if (MLX5_CAP_GEN(sq->channel->mdev, tls_tx))
+		return mlx5e_ktls_handle_tx_skb(tls_ctx, sq, skb, datalen, state);
 
 	skb_seq = ntohl(tcp_hdr(skb)->seq);
 	context = mlx5e_get_tls_tx_context(tls_ctx);
 	expected_seq = context->expected_seq;
 
-	if (unlikely(expected_seq != skb_seq)) {
-		skb = mlx5e_tls_handle_ooo(context, sq, skb, wqe, pi, priv->tls);
-		goto out;
-	}
+	if (unlikely(expected_seq != skb_seq))
+		return mlx5e_tls_handle_ooo(context, sq, skb, priv->tls);
 
 	if (unlikely(mlx5e_tls_add_metadata(skb, context->swid))) {
 		atomic64_inc(&priv->tls->sw_stats.tx_tls_drop_metadata);
 		dev_kfree_skb_any(skb);
-		skb = NULL;
-		goto out;
+		return false;
 	}
 
 	context->expected_seq = skb_seq + datalen;
-out:
-	return skb;
+	return true;
+
+err_out:
+	dev_kfree_skb_any(skb);
+	return false;
+}
+
+void mlx5e_tls_handle_tx_wqe(struct mlx5e_txqsq *sq, struct mlx5_wqe_ctrl_seg *cseg,
+			     struct mlx5e_accel_tx_tls_state *state)
+{
+	cseg->tisn = cpu_to_be32(state->tls_tisn << 8);
 }
 
 static int tls_update_resync_sn(struct net_device *netdev,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
index 90bc1f2384c8..a50d0394df0a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls_rxtx.h
@@ -40,11 +40,14 @@
 #include "en.h"
 #include "en/txrx.h"
 
-struct sk_buff *mlx5e_tls_handle_tx_skb(struct net_device *netdev,
-					struct mlx5e_txqsq *sq,
-					struct sk_buff *skb,
-					struct mlx5e_tx_wqe **wqe,
-					u16 *pi);
+struct mlx5e_accel_tx_tls_state {
+	u32 tls_tisn;
+};
+
+bool mlx5e_tls_handle_tx_skb(struct net_device *netdev, struct mlx5e_txqsq *sq,
+			     struct sk_buff *skb, struct mlx5e_accel_tx_tls_state *state);
+void mlx5e_tls_handle_tx_wqe(struct mlx5e_txqsq *sq, struct mlx5_wqe_ctrl_seg *cseg,
+			     struct mlx5e_accel_tx_tls_state *state);
 
 void mlx5e_tls_handle_rx_skb(struct net_device *netdev, struct sk_buff *skb,
 			     u32 *cqe_bcnt);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 048a4f8601a8..0a9dfc31de3e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1364,13 +1364,12 @@ static void mlx5e_deactivate_txqsq(struct mlx5e_txqsq *sq)
 	/* last doorbell out, godspeed .. */
 	if (mlx5e_wqc_has_room_for(wq, sq->cc, sq->pc, 1)) {
 		u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
-		struct mlx5e_tx_wqe_info *wi;
 		struct mlx5e_tx_wqe *nop;
 
-		wi = &sq->db.wqe_info[pi];
+		sq->db.wqe_info[pi] = (struct mlx5e_tx_wqe_info) {
+			.num_wqebbs = 1,
+		};
 
-		memset(wi, 0, sizeof(*wi));
-		wi->num_wqebbs = 1;
 		nop = mlx5e_post_nop(wq, sq->sqn, &sq->pc);
 		mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &nop->ctrl);
 	}
@@ -1482,20 +1481,21 @@ int mlx5e_open_xdpsq(struct mlx5e_channel *c, struct mlx5e_params *params,
 
 		/* Pre initialize fixed WQE fields */
 		for (i = 0; i < mlx5_wq_cyc_get_size(&sq->wq); i++) {
-			struct mlx5e_xdp_wqe_info *wi  = &sq->db.wqe_info[i];
 			struct mlx5e_tx_wqe      *wqe  = mlx5_wq_cyc_get_wqe(&sq->wq, i);
 			struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
 			struct mlx5_wqe_eth_seg  *eseg = &wqe->eth;
 			struct mlx5_wqe_data_seg *dseg;
 
+			sq->db.wqe_info[i] = (struct mlx5e_xdp_wqe_info) {
+				.num_wqebbs = 1,
+				.num_pkts   = 1,
+			};
+
 			cseg->qpn_ds = cpu_to_be32((sq->sqn << 8) | ds_cnt);
 			eseg->inline_hdr.sz = cpu_to_be16(inline_hdr_sz);
 
 			dseg = (struct mlx5_wqe_data_seg *)cseg + (ds_cnt - 1);
 			dseg->lkey = sq->mkey_be;
-
-			wi->num_wqebbs = 1;
-			wi->num_pkts   = 1;
 		}
 	}
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index d9a5a669b84d..779600bebcca 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -505,9 +505,12 @@ static int mlx5e_alloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix)
 			    MLX5_OPCODE_UMR);
 	umr_wqe->uctrl.xlt_offset = cpu_to_be16(xlt_offset);
 
-	sq->db.wqe_info[pi].opcode = MLX5_OPCODE_UMR;
-	sq->db.wqe_info[pi].num_wqebbs = MLX5E_UMR_WQEBBS;
-	sq->db.wqe_info[pi].umr.rq = rq;
+	sq->db.wqe_info[pi] = (struct mlx5e_icosq_wqe_info) {
+		.wqe_type   = MLX5E_ICOSQ_WQE_UMR_RX,
+		.num_wqebbs = MLX5E_UMR_WQEBBS,
+		.umr.rq     = rq,
+	};
+
 	sq->pc += MLX5E_UMR_WQEBBS;
 
 	sq->doorbell_cseg = &umr_wqe->ctrl;
@@ -616,15 +619,18 @@ int mlx5e_poll_ico_cq(struct mlx5e_cq *cq)
 				break;
 			}
 
-			if (likely(wi->opcode == MLX5_OPCODE_UMR))
+			switch (wi->wqe_type) {
+			case MLX5E_ICOSQ_WQE_UMR_RX:
 				wi->umr.rq->mpwqe.umr_completed++;
-			else if (unlikely(wi->opcode != MLX5_OPCODE_NOP))
+				break;
+			case MLX5E_ICOSQ_WQE_NOP:
+				break;
+			default:
 				netdev_WARN_ONCE(cq->channel->netdev,
-						 "Bad OPCODE in ICOSQ WQE info: 0x%x\n",
-						 wi->opcode);
-
+						 "Bad WQE type in ICOSQ WQE info: 0x%x\n",
+						 wi->wqe_type);
+			}
 		} while (!last_wqe);
-
 	} while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
 
 	sq->cc = sqcc;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
index 583e1b201b75..f79454746d0d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -265,8 +265,8 @@ mlx5e_txwqe_complete(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 		mlx5e_notify_hw(wq, sq->pc, sq->uar_map, cseg);
 }
 
-netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
-			  struct mlx5e_tx_wqe *wqe, u16 pi, bool xmit_more)
+void mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+		   struct mlx5e_tx_wqe *wqe, u16 pi, bool xmit_more)
 {
 	struct mlx5_wq_cyc *wq = &sq->wq;
 	struct mlx5_wqe_ctrl_seg *cseg;
@@ -373,32 +373,38 @@ netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 	mlx5e_txwqe_complete(sq, skb, opcode, ds_cnt, num_wqebbs, num_bytes,
 			     num_dma, wi, cseg, xmit_more);
 
-	return NETDEV_TX_OK;
+	return;
 
 err_drop:
 	stats->dropped++;
 	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_accel_tx_state accel = {};
 	struct mlx5e_tx_wqe *wqe;
 	struct mlx5e_txqsq *sq;
 	u16 pi;
 
 	sq = priv->txq2sq[skb_get_queue_mapping(skb)];
+
+	/* May send SKBs and WQEs. */
+	if (unlikely(!mlx5e_accel_tx_begin(dev, sq, skb, &accel)))
+		goto out;
+
 	pi = mlx5_wq_cyc_ctr2ix(&sq->wq, sq->pc);
 	wqe = MLX5E_TX_FETCH_WQE(sq, pi);
 
-	/* might send skbs and update wqe and pi */
-	skb = mlx5e_accel_handle_tx(skb, sq, dev, &wqe, &pi);
-	if (unlikely(!skb))
-		return NETDEV_TX_OK;
+	/* May update the WQE, but may not post other WQEs. */
+	if (unlikely(!mlx5e_accel_tx_finish(priv, sq, skb, wqe, &accel)))
+		goto out;
 
-	return mlx5e_sq_xmit(sq, skb, wqe, pi, netdev_xmit_more());
+	mlx5e_sq_xmit(sq, skb, wqe, pi, netdev_xmit_more());
+
+out:
+	return NETDEV_TX_OK;
 }
 
 bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq, int napi_budget)
@@ -568,9 +574,8 @@ mlx5i_txwqe_build_datagram(struct mlx5_av *av, u32 dqpn, u32 dqkey,
 	dseg->av.key.qkey.qkey = cpu_to_be32(dqkey);
 }
 
-netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
-			  struct mlx5_av *av, u32 dqpn, u32 dqkey,
-			  bool xmit_more)
+void mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+		   struct mlx5_av *av, u32 dqpn, u32 dqkey, bool xmit_more)
 {
 	struct mlx5i_tx_wqe *wqe;
 
@@ -648,12 +653,10 @@ netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 	mlx5e_txwqe_complete(sq, skb, opcode, ds_cnt, num_wqebbs, num_bytes,
 			     num_dma, wi, cseg, xmit_more);
 
-	return NETDEV_TX_OK;
+	return;
 
 err_drop:
 	stats->dropped++;
 	dev_kfree_skb_any(skb);
-
-	return NETDEV_TX_OK;
 }
 #endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
index 869fd58a6775..8480278f2ee2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -78,8 +78,11 @@ void mlx5e_trigger_irq(struct mlx5e_icosq *sq)
 	struct mlx5e_tx_wqe *nopwqe;
 	u16 pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
 
-	sq->db.wqe_info[pi].opcode = MLX5_OPCODE_NOP;
-	sq->db.wqe_info[pi].num_wqebbs = 1;
+	sq->db.wqe_info[pi] = (struct mlx5e_icosq_wqe_info) {
+		.wqe_type   = MLX5E_ICOSQ_WQE_NOP,
+		.num_wqebbs = 1,
+	};
+
 	nopwqe = mlx5e_post_nop(wq, sq->sqn, &sq->pc);
 	mlx5e_notify_hw(wq, sq->pc, sq->uar_map, &nopwqe->ctrl);
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 068578be00f1..035bd21e5d4e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -663,7 +663,9 @@ static int mlx5i_xmit(struct net_device *dev, struct sk_buff *skb,
 	struct mlx5_ib_ah *mah   = to_mah(address);
 	struct mlx5i_priv *ipriv = epriv->ppriv;
 
-	return mlx5i_sq_xmit(sq, skb, &mah->av, dqpn, ipriv->qkey, netdev_xmit_more());
+	mlx5i_sq_xmit(sq, skb, &mah->av, dqpn, ipriv->qkey, netdev_xmit_more());
+
+	return NETDEV_TX_OK;
 }
 
 static void mlx5i_set_pkey_index(struct net_device *netdev, int id)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
index 7844ab5d0ce7..c4aa47018c0e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.h
@@ -113,9 +113,8 @@ struct mlx5i_tx_wqe {
 #define MLX5I_SQ_FETCH_WQE(sq, pi) \
 	((struct mlx5i_tx_wqe *)mlx5e_fetch_wqe(&(sq)->wq, pi, sizeof(struct mlx5i_tx_wqe)))
 
-netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
-			  struct mlx5_av *av, u32 dqpn, u32 dqkey,
-			  bool xmit_more);
+void mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
+		   struct mlx5_av *av, u32 dqpn, u32 dqkey, bool xmit_more);
 void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe);
 void mlx5i_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats);
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index c6ad5ca46877..874c70e8cc54 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -42,7 +42,7 @@
  * Beware of lock dependencies (preferably, no locks should be acquired
  * under it).
  */
-static DEFINE_MUTEX(lag_mutex);
+static DEFINE_SPINLOCK(lag_lock);
 
 static int mlx5_cmd_create_lag(struct mlx5_core_dev *dev, u8 remap_port1,
 			       u8 remap_port2)
@@ -274,9 +274,9 @@ static void mlx5_do_bond(struct mlx5_lag *ldev)
 	if (!dev0 || !dev1)
 		return;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	tracker = ldev->tracker;
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 
 	do_bond = tracker.is_bonded && mlx5_lag_check_prereq(ldev);
 
@@ -458,9 +458,9 @@ static int mlx5_lag_netdev_event(struct notifier_block *this,
 		break;
 	}
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev->tracker = tracker;
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 
 	if (changed)
 		mlx5_queue_bond_work(ldev, 0);
@@ -502,7 +502,7 @@ static void mlx5_lag_dev_add_pf(struct mlx5_lag *ldev,
 	if (fn >= MLX5_MAX_PORTS)
 		return;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev->pf[fn].dev    = dev;
 	ldev->pf[fn].netdev = netdev;
 	ldev->tracker.netdev_state[fn].link_up = 0;
@@ -510,7 +510,7 @@ static void mlx5_lag_dev_add_pf(struct mlx5_lag *ldev,
 
 	dev->priv.lag = ldev;
 
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 }
 
 static void mlx5_lag_dev_remove_pf(struct mlx5_lag *ldev,
@@ -525,11 +525,11 @@ static void mlx5_lag_dev_remove_pf(struct mlx5_lag *ldev,
 	if (i == MLX5_MAX_PORTS)
 		return;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	memset(&ldev->pf[i], 0, sizeof(*ldev->pf));
 
 	dev->priv.lag = NULL;
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 }
 
 /* Must be called with intf_mutex held */
@@ -607,10 +607,10 @@ bool mlx5_lag_is_roce(struct mlx5_core_dev *dev)
 	struct mlx5_lag *ldev;
 	bool res;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev = mlx5_lag_dev_get(dev);
 	res  = ldev && __mlx5_lag_is_roce(ldev);
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 
 	return res;
 }
@@ -621,10 +621,10 @@ bool mlx5_lag_is_active(struct mlx5_core_dev *dev)
 	struct mlx5_lag *ldev;
 	bool res;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev = mlx5_lag_dev_get(dev);
 	res  = ldev && __mlx5_lag_is_active(ldev);
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 
 	return res;
 }
@@ -635,10 +635,10 @@ bool mlx5_lag_is_sriov(struct mlx5_core_dev *dev)
 	struct mlx5_lag *ldev;
 	bool res;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev = mlx5_lag_dev_get(dev);
 	res  = ldev && __mlx5_lag_is_sriov(ldev);
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 
 	return res;
 }
@@ -664,7 +664,7 @@ struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev)
 	struct net_device *ndev = NULL;
 	struct mlx5_lag *ldev;
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev = mlx5_lag_dev_get(dev);
 
 	if (!(ldev && __mlx5_lag_is_roce(ldev)))
@@ -681,12 +681,36 @@ struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev)
 		dev_hold(ndev);
 
 unlock:
-	mutex_unlock(&lag_mutex);
+	spin_unlock(&lag_lock);
 
 	return ndev;
 }
 EXPORT_SYMBOL(mlx5_lag_get_roce_netdev);
 
+u8 mlx5_lag_get_slave_port(struct mlx5_core_dev *dev,
+			   struct net_device *slave)
+{
+	struct mlx5_lag *ldev;
+	u8 port = 0;
+
+	spin_lock(&lag_lock);
+	ldev = mlx5_lag_dev_get(dev);
+	if (!(ldev && __mlx5_lag_is_roce(ldev)))
+		goto unlock;
+
+	if (ldev->pf[MLX5_LAG_P1].netdev == slave)
+		port = MLX5_LAG_P1;
+	else
+		port = MLX5_LAG_P2;
+
+	port = ldev->v2p_map[port];
+
+unlock:
+	spin_unlock(&lag_lock);
+	return port;
+}
+EXPORT_SYMBOL(mlx5_lag_get_slave_port);
+
 bool mlx5_lag_intf_add(struct mlx5_interface *intf, struct mlx5_priv *priv)
 {
 	struct mlx5_core_dev *dev = container_of(priv, struct mlx5_core_dev,
@@ -723,7 +747,7 @@ int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
 
 	memset(values, 0, sizeof(*values) * num_counters);
 
-	mutex_lock(&lag_mutex);
+	spin_lock(&lag_lock);
 	ldev = mlx5_lag_dev_get(dev);
 	if (ldev && __mlx5_lag_is_roce(ldev)) {
 		num_ports = MLX5_MAX_PORTS;
@@ -733,6 +757,7 @@ int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
 		num_ports = 1;
 		mdev[MLX5_LAG_P1] = dev;
 	}
+	spin_unlock(&lag_lock);
 
 	for (i = 0; i < num_ports; ++i) {
 		u32 in[MLX5_ST_SZ_DW(query_cong_statistics_in)] = {};
@@ -742,14 +767,13 @@ int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
 		ret = mlx5_cmd_exec_inout(mdev[i], query_cong_statistics, in,
 					  out);
 		if (ret)
-			goto unlock;
+			goto free;
 
 		for (j = 0; j < num_counters; ++j)
 			values[j] += be64_to_cpup((__be64 *)(out + offsets[j]));
 	}
 
-unlock:
-	mutex_unlock(&lag_mutex);
+free:
 	kvfree(out);
 	return ret;
 }
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index d82dbbab8179..267dfcc5493e 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -1074,6 +1074,8 @@ bool mlx5_lag_is_sriov(struct mlx5_core_dev *dev);
 bool mlx5_lag_is_multipath(struct mlx5_core_dev *dev);
 bool mlx5_lag_is_active(struct mlx5_core_dev *dev);
 struct net_device *mlx5_lag_get_roce_netdev(struct mlx5_core_dev *dev);
+u8 mlx5_lag_get_slave_port(struct mlx5_core_dev *dev,
+			   struct net_device *slave);
 int mlx5_lag_query_cong_counters(struct mlx5_core_dev *dev,
 				 u64 *values,
 				 int num_counters,
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 7725efd6e48a..6a8f8daef09d 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1148,6 +1148,12 @@ struct netdev_net_notifier {
  * int (*ndo_del_slave)(struct net_device *dev, struct net_device *slave_dev);
  *	Called to release previously enslaved netdev.
  *
+ * struct net_device *(*ndo_get_xmit_slave)(struct net_device *dev,
+ *					    struct sk_buff *skb,
+ *					    bool all_slaves);
+ *	Get the xmit slave of master device. If all_slaves is true, function
+ *	assume all the slaves can transmit.
+ *
  *      Feature/offload setting functions.
  * netdev_features_t (*ndo_fix_features)(struct net_device *dev,
  *		netdev_features_t features);
@@ -1391,6 +1397,9 @@ struct net_device_ops {
 						 struct netlink_ext_ack *extack);
 	int			(*ndo_del_slave)(struct net_device *dev,
 						 struct net_device *slave_dev);
+	struct net_device*	(*ndo_get_xmit_slave)(struct net_device *dev,
+						      struct sk_buff *skb,
+						      bool all_slaves);
 	netdev_features_t	(*ndo_fix_features)(struct net_device *dev,
 						    netdev_features_t features);
 	int			(*ndo_set_features)(struct net_device *dev,
@@ -2745,6 +2754,9 @@ void netdev_freemem(struct net_device *dev);
 void synchronize_net(void);
 int init_dummy_netdev(struct net_device *dev);
 
+struct net_device *netdev_get_xmit_slave(struct net_device *dev,
+					 struct sk_buff *skb,
+					 bool all_slaves);
 struct net_device *dev_get_by_index(struct net *net, int ifindex);
 struct net_device *__dev_get_by_index(struct net *net, int ifindex);
 struct net_device *dev_get_by_index_rcu(struct net *net, int ifindex);
diff --git a/include/net/bond_alb.h b/include/net/bond_alb.h
index b3504fcd773d..f6af76c87a6c 100644
--- a/include/net/bond_alb.h
+++ b/include/net/bond_alb.h
@@ -158,6 +158,10 @@ void bond_alb_handle_link_change(struct bonding *bond, struct slave *slave, char
 void bond_alb_handle_active_change(struct bonding *bond, struct slave *new_slave);
 int bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev);
 int bond_tlb_xmit(struct sk_buff *skb, struct net_device *bond_dev);
+struct slave *bond_xmit_alb_slave_get(struct bonding *bond,
+				      struct sk_buff *skb);
+struct slave *bond_xmit_tlb_slave_get(struct bonding *bond,
+				      struct sk_buff *skb);
 void bond_alb_monitor(struct work_struct *);
 int bond_alb_set_mac_address(struct net_device *bond_dev, void *addr);
 void bond_alb_clear_vlan(struct bonding *bond, unsigned short vlan_id);
diff --git a/include/net/bonding.h b/include/net/bonding.h
index 9b1e76515a9c..aa854a9c01e2 100644
--- a/include/net/bonding.h
+++ b/include/net/bonding.h
@@ -200,7 +200,8 @@ struct bonding {
 	struct   slave __rcu *curr_active_slave;
 	struct   slave __rcu *current_arp_slave;
 	struct   slave __rcu *primary_slave;
-	struct   bond_up_slave __rcu *slave_arr; /* Array of usable slaves */
+	struct   bond_up_slave __rcu *usable_slaves;
+	struct   bond_up_slave __rcu *all_slaves;
 	bool     force_primary;
 	s32      slave_cnt; /* never change this value outside the attach/detach wrappers */
 	int     (*recv_probe)(const struct sk_buff *, struct bonding *,
diff --git a/net/core/dev.c b/net/core/dev.c
index f8d83922a6af..4c91de39890a 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -7861,6 +7861,28 @@ void netdev_bonding_info_change(struct net_device *dev,
 }
 EXPORT_SYMBOL(netdev_bonding_info_change);
 
+/**
+ * netdev_get_xmit_slave - Get the xmit slave of master device
+ * @skb: The packet
+ * @all_slaves: assume all the slaves are active
+ *
+ * The reference counters are not incremented so the caller must be
+ * careful with locks. The caller must hold RCU lock.
+ * %NULL is returned if no slave is found.
+ */
+
+struct net_device *netdev_get_xmit_slave(struct net_device *dev,
+					 struct sk_buff *skb,
+					 bool all_slaves)
+{
+	const struct net_device_ops *ops = dev->netdev_ops;
+
+	if (!ops->ndo_get_xmit_slave)
+		return NULL;
+	return ops->ndo_get_xmit_slave(dev, skb, all_slaves);
+}
+EXPORT_SYMBOL(netdev_get_xmit_slave);
+
 static void netdev_adjacent_add_links(struct net_device *dev)
 {
 	struct netdev_adjacent *iter;