diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
index ba2b1f24ff14..6cc23af66b5b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_priv.h
@@ -94,13 +94,13 @@ struct mlx5e_tc_flow {
 	 * destinations.
 	 */
 	struct encap_flow_item encaps[MLX5_MAX_FLOW_FWD_VPORTS];
-	struct mlx5e_tc_flow *peer_flow;
 	struct mlx5e_hairpin_entry *hpe; /* attached hairpin instance */
 	struct list_head hairpin; /* flows sharing the same hairpin */
-	struct list_head peer;    /* flows with peer flow */
+	struct list_head peer[MLX5_MAX_PORTS];    /* flows with peer flow */
 	struct list_head unready; /* flows not ready to be offloaded (e.g
 				   * due to missing route)
 				   */
+	struct list_head peer_flows; /* flows on peer */
 	struct net_device *orig_dev; /* netdev adding flow first */
 	int tmp_entry_index;
 	struct list_head tmp_list; /* temporary flow list used by neigh update */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index 3e7041bd5705..965a8261c99b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -374,7 +374,9 @@ static void mlx5e_sqs2vport_stop(struct mlx5_eswitch *esw,
 				 struct mlx5_eswitch_rep *rep)
 {
 	struct mlx5e_rep_sq *rep_sq, *tmp;
+	struct mlx5e_rep_sq_peer *sq_peer;
 	struct mlx5e_rep_priv *rpriv;
+	unsigned long i;
 
 	if (esw->mode != MLX5_ESWITCH_OFFLOADS)
 		return;
@@ -382,31 +384,78 @@ static void mlx5e_sqs2vport_stop(struct mlx5_eswitch *esw,
 	rpriv = mlx5e_rep_to_rep_priv(rep);
 	list_for_each_entry_safe(rep_sq, tmp, &rpriv->vport_sqs_list, list) {
 		mlx5_eswitch_del_send_to_vport_rule(rep_sq->send_to_vport_rule);
-		if (rep_sq->send_to_vport_rule_peer)
-			mlx5_eswitch_del_send_to_vport_rule(rep_sq->send_to_vport_rule_peer);
+		xa_for_each(&rep_sq->sq_peer, i, sq_peer) {
+			if (sq_peer->rule)
+				mlx5_eswitch_del_send_to_vport_rule(sq_peer->rule);
+
+			xa_erase(&rep_sq->sq_peer, i);
+			kfree(sq_peer);
+		}
+
+		xa_destroy(&rep_sq->sq_peer);
 		list_del(&rep_sq->list);
 		kfree(rep_sq);
 	}
 }
 
+static int mlx5e_sqs2vport_add_peers_rules(struct mlx5_eswitch *esw, struct mlx5_eswitch_rep *rep,
+					   struct mlx5_devcom *devcom,
+					   struct mlx5e_rep_sq *rep_sq, int i)
+{
+	struct mlx5_eswitch *peer_esw = NULL;
+	struct mlx5_flow_handle *flow_rule;
+	int tmp;
+
+	mlx5_devcom_for_each_peer_entry(devcom, MLX5_DEVCOM_ESW_OFFLOADS,
+					peer_esw, tmp) {
+		int peer_rule_idx = mlx5_get_dev_index(peer_esw->dev);
+		struct mlx5e_rep_sq_peer *sq_peer;
+		int err;
+
+		sq_peer = kzalloc(sizeof(*sq_peer), GFP_KERNEL);
+		if (!sq_peer)
+			return -ENOMEM;
+
+		flow_rule = mlx5_eswitch_add_send_to_vport_rule(peer_esw, esw,
+								rep, rep_sq->sqn);
+		if (IS_ERR(flow_rule)) {
+			kfree(sq_peer);
+			return PTR_ERR(flow_rule);
+		}
+
+		sq_peer->rule = flow_rule;
+		sq_peer->peer = peer_esw;
+		err = xa_insert(&rep_sq->sq_peer, peer_rule_idx, sq_peer, GFP_KERNEL);
+		if (err) {
+			kfree(sq_peer);
+			mlx5_eswitch_del_send_to_vport_rule(flow_rule);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
 static int mlx5e_sqs2vport_start(struct mlx5_eswitch *esw,
 				 struct mlx5_eswitch_rep *rep,
 				 u32 *sqns_array, int sqns_num)
 {
-	struct mlx5_eswitch *peer_esw = NULL;
 	struct mlx5_flow_handle *flow_rule;
 	struct mlx5e_rep_priv *rpriv;
 	struct mlx5e_rep_sq *rep_sq;
+	struct mlx5_devcom *devcom;
+	bool devcom_locked = false;
 	int err;
 	int i;
 
 	if (esw->mode != MLX5_ESWITCH_OFFLOADS)
 		return 0;
 
+	devcom = esw->dev->priv.devcom;
 	rpriv = mlx5e_rep_to_rep_priv(rep);
-	if (mlx5_devcom_is_paired(esw->dev->priv.devcom, MLX5_DEVCOM_ESW_OFFLOADS))
-		peer_esw = mlx5_devcom_get_peer_data(esw->dev->priv.devcom,
-						     MLX5_DEVCOM_ESW_OFFLOADS);
+	if (mlx5_devcom_comp_is_ready(devcom, MLX5_DEVCOM_ESW_OFFLOADS) &&
+	    mlx5_devcom_for_each_peer_begin(devcom, MLX5_DEVCOM_ESW_OFFLOADS))
+		devcom_locked = true;
 
 	for (i = 0; i < sqns_num; i++) {
 		rep_sq = kzalloc(sizeof(*rep_sq), GFP_KERNEL);
@@ -426,31 +475,30 @@ static int mlx5e_sqs2vport_start(struct mlx5_eswitch *esw,
 		rep_sq->send_to_vport_rule = flow_rule;
 		rep_sq->sqn = sqns_array[i];
 
-		if (peer_esw) {
-			flow_rule = mlx5_eswitch_add_send_to_vport_rule(peer_esw, esw,
-									rep, sqns_array[i]);
-			if (IS_ERR(flow_rule)) {
-				err = PTR_ERR(flow_rule);
+		xa_init(&rep_sq->sq_peer);
+		if (devcom_locked) {
+			err = mlx5e_sqs2vport_add_peers_rules(esw, rep, devcom, rep_sq, i);
+			if (err) {
 				mlx5_eswitch_del_send_to_vport_rule(rep_sq->send_to_vport_rule);
+				xa_destroy(&rep_sq->sq_peer);
 				kfree(rep_sq);
 				goto out_err;
 			}
-			rep_sq->send_to_vport_rule_peer = flow_rule;
 		}
 
 		list_add(&rep_sq->list, &rpriv->vport_sqs_list);
 	}
 
-	if (peer_esw)
-		mlx5_devcom_release_peer_data(esw->dev->priv.devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+	if (devcom_locked)
+		mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 
 	return 0;
 
 out_err:
 	mlx5e_sqs2vport_stop(esw, rep);
 
-	if (peer_esw)
-		mlx5_devcom_release_peer_data(esw->dev->priv.devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+	if (devcom_locked)
+		mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 
 	return err;
 }
@@ -1530,17 +1578,24 @@ static void *mlx5e_vport_rep_get_proto_dev(struct mlx5_eswitch_rep *rep)
 	return rpriv->netdev;
 }
 
-static void mlx5e_vport_rep_event_unpair(struct mlx5_eswitch_rep *rep)
+static void mlx5e_vport_rep_event_unpair(struct mlx5_eswitch_rep *rep,
+					 struct mlx5_eswitch *peer_esw)
 {
+	int i = mlx5_get_dev_index(peer_esw->dev);
 	struct mlx5e_rep_priv *rpriv;
 	struct mlx5e_rep_sq *rep_sq;
 
+	WARN_ON_ONCE(!peer_esw);
 	rpriv = mlx5e_rep_to_rep_priv(rep);
 	list_for_each_entry(rep_sq, &rpriv->vport_sqs_list, list) {
-		if (!rep_sq->send_to_vport_rule_peer)
+		struct mlx5e_rep_sq_peer *sq_peer = xa_load(&rep_sq->sq_peer, i);
+
+		if (!sq_peer || sq_peer->peer != peer_esw)
 			continue;
-		mlx5_eswitch_del_send_to_vport_rule(rep_sq->send_to_vport_rule_peer);
-		rep_sq->send_to_vport_rule_peer = NULL;
+
+		mlx5_eswitch_del_send_to_vport_rule(sq_peer->rule);
+		xa_erase(&rep_sq->sq_peer, i);
+		kfree(sq_peer);
 	}
 }
 
@@ -1548,24 +1603,52 @@ static int mlx5e_vport_rep_event_pair(struct mlx5_eswitch *esw,
 				      struct mlx5_eswitch_rep *rep,
 				      struct mlx5_eswitch *peer_esw)
 {
+	int i = mlx5_get_dev_index(peer_esw->dev);
 	struct mlx5_flow_handle *flow_rule;
+	struct mlx5e_rep_sq_peer *sq_peer;
 	struct mlx5e_rep_priv *rpriv;
 	struct mlx5e_rep_sq *rep_sq;
+	int err;
 
 	rpriv = mlx5e_rep_to_rep_priv(rep);
 	list_for_each_entry(rep_sq, &rpriv->vport_sqs_list, list) {
-		if (rep_sq->send_to_vport_rule_peer)
+		sq_peer = xa_load(&rep_sq->sq_peer, i);
+
+		if (sq_peer && sq_peer->peer)
 			continue;
-		flow_rule = mlx5_eswitch_add_send_to_vport_rule(peer_esw, esw, rep, rep_sq->sqn);
-		if (IS_ERR(flow_rule))
+
+		flow_rule = mlx5_eswitch_add_send_to_vport_rule(peer_esw, esw, rep,
+								rep_sq->sqn);
+		if (IS_ERR(flow_rule)) {
+			err = PTR_ERR(flow_rule);
 			goto err_out;
-		rep_sq->send_to_vport_rule_peer = flow_rule;
+		}
+
+		if (sq_peer) {
+			sq_peer->rule = flow_rule;
+			sq_peer->peer = peer_esw;
+			continue;
+		}
+		sq_peer = kzalloc(sizeof(*sq_peer), GFP_KERNEL);
+		if (!sq_peer) {
+			err = -ENOMEM;
+			goto err_sq_alloc;
+		}
+		err = xa_insert(&rep_sq->sq_peer, i, sq_peer, GFP_KERNEL);
+		if (err)
+			goto err_xa;
+		sq_peer->rule = flow_rule;
+		sq_peer->peer = peer_esw;
 	}
 
 	return 0;
+err_xa:
+	kfree(sq_peer);
+err_sq_alloc:
+	mlx5_eswitch_del_send_to_vport_rule(flow_rule);
 err_out:
-	mlx5e_vport_rep_event_unpair(rep);
-	return PTR_ERR(flow_rule);
+	mlx5e_vport_rep_event_unpair(rep, peer_esw);
+	return err;
 }
 
 static int mlx5e_vport_rep_event(struct mlx5_eswitch *esw,
@@ -1578,7 +1661,7 @@ static int mlx5e_vport_rep_event(struct mlx5_eswitch *esw,
 	if (event == MLX5_SWITCHDEV_EVENT_PAIR)
 		err = mlx5e_vport_rep_event_pair(esw, rep, data);
 	else if (event == MLX5_SWITCHDEV_EVENT_UNPAIR)
-		mlx5e_vport_rep_event_unpair(rep);
+		mlx5e_vport_rep_event_unpair(rep, data);
 
 	return err;
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index 80b7f5079a5a..70640fa1ad7b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -225,9 +225,14 @@ struct mlx5e_encap_entry {
 	struct rcu_head rcu;
 };
 
+struct mlx5e_rep_sq_peer {
+	struct mlx5_flow_handle *rule;
+	void *peer;
+};
+
 struct mlx5e_rep_sq {
 	struct mlx5_flow_handle	*send_to_vport_rule;
-	struct mlx5_flow_handle *send_to_vport_rule_peer;
+	struct xarray sq_peer;
 	u32 sqn;
 	struct list_head	 list;
 };
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index fd9f928e25c7..88631fb9f966 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1666,8 +1666,11 @@ int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *ro
 {
 	struct mlx5e_priv *out_priv, *route_priv;
 	struct mlx5_core_dev *route_mdev;
+	struct mlx5_devcom *devcom;
 	struct mlx5_eswitch *esw;
 	u16 vhca_id;
+	int err;
+	int i;
 
 	out_priv = netdev_priv(out_dev);
 	esw = out_priv->mdev->priv.eswitch;
@@ -1675,28 +1678,25 @@ int mlx5e_tc_query_route_vport(struct net_device *out_dev, struct net_device *ro
 	route_mdev = route_priv->mdev;
 
 	vhca_id = MLX5_CAP_GEN(route_mdev, vhca_id);
-	if (mlx5_lag_is_active(out_priv->mdev)) {
-		struct mlx5_devcom *devcom;
-		int err;
-
-		/* In lag case we may get devices from different eswitch instances.
-		 * If we failed to get vport num, it means, mostly, that we on the wrong
-		 * eswitch.
-		 */
-		err = mlx5_eswitch_vhca_id_to_vport(esw, vhca_id, vport);
-		if (err != -ENOENT)
-			return err;
-
-		rcu_read_lock();
-		devcom = out_priv->mdev->priv.devcom;
-		esw = mlx5_devcom_get_peer_data_rcu(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
-		err = esw ? mlx5_eswitch_vhca_id_to_vport(esw, vhca_id, vport) : -ENODEV;
-		rcu_read_unlock();
-
+	err = mlx5_eswitch_vhca_id_to_vport(esw, vhca_id, vport);
+	if (!err)
 		return err;
-	}
 
-	return mlx5_eswitch_vhca_id_to_vport(esw, vhca_id, vport);
+	if (!mlx5_lag_is_active(out_priv->mdev))
+		return err;
+
+	rcu_read_lock();
+	devcom = out_priv->mdev->priv.devcom;
+	err = -ENODEV;
+	mlx5_devcom_for_each_peer_entry_rcu(devcom, MLX5_DEVCOM_ESW_OFFLOADS,
+					    esw, i) {
+		err = mlx5_eswitch_vhca_id_to_vport(esw, vhca_id, vport);
+		if (!err)
+			break;
+	}
+	rcu_read_unlock();
+
+	return err;
 }
 
 static int
@@ -1986,47 +1986,59 @@ void mlx5e_put_flow_list(struct mlx5e_priv *priv, struct list_head *flow_list)
 		mlx5e_flow_put(priv, flow);
 }
 
-static void __mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow)
+static void mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow,
+				       int peer_index)
 {
 	struct mlx5_eswitch *esw = flow->priv->mdev->priv.eswitch;
+	struct mlx5e_tc_flow *peer_flow;
+	struct mlx5e_tc_flow *tmp;
 
 	if (!flow_flag_test(flow, ESWITCH) ||
 	    !flow_flag_test(flow, DUP))
 		return;
 
 	mutex_lock(&esw->offloads.peer_mutex);
-	list_del(&flow->peer);
+	list_del(&flow->peer[peer_index]);
 	mutex_unlock(&esw->offloads.peer_mutex);
 
-	flow_flag_clear(flow, DUP);
-
-	if (refcount_dec_and_test(&flow->peer_flow->refcnt)) {
-		mlx5e_tc_del_fdb_flow(flow->peer_flow->priv, flow->peer_flow);
-		kfree(flow->peer_flow);
+	list_for_each_entry_safe(peer_flow, tmp, &flow->peer_flows, peer_flows) {
+		if (peer_index != mlx5_get_dev_index(peer_flow->priv->mdev))
+			continue;
+		if (refcount_dec_and_test(&peer_flow->refcnt)) {
+			mlx5e_tc_del_fdb_flow(peer_flow->priv, peer_flow);
+			list_del(&peer_flow->peer_flows);
+			kfree(peer_flow);
+		}
 	}
 
-	flow->peer_flow = NULL;
+	if (list_empty(&flow->peer_flows))
+		flow_flag_clear(flow, DUP);
 }
 
-static void mlx5e_tc_del_fdb_peer_flow(struct mlx5e_tc_flow *flow)
+static void mlx5e_tc_del_fdb_peers_flow(struct mlx5e_tc_flow *flow)
 {
-	struct mlx5_core_dev *dev = flow->priv->mdev;
-	struct mlx5_devcom *devcom = dev->priv.devcom;
-	struct mlx5_eswitch *peer_esw;
+	int i;
 
-	peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
-	if (!peer_esw)
-		return;
-
-	__mlx5e_tc_del_fdb_peer_flow(flow);
-	mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+	for (i = 0; i < MLX5_MAX_PORTS; i++) {
+		if (i == mlx5_get_dev_index(flow->priv->mdev))
+			continue;
+		mlx5e_tc_del_fdb_peer_flow(flow, i);
+	}
 }
 
 static void mlx5e_tc_del_flow(struct mlx5e_priv *priv,
 			      struct mlx5e_tc_flow *flow)
 {
 	if (mlx5e_is_eswitch_flow(flow)) {
-		mlx5e_tc_del_fdb_peer_flow(flow);
+		struct mlx5_devcom *devcom = flow->priv->mdev->priv.devcom;
+
+		if (!mlx5_devcom_for_each_peer_begin(devcom, MLX5_DEVCOM_ESW_OFFLOADS)) {
+			mlx5e_tc_del_fdb_flow(priv, flow);
+			return;
+		}
+
+		mlx5e_tc_del_fdb_peers_flow(flow);
+		mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 		mlx5e_tc_del_fdb_flow(priv, flow);
 	} else {
 		mlx5e_tc_del_nic_flow(priv, flow);
@@ -4203,8 +4215,8 @@ static bool is_peer_flow_needed(struct mlx5e_tc_flow *flow)
 		flow_flag_test(flow, INGRESS);
 	bool act_is_encap = !!(attr->action &
 			       MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT);
-	bool esw_paired = mlx5_devcom_is_paired(esw_attr->in_mdev->priv.devcom,
-						MLX5_DEVCOM_ESW_OFFLOADS);
+	bool esw_paired = mlx5_devcom_comp_is_ready(esw_attr->in_mdev->priv.devcom,
+						    MLX5_DEVCOM_ESW_OFFLOADS);
 
 	if (!esw_paired)
 		return false;
@@ -4295,6 +4307,7 @@ mlx5e_alloc_flow(struct mlx5e_priv *priv, int attr_size,
 	INIT_LIST_HEAD(&flow->hairpin);
 	INIT_LIST_HEAD(&flow->l3_to_l2_reformat);
 	INIT_LIST_HEAD(&flow->attrs);
+	INIT_LIST_HEAD(&flow->peer_flows);
 	refcount_set(&flow->refcnt, 1);
 	init_completion(&flow->init_done);
 	init_completion(&flow->del_hw_done);
@@ -4403,22 +4416,19 @@ out:
 
 static int mlx5e_tc_add_fdb_peer_flow(struct flow_cls_offload *f,
 				      struct mlx5e_tc_flow *flow,
-				      unsigned long flow_flags)
+				      unsigned long flow_flags,
+				      struct mlx5_eswitch *peer_esw)
 {
 	struct mlx5e_priv *priv = flow->priv, *peer_priv;
-	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch, *peer_esw;
+	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
 	struct mlx5_esw_flow_attr *attr = flow->attr->esw_attr;
-	struct mlx5_devcom *devcom = priv->mdev->priv.devcom;
 	struct mlx5e_tc_flow_parse_attr *parse_attr;
+	int i = mlx5_get_dev_index(peer_esw->dev);
 	struct mlx5e_rep_priv *peer_urpriv;
 	struct mlx5e_tc_flow *peer_flow;
 	struct mlx5_core_dev *in_mdev;
 	int err = 0;
 
-	peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
-	if (!peer_esw)
-		return -ENODEV;
-
 	peer_urpriv = mlx5_eswitch_get_uplink_priv(peer_esw, REP_ETH);
 	peer_priv = netdev_priv(peer_urpriv->netdev);
 
@@ -4443,14 +4453,13 @@ static int mlx5e_tc_add_fdb_peer_flow(struct flow_cls_offload *f,
 		goto out;
 	}
 
-	flow->peer_flow = peer_flow;
+	list_add_tail(&peer_flow->peer_flows, &flow->peer_flows);
 	flow_flag_set(flow, DUP);
 	mutex_lock(&esw->offloads.peer_mutex);
-	list_add_tail(&flow->peer, &esw->offloads.peer_flows);
+	list_add_tail(&flow->peer[i], &esw->offloads.peer_flows[i]);
 	mutex_unlock(&esw->offloads.peer_mutex);
 
 out:
-	mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 	return err;
 }
 
@@ -4461,30 +4470,48 @@ mlx5e_add_fdb_flow(struct mlx5e_priv *priv,
 		   struct net_device *filter_dev,
 		   struct mlx5e_tc_flow **__flow)
 {
+	struct mlx5_devcom *devcom = priv->mdev->priv.devcom;
 	struct mlx5e_rep_priv *rpriv = priv->ppriv;
 	struct mlx5_eswitch_rep *in_rep = rpriv->rep;
 	struct mlx5_core_dev *in_mdev = priv->mdev;
+	struct mlx5_eswitch *peer_esw;
 	struct mlx5e_tc_flow *flow;
 	int err;
+	int i;
 
 	flow = __mlx5e_add_fdb_flow(priv, f, flow_flags, filter_dev, in_rep,
 				    in_mdev);
 	if (IS_ERR(flow))
 		return PTR_ERR(flow);
 
-	if (is_peer_flow_needed(flow)) {
-		err = mlx5e_tc_add_fdb_peer_flow(f, flow, flow_flags);
-		if (err) {
-			mlx5e_tc_del_fdb_flow(priv, flow);
-			goto out;
-		}
+	if (!is_peer_flow_needed(flow)) {
+		*__flow = flow;
+		return 0;
 	}
 
-	*__flow = flow;
+	if (!mlx5_devcom_for_each_peer_begin(devcom, MLX5_DEVCOM_ESW_OFFLOADS)) {
+		err = -ENODEV;
+		goto clean_flow;
+	}
 
+	mlx5_devcom_for_each_peer_entry(devcom,
+					MLX5_DEVCOM_ESW_OFFLOADS,
+					peer_esw, i) {
+		err = mlx5e_tc_add_fdb_peer_flow(f, flow, flow_flags, peer_esw);
+		if (err)
+			goto peer_clean;
+	}
+
+	mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+
+	*__flow = flow;
 	return 0;
 
-out:
+peer_clean:
+	mlx5e_tc_del_fdb_peers_flow(flow);
+	mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+clean_flow:
+	mlx5e_tc_del_fdb_flow(priv, flow);
 	return err;
 }
 
@@ -4702,7 +4729,6 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv,
 {
 	struct mlx5_devcom *devcom = priv->mdev->priv.devcom;
 	struct rhashtable *tc_ht = get_tc_ht(priv, flags);
-	struct mlx5_eswitch *peer_esw;
 	struct mlx5e_tc_flow *flow;
 	struct mlx5_fc *counter;
 	u64 lastuse = 0;
@@ -4737,23 +4763,29 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv,
 	/* Under multipath it's possible for one rule to be currently
 	 * un-offloaded while the other rule is offloaded.
 	 */
-	peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
-	if (!peer_esw)
+	if (!mlx5_devcom_for_each_peer_begin(devcom, MLX5_DEVCOM_ESW_OFFLOADS))
 		goto out;
 
-	if (flow_flag_test(flow, DUP) &&
-	    flow_flag_test(flow->peer_flow, OFFLOADED)) {
-		u64 bytes2;
-		u64 packets2;
-		u64 lastuse2;
+	if (flow_flag_test(flow, DUP)) {
+		struct mlx5e_tc_flow *peer_flow;
 
-		if (flow_flag_test(flow, USE_ACT_STATS)) {
-			f->use_act_stats = true;
-		} else {
-			counter = mlx5e_tc_get_counter(flow->peer_flow);
+		list_for_each_entry(peer_flow, &flow->peer_flows, peer_flows) {
+			u64 packets2;
+			u64 lastuse2;
+			u64 bytes2;
+
+			if (!flow_flag_test(peer_flow, OFFLOADED))
+				continue;
+			if (flow_flag_test(flow, USE_ACT_STATS)) {
+				f->use_act_stats = true;
+				break;
+			}
+
+			counter = mlx5e_tc_get_counter(peer_flow);
 			if (!counter)
 				goto no_peer_counter;
-			mlx5_fc_query_cached(counter, &bytes2, &packets2, &lastuse2);
+			mlx5_fc_query_cached(counter, &bytes2, &packets2,
+					     &lastuse2);
 
 			bytes += bytes2;
 			packets += packets2;
@@ -4762,7 +4794,7 @@ int mlx5e_stats_flower(struct net_device *dev, struct mlx5e_priv *priv,
 	}
 
 no_peer_counter:
-	mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+	mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 out:
 	flow_stats_update(&f->stats, bytes, packets, 0, lastuse,
 			  FLOW_ACTION_HW_STATS_DELAYED);
@@ -5280,9 +5312,14 @@ int mlx5e_tc_num_filters(struct mlx5e_priv *priv, unsigned long flags)
 void mlx5e_tc_clean_fdb_peer_flows(struct mlx5_eswitch *esw)
 {
 	struct mlx5e_tc_flow *flow, *tmp;
+	int i;
 
-	list_for_each_entry_safe(flow, tmp, &esw->offloads.peer_flows, peer)
-		__mlx5e_tc_del_fdb_peer_flow(flow);
+	for (i = 0; i < MLX5_MAX_PORTS; i++) {
+		if (i == mlx5_get_dev_index(esw->dev))
+			continue;
+		list_for_each_entry_safe(flow, tmp, &esw->offloads.peer_flows[i], peer[i])
+			mlx5e_tc_del_fdb_peers_flow(flow);
+	}
 }
 
 void mlx5e_tc_reoffload_flows_work(struct work_struct *work)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/egress_ofld.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/egress_ofld.c
index 2e504c7461c6..24b1ca4e4ff8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/egress_ofld.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/egress_ofld.c
@@ -15,13 +15,27 @@ static void esw_acl_egress_ofld_fwd2vport_destroy(struct mlx5_vport *vport)
 	vport->egress.offloads.fwd_rule = NULL;
 }
 
-static void esw_acl_egress_ofld_bounce_rule_destroy(struct mlx5_vport *vport)
+void esw_acl_egress_ofld_bounce_rule_destroy(struct mlx5_vport *vport, int rule_index)
 {
-	if (!vport->egress.offloads.bounce_rule)
+	struct mlx5_flow_handle *bounce_rule =
+		xa_load(&vport->egress.offloads.bounce_rules, rule_index);
+
+	if (!bounce_rule)
 		return;
 
-	mlx5_del_flow_rules(vport->egress.offloads.bounce_rule);
-	vport->egress.offloads.bounce_rule = NULL;
+	mlx5_del_flow_rules(bounce_rule);
+	xa_erase(&vport->egress.offloads.bounce_rules, rule_index);
+}
+
+static void esw_acl_egress_ofld_bounce_rules_destroy(struct mlx5_vport *vport)
+{
+	struct mlx5_flow_handle *bounce_rule;
+	unsigned long i;
+
+	xa_for_each(&vport->egress.offloads.bounce_rules, i, bounce_rule) {
+		mlx5_del_flow_rules(bounce_rule);
+		xa_erase(&vport->egress.offloads.bounce_rules, i);
+	}
 }
 
 static int esw_acl_egress_ofld_fwd2vport_create(struct mlx5_eswitch *esw,
@@ -96,7 +110,7 @@ static void esw_acl_egress_ofld_rules_destroy(struct mlx5_vport *vport)
 {
 	esw_acl_egress_vlan_destroy(vport);
 	esw_acl_egress_ofld_fwd2vport_destroy(vport);
-	esw_acl_egress_ofld_bounce_rule_destroy(vport);
+	esw_acl_egress_ofld_bounce_rules_destroy(vport);
 }
 
 static int esw_acl_egress_ofld_groups_create(struct mlx5_eswitch *esw,
@@ -194,6 +208,7 @@ int esw_acl_egress_ofld_setup(struct mlx5_eswitch *esw, struct mlx5_vport *vport
 		vport->egress.acl = NULL;
 		return err;
 	}
+	vport->egress.type = VPORT_EGRESS_ACL_TYPE_DEFAULT;
 
 	err = esw_acl_egress_ofld_groups_create(esw, vport);
 	if (err)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/ofld.h b/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/ofld.h
index c9f8469e9a47..536b04e83618 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/ofld.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/acl/ofld.h
@@ -10,6 +10,7 @@
 /* Eswitch acl egress external APIs */
 int esw_acl_egress_ofld_setup(struct mlx5_eswitch *esw, struct mlx5_vport *vport);
 void esw_acl_egress_ofld_cleanup(struct mlx5_vport *vport);
+void esw_acl_egress_ofld_bounce_rule_destroy(struct mlx5_vport *vport, int rule_index);
 int mlx5_esw_acl_egress_vport_bond(struct mlx5_eswitch *esw, u16 active_vport_num,
 				   u16 passive_vport_num);
 int mlx5_esw_acl_egress_vport_unbond(struct mlx5_eswitch *esw, u16 vport_num);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c
index 1ba03e219111..bea7cc645461 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge.c
@@ -647,22 +647,35 @@ mlx5_esw_bridge_ingress_flow_create(u16 vport_num, const unsigned char *addr,
 }
 
 static struct mlx5_flow_handle *
-mlx5_esw_bridge_ingress_flow_peer_create(u16 vport_num, const unsigned char *addr,
+mlx5_esw_bridge_ingress_flow_peer_create(u16 vport_num, u16 esw_owner_vhca_id,
+					 const unsigned char *addr,
 					 struct mlx5_esw_bridge_vlan *vlan, u32 counter_id,
 					 struct mlx5_esw_bridge *bridge)
 {
 	struct mlx5_devcom *devcom = bridge->br_offloads->esw->dev->priv.devcom;
+	struct mlx5_eswitch *tmp, *peer_esw = NULL;
 	static struct mlx5_flow_handle *handle;
-	struct mlx5_eswitch *peer_esw;
+	int i;
 
-	peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
-	if (!peer_esw)
+	if (!mlx5_devcom_for_each_peer_begin(devcom, MLX5_DEVCOM_ESW_OFFLOADS))
 		return ERR_PTR(-ENODEV);
 
+	mlx5_devcom_for_each_peer_entry(devcom,
+					MLX5_DEVCOM_ESW_OFFLOADS,
+					tmp, i) {
+		if (mlx5_esw_is_owner(tmp, vport_num, esw_owner_vhca_id)) {
+			peer_esw = tmp;
+			break;
+		}
+	}
+	if (!peer_esw) {
+		mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+		return ERR_PTR(-ENODEV);
+	}
+
 	handle = mlx5_esw_bridge_ingress_flow_with_esw_create(vport_num, addr, vlan, counter_id,
 							      bridge, peer_esw);
-
-	mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+	mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 	return handle;
 }
 
@@ -1369,8 +1382,9 @@ mlx5_esw_bridge_fdb_entry_init(struct net_device *dev, u16 vport_num, u16 esw_ow
 	entry->ingress_counter = counter;
 
 	handle = peer ?
-		mlx5_esw_bridge_ingress_flow_peer_create(vport_num, addr, vlan,
-							 mlx5_fc_id(counter), bridge) :
+		mlx5_esw_bridge_ingress_flow_peer_create(vport_num, esw_owner_vhca_id,
+							 addr, vlan, mlx5_fc_id(counter),
+							 bridge) :
 		mlx5_esw_bridge_ingress_flow_create(vport_num, addr, vlan,
 						    mlx5_fc_id(counter), bridge);
 	if (IS_ERR(handle)) {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c
index 2eae594a5e80..2455f8b93c1e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/bridge_mcast.c
@@ -540,16 +540,29 @@ static struct mlx5_flow_handle *
 mlx5_esw_bridge_mcast_filter_flow_peer_create(struct mlx5_esw_bridge_port *port)
 {
 	struct mlx5_devcom *devcom = port->bridge->br_offloads->esw->dev->priv.devcom;
+	struct mlx5_eswitch *tmp, *peer_esw = NULL;
 	static struct mlx5_flow_handle *handle;
-	struct mlx5_eswitch *peer_esw;
+	int i;
 
-	peer_esw = mlx5_devcom_get_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
-	if (!peer_esw)
+	if (!mlx5_devcom_for_each_peer_begin(devcom, MLX5_DEVCOM_ESW_OFFLOADS))
 		return ERR_PTR(-ENODEV);
 
+	mlx5_devcom_for_each_peer_entry(devcom,
+					MLX5_DEVCOM_ESW_OFFLOADS,
+					tmp, i) {
+		if (mlx5_esw_is_owner(tmp, port->vport_num, port->esw_owner_vhca_id)) {
+			peer_esw = tmp;
+			break;
+		}
+	}
+	if (!peer_esw) {
+		mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+		return ERR_PTR(-ENODEV);
+	}
+
 	handle = mlx5_esw_bridge_mcast_flow_with_esw_create(port, peer_esw);
 
-	mlx5_devcom_release_peer_data(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
+	mlx5_devcom_for_each_peer_end(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
 	return handle;
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index f70124ad71cf..c42c16d9ccbc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -123,8 +123,14 @@ struct vport_ingress {
 	} offloads;
 };
 
+enum vport_egress_acl_type {
+	VPORT_EGRESS_ACL_TYPE_DEFAULT,
+	VPORT_EGRESS_ACL_TYPE_SHARED_FDB,
+};
+
 struct vport_egress {
 	struct mlx5_flow_table *acl;
+	enum vport_egress_acl_type type;
 	struct mlx5_flow_handle  *allowed_vlan;
 	struct mlx5_flow_group *vlan_grp;
 	union {
@@ -136,7 +142,7 @@ struct vport_egress {
 		struct {
 			struct mlx5_flow_group *fwd_grp;
 			struct mlx5_flow_handle *fwd_rule;
-			struct mlx5_flow_handle *bounce_rule;
+			struct xarray bounce_rules;
 			struct mlx5_flow_group *bounce_grp;
 		} offloads;
 	};
@@ -218,7 +224,7 @@ struct mlx5_eswitch_fdb {
 			struct mlx5_flow_group *send_to_vport_grp;
 			struct mlx5_flow_group *send_to_vport_meta_grp;
 			struct mlx5_flow_group *peer_miss_grp;
-			struct mlx5_flow_handle **peer_miss_rules;
+			struct mlx5_flow_handle **peer_miss_rules[MLX5_MAX_PORTS];
 			struct mlx5_flow_group *miss_grp;
 			struct mlx5_flow_handle **send_to_vport_meta_rules;
 			struct mlx5_flow_handle *miss_rule_uni;
@@ -249,7 +255,7 @@ struct mlx5_esw_offload {
 	struct mlx5_flow_group *vport_rx_drop_group;
 	struct mlx5_flow_handle *vport_rx_drop_rule;
 	struct xarray vport_reps;
-	struct list_head peer_flows;
+	struct list_head peer_flows[MLX5_MAX_PORTS];
 	struct mutex peer_mutex;
 	struct mutex encap_tbl_lock; /* protects encap_tbl */
 	DECLARE_HASHTABLE(encap_tbl, 8);
@@ -337,6 +343,7 @@ struct mlx5_eswitch {
 	int                     mode;
 	u16                     manager_vport;
 	u16                     first_host_vport;
+	u8			num_peers;
 	struct mlx5_esw_functions esw_funcs;
 	struct {
 		u32             large_group_num;
@@ -578,6 +585,13 @@ mlx5_esw_is_manager_vport(const struct mlx5_eswitch *esw, u16 vport_num)
 	return esw->manager_vport == vport_num;
 }
 
+static inline bool mlx5_esw_is_owner(struct mlx5_eswitch *esw, u16 vport_num,
+				     u16 esw_owner_vhca_id)
+{
+	return esw_owner_vhca_id == MLX5_CAP_GEN(esw->dev, vhca_id) ||
+		(vport_num == MLX5_VPORT_UPLINK && mlx5_lag_is_master(esw->dev));
+}
+
 static inline u16 mlx5_eswitch_first_host_vport_num(struct mlx5_core_dev *dev)
 {
 	return mlx5_core_is_ecpf_esw_manager(dev) ?
@@ -748,9 +762,9 @@ void esw_vport_change_handle_locked(struct mlx5_vport *vport);
 
 bool mlx5_esw_offloads_controller_valid(const struct mlx5_eswitch *esw, u32 controller);
 
-int mlx5_eswitch_offloads_config_single_fdb(struct mlx5_eswitch *master_esw,
-					    struct mlx5_eswitch *slave_esw);
-void mlx5_eswitch_offloads_destroy_single_fdb(struct mlx5_eswitch *master_esw,
+int mlx5_eswitch_offloads_single_fdb_add_one(struct mlx5_eswitch *master_esw,
+					     struct mlx5_eswitch *slave_esw, int max_slaves);
+void mlx5_eswitch_offloads_single_fdb_del_one(struct mlx5_eswitch *master_esw,
 					      struct mlx5_eswitch *slave_esw);
 int mlx5_eswitch_reload_reps(struct mlx5_eswitch *esw);
 
@@ -802,14 +816,14 @@ mlx5_esw_vport_to_devlink_port_index(const struct mlx5_core_dev *dev,
 }
 
 static inline int
-mlx5_eswitch_offloads_config_single_fdb(struct mlx5_eswitch *master_esw,
-					struct mlx5_eswitch *slave_esw)
+mlx5_eswitch_offloads_single_fdb_add_one(struct mlx5_eswitch *master_esw,
+					 struct mlx5_eswitch *slave_esw, int max_slaves)
 {
 	return 0;
 }
 
 static inline void
-mlx5_eswitch_offloads_destroy_single_fdb(struct mlx5_eswitch *master_esw,
+mlx5_eswitch_offloads_single_fdb_del_one(struct mlx5_eswitch *master_esw,
 					 struct mlx5_eswitch *slave_esw) {}
 
 static inline int
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 1b2f5e273525..29de4e759f4f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -1132,7 +1132,7 @@ static int esw_add_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
 		flows[vport->index] = flow;
 	}
 
-	esw->fdb_table.offloads.peer_miss_rules = flows;
+	esw->fdb_table.offloads.peer_miss_rules[mlx5_get_dev_index(peer_dev)] = flows;
 
 	kvfree(spec);
 	return 0;
@@ -1160,13 +1160,14 @@ alloc_flows_err:
 	return err;
 }
 
-static void esw_del_fdb_peer_miss_rules(struct mlx5_eswitch *esw)
+static void esw_del_fdb_peer_miss_rules(struct mlx5_eswitch *esw,
+					struct mlx5_core_dev *peer_dev)
 {
 	struct mlx5_flow_handle **flows;
 	struct mlx5_vport *vport;
 	unsigned long i;
 
-	flows = esw->fdb_table.offloads.peer_miss_rules;
+	flows = esw->fdb_table.offloads.peer_miss_rules[mlx5_get_dev_index(peer_dev)];
 
 	mlx5_esw_for_each_vf_vport(esw, i, vport, mlx5_core_max_vfs(esw->dev))
 		mlx5_del_flow_rules(flows[vport->index]);
@@ -1573,6 +1574,7 @@ esw_create_peer_esw_miss_group(struct mlx5_eswitch *esw,
 			       u32 *flow_group_in,
 			       int *ix)
 {
+	int max_peer_ports = (esw->total_vports - 1) * (MLX5_MAX_PORTS - 1);
 	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
 	struct mlx5_flow_group *g;
 	void *match_criteria;
@@ -1599,8 +1601,8 @@ esw_create_peer_esw_miss_group(struct mlx5_eswitch *esw,
 
 	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, *ix);
 	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index,
-		 *ix + esw->total_vports - 1);
-	*ix += esw->total_vports;
+		 *ix + max_peer_ports);
+	*ix += max_peer_ports + 1;
 
 	g = mlx5_create_flow_group(fdb, flow_group_in);
 	if (IS_ERR(g)) {
@@ -1702,7 +1704,7 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw)
 	 * total vports of the peer (currently is also uses esw->total_vports).
 	 */
 	table_size = MLX5_MAX_PORTS * (esw->total_vports * MAX_SQ_NVPORTS + MAX_PF_SQ) +
-		     esw->total_vports * 2 + MLX5_ESW_MISS_FLOWS;
+		     esw->total_vports * MLX5_MAX_PORTS + MLX5_ESW_MISS_FLOWS;
 
 	/* create the slow path fdb with encap set, so further table instances
 	 * can be created at run time while VFs are probed if the FW allows that.
@@ -2510,6 +2512,7 @@ static int __esw_set_master_egress_rule(struct mlx5_core_dev *master,
 					struct mlx5_vport *vport,
 					struct mlx5_flow_table *acl)
 {
+	u16 slave_index = MLX5_CAP_GEN(slave, vhca_id);
 	struct mlx5_flow_handle *flow_rule = NULL;
 	struct mlx5_flow_destination dest = {};
 	struct mlx5_flow_act flow_act = {};
@@ -2525,8 +2528,7 @@ static int __esw_set_master_egress_rule(struct mlx5_core_dev *master,
 	misc = MLX5_ADDR_OF(fte_match_param, spec->match_value,
 			    misc_parameters);
 	MLX5_SET(fte_match_set_misc, misc, source_port, MLX5_VPORT_UPLINK);
-	MLX5_SET(fte_match_set_misc, misc, source_eswitch_owner_vhca_id,
-		 MLX5_CAP_GEN(slave, vhca_id));
+	MLX5_SET(fte_match_set_misc, misc, source_eswitch_owner_vhca_id, slave_index);
 
 	misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria, misc_parameters);
 	MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
@@ -2541,44 +2543,35 @@ static int __esw_set_master_egress_rule(struct mlx5_core_dev *master,
 
 	flow_rule = mlx5_add_flow_rules(acl, spec, &flow_act,
 					&dest, 1);
-	if (IS_ERR(flow_rule))
+	if (IS_ERR(flow_rule)) {
 		err = PTR_ERR(flow_rule);
-	else
-		vport->egress.offloads.bounce_rule = flow_rule;
+	} else {
+		err = xa_insert(&vport->egress.offloads.bounce_rules,
+				slave_index, flow_rule, GFP_KERNEL);
+		if (err)
+			mlx5_del_flow_rules(flow_rule);
+	}
 
 	kvfree(spec);
 	return err;
 }
 
-static int esw_set_master_egress_rule(struct mlx5_core_dev *master,
-				      struct mlx5_core_dev *slave)
+static int esw_master_egress_create_resources(struct mlx5_flow_namespace *egress_ns,
+					      struct mlx5_vport *vport, size_t count)
 {
 	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
-	struct mlx5_eswitch *esw = master->priv.eswitch;
 	struct mlx5_flow_table_attr ft_attr = {
-		.max_fte = 1, .prio = 0, .level = 0,
+		.max_fte = count, .prio = 0, .level = 0,
 		.flags = MLX5_FLOW_TABLE_OTHER_VPORT,
 	};
-	struct mlx5_flow_namespace *egress_ns;
 	struct mlx5_flow_table *acl;
 	struct mlx5_flow_group *g;
-	struct mlx5_vport *vport;
 	void *match_criteria;
 	u32 *flow_group_in;
 	int err;
 
-	vport = mlx5_eswitch_get_vport(esw, esw->manager_vport);
-	if (IS_ERR(vport))
-		return PTR_ERR(vport);
-
-	egress_ns = mlx5_get_flow_vport_acl_namespace(master,
-						      MLX5_FLOW_NAMESPACE_ESW_EGRESS,
-						      vport->index);
-	if (!egress_ns)
-		return -EINVAL;
-
 	if (vport->egress.acl)
-		return -EINVAL;
+		return 0;
 
 	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
 	if (!flow_group_in)
@@ -2602,7 +2595,7 @@ static int esw_set_master_egress_rule(struct mlx5_core_dev *master,
 	MLX5_SET(create_flow_group_in, flow_group_in,
 		 source_eswitch_owner_vhca_id_valid, 1);
 	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index, 0);
-	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, 0);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index, count);
 
 	g = mlx5_create_flow_group(acl, flow_group_in);
 	if (IS_ERR(g)) {
@@ -2610,19 +2603,15 @@ static int esw_set_master_egress_rule(struct mlx5_core_dev *master,
 		goto err_group;
 	}
 
-	err = __esw_set_master_egress_rule(master, slave, vport, acl);
-	if (err)
-		goto err_rule;
-
 	vport->egress.acl = acl;
 	vport->egress.offloads.bounce_grp = g;
+	vport->egress.type = VPORT_EGRESS_ACL_TYPE_SHARED_FDB;
+	xa_init_flags(&vport->egress.offloads.bounce_rules, XA_FLAGS_ALLOC);
 
 	kvfree(flow_group_in);
 
 	return 0;
 
-err_rule:
-	mlx5_destroy_flow_group(g);
 err_group:
 	mlx5_destroy_flow_table(acl);
 out:
@@ -2630,18 +2619,70 @@ out:
 	return err;
 }
 
-static void esw_unset_master_egress_rule(struct mlx5_core_dev *dev)
+static void esw_master_egress_destroy_resources(struct mlx5_vport *vport)
+{
+	mlx5_destroy_flow_group(vport->egress.offloads.bounce_grp);
+	mlx5_destroy_flow_table(vport->egress.acl);
+}
+
+static int esw_set_master_egress_rule(struct mlx5_core_dev *master,
+				      struct mlx5_core_dev *slave, size_t count)
+{
+	struct mlx5_eswitch *esw = master->priv.eswitch;
+	u16 slave_index = MLX5_CAP_GEN(slave, vhca_id);
+	struct mlx5_flow_namespace *egress_ns;
+	struct mlx5_vport *vport;
+	int err;
+
+	vport = mlx5_eswitch_get_vport(esw, esw->manager_vport);
+	if (IS_ERR(vport))
+		return PTR_ERR(vport);
+
+	egress_ns = mlx5_get_flow_vport_acl_namespace(master,
+						      MLX5_FLOW_NAMESPACE_ESW_EGRESS,
+						      vport->index);
+	if (!egress_ns)
+		return -EINVAL;
+
+	if (vport->egress.acl && vport->egress.type != VPORT_EGRESS_ACL_TYPE_SHARED_FDB)
+		return 0;
+
+	err = esw_master_egress_create_resources(egress_ns, vport, count);
+	if (err)
+		return err;
+
+	if (xa_load(&vport->egress.offloads.bounce_rules, slave_index))
+		return -EINVAL;
+
+	err = __esw_set_master_egress_rule(master, slave, vport, vport->egress.acl);
+	if (err)
+		goto err_rule;
+
+	return 0;
+
+err_rule:
+	esw_master_egress_destroy_resources(vport);
+	return err;
+}
+
+static void esw_unset_master_egress_rule(struct mlx5_core_dev *dev,
+					 struct mlx5_core_dev *slave_dev)
 {
 	struct mlx5_vport *vport;
 
 	vport = mlx5_eswitch_get_vport(dev->priv.eswitch,
 				       dev->priv.eswitch->manager_vport);
 
-	esw_acl_egress_ofld_cleanup(vport);
+	esw_acl_egress_ofld_bounce_rule_destroy(vport, MLX5_CAP_GEN(slave_dev, vhca_id));
+
+	if (xa_empty(&vport->egress.offloads.bounce_rules)) {
+		esw_acl_egress_ofld_cleanup(vport);
+		xa_destroy(&vport->egress.offloads.bounce_rules);
+	}
 }
 
-int mlx5_eswitch_offloads_config_single_fdb(struct mlx5_eswitch *master_esw,
-					    struct mlx5_eswitch *slave_esw)
+int mlx5_eswitch_offloads_single_fdb_add_one(struct mlx5_eswitch *master_esw,
+					     struct mlx5_eswitch *slave_esw, int max_slaves)
 {
 	int err;
 
@@ -2651,7 +2692,7 @@ int mlx5_eswitch_offloads_config_single_fdb(struct mlx5_eswitch *master_esw,
 		return err;
 
 	err = esw_set_master_egress_rule(master_esw->dev,
-					 slave_esw->dev);
+					 slave_esw->dev, max_slaves);
 	if (err)
 		goto err_acl;
 
@@ -2659,21 +2700,21 @@ int mlx5_eswitch_offloads_config_single_fdb(struct mlx5_eswitch *master_esw,
 
 err_acl:
 	esw_set_slave_root_fdb(NULL, slave_esw->dev);
-
 	return err;
 }
 
-void mlx5_eswitch_offloads_destroy_single_fdb(struct mlx5_eswitch *master_esw,
+void mlx5_eswitch_offloads_single_fdb_del_one(struct mlx5_eswitch *master_esw,
 					      struct mlx5_eswitch *slave_esw)
 {
-	esw_unset_master_egress_rule(master_esw->dev);
 	esw_set_slave_root_fdb(NULL, slave_esw->dev);
+	esw_unset_master_egress_rule(master_esw->dev, slave_esw->dev);
 }
 
 #define ESW_OFFLOADS_DEVCOM_PAIR	(0)
 #define ESW_OFFLOADS_DEVCOM_UNPAIR	(1)
 
-static void mlx5_esw_offloads_rep_event_unpair(struct mlx5_eswitch *esw)
+static void mlx5_esw_offloads_rep_event_unpair(struct mlx5_eswitch *esw,
+					       struct mlx5_eswitch *peer_esw)
 {
 	const struct mlx5_eswitch_rep_ops *ops;
 	struct mlx5_eswitch_rep *rep;
@@ -2686,18 +2727,19 @@ static void mlx5_esw_offloads_rep_event_unpair(struct mlx5_eswitch *esw)
 			ops = esw->offloads.rep_ops[rep_type];
 			if (atomic_read(&rep->rep_data[rep_type].state) == REP_LOADED &&
 			    ops->event)
-				ops->event(esw, rep, MLX5_SWITCHDEV_EVENT_UNPAIR, NULL);
+				ops->event(esw, rep, MLX5_SWITCHDEV_EVENT_UNPAIR, peer_esw);
 		}
 	}
 }
 
-static void mlx5_esw_offloads_unpair(struct mlx5_eswitch *esw)
+static void mlx5_esw_offloads_unpair(struct mlx5_eswitch *esw,
+				     struct mlx5_eswitch *peer_esw)
 {
 #if IS_ENABLED(CONFIG_MLX5_CLS_ACT)
 	mlx5e_tc_clean_fdb_peer_flows(esw);
 #endif
-	mlx5_esw_offloads_rep_event_unpair(esw);
-	esw_del_fdb_peer_miss_rules(esw);
+	mlx5_esw_offloads_rep_event_unpair(esw, peer_esw);
+	esw_del_fdb_peer_miss_rules(esw, peer_esw->dev);
 }
 
 static int mlx5_esw_offloads_pair(struct mlx5_eswitch *esw,
@@ -2728,7 +2770,7 @@ static int mlx5_esw_offloads_pair(struct mlx5_eswitch *esw,
 	return 0;
 
 err_out:
-	mlx5_esw_offloads_unpair(esw);
+	mlx5_esw_offloads_unpair(esw, peer_esw);
 	return err;
 }
 
@@ -2736,7 +2778,9 @@ static int mlx5_esw_offloads_set_ns_peer(struct mlx5_eswitch *esw,
 					 struct mlx5_eswitch *peer_esw,
 					 bool pair)
 {
+	u8 peer_idx = mlx5_get_dev_index(peer_esw->dev);
 	struct mlx5_flow_root_namespace *peer_ns;
+	u8 idx = mlx5_get_dev_index(esw->dev);
 	struct mlx5_flow_root_namespace *ns;
 	int err;
 
@@ -2744,18 +2788,18 @@ static int mlx5_esw_offloads_set_ns_peer(struct mlx5_eswitch *esw,
 	ns = esw->dev->priv.steering->fdb_root_ns;
 
 	if (pair) {
-		err = mlx5_flow_namespace_set_peer(ns, peer_ns);
+		err = mlx5_flow_namespace_set_peer(ns, peer_ns, peer_idx);
 		if (err)
 			return err;
 
-		err = mlx5_flow_namespace_set_peer(peer_ns, ns);
+		err = mlx5_flow_namespace_set_peer(peer_ns, ns, idx);
 		if (err) {
-			mlx5_flow_namespace_set_peer(ns, NULL);
+			mlx5_flow_namespace_set_peer(ns, NULL, peer_idx);
 			return err;
 		}
 	} else {
-		mlx5_flow_namespace_set_peer(ns, NULL);
-		mlx5_flow_namespace_set_peer(peer_ns, NULL);
+		mlx5_flow_namespace_set_peer(ns, NULL, peer_idx);
+		mlx5_flow_namespace_set_peer(peer_ns, NULL, idx);
 	}
 
 	return 0;
@@ -2792,18 +2836,23 @@ static int mlx5_esw_offloads_devcom_event(int event,
 
 		esw->paired[mlx5_get_dev_index(peer_esw->dev)] = true;
 		peer_esw->paired[mlx5_get_dev_index(esw->dev)] = true;
-		mlx5_devcom_set_paired(devcom, MLX5_DEVCOM_ESW_OFFLOADS, true);
+		esw->num_peers++;
+		peer_esw->num_peers++;
+		mlx5_devcom_comp_set_ready(devcom, MLX5_DEVCOM_ESW_OFFLOADS, true);
 		break;
 
 	case ESW_OFFLOADS_DEVCOM_UNPAIR:
 		if (!esw->paired[mlx5_get_dev_index(peer_esw->dev)])
 			break;
 
-		mlx5_devcom_set_paired(devcom, MLX5_DEVCOM_ESW_OFFLOADS, false);
+		peer_esw->num_peers--;
+		esw->num_peers--;
+		if (!esw->num_peers && !peer_esw->num_peers)
+			mlx5_devcom_comp_set_ready(devcom, MLX5_DEVCOM_ESW_OFFLOADS, false);
 		esw->paired[mlx5_get_dev_index(peer_esw->dev)] = false;
 		peer_esw->paired[mlx5_get_dev_index(esw->dev)] = false;
-		mlx5_esw_offloads_unpair(peer_esw);
-		mlx5_esw_offloads_unpair(esw);
+		mlx5_esw_offloads_unpair(peer_esw, esw);
+		mlx5_esw_offloads_unpair(esw, peer_esw);
 		mlx5_esw_offloads_set_ns_peer(esw, peer_esw, false);
 		break;
 	}
@@ -2811,7 +2860,7 @@ static int mlx5_esw_offloads_devcom_event(int event,
 	return 0;
 
 err_pair:
-	mlx5_esw_offloads_unpair(esw);
+	mlx5_esw_offloads_unpair(esw, peer_esw);
 err_peer:
 	mlx5_esw_offloads_set_ns_peer(esw, peer_esw, false);
 err_out:
@@ -2823,8 +2872,10 @@ err_out:
 void mlx5_esw_offloads_devcom_init(struct mlx5_eswitch *esw)
 {
 	struct mlx5_devcom *devcom = esw->dev->priv.devcom;
+	int i;
 
-	INIT_LIST_HEAD(&esw->offloads.peer_flows);
+	for (i = 0; i < MLX5_MAX_PORTS; i++)
+		INIT_LIST_HEAD(&esw->offloads.peer_flows[i]);
 	mutex_init(&esw->offloads.peer_mutex);
 
 	if (!MLX5_CAP_ESW(esw->dev, merged_eswitch))
@@ -2838,9 +2889,11 @@ void mlx5_esw_offloads_devcom_init(struct mlx5_eswitch *esw)
 				       mlx5_esw_offloads_devcom_event,
 				       esw);
 
+	esw->num_peers = 0;
 	mlx5_devcom_send_event(devcom,
 			       MLX5_DEVCOM_ESW_OFFLOADS,
-			       ESW_OFFLOADS_DEVCOM_PAIR, esw);
+			       ESW_OFFLOADS_DEVCOM_PAIR,
+			       ESW_OFFLOADS_DEVCOM_UNPAIR, esw);
 }
 
 void mlx5_esw_offloads_devcom_cleanup(struct mlx5_eswitch *esw)
@@ -2854,6 +2907,7 @@ void mlx5_esw_offloads_devcom_cleanup(struct mlx5_eswitch *esw)
 		return;
 
 	mlx5_devcom_send_event(devcom, MLX5_DEVCOM_ESW_OFFLOADS,
+			       ESW_OFFLOADS_DEVCOM_UNPAIR,
 			       ESW_OFFLOADS_DEVCOM_UNPAIR, esw);
 
 	mlx5_devcom_unregister_component(devcom, MLX5_DEVCOM_ESW_OFFLOADS);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
index 144e59480686..11374c3744c5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
@@ -139,7 +139,8 @@ static void mlx5_cmd_stub_modify_header_dealloc(struct mlx5_flow_root_namespace
 }
 
 static int mlx5_cmd_stub_set_peer(struct mlx5_flow_root_namespace *ns,
-				  struct mlx5_flow_root_namespace *peer_ns)
+				  struct mlx5_flow_root_namespace *peer_ns,
+				  u8 peer_idx)
 {
 	return 0;
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h
index 8ef4254b9ea1..b6b9a5a20591 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h
@@ -93,7 +93,8 @@ struct mlx5_flow_cmds {
 				      struct mlx5_modify_hdr *modify_hdr);
 
 	int (*set_peer)(struct mlx5_flow_root_namespace *ns,
-			struct mlx5_flow_root_namespace *peer_ns);
+			struct mlx5_flow_root_namespace *peer_ns,
+			u8 peer_idx);
 
 	int (*create_ns)(struct mlx5_flow_root_namespace *ns);
 	int (*destroy_ns)(struct mlx5_flow_root_namespace *ns);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 19da02c41616..4ef04aa28771 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -3620,7 +3620,8 @@ void mlx5_destroy_match_definer(struct mlx5_core_dev *dev,
 }
 
 int mlx5_flow_namespace_set_peer(struct mlx5_flow_root_namespace *ns,
-				 struct mlx5_flow_root_namespace *peer_ns)
+				 struct mlx5_flow_root_namespace *peer_ns,
+				 u8 peer_idx)
 {
 	if (peer_ns && ns->mode != peer_ns->mode) {
 		mlx5_core_err(ns->dev,
@@ -3628,7 +3629,7 @@ int mlx5_flow_namespace_set_peer(struct mlx5_flow_root_namespace *ns,
 		return -EINVAL;
 	}
 
-	return ns->cmds->set_peer(ns, peer_ns);
+	return ns->cmds->set_peer(ns, peer_ns, peer_idx);
 }
 
 /* This function should be called only at init stage of the namespace.
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
index f137a0611b77..200ec946409c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
@@ -295,7 +295,8 @@ void mlx5_fc_update_sampling_interval(struct mlx5_core_dev *dev,
 const struct mlx5_flow_cmds *mlx5_fs_cmd_get_fw_cmds(void);
 
 int mlx5_flow_namespace_set_peer(struct mlx5_flow_root_namespace *ns,
-				 struct mlx5_flow_root_namespace *peer_ns);
+				 struct mlx5_flow_root_namespace *peer_ns,
+				 u8 peer_idx);
 
 int mlx5_flow_namespace_set_mode(struct mlx5_flow_namespace *ns,
 				 enum mlx5_flow_steering_mode mode);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
index 5d331b940f4d..c820f7d266de 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c
@@ -550,6 +550,29 @@ char *mlx5_get_str_port_sel_mode(enum mlx5_lag_mode mode, unsigned long flags)
 	}
 }
 
+static int mlx5_lag_create_single_fdb(struct mlx5_lag *ldev)
+{
+	struct mlx5_core_dev *dev0 = ldev->pf[MLX5_LAG_P1].dev;
+	struct mlx5_eswitch *master_esw = dev0->priv.eswitch;
+	int err;
+	int i;
+
+	for (i = MLX5_LAG_P1 + 1; i < ldev->ports; i++) {
+		struct mlx5_eswitch *slave_esw = ldev->pf[i].dev->priv.eswitch;
+
+		err = mlx5_eswitch_offloads_single_fdb_add_one(master_esw,
+							       slave_esw, ldev->ports);
+		if (err)
+			goto err;
+	}
+	return 0;
+err:
+	for (; i > MLX5_LAG_P1; i--)
+		mlx5_eswitch_offloads_single_fdb_del_one(master_esw,
+							 ldev->pf[i].dev->priv.eswitch);
+	return err;
+}
+
 static int mlx5_create_lag(struct mlx5_lag *ldev,
 			   struct lag_tracker *tracker,
 			   enum mlx5_lag_mode mode,
@@ -557,7 +580,6 @@ static int mlx5_create_lag(struct mlx5_lag *ldev,
 {
 	bool shared_fdb = test_bit(MLX5_LAG_MODE_FLAG_SHARED_FDB, &flags);
 	struct mlx5_core_dev *dev0 = ldev->pf[MLX5_LAG_P1].dev;
-	struct mlx5_core_dev *dev1 = ldev->pf[MLX5_LAG_P2].dev;
 	u32 in[MLX5_ST_SZ_DW(destroy_lag_in)] = {};
 	int err;
 
@@ -575,8 +597,7 @@ static int mlx5_create_lag(struct mlx5_lag *ldev,
 	}
 
 	if (shared_fdb) {
-		err = mlx5_eswitch_offloads_config_single_fdb(dev0->priv.eswitch,
-							      dev1->priv.eswitch);
+		err = mlx5_lag_create_single_fdb(ldev);
 		if (err)
 			mlx5_core_err(dev0, "Can't enable single FDB mode\n");
 		else
@@ -647,19 +668,21 @@ int mlx5_activate_lag(struct mlx5_lag *ldev,
 int mlx5_deactivate_lag(struct mlx5_lag *ldev)
 {
 	struct mlx5_core_dev *dev0 = ldev->pf[MLX5_LAG_P1].dev;
-	struct mlx5_core_dev *dev1 = ldev->pf[MLX5_LAG_P2].dev;
+	struct mlx5_eswitch *master_esw = dev0->priv.eswitch;
 	u32 in[MLX5_ST_SZ_DW(destroy_lag_in)] = {};
 	bool roce_lag = __mlx5_lag_is_roce(ldev);
 	unsigned long flags = ldev->mode_flags;
 	int err;
+	int i;
 
 	ldev->mode = MLX5_LAG_MODE_NONE;
 	ldev->mode_flags = 0;
 	mlx5_lag_mp_reset(ldev);
 
 	if (test_bit(MLX5_LAG_MODE_FLAG_SHARED_FDB, &flags)) {
-		mlx5_eswitch_offloads_destroy_single_fdb(dev0->priv.eswitch,
-							 dev1->priv.eswitch);
+		for (i = MLX5_LAG_P1 + 1; i < ldev->ports; i++)
+			mlx5_eswitch_offloads_single_fdb_del_one(master_esw,
+								 ldev->pf[i].dev->priv.eswitch);
 		clear_bit(MLX5_LAG_MODE_FLAG_SHARED_FDB, &flags);
 	}
 
@@ -801,8 +824,8 @@ bool mlx5_shared_fdb_supported(struct mlx5_lag *ldev)
 	    is_mdev_switchdev_mode(dev1) &&
 	    mlx5_eswitch_vport_match_metadata_enabled(dev0->priv.eswitch) &&
 	    mlx5_eswitch_vport_match_metadata_enabled(dev1->priv.eswitch) &&
-	    mlx5_devcom_is_paired(dev0->priv.devcom,
-				  MLX5_DEVCOM_ESW_OFFLOADS) &&
+	    mlx5_devcom_comp_is_ready(dev0->priv.devcom,
+				      MLX5_DEVCOM_ESW_OFFLOADS) &&
 	    MLX5_CAP_GEN(dev1, lag_native_fdb_selection) &&
 	    MLX5_CAP_ESW(dev1, root_ft_on_other_esw) &&
 	    MLX5_CAP_ESW(dev0, esw_shared_ingress_acl))
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c
index b7d779d08d83..8472bbb3cd58 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c
@@ -19,7 +19,7 @@ struct mlx5_devcom_component {
 
 	mlx5_devcom_event_handler_t handler;
 	struct rw_semaphore sem;
-	bool paired;
+	bool ready;
 };
 
 struct mlx5_devcom_list {
@@ -193,7 +193,7 @@ void mlx5_devcom_unregister_component(struct mlx5_devcom *devcom,
 
 int mlx5_devcom_send_event(struct mlx5_devcom *devcom,
 			   enum mlx5_devcom_components id,
-			   int event,
+			   int event, int rollback_event,
 			   void *event_data)
 {
 	struct mlx5_devcom_component *comp;
@@ -210,84 +210,134 @@ int mlx5_devcom_send_event(struct mlx5_devcom *devcom,
 
 		if (i != devcom->idx && data) {
 			err = comp->handler(event, data, event_data);
-			break;
+			if (err)
+				goto rollback;
 		}
 	}
 
+	up_write(&comp->sem);
+	return 0;
+
+rollback:
+	while (i--) {
+		void *data = rcu_dereference_protected(comp->device[i].data,
+						       lockdep_is_held(&comp->sem));
+
+		if (i != devcom->idx && data)
+			comp->handler(rollback_event, data, event_data);
+	}
+
 	up_write(&comp->sem);
 	return err;
 }
 
-void mlx5_devcom_set_paired(struct mlx5_devcom *devcom,
-			    enum mlx5_devcom_components id,
-			    bool paired)
+void mlx5_devcom_comp_set_ready(struct mlx5_devcom *devcom,
+				enum mlx5_devcom_components id,
+				bool ready)
 {
 	struct mlx5_devcom_component *comp;
 
 	comp = &devcom->priv->components[id];
 	WARN_ON(!rwsem_is_locked(&comp->sem));
 
-	WRITE_ONCE(comp->paired, paired);
+	WRITE_ONCE(comp->ready, ready);
 }
 
-bool mlx5_devcom_is_paired(struct mlx5_devcom *devcom,
-			   enum mlx5_devcom_components id)
+bool mlx5_devcom_comp_is_ready(struct mlx5_devcom *devcom,
+			       enum mlx5_devcom_components id)
 {
 	if (IS_ERR_OR_NULL(devcom))
 		return false;
 
-	return READ_ONCE(devcom->priv->components[id].paired);
+	return READ_ONCE(devcom->priv->components[id].ready);
 }
 
-void *mlx5_devcom_get_peer_data(struct mlx5_devcom *devcom,
-				enum mlx5_devcom_components id)
+bool mlx5_devcom_for_each_peer_begin(struct mlx5_devcom *devcom,
+				     enum mlx5_devcom_components id)
 {
 	struct mlx5_devcom_component *comp;
-	int i;
 
 	if (IS_ERR_OR_NULL(devcom))
-		return NULL;
+		return false;
 
 	comp = &devcom->priv->components[id];
 	down_read(&comp->sem);
-	if (!READ_ONCE(comp->paired)) {
+	if (!READ_ONCE(comp->ready)) {
 		up_read(&comp->sem);
-		return NULL;
+		return false;
 	}
 
-	for (i = 0; i < MLX5_DEVCOM_PORTS_SUPPORTED; i++)
-		if (i != devcom->idx)
-			break;
-
-	return rcu_dereference_protected(comp->device[i].data, lockdep_is_held(&comp->sem));
+	return true;
 }
 
-void *mlx5_devcom_get_peer_data_rcu(struct mlx5_devcom *devcom, enum mlx5_devcom_components id)
-{
-	struct mlx5_devcom_component *comp;
-	int i;
-
-	if (IS_ERR_OR_NULL(devcom))
-		return NULL;
-
-	for (i = 0; i < MLX5_DEVCOM_PORTS_SUPPORTED; i++)
-		if (i != devcom->idx)
-			break;
-
-	comp = &devcom->priv->components[id];
-	/* This can change concurrently, however 'data' pointer will remain
-	 * valid for the duration of RCU read section.
-	 */
-	if (!READ_ONCE(comp->paired))
-		return NULL;
-
-	return rcu_dereference(comp->device[i].data);
-}
-
-void mlx5_devcom_release_peer_data(struct mlx5_devcom *devcom,
+void mlx5_devcom_for_each_peer_end(struct mlx5_devcom *devcom,
 				   enum mlx5_devcom_components id)
 {
 	struct mlx5_devcom_component *comp = &devcom->priv->components[id];
 
 	up_read(&comp->sem);
 }
+
+void *mlx5_devcom_get_next_peer_data(struct mlx5_devcom *devcom,
+				     enum mlx5_devcom_components id,
+				     int *i)
+{
+	struct mlx5_devcom_component *comp;
+	void *ret;
+	int idx;
+
+	comp = &devcom->priv->components[id];
+
+	if (*i == MLX5_DEVCOM_PORTS_SUPPORTED)
+		return NULL;
+	for (idx = *i; idx < MLX5_DEVCOM_PORTS_SUPPORTED; idx++) {
+		if (idx != devcom->idx) {
+			ret = rcu_dereference_protected(comp->device[idx].data,
+							lockdep_is_held(&comp->sem));
+			if (ret)
+				break;
+		}
+	}
+
+	if (idx == MLX5_DEVCOM_PORTS_SUPPORTED) {
+		*i = idx;
+		return NULL;
+	}
+	*i = idx + 1;
+
+	return ret;
+}
+
+void *mlx5_devcom_get_next_peer_data_rcu(struct mlx5_devcom *devcom,
+					 enum mlx5_devcom_components id,
+					 int *i)
+{
+	struct mlx5_devcom_component *comp;
+	void *ret;
+	int idx;
+
+	comp = &devcom->priv->components[id];
+
+	if (*i == MLX5_DEVCOM_PORTS_SUPPORTED)
+		return NULL;
+	for (idx = *i; idx < MLX5_DEVCOM_PORTS_SUPPORTED; idx++) {
+		if (idx != devcom->idx) {
+			/* This can change concurrently, however 'data' pointer will remain
+			 * valid for the duration of RCU read section.
+			 */
+			if (!READ_ONCE(comp->ready))
+				return NULL;
+			ret = rcu_dereference(comp->device[idx].data);
+			if (ret)
+				break;
+		}
+	}
+
+	if (idx == MLX5_DEVCOM_PORTS_SUPPORTED) {
+		*i = idx;
+		return NULL;
+	}
+	*i = idx + 1;
+
+	return ret;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h
index 9a496f4722da..bb1970ba8730 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h
@@ -30,20 +30,33 @@ void mlx5_devcom_unregister_component(struct mlx5_devcom *devcom,
 
 int mlx5_devcom_send_event(struct mlx5_devcom *devcom,
 			   enum mlx5_devcom_components id,
-			   int event,
+			   int event, int rollback_event,
 			   void *event_data);
 
-void mlx5_devcom_set_paired(struct mlx5_devcom *devcom,
-			    enum mlx5_devcom_components id,
-			    bool paired);
-bool mlx5_devcom_is_paired(struct mlx5_devcom *devcom,
-			   enum mlx5_devcom_components id);
+void mlx5_devcom_comp_set_ready(struct mlx5_devcom *devcom,
+				enum mlx5_devcom_components id,
+				bool ready);
+bool mlx5_devcom_comp_is_ready(struct mlx5_devcom *devcom,
+			       enum mlx5_devcom_components id);
 
-void *mlx5_devcom_get_peer_data(struct mlx5_devcom *devcom,
-				enum mlx5_devcom_components id);
-void *mlx5_devcom_get_peer_data_rcu(struct mlx5_devcom *devcom, enum mlx5_devcom_components id);
-void mlx5_devcom_release_peer_data(struct mlx5_devcom *devcom,
+bool mlx5_devcom_for_each_peer_begin(struct mlx5_devcom *devcom,
+				     enum mlx5_devcom_components id);
+void mlx5_devcom_for_each_peer_end(struct mlx5_devcom *devcom,
 				   enum mlx5_devcom_components id);
+void *mlx5_devcom_get_next_peer_data(struct mlx5_devcom *devcom,
+				     enum mlx5_devcom_components id, int *i);
+
+#define mlx5_devcom_for_each_peer_entry(devcom, id, data, i)			\
+	for (i = 0, data = mlx5_devcom_get_next_peer_data(devcom, id, &i);	\
+	     data;								\
+	     data = mlx5_devcom_get_next_peer_data(devcom, id, &i))
+
+void *mlx5_devcom_get_next_peer_data_rcu(struct mlx5_devcom *devcom,
+					 enum mlx5_devcom_components id, int *i);
+
+#define mlx5_devcom_for_each_peer_entry_rcu(devcom, id, data, i)		\
+	for (i = 0, data = mlx5_devcom_get_next_peer_data_rcu(devcom, id, &i);	\
+	     data;								\
+	     data = mlx5_devcom_get_next_peer_data_rcu(devcom, id, &i))
 
 #endif
-
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
index 0eb9a8d7f282..4e9bc1897a88 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
@@ -2071,8 +2071,9 @@ mlx5dr_action_create_dest_vport(struct mlx5dr_domain *dmn,
 	struct mlx5dr_action *action;
 	u8 peer_vport;
 
-	peer_vport = vhca_id_valid && (vhca_id != dmn->info.caps.gvmi);
-	vport_dmn = peer_vport ? dmn->peer_dmn : dmn;
+	peer_vport = vhca_id_valid && mlx5_core_is_pf(dmn->mdev) &&
+		(vhca_id != dmn->info.caps.gvmi);
+	vport_dmn = peer_vport ? dmn->peer_dmn[vhca_id] : dmn;
 	if (!vport_dmn) {
 		mlx5dr_dbg(dmn, "No peer vport domain for given vhca_id\n");
 		return NULL;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_domain.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_domain.c
index 9a2dfe6ebe31..75dc85dc24ef 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_domain.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_domain.c
@@ -555,17 +555,18 @@ int mlx5dr_domain_destroy(struct mlx5dr_domain *dmn)
 }
 
 void mlx5dr_domain_set_peer(struct mlx5dr_domain *dmn,
-			    struct mlx5dr_domain *peer_dmn)
+			    struct mlx5dr_domain *peer_dmn,
+			    u8 peer_idx)
 {
 	mlx5dr_domain_lock(dmn);
 
-	if (dmn->peer_dmn)
-		refcount_dec(&dmn->peer_dmn->refcount);
+	if (dmn->peer_dmn[peer_idx])
+		refcount_dec(&dmn->peer_dmn[peer_idx]->refcount);
 
-	dmn->peer_dmn = peer_dmn;
+	dmn->peer_dmn[peer_idx] = peer_dmn;
 
-	if (dmn->peer_dmn)
-		refcount_inc(&dmn->peer_dmn->refcount);
+	if (dmn->peer_dmn[peer_idx])
+		refcount_inc(&dmn->peer_dmn[peer_idx]->refcount);
 
 	mlx5dr_domain_unlock(dmn);
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v0.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v0.c
index 2010d4ac6519..69d7a8f3c402 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v0.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v0.c
@@ -1647,6 +1647,7 @@ dr_ste_v0_build_src_gvmi_qpn_tag(struct mlx5dr_match_param *value,
 				 u8 *tag)
 {
 	struct mlx5dr_match_misc *misc = &value->misc;
+	int id = misc->source_eswitch_owner_vhca_id;
 	struct mlx5dr_cmd_vport_cap *vport_cap;
 	struct mlx5dr_domain *dmn = sb->dmn;
 	struct mlx5dr_domain *vport_dmn;
@@ -1657,11 +1658,11 @@ dr_ste_v0_build_src_gvmi_qpn_tag(struct mlx5dr_match_param *value,
 
 	if (sb->vhca_id_valid) {
 		/* Find port GVMI based on the eswitch_owner_vhca_id */
-		if (misc->source_eswitch_owner_vhca_id == dmn->info.caps.gvmi)
+		if (id == dmn->info.caps.gvmi)
 			vport_dmn = dmn;
-		else if (dmn->peer_dmn && (misc->source_eswitch_owner_vhca_id ==
-					   dmn->peer_dmn->info.caps.gvmi))
-			vport_dmn = dmn->peer_dmn;
+		else if (id < MLX5_MAX_PORTS && dmn->peer_dmn[id] &&
+			 (id == dmn->peer_dmn[id]->info.caps.gvmi))
+			vport_dmn = dmn->peer_dmn[id];
 		else
 			return -EINVAL;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c
index 4c0704ad166b..f4ef0b22b991 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste_v1.c
@@ -1979,6 +1979,7 @@ static int dr_ste_v1_build_src_gvmi_qpn_tag(struct mlx5dr_match_param *value,
 					    u8 *tag)
 {
 	struct mlx5dr_match_misc *misc = &value->misc;
+	int id = misc->source_eswitch_owner_vhca_id;
 	struct mlx5dr_cmd_vport_cap *vport_cap;
 	struct mlx5dr_domain *dmn = sb->dmn;
 	struct mlx5dr_domain *vport_dmn;
@@ -1988,11 +1989,11 @@ static int dr_ste_v1_build_src_gvmi_qpn_tag(struct mlx5dr_match_param *value,
 
 	if (sb->vhca_id_valid) {
 		/* Find port GVMI based on the eswitch_owner_vhca_id */
-		if (misc->source_eswitch_owner_vhca_id == dmn->info.caps.gvmi)
+		if (id == dmn->info.caps.gvmi)
 			vport_dmn = dmn;
-		else if (dmn->peer_dmn && (misc->source_eswitch_owner_vhca_id ==
-					   dmn->peer_dmn->info.caps.gvmi))
-			vport_dmn = dmn->peer_dmn;
+		else if (id < MLX5_MAX_PORTS && dmn->peer_dmn[id] &&
+			 (id == dmn->peer_dmn[id]->info.caps.gvmi))
+			vport_dmn = dmn->peer_dmn[id];
 		else
 			return -EINVAL;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
index 678a993ab053..1622dbbe6b97 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
@@ -935,7 +935,7 @@ struct mlx5dr_domain_info {
 };
 
 struct mlx5dr_domain {
-	struct mlx5dr_domain *peer_dmn;
+	struct mlx5dr_domain *peer_dmn[MLX5_MAX_PORTS];
 	struct mlx5_core_dev *mdev;
 	u32 pdn;
 	struct mlx5_uars_page *uar;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c
index 984653756779..c6fda1cbfcff 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c
@@ -770,14 +770,15 @@ restore_fte:
 }
 
 static int mlx5_cmd_dr_set_peer(struct mlx5_flow_root_namespace *ns,
-				struct mlx5_flow_root_namespace *peer_ns)
+				struct mlx5_flow_root_namespace *peer_ns,
+				u8 peer_idx)
 {
 	struct mlx5dr_domain *peer_domain = NULL;
 
 	if (peer_ns)
 		peer_domain = peer_ns->fs_dr_domain.dr_domain;
 	mlx5dr_domain_set_peer(ns->fs_dr_domain.dr_domain,
-			       peer_domain);
+			       peer_domain, peer_idx);
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h
index 9afd268a2573..5ba88f2ecb3f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h
@@ -48,7 +48,8 @@ int mlx5dr_domain_destroy(struct mlx5dr_domain *domain);
 int mlx5dr_domain_sync(struct mlx5dr_domain *domain, u32 flags);
 
 void mlx5dr_domain_set_peer(struct mlx5dr_domain *dmn,
-			    struct mlx5dr_domain *peer_dmn);
+			    struct mlx5dr_domain *peer_dmn,
+			    u8 peer_idx);
 
 struct mlx5dr_table *
 mlx5dr_table_create(struct mlx5dr_domain *domain, u32 level, u32 flags,