Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next

Pull networking updates from David Miller:

 1) Add WireGuard

 2) Add HE and TWT support to ath11k driver, from John Crispin.

 3) Add ESP in TCP encapsulation support, from Sabrina Dubroca.

 4) Add variable window congestion control to TIPC, from Jon Maloy.

 5) Add BCM84881 PHY driver, from Russell King.

 6) Start adding netlink support for ethtool operations, from Michal
    Kubecek.

 7) Add XDP drop and TX action support to ena driver, from Sameeh
    Jubran.

 8) Add new ipv4 route notifications so that mlxsw driver does not have
    to handle identical routes itself. From Ido Schimmel.

 9) Add BPF dynamic program extensions, from Alexei Starovoitov.

10) Support RX and TX timestamping in igc, from Vinicius Costa Gomes.

11) Add support for macsec HW offloading, from Antoine Tenart.

12) Add initial support for MPTCP protocol, from Christoph Paasch,
    Matthieu Baerts, Florian Westphal, Peter Krystad, and many others.

13) Add Octeontx2 PF support, from Sunil Goutham, Geetha sowjanya, Linu
    Cherian, and others.

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1469 commits)
  net: phy: add default ARCH_BCM_IPROC for MDIO_BCM_IPROC
  udp: segment looped gso packets correctly
  netem: change mailing list
  qed: FW 8.42.2.0 debug features
  qed: rt init valid initialization changed
  qed: Debug feature: ilt and mdump
  qed: FW 8.42.2.0 Add fw overlay feature
  qed: FW 8.42.2.0 HSI changes
  qed: FW 8.42.2.0 iscsi/fcoe changes
  qed: Add abstraction for different hsi values per chip
  qed: FW 8.42.2.0 Additional ll2 type
  qed: Use dmae to write to widebus registers in fw_funcs
  qed: FW 8.42.2.0 Parser offsets modified
  qed: FW 8.42.2.0 Queue Manager changes
  qed: FW 8.42.2.0 Expose new registers and change windows
  qed: FW 8.42.2.0 Internal ram offsets modifications
  MAINTAINERS: Add entry for Marvell OcteonTX2 Physical Function driver
  Documentation: net: octeontx2: Add RVU HW and drivers overview
  octeontx2-pf: ethtool RSS config support
  octeontx2-pf: Add basic ethtool support
  ...

53 files changed, 3736 insertions, 1624 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
index 7af75b63245f..43dcbd8214c6 100644
--- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
+++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
@@ -1363,24 +1363,18 @@ static void mlx4_en_delete_rss_steer_rules(struct mlx4_en_priv *priv)
 	}
 }
 
-static void mlx4_en_tx_timeout(struct net_device *dev)
+static void mlx4_en_tx_timeout(struct net_device *dev, unsigned int txqueue)
 {
 	struct mlx4_en_priv *priv = netdev_priv(dev);
 	struct mlx4_en_dev *mdev = priv->mdev;
-	int i;
+	struct mlx4_en_tx_ring *tx_ring = priv->tx_ring[TX][txqueue];
 
 	if (netif_msg_timer(priv))
 		en_warn(priv, "Tx timeout called on port:%d\n", priv->port);
 
-	for (i = 0; i < priv->tx_ring_num[TX]; i++) {
-		struct mlx4_en_tx_ring *tx_ring = priv->tx_ring[TX][i];
-
-		if (!netif_tx_queue_stopped(netdev_get_tx_queue(dev, i)))
-			continue;
-		en_warn(priv, "TX timeout on queue: %d, QP: 0x%x, CQ: 0x%x, Cons: 0x%x, Prod: 0x%x\n",
-			i, tx_ring->qpn, tx_ring->sp_cqn,
-			tx_ring->cons, tx_ring->prod);
-	}
+	en_warn(priv, "TX timeout on queue: %d, QP: 0x%x, CQ: 0x%x, Cons: 0x%x, Prod: 0x%x\n",
+		txqueue, tx_ring->qpn, tx_ring->sp_cqn,
+		tx_ring->cons, tx_ring->prod);
 
 	priv->port_stats.tx_timeout++;
 	en_dbg(DRV, priv, "Scheduling watchdog\n");
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index a6f390fdb971..d3e06cec8317 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -42,7 +42,7 @@ mlx5_core-$(CONFIG_PCI_HYPERV_INTERFACE) += en/hv_vhca_stats.o
 # Core extra
 #
 mlx5_core-$(CONFIG_MLX5_ESWITCH)   += eswitch.o eswitch_offloads.o eswitch_offloads_termtbl.o \
-				      ecpf.o rdma.o
+				      ecpf.o rdma.o eswitch_offloads_chains.o
 mlx5_core-$(CONFIG_MLX5_MPFS)      += lib/mpfs.o
 mlx5_core-$(CONFIG_VXLAN)          += lib/vxlan.o
 mlx5_core-$(CONFIG_PTP_1588_CLOCK) += lib/clock.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/alloc.c b/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
index 549f962cd86e..42198e64a7f4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/alloc.c
@@ -71,8 +71,8 @@ static void *mlx5_dma_zalloc_coherent_node(struct mlx5_core_dev *dev,
 	return cpu_handle;
 }
 
-int mlx5_buf_alloc_node(struct mlx5_core_dev *dev, int size,
-			struct mlx5_frag_buf *buf, int node)
+static int mlx5_buf_alloc_node(struct mlx5_core_dev *dev, int size,
+			       struct mlx5_frag_buf *buf, int node)
 {
 	dma_addr_t t;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 9c8427698238..220ef9f06f84 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -135,7 +135,7 @@ struct page_pool;
 #define MLX5E_LOG_INDIR_RQT_SIZE       0x7
 #define MLX5E_INDIR_RQT_SIZE           BIT(MLX5E_LOG_INDIR_RQT_SIZE)
 #define MLX5E_MIN_NUM_CHANNELS         0x1
-#define MLX5E_MAX_NUM_CHANNELS         (MLX5E_INDIR_RQT_SIZE >> 1)
+#define MLX5E_MAX_NUM_CHANNELS         MLX5E_INDIR_RQT_SIZE
 #define MLX5E_MAX_NUM_SQS              (MLX5E_MAX_NUM_CHANNELS * MLX5E_MAX_NUM_TC)
 #define MLX5E_TX_CQ_POLL_BUDGET        128
 #define MLX5E_TX_XSK_POLL_BUDGET       64
@@ -892,6 +892,8 @@ struct mlx5e_profile {
 	int	(*update_rx)(struct mlx5e_priv *priv);
 	void	(*update_stats)(struct mlx5e_priv *priv);
 	void	(*update_carrier)(struct mlx5e_priv *priv);
+	unsigned int (*stats_grps_num)(struct mlx5e_priv *priv);
+	mlx5e_stats_grp_t *stats_grps;
 	struct {
 		mlx5e_fp_handle_rx_cqe handle_rx_cqe;
 		mlx5e_fp_handle_rx_cqe handle_rx_cqe_mpwqe;
@@ -964,7 +966,6 @@ struct sk_buff *
 mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
 			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
 
-void mlx5e_update_stats(struct mlx5e_priv *priv);
 void mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats);
 void mlx5e_fold_sw_stats64(struct mlx5e_priv *priv, struct rtnl_link_stats64 *s);
 
@@ -1175,11 +1176,11 @@ int mlx5e_attach_netdev(struct mlx5e_priv *priv);
 void mlx5e_detach_netdev(struct mlx5e_priv *priv);
 void mlx5e_destroy_netdev(struct mlx5e_priv *priv);
 void mlx5e_set_netdev_mtu_boundaries(struct mlx5e_priv *priv);
-void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
+void mlx5e_build_nic_params(struct mlx5e_priv *priv,
 			    struct mlx5e_xsk *xsk,
 			    struct mlx5e_rss_params *rss_params,
 			    struct mlx5e_params *params,
-			    u16 max_channels, u16 mtu);
+			    u16 mtu);
 void mlx5e_build_rq_params(struct mlx5_core_dev *mdev,
 			   struct mlx5e_params *params);
 void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h
index d48292ccda29..0416f7712109 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/fs.h
@@ -21,6 +21,7 @@ struct mlx5e_tc_table {
 	DECLARE_HASHTABLE(hairpin_tbl, 8);
 
 	struct notifier_block     netdevice_nb;
+	struct netdev_net_notifier	netdevice_nn;
 };
 
 struct mlx5e_flow_table {
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c
index 475b6bd5d29b..62fc8a128a8d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/xsk/rx.c
@@ -35,7 +35,7 @@ int mlx5e_xsk_page_alloc_umem(struct mlx5e_rq *rq,
 	 */
 	dma_info->addr = xdp_umem_get_dma(umem, handle);
 
-	xsk_umem_discard_addr_rq(umem);
+	xsk_umem_release_addr_rq(umem);
 
 	dma_sync_single_for_device(rq->pdev, dma_info->addr, PAGE_SIZE,
 				   DMA_BIDIRECTIONAL);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index c6776f308d5e..d674cb679895 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -218,13 +218,9 @@ static const struct pflag_desc mlx5e_priv_flags[MLX5E_NUM_PFLAGS];
 
 int mlx5e_ethtool_get_sset_count(struct mlx5e_priv *priv, int sset)
 {
-	int i, num_stats = 0;
-
 	switch (sset) {
 	case ETH_SS_STATS:
-		for (i = 0; i < mlx5e_num_stats_grps; i++)
-			num_stats += mlx5e_stats_grps[i].get_num_stats(priv);
-		return num_stats;
+		return mlx5e_stats_total_num(priv);
 	case ETH_SS_PRIV_FLAGS:
 		return MLX5E_NUM_PFLAGS;
 	case ETH_SS_TEST:
@@ -242,14 +238,6 @@ static int mlx5e_get_sset_count(struct net_device *dev, int sset)
 	return mlx5e_ethtool_get_sset_count(priv, sset);
 }
 
-static void mlx5e_fill_stats_strings(struct mlx5e_priv *priv, u8 *data)
-{
-	int i, idx = 0;
-
-	for (i = 0; i < mlx5e_num_stats_grps; i++)
-		idx = mlx5e_stats_grps[i].fill_strings(priv, data, idx);
-}
-
 void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv, u32 stringset, u8 *data)
 {
 	int i;
@@ -268,7 +256,7 @@ void mlx5e_ethtool_get_strings(struct mlx5e_priv *priv, u32 stringset, u8 *data)
 		break;
 
 	case ETH_SS_STATS:
-		mlx5e_fill_stats_strings(priv, data);
+		mlx5e_stats_fill_strings(priv, data);
 		break;
 	}
 }
@@ -283,14 +271,13 @@ static void mlx5e_get_strings(struct net_device *dev, u32 stringset, u8 *data)
 void mlx5e_ethtool_get_ethtool_stats(struct mlx5e_priv *priv,
 				     struct ethtool_stats *stats, u64 *data)
 {
-	int i, idx = 0;
+	int idx = 0;
 
 	mutex_lock(&priv->state_lock);
-	mlx5e_update_stats(priv);
+	mlx5e_stats_update(priv);
 	mutex_unlock(&priv->state_lock);
 
-	for (i = 0; i < mlx5e_num_stats_grps; i++)
-		idx = mlx5e_stats_grps[i].fill_stats(priv, data, idx);
+	mlx5e_stats_fill(priv, data, idx);
 }
 
 static void mlx5e_get_ethtool_stats(struct net_device *dev,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
index acd946f2ddbe..3bc2ac3d53fc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs_ethtool.c
@@ -58,6 +58,7 @@ static struct mlx5e_ethtool_table *get_flow_table(struct mlx5e_priv *priv,
 						  struct ethtool_rx_flow_spec *fs,
 						  int num_tuples)
 {
+	struct mlx5_flow_table_attr ft_attr = {};
 	struct mlx5e_ethtool_table *eth_ft;
 	struct mlx5_flow_namespace *ns;
 	struct mlx5_flow_table *ft;
@@ -102,9 +103,11 @@ static struct mlx5e_ethtool_table *get_flow_table(struct mlx5e_priv *priv,
 	table_size = min_t(u32, BIT(MLX5_CAP_FLOWTABLE(priv->mdev,
 						       flow_table_properties_nic_receive.log_max_ft_size)),
 			   MLX5E_ETHTOOL_NUM_ENTRIES);
-	ft = mlx5_create_auto_grouped_flow_table(ns, prio,
-						 table_size,
-						 MLX5E_ETHTOOL_NUM_GROUPS, 0, 0);
+
+	ft_attr.prio = prio;
+	ft_attr.max_fte = table_size;
+	ft_attr.autogroup.max_num_groups = MLX5E_ETHTOOL_NUM_GROUPS;
+	ft = mlx5_create_auto_grouped_flow_table(ns, &ft_attr);
 	if (IS_ERR(ft))
 		return (void *)ft;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 4997b8a51994..454d3459bd8b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -159,23 +159,14 @@ static void mlx5e_update_carrier_work(struct work_struct *work)
 	mutex_unlock(&priv->state_lock);
 }
 
-void mlx5e_update_stats(struct mlx5e_priv *priv)
-{
-	int i;
-
-	for (i = mlx5e_num_stats_grps - 1; i >= 0; i--)
-		if (mlx5e_stats_grps[i].update_stats)
-			mlx5e_stats_grps[i].update_stats(priv);
-}
-
 void mlx5e_update_ndo_stats(struct mlx5e_priv *priv)
 {
 	int i;
 
-	for (i = mlx5e_num_stats_grps - 1; i >= 0; i--)
-		if (mlx5e_stats_grps[i].update_stats_mask &
+	for (i = mlx5e_nic_stats_grps_num(priv) - 1; i >= 0; i--)
+		if (mlx5e_nic_stats_grps[i]->update_stats_mask &
 		    MLX5E_NDO_UPDATE_STATS)
-			mlx5e_stats_grps[i].update_stats(priv);
+			mlx5e_nic_stats_grps[i]->update_stats(priv);
 }
 
 static void mlx5e_update_stats_work(struct work_struct *work)
@@ -4325,7 +4316,7 @@ unlock:
 	rtnl_unlock();
 }
 
-static void mlx5e_tx_timeout(struct net_device *dev)
+static void mlx5e_tx_timeout(struct net_device *dev, unsigned int txqueue)
 {
 	struct mlx5e_priv *priv = netdev_priv(dev);
 
@@ -4739,17 +4730,19 @@ void mlx5e_build_rss_params(struct mlx5e_rss_params *rss_params,
 			tirc_default_config[tt].rx_hash_fields;
 }
 
-void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
+void mlx5e_build_nic_params(struct mlx5e_priv *priv,
 			    struct mlx5e_xsk *xsk,
 			    struct mlx5e_rss_params *rss_params,
 			    struct mlx5e_params *params,
-			    u16 max_channels, u16 mtu)
+			    u16 mtu)
 {
+	struct mlx5_core_dev *mdev = priv->mdev;
 	u8 rx_cq_period_mode;
 
 	params->sw_mtu = mtu;
 	params->hard_mtu = MLX5E_ETH_HARD_MTU;
-	params->num_channels = max_channels;
+	params->num_channels = min_t(unsigned int, MLX5E_MAX_NUM_CHANNELS / 2,
+				     priv->max_nch);
 	params->num_tc       = 1;
 
 	/* SQ */
@@ -4876,6 +4869,8 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev)
 		netdev->hw_enc_features |= NETIF_F_GSO_UDP_TUNNEL |
 					   NETIF_F_GSO_UDP_TUNNEL_CSUM;
 		netdev->gso_partial_features = NETIF_F_GSO_UDP_TUNNEL_CSUM;
+		netdev->vlan_features |= NETIF_F_GSO_UDP_TUNNEL |
+					 NETIF_F_GSO_UDP_TUNNEL_CSUM;
 	}
 
 	if (mlx5e_tunnel_proto_supported(mdev, IPPROTO_GRE)) {
@@ -4986,8 +4981,8 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
 	if (err)
 		return err;
 
-	mlx5e_build_nic_params(mdev, &priv->xsk, rss, &priv->channels.params,
-			       priv->max_nch, netdev->mtu);
+	mlx5e_build_nic_params(priv, &priv->xsk, rss, &priv->channels.params,
+			       netdev->mtu);
 
 	mlx5e_timestamp_init(priv);
 
@@ -5149,6 +5144,7 @@ static void mlx5e_nic_enable(struct mlx5e_priv *priv)
 
 static void mlx5e_nic_disable(struct mlx5e_priv *priv)
 {
+	struct net_device *netdev = priv->netdev;
 	struct mlx5_core_dev *mdev = priv->mdev;
 
 #ifdef CONFIG_MLX5_CORE_EN_DCB
@@ -5169,7 +5165,7 @@ static void mlx5e_nic_disable(struct mlx5e_priv *priv)
 		mlx5e_monitor_counter_cleanup(priv);
 
 	mlx5e_disable_async_events(priv);
-	mlx5_lag_remove(mdev);
+	mlx5_lag_remove(mdev, netdev);
 }
 
 int mlx5e_update_nic_rx(struct mlx5e_priv *priv)
@@ -5193,6 +5189,8 @@ static const struct mlx5e_profile mlx5e_nic_profile = {
 	.rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq,
 	.max_tc		   = MLX5E_MAX_NUM_TC,
 	.rq_groups	   = MLX5E_NUM_RQ_GROUPS(XSK),
+	.stats_grps	   = mlx5e_nic_stats_grps,
+	.stats_grps_num	   = mlx5e_nic_stats_grps_num,
 };
 
 /* mlx5e generic netdev management API (move to en_common.c) */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index f175cb24bb67..7b48ccacebe2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -41,6 +41,7 @@
 #include <net/ipv6_stubs.h>
 
 #include "eswitch.h"
+#include "eswitch_offloads_chains.h"
 #include "en.h"
 #include "en_rep.h"
 #include "en_tc.h"
@@ -116,24 +117,71 @@ static const struct counter_desc vport_rep_stats_desc[] = {
 #define NUM_VPORT_REP_SW_COUNTERS ARRAY_SIZE(sw_rep_stats_desc)
 #define NUM_VPORT_REP_HW_COUNTERS ARRAY_SIZE(vport_rep_stats_desc)
 
-static void mlx5e_rep_get_strings(struct net_device *dev,
-				  u32 stringset, uint8_t *data)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(sw_rep)
 {
-	int i, j;
+	return NUM_VPORT_REP_SW_COUNTERS;
+}
 
-	switch (stringset) {
-	case ETH_SS_STATS:
-		for (i = 0; i < NUM_VPORT_REP_SW_COUNTERS; i++)
-			strcpy(data + (i * ETH_GSTRING_LEN),
-			       sw_rep_stats_desc[i].format);
-		for (j = 0; j < NUM_VPORT_REP_HW_COUNTERS; j++, i++)
-			strcpy(data + (i * ETH_GSTRING_LEN),
-			       vport_rep_stats_desc[j].format);
-		break;
-	}
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(sw_rep)
+{
+	int i;
+
+	for (i = 0; i < NUM_VPORT_REP_SW_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN,
+		       sw_rep_stats_desc[i].format);
+	return idx;
+}
+
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(sw_rep)
+{
+	int i;
+
+	for (i = 0; i < NUM_VPORT_REP_SW_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_CPU(&priv->stats.sw,
+						   sw_rep_stats_desc, i);
+	return idx;
+}
+
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(sw_rep)
+{
+	struct mlx5e_sw_stats *s = &priv->stats.sw;
+	struct rtnl_link_stats64 stats64 = {};
+
+	memset(s, 0, sizeof(*s));
+	mlx5e_fold_sw_stats64(priv, &stats64);
+
+	s->rx_packets = stats64.rx_packets;
+	s->rx_bytes   = stats64.rx_bytes;
+	s->tx_packets = stats64.tx_packets;
+	s->tx_bytes   = stats64.tx_bytes;
+	s->tx_queue_dropped = stats64.tx_dropped;
+}
+
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(vport_rep)
+{
+	return NUM_VPORT_REP_HW_COUNTERS;
+}
+
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(vport_rep)
+{
+	int i;
+
+	for (i = 0; i < NUM_VPORT_REP_HW_COUNTERS; i++)
+		strcpy(data + (idx++) * ETH_GSTRING_LEN, vport_rep_stats_desc[i].format);
+	return idx;
 }
 
-static void mlx5e_rep_update_hw_counters(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(vport_rep)
+{
+	int i;
+
+	for (i = 0; i < NUM_VPORT_REP_HW_COUNTERS; i++)
+		data[idx++] = MLX5E_READ_CTR64_CPU(&priv->stats.vf_vport,
+						   vport_rep_stats_desc, i);
+	return idx;
+}
+
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(vport_rep)
 {
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
 	struct mlx5e_rep_priv *rpriv = priv->ppriv;
@@ -156,64 +204,33 @@ static void mlx5e_rep_update_hw_counters(struct mlx5e_priv *priv)
 	vport_stats->tx_bytes   = vf_stats.rx_bytes;
 }
 
-static void mlx5e_uplink_rep_update_hw_counters(struct mlx5e_priv *priv)
-{
-	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
-	struct rtnl_link_stats64 *vport_stats;
-
-	mlx5e_grp_802_3_update_stats(priv);
-
-	vport_stats = &priv->stats.vf_vport;
-
-	vport_stats->rx_packets = PPORT_802_3_GET(pstats, a_frames_received_ok);
-	vport_stats->rx_bytes   = PPORT_802_3_GET(pstats, a_octets_received_ok);
-	vport_stats->tx_packets = PPORT_802_3_GET(pstats, a_frames_transmitted_ok);
-	vport_stats->tx_bytes   = PPORT_802_3_GET(pstats, a_octets_transmitted_ok);
-}
-
-static void mlx5e_rep_update_sw_counters(struct mlx5e_priv *priv)
+static void mlx5e_rep_get_strings(struct net_device *dev,
+				  u32 stringset, uint8_t *data)
 {
-	struct mlx5e_sw_stats *s = &priv->stats.sw;
-	struct rtnl_link_stats64 stats64 = {};
-
-	memset(s, 0, sizeof(*s));
-	mlx5e_fold_sw_stats64(priv, &stats64);
+	struct mlx5e_priv *priv = netdev_priv(dev);
 
-	s->rx_packets = stats64.rx_packets;
-	s->rx_bytes   = stats64.rx_bytes;
-	s->tx_packets = stats64.tx_packets;
-	s->tx_bytes   = stats64.tx_bytes;
-	s->tx_queue_dropped = stats64.tx_dropped;
+	switch (stringset) {
+	case ETH_SS_STATS:
+		mlx5e_stats_fill_strings(priv, data);
+		break;
+	}
 }
 
 static void mlx5e_rep_get_ethtool_stats(struct net_device *dev,
 					struct ethtool_stats *stats, u64 *data)
 {
 	struct mlx5e_priv *priv = netdev_priv(dev);
-	int i, j;
-
-	if (!data)
-		return;
-
-	mutex_lock(&priv->state_lock);
-	mlx5e_rep_update_sw_counters(priv);
-	priv->profile->update_stats(priv);
-	mutex_unlock(&priv->state_lock);
 
-	for (i = 0; i < NUM_VPORT_REP_SW_COUNTERS; i++)
-		data[i] = MLX5E_READ_CTR64_CPU(&priv->stats.sw,
-					       sw_rep_stats_desc, i);
-
-	for (j = 0; j < NUM_VPORT_REP_HW_COUNTERS; j++, i++)
-		data[i] = MLX5E_READ_CTR64_CPU(&priv->stats.vf_vport,
-					       vport_rep_stats_desc, j);
+	mlx5e_ethtool_get_ethtool_stats(priv, stats, data);
 }
 
 static int mlx5e_rep_get_sset_count(struct net_device *dev, int sset)
 {
+	struct mlx5e_priv *priv = netdev_priv(dev);
+
 	switch (sset) {
 	case ETH_SS_STATS:
-		return NUM_VPORT_REP_SW_COUNTERS + NUM_VPORT_REP_HW_COUNTERS;
+		return mlx5e_stats_total_num(priv);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -1247,8 +1264,7 @@ static int mlx5e_rep_setup_tc_cb(enum tc_setup_type type, void *type_data,
 static int mlx5e_rep_setup_ft_cb(enum tc_setup_type type, void *type_data,
 				 void *cb_priv)
 {
-	struct flow_cls_offload *f = type_data;
-	struct flow_cls_offload cls_flower;
+	struct flow_cls_offload tmp, *f = type_data;
 	struct mlx5e_priv *priv = cb_priv;
 	struct mlx5_eswitch *esw;
 	unsigned long flags;
@@ -1261,16 +1277,30 @@ static int mlx5e_rep_setup_ft_cb(enum tc_setup_type type, void *type_data,
 
 	switch (type) {
 	case TC_SETUP_CLSFLOWER:
-		if (!mlx5_eswitch_prios_supported(esw) || f->common.chain_index)
+		memcpy(&tmp, f, sizeof(*f));
+
+		if (!mlx5_esw_chains_prios_supported(esw) ||
+		    tmp.common.chain_index)
 			return -EOPNOTSUPP;
 
 		/* Re-use tc offload path by moving the ft flow to the
 		 * reserved ft chain.
+		 *
+		 * FT offload can use prio range [0, INT_MAX], so we normalize
+		 * it to range [1, mlx5_esw_chains_get_prio_range(esw)]
+		 * as with tc, where prio 0 isn't supported.
+		 *
+		 * We only support chain 0 of FT offload.
 		 */
-		memcpy(&cls_flower, f, sizeof(*f));
-		cls_flower.common.chain_index = FDB_FT_CHAIN;
-		err = mlx5e_rep_setup_tc_cls_flower(priv, &cls_flower, flags);
-		memcpy(&f->stats, &cls_flower.stats, sizeof(f->stats));
+		if (tmp.common.prio >= mlx5_esw_chains_get_prio_range(esw))
+			return -EOPNOTSUPP;
+		if (tmp.common.chain_index != 0)
+			return -EOPNOTSUPP;
+
+		tmp.common.chain_index = mlx5_esw_chains_get_ft_chain(esw);
+		tmp.common.prio++;
+		err = mlx5e_rep_setup_tc_cls_flower(priv, &tmp, flags);
+		memcpy(&f->stats, &tmp.stats, sizeof(f->stats));
 		return err;
 	default:
 		return -EOPNOTSUPP;
@@ -1660,10 +1690,65 @@ static void mlx5e_cleanup_rep_rx(struct mlx5e_priv *priv)
 	mlx5e_close_drop_rq(&priv->drop_rq);
 }
 
+static int mlx5e_init_ul_rep_rx(struct mlx5e_priv *priv)
+{
+	int err = mlx5e_init_rep_rx(priv);
+
+	if (err)
+		return err;
+
+	mlx5e_create_q_counters(priv);
+	return 0;
+}
+
+static void mlx5e_cleanup_ul_rep_rx(struct mlx5e_priv *priv)
+{
+	mlx5e_destroy_q_counters(priv);
+	mlx5e_cleanup_rep_rx(priv);
+}
+
+static int mlx5e_init_uplink_rep_tx(struct mlx5e_rep_priv *rpriv)
+{
+	struct mlx5_rep_uplink_priv *uplink_priv;
+	struct net_device *netdev;
+	struct mlx5e_priv *priv;
+	int err;
+
+	netdev = rpriv->netdev;
+	priv = netdev_priv(netdev);
+	uplink_priv = &rpriv->uplink_priv;
+
+	mutex_init(&uplink_priv->unready_flows_lock);
+	INIT_LIST_HEAD(&uplink_priv->unready_flows);
+
+	/* init shared tc flow table */
+	err = mlx5e_tc_esw_init(&uplink_priv->tc_ht);
+	if (err)
+		return err;
+
+	mlx5_init_port_tun_entropy(&uplink_priv->tun_entropy, priv->mdev);
+
+	/* init indirect block notifications */
+	INIT_LIST_HEAD(&uplink_priv->tc_indr_block_priv_list);
+	uplink_priv->netdevice_nb.notifier_call = mlx5e_nic_rep_netdevice_event;
+	err = register_netdevice_notifier_dev_net(rpriv->netdev,
+						  &uplink_priv->netdevice_nb,
+						  &uplink_priv->netdevice_nn);
+	if (err) {
+		mlx5_core_err(priv->mdev, "Failed to register netdev notifier\n");
+		goto tc_esw_cleanup;
+	}
+
+	return 0;
+
+tc_esw_cleanup:
+	mlx5e_tc_esw_cleanup(&uplink_priv->tc_ht);
+	return err;
+}
+
 static int mlx5e_init_rep_tx(struct mlx5e_priv *priv)
 {
 	struct mlx5e_rep_priv *rpriv = priv->ppriv;
-	struct mlx5_rep_uplink_priv *uplink_priv;
 	int err;
 
 	err = mlx5e_create_tises(priv);
@@ -1673,52 +1758,41 @@ static int mlx5e_init_rep_tx(struct mlx5e_priv *priv)
 	}
 
 	if (rpriv->rep->vport == MLX5_VPORT_UPLINK) {
-		uplink_priv = &rpriv->uplink_priv;
-
-		mutex_init(&uplink_priv->unready_flows_lock);
-		INIT_LIST_HEAD(&uplink_priv->unready_flows);
-
-		/* init shared tc flow table */
-		err = mlx5e_tc_esw_init(&uplink_priv->tc_ht);
+		err = mlx5e_init_uplink_rep_tx(rpriv);
 		if (err)
 			goto destroy_tises;
-
-		mlx5_init_port_tun_entropy(&uplink_priv->tun_entropy, priv->mdev);
-
-		/* init indirect block notifications */
-		INIT_LIST_HEAD(&uplink_priv->tc_indr_block_priv_list);
-		uplink_priv->netdevice_nb.notifier_call = mlx5e_nic_rep_netdevice_event;
-		err = register_netdevice_notifier(&uplink_priv->netdevice_nb);
-		if (err) {
-			mlx5_core_err(priv->mdev, "Failed to register netdev notifier\n");
-			goto tc_esw_cleanup;
-		}
 	}
 
 	return 0;
 
-tc_esw_cleanup:
-	mlx5e_tc_esw_cleanup(&uplink_priv->tc_ht);
 destroy_tises:
 	mlx5e_destroy_tises(priv);
 	return err;
 }
 
+static void mlx5e_cleanup_uplink_rep_tx(struct mlx5e_rep_priv *rpriv)
+{
+	struct mlx5_rep_uplink_priv *uplink_priv = &rpriv->uplink_priv;
+
+	/* clean indirect TC block notifications */
+	unregister_netdevice_notifier_dev_net(rpriv->netdev,
+					      &uplink_priv->netdevice_nb,
+					      &uplink_priv->netdevice_nn);
+	mlx5e_rep_indr_clean_block_privs(rpriv);
+
+	/* delete shared tc flow table */
+	mlx5e_tc_esw_cleanup(&rpriv->uplink_priv.tc_ht);
+	mutex_destroy(&rpriv->uplink_priv.unready_flows_lock);
+}
+
 static void mlx5e_cleanup_rep_tx(struct mlx5e_priv *priv)
 {
 	struct mlx5e_rep_priv *rpriv = priv->ppriv;
 
 	mlx5e_destroy_tises(priv);
 
-	if (rpriv->rep->vport == MLX5_VPORT_UPLINK) {
-		/* clean indirect TC block notifications */
-		unregister_netdevice_notifier(&rpriv->uplink_priv.netdevice_nb);
-		mlx5e_rep_indr_clean_block_privs(rpriv);
-
-		/* delete shared tc flow table */
-		mlx5e_tc_esw_cleanup(&rpriv->uplink_priv.tc_ht);
-		mutex_destroy(&rpriv->uplink_priv.unready_flows_lock);
-	}
+	if (rpriv->rep->vport == MLX5_VPORT_UPLINK)
+		mlx5e_cleanup_uplink_rep_tx(rpriv);
 }
 
 static void mlx5e_rep_enable(struct mlx5e_priv *priv)
@@ -1787,6 +1861,7 @@ static void mlx5e_uplink_rep_enable(struct mlx5e_priv *priv)
 
 static void mlx5e_uplink_rep_disable(struct mlx5e_priv *priv)
 {
+	struct net_device *netdev = priv->netdev;
 	struct mlx5_core_dev *mdev = priv->mdev;
 	struct mlx5e_rep_priv *rpriv = priv->ppriv;
 
@@ -1795,7 +1870,44 @@ static void mlx5e_uplink_rep_disable(struct mlx5e_priv *priv)
 #endif
 	mlx5_notifier_unregister(mdev, &priv->events_nb);
 	cancel_work_sync(&rpriv->uplink_priv.reoffload_flows_work);
-	mlx5_lag_remove(mdev);
+	mlx5_lag_remove(mdev, netdev);
+}
+
+static MLX5E_DEFINE_STATS_GRP(sw_rep, 0);
+static MLX5E_DEFINE_STATS_GRP(vport_rep, MLX5E_NDO_UPDATE_STATS);
+
+/* The stats groups order is opposite to the update_stats() order calls */
+static mlx5e_stats_grp_t mlx5e_rep_stats_grps[] = {
+	&MLX5E_STATS_GRP(sw_rep),
+	&MLX5E_STATS_GRP(vport_rep),
+};
+
+static unsigned int mlx5e_rep_stats_grps_num(struct mlx5e_priv *priv)
+{
+	return ARRAY_SIZE(mlx5e_rep_stats_grps);
+}
+
+/* The stats groups order is opposite to the update_stats() order calls */
+static mlx5e_stats_grp_t mlx5e_ul_rep_stats_grps[] = {
+	&MLX5E_STATS_GRP(sw),
+	&MLX5E_STATS_GRP(qcnt),
+	&MLX5E_STATS_GRP(vnic_env),
+	&MLX5E_STATS_GRP(vport),
+	&MLX5E_STATS_GRP(802_3),
+	&MLX5E_STATS_GRP(2863),
+	&MLX5E_STATS_GRP(2819),
+	&MLX5E_STATS_GRP(phy),
+	&MLX5E_STATS_GRP(eth_ext),
+	&MLX5E_STATS_GRP(pcie),
+	&MLX5E_STATS_GRP(per_prio),
+	&MLX5E_STATS_GRP(pme),
+	&MLX5E_STATS_GRP(channels),
+	&MLX5E_STATS_GRP(per_port_buff_congest),
+};
+
+static unsigned int mlx5e_ul_rep_stats_grps_num(struct mlx5e_priv *priv)
+{
+	return ARRAY_SIZE(mlx5e_ul_rep_stats_grps);
 }
 
 static const struct mlx5e_profile mlx5e_rep_profile = {
@@ -1807,29 +1919,33 @@ static const struct mlx5e_profile mlx5e_rep_profile = {
 	.cleanup_tx		= mlx5e_cleanup_rep_tx,
 	.enable		        = mlx5e_rep_enable,
 	.update_rx		= mlx5e_update_rep_rx,
-	.update_stats           = mlx5e_rep_update_hw_counters,
+	.update_stats           = mlx5e_update_ndo_stats,
 	.rx_handlers.handle_rx_cqe       = mlx5e_handle_rx_cqe_rep,
 	.rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq,
 	.max_tc			= 1,
 	.rq_groups		= MLX5E_NUM_RQ_GROUPS(REGULAR),
+	.stats_grps		= mlx5e_rep_stats_grps,
+	.stats_grps_num		= mlx5e_rep_stats_grps_num,
 };
 
 static const struct mlx5e_profile mlx5e_uplink_rep_profile = {
 	.init			= mlx5e_init_rep,
 	.cleanup		= mlx5e_cleanup_rep,
-	.init_rx		= mlx5e_init_rep_rx,
-	.cleanup_rx		= mlx5e_cleanup_rep_rx,
+	.init_rx		= mlx5e_init_ul_rep_rx,
+	.cleanup_rx		= mlx5e_cleanup_ul_rep_rx,
 	.init_tx		= mlx5e_init_rep_tx,
 	.cleanup_tx		= mlx5e_cleanup_rep_tx,
 	.enable		        = mlx5e_uplink_rep_enable,
 	.disable	        = mlx5e_uplink_rep_disable,
 	.update_rx		= mlx5e_update_rep_rx,
-	.update_stats           = mlx5e_uplink_rep_update_hw_counters,
+	.update_stats           = mlx5e_update_ndo_stats,
 	.update_carrier	        = mlx5e_update_carrier,
 	.rx_handlers.handle_rx_cqe       = mlx5e_handle_rx_cqe_rep,
 	.rx_handlers.handle_rx_cqe_mpwqe = mlx5e_handle_rx_cqe_mpwrq,
 	.max_tc			= MLX5E_MAX_NUM_TC,
 	.rq_groups		= MLX5E_NUM_RQ_GROUPS(REGULAR),
+	.stats_grps		= mlx5e_ul_rep_stats_grps,
+	.stats_grps_num		= mlx5e_ul_rep_stats_grps_num,
 };
 
 static bool
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
index 31f83c8adcc9..3f756d51435f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.h
@@ -73,6 +73,7 @@ struct mlx5_rep_uplink_priv {
 	 */
 	struct list_head	    tc_indr_block_priv_list;
 	struct notifier_block	    netdevice_nb;
+	struct netdev_net_notifier  netdevice_nn;
 
 	struct mlx5_tun_entropy tun_entropy;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
index 9f09253f9f46..30b216d9284c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c
@@ -35,6 +35,58 @@
 #include "en_accel/ipsec.h"
 #include "en_accel/tls.h"
 
+static unsigned int stats_grps_num(struct mlx5e_priv *priv)
+{
+	return !priv->profile->stats_grps_num ? 0 :
+		priv->profile->stats_grps_num(priv);
+}
+
+unsigned int mlx5e_stats_total_num(struct mlx5e_priv *priv)
+{
+	mlx5e_stats_grp_t *stats_grps = priv->profile->stats_grps;
+	const unsigned int num_stats_grps = stats_grps_num(priv);
+	unsigned int total = 0;
+	int i;
+
+	for (i = 0; i < num_stats_grps; i++)
+		total += stats_grps[i]->get_num_stats(priv);
+
+	return total;
+}
+
+void mlx5e_stats_update(struct mlx5e_priv *priv)
+{
+	mlx5e_stats_grp_t *stats_grps = priv->profile->stats_grps;
+	const unsigned int num_stats_grps = stats_grps_num(priv);
+	int i;
+
+	for (i = num_stats_grps - 1; i >= 0; i--)
+		if (stats_grps[i]->update_stats)
+			stats_grps[i]->update_stats(priv);
+}
+
+void mlx5e_stats_fill(struct mlx5e_priv *priv, u64 *data, int idx)
+{
+	mlx5e_stats_grp_t *stats_grps = priv->profile->stats_grps;
+	const unsigned int num_stats_grps = stats_grps_num(priv);
+	int i;
+
+	for (i = 0; i < num_stats_grps; i++)
+		idx = stats_grps[i]->fill_stats(priv, data, idx);
+}
+
+void mlx5e_stats_fill_strings(struct mlx5e_priv *priv, u8 *data)
+{
+	mlx5e_stats_grp_t *stats_grps = priv->profile->stats_grps;
+	const unsigned int num_stats_grps = stats_grps_num(priv);
+	int i, idx = 0;
+
+	for (i = 0; i < num_stats_grps; i++)
+		idx = stats_grps[i]->fill_strings(priv, data, idx);
+}
+
+/* Concrete NIC Stats */
+
 static const struct counter_desc sw_stats_desc[] = {
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_packets) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_bytes) },
@@ -146,12 +198,12 @@ static const struct counter_desc sw_stats_desc[] = {
 
 #define NUM_SW_COUNTERS			ARRAY_SIZE(sw_stats_desc)
 
-static int mlx5e_grp_sw_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(sw)
 {
 	return NUM_SW_COUNTERS;
 }
 
-static int mlx5e_grp_sw_fill_strings(struct mlx5e_priv *priv, u8 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(sw)
 {
 	int i;
 
@@ -160,7 +212,7 @@ static int mlx5e_grp_sw_fill_strings(struct mlx5e_priv *priv, u8 *data, int idx)
 	return idx;
 }
 
-static int mlx5e_grp_sw_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(sw)
 {
 	int i;
 
@@ -169,7 +221,7 @@ static int mlx5e_grp_sw_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
 	return idx;
 }
 
-static void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(sw)
 {
 	struct mlx5e_sw_stats *s = &priv->stats.sw;
 	int i;
@@ -297,6 +349,9 @@ static void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv)
 			s->tx_tls_drop_bypass_req   += sq_stats->tls_drop_bypass_req;
 #endif
 			s->tx_cqes		+= sq_stats->cqes;
+
+			/* https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92657 */
+			barrier();
 		}
 	}
 }
@@ -312,7 +367,7 @@ static const struct counter_desc drop_rq_stats_desc[] = {
 #define NUM_Q_COUNTERS			ARRAY_SIZE(q_stats_desc)
 #define NUM_DROP_RQ_COUNTERS		ARRAY_SIZE(drop_rq_stats_desc)
 
-static int mlx5e_grp_q_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(qcnt)
 {
 	int num_stats = 0;
 
@@ -325,7 +380,7 @@ static int mlx5e_grp_q_get_num_stats(struct mlx5e_priv *priv)
 	return num_stats;
 }
 
-static int mlx5e_grp_q_fill_strings(struct mlx5e_priv *priv, u8 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(qcnt)
 {
 	int i;
 
@@ -340,7 +395,7 @@ static int mlx5e_grp_q_fill_strings(struct mlx5e_priv *priv, u8 *data, int idx)
 	return idx;
 }
 
-static int mlx5e_grp_q_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(qcnt)
 {
 	int i;
 
@@ -353,7 +408,7 @@ static int mlx5e_grp_q_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
 	return idx;
 }
 
-static void mlx5e_grp_q_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(qcnt)
 {
 	struct mlx5e_qcounter_stats *qcnt = &priv->stats.qcnt;
 	u32 out[MLX5_ST_SZ_DW(query_q_counter_out)];
@@ -388,14 +443,13 @@ static const struct counter_desc vnic_env_stats_dev_oob_desc[] = {
 	(MLX5_CAP_GEN(dev, vnic_env_int_rq_oob) ? \
 	 ARRAY_SIZE(vnic_env_stats_dev_oob_desc) : 0)
 
-static int mlx5e_grp_vnic_env_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(vnic_env)
 {
 	return NUM_VNIC_ENV_STEER_COUNTERS(priv->mdev) +
 		NUM_VNIC_ENV_DEV_OOB_COUNTERS(priv->mdev);
 }
 
-static int mlx5e_grp_vnic_env_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					   int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(vnic_env)
 {
 	int i;
 
@@ -409,8 +463,7 @@ static int mlx5e_grp_vnic_env_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_vnic_env_fill_stats(struct mlx5e_priv *priv, u64 *data,
-					 int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(vnic_env)
 {
 	int i;
 
@@ -424,7 +477,7 @@ static int mlx5e_grp_vnic_env_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static void mlx5e_grp_vnic_env_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(vnic_env)
 {
 	u32 *out = (u32 *)priv->stats.vnic.query_vnic_env_out;
 	int outlen = MLX5_ST_SZ_BYTES(query_vnic_env_out);
@@ -487,13 +540,12 @@ static const struct counter_desc vport_stats_desc[] = {
 
 #define NUM_VPORT_COUNTERS		ARRAY_SIZE(vport_stats_desc)
 
-static int mlx5e_grp_vport_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(vport)
 {
 	return NUM_VPORT_COUNTERS;
 }
 
-static int mlx5e_grp_vport_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(vport)
 {
 	int i;
 
@@ -502,8 +554,7 @@ static int mlx5e_grp_vport_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_vport_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				      int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(vport)
 {
 	int i;
 
@@ -513,7 +564,7 @@ static int mlx5e_grp_vport_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static void mlx5e_grp_vport_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(vport)
 {
 	int outlen = MLX5_ST_SZ_BYTES(query_vport_counter_out);
 	u32 *out = (u32 *)priv->stats.vport.query_vport_out;
@@ -552,13 +603,12 @@ static const struct counter_desc pport_802_3_stats_desc[] = {
 
 #define NUM_PPORT_802_3_COUNTERS	ARRAY_SIZE(pport_802_3_stats_desc)
 
-static int mlx5e_grp_802_3_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(802_3)
 {
 	return NUM_PPORT_802_3_COUNTERS;
 }
 
-static int mlx5e_grp_802_3_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(802_3)
 {
 	int i;
 
@@ -567,8 +617,7 @@ static int mlx5e_grp_802_3_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_802_3_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				      int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(802_3)
 {
 	int i;
 
@@ -581,7 +630,7 @@ static int mlx5e_grp_802_3_fill_stats(struct mlx5e_priv *priv, u64 *data,
 #define MLX5_BASIC_PPCNT_SUPPORTED(mdev) \
 	(MLX5_CAP_GEN(mdev, pcam_reg) ? MLX5_CAP_PCAM_REG(mdev, ppcnt) : 1)
 
-void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(802_3)
 {
 	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -609,13 +658,12 @@ static const struct counter_desc pport_2863_stats_desc[] = {
 
 #define NUM_PPORT_2863_COUNTERS		ARRAY_SIZE(pport_2863_stats_desc)
 
-static int mlx5e_grp_2863_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(2863)
 {
 	return NUM_PPORT_2863_COUNTERS;
 }
 
-static int mlx5e_grp_2863_fill_strings(struct mlx5e_priv *priv, u8 *data,
-				       int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(2863)
 {
 	int i;
 
@@ -624,8 +672,7 @@ static int mlx5e_grp_2863_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_2863_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				     int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(2863)
 {
 	int i;
 
@@ -635,7 +682,7 @@ static int mlx5e_grp_2863_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static void mlx5e_grp_2863_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(2863)
 {
 	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -670,13 +717,12 @@ static const struct counter_desc pport_2819_stats_desc[] = {
 
 #define NUM_PPORT_2819_COUNTERS		ARRAY_SIZE(pport_2819_stats_desc)
 
-static int mlx5e_grp_2819_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(2819)
 {
 	return NUM_PPORT_2819_COUNTERS;
 }
 
-static int mlx5e_grp_2819_fill_strings(struct mlx5e_priv *priv, u8 *data,
-				       int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(2819)
 {
 	int i;
 
@@ -685,8 +731,7 @@ static int mlx5e_grp_2819_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_2819_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				     int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(2819)
 {
 	int i;
 
@@ -696,7 +741,7 @@ static int mlx5e_grp_2819_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static void mlx5e_grp_2819_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(2819)
 {
 	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -734,7 +779,7 @@ pport_phy_statistical_err_lanes_stats_desc[] = {
 #define NUM_PPORT_PHY_STATISTICAL_PER_LANE_COUNTERS \
 	ARRAY_SIZE(pport_phy_statistical_err_lanes_stats_desc)
 
-static int mlx5e_grp_phy_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(phy)
 {
 	struct mlx5_core_dev *mdev = priv->mdev;
 	int num_stats;
@@ -751,8 +796,7 @@ static int mlx5e_grp_phy_get_num_stats(struct mlx5e_priv *priv)
 	return num_stats;
 }
 
-static int mlx5e_grp_phy_fill_strings(struct mlx5e_priv *priv, u8 *data,
-				      int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(phy)
 {
 	struct mlx5_core_dev *mdev = priv->mdev;
 	int i;
@@ -774,7 +818,7 @@ static int mlx5e_grp_phy_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_phy_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(phy)
 {
 	struct mlx5_core_dev *mdev = priv->mdev;
 	int i;
@@ -800,7 +844,7 @@ static int mlx5e_grp_phy_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
 	return idx;
 }
 
-static void mlx5e_grp_phy_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(phy)
 {
 	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -830,7 +874,7 @@ static const struct counter_desc pport_eth_ext_stats_desc[] = {
 
 #define NUM_PPORT_ETH_EXT_COUNTERS	ARRAY_SIZE(pport_eth_ext_stats_desc)
 
-static int mlx5e_grp_eth_ext_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(eth_ext)
 {
 	if (MLX5_CAP_PCAM_FEATURE((priv)->mdev, rx_buffer_fullness_counters))
 		return NUM_PPORT_ETH_EXT_COUNTERS;
@@ -838,8 +882,7 @@ static int mlx5e_grp_eth_ext_get_num_stats(struct mlx5e_priv *priv)
 	return 0;
 }
 
-static int mlx5e_grp_eth_ext_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					  int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(eth_ext)
 {
 	int i;
 
@@ -850,8 +893,7 @@ static int mlx5e_grp_eth_ext_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_eth_ext_fill_stats(struct mlx5e_priv *priv, u64 *data,
-					int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(eth_ext)
 {
 	int i;
 
@@ -863,7 +905,7 @@ static int mlx5e_grp_eth_ext_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static void mlx5e_grp_eth_ext_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(eth_ext)
 {
 	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -904,7 +946,7 @@ static const struct counter_desc pcie_perf_stall_stats_desc[] = {
 #define NUM_PCIE_PERF_COUNTERS64	ARRAY_SIZE(pcie_perf_stats_desc64)
 #define NUM_PCIE_PERF_STALL_COUNTERS	ARRAY_SIZE(pcie_perf_stall_stats_desc)
 
-static int mlx5e_grp_pcie_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(pcie)
 {
 	int num_stats = 0;
 
@@ -920,8 +962,7 @@ static int mlx5e_grp_pcie_get_num_stats(struct mlx5e_priv *priv)
 	return num_stats;
 }
 
-static int mlx5e_grp_pcie_fill_strings(struct mlx5e_priv *priv, u8 *data,
-				       int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(pcie)
 {
 	int i;
 
@@ -942,8 +983,7 @@ static int mlx5e_grp_pcie_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_pcie_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				     int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(pcie)
 {
 	int i;
 
@@ -967,7 +1007,7 @@ static int mlx5e_grp_pcie_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static void mlx5e_grp_pcie_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(pcie)
 {
 	struct mlx5e_pcie_stats *pcie_stats = &priv->stats.pcie;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -1015,8 +1055,7 @@ static int mlx5e_grp_per_tc_prio_get_num_stats(struct mlx5e_priv *priv)
 	return NUM_PPORT_PER_TC_PRIO_COUNTERS * NUM_PPORT_PRIO;
 }
 
-static int mlx5e_grp_per_port_buffer_congest_fill_strings(struct mlx5e_priv *priv,
-							  u8 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(per_port_buff_congest)
 {
 	struct mlx5_core_dev *mdev = priv->mdev;
 	int i, prio;
@@ -1036,8 +1075,7 @@ static int mlx5e_grp_per_port_buffer_congest_fill_strings(struct mlx5e_priv *pri
 	return idx;
 }
 
-static int mlx5e_grp_per_port_buffer_congest_fill_stats(struct mlx5e_priv *priv,
-							u64 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(per_port_buff_congest)
 {
 	struct mlx5e_pport_stats *pport = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -1112,13 +1150,13 @@ static void mlx5e_grp_per_tc_congest_prio_update_stats(struct mlx5e_priv *priv)
 	}
 }
 
-static int mlx5e_grp_per_port_buffer_congest_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(per_port_buff_congest)
 {
 	return mlx5e_grp_per_tc_prio_get_num_stats(priv) +
 		mlx5e_grp_per_tc_congest_prio_get_num_stats(priv);
 }
 
-static void mlx5e_grp_per_port_buffer_congest_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(per_port_buff_congest)
 {
 	mlx5e_grp_per_tc_prio_update_stats(priv);
 	mlx5e_grp_per_tc_congest_prio_update_stats(priv);
@@ -1130,6 +1168,7 @@ static void mlx5e_grp_per_port_buffer_congest_update_stats(struct mlx5e_priv *pr
 static const struct counter_desc pport_per_prio_traffic_stats_desc[] = {
 	{ "rx_prio%d_bytes", PPORT_PER_PRIO_OFF(rx_octets) },
 	{ "rx_prio%d_packets", PPORT_PER_PRIO_OFF(rx_frames) },
+	{ "rx_prio%d_discards", PPORT_PER_PRIO_OFF(rx_discards) },
 	{ "tx_prio%d_bytes", PPORT_PER_PRIO_OFF(tx_octets) },
 	{ "tx_prio%d_packets", PPORT_PER_PRIO_OFF(tx_frames) },
 };
@@ -1292,29 +1331,27 @@ static int mlx5e_grp_per_prio_pfc_fill_stats(struct mlx5e_priv *priv,
 	return idx;
 }
 
-static int mlx5e_grp_per_prio_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(per_prio)
 {
 	return mlx5e_grp_per_prio_traffic_get_num_stats() +
 		mlx5e_grp_per_prio_pfc_get_num_stats(priv);
 }
 
-static int mlx5e_grp_per_prio_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					   int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(per_prio)
 {
 	idx = mlx5e_grp_per_prio_traffic_fill_strings(priv, data, idx);
 	idx = mlx5e_grp_per_prio_pfc_fill_strings(priv, data, idx);
 	return idx;
 }
 
-static int mlx5e_grp_per_prio_fill_stats(struct mlx5e_priv *priv, u64 *data,
-					 int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(per_prio)
 {
 	idx = mlx5e_grp_per_prio_traffic_fill_stats(priv, data, idx);
 	idx = mlx5e_grp_per_prio_pfc_fill_stats(priv, data, idx);
 	return idx;
 }
 
-static void mlx5e_grp_per_prio_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(per_prio)
 {
 	struct mlx5e_pport_stats *pstats = &priv->stats.pport;
 	struct mlx5_core_dev *mdev = priv->mdev;
@@ -1349,13 +1386,12 @@ static const struct counter_desc mlx5e_pme_error_desc[] = {
 #define NUM_PME_STATUS_STATS		ARRAY_SIZE(mlx5e_pme_status_desc)
 #define NUM_PME_ERR_STATS		ARRAY_SIZE(mlx5e_pme_error_desc)
 
-static int mlx5e_grp_pme_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(pme)
 {
 	return NUM_PME_STATUS_STATS + NUM_PME_ERR_STATS;
 }
 
-static int mlx5e_grp_pme_fill_strings(struct mlx5e_priv *priv, u8 *data,
-				      int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(pme)
 {
 	int i;
 
@@ -1368,8 +1404,7 @@ static int mlx5e_grp_pme_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_pme_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				    int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(pme)
 {
 	struct mlx5_pme_stats pme_stats;
 	int i;
@@ -1387,45 +1422,46 @@ static int mlx5e_grp_pme_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
-static int mlx5e_grp_ipsec_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(pme) { return; }
+
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(ipsec)
 {
 	return mlx5e_ipsec_get_count(priv);
 }
 
-static int mlx5e_grp_ipsec_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(ipsec)
 {
 	return idx + mlx5e_ipsec_get_strings(priv,
 					     data + idx * ETH_GSTRING_LEN);
 }
 
-static int mlx5e_grp_ipsec_fill_stats(struct mlx5e_priv *priv, u64 *data,
-				      int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(ipsec)
 {
 	return idx + mlx5e_ipsec_get_stats(priv, data + idx);
 }
 
-static void mlx5e_grp_ipsec_update_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(ipsec)
 {
 	mlx5e_ipsec_update_stats(priv);
 }
 
-static int mlx5e_grp_tls_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(tls)
 {
 	return mlx5e_tls_get_count(priv);
 }
 
-static int mlx5e_grp_tls_fill_strings(struct mlx5e_priv *priv, u8 *data,
-				      int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(tls)
 {
 	return idx + mlx5e_tls_get_strings(priv, data + idx * ETH_GSTRING_LEN);
 }
 
-static int mlx5e_grp_tls_fill_stats(struct mlx5e_priv *priv, u64 *data, int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(tls)
 {
 	return idx + mlx5e_tls_get_stats(priv, data + idx);
 }
 
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(tls) { return; }
+
 static const struct counter_desc rq_stats_desc[] = {
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, packets) },
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, bytes) },
@@ -1559,7 +1595,7 @@ static const struct counter_desc ch_stats_desc[] = {
 #define NUM_XSKSQ_STATS			ARRAY_SIZE(xsksq_stats_desc)
 #define NUM_CH_STATS			ARRAY_SIZE(ch_stats_desc)
 
-static int mlx5e_grp_channels_get_num_stats(struct mlx5e_priv *priv)
+static MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(channels)
 {
 	int max_nch = priv->max_nch;
 
@@ -1572,8 +1608,7 @@ static int mlx5e_grp_channels_get_num_stats(struct mlx5e_priv *priv)
 	       (NUM_XSKSQ_STATS * max_nch * priv->xsk.ever_used);
 }
 
-static int mlx5e_grp_channels_fill_strings(struct mlx5e_priv *priv, u8 *data,
-					   int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(channels)
 {
 	bool is_xsk = priv->xsk.ever_used;
 	int max_nch = priv->max_nch;
@@ -1615,8 +1650,7 @@ static int mlx5e_grp_channels_fill_strings(struct mlx5e_priv *priv, u8 *data,
 	return idx;
 }
 
-static int mlx5e_grp_channels_fill_stats(struct mlx5e_priv *priv, u64 *data,
-					 int idx)
+static MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(channels)
 {
 	bool is_xsk = priv->xsk.ever_used;
 	int max_nch = priv->max_nch;
@@ -1664,104 +1698,46 @@ static int mlx5e_grp_channels_fill_stats(struct mlx5e_priv *priv, u64 *data,
 	return idx;
 }
 
+static MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(channels) { return; }
+
+MLX5E_DEFINE_STATS_GRP(sw, 0);
+MLX5E_DEFINE_STATS_GRP(qcnt, MLX5E_NDO_UPDATE_STATS);
+MLX5E_DEFINE_STATS_GRP(vnic_env, 0);
+MLX5E_DEFINE_STATS_GRP(vport, MLX5E_NDO_UPDATE_STATS);
+MLX5E_DEFINE_STATS_GRP(802_3, MLX5E_NDO_UPDATE_STATS);
+MLX5E_DEFINE_STATS_GRP(2863, 0);
+MLX5E_DEFINE_STATS_GRP(2819, 0);
+MLX5E_DEFINE_STATS_GRP(phy, 0);
+MLX5E_DEFINE_STATS_GRP(pcie, 0);
+MLX5E_DEFINE_STATS_GRP(per_prio, 0);
+MLX5E_DEFINE_STATS_GRP(pme, 0);
+MLX5E_DEFINE_STATS_GRP(channels, 0);
+MLX5E_DEFINE_STATS_GRP(per_port_buff_congest, 0);
+MLX5E_DEFINE_STATS_GRP(eth_ext, 0);
+static MLX5E_DEFINE_STATS_GRP(ipsec, 0);
+static MLX5E_DEFINE_STATS_GRP(tls, 0);
+
 /* The stats groups order is opposite to the update_stats() order calls */
-const struct mlx5e_stats_grp mlx5e_stats_grps[] = {
-	{
-		.get_num_stats = mlx5e_grp_sw_get_num_stats,
-		.fill_strings = mlx5e_grp_sw_fill_strings,
-		.fill_stats = mlx5e_grp_sw_fill_stats,
-		.update_stats = mlx5e_grp_sw_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_q_get_num_stats,
-		.fill_strings = mlx5e_grp_q_fill_strings,
-		.fill_stats = mlx5e_grp_q_fill_stats,
-		.update_stats_mask = MLX5E_NDO_UPDATE_STATS,
-		.update_stats = mlx5e_grp_q_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_vnic_env_get_num_stats,
-		.fill_strings = mlx5e_grp_vnic_env_fill_strings,
-		.fill_stats = mlx5e_grp_vnic_env_fill_stats,
-		.update_stats = mlx5e_grp_vnic_env_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_vport_get_num_stats,
-		.fill_strings = mlx5e_grp_vport_fill_strings,
-		.fill_stats = mlx5e_grp_vport_fill_stats,
-		.update_stats_mask = MLX5E_NDO_UPDATE_STATS,
-		.update_stats = mlx5e_grp_vport_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_802_3_get_num_stats,
-		.fill_strings = mlx5e_grp_802_3_fill_strings,
-		.fill_stats = mlx5e_grp_802_3_fill_stats,
-		.update_stats_mask = MLX5E_NDO_UPDATE_STATS,
-		.update_stats = mlx5e_grp_802_3_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_2863_get_num_stats,
-		.fill_strings = mlx5e_grp_2863_fill_strings,
-		.fill_stats = mlx5e_grp_2863_fill_stats,
-		.update_stats = mlx5e_grp_2863_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_2819_get_num_stats,
-		.fill_strings = mlx5e_grp_2819_fill_strings,
-		.fill_stats = mlx5e_grp_2819_fill_stats,
-		.update_stats = mlx5e_grp_2819_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_phy_get_num_stats,
-		.fill_strings = mlx5e_grp_phy_fill_strings,
-		.fill_stats = mlx5e_grp_phy_fill_stats,
-		.update_stats = mlx5e_grp_phy_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_eth_ext_get_num_stats,
-		.fill_strings = mlx5e_grp_eth_ext_fill_strings,
-		.fill_stats = mlx5e_grp_eth_ext_fill_stats,
-		.update_stats = mlx5e_grp_eth_ext_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_pcie_get_num_stats,
-		.fill_strings = mlx5e_grp_pcie_fill_strings,
-		.fill_stats = mlx5e_grp_pcie_fill_stats,
-		.update_stats = mlx5e_grp_pcie_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_per_prio_get_num_stats,
-		.fill_strings = mlx5e_grp_per_prio_fill_strings,
-		.fill_stats = mlx5e_grp_per_prio_fill_stats,
-		.update_stats = mlx5e_grp_per_prio_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_pme_get_num_stats,
-		.fill_strings = mlx5e_grp_pme_fill_strings,
-		.fill_stats = mlx5e_grp_pme_fill_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_ipsec_get_num_stats,
-		.fill_strings = mlx5e_grp_ipsec_fill_strings,
-		.fill_stats = mlx5e_grp_ipsec_fill_stats,
-		.update_stats = mlx5e_grp_ipsec_update_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_tls_get_num_stats,
-		.fill_strings = mlx5e_grp_tls_fill_strings,
-		.fill_stats = mlx5e_grp_tls_fill_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_channels_get_num_stats,
-		.fill_strings = mlx5e_grp_channels_fill_strings,
-		.fill_stats = mlx5e_grp_channels_fill_stats,
-	},
-	{
-		.get_num_stats = mlx5e_grp_per_port_buffer_congest_get_num_stats,
-		.fill_strings = mlx5e_grp_per_port_buffer_congest_fill_strings,
-		.fill_stats = mlx5e_grp_per_port_buffer_congest_fill_stats,
-		.update_stats = mlx5e_grp_per_port_buffer_congest_update_stats,
-	},
+mlx5e_stats_grp_t mlx5e_nic_stats_grps[] = {
+	&MLX5E_STATS_GRP(sw),
+	&MLX5E_STATS_GRP(qcnt),
+	&MLX5E_STATS_GRP(vnic_env),
+	&MLX5E_STATS_GRP(vport),
+	&MLX5E_STATS_GRP(802_3),
+	&MLX5E_STATS_GRP(2863),
+	&MLX5E_STATS_GRP(2819),
+	&MLX5E_STATS_GRP(phy),
+	&MLX5E_STATS_GRP(eth_ext),
+	&MLX5E_STATS_GRP(pcie),
+	&MLX5E_STATS_GRP(per_prio),
+	&MLX5E_STATS_GRP(pme),
+	&MLX5E_STATS_GRP(ipsec),
+	&MLX5E_STATS_GRP(tls),
+	&MLX5E_STATS_GRP(channels),
+	&MLX5E_STATS_GRP(per_port_buff_congest),
 };
 
-const int mlx5e_num_stats_grps = ARRAY_SIZE(mlx5e_stats_grps);
+unsigned int mlx5e_nic_stats_grps_num(struct mlx5e_priv *priv)
+{
+	return ARRAY_SIZE(mlx5e_nic_stats_grps);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
index 869f3502f631..092b39ffa32a 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.h
@@ -29,6 +29,7 @@
  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
+
 #ifndef __MLX5_EN_STATS_H__
 #define __MLX5_EN_STATS_H__
 
@@ -55,6 +56,56 @@ struct counter_desc {
 	size_t		offset; /* Byte offset */
 };
 
+enum {
+	MLX5E_NDO_UPDATE_STATS = BIT(0x1),
+};
+
+struct mlx5e_priv;
+struct mlx5e_stats_grp {
+	u16 update_stats_mask;
+	int (*get_num_stats)(struct mlx5e_priv *priv);
+	int (*fill_strings)(struct mlx5e_priv *priv, u8 *data, int idx);
+	int (*fill_stats)(struct mlx5e_priv *priv, u64 *data, int idx);
+	void (*update_stats)(struct mlx5e_priv *priv);
+};
+
+typedef const struct mlx5e_stats_grp *const mlx5e_stats_grp_t;
+
+#define MLX5E_STATS_GRP_OP(grp, name) mlx5e_stats_grp_ ## grp ## _ ## name
+
+#define MLX5E_DECLARE_STATS_GRP_OP_NUM_STATS(grp) \
+	int MLX5E_STATS_GRP_OP(grp, num_stats)(struct mlx5e_priv *priv)
+
+#define MLX5E_DECLARE_STATS_GRP_OP_UPDATE_STATS(grp) \
+	void MLX5E_STATS_GRP_OP(grp, update_stats)(struct mlx5e_priv *priv)
+
+#define MLX5E_DECLARE_STATS_GRP_OP_FILL_STRS(grp) \
+	int MLX5E_STATS_GRP_OP(grp, fill_strings)(struct mlx5e_priv *priv, u8 *data, int idx)
+
+#define MLX5E_DECLARE_STATS_GRP_OP_FILL_STATS(grp) \
+	int MLX5E_STATS_GRP_OP(grp, fill_stats)(struct mlx5e_priv *priv, u64 *data, int idx)
+
+#define MLX5E_STATS_GRP(grp) mlx5e_stats_grp_ ## grp
+
+#define MLX5E_DECLARE_STATS_GRP(grp) \
+	const struct mlx5e_stats_grp MLX5E_STATS_GRP(grp)
+
+#define MLX5E_DEFINE_STATS_GRP(grp, mask) \
+MLX5E_DECLARE_STATS_GRP(grp) = { \
+	.get_num_stats = MLX5E_STATS_GRP_OP(grp, num_stats), \
+	.fill_stats    = MLX5E_STATS_GRP_OP(grp, fill_stats), \
+	.fill_strings  = MLX5E_STATS_GRP_OP(grp, fill_strings), \
+	.update_stats  = MLX5E_STATS_GRP_OP(grp, update_stats), \
+	.update_stats_mask = mask, \
+}
+
+unsigned int mlx5e_stats_total_num(struct mlx5e_priv *priv);
+void mlx5e_stats_update(struct mlx5e_priv *priv);
+void mlx5e_stats_fill(struct mlx5e_priv *priv, u64 *data, int idx);
+void mlx5e_stats_fill_strings(struct mlx5e_priv *priv, u8 *data);
+
+/* Concrete NIC Stats */
+
 struct mlx5e_sw_stats {
 	u64 rx_packets;
 	u64 rx_bytes;
@@ -322,22 +373,22 @@ struct mlx5e_stats {
 	struct mlx5e_pcie_stats pcie;
 };
 
-enum {
-	MLX5E_NDO_UPDATE_STATS = BIT(0x1),
-};
-
-struct mlx5e_priv;
-struct mlx5e_stats_grp {
-	u16 update_stats_mask;
-	int (*get_num_stats)(struct mlx5e_priv *priv);
-	int (*fill_strings)(struct mlx5e_priv *priv, u8 *data, int idx);
-	int (*fill_stats)(struct mlx5e_priv *priv, u64 *data, int idx);
-	void (*update_stats)(struct mlx5e_priv *priv);
-};
-
-extern const struct mlx5e_stats_grp mlx5e_stats_grps[];
-extern const int mlx5e_num_stats_grps;
+extern mlx5e_stats_grp_t mlx5e_nic_stats_grps[];
+unsigned int mlx5e_nic_stats_grps_num(struct mlx5e_priv *priv);
 
-void mlx5e_grp_802_3_update_stats(struct mlx5e_priv *priv);
+extern MLX5E_DECLARE_STATS_GRP(sw);
+extern MLX5E_DECLARE_STATS_GRP(qcnt);
+extern MLX5E_DECLARE_STATS_GRP(vnic_env);
+extern MLX5E_DECLARE_STATS_GRP(vport);
+extern MLX5E_DECLARE_STATS_GRP(802_3);
+extern MLX5E_DECLARE_STATS_GRP(2863);
+extern MLX5E_DECLARE_STATS_GRP(2819);
+extern MLX5E_DECLARE_STATS_GRP(phy);
+extern MLX5E_DECLARE_STATS_GRP(eth_ext);
+extern MLX5E_DECLARE_STATS_GRP(pcie);
+extern MLX5E_DECLARE_STATS_GRP(per_prio);
+extern MLX5E_DECLARE_STATS_GRP(pme);
+extern MLX5E_DECLARE_STATS_GRP(channels);
+extern MLX5E_DECLARE_STATS_GRP(per_port_buff_congest);
 
 #endif /* __MLX5_EN_STATS_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 7e32b9e3667c..74091f72c9a8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -51,6 +51,7 @@
 #include "en_rep.h"
 #include "en_tc.h"
 #include "eswitch.h"
+#include "eswitch_offloads_chains.h"
 #include "fs_core.h"
 #include "en/port.h"
 #include "en/tc_tun.h"
@@ -960,7 +961,8 @@ mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv,
 
 	mutex_lock(&priv->fs.tc.t_lock);
 	if (IS_ERR_OR_NULL(priv->fs.tc.t)) {
-		int tc_grp_size, tc_tbl_size;
+		struct mlx5_flow_table_attr ft_attr = {};
+		int tc_grp_size, tc_tbl_size, tc_num_grps;
 		u32 max_flow_counter;
 
 		max_flow_counter = (MLX5_CAP_GEN(dev, max_flow_counter_31_16) << 16) |
@@ -970,13 +972,15 @@ mlx5e_tc_add_nic_flow(struct mlx5e_priv *priv,
 
 		tc_tbl_size = min_t(int, tc_grp_size * MLX5E_TC_TABLE_NUM_GROUPS,
 				    BIT(MLX5_CAP_FLOWTABLE_NIC_RX(dev, log_max_ft_size)));
+		tc_num_grps = MLX5E_TC_TABLE_NUM_GROUPS;
 
+		ft_attr.prio = MLX5E_TC_PRIO;
+		ft_attr.max_fte = tc_tbl_size;
+		ft_attr.level = MLX5E_TC_FT_LEVEL;
+		ft_attr.autogroup.max_num_groups = tc_num_grps;
 		priv->fs.tc.t =
 			mlx5_create_auto_grouped_flow_table(priv->fs.ns,
-							    MLX5E_TC_PRIO,
-							    tc_tbl_size,
-							    MLX5E_TC_TABLE_NUM_GROUPS,
-							    MLX5E_TC_FT_LEVEL, 0);
+							    &ft_attr);
 		if (IS_ERR(priv->fs.tc.t)) {
 			mutex_unlock(&priv->fs.tc.t_lock);
 			NL_SET_ERR_MSG_MOD(extack,
@@ -1080,7 +1084,7 @@ mlx5e_tc_offload_to_slow_path(struct mlx5_eswitch *esw,
 	memcpy(slow_attr, flow->esw_attr, sizeof(*slow_attr));
 	slow_attr->action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 	slow_attr->split_count = 0;
-	slow_attr->dest_chain = FDB_TC_SLOW_PATH_CHAIN;
+	slow_attr->flags |= MLX5_ESW_ATTR_FLAG_SLOW_PATH;
 
 	rule = mlx5e_tc_offload_fdb_rules(esw, flow, spec, slow_attr);
 	if (!IS_ERR(rule))
@@ -1097,7 +1101,7 @@ mlx5e_tc_unoffload_from_slow_path(struct mlx5_eswitch *esw,
 	memcpy(slow_attr, flow->esw_attr, sizeof(*slow_attr));
 	slow_attr->action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 	slow_attr->split_count = 0;
-	slow_attr->dest_chain = FDB_TC_SLOW_PATH_CHAIN;
+	slow_attr->flags |= MLX5_ESW_ATTR_FLAG_SLOW_PATH;
 	mlx5e_tc_unoffload_fdb_rules(esw, flow, slow_attr);
 	flow_flag_clear(flow, SLOW);
 }
@@ -1157,19 +1161,18 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
 		      struct netlink_ext_ack *extack)
 {
 	struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
-	u32 max_chain = mlx5_eswitch_get_chain_range(esw);
 	struct mlx5_esw_flow_attr *attr = flow->esw_attr;
 	struct mlx5e_tc_flow_parse_attr *parse_attr = attr->parse_attr;
-	u16 max_prio = mlx5_eswitch_get_prio_range(esw);
 	struct net_device *out_dev, *encap_dev = NULL;
 	struct mlx5_fc *counter = NULL;
 	struct mlx5e_rep_priv *rpriv;
 	struct mlx5e_priv *out_priv;
 	bool encap_valid = true;
+	u32 max_prio, max_chain;
 	int err = 0;
 	int out_index;
 
-	if (!mlx5_eswitch_prios_supported(esw) && attr->prio != 1) {
+	if (!mlx5_esw_chains_prios_supported(esw) && attr->prio != 1) {
 		NL_SET_ERR_MSG(extack, "E-switch priorities unsupported, upgrade FW");
 		return -EOPNOTSUPP;
 	}
@@ -1179,11 +1182,13 @@ mlx5e_tc_add_fdb_flow(struct mlx5e_priv *priv,
 	 * FDB_FT_CHAIN which is outside tc range.
 	 * See mlx5e_rep_setup_ft_cb().
 	 */
+	max_chain = mlx5_esw_chains_get_chain_range(esw);
 	if (!mlx5e_is_ft_flow(flow) && attr->chain > max_chain) {
 		NL_SET_ERR_MSG(extack, "Requested chain is out of supported range");
 		return -EOPNOTSUPP;
 	}
 
+	max_prio = mlx5_esw_chains_get_prio_range(esw);
 	if (attr->prio > max_prio) {
 		NL_SET_ERR_MSG(extack, "Requested priority is out of supported range");
 		return -EOPNOTSUPP;
@@ -1805,6 +1810,40 @@ static void *get_match_headers_value(u32 flags,
 			     outer_headers);
 }
 
+static int mlx5e_flower_parse_meta(struct net_device *filter_dev,
+				   struct flow_cls_offload *f)
+{
+	struct flow_rule *rule = flow_cls_offload_flow_rule(f);
+	struct netlink_ext_ack *extack = f->common.extack;
+	struct net_device *ingress_dev;
+	struct flow_match_meta match;
+
+	if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_META))
+		return 0;
+
+	flow_rule_match_meta(rule, &match);
+	if (match.mask->ingress_ifindex != 0xFFFFFFFF) {
+		NL_SET_ERR_MSG_MOD(extack, "Unsupported ingress ifindex mask");
+		return -EINVAL;
+	}
+
+	ingress_dev = __dev_get_by_index(dev_net(filter_dev),
+					 match.key->ingress_ifindex);
+	if (!ingress_dev) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Can't find the ingress port to match on");
+		return -EINVAL;
+	}
+
+	if (ingress_dev != filter_dev) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Can't match on the ingress filter port");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
 static int __parse_cls_flower(struct mlx5e_priv *priv,
 			      struct mlx5_flow_spec *spec,
 			      struct flow_cls_offload *f,
@@ -1825,6 +1864,7 @@ static int __parse_cls_flower(struct mlx5e_priv *priv,
 	u16 addr_type = 0;
 	u8 ip_proto = 0;
 	u8 *match_level;
+	int err;
 
 	match_level = outer_match_level;
 
@@ -1868,6 +1908,10 @@ static int __parse_cls_flower(struct mlx5e_priv *priv,
 						    spec);
 	}
 
+	err = mlx5e_flower_parse_meta(filter_dev, f);
+	if (err)
+		return err;
+
 	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_BASIC)) {
 		struct flow_match_basic match;
 
@@ -2842,6 +2886,10 @@ static int parse_tc_nic_actions(struct mlx5e_priv *priv,
 
 	flow_action_for_each(i, act, flow_action) {
 		switch (act->id) {
+		case FLOW_ACTION_ACCEPT:
+			action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST |
+				  MLX5_FLOW_CONTEXT_ACTION_COUNT;
+			break;
 		case FLOW_ACTION_DROP:
 			action |= MLX5_FLOW_CONTEXT_ACTION_DROP;
 			if (MLX5_CAP_FLOWTABLE(priv->mdev,
@@ -3462,7 +3510,7 @@ static int parse_tc_fdb_actions(struct mlx5e_priv *priv,
 			break;
 		case FLOW_ACTION_GOTO: {
 			u32 dest_chain = act->chain_index;
-			u32 max_chain = mlx5_eswitch_get_chain_range(esw);
+			u32 max_chain = mlx5_esw_chains_get_chain_range(esw);
 
 			if (ft_flow) {
 				NL_SET_ERR_MSG_MOD(extack, "Goto action is not supported");
@@ -4203,7 +4251,10 @@ int mlx5e_tc_nic_init(struct mlx5e_priv *priv)
 		return err;
 
 	tc->netdevice_nb.notifier_call = mlx5e_tc_netdev_event;
-	if (register_netdevice_notifier(&tc->netdevice_nb)) {
+	err = register_netdevice_notifier_dev_net(priv->netdev,
+						  &tc->netdevice_nb,
+						  &tc->netdevice_nn);
+	if (err) {
 		tc->netdevice_nb.notifier_call = NULL;
 		mlx5_core_warn(priv->mdev, "Failed to register netdev notifier\n");
 	}
@@ -4225,7 +4276,9 @@ void mlx5e_tc_nic_cleanup(struct mlx5e_priv *priv)
 	struct mlx5e_tc_table *tc = &priv->fs.tc;
 
 	if (tc->netdevice_nb.notifier_call)
-		unregister_netdevice_notifier(&tc->netdevice_nb);
+		unregister_netdevice_notifier_dev_net(priv->netdev,
+						      &tc->netdevice_nb,
+						      &tc->netdevice_nn);
 
 	mutex_destroy(&tc->mod_hdr.lock);
 	mutex_destroy(&tc->hairpin_tbl_lock);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
index 580c71cb9dfa..cccea3a8eddd 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
@@ -156,7 +156,8 @@ static int mlx5_eq_comp_int(struct notifier_block *nb,
 			cq->comp(cq, eqe);
 			mlx5_cq_put(cq);
 		} else {
-			mlx5_core_warn(eq->dev, "Completion event for bogus CQ 0x%x\n", cqn);
+			dev_dbg_ratelimited(eq->dev->device,
+					    "Completion event for bogus CQ 0x%x\n", cqn);
 		}
 
 		++eq->cons_index;
@@ -563,6 +564,39 @@ static void gather_async_events_mask(struct mlx5_core_dev *dev, u64 mask[4])
 		gather_user_async_events(dev, mask);
 }
 
+static int
+setup_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq_async *eq,
+	       struct mlx5_eq_param *param, const char *name)
+{
+	int err;
+
+	eq->irq_nb.notifier_call = mlx5_eq_async_int;
+
+	err = create_async_eq(dev, &eq->core, param);
+	if (err) {
+		mlx5_core_warn(dev, "failed to create %s EQ %d\n", name, err);
+		return err;
+	}
+	err = mlx5_eq_enable(dev, &eq->core, &eq->irq_nb);
+	if (err) {
+		mlx5_core_warn(dev, "failed to enable %s EQ %d\n", name, err);
+		destroy_async_eq(dev, &eq->core);
+	}
+	return err;
+}
+
+static void cleanup_async_eq(struct mlx5_core_dev *dev,
+			     struct mlx5_eq_async *eq, const char *name)
+{
+	int err;
+
+	mlx5_eq_disable(dev, &eq->core, &eq->irq_nb);
+	err = destroy_async_eq(dev, &eq->core);
+	if (err)
+		mlx5_core_err(dev, "failed to destroy %s eq, err(%d)\n",
+			      name, err);
+}
+
 static int create_async_eqs(struct mlx5_core_dev *dev)
 {
 	struct mlx5_eq_table *table = dev->priv.eq_table;
@@ -572,77 +606,45 @@ static int create_async_eqs(struct mlx5_core_dev *dev)
 	MLX5_NB_INIT(&table->cq_err_nb, cq_err_event_notifier, CQ_ERROR);
 	mlx5_eq_notifier_register(dev, &table->cq_err_nb);
 
-	table->cmd_eq.irq_nb.notifier_call = mlx5_eq_async_int;
 	param = (struct mlx5_eq_param) {
 		.irq_index = 0,
 		.nent = MLX5_NUM_CMD_EQE,
+		.mask[0] = 1ull << MLX5_EVENT_TYPE_CMD,
 	};
-
-	param.mask[0] = 1ull << MLX5_EVENT_TYPE_CMD;
-	err = create_async_eq(dev, &table->cmd_eq.core, &param);
-	if (err) {
-		mlx5_core_warn(dev, "failed to create cmd EQ %d\n", err);
-		goto err0;
-	}
-	err = mlx5_eq_enable(dev, &table->cmd_eq.core, &table->cmd_eq.irq_nb);
-	if (err) {
-		mlx5_core_warn(dev, "failed to enable cmd EQ %d\n", err);
+	err = setup_async_eq(dev, &table->cmd_eq, &param, "cmd");
+	if (err)
 		goto err1;
-	}
+
 	mlx5_cmd_use_events(dev);
 
-	table->async_eq.irq_nb.notifier_call = mlx5_eq_async_int;
 	param = (struct mlx5_eq_param) {
 		.irq_index = 0,
 		.nent = MLX5_NUM_ASYNC_EQE,
 	};
 
 	gather_async_events_mask(dev, param.mask);
-	err = create_async_eq(dev, &table->async_eq.core, &param);
-	if (err) {
-		mlx5_core_warn(dev, "failed to create async EQ %d\n", err);
+	err = setup_async_eq(dev, &table->async_eq, &param, "async");
+	if (err)
 		goto err2;
-	}
-	err = mlx5_eq_enable(dev, &table->async_eq.core,
-			     &table->async_eq.irq_nb);
-	if (err) {
-		mlx5_core_warn(dev, "failed to enable async EQ %d\n", err);
-		goto err3;
-	}
 
-	table->pages_eq.irq_nb.notifier_call = mlx5_eq_async_int;
 	param = (struct mlx5_eq_param) {
 		.irq_index = 0,
 		.nent = /* TODO: sriov max_vf + */ 1,
+		.mask[0] = 1ull << MLX5_EVENT_TYPE_PAGE_REQUEST,
 	};
 
-	param.mask[0] = 1ull << MLX5_EVENT_TYPE_PAGE_REQUEST;
-	err = create_async_eq(dev, &table->pages_eq.core, &param);
-	if (err) {
-		mlx5_core_warn(dev, "failed to create pages EQ %d\n", err);
-		goto err4;
-	}
-	err = mlx5_eq_enable(dev, &table->pages_eq.core,
-			     &table->pages_eq.irq_nb);
-	if (err) {
-		mlx5_core_warn(dev, "failed to enable pages EQ %d\n", err);
-		goto err5;
-	}
+	err = setup_async_eq(dev, &table->pages_eq, &param, "pages");
+	if (err)
+		goto err3;
 
-	return err;
+	return 0;
 
-err5:
-	destroy_async_eq(dev, &table->pages_eq.core);
-err4:
-	mlx5_eq_disable(dev, &table->async_eq.core, &table->async_eq.irq_nb);
 err3:
-	destroy_async_eq(dev, &table->async_eq.core);
+	cleanup_async_eq(dev, &table->async_eq, "async");
 err2:
 	mlx5_cmd_use_polling(dev);
-	mlx5_eq_disable(dev, &table->cmd_eq.core, &table->cmd_eq.irq_nb);
+	cleanup_async_eq(dev, &table->cmd_eq, "cmd");
 err1:
-	destroy_async_eq(dev, &table->cmd_eq.core);
-err0:
 	mlx5_eq_notifier_unregister(dev, &table->cq_err_nb);
 	return err;
 }
@@ -650,28 +652,11 @@ err0:
 static void destroy_async_eqs(struct mlx5_core_dev *dev)
 {
 	struct mlx5_eq_table *table = dev->priv.eq_table;
-	int err;
-
-	mlx5_eq_disable(dev, &table->pages_eq.core, &table->pages_eq.irq_nb);
-	err = destroy_async_eq(dev, &table->pages_eq.core);
-	if (err)
-		mlx5_core_err(dev, "failed to destroy pages eq, err(%d)\n",
-			      err);
-
-	mlx5_eq_disable(dev, &table->async_eq.core, &table->async_eq.irq_nb);
-	err = destroy_async_eq(dev, &table->async_eq.core);
-	if (err)
-		mlx5_core_err(dev, "failed to destroy async eq, err(%d)\n",
-			      err);
 
+	cleanup_async_eq(dev, &table->pages_eq, "pages");
+	cleanup_async_eq(dev, &table->async_eq, "async");
 	mlx5_cmd_use_polling(dev);
-
-	mlx5_eq_disable(dev, &table->cmd_eq.core, &table->cmd_eq.irq_nb);
-	err = destroy_async_eq(dev, &table->cmd_eq.core);
-	if (err)
-		mlx5_core_err(dev, "failed to destroy command eq, err(%d)\n",
-			      err);
-
+	cleanup_async_eq(dev, &table->cmd_eq, "cmd");
 	mlx5_eq_notifier_unregister(dev, &table->cq_err_nb);
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
index 3df3604e8929..5acf60b1bbfe 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c
@@ -277,6 +277,7 @@ enum {
 
 static int esw_create_legacy_vepa_table(struct mlx5_eswitch *esw)
 {
+	struct mlx5_flow_table_attr ft_attr = {};
 	struct mlx5_core_dev *dev = esw->dev;
 	struct mlx5_flow_namespace *root_ns;
 	struct mlx5_flow_table *fdb;
@@ -289,8 +290,10 @@ static int esw_create_legacy_vepa_table(struct mlx5_eswitch *esw)
 	}
 
 	/* num FTE 2, num FG 2 */
-	fdb = mlx5_create_auto_grouped_flow_table(root_ns, LEGACY_VEPA_PRIO,
-						  2, 2, 0, 0);
+	ft_attr.prio = LEGACY_VEPA_PRIO;
+	ft_attr.max_fte = 2;
+	ft_attr.autogroup.max_num_groups = 2;
+	fdb = mlx5_create_auto_grouped_flow_table(root_ns, &ft_attr);
 	if (IS_ERR(fdb)) {
 		err = PTR_ERR(fdb);
 		esw_warn(dev, "Failed to create VEPA FDB err %d\n", err);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
index ffcff3ba3701..4472710ccc9c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h
@@ -157,7 +157,7 @@ enum offloads_fdb_flags {
 	ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED = BIT(0),
 };
 
-extern const unsigned int ESW_POOLS[4];
+struct mlx5_esw_chains_priv;
 
 struct mlx5_eswitch_fdb {
 	union {
@@ -182,14 +182,7 @@ struct mlx5_eswitch_fdb {
 			struct mlx5_flow_handle *miss_rule_multi;
 			int vlan_push_pop_refcount;
 
-			struct {
-				struct mlx5_flow_table *fdb;
-				u32 num_rules;
-			} fdb_prio[FDB_NUM_CHAINS][FDB_TC_MAX_PRIO + 1][FDB_TC_LEVELS_PER_PRIO];
-			/* Protects fdb_prio table */
-			struct mutex fdb_prio_lock;
-
-			int fdb_left[ARRAY_SIZE(ESW_POOLS)];
+			struct mlx5_esw_chains_priv *esw_chains_priv;
 		} offloads;
 	};
 	u32 flags;
@@ -355,15 +348,6 @@ mlx5_eswitch_del_fwd_rule(struct mlx5_eswitch *esw,
 			  struct mlx5_flow_handle *rule,
 			  struct mlx5_esw_flow_attr *attr);
 
-bool
-mlx5_eswitch_prios_supported(struct mlx5_eswitch *esw);
-
-u16
-mlx5_eswitch_get_prio_range(struct mlx5_eswitch *esw);
-
-u32
-mlx5_eswitch_get_chain_range(struct mlx5_eswitch *esw);
-
 struct mlx5_flow_handle *
 mlx5_eswitch_create_vport_rx_rule(struct mlx5_eswitch *esw, u16 vport,
 				  struct mlx5_flow_destination *dest);
@@ -388,6 +372,11 @@ enum {
 	MLX5_ESW_DEST_ENCAP_VALID   = BIT(1),
 };
 
+enum {
+	MLX5_ESW_ATTR_FLAG_VLAN_HANDLED  = BIT(0),
+	MLX5_ESW_ATTR_FLAG_SLOW_PATH     = BIT(1),
+};
+
 struct mlx5_esw_flow_attr {
 	struct mlx5_eswitch_rep *in_rep;
 	struct mlx5_core_dev	*in_mdev;
@@ -401,7 +390,6 @@ struct mlx5_esw_flow_attr {
 	u16	vlan_vid[MLX5_FS_VLAN_DEPTH];
 	u8	vlan_prio[MLX5_FS_VLAN_DEPTH];
 	u8	total_vlan;
-	bool	vlan_handled;
 	struct {
 		u32 flags;
 		struct mlx5_eswitch_rep *rep;
@@ -416,6 +404,7 @@ struct mlx5_esw_flow_attr {
 	u32	chain;
 	u16	prio;
 	u32	dest_chain;
+	u32	flags;
 	struct mlx5e_tc_flow_parse_attr *parse_attr;
 };
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 3e6412783078..979f13bdc203 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -37,6 +37,7 @@
 #include <linux/mlx5/fs.h>
 #include "mlx5_core.h"
 #include "eswitch.h"
+#include "eswitch_offloads_chains.h"
 #include "rdma.h"
 #include "en.h"
 #include "fs_core.h"
@@ -47,10 +48,6 @@
  * one for multicast.
  */
 #define MLX5_ESW_MISS_FLOWS (2)
-
-#define fdb_prio_table(esw, chain, prio, level) \
-	(esw)->fdb_table.offloads.fdb_prio[(chain)][(prio)][(level)]
-
 #define UPLINK_REP_INDEX 0
 
 static struct mlx5_eswitch_rep *mlx5_eswitch_get_rep(struct mlx5_eswitch *esw,
@@ -62,32 +59,6 @@ static struct mlx5_eswitch_rep *mlx5_eswitch_get_rep(struct mlx5_eswitch *esw,
 	return &esw->offloads.vport_reps[idx];
 }
 
-static struct mlx5_flow_table *
-esw_get_prio_table(struct mlx5_eswitch *esw, u32 chain, u16 prio, int level);
-static void
-esw_put_prio_table(struct mlx5_eswitch *esw, u32 chain, u16 prio, int level);
-
-bool mlx5_eswitch_prios_supported(struct mlx5_eswitch *esw)
-{
-	return (!!(esw->fdb_table.flags & ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED));
-}
-
-u32 mlx5_eswitch_get_chain_range(struct mlx5_eswitch *esw)
-{
-	if (esw->fdb_table.flags & ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED)
-		return FDB_TC_MAX_CHAIN;
-
-	return 0;
-}
-
-u16 mlx5_eswitch_get_prio_range(struct mlx5_eswitch *esw)
-{
-	if (esw->fdb_table.flags & ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED)
-		return FDB_TC_MAX_PRIO;
-
-	return 1;
-}
-
 static bool
 esw_check_ingress_prio_tag_enabled(const struct mlx5_eswitch *esw,
 				   const struct mlx5_vport *vport)
@@ -175,10 +146,17 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
 	}
 
 	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
-		if (attr->dest_chain) {
-			struct mlx5_flow_table *ft;
+		struct mlx5_flow_table *ft;
 
-			ft = esw_get_prio_table(esw, attr->dest_chain, 1, 0);
+		if (attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) {
+			flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
+			dest[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+			dest[i].ft = mlx5_esw_chains_get_tc_end_ft(esw);
+			i++;
+		} else if (attr->dest_chain) {
+			flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
+			ft = mlx5_esw_chains_get_table(esw, attr->dest_chain,
+						       1, 0);
 			if (IS_ERR(ft)) {
 				rule = ERR_CAST(ft);
 				goto err_create_goto_table;
@@ -223,7 +201,8 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
 	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_MOD_HDR)
 		flow_act.modify_hdr = attr->modify_hdr;
 
-	fdb = esw_get_prio_table(esw, attr->chain, attr->prio, !!split);
+	fdb = mlx5_esw_chains_get_table(esw, attr->chain, attr->prio,
+					!!split);
 	if (IS_ERR(fdb)) {
 		rule = ERR_CAST(fdb);
 		goto err_esw_get;
@@ -242,10 +221,10 @@ mlx5_eswitch_add_offloaded_rule(struct mlx5_eswitch *esw,
 	return rule;
 
 err_add_rule:
-	esw_put_prio_table(esw, attr->chain, attr->prio, !!split);
+	mlx5_esw_chains_put_table(esw, attr->chain, attr->prio, !!split);
 err_esw_get:
-	if (attr->dest_chain)
-		esw_put_prio_table(esw, attr->dest_chain, 1, 0);
+	if (!(attr->flags & MLX5_ESW_ATTR_FLAG_SLOW_PATH) && attr->dest_chain)
+		mlx5_esw_chains_put_table(esw, attr->dest_chain, 1, 0);
 err_create_goto_table:
 	return rule;
 }
@@ -262,13 +241,13 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
 	struct mlx5_flow_handle *rule;
 	int i;
 
-	fast_fdb = esw_get_prio_table(esw, attr->chain, attr->prio, 0);
+	fast_fdb = mlx5_esw_chains_get_table(esw, attr->chain, attr->prio, 0);
 	if (IS_ERR(fast_fdb)) {
 		rule = ERR_CAST(fast_fdb);
 		goto err_get_fast;
 	}
 
-	fwd_fdb = esw_get_prio_table(esw, attr->chain, attr->prio, 1);
+	fwd_fdb = mlx5_esw_chains_get_table(esw, attr->chain, attr->prio, 1);
 	if (IS_ERR(fwd_fdb)) {
 		rule = ERR_CAST(fwd_fdb);
 		goto err_get_fwd;
@@ -296,6 +275,7 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
 	if (attr->outer_match_level != MLX5_MATCH_NONE)
 		spec->match_criteria_enable |= MLX5_MATCH_OUTER_HEADERS;
 
+	flow_act.flags |= FLOW_ACT_IGNORE_FLOW_LEVEL;
 	rule = mlx5_add_flow_rules(fast_fdb, spec, &flow_act, dest, i);
 
 	if (IS_ERR(rule))
@@ -305,9 +285,9 @@ mlx5_eswitch_add_fwd_rule(struct mlx5_eswitch *esw,
 
 	return rule;
 add_err:
-	esw_put_prio_table(esw, attr->chain, attr->prio, 1);
+	mlx5_esw_chains_put_table(esw, attr->chain, attr->prio, 1);
 err_get_fwd:
-	esw_put_prio_table(esw, attr->chain, attr->prio, 0);
+	mlx5_esw_chains_put_table(esw, attr->chain, attr->prio, 0);
 err_get_fast:
 	return rule;
 }
@@ -332,12 +312,13 @@ __mlx5_eswitch_del_rule(struct mlx5_eswitch *esw,
 	atomic64_dec(&esw->offloads.num_flows);
 
 	if (fwd_rule)  {
-		esw_put_prio_table(esw, attr->chain, attr->prio, 1);
-		esw_put_prio_table(esw, attr->chain, attr->prio, 0);
+		mlx5_esw_chains_put_table(esw, attr->chain, attr->prio, 1);
+		mlx5_esw_chains_put_table(esw, attr->chain, attr->prio, 0);
 	} else {
-		esw_put_prio_table(esw, attr->chain, attr->prio, !!split);
+		mlx5_esw_chains_put_table(esw, attr->chain, attr->prio,
+					  !!split);
 		if (attr->dest_chain)
-			esw_put_prio_table(esw, attr->dest_chain, 1, 0);
+			mlx5_esw_chains_put_table(esw, attr->dest_chain, 1, 0);
 	}
 }
 
@@ -451,7 +432,7 @@ int mlx5_eswitch_add_vlan_action(struct mlx5_eswitch *esw,
 	if (err)
 		goto unlock;
 
-	attr->vlan_handled = false;
+	attr->flags &= ~MLX5_ESW_ATTR_FLAG_VLAN_HANDLED;
 
 	vport = esw_vlan_action_get_vport(attr, push, pop);
 
@@ -459,7 +440,7 @@ int mlx5_eswitch_add_vlan_action(struct mlx5_eswitch *esw,
 		/* tracks VF --> wire rules without vlan push action */
 		if (attr->dests[0].rep->vport == MLX5_VPORT_UPLINK) {
 			vport->vlan_refcount++;
-			attr->vlan_handled = true;
+			attr->flags |= MLX5_ESW_ATTR_FLAG_VLAN_HANDLED;
 		}
 
 		goto unlock;
@@ -490,7 +471,7 @@ skip_set_push:
 	}
 out:
 	if (!err)
-		attr->vlan_handled = true;
+		attr->flags |= MLX5_ESW_ATTR_FLAG_VLAN_HANDLED;
 unlock:
 	mutex_unlock(&esw->state_lock);
 	return err;
@@ -508,7 +489,7 @@ int mlx5_eswitch_del_vlan_action(struct mlx5_eswitch *esw,
 	if (mlx5_eswitch_vlan_actions_supported(esw->dev, 1))
 		return 0;
 
-	if (!attr->vlan_handled)
+	if (!(attr->flags & MLX5_ESW_ATTR_FLAG_VLAN_HANDLED))
 		return 0;
 
 	push = !!(attr->action & MLX5_FLOW_CONTEXT_ACTION_VLAN_PUSH);
@@ -582,8 +563,8 @@ mlx5_eswitch_add_send_to_vport_rule(struct mlx5_eswitch *esw, u16 vport,
 	dest.vport.num = vport;
 	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 
-	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
-					&flow_act, &dest, 1);
+	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
+					spec, &flow_act, &dest, 1);
 	if (IS_ERR(flow_rule))
 		esw_warn(esw->dev, "FDB: Failed to add send to vport rule err %ld\n", PTR_ERR(flow_rule));
 out:
@@ -824,8 +805,8 @@ static int esw_add_fdb_miss_rule(struct mlx5_eswitch *esw)
 	dest.vport.num = esw->manager_vport;
 	flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
 
-	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
-					&flow_act, &dest, 1);
+	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
+					spec, &flow_act, &dest, 1);
 	if (IS_ERR(flow_rule)) {
 		err = PTR_ERR(flow_rule);
 		esw_warn(esw->dev,  "FDB: Failed to add unicast miss flow rule err %d\n", err);
@@ -839,8 +820,8 @@ static int esw_add_fdb_miss_rule(struct mlx5_eswitch *esw)
 	dmac_v = MLX5_ADDR_OF(fte_match_param, headers_v,
 			      outer_headers.dmac_47_16);
 	dmac_v[0] = 0x01;
-	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb, spec,
-					&flow_act, &dest, 1);
+	flow_rule = mlx5_add_flow_rules(esw->fdb_table.offloads.slow_fdb,
+					spec, &flow_act, &dest, 1);
 	if (IS_ERR(flow_rule)) {
 		err = PTR_ERR(flow_rule);
 		esw_warn(esw->dev, "FDB: Failed to add multicast miss flow rule err %d\n", err);
@@ -855,174 +836,6 @@ out:
 	return err;
 }
 
-#define ESW_OFFLOADS_NUM_GROUPS  4
-
-/* Firmware currently has 4 pool of 4 sizes that it supports (ESW_POOLS),
- * and a virtual memory region of 16M (ESW_SIZE), this region is duplicated
- * for each flow table pool. We can allocate up to 16M of each pool,
- * and we keep track of how much we used via put/get_sz_to_pool.
- * Firmware doesn't report any of this for now.
- * ESW_POOL is expected to be sorted from large to small
- */
-#define ESW_SIZE (16 * 1024 * 1024)
-const unsigned int ESW_POOLS[4] = { 4 * 1024 * 1024, 1 * 1024 * 1024,
-				    64 * 1024, 128 };
-
-static int
-get_sz_from_pool(struct mlx5_eswitch *esw)
-{
-	int sz = 0, i;
-
-	for (i = 0; i < ARRAY_SIZE(ESW_POOLS); i++) {
-		if (esw->fdb_table.offloads.fdb_left[i]) {
-			--esw->fdb_table.offloads.fdb_left[i];
-			sz = ESW_POOLS[i];
-			break;
-		}
-	}
-
-	return sz;
-}
-
-static void
-put_sz_to_pool(struct mlx5_eswitch *esw, int sz)
-{
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(ESW_POOLS); i++) {
-		if (sz >= ESW_POOLS[i]) {
-			++esw->fdb_table.offloads.fdb_left[i];
-			break;
-		}
-	}
-}
-
-static struct mlx5_flow_table *
-create_next_size_table(struct mlx5_eswitch *esw,
-		       struct mlx5_flow_namespace *ns,
-		       u16 table_prio,
-		       int level,
-		       u32 flags)
-{
-	struct mlx5_flow_table *fdb;
-	int sz;
-
-	sz = get_sz_from_pool(esw);
-	if (!sz)
-		return ERR_PTR(-ENOSPC);
-
-	fdb = mlx5_create_auto_grouped_flow_table(ns,
-						  table_prio,
-						  sz,
-						  ESW_OFFLOADS_NUM_GROUPS,
-						  level,
-						  flags);
-	if (IS_ERR(fdb)) {
-		esw_warn(esw->dev, "Failed to create FDB Table err %d (table prio: %d, level: %d, size: %d)\n",
-			 (int)PTR_ERR(fdb), table_prio, level, sz);
-		put_sz_to_pool(esw, sz);
-	}
-
-	return fdb;
-}
-
-static struct mlx5_flow_table *
-esw_get_prio_table(struct mlx5_eswitch *esw, u32 chain, u16 prio, int level)
-{
-	struct mlx5_core_dev *dev = esw->dev;
-	struct mlx5_flow_table *fdb = NULL;
-	struct mlx5_flow_namespace *ns;
-	int table_prio, l = 0;
-	u32 flags = 0;
-
-	if (chain == FDB_TC_SLOW_PATH_CHAIN)
-		return esw->fdb_table.offloads.slow_fdb;
-
-	mutex_lock(&esw->fdb_table.offloads.fdb_prio_lock);
-
-	fdb = fdb_prio_table(esw, chain, prio, level).fdb;
-	if (fdb) {
-		/* take ref on earlier levels as well */
-		while (level >= 0)
-			fdb_prio_table(esw, chain, prio, level--).num_rules++;
-		mutex_unlock(&esw->fdb_table.offloads.fdb_prio_lock);
-		return fdb;
-	}
-
-	ns = mlx5_get_fdb_sub_ns(dev, chain);
-	if (!ns) {
-		esw_warn(dev, "Failed to get FDB sub namespace\n");
-		mutex_unlock(&esw->fdb_table.offloads.fdb_prio_lock);
-		return ERR_PTR(-EOPNOTSUPP);
-	}
-
-	if (esw->offloads.encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE)
-		flags |= (MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT |
-			  MLX5_FLOW_TABLE_TUNNEL_EN_DECAP);
-
-	table_prio = prio - 1;
-
-	/* create earlier levels for correct fs_core lookup when
-	 * connecting tables
-	 */
-	for (l = 0; l <= level; l++) {
-		if (fdb_prio_table(esw, chain, prio, l).fdb) {
-			fdb_prio_table(esw, chain, prio, l).num_rules++;
-			continue;
-		}
-
-		fdb = create_next_size_table(esw, ns, table_prio, l, flags);
-		if (IS_ERR(fdb)) {
-			l--;
-			goto err_create_fdb;
-		}
-
-		fdb_prio_table(esw, chain, prio, l).fdb = fdb;
-		fdb_prio_table(esw, chain, prio, l).num_rules = 1;
-	}
-
-	mutex_unlock(&esw->fdb_table.offloads.fdb_prio_lock);
-	return fdb;
-
-err_create_fdb:
-	mutex_unlock(&esw->fdb_table.offloads.fdb_prio_lock);
-	if (l >= 0)
-		esw_put_prio_table(esw, chain, prio, l);
-
-	return fdb;
-}
-
-static void
-esw_put_prio_table(struct mlx5_eswitch *esw, u32 chain, u16 prio, int level)
-{
-	int l;
-
-	if (chain == FDB_TC_SLOW_PATH_CHAIN)
-		return;
-
-	mutex_lock(&esw->fdb_table.offloads.fdb_prio_lock);
-
-	for (l = level; l >= 0; l--) {
-		if (--(fdb_prio_table(esw, chain, prio, l).num_rules) > 0)
-			continue;
-
-		put_sz_to_pool(esw, fdb_prio_table(esw, chain, prio, l).fdb->max_fte);
-		mlx5_destroy_flow_table(fdb_prio_table(esw, chain, prio, l).fdb);
-		fdb_prio_table(esw, chain, prio, l).fdb = NULL;
-	}
-
-	mutex_unlock(&esw->fdb_table.offloads.fdb_prio_lock);
-}
-
-static void esw_destroy_offloads_fast_fdb_tables(struct mlx5_eswitch *esw)
-{
-	/* If lazy creation isn't supported, deref the fast path tables */
-	if (!(esw->fdb_table.flags & ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED)) {
-		esw_put_prio_table(esw, 0, 1, 1);
-		esw_put_prio_table(esw, 0, 1, 0);
-	}
-}
-
 #define MAX_PF_SQ 256
 #define MAX_SQ_NVPORTS 32
 
@@ -1055,16 +868,16 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports)
 	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
 	struct mlx5_flow_table_attr ft_attr = {};
 	struct mlx5_core_dev *dev = esw->dev;
-	u32 *flow_group_in, max_flow_counter;
 	struct mlx5_flow_namespace *root_ns;
 	struct mlx5_flow_table *fdb = NULL;
-	int table_size, ix, err = 0, i;
+	u32 flags = 0, *flow_group_in;
+	int table_size, ix, err = 0;
 	struct mlx5_flow_group *g;
-	u32 flags = 0, fdb_max;
 	void *match_criteria;
 	u8 *dmac;
 
 	esw_debug(esw->dev, "Create offloads FDB Tables\n");
+
 	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
 	if (!flow_group_in)
 		return -ENOMEM;
@@ -1083,19 +896,6 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports)
 		goto ns_err;
 	}
 
-	max_flow_counter = (MLX5_CAP_GEN(dev, max_flow_counter_31_16) << 16) |
-			    MLX5_CAP_GEN(dev, max_flow_counter_15_0);
-	fdb_max = 1 << MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size);
-
-	esw_debug(dev, "Create offloads FDB table, min (max esw size(2^%d), max counters(%d), groups(%d), max flow table size(%d))\n",
-		  MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size),
-		  max_flow_counter, ESW_OFFLOADS_NUM_GROUPS,
-		  fdb_max);
-
-	for (i = 0; i < ARRAY_SIZE(ESW_POOLS); i++)
-		esw->fdb_table.offloads.fdb_left[i] =
-			ESW_POOLS[i] <= fdb_max ? ESW_SIZE / ESW_POOLS[i] : 0;
-
 	table_size = nvports * MAX_SQ_NVPORTS + MAX_PF_SQ +
 		MLX5_ESW_MISS_FLOWS + esw->total_vports;
 
@@ -1118,16 +918,10 @@ static int esw_create_offloads_fdb_tables(struct mlx5_eswitch *esw, int nvports)
 	}
 	esw->fdb_table.offloads.slow_fdb = fdb;
 
-	/* If lazy creation isn't supported, open the fast path tables now */
-	if (!MLX5_CAP_ESW_FLOWTABLE(esw->dev, multi_fdb_encap) &&
-	    esw->offloads.encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE) {
-		esw->fdb_table.flags &= ~ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED;
-		esw_warn(dev, "Lazy creation of flow tables isn't supported, ignoring priorities\n");
-		esw_get_prio_table(esw, 0, 1, 0);
-		esw_get_prio_table(esw, 0, 1, 1);
-	} else {
-		esw_debug(dev, "Lazy creation of flow tables supported, deferring table opening\n");
-		esw->fdb_table.flags |= ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED;
+	err = mlx5_esw_chains_create(esw);
+	if (err) {
+		esw_warn(dev, "Failed to create fdb chains err(%d)\n", err);
+		goto fdb_chains_err;
 	}
 
 	/* create send-to-vport group */
@@ -1218,7 +1012,8 @@ miss_err:
 peer_miss_err:
 	mlx5_destroy_flow_group(esw->fdb_table.offloads.send_to_vport_grp);
 send_vport_err:
-	esw_destroy_offloads_fast_fdb_tables(esw);
+	mlx5_esw_chains_destroy(esw);
+fdb_chains_err:
 	mlx5_destroy_flow_table(esw->fdb_table.offloads.slow_fdb);
 slow_fdb_err:
 	/* Holds true only as long as DMFS is the default */
@@ -1240,8 +1035,8 @@ static void esw_destroy_offloads_fdb_tables(struct mlx5_eswitch *esw)
 	mlx5_destroy_flow_group(esw->fdb_table.offloads.peer_miss_grp);
 	mlx5_destroy_flow_group(esw->fdb_table.offloads.miss_grp);
 
+	mlx5_esw_chains_destroy(esw);
 	mlx5_destroy_flow_table(esw->fdb_table.offloads.slow_fdb);
-	esw_destroy_offloads_fast_fdb_tables(esw);
 	/* Holds true only as long as DMFS is the default */
 	mlx5_flow_namespace_set_mode(esw->fdb_table.offloads.ns,
 				     MLX5_FLOW_STEERING_MODE_DMFS);
@@ -2111,7 +1906,6 @@ static int esw_offloads_steering_init(struct mlx5_eswitch *esw)
 		total_vports = num_vfs + MLX5_SPECIAL_VPORTS(esw->dev);
 
 	memset(&esw->fdb_table.offloads, 0, sizeof(struct offloads_fdb));
-	mutex_init(&esw->fdb_table.offloads.fdb_prio_lock);
 
 	err = esw_create_uplink_offloads_acl_tables(esw);
 	if (err)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c
new file mode 100644
index 000000000000..c5a446e295aa
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.c
@@ -0,0 +1,758 @@
+// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
+// Copyright (c) 2020 Mellanox Technologies.
+
+#include <linux/mlx5/driver.h>
+#include <linux/mlx5/mlx5_ifc.h>
+#include <linux/mlx5/fs.h>
+
+#include "eswitch_offloads_chains.h"
+#include "mlx5_core.h"
+#include "fs_core.h"
+#include "eswitch.h"
+#include "en.h"
+
+#define esw_chains_priv(esw) ((esw)->fdb_table.offloads.esw_chains_priv)
+#define esw_chains_lock(esw) (esw_chains_priv(esw)->lock)
+#define esw_chains_ht(esw) (esw_chains_priv(esw)->chains_ht)
+#define esw_prios_ht(esw) (esw_chains_priv(esw)->prios_ht)
+#define fdb_pool_left(esw) (esw_chains_priv(esw)->fdb_left)
+#define tc_slow_fdb(esw) ((esw)->fdb_table.offloads.slow_fdb)
+#define tc_end_fdb(esw) (esw_chains_priv(esw)->tc_end_fdb)
+#define fdb_ignore_flow_level_supported(esw) \
+	(MLX5_CAP_ESW_FLOWTABLE_FDB((esw)->dev, ignore_flow_level))
+
+#define ESW_OFFLOADS_NUM_GROUPS  4
+
+/* Firmware currently has 4 pool of 4 sizes that it supports (ESW_POOLS),
+ * and a virtual memory region of 16M (ESW_SIZE), this region is duplicated
+ * for each flow table pool. We can allocate up to 16M of each pool,
+ * and we keep track of how much we used via get_next_avail_sz_from_pool.
+ * Firmware doesn't report any of this for now.
+ * ESW_POOL is expected to be sorted from large to small and match firmware
+ * pools.
+ */
+#define ESW_SIZE (16 * 1024 * 1024)
+static const unsigned int ESW_POOLS[] = { 4 * 1024 * 1024,
+					  1 * 1024 * 1024,
+					  64 * 1024,
+					  4 * 1024, };
+
+struct mlx5_esw_chains_priv {
+	struct rhashtable chains_ht;
+	struct rhashtable prios_ht;
+	/* Protects above chains_ht and prios_ht */
+	struct mutex lock;
+
+	struct mlx5_flow_table *tc_end_fdb;
+
+	int fdb_left[ARRAY_SIZE(ESW_POOLS)];
+};
+
+struct fdb_chain {
+	struct rhash_head node;
+
+	u32 chain;
+
+	int ref;
+
+	struct mlx5_eswitch *esw;
+	struct list_head prios_list;
+};
+
+struct fdb_prio_key {
+	u32 chain;
+	u32 prio;
+	u32 level;
+};
+
+struct fdb_prio {
+	struct rhash_head node;
+	struct list_head list;
+
+	struct fdb_prio_key key;
+
+	int ref;
+
+	struct fdb_chain *fdb_chain;
+	struct mlx5_flow_table *fdb;
+	struct mlx5_flow_table *next_fdb;
+	struct mlx5_flow_group *miss_group;
+	struct mlx5_flow_handle *miss_rule;
+};
+
+static const struct rhashtable_params chain_params = {
+	.head_offset = offsetof(struct fdb_chain, node),
+	.key_offset = offsetof(struct fdb_chain, chain),
+	.key_len = sizeof_field(struct fdb_chain, chain),
+	.automatic_shrinking = true,
+};
+
+static const struct rhashtable_params prio_params = {
+	.head_offset = offsetof(struct fdb_prio, node),
+	.key_offset = offsetof(struct fdb_prio, key),
+	.key_len = sizeof_field(struct fdb_prio, key),
+	.automatic_shrinking = true,
+};
+
+bool mlx5_esw_chains_prios_supported(struct mlx5_eswitch *esw)
+{
+	return esw->fdb_table.flags & ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED;
+}
+
+u32 mlx5_esw_chains_get_chain_range(struct mlx5_eswitch *esw)
+{
+	if (!mlx5_esw_chains_prios_supported(esw))
+		return 1;
+
+	if (fdb_ignore_flow_level_supported(esw))
+		return UINT_MAX - 1;
+
+	return FDB_TC_MAX_CHAIN;
+}
+
+u32 mlx5_esw_chains_get_ft_chain(struct mlx5_eswitch *esw)
+{
+	return mlx5_esw_chains_get_chain_range(esw) + 1;
+}
+
+u32 mlx5_esw_chains_get_prio_range(struct mlx5_eswitch *esw)
+{
+	if (!mlx5_esw_chains_prios_supported(esw))
+		return 1;
+
+	if (fdb_ignore_flow_level_supported(esw))
+		return UINT_MAX;
+
+	return FDB_TC_MAX_PRIO;
+}
+
+static unsigned int mlx5_esw_chains_get_level_range(struct mlx5_eswitch *esw)
+{
+	if (fdb_ignore_flow_level_supported(esw))
+		return UINT_MAX;
+
+	return FDB_TC_LEVELS_PER_PRIO;
+}
+
+#define POOL_NEXT_SIZE 0
+static int
+mlx5_esw_chains_get_avail_sz_from_pool(struct mlx5_eswitch *esw,
+				       int desired_size)
+{
+	int i, found_i = -1;
+
+	for (i = ARRAY_SIZE(ESW_POOLS) - 1; i >= 0; i--) {
+		if (fdb_pool_left(esw)[i] && ESW_POOLS[i] > desired_size) {
+			found_i = i;
+			if (desired_size != POOL_NEXT_SIZE)
+				break;
+		}
+	}
+
+	if (found_i != -1) {
+		--fdb_pool_left(esw)[found_i];
+		return ESW_POOLS[found_i];
+	}
+
+	return 0;
+}
+
+static void
+mlx5_esw_chains_put_sz_to_pool(struct mlx5_eswitch *esw, int sz)
+{
+	int i;
+
+	for (i = ARRAY_SIZE(ESW_POOLS) - 1; i >= 0; i--) {
+		if (sz == ESW_POOLS[i]) {
+			++fdb_pool_left(esw)[i];
+			return;
+		}
+	}
+
+	WARN_ONCE(1, "Couldn't find size %d in fdb size pool", sz);
+}
+
+static void
+mlx5_esw_chains_init_sz_pool(struct mlx5_eswitch *esw)
+{
+	u32 fdb_max;
+	int i;
+
+	fdb_max = 1 << MLX5_CAP_ESW_FLOWTABLE_FDB(esw->dev, log_max_ft_size);
+
+	for (i = ARRAY_SIZE(ESW_POOLS) - 1; i >= 0; i--)
+		fdb_pool_left(esw)[i] =
+			ESW_POOLS[i] <= fdb_max ? ESW_SIZE / ESW_POOLS[i] : 0;
+}
+
+static struct mlx5_flow_table *
+mlx5_esw_chains_create_fdb_table(struct mlx5_eswitch *esw,
+				 u32 chain, u32 prio, u32 level)
+{
+	struct mlx5_flow_table_attr ft_attr = {};
+	struct mlx5_flow_namespace *ns;
+	struct mlx5_flow_table *fdb;
+	int sz;
+
+	if (esw->offloads.encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE)
+		ft_attr.flags |= (MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT |
+				  MLX5_FLOW_TABLE_TUNNEL_EN_DECAP);
+
+	sz = mlx5_esw_chains_get_avail_sz_from_pool(esw, POOL_NEXT_SIZE);
+	if (!sz)
+		return ERR_PTR(-ENOSPC);
+	ft_attr.max_fte = sz;
+
+	/* We use tc_slow_fdb(esw) as the table's next_ft till
+	 * ignore_flow_level is allowed on FT creation and not just for FTEs.
+	 * Instead caller should add an explicit miss rule if needed.
+	 */
+	ft_attr.next_ft = tc_slow_fdb(esw);
+
+	/* The root table(chain 0, prio 1, level 0) is required to be
+	 * connected to the previous prio (FDB_BYPASS_PATH if exists).
+	 * We always create it, as a managed table, in order to align with
+	 * fs_core logic.
+	 */
+	if (!fdb_ignore_flow_level_supported(esw) ||
+	    (chain == 0 && prio == 1 && level == 0)) {
+		ft_attr.level = level;
+		ft_attr.prio = prio - 1;
+		ns = mlx5_get_fdb_sub_ns(esw->dev, chain);
+	} else {
+		ft_attr.flags |= MLX5_FLOW_TABLE_UNMANAGED;
+		ft_attr.prio = FDB_TC_OFFLOAD;
+		/* Firmware doesn't allow us to create another level 0 table,
+		 * so we create all unmanaged tables as level 1.
+		 *
+		 * To connect them, we use explicit miss rules with
+		 * ignore_flow_level. Caller is responsible to create
+		 * these rules (if needed).
+		 */
+		ft_attr.level = 1;
+		ns = mlx5_get_flow_namespace(esw->dev, MLX5_FLOW_NAMESPACE_FDB);
+	}
+
+	ft_attr.autogroup.num_reserved_entries = 2;
+	ft_attr.autogroup.max_num_groups = ESW_OFFLOADS_NUM_GROUPS;
+	fdb = mlx5_create_auto_grouped_flow_table(ns, &ft_attr);
+	if (IS_ERR(fdb)) {
+		esw_warn(esw->dev,
+			 "Failed to create FDB table err %d (chain: %d, prio: %d, level: %d, size: %d)\n",
+			 (int)PTR_ERR(fdb), chain, prio, level, sz);
+		mlx5_esw_chains_put_sz_to_pool(esw, sz);
+		return fdb;
+	}
+
+	return fdb;
+}
+
+static void
+mlx5_esw_chains_destroy_fdb_table(struct mlx5_eswitch *esw,
+				  struct mlx5_flow_table *fdb)
+{
+	mlx5_esw_chains_put_sz_to_pool(esw, fdb->max_fte);
+	mlx5_destroy_flow_table(fdb);
+}
+
+static struct fdb_chain *
+mlx5_esw_chains_create_fdb_chain(struct mlx5_eswitch *esw, u32 chain)
+{
+	struct fdb_chain *fdb_chain = NULL;
+	int err;
+
+	fdb_chain = kvzalloc(sizeof(*fdb_chain), GFP_KERNEL);
+	if (!fdb_chain)
+		return ERR_PTR(-ENOMEM);
+
+	fdb_chain->esw = esw;
+	fdb_chain->chain = chain;
+	INIT_LIST_HEAD(&fdb_chain->prios_list);
+
+	err = rhashtable_insert_fast(&esw_chains_ht(esw), &fdb_chain->node,
+				     chain_params);
+	if (err)
+		goto err_insert;
+
+	return fdb_chain;
+
+err_insert:
+	kvfree(fdb_chain);
+	return ERR_PTR(err);
+}
+
+static void
+mlx5_esw_chains_destroy_fdb_chain(struct fdb_chain *fdb_chain)
+{
+	struct mlx5_eswitch *esw = fdb_chain->esw;
+
+	rhashtable_remove_fast(&esw_chains_ht(esw), &fdb_chain->node,
+			       chain_params);
+	kvfree(fdb_chain);
+}
+
+static struct fdb_chain *
+mlx5_esw_chains_get_fdb_chain(struct mlx5_eswitch *esw, u32 chain)
+{
+	struct fdb_chain *fdb_chain;
+
+	fdb_chain = rhashtable_lookup_fast(&esw_chains_ht(esw), &chain,
+					   chain_params);
+	if (!fdb_chain) {
+		fdb_chain = mlx5_esw_chains_create_fdb_chain(esw, chain);
+		if (IS_ERR(fdb_chain))
+			return fdb_chain;
+	}
+
+	fdb_chain->ref++;
+
+	return fdb_chain;
+}
+
+static struct mlx5_flow_handle *
+mlx5_esw_chains_add_miss_rule(struct mlx5_flow_table *fdb,
+			      struct mlx5_flow_table *next_fdb)
+{
+	static const struct mlx5_flow_spec spec = {};
+	struct mlx5_flow_destination dest = {};
+	struct mlx5_flow_act act = {};
+
+	act.flags  = FLOW_ACT_IGNORE_FLOW_LEVEL | FLOW_ACT_NO_APPEND;
+	act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+	dest.type  = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	dest.ft = next_fdb;
+
+	return mlx5_add_flow_rules(fdb, &spec, &act, &dest, 1);
+}
+
+static int
+mlx5_esw_chains_update_prio_prevs(struct fdb_prio *fdb_prio,
+				  struct mlx5_flow_table *next_fdb)
+{
+	struct mlx5_flow_handle *miss_rules[FDB_TC_LEVELS_PER_PRIO + 1] = {};
+	struct fdb_chain *fdb_chain = fdb_prio->fdb_chain;
+	struct fdb_prio *pos;
+	int n = 0, err;
+
+	if (fdb_prio->key.level)
+		return 0;
+
+	/* Iterate in reverse order until reaching the level 0 rule of
+	 * the previous priority, adding all the miss rules first, so we can
+	 * revert them if any of them fails.
+	 */
+	pos = fdb_prio;
+	list_for_each_entry_continue_reverse(pos,
+					     &fdb_chain->prios_list,
+					     list) {
+		miss_rules[n] = mlx5_esw_chains_add_miss_rule(pos->fdb,
+							      next_fdb);
+		if (IS_ERR(miss_rules[n])) {
+			err = PTR_ERR(miss_rules[n]);
+			goto err_prev_rule;
+		}
+
+		n++;
+		if (!pos->key.level)
+			break;
+	}
+
+	/* Success, delete old miss rules, and update the pointers. */
+	n = 0;
+	pos = fdb_prio;
+	list_for_each_entry_continue_reverse(pos,
+					     &fdb_chain->prios_list,
+					     list) {
+		mlx5_del_flow_rules(pos->miss_rule);
+
+		pos->miss_rule = miss_rules[n];
+		pos->next_fdb = next_fdb;
+
+		n++;
+		if (!pos->key.level)
+			break;
+	}
+
+	return 0;
+
+err_prev_rule:
+	while (--n >= 0)
+		mlx5_del_flow_rules(miss_rules[n]);
+
+	return err;
+}
+
+static void
+mlx5_esw_chains_put_fdb_chain(struct fdb_chain *fdb_chain)
+{
+	if (--fdb_chain->ref == 0)
+		mlx5_esw_chains_destroy_fdb_chain(fdb_chain);
+}
+
+static struct fdb_prio *
+mlx5_esw_chains_create_fdb_prio(struct mlx5_eswitch *esw,
+				u32 chain, u32 prio, u32 level)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	struct mlx5_flow_handle *miss_rule = NULL;
+	struct mlx5_flow_group *miss_group;
+	struct fdb_prio *fdb_prio = NULL;
+	struct mlx5_flow_table *next_fdb;
+	struct fdb_chain *fdb_chain;
+	struct mlx5_flow_table *fdb;
+	struct list_head *pos;
+	u32 *flow_group_in;
+	int err;
+
+	fdb_chain = mlx5_esw_chains_get_fdb_chain(esw, chain);
+	if (IS_ERR(fdb_chain))
+		return ERR_CAST(fdb_chain);
+
+	fdb_prio = kvzalloc(sizeof(*fdb_prio), GFP_KERNEL);
+	flow_group_in = kvzalloc(inlen, GFP_KERNEL);
+	if (!fdb_prio || !flow_group_in) {
+		err = -ENOMEM;
+		goto err_alloc;
+	}
+
+	/* Chain's prio list is sorted by prio and level.
+	 * And all levels of some prio point to the next prio's level 0.
+	 * Example list (prio, level):
+	 * (3,0)->(3,1)->(5,0)->(5,1)->(6,1)->(7,0)
+	 * In hardware, we will we have the following pointers:
+	 * (3,0) -> (5,0) -> (7,0) -> Slow path
+	 * (3,1) -> (5,0)
+	 * (5,1) -> (7,0)
+	 * (6,1) -> (7,0)
+	 */
+
+	/* Default miss for each chain: */
+	next_fdb = (chain == mlx5_esw_chains_get_ft_chain(esw)) ?
+		    tc_slow_fdb(esw) :
+		    tc_end_fdb(esw);
+	list_for_each(pos, &fdb_chain->prios_list) {
+		struct fdb_prio *p = list_entry(pos, struct fdb_prio, list);
+
+		/* exit on first pos that is larger */
+		if (prio < p->key.prio || (prio == p->key.prio &&
+					   level < p->key.level)) {
+			/* Get next level 0 table */
+			next_fdb = p->key.level == 0 ? p->fdb : p->next_fdb;
+			break;
+		}
+	}
+
+	fdb = mlx5_esw_chains_create_fdb_table(esw, chain, prio, level);
+	if (IS_ERR(fdb)) {
+		err = PTR_ERR(fdb);
+		goto err_create;
+	}
+
+	MLX5_SET(create_flow_group_in, flow_group_in, start_flow_index,
+		 fdb->max_fte - 2);
+	MLX5_SET(create_flow_group_in, flow_group_in, end_flow_index,
+		 fdb->max_fte - 1);
+	miss_group = mlx5_create_flow_group(fdb, flow_group_in);
+	if (IS_ERR(miss_group)) {
+		err = PTR_ERR(miss_group);
+		goto err_group;
+	}
+
+	/* Add miss rule to next_fdb */
+	miss_rule = mlx5_esw_chains_add_miss_rule(fdb, next_fdb);
+	if (IS_ERR(miss_rule)) {
+		err = PTR_ERR(miss_rule);
+		goto err_miss_rule;
+	}
+
+	fdb_prio->miss_group = miss_group;
+	fdb_prio->miss_rule = miss_rule;
+	fdb_prio->next_fdb = next_fdb;
+	fdb_prio->fdb_chain = fdb_chain;
+	fdb_prio->key.chain = chain;
+	fdb_prio->key.prio = prio;
+	fdb_prio->key.level = level;
+	fdb_prio->fdb = fdb;
+
+	err = rhashtable_insert_fast(&esw_prios_ht(esw), &fdb_prio->node,
+				     prio_params);
+	if (err)
+		goto err_insert;
+
+	list_add(&fdb_prio->list, pos->prev);
+
+	/* Table is ready, connect it */
+	err = mlx5_esw_chains_update_prio_prevs(fdb_prio, fdb);
+	if (err)
+		goto err_update;
+
+	kvfree(flow_group_in);
+	return fdb_prio;
+
+err_update:
+	list_del(&fdb_prio->list);
+	rhashtable_remove_fast(&esw_prios_ht(esw), &fdb_prio->node,
+			       prio_params);
+err_insert:
+	mlx5_del_flow_rules(miss_rule);
+err_miss_rule:
+	mlx5_destroy_flow_group(miss_group);
+err_group:
+	mlx5_esw_chains_destroy_fdb_table(esw, fdb);
+err_create:
+err_alloc:
+	kvfree(fdb_prio);
+	kvfree(flow_group_in);
+	mlx5_esw_chains_put_fdb_chain(fdb_chain);
+	return ERR_PTR(err);
+}
+
+static void
+mlx5_esw_chains_destroy_fdb_prio(struct mlx5_eswitch *esw,
+				 struct fdb_prio *fdb_prio)
+{
+	struct fdb_chain *fdb_chain = fdb_prio->fdb_chain;
+
+	WARN_ON(mlx5_esw_chains_update_prio_prevs(fdb_prio,
+						  fdb_prio->next_fdb));
+
+	list_del(&fdb_prio->list);
+	rhashtable_remove_fast(&esw_prios_ht(esw), &fdb_prio->node,
+			       prio_params);
+	mlx5_del_flow_rules(fdb_prio->miss_rule);
+	mlx5_destroy_flow_group(fdb_prio->miss_group);
+	mlx5_esw_chains_destroy_fdb_table(esw, fdb_prio->fdb);
+	mlx5_esw_chains_put_fdb_chain(fdb_chain);
+	kvfree(fdb_prio);
+}
+
+struct mlx5_flow_table *
+mlx5_esw_chains_get_table(struct mlx5_eswitch *esw, u32 chain, u32 prio,
+			  u32 level)
+{
+	struct mlx5_flow_table *prev_fts;
+	struct fdb_prio *fdb_prio;
+	struct fdb_prio_key key;
+	int l = 0;
+
+	if ((chain > mlx5_esw_chains_get_chain_range(esw) &&
+	     chain != mlx5_esw_chains_get_ft_chain(esw)) ||
+	    prio > mlx5_esw_chains_get_prio_range(esw) ||
+	    level > mlx5_esw_chains_get_level_range(esw))
+		return ERR_PTR(-EOPNOTSUPP);
+
+	/* create earlier levels for correct fs_core lookup when
+	 * connecting tables.
+	 */
+	for (l = 0; l < level; l++) {
+		prev_fts = mlx5_esw_chains_get_table(esw, chain, prio, l);
+		if (IS_ERR(prev_fts)) {
+			fdb_prio = ERR_CAST(prev_fts);
+			goto err_get_prevs;
+		}
+	}
+
+	key.chain = chain;
+	key.prio = prio;
+	key.level = level;
+
+	mutex_lock(&esw_chains_lock(esw));
+	fdb_prio = rhashtable_lookup_fast(&esw_prios_ht(esw), &key,
+					  prio_params);
+	if (!fdb_prio) {
+		fdb_prio = mlx5_esw_chains_create_fdb_prio(esw, chain,
+							   prio, level);
+		if (IS_ERR(fdb_prio))
+			goto err_create_prio;
+	}
+
+	++fdb_prio->ref;
+	mutex_unlock(&esw_chains_lock(esw));
+
+	return fdb_prio->fdb;
+
+err_create_prio:
+	mutex_unlock(&esw_chains_lock(esw));
+err_get_prevs:
+	while (--l >= 0)
+		mlx5_esw_chains_put_table(esw, chain, prio, l);
+	return ERR_CAST(fdb_prio);
+}
+
+void
+mlx5_esw_chains_put_table(struct mlx5_eswitch *esw, u32 chain, u32 prio,
+			  u32 level)
+{
+	struct fdb_prio *fdb_prio;
+	struct fdb_prio_key key;
+
+	key.chain = chain;
+	key.prio = prio;
+	key.level = level;
+
+	mutex_lock(&esw_chains_lock(esw));
+	fdb_prio = rhashtable_lookup_fast(&esw_prios_ht(esw), &key,
+					  prio_params);
+	if (!fdb_prio)
+		goto err_get_prio;
+
+	if (--fdb_prio->ref == 0)
+		mlx5_esw_chains_destroy_fdb_prio(esw, fdb_prio);
+	mutex_unlock(&esw_chains_lock(esw));
+
+	while (level-- > 0)
+		mlx5_esw_chains_put_table(esw, chain, prio, level);
+
+	return;
+
+err_get_prio:
+	mutex_unlock(&esw_chains_lock(esw));
+	WARN_ONCE(1,
+		  "Couldn't find table: (chain: %d prio: %d level: %d)",
+		  chain, prio, level);
+}
+
+struct mlx5_flow_table *
+mlx5_esw_chains_get_tc_end_ft(struct mlx5_eswitch *esw)
+{
+	return tc_end_fdb(esw);
+}
+
+static int
+mlx5_esw_chains_init(struct mlx5_eswitch *esw)
+{
+	struct mlx5_esw_chains_priv *chains_priv;
+	struct mlx5_core_dev *dev = esw->dev;
+	u32 max_flow_counter, fdb_max;
+	int err;
+
+	chains_priv = kzalloc(sizeof(*chains_priv), GFP_KERNEL);
+	if (!chains_priv)
+		return -ENOMEM;
+	esw_chains_priv(esw) = chains_priv;
+
+	max_flow_counter = (MLX5_CAP_GEN(dev, max_flow_counter_31_16) << 16) |
+			    MLX5_CAP_GEN(dev, max_flow_counter_15_0);
+	fdb_max = 1 << MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size);
+
+	esw_debug(dev,
+		  "Init esw offloads chains, max counters(%d), groups(%d), max flow table size(%d)\n",
+		  max_flow_counter, ESW_OFFLOADS_NUM_GROUPS, fdb_max);
+
+	mlx5_esw_chains_init_sz_pool(esw);
+
+	if (!MLX5_CAP_ESW_FLOWTABLE(esw->dev, multi_fdb_encap) &&
+	    esw->offloads.encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE) {
+		esw->fdb_table.flags &= ~ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED;
+		esw_warn(dev, "Tc chains and priorities offload aren't supported, update firmware if needed\n");
+	} else {
+		esw->fdb_table.flags |= ESW_FDB_CHAINS_AND_PRIOS_SUPPORTED;
+		esw_info(dev, "Supported tc offload range - chains: %u, prios: %u\n",
+			 mlx5_esw_chains_get_chain_range(esw),
+			 mlx5_esw_chains_get_prio_range(esw));
+	}
+
+	err = rhashtable_init(&esw_chains_ht(esw), &chain_params);
+	if (err)
+		goto init_chains_ht_err;
+
+	err = rhashtable_init(&esw_prios_ht(esw), &prio_params);
+	if (err)
+		goto init_prios_ht_err;
+
+	mutex_init(&esw_chains_lock(esw));
+
+	return 0;
+
+init_prios_ht_err:
+	rhashtable_destroy(&esw_chains_ht(esw));
+init_chains_ht_err:
+	kfree(chains_priv);
+	return err;
+}
+
+static void
+mlx5_esw_chains_cleanup(struct mlx5_eswitch *esw)
+{
+	mutex_destroy(&esw_chains_lock(esw));
+	rhashtable_destroy(&esw_prios_ht(esw));
+	rhashtable_destroy(&esw_chains_ht(esw));
+
+	kfree(esw_chains_priv(esw));
+}
+
+static int
+mlx5_esw_chains_open(struct mlx5_eswitch *esw)
+{
+	struct mlx5_flow_table *ft;
+	int err;
+
+	/* Create tc_end_fdb(esw) which is the always created ft chain */
+	ft = mlx5_esw_chains_get_table(esw, mlx5_esw_chains_get_ft_chain(esw),
+				       1, 0);
+	if (IS_ERR(ft))
+		return PTR_ERR(ft);
+
+	tc_end_fdb(esw) = ft;
+
+	/* Always open the root for fast path */
+	ft = mlx5_esw_chains_get_table(esw, 0, 1, 0);
+	if (IS_ERR(ft)) {
+		err = PTR_ERR(ft);
+		goto level_0_err;
+	}
+
+	/* Open level 1 for split rules now if prios isn't supported  */
+	if (!mlx5_esw_chains_prios_supported(esw)) {
+		ft = mlx5_esw_chains_get_table(esw, 0, 1, 1);
+
+		if (IS_ERR(ft)) {
+			err = PTR_ERR(ft);
+			goto level_1_err;
+		}
+	}
+
+	return 0;
+
+level_1_err:
+	mlx5_esw_chains_put_table(esw, 0, 1, 0);
+level_0_err:
+	mlx5_esw_chains_put_table(esw, mlx5_esw_chains_get_ft_chain(esw), 1, 0);
+	return err;
+}
+
+static void
+mlx5_esw_chains_close(struct mlx5_eswitch *esw)
+{
+	if (!mlx5_esw_chains_prios_supported(esw))
+		mlx5_esw_chains_put_table(esw, 0, 1, 1);
+	mlx5_esw_chains_put_table(esw, 0, 1, 0);
+	mlx5_esw_chains_put_table(esw, mlx5_esw_chains_get_ft_chain(esw), 1, 0);
+}
+
+int
+mlx5_esw_chains_create(struct mlx5_eswitch *esw)
+{
+	int err;
+
+	err = mlx5_esw_chains_init(esw);
+	if (err)
+		return err;
+
+	err = mlx5_esw_chains_open(esw);
+	if (err)
+		goto err_open;
+
+	return 0;
+
+err_open:
+	mlx5_esw_chains_cleanup(esw);
+	return err;
+}
+
+void
+mlx5_esw_chains_destroy(struct mlx5_eswitch *esw)
+{
+	mlx5_esw_chains_close(esw);
+	mlx5_esw_chains_cleanup(esw);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.h
new file mode 100644
index 000000000000..2e13097fe348
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_chains.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2020 Mellanox Technologies. */
+
+#ifndef __ML5_ESW_CHAINS_H__
+#define __ML5_ESW_CHAINS_H__
+
+bool
+mlx5_esw_chains_prios_supported(struct mlx5_eswitch *esw);
+u32
+mlx5_esw_chains_get_prio_range(struct mlx5_eswitch *esw);
+u32
+mlx5_esw_chains_get_chain_range(struct mlx5_eswitch *esw);
+u32
+mlx5_esw_chains_get_ft_chain(struct mlx5_eswitch *esw);
+
+struct mlx5_flow_table *
+mlx5_esw_chains_get_table(struct mlx5_eswitch *esw, u32 chain, u32 prio,
+			  u32 level);
+void
+mlx5_esw_chains_put_table(struct mlx5_eswitch *esw, u32 chain, u32 prio,
+			  u32 level);
+
+struct mlx5_flow_table *
+mlx5_esw_chains_get_tc_end_ft(struct mlx5_eswitch *esw);
+
+int mlx5_esw_chains_create(struct mlx5_eswitch *esw);
+void mlx5_esw_chains_destroy(struct mlx5_eswitch *esw);
+
+#endif /* __ML5_ESW_CHAINS_H__ */
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c
index 366bda1bb1c3..dc08ed9339ab 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c
@@ -50,8 +50,8 @@ mlx5_eswitch_termtbl_create(struct mlx5_core_dev *dev,
 			    struct mlx5_flow_act *flow_act)
 {
 	static const struct mlx5_flow_spec spec = {};
+	struct mlx5_flow_table_attr ft_attr = {};
 	struct mlx5_flow_namespace *root_ns;
-	int prio, flags;
 	int err;
 
 	root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_FDB);
@@ -63,10 +63,11 @@ mlx5_eswitch_termtbl_create(struct mlx5_core_dev *dev,
 	/* As this is the terminating action then the termination table is the
 	 * same prio as the slow path
 	 */
-	prio = FDB_SLOW_PATH;
-	flags = MLX5_FLOW_TABLE_TERMINATION;
-	tt->termtbl = mlx5_create_auto_grouped_flow_table(root_ns, prio, 1, 1,
-							  0, flags);
+	ft_attr.flags = MLX5_FLOW_TABLE_TERMINATION;
+	ft_attr.prio = FDB_SLOW_PATH;
+	ft_attr.max_fte = 1;
+	ft_attr.autogroup.max_num_groups = 1;
+	tt->termtbl = mlx5_create_auto_grouped_flow_table(root_ns, &ft_attr);
 	if (IS_ERR(tt->termtbl)) {
 		esw_warn(dev, "Failed to create termination table\n");
 		return -EOPNOTSUPP;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
index 3c816e81f8d9..b25465d9e030 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c
@@ -432,6 +432,9 @@ static int mlx5_cmd_set_fte(struct mlx5_core_dev *dev,
 	MLX5_SET(set_fte_in, in, table_type, ft->type);
 	MLX5_SET(set_fte_in, in, table_id,   ft->id);
 	MLX5_SET(set_fte_in, in, flow_index, fte->index);
+	MLX5_SET(set_fte_in, in, ignore_flow_level,
+		 !!(fte->action.flags & FLOW_ACT_IGNORE_FLOW_LEVEL));
+
 	if (ft->vport) {
 		MLX5_SET(set_fte_in, in, vport_number, ft->vport);
 		MLX5_SET(set_fte_in, in, other_vport, 1);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 8c5df6c7d7b6..c7a16ae05fa8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -579,7 +579,9 @@ static void del_sw_flow_group(struct fs_node *node)
 
 	rhashtable_destroy(&fg->ftes_hash);
 	ida_destroy(&fg->fte_allocator);
-	if (ft->autogroup.active && fg->max_ftes == ft->autogroup.group_size)
+	if (ft->autogroup.active &&
+	    fg->max_ftes == ft->autogroup.group_size &&
+	    fg->start_index < ft->autogroup.max_fte)
 		ft->autogroup.num_groups--;
 	err = rhltable_remove(&ft->fgs_hash,
 			      &fg->hash,
@@ -1006,7 +1008,8 @@ static struct mlx5_flow_table *__mlx5_create_flow_table(struct mlx5_flow_namespa
 							u16 vport)
 {
 	struct mlx5_flow_root_namespace *root = find_root(&ns->node);
-	struct mlx5_flow_table *next_ft = NULL;
+	bool unmanaged = ft_attr->flags & MLX5_FLOW_TABLE_UNMANAGED;
+	struct mlx5_flow_table *next_ft;
 	struct fs_prio *fs_prio = NULL;
 	struct mlx5_flow_table *ft;
 	int log_table_sz;
@@ -1023,14 +1026,21 @@ static struct mlx5_flow_table *__mlx5_create_flow_table(struct mlx5_flow_namespa
 		err = -EINVAL;
 		goto unlock_root;
 	}
-	if (ft_attr->level >= fs_prio->num_levels) {
-		err = -ENOSPC;
-		goto unlock_root;
+	if (!unmanaged) {
+		/* The level is related to the
+		 * priority level range.
+		 */
+		if (ft_attr->level >= fs_prio->num_levels) {
+			err = -ENOSPC;
+			goto unlock_root;
+		}
+
+		ft_attr->level += fs_prio->start_level;
 	}
+
 	/* The level is related to the
 	 * priority level range.
 	 */
-	ft_attr->level += fs_prio->start_level;
 	ft = alloc_flow_table(ft_attr->level,
 			      vport,
 			      ft_attr->max_fte ? roundup_pow_of_two(ft_attr->max_fte) : 0,
@@ -1043,19 +1053,27 @@ static struct mlx5_flow_table *__mlx5_create_flow_table(struct mlx5_flow_namespa
 
 	tree_init_node(&ft->node, del_hw_flow_table, del_sw_flow_table);
 	log_table_sz = ft->max_fte ? ilog2(ft->max_fte) : 0;
-	next_ft = find_next_chained_ft(fs_prio);
+	next_ft = unmanaged ? ft_attr->next_ft :
+			      find_next_chained_ft(fs_prio);
 	ft->def_miss_action = ns->def_miss_action;
 	err = root->cmds->create_flow_table(root, ft, log_table_sz, next_ft);
 	if (err)
 		goto free_ft;
 
-	err = connect_flow_table(root->dev, ft, fs_prio);
-	if (err)
-		goto destroy_ft;
+	if (!unmanaged) {
+		err = connect_flow_table(root->dev, ft, fs_prio);
+		if (err)
+			goto destroy_ft;
+	}
+
 	ft->node.active = true;
 	down_write_ref_node(&fs_prio->node, false);
-	tree_add_node(&ft->node, &fs_prio->node);
-	list_add_flow_table(ft, fs_prio);
+	if (!unmanaged) {
+		tree_add_node(&ft->node, &fs_prio->node);
+		list_add_flow_table(ft, fs_prio);
+	} else {
+		ft->node.root = fs_prio->node.root;
+	}
 	fs_prio->num_ft++;
 	up_write_ref_node(&fs_prio->node, false);
 	mutex_unlock(&root->chain_lock);
@@ -1103,31 +1121,27 @@ EXPORT_SYMBOL(mlx5_create_lag_demux_flow_table);
 
 struct mlx5_flow_table*
 mlx5_create_auto_grouped_flow_table(struct mlx5_flow_namespace *ns,
-				    int prio,
-				    int num_flow_table_entries,
-				    int max_num_groups,
-				    u32 level,
-				    u32 flags)
+				    struct mlx5_flow_table_attr *ft_attr)
 {
-	struct mlx5_flow_table_attr ft_attr = {};
+	int num_reserved_entries = ft_attr->autogroup.num_reserved_entries;
+	int autogroups_max_fte = ft_attr->max_fte - num_reserved_entries;
+	int max_num_groups = ft_attr->autogroup.max_num_groups;
 	struct mlx5_flow_table *ft;
 
-	if (max_num_groups > num_flow_table_entries)
+	if (max_num_groups > autogroups_max_fte)
+		return ERR_PTR(-EINVAL);
+	if (num_reserved_entries > ft_attr->max_fte)
 		return ERR_PTR(-EINVAL);
 
-	ft_attr.max_fte = num_flow_table_entries;
-	ft_attr.prio    = prio;
-	ft_attr.level   = level;
-	ft_attr.flags   = flags;
-
-	ft = mlx5_create_flow_table(ns, &ft_attr);
+	ft = mlx5_create_flow_table(ns, ft_attr);
 	if (IS_ERR(ft))
 		return ft;
 
 	ft->autogroup.active = true;
 	ft->autogroup.required_groups = max_num_groups;
+	ft->autogroup.max_fte = autogroups_max_fte;
 	/* We save place for flow groups in addition to max types */
-	ft->autogroup.group_size = ft->max_fte / (max_num_groups + 1);
+	ft->autogroup.group_size = autogroups_max_fte / (max_num_groups + 1);
 
 	return ft;
 }
@@ -1149,7 +1163,7 @@ struct mlx5_flow_group *mlx5_create_flow_group(struct mlx5_flow_table *ft,
 	struct mlx5_flow_group *fg;
 	int err;
 
-	if (ft->autogroup.active)
+	if (ft->autogroup.active && start_index < ft->autogroup.max_fte)
 		return ERR_PTR(-EPERM);
 
 	down_write_ref_node(&ft->node, false);
@@ -1322,9 +1336,10 @@ static struct mlx5_flow_group *alloc_auto_flow_group(struct mlx5_flow_table  *ft
 						     const struct mlx5_flow_spec *spec)
 {
 	struct list_head *prev = &ft->node.children;
-	struct mlx5_flow_group *fg;
+	u32 max_fte = ft->autogroup.max_fte;
 	unsigned int candidate_index = 0;
 	unsigned int group_size = 0;
+	struct mlx5_flow_group *fg;
 
 	if (!ft->autogroup.active)
 		return ERR_PTR(-ENOENT);
@@ -1332,7 +1347,7 @@ static struct mlx5_flow_group *alloc_auto_flow_group(struct mlx5_flow_table  *ft
 	if (ft->autogroup.num_groups < ft->autogroup.required_groups)
 		group_size = ft->autogroup.group_size;
 
-	/*  ft->max_fte == ft->autogroup.max_types */
+	/*  max_fte == ft->autogroup.max_types */
 	if (group_size == 0)
 		group_size = 1;
 
@@ -1345,7 +1360,7 @@ static struct mlx5_flow_group *alloc_auto_flow_group(struct mlx5_flow_table  *ft
 		prev = &fg->node.list;
 	}
 
-	if (candidate_index + group_size > ft->max_fte)
+	if (candidate_index + group_size > max_fte)
 		return ERR_PTR(-ENOSPC);
 
 	fg = alloc_insert_flow_group(ft,
@@ -1529,18 +1544,30 @@ static bool counter_is_valid(u32 action)
 }
 
 static bool dest_is_valid(struct mlx5_flow_destination *dest,
-			  u32 action,
+			  struct mlx5_flow_act *flow_act,
 			  struct mlx5_flow_table *ft)
 {
+	bool ignore_level = flow_act->flags & FLOW_ACT_IGNORE_FLOW_LEVEL;
+	u32 action = flow_act->action;
+
 	if (dest && (dest->type == MLX5_FLOW_DESTINATION_TYPE_COUNTER))
 		return counter_is_valid(action);
 
 	if (!(action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST))
 		return true;
 
+	if (ignore_level) {
+		if (ft->type != FS_FT_FDB)
+			return false;
+
+		if (dest->type == MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE &&
+		    dest->ft->type != FS_FT_FDB)
+			return false;
+	}
+
 	if (!dest || ((dest->type ==
 	    MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE) &&
-	    (dest->ft->level <= ft->level)))
+	    (dest->ft->level <= ft->level && !ignore_level)))
 		return false;
 	return true;
 }
@@ -1770,7 +1797,7 @@ _mlx5_add_flow_rules(struct mlx5_flow_table *ft,
 		return ERR_PTR(-EINVAL);
 
 	for (i = 0; i < dest_num; i++) {
-		if (!dest_is_valid(&dest[i], flow_act->action, ft))
+		if (!dest_is_valid(&dest[i], flow_act, ft))
 			return ERR_PTR(-EINVAL);
 	}
 	nested_down_read_ref_node(&ft->node, FS_LOCK_GRANDPARENT);
@@ -2033,7 +2060,8 @@ int mlx5_destroy_flow_table(struct mlx5_flow_table *ft)
 	int err = 0;
 
 	mutex_lock(&root->chain_lock);
-	err = disconnect_flow_table(ft);
+	if (!(ft->flags & MLX5_FLOW_TABLE_UNMANAGED))
+		err = disconnect_flow_table(ft);
 	if (err) {
 		mutex_unlock(&root->chain_lock);
 		return err;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
index c2621b911563..be5f5e32c1e8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h
@@ -164,6 +164,7 @@ struct mlx5_flow_table {
 		unsigned int		required_groups;
 		unsigned int		group_size;
 		unsigned int		num_groups;
+		unsigned int		max_fte;
 	} autogroup;
 	/* Protect fwd_rules */
 	struct mutex			lock;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fw.c b/drivers/net/ethernet/mellanox/mlx5/core/fw.c
index a19790dee7b2..d89ff1d09119 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fw.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fw.c
@@ -131,11 +131,11 @@ static int mlx5_get_pcam_reg(struct mlx5_core_dev *dev)
 				   MLX5_PCAM_REGS_5000_TO_507F);
 }
 
-static int mlx5_get_mcam_reg(struct mlx5_core_dev *dev)
+static int mlx5_get_mcam_access_reg_group(struct mlx5_core_dev *dev,
+					  enum mlx5_mcam_reg_groups group)
 {
-	return mlx5_query_mcam_reg(dev, dev->caps.mcam,
-				   MLX5_MCAM_FEATURE_ENHANCED_FEATURES,
-				   MLX5_MCAM_REGS_FIRST_128);
+	return mlx5_query_mcam_reg(dev, dev->caps.mcam[group],
+				   MLX5_MCAM_FEATURE_ENHANCED_FEATURES, group);
 }
 
 static int mlx5_get_qcam_reg(struct mlx5_core_dev *dev)
@@ -221,8 +221,11 @@ int mlx5_query_hca_caps(struct mlx5_core_dev *dev)
 	if (MLX5_CAP_GEN(dev, pcam_reg))
 		mlx5_get_pcam_reg(dev);
 
-	if (MLX5_CAP_GEN(dev, mcam_reg))
-		mlx5_get_mcam_reg(dev);
+	if (MLX5_CAP_GEN(dev, mcam_reg)) {
+		mlx5_get_mcam_access_reg_group(dev, MLX5_MCAM_REGS_FIRST_128);
+		mlx5_get_mcam_access_reg_group(dev, MLX5_MCAM_REGS_0x9080_0x90FF);
+		mlx5_get_mcam_access_reg_group(dev, MLX5_MCAM_REGS_0x9100_0x917F);
+	}
 
 	if (MLX5_CAP_GEN(dev, qcam_reg))
 		mlx5_get_qcam_reg(dev);
@@ -245,6 +248,13 @@ int mlx5_query_hca_caps(struct mlx5_core_dev *dev)
 			return err;
 	}
 
+	if (MLX5_CAP_GEN_64(dev, general_obj_types) &
+		MLX5_GENERAL_OBJ_TYPES_CAP_VIRTIO_NET_Q) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_VDPA_EMULATION);
+		if (err)
+			return err;
+	}
+
 	return 0;
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
index 3ed8ab2d703d..56078b23f1a0 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib.c
@@ -87,8 +87,8 @@ int mlx5i_init(struct mlx5_core_dev *mdev,
 	mlx5e_set_netdev_mtu_boundaries(priv);
 	netdev->mtu = netdev->max_mtu;
 
-	mlx5e_build_nic_params(mdev, NULL, &priv->rss_params, &priv->channels.params,
-			       priv->max_nch, netdev->mtu);
+	mlx5e_build_nic_params(priv, NULL, &priv->rss_params, &priv->channels.params,
+			       netdev->mtu);
 	mlx5i_build_nic_params(mdev, &priv->channels.params);
 
 	mlx5e_timestamp_init(priv);
@@ -419,6 +419,28 @@ static void mlx5i_cleanup_rx(struct mlx5e_priv *priv)
 	mlx5e_destroy_q_counters(priv);
 }
 
+/* The stats groups order is opposite to the update_stats() order calls */
+static mlx5e_stats_grp_t mlx5i_stats_grps[] = {
+	&MLX5E_STATS_GRP(sw),
+	&MLX5E_STATS_GRP(qcnt),
+	&MLX5E_STATS_GRP(vnic_env),
+	&MLX5E_STATS_GRP(vport),
+	&MLX5E_STATS_GRP(802_3),
+	&MLX5E_STATS_GRP(2863),
+	&MLX5E_STATS_GRP(2819),
+	&MLX5E_STATS_GRP(phy),
+	&MLX5E_STATS_GRP(pcie),
+	&MLX5E_STATS_GRP(per_prio),
+	&MLX5E_STATS_GRP(pme),
+	&MLX5E_STATS_GRP(channels),
+	&MLX5E_STATS_GRP(per_port_buff_congest),
+};
+
+static unsigned int mlx5i_stats_grps_num(struct mlx5e_priv *priv)
+{
+	return ARRAY_SIZE(mlx5i_stats_grps);
+}
+
 static const struct mlx5e_profile mlx5i_nic_profile = {
 	.init		   = mlx5i_init,
 	.cleanup	   = mlx5i_cleanup,
@@ -435,6 +457,8 @@ static const struct mlx5e_profile mlx5i_nic_profile = {
 	.rx_handlers.handle_rx_cqe_mpwqe = NULL, /* Not supported */
 	.max_tc		   = MLX5I_MAX_NUM_TC,
 	.rq_groups	   = MLX5E_NUM_RQ_GROUPS(REGULAR),
+	.stats_grps        = mlx5i_stats_grps,
+	.stats_grps_num    = mlx5i_stats_grps_num,
 };
 
 /* mlx5i netdev NDos */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
index fc0d9583475d..b91eabc09fbc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c
@@ -586,7 +586,8 @@ void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev)
 
 	if (!ldev->nb.notifier_call) {
 		ldev->nb.notifier_call = mlx5_lag_netdev_event;
-		if (register_netdevice_notifier(&ldev->nb)) {
+		if (register_netdevice_notifier_dev_net(netdev, &ldev->nb,
+							&ldev->nn)) {
 			ldev->nb.notifier_call = NULL;
 			mlx5_core_err(dev, "Failed to register LAG netdev notifier\n");
 		}
@@ -599,7 +600,7 @@ void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev)
 }
 
 /* Must be called with intf_mutex held */
-void mlx5_lag_remove(struct mlx5_core_dev *dev)
+void mlx5_lag_remove(struct mlx5_core_dev *dev, struct net_device *netdev)
 {
 	struct mlx5_lag *ldev;
 	int i;
@@ -619,7 +620,8 @@ void mlx5_lag_remove(struct mlx5_core_dev *dev)
 
 	if (i == MLX5_MAX_PORTS) {
 		if (ldev->nb.notifier_call)
-			unregister_netdevice_notifier(&ldev->nb);
+			unregister_netdevice_notifier_dev_net(netdev, &ldev->nb,
+							      &ldev->nn);
 		mlx5_lag_mp_cleanup(ldev);
 		cancel_delayed_work_sync(&ldev->bond_work);
 		mlx5_lag_dev_free(ldev);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.h b/drivers/net/ethernet/mellanox/mlx5/core/lag.h
index f1068aac6406..316ab09e2664 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.h
@@ -44,6 +44,7 @@ struct mlx5_lag {
 	struct workqueue_struct   *wq;
 	struct delayed_work       bond_work;
 	struct notifier_block     nb;
+	struct netdev_net_notifier	nn;
 	struct lag_mp             lag_mp;
 };
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
index b70afa310ad2..416676c35b1f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c
@@ -200,8 +200,6 @@ static void mlx5_lag_fib_update(struct work_struct *work)
 	rtnl_lock();
 	switch (fib_work->event) {
 	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_APPEND: /* fall through */
-	case FIB_EVENT_ENTRY_ADD: /* fall through */
 	case FIB_EVENT_ENTRY_DEL:
 		mlx5_lag_fib_route_event(ldev, fib_work->event,
 					 fib_work->fen_info.fi);
@@ -259,8 +257,6 @@ static int mlx5_lag_fib_event(struct notifier_block *nb,
 
 	switch (event) {
 	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_APPEND: /* fall through */
-	case FIB_EVENT_ENTRY_ADD: /* fall through */
 	case FIB_EVENT_ENTRY_DEL:
 		fen_info = container_of(info, struct fib_entry_notifier_info,
 					info);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
index da67b28d6e23..fcce9e0fc82c 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h
@@ -157,7 +157,7 @@ int mlx5_query_qcam_reg(struct mlx5_core_dev *mdev, u32 *qcam,
 			u8 feature_group, u8 access_reg_group);
 
 void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev);
-void mlx5_lag_remove(struct mlx5_core_dev *dev);
+void mlx5_lag_remove(struct mlx5_core_dev *dev, struct net_device *netdev);
 
 int mlx5_irq_table_init(struct mlx5_core_dev *dev);
 void mlx5_irq_table_cleanup(struct mlx5_core_dev *dev);
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 004c56c2fc0c..6dec2a550a10 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_action.c
@@ -677,9 +677,12 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher,
 					goto out_invalid_arg;
 				}
 				if (action->dest_tbl.tbl->level <= matcher->tbl->level) {
+					mlx5_core_warn_once(dmn->mdev,
+							    "Connecting table to a lower/same level destination table\n");
 					mlx5dr_dbg(dmn,
-						   "Destination table level should be higher than source table\n");
-					goto out_invalid_arg;
+						   "Connecting table at level %d to a destination table at level %d\n",
+						   matcher->tbl->level,
+						   action->dest_tbl.tbl->level);
 				}
 				attr.final_icm_addr = rx_rule ?
 					action->dest_tbl.tbl->rx.s_anchor->chunk->icm_addr :
@@ -690,9 +693,9 @@ int mlx5dr_actions_build_ste_arr(struct mlx5dr_matcher *matcher,
 
 				/* get the relevant addresses */
 				if (!action->dest_tbl.fw_tbl.rx_icm_addr) {
-					ret = mlx5dr_cmd_query_flow_table(action->dest_tbl.fw_tbl.mdev,
-									  action->dest_tbl.fw_tbl.ft->type,
-									  action->dest_tbl.fw_tbl.ft->id,
+					ret = mlx5dr_cmd_query_flow_table(dmn->mdev,
+									  action->dest_tbl.fw_tbl.type,
+									  action->dest_tbl.fw_tbl.id,
 									  &output);
 					if (!ret) {
 						action->dest_tbl.fw_tbl.tx_icm_addr =
@@ -982,8 +985,106 @@ dec_ref:
 }
 
 struct mlx5dr_action *
-mlx5dr_create_action_dest_flow_fw_table(struct mlx5_flow_table *ft,
-					struct mlx5_core_dev *mdev)
+mlx5dr_action_create_mult_dest_tbl(struct mlx5dr_domain *dmn,
+				   struct mlx5dr_action_dest *dests,
+				   u32 num_of_dests)
+{
+	struct mlx5dr_cmd_flow_destination_hw_info *hw_dests;
+	struct mlx5dr_action **ref_actions;
+	struct mlx5dr_action *action;
+	bool reformat_req = false;
+	u32 num_of_ref = 0;
+	int ret;
+	int i;
+
+	if (dmn->type != MLX5DR_DOMAIN_TYPE_FDB) {
+		mlx5dr_err(dmn, "Multiple destination support is for FDB only\n");
+		return NULL;
+	}
+
+	hw_dests = kzalloc(sizeof(*hw_dests) * num_of_dests, GFP_KERNEL);
+	if (!hw_dests)
+		return NULL;
+
+	ref_actions = kzalloc(sizeof(*ref_actions) * num_of_dests * 2, GFP_KERNEL);
+	if (!ref_actions)
+		goto free_hw_dests;
+
+	for (i = 0; i < num_of_dests; i++) {
+		struct mlx5dr_action *reformat_action = dests[i].reformat;
+		struct mlx5dr_action *dest_action = dests[i].dest;
+
+		ref_actions[num_of_ref++] = dest_action;
+
+		switch (dest_action->action_type) {
+		case DR_ACTION_TYP_VPORT:
+			hw_dests[i].vport.flags = MLX5_FLOW_DEST_VPORT_VHCA_ID;
+			hw_dests[i].type = MLX5_FLOW_DESTINATION_TYPE_VPORT;
+			hw_dests[i].vport.num = dest_action->vport.caps->num;
+			hw_dests[i].vport.vhca_id = dest_action->vport.caps->vhca_gvmi;
+			if (reformat_action) {
+				reformat_req = true;
+				hw_dests[i].vport.reformat_id =
+					reformat_action->reformat.reformat_id;
+				ref_actions[num_of_ref++] = reformat_action;
+				hw_dests[i].vport.flags |= MLX5_FLOW_DEST_VPORT_REFORMAT_ID;
+			}
+			break;
+
+		case DR_ACTION_TYP_FT:
+			hw_dests[i].type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+			if (dest_action->dest_tbl.is_fw_tbl)
+				hw_dests[i].ft_id = dest_action->dest_tbl.fw_tbl.id;
+			else
+				hw_dests[i].ft_id = dest_action->dest_tbl.tbl->table_id;
+			break;
+
+		default:
+			mlx5dr_dbg(dmn, "Invalid multiple destinations action\n");
+			goto free_ref_actions;
+		}
+	}
+
+	action = dr_action_create_generic(DR_ACTION_TYP_FT);
+	if (!action)
+		goto free_ref_actions;
+
+	ret = mlx5dr_fw_create_md_tbl(dmn,
+				      hw_dests,
+				      num_of_dests,
+				      reformat_req,
+				      &action->dest_tbl.fw_tbl.id,
+				      &action->dest_tbl.fw_tbl.group_id);
+	if (ret)
+		goto free_action;
+
+	refcount_inc(&dmn->refcount);
+
+	for (i = 0; i < num_of_ref; i++)
+		refcount_inc(&ref_actions[i]->refcount);
+
+	action->dest_tbl.is_fw_tbl = true;
+	action->dest_tbl.fw_tbl.dmn = dmn;
+	action->dest_tbl.fw_tbl.type = FS_FT_FDB;
+	action->dest_tbl.fw_tbl.ref_actions = ref_actions;
+	action->dest_tbl.fw_tbl.num_of_ref_actions = num_of_ref;
+
+	kfree(hw_dests);
+
+	return action;
+
+free_action:
+	kfree(action);
+free_ref_actions:
+	kfree(ref_actions);
+free_hw_dests:
+	kfree(hw_dests);
+	return NULL;
+}
+
+struct mlx5dr_action *
+mlx5dr_action_create_dest_flow_fw_table(struct mlx5dr_domain *dmn,
+					struct mlx5_flow_table *ft)
 {
 	struct mlx5dr_action *action;
 
@@ -992,8 +1093,11 @@ mlx5dr_create_action_dest_flow_fw_table(struct mlx5_flow_table *ft,
 		return NULL;
 
 	action->dest_tbl.is_fw_tbl = 1;
-	action->dest_tbl.fw_tbl.ft = ft;
-	action->dest_tbl.fw_tbl.mdev = mdev;
+	action->dest_tbl.fw_tbl.type = ft->type;
+	action->dest_tbl.fw_tbl.id = ft->id;
+	action->dest_tbl.fw_tbl.dmn = dmn;
+
+	refcount_inc(&dmn->refcount);
 
 	return action;
 }
@@ -1213,58 +1317,85 @@ not_found:
 }
 
 static int
-dr_action_modify_sw_to_hw(struct mlx5dr_domain *dmn,
-			  __be64 *sw_action,
-			  __be64 *hw_action,
-			  const struct dr_action_modify_field_conv **ret_hw_info)
+dr_action_modify_sw_to_hw_add(struct mlx5dr_domain *dmn,
+			      __be64 *sw_action,
+			      __be64 *hw_action,
+			      const struct dr_action_modify_field_conv **ret_hw_info)
 {
 	const struct dr_action_modify_field_conv *hw_action_info;
-	u8 offset, length, max_length, action;
+	u8 max_length;
 	u16 sw_field;
-	u8 hw_opcode;
 	u32 data;
 
 	/* Get SW modify action data */
-	action = MLX5_GET(set_action_in, sw_action, action_type);
-	length = MLX5_GET(set_action_in, sw_action, length);
-	offset = MLX5_GET(set_action_in, sw_action, offset);
 	sw_field = MLX5_GET(set_action_in, sw_action, field);
 	data = MLX5_GET(set_action_in, sw_action, data);
 
 	/* Convert SW data to HW modify action format */
 	hw_action_info = dr_action_modify_get_hw_info(sw_field);
 	if (!hw_action_info) {
-		mlx5dr_dbg(dmn, "Modify action invalid field given\n");
+		mlx5dr_dbg(dmn, "Modify add action invalid field given\n");
 		return -EINVAL;
 	}
 
 	max_length = hw_action_info->end - hw_action_info->start + 1;
 
-	switch (action) {
-	case MLX5_ACTION_TYPE_SET:
-		hw_opcode = MLX5DR_ACTION_MDFY_HW_OP_SET;
-		/* PRM defines that length zero specific length of 32bits */
-		if (!length)
-			length = 32;
+	MLX5_SET(dr_action_hw_set, hw_action,
+		 opcode, MLX5DR_ACTION_MDFY_HW_OP_ADD);
 
-		if (length + offset > max_length) {
-			mlx5dr_dbg(dmn, "Modify action length + offset exceeds limit\n");
-			return -EINVAL;
-		}
-		break;
+	MLX5_SET(dr_action_hw_set, hw_action, destination_field_code,
+		 hw_action_info->hw_field);
 
-	case MLX5_ACTION_TYPE_ADD:
-		hw_opcode = MLX5DR_ACTION_MDFY_HW_OP_ADD;
-		offset = 0;
-		length = max_length;
-		break;
+	MLX5_SET(dr_action_hw_set, hw_action, destination_left_shifter,
+		 hw_action_info->start);
 
-	default:
-		mlx5dr_info(dmn, "Unsupported action_type for modify action\n");
-		return -EOPNOTSUPP;
+	/* PRM defines that length zero specific length of 32bits */
+	MLX5_SET(dr_action_hw_set, hw_action, destination_length,
+		 max_length == 32 ? 0 : max_length);
+
+	MLX5_SET(dr_action_hw_set, hw_action, inline_data, data);
+
+	*ret_hw_info = hw_action_info;
+
+	return 0;
+}
+
+static int
+dr_action_modify_sw_to_hw_set(struct mlx5dr_domain *dmn,
+			      __be64 *sw_action,
+			      __be64 *hw_action,
+			      const struct dr_action_modify_field_conv **ret_hw_info)
+{
+	const struct dr_action_modify_field_conv *hw_action_info;
+	u8 offset, length, max_length;
+	u16 sw_field;
+	u32 data;
+
+	/* Get SW modify action data */
+	length = MLX5_GET(set_action_in, sw_action, length);
+	offset = MLX5_GET(set_action_in, sw_action, offset);
+	sw_field = MLX5_GET(set_action_in, sw_action, field);
+	data = MLX5_GET(set_action_in, sw_action, data);
+
+	/* Convert SW data to HW modify action format */
+	hw_action_info = dr_action_modify_get_hw_info(sw_field);
+	if (!hw_action_info) {
+		mlx5dr_dbg(dmn, "Modify set action invalid field given\n");
+		return -EINVAL;
+	}
+
+	/* PRM defines that length zero specific length of 32bits */
+	length = length ? length : 32;
+
+	max_length = hw_action_info->end - hw_action_info->start + 1;
+
+	if (length + offset > max_length) {
+		mlx5dr_dbg(dmn, "Modify action length + offset exceeds limit\n");
+		return -EINVAL;
 	}
 
-	MLX5_SET(dr_action_hw_set, hw_action, opcode, hw_opcode);
+	MLX5_SET(dr_action_hw_set, hw_action,
+		 opcode, MLX5DR_ACTION_MDFY_HW_OP_SET);
 
 	MLX5_SET(dr_action_hw_set, hw_action, destination_field_code,
 		 hw_action_info->hw_field);
@@ -1283,48 +1414,236 @@ dr_action_modify_sw_to_hw(struct mlx5dr_domain *dmn,
 }
 
 static int
-dr_action_modify_check_field_limitation(struct mlx5dr_domain *dmn,
-					const __be64 *sw_action)
+dr_action_modify_sw_to_hw_copy(struct mlx5dr_domain *dmn,
+			       __be64 *sw_action,
+			       __be64 *hw_action,
+			       const struct dr_action_modify_field_conv **ret_dst_hw_info,
+			       const struct dr_action_modify_field_conv **ret_src_hw_info)
+{
+	u8 src_offset, dst_offset, src_max_length, dst_max_length, length;
+	const struct dr_action_modify_field_conv *hw_dst_action_info;
+	const struct dr_action_modify_field_conv *hw_src_action_info;
+	u16 src_field, dst_field;
+
+	/* Get SW modify action data */
+	src_field = MLX5_GET(copy_action_in, sw_action, src_field);
+	dst_field = MLX5_GET(copy_action_in, sw_action, dst_field);
+	src_offset = MLX5_GET(copy_action_in, sw_action, src_offset);
+	dst_offset = MLX5_GET(copy_action_in, sw_action, dst_offset);
+	length = MLX5_GET(copy_action_in, sw_action, length);
+
+	/* Convert SW data to HW modify action format */
+	hw_src_action_info = dr_action_modify_get_hw_info(src_field);
+	hw_dst_action_info = dr_action_modify_get_hw_info(dst_field);
+	if (!hw_src_action_info || !hw_dst_action_info) {
+		mlx5dr_dbg(dmn, "Modify copy action invalid field given\n");
+		return -EINVAL;
+	}
+
+	/* PRM defines that length zero specific length of 32bits */
+	length = length ? length : 32;
+
+	src_max_length = hw_src_action_info->end -
+			 hw_src_action_info->start + 1;
+	dst_max_length = hw_dst_action_info->end -
+			 hw_dst_action_info->start + 1;
+
+	if (length + src_offset > src_max_length ||
+	    length + dst_offset > dst_max_length) {
+		mlx5dr_dbg(dmn, "Modify action length + offset exceeds limit\n");
+		return -EINVAL;
+	}
+
+	MLX5_SET(dr_action_hw_copy, hw_action,
+		 opcode, MLX5DR_ACTION_MDFY_HW_OP_COPY);
+
+	MLX5_SET(dr_action_hw_copy, hw_action, destination_field_code,
+		 hw_dst_action_info->hw_field);
+
+	MLX5_SET(dr_action_hw_copy, hw_action, destination_left_shifter,
+		 hw_dst_action_info->start + dst_offset);
+
+	MLX5_SET(dr_action_hw_copy, hw_action, destination_length,
+		 length == 32 ? 0 : length);
+
+	MLX5_SET(dr_action_hw_copy, hw_action, source_field_code,
+		 hw_src_action_info->hw_field);
+
+	MLX5_SET(dr_action_hw_copy, hw_action, source_left_shifter,
+		 hw_src_action_info->start + dst_offset);
+
+	*ret_dst_hw_info = hw_dst_action_info;
+	*ret_src_hw_info = hw_src_action_info;
+
+	return 0;
+}
+
+static int
+dr_action_modify_sw_to_hw(struct mlx5dr_domain *dmn,
+			  __be64 *sw_action,
+			  __be64 *hw_action,
+			  const struct dr_action_modify_field_conv **ret_dst_hw_info,
+			  const struct dr_action_modify_field_conv **ret_src_hw_info)
 {
-	u16 sw_field;
 	u8 action;
+	int ret;
 
-	sw_field = MLX5_GET(set_action_in, sw_action, field);
+	*hw_action = 0;
+	*ret_src_hw_info = NULL;
+
+	/* Get SW modify action type */
 	action = MLX5_GET(set_action_in, sw_action, action_type);
 
-	/* Check if SW field is supported in current domain (RX/TX) */
-	if (action == MLX5_ACTION_TYPE_SET) {
-		if (sw_field == MLX5_ACTION_IN_FIELD_METADATA_REG_A) {
+	switch (action) {
+	case MLX5_ACTION_TYPE_SET:
+		ret = dr_action_modify_sw_to_hw_set(dmn, sw_action,
+						    hw_action,
+						    ret_dst_hw_info);
+		break;
+
+	case MLX5_ACTION_TYPE_ADD:
+		ret = dr_action_modify_sw_to_hw_add(dmn, sw_action,
+						    hw_action,
+						    ret_dst_hw_info);
+		break;
+
+	case MLX5_ACTION_TYPE_COPY:
+		ret = dr_action_modify_sw_to_hw_copy(dmn, sw_action,
+						     hw_action,
+						     ret_dst_hw_info,
+						     ret_src_hw_info);
+		break;
+
+	default:
+		mlx5dr_info(dmn, "Unsupported action_type for modify action\n");
+		ret = -EOPNOTSUPP;
+	}
+
+	return ret;
+}
+
+static int
+dr_action_modify_check_set_field_limitation(struct mlx5dr_action *action,
+					    const __be64 *sw_action)
+{
+	u16 sw_field = MLX5_GET(set_action_in, sw_action, field);
+	struct mlx5dr_domain *dmn = action->rewrite.dmn;
+
+	if (sw_field == MLX5_ACTION_IN_FIELD_METADATA_REG_A) {
+		action->rewrite.allow_rx = 0;
+		if (dmn->type != MLX5DR_DOMAIN_TYPE_NIC_TX) {
+			mlx5dr_dbg(dmn, "Unsupported field %d for RX/FDB set action\n",
+				   sw_field);
+			return -EINVAL;
+		}
+	} else if (sw_field == MLX5_ACTION_IN_FIELD_METADATA_REG_B) {
+		action->rewrite.allow_tx = 0;
+		if (dmn->type != MLX5DR_DOMAIN_TYPE_NIC_RX) {
+			mlx5dr_dbg(dmn, "Unsupported field %d for TX/FDB set action\n",
+				   sw_field);
+			return -EINVAL;
+		}
+	}
+
+	if (!action->rewrite.allow_rx && !action->rewrite.allow_tx) {
+		mlx5dr_dbg(dmn, "Modify SET actions not supported on both RX and TX\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+dr_action_modify_check_add_field_limitation(struct mlx5dr_action *action,
+					    const __be64 *sw_action)
+{
+	u16 sw_field = MLX5_GET(set_action_in, sw_action, field);
+	struct mlx5dr_domain *dmn = action->rewrite.dmn;
+
+	if (sw_field != MLX5_ACTION_IN_FIELD_OUT_IP_TTL &&
+	    sw_field != MLX5_ACTION_IN_FIELD_OUT_IPV6_HOPLIMIT &&
+	    sw_field != MLX5_ACTION_IN_FIELD_OUT_TCP_SEQ_NUM &&
+	    sw_field != MLX5_ACTION_IN_FIELD_OUT_TCP_ACK_NUM) {
+		mlx5dr_dbg(dmn, "Unsupported field %d for add action\n",
+			   sw_field);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+dr_action_modify_check_copy_field_limitation(struct mlx5dr_action *action,
+					     const __be64 *sw_action)
+{
+	struct mlx5dr_domain *dmn = action->rewrite.dmn;
+	u16 sw_fields[2];
+	int i;
+
+	sw_fields[0] = MLX5_GET(copy_action_in, sw_action, src_field);
+	sw_fields[1] = MLX5_GET(copy_action_in, sw_action, dst_field);
+
+	for (i = 0; i < 2; i++) {
+		if (sw_fields[i] == MLX5_ACTION_IN_FIELD_METADATA_REG_A) {
+			action->rewrite.allow_rx = 0;
 			if (dmn->type != MLX5DR_DOMAIN_TYPE_NIC_TX) {
 				mlx5dr_dbg(dmn, "Unsupported field %d for RX/FDB set action\n",
-					   sw_field);
+					   sw_fields[i]);
 				return -EINVAL;
 			}
-		}
-
-		if (sw_field == MLX5_ACTION_IN_FIELD_METADATA_REG_B) {
+		} else if (sw_fields[i] == MLX5_ACTION_IN_FIELD_METADATA_REG_B) {
+			action->rewrite.allow_tx = 0;
 			if (dmn->type != MLX5DR_DOMAIN_TYPE_NIC_RX) {
 				mlx5dr_dbg(dmn, "Unsupported field %d for TX/FDB set action\n",
-					   sw_field);
+					   sw_fields[i]);
 				return -EINVAL;
 			}
 		}
-	} else if (action == MLX5_ACTION_TYPE_ADD) {
-		if (sw_field != MLX5_ACTION_IN_FIELD_OUT_IP_TTL &&
-		    sw_field != MLX5_ACTION_IN_FIELD_OUT_IPV6_HOPLIMIT &&
-		    sw_field != MLX5_ACTION_IN_FIELD_OUT_TCP_SEQ_NUM &&
-		    sw_field != MLX5_ACTION_IN_FIELD_OUT_TCP_ACK_NUM) {
-			mlx5dr_dbg(dmn, "Unsupported field %d for add action\n", sw_field);
-			return -EINVAL;
-		}
-	} else {
-		mlx5dr_info(dmn, "Unsupported action %d modify action\n", action);
-		return -EOPNOTSUPP;
+	}
+
+	if (!action->rewrite.allow_rx && !action->rewrite.allow_tx) {
+		mlx5dr_dbg(dmn, "Modify copy actions not supported on both RX and TX\n");
+		return -EINVAL;
 	}
 
 	return 0;
 }
 
+static int
+dr_action_modify_check_field_limitation(struct mlx5dr_action *action,
+					const __be64 *sw_action)
+{
+	struct mlx5dr_domain *dmn = action->rewrite.dmn;
+	u8 action_type;
+	int ret;
+
+	action_type = MLX5_GET(set_action_in, sw_action, action_type);
+
+	switch (action_type) {
+	case MLX5_ACTION_TYPE_SET:
+		ret = dr_action_modify_check_set_field_limitation(action,
+								  sw_action);
+		break;
+
+	case MLX5_ACTION_TYPE_ADD:
+		ret = dr_action_modify_check_add_field_limitation(action,
+								  sw_action);
+		break;
+
+	case MLX5_ACTION_TYPE_COPY:
+		ret = dr_action_modify_check_copy_field_limitation(action,
+								   sw_action);
+		break;
+
+	default:
+		mlx5dr_info(dmn, "Unsupported action %d modify action\n",
+			    action_type);
+		ret = -EOPNOTSUPP;
+	}
+
+	return ret;
+}
+
 static bool
 dr_action_modify_check_is_ttl_modify(const u64 *sw_action)
 {
@@ -1333,7 +1652,7 @@ dr_action_modify_check_is_ttl_modify(const u64 *sw_action)
 	return sw_field == MLX5_ACTION_IN_FIELD_OUT_IP_TTL;
 }
 
-static int dr_actions_convert_modify_header(struct mlx5dr_domain *dmn,
+static int dr_actions_convert_modify_header(struct mlx5dr_action *action,
 					    u32 max_hw_actions,
 					    u32 num_sw_actions,
 					    __be64 sw_actions[],
@@ -1341,20 +1660,26 @@ static int dr_actions_convert_modify_header(struct mlx5dr_domain *dmn,
 					    u32 *num_hw_actions,
 					    bool *modify_ttl)
 {
-	const struct dr_action_modify_field_conv *hw_action_info;
+	const struct dr_action_modify_field_conv *hw_dst_action_info;
+	const struct dr_action_modify_field_conv *hw_src_action_info;
 	u16 hw_field = MLX5DR_ACTION_MDFY_HW_FLD_RESERVED;
 	u32 l3_type = MLX5DR_ACTION_MDFY_HW_HDR_L3_NONE;
 	u32 l4_type = MLX5DR_ACTION_MDFY_HW_HDR_L4_NONE;
+	struct mlx5dr_domain *dmn = action->rewrite.dmn;
 	int ret, i, hw_idx = 0;
 	__be64 *sw_action;
 	__be64 hw_action;
 
 	*modify_ttl = false;
 
+	action->rewrite.allow_rx = 1;
+	action->rewrite.allow_tx = 1;
+
 	for (i = 0; i < num_sw_actions; i++) {
 		sw_action = &sw_actions[i];
 
-		ret = dr_action_modify_check_field_limitation(dmn, sw_action);
+		ret = dr_action_modify_check_field_limitation(action,
+							      sw_action);
 		if (ret)
 			return ret;
 
@@ -1365,32 +1690,35 @@ static int dr_actions_convert_modify_header(struct mlx5dr_domain *dmn,
 		ret = dr_action_modify_sw_to_hw(dmn,
 						sw_action,
 						&hw_action,
-						&hw_action_info);
+						&hw_dst_action_info,
+						&hw_src_action_info);
 		if (ret)
 			return ret;
 
 		/* Due to a HW limitation we cannot modify 2 different L3 types */
-		if (l3_type && hw_action_info->l3_type &&
-		    hw_action_info->l3_type != l3_type) {
+		if (l3_type && hw_dst_action_info->l3_type &&
+		    hw_dst_action_info->l3_type != l3_type) {
 			mlx5dr_dbg(dmn, "Action list can't support two different L3 types\n");
 			return -EINVAL;
 		}
-		if (hw_action_info->l3_type)
-			l3_type = hw_action_info->l3_type;
+		if (hw_dst_action_info->l3_type)
+			l3_type = hw_dst_action_info->l3_type;
 
 		/* Due to a HW limitation we cannot modify two different L4 types */
-		if (l4_type && hw_action_info->l4_type &&
-		    hw_action_info->l4_type != l4_type) {
+		if (l4_type && hw_dst_action_info->l4_type &&
+		    hw_dst_action_info->l4_type != l4_type) {
 			mlx5dr_dbg(dmn, "Action list can't support two different L4 types\n");
 			return -EINVAL;
 		}
-		if (hw_action_info->l4_type)
-			l4_type = hw_action_info->l4_type;
+		if (hw_dst_action_info->l4_type)
+			l4_type = hw_dst_action_info->l4_type;
 
 		/* HW reads and executes two actions at once this means we
 		 * need to create a gap if two actions access the same field
 		 */
-		if ((hw_idx % 2) && hw_field == hw_action_info->hw_field) {
+		if ((hw_idx % 2) && (hw_field == hw_dst_action_info->hw_field ||
+				     (hw_src_action_info &&
+				      hw_field == hw_src_action_info->hw_field))) {
 			/* Check if after gap insertion the total number of HW
 			 * modify actions doesn't exceeds the limit
 			 */
@@ -1400,7 +1728,7 @@ static int dr_actions_convert_modify_header(struct mlx5dr_domain *dmn,
 				return -EINVAL;
 			}
 		}
-		hw_field = hw_action_info->hw_field;
+		hw_field = hw_dst_action_info->hw_field;
 
 		hw_actions[hw_idx] = hw_action;
 		hw_idx++;
@@ -1443,7 +1771,7 @@ static int dr_action_create_modify_action(struct mlx5dr_domain *dmn,
 		goto free_chunk;
 	}
 
-	ret = dr_actions_convert_modify_header(dmn,
+	ret = dr_actions_convert_modify_header(action,
 					       max_hw_actions,
 					       num_sw_actions,
 					       actions,
@@ -1559,8 +1887,26 @@ int mlx5dr_action_destroy(struct mlx5dr_action *action)
 
 	switch (action->action_type) {
 	case DR_ACTION_TYP_FT:
-		if (!action->dest_tbl.is_fw_tbl)
+		if (action->dest_tbl.is_fw_tbl)
+			refcount_dec(&action->dest_tbl.fw_tbl.dmn->refcount);
+		else
 			refcount_dec(&action->dest_tbl.tbl->refcount);
+
+		if (action->dest_tbl.is_fw_tbl &&
+		    action->dest_tbl.fw_tbl.num_of_ref_actions) {
+			struct mlx5dr_action **ref_actions;
+			int i;
+
+			ref_actions = action->dest_tbl.fw_tbl.ref_actions;
+			for (i = 0; i < action->dest_tbl.fw_tbl.num_of_ref_actions; i++)
+				refcount_dec(&ref_actions[i]->refcount);
+
+			kfree(ref_actions);
+
+			mlx5dr_fw_destroy_md_tbl(action->dest_tbl.fw_tbl.dmn,
+						 action->dest_tbl.fw_tbl.id,
+						 action->dest_tbl.fw_tbl.group_id);
+		}
 		break;
 	case DR_ACTION_TYP_TNL_L2_TO_L2:
 		refcount_dec(&action->reformat.dmn->refcount);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
index 41662c4e2664..461b39376daf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c
@@ -320,12 +320,7 @@ int mlx5dr_cmd_destroy_flow_group(struct mlx5_core_dev *mdev,
 }
 
 int mlx5dr_cmd_create_flow_table(struct mlx5_core_dev *mdev,
-				 u32 table_type,
-				 u64 icm_addr_rx,
-				 u64 icm_addr_tx,
-				 u8 level,
-				 bool sw_owner,
-				 bool term_tbl,
+				 struct mlx5dr_cmd_create_flow_table_attr *attr,
 				 u64 *fdb_rx_icm_addr,
 				 u32 *table_id)
 {
@@ -335,37 +330,43 @@ int mlx5dr_cmd_create_flow_table(struct mlx5_core_dev *mdev,
 	int err;
 
 	MLX5_SET(create_flow_table_in, in, opcode, MLX5_CMD_OP_CREATE_FLOW_TABLE);
-	MLX5_SET(create_flow_table_in, in, table_type, table_type);
+	MLX5_SET(create_flow_table_in, in, table_type, attr->table_type);
 
 	ft_mdev = MLX5_ADDR_OF(create_flow_table_in, in, flow_table_context);
-	MLX5_SET(flow_table_context, ft_mdev, termination_table, term_tbl);
-	MLX5_SET(flow_table_context, ft_mdev, sw_owner, sw_owner);
-	MLX5_SET(flow_table_context, ft_mdev, level, level);
+	MLX5_SET(flow_table_context, ft_mdev, termination_table, attr->term_tbl);
+	MLX5_SET(flow_table_context, ft_mdev, sw_owner, attr->sw_owner);
+	MLX5_SET(flow_table_context, ft_mdev, level, attr->level);
 
-	if (sw_owner) {
+	if (attr->sw_owner) {
 		/* icm_addr_0 used for FDB RX / NIC TX / NIC_RX
 		 * icm_addr_1 used for FDB TX
 		 */
-		if (table_type == MLX5_FLOW_TABLE_TYPE_NIC_RX) {
+		if (attr->table_type == MLX5_FLOW_TABLE_TYPE_NIC_RX) {
 			MLX5_SET64(flow_table_context, ft_mdev,
-				   sw_owner_icm_root_0, icm_addr_rx);
-		} else if (table_type == MLX5_FLOW_TABLE_TYPE_NIC_TX) {
+				   sw_owner_icm_root_0, attr->icm_addr_rx);
+		} else if (attr->table_type == MLX5_FLOW_TABLE_TYPE_NIC_TX) {
 			MLX5_SET64(flow_table_context, ft_mdev,
-				   sw_owner_icm_root_0, icm_addr_tx);
-		} else if (table_type == MLX5_FLOW_TABLE_TYPE_FDB) {
+				   sw_owner_icm_root_0, attr->icm_addr_tx);
+		} else if (attr->table_type == MLX5_FLOW_TABLE_TYPE_FDB) {
 			MLX5_SET64(flow_table_context, ft_mdev,
-				   sw_owner_icm_root_0, icm_addr_rx);
+				   sw_owner_icm_root_0, attr->icm_addr_rx);
 			MLX5_SET64(flow_table_context, ft_mdev,
-				   sw_owner_icm_root_1, icm_addr_tx);
+				   sw_owner_icm_root_1, attr->icm_addr_tx);
 		}
 	}
 
+	MLX5_SET(create_flow_table_in, in, flow_table_context.decap_en,
+		 attr->decap_en);
+	MLX5_SET(create_flow_table_in, in, flow_table_context.reformat_en,
+		 attr->reformat_en);
+
 	err = mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
 	if (err)
 		return err;
 
 	*table_id = MLX5_GET(create_flow_table_out, out, table_id);
-	if (!sw_owner && table_type == MLX5_FLOW_TABLE_TYPE_FDB)
+	if (!attr->sw_owner && attr->table_type == MLX5_FLOW_TABLE_TYPE_FDB &&
+	    fdb_rx_icm_addr)
 		*fdb_rx_icm_addr =
 		(u64)MLX5_GET(create_flow_table_out, out, icm_address_31_0) |
 		(u64)MLX5_GET(create_flow_table_out, out, icm_address_39_32) << 32 |
@@ -478,3 +479,208 @@ int mlx5dr_cmd_query_gid(struct mlx5_core_dev *mdev, u8 vhca_port_num,
 
 	return 0;
 }
+
+static int mlx5dr_cmd_set_extended_dest(struct mlx5_core_dev *dev,
+					struct mlx5dr_cmd_fte_info *fte,
+					bool *extended_dest)
+{
+	int fw_log_max_fdb_encap_uplink = MLX5_CAP_ESW(dev, log_max_fdb_encap_uplink);
+	int num_fwd_destinations = 0;
+	int num_encap = 0;
+	int i;
+
+	*extended_dest = false;
+	if (!(fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST))
+		return 0;
+	for (i = 0; i < fte->dests_size; i++) {
+		if (fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+			continue;
+		if (fte->dest_arr[i].type == MLX5_FLOW_DESTINATION_TYPE_VPORT &&
+		    fte->dest_arr[i].vport.flags & MLX5_FLOW_DEST_VPORT_REFORMAT_ID)
+			num_encap++;
+		num_fwd_destinations++;
+	}
+
+	if (num_fwd_destinations > 1 && num_encap > 0)
+		*extended_dest = true;
+
+	if (*extended_dest && !fw_log_max_fdb_encap_uplink) {
+		mlx5_core_warn(dev, "FW does not support extended destination");
+		return -EOPNOTSUPP;
+	}
+	if (num_encap > (1 << fw_log_max_fdb_encap_uplink)) {
+		mlx5_core_warn(dev, "FW does not support more than %d encaps",
+			       1 << fw_log_max_fdb_encap_uplink);
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
+		       int opmod, int modify_mask,
+		       struct mlx5dr_cmd_ft_info *ft,
+		       u32 group_id,
+		       struct mlx5dr_cmd_fte_info *fte)
+{
+	u32 out[MLX5_ST_SZ_DW(set_fte_out)] = {};
+	void *in_flow_context, *vlan;
+	bool extended_dest = false;
+	void *in_match_value;
+	unsigned int inlen;
+	int dst_cnt_size;
+	void *in_dests;
+	u32 *in;
+	int err;
+	int i;
+
+	if (mlx5dr_cmd_set_extended_dest(dev, fte, &extended_dest))
+		return -EOPNOTSUPP;
+
+	if (!extended_dest)
+		dst_cnt_size = MLX5_ST_SZ_BYTES(dest_format_struct);
+	else
+		dst_cnt_size = MLX5_ST_SZ_BYTES(extended_dest_format);
+
+	inlen = MLX5_ST_SZ_BYTES(set_fte_in) + fte->dests_size * dst_cnt_size;
+	in = kvzalloc(inlen, GFP_KERNEL);
+	if (!in)
+		return -ENOMEM;
+
+	MLX5_SET(set_fte_in, in, opcode, MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY);
+	MLX5_SET(set_fte_in, in, op_mod, opmod);
+	MLX5_SET(set_fte_in, in, modify_enable_mask, modify_mask);
+	MLX5_SET(set_fte_in, in, table_type, ft->type);
+	MLX5_SET(set_fte_in, in, table_id, ft->id);
+	MLX5_SET(set_fte_in, in, flow_index, fte->index);
+	if (ft->vport) {
+		MLX5_SET(set_fte_in, in, vport_number, ft->vport);
+		MLX5_SET(set_fte_in, in, other_vport, 1);
+	}
+
+	in_flow_context = MLX5_ADDR_OF(set_fte_in, in, flow_context);
+	MLX5_SET(flow_context, in_flow_context, group_id, group_id);
+
+	MLX5_SET(flow_context, in_flow_context, flow_tag,
+		 fte->flow_context.flow_tag);
+	MLX5_SET(flow_context, in_flow_context, flow_source,
+		 fte->flow_context.flow_source);
+
+	MLX5_SET(flow_context, in_flow_context, extended_destination,
+		 extended_dest);
+	if (extended_dest) {
+		u32 action;
+
+		action = fte->action.action &
+			~MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
+		MLX5_SET(flow_context, in_flow_context, action, action);
+	} else {
+		MLX5_SET(flow_context, in_flow_context, action,
+			 fte->action.action);
+		if (fte->action.pkt_reformat)
+			MLX5_SET(flow_context, in_flow_context, packet_reformat_id,
+				 fte->action.pkt_reformat->id);
+	}
+	if (fte->action.modify_hdr)
+		MLX5_SET(flow_context, in_flow_context, modify_header_id,
+			 fte->action.modify_hdr->id);
+
+	vlan = MLX5_ADDR_OF(flow_context, in_flow_context, push_vlan);
+
+	MLX5_SET(vlan, vlan, ethtype, fte->action.vlan[0].ethtype);
+	MLX5_SET(vlan, vlan, vid, fte->action.vlan[0].vid);
+	MLX5_SET(vlan, vlan, prio, fte->action.vlan[0].prio);
+
+	vlan = MLX5_ADDR_OF(flow_context, in_flow_context, push_vlan_2);
+
+	MLX5_SET(vlan, vlan, ethtype, fte->action.vlan[1].ethtype);
+	MLX5_SET(vlan, vlan, vid, fte->action.vlan[1].vid);
+	MLX5_SET(vlan, vlan, prio, fte->action.vlan[1].prio);
+
+	in_match_value = MLX5_ADDR_OF(flow_context, in_flow_context,
+				      match_value);
+	memcpy(in_match_value, fte->val, sizeof(u32) * MLX5_ST_SZ_DW_MATCH_PARAM);
+
+	in_dests = MLX5_ADDR_OF(flow_context, in_flow_context, destination);
+	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		int list_size = 0;
+
+		for (i = 0; i < fte->dests_size; i++) {
+			unsigned int id, type = fte->dest_arr[i].type;
+
+			if (type == MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+				continue;
+
+			switch (type) {
+			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE_NUM:
+				id = fte->dest_arr[i].ft_num;
+				type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+				break;
+			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
+				id = fte->dest_arr[i].ft_id;
+				break;
+			case MLX5_FLOW_DESTINATION_TYPE_VPORT:
+				id = fte->dest_arr[i].vport.num;
+				MLX5_SET(dest_format_struct, in_dests,
+					 destination_eswitch_owner_vhca_id_valid,
+					 !!(fte->dest_arr[i].vport.flags &
+					    MLX5_FLOW_DEST_VPORT_VHCA_ID));
+				MLX5_SET(dest_format_struct, in_dests,
+					 destination_eswitch_owner_vhca_id,
+					 fte->dest_arr[i].vport.vhca_id);
+				if (extended_dest && (fte->dest_arr[i].vport.flags &
+						    MLX5_FLOW_DEST_VPORT_REFORMAT_ID)) {
+					MLX5_SET(dest_format_struct, in_dests,
+						 packet_reformat,
+						 !!(fte->dest_arr[i].vport.flags &
+						    MLX5_FLOW_DEST_VPORT_REFORMAT_ID));
+					MLX5_SET(extended_dest_format, in_dests,
+						 packet_reformat_id,
+						 fte->dest_arr[i].vport.reformat_id);
+				}
+				break;
+			default:
+				id = fte->dest_arr[i].tir_num;
+			}
+
+			MLX5_SET(dest_format_struct, in_dests, destination_type,
+				 type);
+			MLX5_SET(dest_format_struct, in_dests, destination_id, id);
+			in_dests += dst_cnt_size;
+			list_size++;
+		}
+
+		MLX5_SET(flow_context, in_flow_context, destination_list_size,
+			 list_size);
+	}
+
+	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_COUNT) {
+		int max_list_size = BIT(MLX5_CAP_FLOWTABLE_TYPE(dev,
+					log_max_flow_counter,
+					ft->type));
+		int list_size = 0;
+
+		for (i = 0; i < fte->dests_size; i++) {
+			if (fte->dest_arr[i].type !=
+			    MLX5_FLOW_DESTINATION_TYPE_COUNTER)
+				continue;
+
+			MLX5_SET(flow_counter_list, in_dests, flow_counter_id,
+				 fte->dest_arr[i].counter_id);
+			in_dests += dst_cnt_size;
+			list_size++;
+		}
+		if (list_size > max_list_size) {
+			err = -EINVAL;
+			goto err_out;
+		}
+
+		MLX5_SET(flow_context, in_flow_context, flow_counter_list_size,
+			 list_size);
+	}
+
+	err = mlx5_cmd_exec(dev, in, inlen, out, sizeof(out));
+err_out:
+	kvfree(in);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c
index 60ef6e6171e3..1fbcd012bb85 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_fw.c
@@ -7,6 +7,7 @@
 struct mlx5dr_fw_recalc_cs_ft *
 mlx5dr_fw_create_recalc_cs_ft(struct mlx5dr_domain *dmn, u32 vport_num)
 {
+	struct mlx5dr_cmd_create_flow_table_attr ft_attr = {};
 	struct mlx5dr_fw_recalc_cs_ft *recalc_cs_ft;
 	u32 table_id, group_id, modify_hdr_id;
 	u64 rx_icm_addr, modify_ttl_action;
@@ -16,9 +17,14 @@ mlx5dr_fw_create_recalc_cs_ft(struct mlx5dr_domain *dmn, u32 vport_num)
 	if (!recalc_cs_ft)
 		return NULL;
 
-	ret = mlx5dr_cmd_create_flow_table(dmn->mdev, MLX5_FLOW_TABLE_TYPE_FDB,
-					   0, 0, dmn->info.caps.max_ft_level - 1,
-					   false, true, &rx_icm_addr, &table_id);
+	ft_attr.table_type = MLX5_FLOW_TABLE_TYPE_FDB;
+	ft_attr.level = dmn->info.caps.max_ft_level - 1;
+	ft_attr.term_tbl = true;
+
+	ret = mlx5dr_cmd_create_flow_table(dmn->mdev,
+					   &ft_attr,
+					   &rx_icm_addr,
+					   &table_id);
 	if (ret) {
 		mlx5dr_err(dmn, "Failed creating TTL W/A FW flow table %d\n", ret);
 		goto free_ttl_tbl;
@@ -91,3 +97,70 @@ void mlx5dr_fw_destroy_recalc_cs_ft(struct mlx5dr_domain *dmn,
 
 	kfree(recalc_cs_ft);
 }
+
+int mlx5dr_fw_create_md_tbl(struct mlx5dr_domain *dmn,
+			    struct mlx5dr_cmd_flow_destination_hw_info *dest,
+			    int num_dest,
+			    bool reformat_req,
+			    u32 *tbl_id,
+			    u32 *group_id)
+{
+	struct mlx5dr_cmd_create_flow_table_attr ft_attr = {};
+	struct mlx5dr_cmd_fte_info fte_info = {};
+	u32 val[MLX5_ST_SZ_DW_MATCH_PARAM] = {};
+	struct mlx5dr_cmd_ft_info ft_info = {};
+	int ret;
+
+	ft_attr.table_type = MLX5_FLOW_TABLE_TYPE_FDB;
+	ft_attr.level = dmn->info.caps.max_ft_level - 2;
+	ft_attr.reformat_en = reformat_req;
+	ft_attr.decap_en = reformat_req;
+
+	ret = mlx5dr_cmd_create_flow_table(dmn->mdev, &ft_attr, NULL, tbl_id);
+	if (ret) {
+		mlx5dr_err(dmn, "Failed creating multi dest FW flow table %d\n", ret);
+		return ret;
+	}
+
+	ret = mlx5dr_cmd_create_empty_flow_group(dmn->mdev,
+						 MLX5_FLOW_TABLE_TYPE_FDB,
+						 *tbl_id, group_id);
+	if (ret) {
+		mlx5dr_err(dmn, "Failed creating multi dest FW flow group %d\n", ret);
+		goto free_flow_table;
+	}
+
+	ft_info.id = *tbl_id;
+	ft_info.type = FS_FT_FDB;
+	fte_info.action.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+	fte_info.dests_size = num_dest;
+	fte_info.val = val;
+	fte_info.dest_arr = dest;
+
+	ret = mlx5dr_cmd_set_fte(dmn->mdev, 0, 0, &ft_info, *group_id, &fte_info);
+	if (ret) {
+		mlx5dr_err(dmn, "Failed setting fte into table %d\n", ret);
+		goto free_flow_group;
+	}
+
+	return 0;
+
+free_flow_group:
+	mlx5dr_cmd_destroy_flow_group(dmn->mdev, MLX5_FLOW_TABLE_TYPE_FDB,
+				      *tbl_id, *group_id);
+free_flow_table:
+	mlx5dr_cmd_destroy_flow_table(dmn->mdev, *tbl_id,
+				      MLX5_FLOW_TABLE_TYPE_FDB);
+	return ret;
+}
+
+void mlx5dr_fw_destroy_md_tbl(struct mlx5dr_domain *dmn,
+			      u32 tbl_id, u32 group_id)
+{
+	mlx5dr_cmd_del_flow_table_entry(dmn->mdev, FS_FT_FDB, tbl_id);
+	mlx5dr_cmd_destroy_flow_group(dmn->mdev,
+				      MLX5_FLOW_TABLE_TYPE_FDB,
+				      tbl_id, group_id);
+	mlx5dr_cmd_destroy_flow_table(dmn->mdev, tbl_id,
+				      MLX5_FLOW_TABLE_TYPE_FDB);
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_table.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_table.c
index e178d8d3dbc9..14ce2d7dbb66 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_table.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_table.c
@@ -211,6 +211,9 @@ static int dr_table_destroy_sw_owned_tbl(struct mlx5dr_table *tbl)
 
 static int dr_table_create_sw_owned_tbl(struct mlx5dr_table *tbl)
 {
+	bool en_encap = !!(tbl->flags & MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT);
+	bool en_decap = !!(tbl->flags & MLX5_FLOW_TABLE_TUNNEL_EN_DECAP);
+	struct mlx5dr_cmd_create_flow_table_attr ft_attr = {};
 	u64 icm_addr_rx = 0;
 	u64 icm_addr_tx = 0;
 	int ret;
@@ -221,18 +224,21 @@ static int dr_table_create_sw_owned_tbl(struct mlx5dr_table *tbl)
 	if (tbl->tx.s_anchor)
 		icm_addr_tx = tbl->tx.s_anchor->chunk->icm_addr;
 
-	ret = mlx5dr_cmd_create_flow_table(tbl->dmn->mdev,
-					   tbl->table_type,
-					   icm_addr_rx,
-					   icm_addr_tx,
-					   tbl->dmn->info.caps.max_ft_level - 1,
-					   true, false, NULL,
-					   &tbl->table_id);
+	ft_attr.table_type = tbl->table_type;
+	ft_attr.icm_addr_rx = icm_addr_rx;
+	ft_attr.icm_addr_tx = icm_addr_tx;
+	ft_attr.level = tbl->dmn->info.caps.max_ft_level - 1;
+	ft_attr.sw_owner = true;
+	ft_attr.decap_en = en_decap;
+	ft_attr.reformat_en = en_encap;
+
+	ret = mlx5dr_cmd_create_flow_table(tbl->dmn->mdev, &ft_attr,
+					   NULL, &tbl->table_id);
 
 	return ret;
 }
 
-struct mlx5dr_table *mlx5dr_table_create(struct mlx5dr_domain *dmn, u32 level)
+struct mlx5dr_table *mlx5dr_table_create(struct mlx5dr_domain *dmn, u32 level, u32 flags)
 {
 	struct mlx5dr_table *tbl;
 	int ret;
@@ -245,6 +251,7 @@ struct mlx5dr_table *mlx5dr_table_create(struct mlx5dr_domain *dmn, u32 level)
 
 	tbl->dmn = dmn;
 	tbl->level = level;
+	tbl->flags = flags;
 	refcount_set(&tbl->refcount, 1);
 
 	ret = dr_table_init(tbl);
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 3fdf4a5eb031..dffe35145d19 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h
@@ -681,6 +681,7 @@ struct mlx5dr_table {
 	u32 level;
 	u32 table_type;
 	u32 table_id;
+	u32 flags;
 	struct list_head matcher_list;
 	struct mlx5dr_action *miss_action;
 	refcount_t refcount;
@@ -744,10 +745,14 @@ struct mlx5dr_action {
 			union {
 				struct mlx5dr_table *tbl;
 				struct {
-					struct mlx5_flow_table *ft;
+					struct mlx5dr_domain *dmn;
+					u32 id;
+					u32 group_id;
+					enum fs_flow_table_type type;
 					u64 rx_icm_addr;
 					u64 tx_icm_addr;
-					struct mlx5_core_dev *mdev;
+					struct mlx5dr_action **ref_actions;
+					u32 num_of_ref_actions;
 				} fw_tbl;
 			};
 		} dest_tbl;
@@ -869,6 +874,17 @@ struct mlx5dr_cmd_query_flow_table_details {
 	u64 sw_owner_icm_root_0;
 };
 
+struct mlx5dr_cmd_create_flow_table_attr {
+	u32 table_type;
+	u64 icm_addr_rx;
+	u64 icm_addr_tx;
+	u8 level;
+	bool sw_owner;
+	bool term_tbl;
+	bool decap_en;
+	bool reformat_en;
+};
+
 /* internal API functions */
 int mlx5dr_cmd_query_device(struct mlx5_core_dev *mdev,
 			    struct mlx5dr_cmd_caps *caps);
@@ -906,12 +922,7 @@ int mlx5dr_cmd_destroy_flow_group(struct mlx5_core_dev *mdev,
 				  u32 table_id,
 				  u32 group_id);
 int mlx5dr_cmd_create_flow_table(struct mlx5_core_dev *mdev,
-				 u32 table_type,
-				 u64 icm_addr_rx,
-				 u64 icm_addr_tx,
-				 u8 level,
-				 bool sw_owner,
-				 bool term_tbl,
+				 struct mlx5dr_cmd_create_flow_table_attr *attr,
 				 u64 *fdb_rx_icm_addr,
 				 u32 *table_id);
 int mlx5dr_cmd_destroy_flow_table(struct mlx5_core_dev *mdev,
@@ -1053,6 +1064,43 @@ int mlx5dr_send_postsend_formatted_htbl(struct mlx5dr_domain *dmn,
 int mlx5dr_send_postsend_action(struct mlx5dr_domain *dmn,
 				struct mlx5dr_action *action);
 
+struct mlx5dr_cmd_ft_info {
+	u32 id;
+	u16 vport;
+	enum fs_flow_table_type type;
+};
+
+struct mlx5dr_cmd_flow_destination_hw_info {
+	enum mlx5_flow_destination_type type;
+	union {
+		u32 tir_num;
+		u32 ft_num;
+		u32 ft_id;
+		u32 counter_id;
+		struct {
+			u16 num;
+			u16 vhca_id;
+			u32 reformat_id;
+			u8 flags;
+		} vport;
+	};
+};
+
+struct mlx5dr_cmd_fte_info {
+	u32 dests_size;
+	u32 index;
+	struct mlx5_flow_context flow_context;
+	u32 *val;
+	struct mlx5_flow_act action;
+	struct mlx5dr_cmd_flow_destination_hw_info *dest_arr;
+};
+
+int mlx5dr_cmd_set_fte(struct mlx5_core_dev *dev,
+		       int opmod, int modify_mask,
+		       struct mlx5dr_cmd_ft_info *ft,
+		       u32 group_id,
+		       struct mlx5dr_cmd_fte_info *fte);
+
 struct mlx5dr_fw_recalc_cs_ft {
 	u64 rx_icm_addr;
 	u32 table_id;
@@ -1067,4 +1115,12 @@ void mlx5dr_fw_destroy_recalc_cs_ft(struct mlx5dr_domain *dmn,
 int mlx5dr_domain_cache_get_recalc_cs_ft_addr(struct mlx5dr_domain *dmn,
 					      u32 vport_num,
 					      u64 *rx_icm_addr);
+int mlx5dr_fw_create_md_tbl(struct mlx5dr_domain *dmn,
+			    struct mlx5dr_cmd_flow_destination_hw_info *dest,
+			    int num_dest,
+			    bool reformat_req,
+			    u32 *tbl_id,
+			    u32 *group_id);
+void mlx5dr_fw_destroy_md_tbl(struct mlx5dr_domain *dmn, u32 tbl_id,
+			      u32 group_id);
 #endif  /* _DR_TYPES_H_ */
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 1e32e2443f73..3abfc8125926 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/fs_dr.c
@@ -74,7 +74,7 @@ static int mlx5_cmd_dr_create_flow_table(struct mlx5_flow_root_namespace *ns,
 								    next_ft);
 
 	tbl = mlx5dr_table_create(ns->fs_dr_domain.dr_domain,
-				  ft->level);
+				  ft->level, ft->flags);
 	if (!tbl) {
 		mlx5_core_err(ns->dev, "Failed creating dr flow_table\n");
 		return -EINVAL;
@@ -184,13 +184,13 @@ static struct mlx5dr_action *create_vport_action(struct mlx5dr_domain *domain,
 					       dest_attr->vport.vhca_id);
 }
 
-static struct mlx5dr_action *create_ft_action(struct mlx5_core_dev *dev,
+static struct mlx5dr_action *create_ft_action(struct mlx5dr_domain *domain,
 					      struct mlx5_flow_rule *dst)
 {
 	struct mlx5_flow_table *dest_ft = dst->dest_attr.ft;
 
 	if (mlx5_dr_is_fw_table(dest_ft->flags))
-		return mlx5dr_create_action_dest_flow_fw_table(dest_ft, dev);
+		return mlx5dr_action_create_dest_flow_fw_table(domain, dest_ft);
 	return mlx5dr_action_create_dest_table(dest_ft->fs_dr_table.dr_table);
 }
 
@@ -206,6 +206,12 @@ static struct mlx5dr_action *create_action_push_vlan(struct mlx5dr_domain *domai
 	return mlx5dr_action_create_push_vlan(domain, htonl(vlan_hdr));
 }
 
+static bool contain_vport_reformat_action(struct mlx5_flow_rule *dst)
+{
+	return dst->dest_attr.type == MLX5_FLOW_DESTINATION_TYPE_VPORT &&
+		dst->dest_attr.vport.flags & MLX5_FLOW_DEST_VPORT_REFORMAT_ID;
+}
+
 #define MLX5_FLOW_CONTEXT_ACTION_MAX  20
 static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 				  struct mlx5_flow_table *ft,
@@ -213,7 +219,7 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 				  struct fs_fte *fte)
 {
 	struct mlx5dr_domain *domain = ns->fs_dr_domain.dr_domain;
-	struct mlx5dr_action *term_action = NULL;
+	struct mlx5dr_action_dest *term_actions;
 	struct mlx5dr_match_parameters params;
 	struct mlx5_core_dev *dev = ns->dev;
 	struct mlx5dr_action **fs_dr_actions;
@@ -223,6 +229,7 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 	struct mlx5dr_rule *rule;
 	struct mlx5_flow_rule *dst;
 	int fs_dr_num_actions = 0;
+	int num_term_actions = 0;
 	int num_actions = 0;
 	size_t match_sz;
 	int err = 0;
@@ -233,18 +240,38 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 
 	actions = kcalloc(MLX5_FLOW_CONTEXT_ACTION_MAX, sizeof(*actions),
 			  GFP_KERNEL);
-	if (!actions)
-		return -ENOMEM;
+	if (!actions) {
+		err = -ENOMEM;
+		goto out_err;
+	}
 
 	fs_dr_actions = kcalloc(MLX5_FLOW_CONTEXT_ACTION_MAX,
 				sizeof(*fs_dr_actions), GFP_KERNEL);
 	if (!fs_dr_actions) {
-		kfree(actions);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto free_actions_alloc;
+	}
+
+	term_actions = kcalloc(MLX5_FLOW_CONTEXT_ACTION_MAX,
+			       sizeof(*term_actions), GFP_KERNEL);
+	if (!term_actions) {
+		err = -ENOMEM;
+		goto free_fs_dr_actions_alloc;
 	}
 
 	match_sz = sizeof(fte->val);
 
+	/* Drop reformat action bit if destination vport set with reformat */
+	if (fte->action.action & MLX5_FLOW_CONTEXT_ACTION_FWD_DEST) {
+		list_for_each_entry(dst, &fte->node.children, node.list) {
+			if (!contain_vport_reformat_action(dst))
+				continue;
+
+			fte->action.action &= ~MLX5_FLOW_CONTEXT_ACTION_PACKET_REFORMAT;
+			break;
+		}
+	}
+
 	/* The order of the actions are must to be keep, only the following
 	 * order is supported by SW steering:
 	 * TX: push vlan -> modify header -> encap
@@ -335,7 +362,7 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 			goto free_actions;
 		}
 		fs_dr_actions[fs_dr_num_actions++] = tmp_action;
-		term_action = tmp_action;
+		term_actions[num_term_actions++].dest = tmp_action;
 	}
 
 	if (fte->flow_context.flow_tag) {
@@ -353,7 +380,8 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 		list_for_each_entry(dst, &fte->node.children, node.list) {
 			enum mlx5_flow_destination_type type = dst->dest_attr.type;
 
-			if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX) {
+			if (num_actions == MLX5_FLOW_CONTEXT_ACTION_MAX ||
+			    num_term_actions >= MLX5_FLOW_CONTEXT_ACTION_MAX) {
 				err = -ENOSPC;
 				goto free_actions;
 			}
@@ -363,13 +391,13 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 
 			switch (type) {
 			case MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE:
-				tmp_action = create_ft_action(dev, dst);
+				tmp_action = create_ft_action(domain, dst);
 				if (!tmp_action) {
 					err = -ENOMEM;
 					goto free_actions;
 				}
 				fs_dr_actions[fs_dr_num_actions++] = tmp_action;
-				term_action = tmp_action;
+				term_actions[num_term_actions++].dest = tmp_action;
 				break;
 			case MLX5_FLOW_DESTINATION_TYPE_VPORT:
 				tmp_action = create_vport_action(domain, dst);
@@ -378,7 +406,14 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 					goto free_actions;
 				}
 				fs_dr_actions[fs_dr_num_actions++] = tmp_action;
-				term_action = tmp_action;
+				term_actions[num_term_actions].dest = tmp_action;
+
+				if (dst->dest_attr.vport.flags &
+				    MLX5_FLOW_DEST_VPORT_REFORMAT_ID)
+					term_actions[num_term_actions].reformat =
+						dst->dest_attr.vport.pkt_reformat->action.dr_action;
+
+				num_term_actions++;
 				break;
 			default:
 				err = -EOPNOTSUPP;
@@ -415,9 +450,22 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 
 	params.match_sz = match_sz;
 	params.match_buf = (u64 *)fte->val;
-
-	if (term_action)
-		actions[num_actions++] = term_action;
+	if (num_term_actions == 1) {
+		if (term_actions->reformat)
+			actions[num_actions++] = term_actions->reformat;
+
+		actions[num_actions++] = term_actions->dest;
+	} else if (num_term_actions > 1) {
+		tmp_action = mlx5dr_action_create_mult_dest_tbl(domain,
+								term_actions,
+								num_term_actions);
+		if (!tmp_action) {
+			err = -EOPNOTSUPP;
+			goto free_actions;
+		}
+		fs_dr_actions[fs_dr_num_actions++] = tmp_action;
+		actions[num_actions++] = tmp_action;
+	}
 
 	rule = mlx5dr_rule_create(group->fs_dr_matcher.dr_matcher,
 				  &params,
@@ -428,7 +476,9 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 		goto free_actions;
 	}
 
+	kfree(term_actions);
 	kfree(actions);
+
 	fte->fs_dr_rule.dr_rule = rule;
 	fte->fs_dr_rule.num_actions = fs_dr_num_actions;
 	fte->fs_dr_rule.dr_actions = fs_dr_actions;
@@ -436,13 +486,18 @@ static int mlx5_cmd_dr_create_fte(struct mlx5_flow_root_namespace *ns,
 	return 0;
 
 free_actions:
-	for (i = 0; i < fs_dr_num_actions; i++)
+	/* Free in reverse order to handle action dependencies */
+	for (i = fs_dr_num_actions - 1; i >= 0; i--)
 		if (!IS_ERR_OR_NULL(fs_dr_actions[i]))
 			mlx5dr_action_destroy(fs_dr_actions[i]);
 
-	mlx5_core_err(dev, "Failed to create dr rule err(%d)\n", err);
-	kfree(actions);
+	kfree(term_actions);
+free_fs_dr_actions_alloc:
 	kfree(fs_dr_actions);
+free_actions_alloc:
+	kfree(actions);
+out_err:
+	mlx5_core_err(dev, "Failed to create dr rule err(%d)\n", err);
 	return err;
 }
 
@@ -549,7 +604,8 @@ static int mlx5_cmd_dr_delete_fte(struct mlx5_flow_root_namespace *ns,
 	if (err)
 		return err;
 
-	for (i = 0; i < rule->num_actions; i++)
+	/* Free in reverse order to handle action dependencies */
+	for (i = rule->num_actions - 1; i >= 0; i--)
 		if (!IS_ERR_OR_NULL(rule->dr_actions[i]))
 			mlx5dr_action_destroy(rule->dr_actions[i]);
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5_ifc_dr.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5_ifc_dr.h
index 1722f4668269..e01c3766c7de 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5_ifc_dr.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5_ifc_dr.h
@@ -32,6 +32,7 @@ enum {
 };
 
 enum {
+	MLX5DR_ACTION_MDFY_HW_OP_COPY		= 0x1,
 	MLX5DR_ACTION_MDFY_HW_OP_SET		= 0x2,
 	MLX5DR_ACTION_MDFY_HW_OP_ADD		= 0x3,
 };
@@ -625,4 +626,19 @@ struct mlx5_ifc_dr_action_hw_set_bits {
 	u8         inline_data[0x20];
 };
 
+struct mlx5_ifc_dr_action_hw_copy_bits {
+	u8         opcode[0x8];
+	u8         destination_field_code[0x8];
+	u8         reserved_at_10[0x2];
+	u8         destination_left_shifter[0x6];
+	u8         reserved_at_18[0x2];
+	u8         destination_length[0x6];
+
+	u8         reserved_at_20[0x8];
+	u8         source_field_code[0x8];
+	u8         reserved_at_30[0x2];
+	u8         source_left_shifter[0x6];
+	u8         reserved_at_38[0x8];
+};
+
 #endif /* MLX5_IFC_DR_H */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h
index adda9cbfba45..e1edc9c247b7 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5dr.h
@@ -33,6 +33,11 @@ struct mlx5dr_match_parameters {
 	u64 *match_buf; /* Device spec format */
 };
 
+struct mlx5dr_action_dest {
+	struct mlx5dr_action *dest;
+	struct mlx5dr_action *reformat;
+};
+
 #ifdef CONFIG_MLX5_SW_STEERING
 
 struct mlx5dr_domain *
@@ -46,7 +51,7 @@ void mlx5dr_domain_set_peer(struct mlx5dr_domain *dmn,
 			    struct mlx5dr_domain *peer_dmn);
 
 struct mlx5dr_table *
-mlx5dr_table_create(struct mlx5dr_domain *domain, u32 level);
+mlx5dr_table_create(struct mlx5dr_domain *domain, u32 level, u32 flags);
 
 int mlx5dr_table_destroy(struct mlx5dr_table *table);
 
@@ -75,14 +80,19 @@ struct mlx5dr_action *
 mlx5dr_action_create_dest_table(struct mlx5dr_table *table);
 
 struct mlx5dr_action *
-mlx5dr_create_action_dest_flow_fw_table(struct mlx5_flow_table *ft,
-					struct mlx5_core_dev *mdev);
+mlx5dr_action_create_dest_flow_fw_table(struct mlx5dr_domain *domain,
+					struct mlx5_flow_table *ft);
 
 struct mlx5dr_action *
 mlx5dr_action_create_dest_vport(struct mlx5dr_domain *domain,
 				u32 vport, u8 vhca_id_valid,
 				u16 vhca_id);
 
+struct mlx5dr_action *
+mlx5dr_action_create_mult_dest_tbl(struct mlx5dr_domain *dmn,
+				   struct mlx5dr_action_dest *dests,
+				   u32 num_of_dests);
+
 struct mlx5dr_action *mlx5dr_action_create_drop(void);
 
 struct mlx5dr_action *mlx5dr_action_create_tag(u32 tag_value);
@@ -131,7 +141,7 @@ mlx5dr_domain_set_peer(struct mlx5dr_domain *dmn,
 		       struct mlx5dr_domain *peer_dmn) { }
 
 static inline struct mlx5dr_table *
-mlx5dr_table_create(struct mlx5dr_domain *domain, u32 level) { return NULL; }
+mlx5dr_table_create(struct mlx5dr_domain *domain, u32 level, u32 flags) { return NULL; }
 
 static inline int
 mlx5dr_table_destroy(struct mlx5dr_table *table) { return 0; }
@@ -165,8 +175,8 @@ static inline struct mlx5dr_action *
 mlx5dr_action_create_dest_table(struct mlx5dr_table *table) { return NULL; }
 
 static inline struct mlx5dr_action *
-mlx5dr_create_action_dest_flow_fw_table(struct mlx5_flow_table *ft,
-					struct mlx5_core_dev *mdev) { return NULL; }
+mlx5dr_action_create_dest_flow_fw_table(struct mlx5dr_domain *domain,
+					struct mlx5_flow_table *ft) { return NULL; }
 
 static inline struct mlx5dr_action *
 mlx5dr_action_create_dest_vport(struct mlx5dr_domain *domain,
@@ -174,6 +184,11 @@ mlx5dr_action_create_dest_vport(struct mlx5dr_domain *domain,
 				u16 vhca_id) { return NULL; }
 
 static inline struct mlx5dr_action *
+mlx5dr_action_create_mult_dest_tbl(struct mlx5dr_domain *dmn,
+				   struct mlx5dr_action_dest *dests,
+				   u32 num_of_dests)  { return NULL; }
+
+static inline struct mlx5dr_action *
 mlx5dr_action_create_drop(void) { return NULL; }
 
 static inline struct mlx5dr_action *
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/wq.c b/drivers/net/ethernet/mellanox/mlx5/core/wq.c
index f2a0e72285ba..02f7e4a39578 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/wq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/wq.c
@@ -89,7 +89,7 @@ void mlx5_wq_cyc_wqe_dump(struct mlx5_wq_cyc *wq, u16 ix, u8 nstrides)
 	len = nstrides << wq->fbc.log_stride;
 	wqe = mlx5_wq_cyc_get_wqe(wq, ix);
 
-	pr_info("WQE DUMP: WQ size %d WQ cur size %d, WQE index 0x%x, len: %ld\n",
+	pr_info("WQE DUMP: WQ size %d WQ cur size %d, WQE index 0x%x, len: %zu\n",
 		mlx5_wq_cyc_get_size(wq), wq->cur_sz, ix, len);
 	print_hex_dump(KERN_WARNING, "", DUMP_PREFIX_OFFSET, 16, 1, wqe, len, false);
 }
diff --git a/drivers/net/ethernet/mellanox/mlxsw/minimal.c b/drivers/net/ethernet/mellanox/mlxsw/minimal.c
index 2b543911ae00..c4caeeadcba9 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/minimal.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/minimal.c
@@ -213,8 +213,8 @@ mlxsw_m_port_create(struct mlxsw_m *mlxsw_m, u8 local_port, u8 module)
 
 err_register_netdev:
 	mlxsw_m->ports[local_port] = NULL;
-	free_netdev(dev);
 err_dev_addr_get:
+	free_netdev(dev);
 err_alloc_etherdev:
 	mlxsw_core_port_fini(mlxsw_m->core, local_port);
 	return err;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/reg.h b/drivers/net/ethernet/mellanox/mlxsw/reg.h
index af30e8a76682..dd6685156396 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/reg.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/reg.h
@@ -3477,10 +3477,10 @@ MLXSW_REG_DEFINE(qeec, MLXSW_REG_QEEC_ID, MLXSW_REG_QEEC_LEN);
 MLXSW_ITEM32(reg, qeec, local_port, 0x00, 16, 8);
 
 enum mlxsw_reg_qeec_hr {
-	MLXSW_REG_QEEC_HIERARCY_PORT,
-	MLXSW_REG_QEEC_HIERARCY_GROUP,
-	MLXSW_REG_QEEC_HIERARCY_SUBGROUP,
-	MLXSW_REG_QEEC_HIERARCY_TC,
+	MLXSW_REG_QEEC_HR_PORT,
+	MLXSW_REG_QEEC_HR_GROUP,
+	MLXSW_REG_QEEC_HR_SUBGROUP,
+	MLXSW_REG_QEEC_HR_TC,
 };
 
 /* reg_qeec_element_hierarchy
@@ -3563,8 +3563,8 @@ MLXSW_ITEM32(reg, qeec, min_shaper_rate, 0x0C, 0, 28);
  */
 MLXSW_ITEM32(reg, qeec, mase, 0x10, 31, 1);
 
-/* A large max rate will disable the max shaper. */
-#define MLXSW_REG_QEEC_MAS_DIS	200000000	/* Kbps */
+/* The largest max shaper value possible to disable the shaper. */
+#define MLXSW_REG_QEEC_MAS_DIS	((1u << 31) - 1)	/* Kbps */
 
 /* reg_qeec_max_shaper_rate
  * Max shaper information rate.
@@ -3602,6 +3602,21 @@ MLXSW_ITEM32(reg, qeec, dwrr, 0x18, 15, 1);
  */
 MLXSW_ITEM32(reg, qeec, dwrr_weight, 0x18, 0, 8);
 
+/* reg_qeec_max_shaper_bs
+ * Max shaper burst size
+ * Burst size is 2^max_shaper_bs * 512 bits
+ * For Spectrum-1: Range is: 5..25
+ * For Spectrum-2: Range is: 11..25
+ * Reserved when ptps = 1
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, qeec, max_shaper_bs, 0x1C, 0, 6);
+
+#define MLXSW_REG_QEEC_HIGHEST_SHAPER_BS	25
+#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP1	5
+#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP2	11
+#define MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP3	5
+
 static inline void mlxsw_reg_qeec_pack(char *payload, u8 local_port,
 				       enum mlxsw_reg_qeec_hr hr, u8 index,
 				       u8 next_index)
@@ -3618,8 +3633,7 @@ static inline void mlxsw_reg_qeec_ptps_pack(char *payload, u8 local_port,
 {
 	MLXSW_REG_ZERO(qeec, payload);
 	mlxsw_reg_qeec_local_port_set(payload, local_port);
-	mlxsw_reg_qeec_element_hierarchy_set(payload,
-					     MLXSW_REG_QEEC_HIERARCY_PORT);
+	mlxsw_reg_qeec_element_hierarchy_set(payload, MLXSW_REG_QEEC_HR_PORT);
 	mlxsw_reg_qeec_ptps_set(payload, ptps);
 }
 
@@ -3749,6 +3763,38 @@ mlxsw_reg_qpdsm_prio_pack(char *payload, unsigned short prio, u8 dscp)
 	mlxsw_reg_qpdsm_prio_entry_color2_dscp_set(payload, prio, dscp);
 }
 
+/* QPDP - QoS Port DSCP to Priority Mapping Register
+ * -------------------------------------------------
+ * This register controls the port default Switch Priority and Color. The
+ * default Switch Priority and Color are used for frames where the trust state
+ * uses default values. All member ports of a LAG should be configured with the
+ * same default values.
+ */
+#define MLXSW_REG_QPDP_ID 0x4007
+#define MLXSW_REG_QPDP_LEN 0x8
+
+MLXSW_REG_DEFINE(qpdp, MLXSW_REG_QPDP_ID, MLXSW_REG_QPDP_LEN);
+
+/* reg_qpdp_local_port
+ * Local Port. Supported for data packets from CPU port.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, qpdp, local_port, 0x00, 16, 8);
+
+/* reg_qpdp_switch_prio
+ * Default port Switch Priority (default 0)
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, qpdp, switch_prio, 0x04, 0, 4);
+
+static inline void mlxsw_reg_qpdp_pack(char *payload, u8 local_port,
+				       u8 switch_prio)
+{
+	MLXSW_REG_ZERO(qpdp, payload);
+	mlxsw_reg_qpdp_local_port_set(payload, local_port);
+	mlxsw_reg_qpdp_switch_prio_set(payload, switch_prio);
+}
+
 /* QPDPM - QoS Port DSCP to Priority Mapping Register
  * --------------------------------------------------
  * This register controls the mapping from DSCP field to
@@ -5482,6 +5528,7 @@ enum mlxsw_reg_htgt_discard_trap_group {
 	MLXSW_REG_HTGT_DISCARD_TRAP_GROUP_BASE = MLXSW_REG_HTGT_TRAP_GROUP_MAX,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_L2_DISCARDS,
 	MLXSW_REG_HTGT_TRAP_GROUP_SP_L3_DISCARDS,
+	MLXSW_REG_HTGT_TRAP_GROUP_SP_TUNNEL_DISCARDS,
 };
 
 /* reg_htgt_trap_group
@@ -10109,6 +10156,92 @@ static inline void mlxsw_reg_tigcr_pack(char *payload, bool ttlc, u8 ttl_uc)
 	mlxsw_reg_tigcr_ttl_uc_set(payload, ttl_uc);
 }
 
+/* TIEEM - Tunneling IPinIP Encapsulation ECN Mapping Register
+ * -----------------------------------------------------------
+ * The TIEEM register maps ECN of the IP header at the ingress to the
+ * encapsulation to the ECN of the underlay network.
+ */
+#define MLXSW_REG_TIEEM_ID 0xA812
+#define MLXSW_REG_TIEEM_LEN 0x0C
+
+MLXSW_REG_DEFINE(tieem, MLXSW_REG_TIEEM_ID, MLXSW_REG_TIEEM_LEN);
+
+/* reg_tieem_overlay_ecn
+ * ECN of the IP header in the overlay network.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, tieem, overlay_ecn, 0x04, 24, 2);
+
+/* reg_tineem_underlay_ecn
+ * ECN of the IP header in the underlay network.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, tieem, underlay_ecn, 0x04, 16, 2);
+
+static inline void mlxsw_reg_tieem_pack(char *payload, u8 overlay_ecn,
+					u8 underlay_ecn)
+{
+	MLXSW_REG_ZERO(tieem, payload);
+	mlxsw_reg_tieem_overlay_ecn_set(payload, overlay_ecn);
+	mlxsw_reg_tieem_underlay_ecn_set(payload, underlay_ecn);
+}
+
+/* TIDEM - Tunneling IPinIP Decapsulation ECN Mapping Register
+ * -----------------------------------------------------------
+ * The TIDEM register configures the actions that are done in the
+ * decapsulation.
+ */
+#define MLXSW_REG_TIDEM_ID 0xA813
+#define MLXSW_REG_TIDEM_LEN 0x0C
+
+MLXSW_REG_DEFINE(tidem, MLXSW_REG_TIDEM_ID, MLXSW_REG_TIDEM_LEN);
+
+/* reg_tidem_underlay_ecn
+ * ECN field of the IP header in the underlay network.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, tidem, underlay_ecn, 0x04, 24, 2);
+
+/* reg_tidem_overlay_ecn
+ * ECN field of the IP header in the overlay network.
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, tidem, overlay_ecn, 0x04, 16, 2);
+
+/* reg_tidem_eip_ecn
+ * Egress IP ECN. ECN field of the IP header of the packet which goes out
+ * from the decapsulation.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, tidem, eip_ecn, 0x04, 8, 2);
+
+/* reg_tidem_trap_en
+ * Trap enable:
+ * 0 - No trap due to decap ECN
+ * 1 - Trap enable with trap_id
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, tidem, trap_en, 0x08, 28, 4);
+
+/* reg_tidem_trap_id
+ * Trap ID. Either DECAP_ECN0 or DECAP_ECN1.
+ * Reserved when trap_en is '0'.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, tidem, trap_id, 0x08, 0, 9);
+
+static inline void mlxsw_reg_tidem_pack(char *payload, u8 underlay_ecn,
+					u8 overlay_ecn, u8 eip_ecn,
+					bool trap_en, u16 trap_id)
+{
+	MLXSW_REG_ZERO(tidem, payload);
+	mlxsw_reg_tidem_underlay_ecn_set(payload, underlay_ecn);
+	mlxsw_reg_tidem_overlay_ecn_set(payload, overlay_ecn);
+	mlxsw_reg_tidem_eip_ecn_set(payload, eip_ecn);
+	mlxsw_reg_tidem_trap_en_set(payload, trap_en);
+	mlxsw_reg_tidem_trap_id_set(payload, trap_id);
+}
+
 /* SBPR - Shared Buffer Pools Register
  * -----------------------------------
  * The SBPR configures and retrieves the shared buffer pools and configuration.
@@ -10581,6 +10714,7 @@ static const struct mlxsw_reg_info *mlxsw_reg_infos[] = {
 	MLXSW_REG(qeec),
 	MLXSW_REG(qrwe),
 	MLXSW_REG(qpdsm),
+	MLXSW_REG(qpdp),
 	MLXSW_REG(qpdpm),
 	MLXSW_REG(qtctm),
 	MLXSW_REG(qpsc),
@@ -10652,6 +10786,8 @@ static const struct mlxsw_reg_info *mlxsw_reg_infos[] = {
 	MLXSW_REG(tndem),
 	MLXSW_REG(tnpc),
 	MLXSW_REG(tigcr),
+	MLXSW_REG(tieem),
+	MLXSW_REG(tidem),
 	MLXSW_REG(sbpr),
 	MLXSW_REG(sbcm),
 	MLXSW_REG(sbpm),
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
index 8ed15199eb4f..7358b5bc7eb6 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.c
@@ -45,11 +45,9 @@
 #include "spectrum_ptp.h"
 #include "../mlxfw/mlxfw.h"
 
-#define MLXSW_SP_FWREV_MINOR_TO_BRANCH(minor) ((minor) / 100)
-
 #define MLXSW_SP1_FWREV_MAJOR 13
 #define MLXSW_SP1_FWREV_MINOR 2000
-#define MLXSW_SP1_FWREV_SUBMINOR 2308
+#define MLXSW_SP1_FWREV_SUBMINOR 2714
 #define MLXSW_SP1_FWREV_CAN_RESET_MINOR 1702
 
 static const struct mlxsw_fw_rev mlxsw_sp1_fw_rev = {
@@ -66,7 +64,7 @@ static const struct mlxsw_fw_rev mlxsw_sp1_fw_rev = {
 
 #define MLXSW_SP2_FWREV_MAJOR 29
 #define MLXSW_SP2_FWREV_MINOR 2000
-#define MLXSW_SP2_FWREV_SUBMINOR 2308
+#define MLXSW_SP2_FWREV_SUBMINOR 2714
 
 static const struct mlxsw_fw_rev mlxsw_sp2_fw_rev = {
 	.major = MLXSW_SP2_FWREV_MAJOR,
@@ -197,6 +195,10 @@ struct mlxsw_sp_ptp_ops {
 			  u64 *data, int data_index);
 };
 
+struct mlxsw_sp_span_ops {
+	u32 (*buffsize_get)(int mtu, u32 speed);
+};
+
 static int mlxsw_sp_component_query(struct mlxfw_dev *mlxfw_dev,
 				    u16 component_index, u32 *p_max_size,
 				    u8 *p_align_bits, u16 *p_max_write_size)
@@ -423,13 +425,12 @@ static int mlxsw_sp_fw_rev_validate(struct mlxsw_sp *mlxsw_sp)
 		     rev->major, req_rev->major);
 		return -EINVAL;
 	}
-	if (MLXSW_SP_FWREV_MINOR_TO_BRANCH(rev->minor) ==
-	    MLXSW_SP_FWREV_MINOR_TO_BRANCH(req_rev->minor) &&
-	    mlxsw_core_fw_rev_minor_subminor_validate(rev, req_rev))
+	if (mlxsw_core_fw_rev_minor_subminor_validate(rev, req_rev))
 		return 0;
 
-	dev_info(mlxsw_sp->bus_info->dev, "The firmware version %d.%d.%d is incompatible with the driver\n",
-		 rev->major, rev->minor, rev->subminor);
+	dev_err(mlxsw_sp->bus_info->dev, "The firmware version %d.%d.%d is incompatible with the driver (required >= %d.%d.%d)\n",
+		rev->major, rev->minor, rev->subminor, req_rev->major,
+		req_rev->minor, req_rev->subminor);
 	dev_info(mlxsw_sp->bus_info->dev, "Flashing firmware using file %s\n",
 		 fw_filename);
 
@@ -1793,6 +1794,10 @@ static int mlxsw_sp_setup_tc(struct net_device *dev, enum tc_setup_type type,
 		return mlxsw_sp_setup_tc_red(mlxsw_sp_port, type_data);
 	case TC_SETUP_QDISC_PRIO:
 		return mlxsw_sp_setup_tc_prio(mlxsw_sp_port, type_data);
+	case TC_SETUP_QDISC_ETS:
+		return mlxsw_sp_setup_tc_ets(mlxsw_sp_port, type_data);
+	case TC_SETUP_QDISC_TBF:
+		return mlxsw_sp_setup_tc_tbf(mlxsw_sp_port, type_data);
 	default:
 		return -EOPNOTSUPP;
 	}
@@ -3536,6 +3541,27 @@ mlxsw_sp_port_speed_by_width_set(struct mlxsw_sp_port *mlxsw_sp_port)
 	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ptys), ptys_pl);
 }
 
+int mlxsw_sp_port_speed_get(struct mlxsw_sp_port *mlxsw_sp_port, u32 *speed)
+{
+	const struct mlxsw_sp_port_type_speed_ops *port_type_speed_ops;
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char ptys_pl[MLXSW_REG_PTYS_LEN];
+	u32 eth_proto_oper;
+	int err;
+
+	port_type_speed_ops = mlxsw_sp->port_type_speed_ops;
+	port_type_speed_ops->reg_ptys_eth_pack(mlxsw_sp, ptys_pl,
+					       mlxsw_sp_port->local_port, 0,
+					       false);
+	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ptys), ptys_pl);
+	if (err)
+		return err;
+	port_type_speed_ops->reg_ptys_eth_unpack(mlxsw_sp, ptys_pl, NULL, NULL,
+						 &eth_proto_oper);
+	*speed = port_type_speed_ops->from_ptys_speed(mlxsw_sp, eth_proto_oper);
+	return 0;
+}
+
 int mlxsw_sp_port_ets_set(struct mlxsw_sp_port *mlxsw_sp_port,
 			  enum mlxsw_reg_qeec_hr hr, u8 index, u8 next_index,
 			  bool dwrr, u8 dwrr_weight)
@@ -3553,7 +3579,7 @@ int mlxsw_sp_port_ets_set(struct mlxsw_sp_port *mlxsw_sp_port,
 
 int mlxsw_sp_port_ets_maxrate_set(struct mlxsw_sp_port *mlxsw_sp_port,
 				  enum mlxsw_reg_qeec_hr hr, u8 index,
-				  u8 next_index, u32 maxrate)
+				  u8 next_index, u32 maxrate, u8 burst_size)
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
 	char qeec_pl[MLXSW_REG_QEEC_LEN];
@@ -3562,6 +3588,7 @@ int mlxsw_sp_port_ets_maxrate_set(struct mlxsw_sp_port *mlxsw_sp_port,
 			    next_index);
 	mlxsw_reg_qeec_mase_set(qeec_pl, true);
 	mlxsw_reg_qeec_max_shaper_rate_set(qeec_pl, maxrate);
+	mlxsw_reg_qeec_max_shaper_bs_set(qeec_pl, burst_size);
 	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(qeec), qeec_pl);
 }
 
@@ -3599,26 +3626,25 @@ static int mlxsw_sp_port_ets_init(struct mlxsw_sp_port *mlxsw_sp_port)
 	 * one subgroup, which are all member in the same group.
 	 */
 	err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
-				    MLXSW_REG_QEEC_HIERARCY_GROUP, 0, 0, false,
-				    0);
+				    MLXSW_REG_QEEC_HR_GROUP, 0, 0, false, 0);
 	if (err)
 		return err;
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
-					    MLXSW_REG_QEEC_HIERARCY_SUBGROUP, i,
+					    MLXSW_REG_QEEC_HR_SUBGROUP, i,
 					    0, false, 0);
 		if (err)
 			return err;
 	}
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
-					    MLXSW_REG_QEEC_HIERARCY_TC, i, i,
+					    MLXSW_REG_QEEC_HR_TC, i, i,
 					    false, 0);
 		if (err)
 			return err;
 
 		err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
-					    MLXSW_REG_QEEC_HIERARCY_TC,
+					    MLXSW_REG_QEEC_HR_TC,
 					    i + 8, i,
 					    true, 100);
 		if (err)
@@ -3630,30 +3656,30 @@ static int mlxsw_sp_port_ets_init(struct mlxsw_sp_port *mlxsw_sp_port)
 	 * for the initial configuration.
 	 */
 	err = mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
-					    MLXSW_REG_QEEC_HIERARCY_PORT, 0, 0,
-					    MLXSW_REG_QEEC_MAS_DIS);
+					    MLXSW_REG_QEEC_HR_PORT, 0, 0,
+					    MLXSW_REG_QEEC_MAS_DIS, 0);
 	if (err)
 		return err;
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		err = mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
-						    MLXSW_REG_QEEC_HIERARCY_SUBGROUP,
+						    MLXSW_REG_QEEC_HR_SUBGROUP,
 						    i, 0,
-						    MLXSW_REG_QEEC_MAS_DIS);
+						    MLXSW_REG_QEEC_MAS_DIS, 0);
 		if (err)
 			return err;
 	}
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		err = mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
-						    MLXSW_REG_QEEC_HIERARCY_TC,
+						    MLXSW_REG_QEEC_HR_TC,
 						    i, i,
-						    MLXSW_REG_QEEC_MAS_DIS);
+						    MLXSW_REG_QEEC_MAS_DIS, 0);
 		if (err)
 			return err;
 
 		err = mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
-						    MLXSW_REG_QEEC_HIERARCY_TC,
+						    MLXSW_REG_QEEC_HR_TC,
 						    i + 8, i,
-						    MLXSW_REG_QEEC_MAS_DIS);
+						    MLXSW_REG_QEEC_MAS_DIS, 0);
 		if (err)
 			return err;
 	}
@@ -3661,7 +3687,7 @@ static int mlxsw_sp_port_ets_init(struct mlxsw_sp_port *mlxsw_sp_port)
 	/* Configure the min shaper for multicast TCs. */
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		err = mlxsw_sp_port_min_bw_set(mlxsw_sp_port,
-					       MLXSW_REG_QEEC_HIERARCY_TC,
+					       MLXSW_REG_QEEC_HR_TC,
 					       i + 8, i,
 					       MLXSW_REG_QEEC_MIS_MIN);
 		if (err)
@@ -3885,6 +3911,8 @@ static int mlxsw_sp_port_create(struct mlxsw_sp *mlxsw_sp, u8 local_port,
 
 	INIT_DELAYED_WORK(&mlxsw_sp_port->ptp.shaper_dw,
 			  mlxsw_sp->ptp_ops->shaper_work);
+	INIT_DELAYED_WORK(&mlxsw_sp_port->span.speed_update_dw,
+			  mlxsw_sp_span_speed_update_work);
 
 	mlxsw_sp->ports[local_port] = mlxsw_sp_port;
 	err = register_netdev(dev);
@@ -3941,6 +3969,7 @@ static void mlxsw_sp_port_remove(struct mlxsw_sp *mlxsw_sp, u8 local_port)
 	struct mlxsw_sp_port *mlxsw_sp_port = mlxsw_sp->ports[local_port];
 
 	cancel_delayed_work_sync(&mlxsw_sp_port->periodic_hw_stats.update_dw);
+	cancel_delayed_work_sync(&mlxsw_sp_port->span.speed_update_dw);
 	cancel_delayed_work_sync(&mlxsw_sp_port->ptp.shaper_dw);
 	mlxsw_sp_port_ptp_clear(mlxsw_sp_port);
 	mlxsw_core_port_clear(mlxsw_sp->core, local_port, mlxsw_sp);
@@ -4348,6 +4377,7 @@ static void mlxsw_sp_pude_event_func(const struct mlxsw_reg_info *reg,
 		netdev_info(mlxsw_sp_port->dev, "link up\n");
 		netif_carrier_on(mlxsw_sp_port->dev);
 		mlxsw_core_schedule_dw(&mlxsw_sp_port->ptp.shaper_dw, 0);
+		mlxsw_core_schedule_dw(&mlxsw_sp_port->span.speed_update_dw, 0);
 	} else {
 		netdev_info(mlxsw_sp_port->dev, "link down\n");
 		netif_carrier_off(mlxsw_sp_port->dev);
@@ -4547,10 +4577,16 @@ static const struct mlxsw_listener mlxsw_sp_listener[] = {
 			  false),
 	MLXSW_SP_RXL_MARK(ROUTER_ALERT_IPV4, TRAP_TO_CPU, ROUTER_EXP, false),
 	MLXSW_SP_RXL_MARK(ROUTER_ALERT_IPV6, TRAP_TO_CPU, ROUTER_EXP, false),
-	MLXSW_SP_RXL_MARK(IPIP_DECAP_ERROR, TRAP_TO_CPU, ROUTER_EXP, false),
-	MLXSW_SP_RXL_MARK(DECAP_ECN0, TRAP_TO_CPU, ROUTER_EXP, false),
 	MLXSW_SP_RXL_MARK(IPV4_VRRP, TRAP_TO_CPU, VRRP, false),
 	MLXSW_SP_RXL_MARK(IPV6_VRRP, TRAP_TO_CPU, VRRP, false),
+	MLXSW_SP_RXL_NO_MARK(DISCARD_ING_ROUTER_SIP_CLASS_E, FORWARD,
+			     ROUTER_EXP, false),
+	MLXSW_SP_RXL_NO_MARK(DISCARD_ING_ROUTER_MC_DMAC, FORWARD,
+			     ROUTER_EXP, false),
+	MLXSW_SP_RXL_NO_MARK(DISCARD_ING_ROUTER_SIP_DIP, FORWARD,
+			     ROUTER_EXP, false),
+	MLXSW_SP_RXL_NO_MARK(DISCARD_ING_ROUTER_DIP_LINK_LOCAL, FORWARD,
+			     ROUTER_EXP, false),
 	/* PKT Sample trap */
 	MLXSW_RXL(mlxsw_sp_rx_listener_sample_func, PKT_SAMPLE, MIRROR_TO_CPU,
 		  false, SP_IP2ME, DISCARD),
@@ -4889,6 +4925,33 @@ static const struct mlxsw_sp_ptp_ops mlxsw_sp2_ptp_ops = {
 	.get_stats	= mlxsw_sp2_get_stats,
 };
 
+static u32 mlxsw_sp1_span_buffsize_get(int mtu, u32 speed)
+{
+	return mtu * 5 / 2;
+}
+
+static const struct mlxsw_sp_span_ops mlxsw_sp1_span_ops = {
+	.buffsize_get = mlxsw_sp1_span_buffsize_get,
+};
+
+#define MLXSW_SP2_SPAN_EG_MIRROR_BUFFER_FACTOR 38
+
+static u32 mlxsw_sp2_span_buffsize_get(int mtu, u32 speed)
+{
+	return 3 * mtu + MLXSW_SP2_SPAN_EG_MIRROR_BUFFER_FACTOR * speed / 1000;
+}
+
+static const struct mlxsw_sp_span_ops mlxsw_sp2_span_ops = {
+	.buffsize_get = mlxsw_sp2_span_buffsize_get,
+};
+
+u32 mlxsw_sp_span_buffsize_get(struct mlxsw_sp *mlxsw_sp, int mtu, u32 speed)
+{
+	u32 buffsize = mlxsw_sp->span_ops->buffsize_get(speed, mtu);
+
+	return mlxsw_sp_bytes_cells(mlxsw_sp, buffsize) + 1;
+}
+
 static int mlxsw_sp_netdevice_event(struct notifier_block *unused,
 				    unsigned long event, void *ptr);
 
@@ -5110,8 +5173,10 @@ static int mlxsw_sp1_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->sb_vals = &mlxsw_sp1_sb_vals;
 	mlxsw_sp->port_type_speed_ops = &mlxsw_sp1_port_type_speed_ops;
 	mlxsw_sp->ptp_ops = &mlxsw_sp1_ptp_ops;
+	mlxsw_sp->span_ops = &mlxsw_sp1_span_ops;
 	mlxsw_sp->listeners = mlxsw_sp1_listener;
 	mlxsw_sp->listeners_count = ARRAY_SIZE(mlxsw_sp1_listener);
+	mlxsw_sp->lowest_shaper_bs = MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP1;
 
 	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
 }
@@ -5135,6 +5200,8 @@ static int mlxsw_sp2_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->sb_vals = &mlxsw_sp2_sb_vals;
 	mlxsw_sp->port_type_speed_ops = &mlxsw_sp2_port_type_speed_ops;
 	mlxsw_sp->ptp_ops = &mlxsw_sp2_ptp_ops;
+	mlxsw_sp->span_ops = &mlxsw_sp2_span_ops;
+	mlxsw_sp->lowest_shaper_bs = MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP2;
 
 	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
 }
@@ -5156,6 +5223,8 @@ static int mlxsw_sp3_init(struct mlxsw_core *mlxsw_core,
 	mlxsw_sp->sb_vals = &mlxsw_sp2_sb_vals;
 	mlxsw_sp->port_type_speed_ops = &mlxsw_sp2_port_type_speed_ops;
 	mlxsw_sp->ptp_ops = &mlxsw_sp2_ptp_ops;
+	mlxsw_sp->span_ops = &mlxsw_sp2_span_ops;
+	mlxsw_sp->lowest_shaper_bs = MLXSW_REG_QEEC_LOWEST_SHAPER_BS_SP3;
 
 	return mlxsw_sp_init(mlxsw_core, mlxsw_bus_info, extack);
 }
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index 347bec9d1ecf..a0f1f9dceec5 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -140,6 +140,7 @@ struct mlxsw_sp_sb_vals;
 struct mlxsw_sp_port_type_speed_ops;
 struct mlxsw_sp_ptp_state;
 struct mlxsw_sp_ptp_ops;
+struct mlxsw_sp_span_ops;
 
 struct mlxsw_sp_port_mapping {
 	u8 module;
@@ -185,8 +186,10 @@ struct mlxsw_sp {
 	const struct mlxsw_sp_sb_vals *sb_vals;
 	const struct mlxsw_sp_port_type_speed_ops *port_type_speed_ops;
 	const struct mlxsw_sp_ptp_ops *ptp_ops;
+	const struct mlxsw_sp_span_ops *span_ops;
 	const struct mlxsw_listener *listeners;
 	size_t listeners_count;
+	u32 lowest_shaper_bs;
 };
 
 static inline struct mlxsw_sp_upper *
@@ -292,6 +295,9 @@ struct mlxsw_sp_port {
 		struct mlxsw_sp_ptp_port_stats stats;
 	} ptp;
 	u8 split_base_local_port;
+	struct {
+		struct delayed_work speed_update_dw;
+	} span;
 };
 
 struct mlxsw_sp_port_type_speed_ops {
@@ -471,6 +477,7 @@ extern struct notifier_block mlxsw_sp_switchdev_notifier;
 /* spectrum.c */
 void mlxsw_sp_rx_listener_no_mark_func(struct sk_buff *skb,
 				       u8 local_port, void *priv);
+int mlxsw_sp_port_speed_get(struct mlxsw_sp_port *mlxsw_sp_port, u32 *speed);
 int mlxsw_sp_port_ets_set(struct mlxsw_sp_port *mlxsw_sp_port,
 			  enum mlxsw_reg_qeec_hr hr, u8 index, u8 next_index,
 			  bool dwrr, u8 dwrr_weight);
@@ -481,7 +488,7 @@ int __mlxsw_sp_port_headroom_set(struct mlxsw_sp_port *mlxsw_sp_port, int mtu,
 				 struct ieee_pfc *my_pfc);
 int mlxsw_sp_port_ets_maxrate_set(struct mlxsw_sp_port *mlxsw_sp_port,
 				  enum mlxsw_reg_qeec_hr hr, u8 index,
-				  u8 next_index, u32 maxrate);
+				  u8 next_index, u32 maxrate, u8 burst_size);
 enum mlxsw_reg_spms_state mlxsw_sp_stp_spms_state(u8 stp_state);
 int mlxsw_sp_port_vid_stp_set(struct mlxsw_sp_port *mlxsw_sp_port, u16 vid,
 			      u8 state);
@@ -501,6 +508,7 @@ int mlxsw_sp_flow_counter_alloc(struct mlxsw_sp *mlxsw_sp,
 				unsigned int *p_counter_index);
 void mlxsw_sp_flow_counter_free(struct mlxsw_sp *mlxsw_sp,
 				unsigned int counter_index);
+u32 mlxsw_sp_span_buffsize_get(struct mlxsw_sp *mlxsw_sp, int mtu, u32 speed);
 bool mlxsw_sp_port_dev_check(const struct net_device *dev);
 struct mlxsw_sp *mlxsw_sp_lower_get(struct net_device *dev);
 struct mlxsw_sp_port *mlxsw_sp_port_dev_lower_find(struct net_device *dev);
@@ -852,6 +860,10 @@ int mlxsw_sp_setup_tc_red(struct mlxsw_sp_port *mlxsw_sp_port,
 			  struct tc_red_qopt_offload *p);
 int mlxsw_sp_setup_tc_prio(struct mlxsw_sp_port *mlxsw_sp_port,
 			   struct tc_prio_qopt_offload *p);
+int mlxsw_sp_setup_tc_ets(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct tc_ets_qopt_offload *p);
+int mlxsw_sp_setup_tc_tbf(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct tc_tbf_qopt_offload *p);
 
 /* spectrum_fid.c */
 bool mlxsw_sp_fid_is_dummy(struct mlxsw_sp *mlxsw_sp, u16 fid_index);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c
index 21296fa7f7fb..49a72a8f1f57 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_dcb.c
@@ -160,7 +160,7 @@ static int __mlxsw_sp_dcbnl_ieee_setets(struct mlxsw_sp_port *mlxsw_sp_port,
 		u8 weight = ets->tc_tx_bw[i];
 
 		err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
-					    MLXSW_REG_QEEC_HIERARCY_SUBGROUP, i,
+					    MLXSW_REG_QEEC_HR_SUBGROUP, i,
 					    0, dwrr, weight);
 		if (err) {
 			netdev_err(dev, "Failed to link subgroup ETS element %d to group\n",
@@ -198,7 +198,7 @@ err_port_ets_set:
 		u8 weight = my_ets->tc_tx_bw[i];
 
 		err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
-					    MLXSW_REG_QEEC_HIERARCY_SUBGROUP, i,
+					    MLXSW_REG_QEEC_HR_SUBGROUP, i,
 					    0, dwrr, weight);
 	}
 	return err;
@@ -369,6 +369,17 @@ err_update_qrwe:
 }
 
 static int
+mlxsw_sp_port_dcb_app_update_qpdp(struct mlxsw_sp_port *mlxsw_sp_port,
+				  u8 default_prio)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char qpdp_pl[MLXSW_REG_QPDP_LEN];
+
+	mlxsw_reg_qpdp_pack(qpdp_pl, mlxsw_sp_port->local_port, default_prio);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(qpdp), qpdp_pl);
+}
+
+static int
 mlxsw_sp_port_dcb_app_update_qpdpm(struct mlxsw_sp_port *mlxsw_sp_port,
 				   struct dcb_ieee_app_dscp_map *map)
 {
@@ -405,6 +416,12 @@ static int mlxsw_sp_port_dcb_app_update(struct mlxsw_sp_port *mlxsw_sp_port)
 	int err;
 
 	default_prio = mlxsw_sp_port_dcb_app_default_prio(mlxsw_sp_port);
+	err = mlxsw_sp_port_dcb_app_update_qpdp(mlxsw_sp_port, default_prio);
+	if (err) {
+		netdev_err(mlxsw_sp_port->dev, "Couldn't configure port default priority\n");
+		return err;
+	}
+
 	have_dscp = mlxsw_sp_port_dcb_app_prio_dscp_map(mlxsw_sp_port,
 							&prio_map);
 
@@ -507,9 +524,9 @@ static int mlxsw_sp_dcbnl_ieee_setmaxrate(struct net_device *dev,
 
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		err = mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
-						    MLXSW_REG_QEEC_HIERARCY_SUBGROUP,
+						    MLXSW_REG_QEEC_HR_SUBGROUP,
 						    i, 0,
-						    maxrate->tc_maxrate[i]);
+						    maxrate->tc_maxrate[i], 0);
 		if (err) {
 			netdev_err(dev, "Failed to set maxrate for TC %d\n", i);
 			goto err_port_ets_maxrate_set;
@@ -523,8 +540,9 @@ static int mlxsw_sp_dcbnl_ieee_setmaxrate(struct net_device *dev,
 err_port_ets_maxrate_set:
 	for (i--; i >= 0; i--)
 		mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
-					      MLXSW_REG_QEEC_HIERARCY_SUBGROUP,
-					      i, 0, my_maxrate->tc_maxrate[i]);
+					      MLXSW_REG_QEEC_HR_SUBGROUP,
+					      i, 0,
+					      my_maxrate->tc_maxrate[i], 0);
 	return err;
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c
index 6400cd644b7a..a8525992528f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ipip.c
@@ -3,8 +3,10 @@
 
 #include <net/ip_tunnels.h>
 #include <net/ip6_tunnel.h>
+#include <net/inet_ecn.h>
 
 #include "spectrum_ipip.h"
+#include "reg.h"
 
 struct ip_tunnel_parm
 mlxsw_sp_ipip_netdev_parms4(const struct net_device *ol_dev)
@@ -338,3 +340,61 @@ static const struct mlxsw_sp_ipip_ops mlxsw_sp_ipip_gre4_ops = {
 const struct mlxsw_sp_ipip_ops *mlxsw_sp_ipip_ops_arr[] = {
 	[MLXSW_SP_IPIP_TYPE_GRE4] = &mlxsw_sp_ipip_gre4_ops,
 };
+
+static int mlxsw_sp_ipip_ecn_encap_init_one(struct mlxsw_sp *mlxsw_sp,
+					    u8 inner_ecn, u8 outer_ecn)
+{
+	char tieem_pl[MLXSW_REG_TIEEM_LEN];
+
+	mlxsw_reg_tieem_pack(tieem_pl, inner_ecn, outer_ecn);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(tieem), tieem_pl);
+}
+
+int mlxsw_sp_ipip_ecn_encap_init(struct mlxsw_sp *mlxsw_sp)
+{
+	int i;
+
+	/* Iterate over inner ECN values */
+	for (i = INET_ECN_NOT_ECT; i <= INET_ECN_CE; i++) {
+		u8 outer_ecn = INET_ECN_encapsulate(0, i);
+		int err;
+
+		err = mlxsw_sp_ipip_ecn_encap_init_one(mlxsw_sp, i, outer_ecn);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int mlxsw_sp_ipip_ecn_decap_init_one(struct mlxsw_sp *mlxsw_sp,
+					    u8 inner_ecn, u8 outer_ecn)
+{
+	char tidem_pl[MLXSW_REG_TIDEM_LEN];
+	bool trap_en, set_ce = false;
+	u8 new_inner_ecn;
+
+	trap_en = __INET_ECN_decapsulate(outer_ecn, inner_ecn, &set_ce);
+	new_inner_ecn = set_ce ? INET_ECN_CE : inner_ecn;
+
+	mlxsw_reg_tidem_pack(tidem_pl, outer_ecn, inner_ecn, new_inner_ecn,
+			     trap_en, trap_en ? MLXSW_TRAP_ID_DECAP_ECN0 : 0);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(tidem), tidem_pl);
+}
+
+int mlxsw_sp_ipip_ecn_decap_init(struct mlxsw_sp *mlxsw_sp)
+{
+	int i, j, err;
+
+	/* Iterate over inner ECN values */
+	for (i = INET_ECN_NOT_ECT; i <= INET_ECN_CE; i++) {
+		/* Iterate over outer ECN values */
+		for (j = INET_ECN_NOT_ECT; j <= INET_ECN_CE; j++) {
+			err = mlxsw_sp_ipip_ecn_decap_init_one(mlxsw_sp, i, j);
+			if (err)
+				return err;
+		}
+	}
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ptp.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ptp.c
index ec2ff3d7f41c..34f7c3501b08 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_ptp.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_ptp.c
@@ -920,6 +920,7 @@ static int mlxsw_sp_ptp_get_message_types(const struct hwtstamp_config *config,
 		egr_types = 0xff;
 		break;
 	case HWTSTAMP_TX_ONESTEP_SYNC:
+	case HWTSTAMP_TX_ONESTEP_P2P:
 		return -ERANGE;
 	}
 
@@ -1015,27 +1016,17 @@ mlxsw_sp1_ptp_port_shaper_set(struct mlxsw_sp_port *mlxsw_sp_port, bool enable)
 
 static int mlxsw_sp1_ptp_port_shaper_check(struct mlxsw_sp_port *mlxsw_sp_port)
 {
-	const struct mlxsw_sp_port_type_speed_ops *port_type_speed_ops;
-	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
-	char ptys_pl[MLXSW_REG_PTYS_LEN];
-	u32 eth_proto_oper, speed;
 	bool ptps = false;
 	int err, i;
+	u32 speed;
 
 	if (!mlxsw_sp1_ptp_hwtstamp_enabled(mlxsw_sp_port))
 		return mlxsw_sp1_ptp_port_shaper_set(mlxsw_sp_port, false);
 
-	port_type_speed_ops = mlxsw_sp->port_type_speed_ops;
-	port_type_speed_ops->reg_ptys_eth_pack(mlxsw_sp, ptys_pl,
-					       mlxsw_sp_port->local_port, 0,
-					       false);
-	err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(ptys), ptys_pl);
+	err = mlxsw_sp_port_speed_get(mlxsw_sp_port, &speed);
 	if (err)
 		return err;
-	port_type_speed_ops->reg_ptys_eth_unpack(mlxsw_sp, ptys_pl, NULL, NULL,
-						 &eth_proto_oper);
 
-	speed = port_type_speed_ops->from_ptys_speed(mlxsw_sp, eth_proto_oper);
 	for (i = 0; i < MLXSW_SP1_PTP_SHAPER_PARAMS_LEN; i++) {
 		if (mlxsw_sp1_ptp_shaper_params[i].ethtool_speed == speed) {
 			ptps = true;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
index 0124bfe1963b..79a2801d59f6 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
@@ -18,6 +18,8 @@ enum mlxsw_sp_qdisc_type {
 	MLXSW_SP_QDISC_NO_QDISC,
 	MLXSW_SP_QDISC_RED,
 	MLXSW_SP_QDISC_PRIO,
+	MLXSW_SP_QDISC_ETS,
+	MLXSW_SP_QDISC_TBF,
 };
 
 struct mlxsw_sp_qdisc_ops {
@@ -226,6 +228,70 @@ mlxsw_sp_qdisc_bstats_per_priority_get(struct mlxsw_sp_port_xstats *xstats,
 	}
 }
 
+static void
+mlxsw_sp_qdisc_collect_tc_stats(struct mlxsw_sp_port *mlxsw_sp_port,
+				struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+				u64 *p_tx_bytes, u64 *p_tx_packets,
+				u64 *p_drops, u64 *p_backlog)
+{
+	u8 tclass_num = mlxsw_sp_qdisc->tclass_num;
+	struct mlxsw_sp_port_xstats *xstats;
+	u64 tx_bytes, tx_packets;
+
+	xstats = &mlxsw_sp_port->periodic_hw_stats.xstats;
+	mlxsw_sp_qdisc_bstats_per_priority_get(xstats,
+					       mlxsw_sp_qdisc->prio_bitmap,
+					       &tx_packets, &tx_bytes);
+
+	*p_tx_packets += tx_packets;
+	*p_tx_bytes += tx_bytes;
+	*p_drops += xstats->wred_drop[tclass_num] +
+		    mlxsw_sp_xstats_tail_drop(xstats, tclass_num);
+	*p_backlog += mlxsw_sp_xstats_backlog(xstats, tclass_num);
+}
+
+static void
+mlxsw_sp_qdisc_update_stats(struct mlxsw_sp *mlxsw_sp,
+			    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			    u64 tx_bytes, u64 tx_packets,
+			    u64 drops, u64 backlog,
+			    struct tc_qopt_offload_stats *stats_ptr)
+{
+	struct mlxsw_sp_qdisc_stats *stats_base = &mlxsw_sp_qdisc->stats_base;
+
+	tx_bytes -= stats_base->tx_bytes;
+	tx_packets -= stats_base->tx_packets;
+	drops -= stats_base->drops;
+	backlog -= stats_base->backlog;
+
+	_bstats_update(stats_ptr->bstats, tx_bytes, tx_packets);
+	stats_ptr->qstats->drops += drops;
+	stats_ptr->qstats->backlog += mlxsw_sp_cells_bytes(mlxsw_sp, backlog);
+
+	stats_base->backlog += backlog;
+	stats_base->drops += drops;
+	stats_base->tx_bytes += tx_bytes;
+	stats_base->tx_packets += tx_packets;
+}
+
+static void
+mlxsw_sp_qdisc_get_tc_stats(struct mlxsw_sp_port *mlxsw_sp_port,
+			    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			    struct tc_qopt_offload_stats *stats_ptr)
+{
+	u64 tx_packets = 0;
+	u64 tx_bytes = 0;
+	u64 backlog = 0;
+	u64 drops = 0;
+
+	mlxsw_sp_qdisc_collect_tc_stats(mlxsw_sp_port, mlxsw_sp_qdisc,
+					&tx_bytes, &tx_packets,
+					&drops, &backlog);
+	mlxsw_sp_qdisc_update_stats(mlxsw_sp_port->mlxsw_sp, mlxsw_sp_qdisc,
+				    tx_bytes, tx_packets, drops, backlog,
+				    stats_ptr);
+}
+
 static int
 mlxsw_sp_tclass_congestion_enable(struct mlxsw_sp_port *mlxsw_sp_port,
 				  int tclass_num, u32 min, u32 max,
@@ -356,19 +422,28 @@ mlxsw_sp_qdisc_red_replace(struct mlxsw_sp_port *mlxsw_sp_port,
 }
 
 static void
-mlxsw_sp_qdisc_red_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
-			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
-			     void *params)
+mlxsw_sp_qdisc_leaf_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
+			      struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			      struct gnet_stats_queue *qstats)
 {
-	struct tc_red_qopt_offload_params *p = params;
 	u64 backlog;
 
 	backlog = mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
 				       mlxsw_sp_qdisc->stats_base.backlog);
-	p->qstats->backlog -= backlog;
+	qstats->backlog -= backlog;
 	mlxsw_sp_qdisc->stats_base.backlog = 0;
 }
 
+static void
+mlxsw_sp_qdisc_red_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
+			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			     void *params)
+{
+	struct tc_red_qopt_offload_params *p = params;
+
+	mlxsw_sp_qdisc_leaf_unoffload(mlxsw_sp_port, mlxsw_sp_qdisc, p->qstats);
+}
+
 static int
 mlxsw_sp_qdisc_get_red_xstats(struct mlxsw_sp_port *mlxsw_sp_port,
 			      struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
@@ -402,41 +477,21 @@ mlxsw_sp_qdisc_get_red_stats(struct mlxsw_sp_port *mlxsw_sp_port,
 			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
 			     struct tc_qopt_offload_stats *stats_ptr)
 {
-	u64 tx_bytes, tx_packets, overlimits, drops, backlog;
 	u8 tclass_num = mlxsw_sp_qdisc->tclass_num;
 	struct mlxsw_sp_qdisc_stats *stats_base;
 	struct mlxsw_sp_port_xstats *xstats;
+	u64 overlimits;
 
 	xstats = &mlxsw_sp_port->periodic_hw_stats.xstats;
 	stats_base = &mlxsw_sp_qdisc->stats_base;
 
-	mlxsw_sp_qdisc_bstats_per_priority_get(xstats,
-					       mlxsw_sp_qdisc->prio_bitmap,
-					       &tx_packets, &tx_bytes);
-	tx_bytes = tx_bytes - stats_base->tx_bytes;
-	tx_packets = tx_packets - stats_base->tx_packets;
-
+	mlxsw_sp_qdisc_get_tc_stats(mlxsw_sp_port, mlxsw_sp_qdisc, stats_ptr);
 	overlimits = xstats->wred_drop[tclass_num] + xstats->ecn -
 		     stats_base->overlimits;
-	drops = xstats->wred_drop[tclass_num] +
-		mlxsw_sp_xstats_tail_drop(xstats, tclass_num) -
-		stats_base->drops;
-	backlog = mlxsw_sp_xstats_backlog(xstats, tclass_num);
 
-	_bstats_update(stats_ptr->bstats, tx_bytes, tx_packets);
 	stats_ptr->qstats->overlimits += overlimits;
-	stats_ptr->qstats->drops += drops;
-	stats_ptr->qstats->backlog +=
-				mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
-						     backlog) -
-				mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
-						     stats_base->backlog);
-
-	stats_base->backlog = backlog;
-	stats_base->drops +=  drops;
 	stats_base->overlimits += overlimits;
-	stats_base->tx_bytes += tx_bytes;
-	stats_base->tx_packets += tx_packets;
+
 	return 0;
 }
 
@@ -486,15 +541,215 @@ int mlxsw_sp_setup_tc_red(struct mlxsw_sp_port *mlxsw_sp_port,
 	}
 }
 
+static void
+mlxsw_sp_setup_tc_qdisc_leaf_clean_stats(struct mlxsw_sp_port *mlxsw_sp_port,
+					 struct mlxsw_sp_qdisc *mlxsw_sp_qdisc)
+{
+	u64 backlog_cells = 0;
+	u64 tx_packets = 0;
+	u64 tx_bytes = 0;
+	u64 drops = 0;
+
+	mlxsw_sp_qdisc_collect_tc_stats(mlxsw_sp_port, mlxsw_sp_qdisc,
+					&tx_bytes, &tx_packets,
+					&drops, &backlog_cells);
+
+	mlxsw_sp_qdisc->stats_base.tx_packets = tx_packets;
+	mlxsw_sp_qdisc->stats_base.tx_bytes = tx_bytes;
+	mlxsw_sp_qdisc->stats_base.drops = drops;
+	mlxsw_sp_qdisc->stats_base.backlog = 0;
+}
+
 static int
-mlxsw_sp_qdisc_prio_destroy(struct mlxsw_sp_port *mlxsw_sp_port,
-			    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc)
+mlxsw_sp_qdisc_tbf_destroy(struct mlxsw_sp_port *mlxsw_sp_port,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc)
+{
+	struct mlxsw_sp_qdisc *root_qdisc = mlxsw_sp_port->root_qdisc;
+
+	if (root_qdisc != mlxsw_sp_qdisc)
+		root_qdisc->stats_base.backlog -=
+					mlxsw_sp_qdisc->stats_base.backlog;
+
+	return mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
+					     MLXSW_REG_QEEC_HR_SUBGROUP,
+					     mlxsw_sp_qdisc->tclass_num, 0,
+					     MLXSW_REG_QEEC_MAS_DIS, 0);
+}
+
+static int
+mlxsw_sp_qdisc_tbf_bs(struct mlxsw_sp_port *mlxsw_sp_port,
+		      u32 max_size, u8 *p_burst_size)
+{
+	/* TBF burst size is configured in bytes. The ASIC burst size value is
+	 * ((2 ^ bs) * 512 bits. Convert the TBF bytes to 512-bit units.
+	 */
+	u32 bs512 = max_size / 64;
+	u8 bs = fls(bs512);
+
+	if (!bs)
+		return -EINVAL;
+	--bs;
+
+	/* Demand a power of two. */
+	if ((1 << bs) != bs512)
+		return -EINVAL;
+
+	if (bs < mlxsw_sp_port->mlxsw_sp->lowest_shaper_bs ||
+	    bs > MLXSW_REG_QEEC_HIGHEST_SHAPER_BS)
+		return -EINVAL;
+
+	*p_burst_size = bs;
+	return 0;
+}
+
+static u32
+mlxsw_sp_qdisc_tbf_max_size(u8 bs)
+{
+	return (1U << bs) * 64;
+}
+
+static u64
+mlxsw_sp_qdisc_tbf_rate_kbps(struct tc_tbf_qopt_offload_replace_params *p)
+{
+	/* TBF interface is in bytes/s, whereas Spectrum ASIC is configured in
+	 * Kbits/s.
+	 */
+	return p->rate.rate_bytes_ps / 1000 * 8;
+}
+
+static int
+mlxsw_sp_qdisc_tbf_check_params(struct mlxsw_sp_port *mlxsw_sp_port,
+				struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+				void *params)
+{
+	struct tc_tbf_qopt_offload_replace_params *p = params;
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	u64 rate_kbps = mlxsw_sp_qdisc_tbf_rate_kbps(p);
+	u8 burst_size;
+	int err;
+
+	if (rate_kbps >= MLXSW_REG_QEEC_MAS_DIS) {
+		dev_err(mlxsw_sp_port->mlxsw_sp->bus_info->dev,
+			"spectrum: TBF: rate of %lluKbps must be below %u\n",
+			rate_kbps, MLXSW_REG_QEEC_MAS_DIS);
+		return -EINVAL;
+	}
+
+	err = mlxsw_sp_qdisc_tbf_bs(mlxsw_sp_port, p->max_size, &burst_size);
+	if (err) {
+		u8 highest_shaper_bs = MLXSW_REG_QEEC_HIGHEST_SHAPER_BS;
+
+		dev_err(mlxsw_sp->bus_info->dev,
+			"spectrum: TBF: invalid burst size of %u, must be a power of two between %u and %u",
+			p->max_size,
+			mlxsw_sp_qdisc_tbf_max_size(mlxsw_sp->lowest_shaper_bs),
+			mlxsw_sp_qdisc_tbf_max_size(highest_shaper_bs));
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+mlxsw_sp_qdisc_tbf_replace(struct mlxsw_sp_port *mlxsw_sp_port,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			   void *params)
+{
+	struct tc_tbf_qopt_offload_replace_params *p = params;
+	u64 rate_kbps = mlxsw_sp_qdisc_tbf_rate_kbps(p);
+	u8 burst_size;
+	int err;
+
+	err = mlxsw_sp_qdisc_tbf_bs(mlxsw_sp_port, p->max_size, &burst_size);
+	if (WARN_ON_ONCE(err))
+		/* check_params above was supposed to reject this value. */
+		return -EINVAL;
+
+	/* Configure subgroup shaper, so that both UC and MC traffic is subject
+	 * to shaping. That is unlike RED, however UC queue lengths are going to
+	 * be different than MC ones due to different pool and quota
+	 * configurations, so the configuration is not applicable. For shaper on
+	 * the other hand, subjecting the overall stream to the configured
+	 * shaper makes sense. Also note that that is what we do for
+	 * ieee_setmaxrate().
+	 */
+	return mlxsw_sp_port_ets_maxrate_set(mlxsw_sp_port,
+					     MLXSW_REG_QEEC_HR_SUBGROUP,
+					     mlxsw_sp_qdisc->tclass_num, 0,
+					     rate_kbps, burst_size);
+}
+
+static void
+mlxsw_sp_qdisc_tbf_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
+			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			     void *params)
+{
+	struct tc_tbf_qopt_offload_replace_params *p = params;
+
+	mlxsw_sp_qdisc_leaf_unoffload(mlxsw_sp_port, mlxsw_sp_qdisc, p->qstats);
+}
+
+static int
+mlxsw_sp_qdisc_get_tbf_stats(struct mlxsw_sp_port *mlxsw_sp_port,
+			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			     struct tc_qopt_offload_stats *stats_ptr)
+{
+	mlxsw_sp_qdisc_get_tc_stats(mlxsw_sp_port, mlxsw_sp_qdisc,
+				    stats_ptr);
+	return 0;
+}
+
+static struct mlxsw_sp_qdisc_ops mlxsw_sp_qdisc_ops_tbf = {
+	.type = MLXSW_SP_QDISC_TBF,
+	.check_params = mlxsw_sp_qdisc_tbf_check_params,
+	.replace = mlxsw_sp_qdisc_tbf_replace,
+	.unoffload = mlxsw_sp_qdisc_tbf_unoffload,
+	.destroy = mlxsw_sp_qdisc_tbf_destroy,
+	.get_stats = mlxsw_sp_qdisc_get_tbf_stats,
+	.clean_stats = mlxsw_sp_setup_tc_qdisc_leaf_clean_stats,
+};
+
+int mlxsw_sp_setup_tc_tbf(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct tc_tbf_qopt_offload *p)
+{
+	struct mlxsw_sp_qdisc *mlxsw_sp_qdisc;
+
+	mlxsw_sp_qdisc = mlxsw_sp_qdisc_find(mlxsw_sp_port, p->parent, false);
+	if (!mlxsw_sp_qdisc)
+		return -EOPNOTSUPP;
+
+	if (p->command == TC_TBF_REPLACE)
+		return mlxsw_sp_qdisc_replace(mlxsw_sp_port, p->handle,
+					      mlxsw_sp_qdisc,
+					      &mlxsw_sp_qdisc_ops_tbf,
+					      &p->replace_params);
+
+	if (!mlxsw_sp_qdisc_compare(mlxsw_sp_qdisc, p->handle,
+				    MLXSW_SP_QDISC_TBF))
+		return -EOPNOTSUPP;
+
+	switch (p->command) {
+	case TC_TBF_DESTROY:
+		return mlxsw_sp_qdisc_destroy(mlxsw_sp_port, mlxsw_sp_qdisc);
+	case TC_TBF_STATS:
+		return mlxsw_sp_qdisc_get_stats(mlxsw_sp_port, mlxsw_sp_qdisc,
+						&p->stats);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static int
+__mlxsw_sp_qdisc_ets_destroy(struct mlxsw_sp_port *mlxsw_sp_port)
 {
 	int i;
 
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
 		mlxsw_sp_port_prio_tc_set(mlxsw_sp_port, i,
 					  MLXSW_SP_PORT_DEFAULT_TCLASS);
+		mlxsw_sp_port_ets_set(mlxsw_sp_port,
+				      MLXSW_REG_QEEC_HR_SUBGROUP,
+				      i, 0, false, 0);
 		mlxsw_sp_qdisc_destroy(mlxsw_sp_port,
 				       &mlxsw_sp_port->tclass_qdiscs[i]);
 		mlxsw_sp_port->tclass_qdiscs[i].prio_bitmap = 0;
@@ -504,36 +759,58 @@ mlxsw_sp_qdisc_prio_destroy(struct mlxsw_sp_port *mlxsw_sp_port,
 }
 
 static int
-mlxsw_sp_qdisc_prio_check_params(struct mlxsw_sp_port *mlxsw_sp_port,
-				 struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
-				 void *params)
+mlxsw_sp_qdisc_prio_destroy(struct mlxsw_sp_port *mlxsw_sp_port,
+			    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc)
 {
-	struct tc_prio_qopt_offload_params *p = params;
+	return __mlxsw_sp_qdisc_ets_destroy(mlxsw_sp_port);
+}
 
-	if (p->bands > IEEE_8021QAZ_MAX_TCS)
+static int
+__mlxsw_sp_qdisc_ets_check_params(unsigned int nbands)
+{
+	if (nbands > IEEE_8021QAZ_MAX_TCS)
 		return -EOPNOTSUPP;
 
 	return 0;
 }
 
 static int
-mlxsw_sp_qdisc_prio_replace(struct mlxsw_sp_port *mlxsw_sp_port,
-			    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
-			    void *params)
+mlxsw_sp_qdisc_prio_check_params(struct mlxsw_sp_port *mlxsw_sp_port,
+				 struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+				 void *params)
 {
 	struct tc_prio_qopt_offload_params *p = params;
+
+	return __mlxsw_sp_qdisc_ets_check_params(p->bands);
+}
+
+static int
+__mlxsw_sp_qdisc_ets_replace(struct mlxsw_sp_port *mlxsw_sp_port,
+			     unsigned int nbands,
+			     const unsigned int *quanta,
+			     const unsigned int *weights,
+			     const u8 *priomap)
+{
 	struct mlxsw_sp_qdisc *child_qdisc;
 	int tclass, i, band, backlog;
 	u8 old_priomap;
 	int err;
 
-	for (band = 0; band < p->bands; band++) {
+	for (band = 0; band < nbands; band++) {
 		tclass = MLXSW_SP_PRIO_BAND_TO_TCLASS(band);
 		child_qdisc = &mlxsw_sp_port->tclass_qdiscs[tclass];
 		old_priomap = child_qdisc->prio_bitmap;
 		child_qdisc->prio_bitmap = 0;
+
+		err = mlxsw_sp_port_ets_set(mlxsw_sp_port,
+					    MLXSW_REG_QEEC_HR_SUBGROUP,
+					    tclass, 0, !!quanta[band],
+					    weights[band]);
+		if (err)
+			return err;
+
 		for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
-			if (p->priomap[i] == band) {
+			if (priomap[i] == band) {
 				child_qdisc->prio_bitmap |= BIT(i);
 				if (BIT(i) & old_priomap)
 					continue;
@@ -556,21 +833,46 @@ mlxsw_sp_qdisc_prio_replace(struct mlxsw_sp_port *mlxsw_sp_port,
 		child_qdisc = &mlxsw_sp_port->tclass_qdiscs[tclass];
 		child_qdisc->prio_bitmap = 0;
 		mlxsw_sp_qdisc_destroy(mlxsw_sp_port, child_qdisc);
+		mlxsw_sp_port_ets_set(mlxsw_sp_port,
+				      MLXSW_REG_QEEC_HR_SUBGROUP,
+				      tclass, 0, false, 0);
 	}
 	return 0;
 }
 
+static int
+mlxsw_sp_qdisc_prio_replace(struct mlxsw_sp_port *mlxsw_sp_port,
+			    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			    void *params)
+{
+	struct tc_prio_qopt_offload_params *p = params;
+	unsigned int zeroes[TCQ_ETS_MAX_BANDS] = {0};
+
+	return __mlxsw_sp_qdisc_ets_replace(mlxsw_sp_port, p->bands,
+					    zeroes, zeroes, p->priomap);
+}
+
+static void
+__mlxsw_sp_qdisc_ets_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
+			       struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			       struct gnet_stats_queue *qstats)
+{
+	u64 backlog;
+
+	backlog = mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
+				       mlxsw_sp_qdisc->stats_base.backlog);
+	qstats->backlog -= backlog;
+}
+
 static void
 mlxsw_sp_qdisc_prio_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
 			      struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
 			      void *params)
 {
 	struct tc_prio_qopt_offload_params *p = params;
-	u64 backlog;
 
-	backlog = mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
-				       mlxsw_sp_qdisc->stats_base.backlog);
-	p->qstats->backlog -= backlog;
+	__mlxsw_sp_qdisc_ets_unoffload(mlxsw_sp_port, mlxsw_sp_qdisc,
+				       p->qstats);
 }
 
 static int
@@ -578,37 +880,23 @@ mlxsw_sp_qdisc_get_prio_stats(struct mlxsw_sp_port *mlxsw_sp_port,
 			      struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
 			      struct tc_qopt_offload_stats *stats_ptr)
 {
-	u64 tx_bytes, tx_packets, drops = 0, backlog = 0;
-	struct mlxsw_sp_qdisc_stats *stats_base;
-	struct mlxsw_sp_port_xstats *xstats;
-	struct rtnl_link_stats64 *stats;
+	struct mlxsw_sp_qdisc *tc_qdisc;
+	u64 tx_packets = 0;
+	u64 tx_bytes = 0;
+	u64 backlog = 0;
+	u64 drops = 0;
 	int i;
 
-	xstats = &mlxsw_sp_port->periodic_hw_stats.xstats;
-	stats = &mlxsw_sp_port->periodic_hw_stats.stats;
-	stats_base = &mlxsw_sp_qdisc->stats_base;
-
-	tx_bytes = stats->tx_bytes - stats_base->tx_bytes;
-	tx_packets = stats->tx_packets - stats_base->tx_packets;
-
 	for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++) {
-		drops += mlxsw_sp_xstats_tail_drop(xstats, i);
-		drops += xstats->wred_drop[i];
-		backlog += mlxsw_sp_xstats_backlog(xstats, i);
+		tc_qdisc = &mlxsw_sp_port->tclass_qdiscs[i];
+		mlxsw_sp_qdisc_collect_tc_stats(mlxsw_sp_port, tc_qdisc,
+						&tx_bytes, &tx_packets,
+						&drops, &backlog);
 	}
-	drops = drops - stats_base->drops;
 
-	_bstats_update(stats_ptr->bstats, tx_bytes, tx_packets);
-	stats_ptr->qstats->drops += drops;
-	stats_ptr->qstats->backlog +=
-				mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
-						     backlog) -
-				mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
-						     stats_base->backlog);
-	stats_base->backlog = backlog;
-	stats_base->drops += drops;
-	stats_base->tx_bytes += tx_bytes;
-	stats_base->tx_packets += tx_packets;
+	mlxsw_sp_qdisc_update_stats(mlxsw_sp_port->mlxsw_sp, mlxsw_sp_qdisc,
+				    tx_bytes, tx_packets, drops, backlog,
+				    stats_ptr);
 	return 0;
 }
 
@@ -647,27 +935,93 @@ static struct mlxsw_sp_qdisc_ops mlxsw_sp_qdisc_ops_prio = {
 	.clean_stats = mlxsw_sp_setup_tc_qdisc_prio_clean_stats,
 };
 
-/* Grafting is not supported in mlxsw. It will result in un-offloading of the
- * grafted qdisc as well as the qdisc in the qdisc new location.
- * (However, if the graft is to the location where the qdisc is already at, it
- * will be ignored completely and won't cause un-offloading).
+static int
+mlxsw_sp_qdisc_ets_check_params(struct mlxsw_sp_port *mlxsw_sp_port,
+				struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+				void *params)
+{
+	struct tc_ets_qopt_offload_replace_params *p = params;
+
+	return __mlxsw_sp_qdisc_ets_check_params(p->bands);
+}
+
+static int
+mlxsw_sp_qdisc_ets_replace(struct mlxsw_sp_port *mlxsw_sp_port,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			   void *params)
+{
+	struct tc_ets_qopt_offload_replace_params *p = params;
+
+	return __mlxsw_sp_qdisc_ets_replace(mlxsw_sp_port, p->bands,
+					    p->quanta, p->weights, p->priomap);
+}
+
+static void
+mlxsw_sp_qdisc_ets_unoffload(struct mlxsw_sp_port *mlxsw_sp_port,
+			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			     void *params)
+{
+	struct tc_ets_qopt_offload_replace_params *p = params;
+
+	__mlxsw_sp_qdisc_ets_unoffload(mlxsw_sp_port, mlxsw_sp_qdisc,
+				       p->qstats);
+}
+
+static int
+mlxsw_sp_qdisc_ets_destroy(struct mlxsw_sp_port *mlxsw_sp_port,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc)
+{
+	return __mlxsw_sp_qdisc_ets_destroy(mlxsw_sp_port);
+}
+
+static struct mlxsw_sp_qdisc_ops mlxsw_sp_qdisc_ops_ets = {
+	.type = MLXSW_SP_QDISC_ETS,
+	.check_params = mlxsw_sp_qdisc_ets_check_params,
+	.replace = mlxsw_sp_qdisc_ets_replace,
+	.unoffload = mlxsw_sp_qdisc_ets_unoffload,
+	.destroy = mlxsw_sp_qdisc_ets_destroy,
+	.get_stats = mlxsw_sp_qdisc_get_prio_stats,
+	.clean_stats = mlxsw_sp_setup_tc_qdisc_prio_clean_stats,
+};
+
+/* Linux allows linking of Qdiscs to arbitrary classes (so long as the resulting
+ * graph is free of cycles). These operations do not change the parent handle
+ * though, which means it can be incomplete (if there is more than one class
+ * where the Qdisc in question is grafted) or outright wrong (if the Qdisc was
+ * linked to a different class and then removed from the original class).
+ *
+ * E.g. consider this sequence of operations:
+ *
+ *  # tc qdisc add dev swp1 root handle 1: prio
+ *  # tc qdisc add dev swp1 parent 1:3 handle 13: red limit 1000000 avpkt 10000
+ *  RED: set bandwidth to 10Mbit
+ *  # tc qdisc link dev swp1 handle 13: parent 1:2
+ *
+ * At this point, both 1:2 and 1:3 have the same RED Qdisc instance as their
+ * child. But RED will still only claim that 1:3 is its parent. If it's removed
+ * from that band, its only parent will be 1:2, but it will continue to claim
+ * that it is in fact 1:3.
+ *
+ * The notification for child Qdisc replace (e.g. TC_RED_REPLACE) comes before
+ * the notification for parent graft (e.g. TC_PRIO_GRAFT). We take the replace
+ * notification to offload the child Qdisc, based on its parent handle, and use
+ * the graft operation to validate that the class where the child is actually
+ * grafted corresponds to the parent handle. If the two don't match, we
+ * unoffload the child.
  */
 static int
-mlxsw_sp_qdisc_prio_graft(struct mlxsw_sp_port *mlxsw_sp_port,
-			  struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
-			  struct tc_prio_qopt_offload_graft_params *p)
+__mlxsw_sp_qdisc_ets_graft(struct mlxsw_sp_port *mlxsw_sp_port,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			   u8 band, u32 child_handle)
 {
-	int tclass_num = MLXSW_SP_PRIO_BAND_TO_TCLASS(p->band);
+	int tclass_num = MLXSW_SP_PRIO_BAND_TO_TCLASS(band);
 	struct mlxsw_sp_qdisc *old_qdisc;
 
-	/* Check if the grafted qdisc is already in its "new" location. If so -
-	 * nothing needs to be done.
-	 */
-	if (p->band < IEEE_8021QAZ_MAX_TCS &&
-	    mlxsw_sp_port->tclass_qdiscs[tclass_num].handle == p->child_handle)
+	if (band < IEEE_8021QAZ_MAX_TCS &&
+	    mlxsw_sp_port->tclass_qdiscs[tclass_num].handle == child_handle)
 		return 0;
 
-	if (!p->child_handle) {
+	if (!child_handle) {
 		/* This is an invisible FIFO replacing the original Qdisc.
 		 * Ignore it--the original Qdisc's destroy will follow.
 		 */
@@ -678,7 +1032,7 @@ mlxsw_sp_qdisc_prio_graft(struct mlxsw_sp_port *mlxsw_sp_port,
 	 * unoffload it.
 	 */
 	old_qdisc = mlxsw_sp_qdisc_find_by_handle(mlxsw_sp_port,
-						  p->child_handle);
+						  child_handle);
 	if (old_qdisc)
 		mlxsw_sp_qdisc_destroy(mlxsw_sp_port, old_qdisc);
 
@@ -687,6 +1041,15 @@ mlxsw_sp_qdisc_prio_graft(struct mlxsw_sp_port *mlxsw_sp_port,
 	return -EOPNOTSUPP;
 }
 
+static int
+mlxsw_sp_qdisc_prio_graft(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			  struct tc_prio_qopt_offload_graft_params *p)
+{
+	return __mlxsw_sp_qdisc_ets_graft(mlxsw_sp_port, mlxsw_sp_qdisc,
+					  p->band, p->child_handle);
+}
+
 int mlxsw_sp_setup_tc_prio(struct mlxsw_sp_port *mlxsw_sp_port,
 			   struct tc_prio_qopt_offload *p)
 {
@@ -720,6 +1083,40 @@ int mlxsw_sp_setup_tc_prio(struct mlxsw_sp_port *mlxsw_sp_port,
 	}
 }
 
+int mlxsw_sp_setup_tc_ets(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct tc_ets_qopt_offload *p)
+{
+	struct mlxsw_sp_qdisc *mlxsw_sp_qdisc;
+
+	mlxsw_sp_qdisc = mlxsw_sp_qdisc_find(mlxsw_sp_port, p->parent, true);
+	if (!mlxsw_sp_qdisc)
+		return -EOPNOTSUPP;
+
+	if (p->command == TC_ETS_REPLACE)
+		return mlxsw_sp_qdisc_replace(mlxsw_sp_port, p->handle,
+					      mlxsw_sp_qdisc,
+					      &mlxsw_sp_qdisc_ops_ets,
+					      &p->replace_params);
+
+	if (!mlxsw_sp_qdisc_compare(mlxsw_sp_qdisc, p->handle,
+				    MLXSW_SP_QDISC_ETS))
+		return -EOPNOTSUPP;
+
+	switch (p->command) {
+	case TC_ETS_DESTROY:
+		return mlxsw_sp_qdisc_destroy(mlxsw_sp_port, mlxsw_sp_qdisc);
+	case TC_ETS_STATS:
+		return mlxsw_sp_qdisc_get_stats(mlxsw_sp_port, mlxsw_sp_qdisc,
+						&p->stats);
+	case TC_ETS_GRAFT:
+		return __mlxsw_sp_qdisc_ets_graft(mlxsw_sp_port, mlxsw_sp_qdisc,
+						  p->graft_params.band,
+						  p->graft_params.child_handle);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 int mlxsw_sp_tc_qdisc_init(struct mlxsw_sp_port *mlxsw_sp_port)
 {
 	struct mlxsw_sp_qdisc *mlxsw_sp_qdisc;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
index 8290e82240fc..ce707723f8cf 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c
@@ -382,9 +382,10 @@ enum mlxsw_sp_fib_entry_type {
 };
 
 struct mlxsw_sp_nexthop_group;
+struct mlxsw_sp_fib_entry;
 
 struct mlxsw_sp_fib_node {
-	struct list_head entry_list;
+	struct mlxsw_sp_fib_entry *fib_entry;
 	struct list_head list;
 	struct rhash_head ht_node;
 	struct mlxsw_sp_fib *fib;
@@ -397,7 +398,6 @@ struct mlxsw_sp_fib_entry_decap {
 };
 
 struct mlxsw_sp_fib_entry {
-	struct list_head list;
 	struct mlxsw_sp_fib_node *fib_node;
 	enum mlxsw_sp_fib_entry_type type;
 	struct list_head nexthop_group_node;
@@ -1162,7 +1162,6 @@ mlxsw_sp_router_ip2me_fib_entry_find(struct mlxsw_sp *mlxsw_sp, u32 tb_id,
 				     const union mlxsw_sp_l3addr *addr,
 				     enum mlxsw_sp_fib_entry_type type)
 {
-	struct mlxsw_sp_fib_entry *fib_entry;
 	struct mlxsw_sp_fib_node *fib_node;
 	unsigned char addr_prefix_len;
 	struct mlxsw_sp_fib *fib;
@@ -1191,15 +1190,10 @@ mlxsw_sp_router_ip2me_fib_entry_find(struct mlxsw_sp *mlxsw_sp, u32 tb_id,
 
 	fib_node = mlxsw_sp_fib_node_lookup(fib, addrp, addr_len,
 					    addr_prefix_len);
-	if (!fib_node || list_empty(&fib_node->entry_list))
+	if (!fib_node || fib_node->fib_entry->type != type)
 		return NULL;
 
-	fib_entry = list_first_entry(&fib_node->entry_list,
-				     struct mlxsw_sp_fib_entry, list);
-	if (fib_entry->type != type)
-		return NULL;
-
-	return fib_entry;
+	return fib_node->fib_entry;
 }
 
 /* Given an IPIP entry, find the corresponding decap route. */
@@ -1209,7 +1203,6 @@ mlxsw_sp_ipip_entry_find_decap(struct mlxsw_sp *mlxsw_sp,
 {
 	static struct mlxsw_sp_fib_node *fib_node;
 	const struct mlxsw_sp_ipip_ops *ipip_ops;
-	struct mlxsw_sp_fib_entry *fib_entry;
 	unsigned char saddr_prefix_len;
 	union mlxsw_sp_l3addr saddr;
 	struct mlxsw_sp_fib *ul_fib;
@@ -1244,15 +1237,11 @@ mlxsw_sp_ipip_entry_find_decap(struct mlxsw_sp *mlxsw_sp,
 
 	fib_node = mlxsw_sp_fib_node_lookup(ul_fib, saddrp, saddr_len,
 					    saddr_prefix_len);
-	if (!fib_node || list_empty(&fib_node->entry_list))
+	if (!fib_node ||
+	    fib_node->fib_entry->type != MLXSW_SP_FIB_ENTRY_TYPE_TRAP)
 		return NULL;
 
-	fib_entry = list_first_entry(&fib_node->entry_list,
-				     struct mlxsw_sp_fib_entry, list);
-	if (fib_entry->type != MLXSW_SP_FIB_ENTRY_TYPE_TRAP)
-		return NULL;
-
-	return fib_entry;
+	return fib_node->fib_entry;
 }
 
 static struct mlxsw_sp_ipip_entry *
@@ -3231,10 +3220,6 @@ mlxsw_sp_nexthop_group_update(struct mlxsw_sp *mlxsw_sp,
 	return 0;
 }
 
-static bool
-mlxsw_sp_fib_node_entry_is_first(const struct mlxsw_sp_fib_node *fib_node,
-				 const struct mlxsw_sp_fib_entry *fib_entry);
-
 static int
 mlxsw_sp_nexthop_fib_entries_update(struct mlxsw_sp *mlxsw_sp,
 				    struct mlxsw_sp_nexthop_group *nh_grp)
@@ -3243,9 +3228,6 @@ mlxsw_sp_nexthop_fib_entries_update(struct mlxsw_sp *mlxsw_sp,
 	int err;
 
 	list_for_each_entry(fib_entry, &nh_grp->fib_list, nexthop_group_node) {
-		if (!mlxsw_sp_fib_node_entry_is_first(fib_entry->fib_node,
-						      fib_entry))
-			continue;
 		err = mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
 		if (err)
 			return err;
@@ -3253,24 +3235,6 @@ mlxsw_sp_nexthop_fib_entries_update(struct mlxsw_sp *mlxsw_sp,
 	return 0;
 }
 
-static void
-mlxsw_sp_fib_entry_offload_refresh(struct mlxsw_sp_fib_entry *fib_entry,
-				   enum mlxsw_reg_ralue_op op, int err);
-
-static void
-mlxsw_sp_nexthop_fib_entries_refresh(struct mlxsw_sp_nexthop_group *nh_grp)
-{
-	enum mlxsw_reg_ralue_op op = MLXSW_REG_RALUE_OP_WRITE_WRITE;
-	struct mlxsw_sp_fib_entry *fib_entry;
-
-	list_for_each_entry(fib_entry, &nh_grp->fib_list, nexthop_group_node) {
-		if (!mlxsw_sp_fib_node_entry_is_first(fib_entry->fib_node,
-						      fib_entry))
-			continue;
-		mlxsw_sp_fib_entry_offload_refresh(fib_entry, op, 0);
-	}
-}
-
 static void mlxsw_sp_adj_grp_size_round_up(u16 *p_adj_grp_size)
 {
 	/* Valid sizes for an adjacency group are:
@@ -3374,6 +3338,73 @@ mlxsw_sp_nexthop_group_rebalance(struct mlxsw_sp_nexthop_group *nh_grp)
 	}
 }
 
+static struct mlxsw_sp_nexthop *
+mlxsw_sp_rt6_nexthop(struct mlxsw_sp_nexthop_group *nh_grp,
+		     const struct mlxsw_sp_rt6 *mlxsw_sp_rt6);
+
+static void
+mlxsw_sp_nexthop4_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
+					struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	int i;
+
+	for (i = 0; i < nh_grp->count; i++) {
+		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
+
+		if (nh->offloaded)
+			nh->key.fib_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
+		else
+			nh->key.fib_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
+	}
+}
+
+static void
+__mlxsw_sp_nexthop6_group_offload_refresh(struct mlxsw_sp_nexthop_group *nh_grp,
+					  struct mlxsw_sp_fib6_entry *fib6_entry)
+{
+	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
+
+	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
+		struct fib6_nh *fib6_nh = mlxsw_sp_rt6->rt->fib6_nh;
+		struct mlxsw_sp_nexthop *nh;
+
+		nh = mlxsw_sp_rt6_nexthop(nh_grp, mlxsw_sp_rt6);
+		if (nh && nh->offloaded)
+			fib6_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
+		else
+			fib6_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
+	}
+}
+
+static void
+mlxsw_sp_nexthop6_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
+					struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	struct mlxsw_sp_fib6_entry *fib6_entry;
+
+	/* Unfortunately, in IPv6 the route and the nexthop are described by
+	 * the same struct, so we need to iterate over all the routes using the
+	 * nexthop group and set / clear the offload indication for them.
+	 */
+	list_for_each_entry(fib6_entry, &nh_grp->fib_list,
+			    common.nexthop_group_node)
+		__mlxsw_sp_nexthop6_group_offload_refresh(nh_grp, fib6_entry);
+}
+
+static void
+mlxsw_sp_nexthop_group_offload_refresh(struct mlxsw_sp *mlxsw_sp,
+				       struct mlxsw_sp_nexthop_group *nh_grp)
+{
+	switch (mlxsw_sp_nexthop_group_type(nh_grp)) {
+	case AF_INET:
+		mlxsw_sp_nexthop4_group_offload_refresh(mlxsw_sp, nh_grp);
+		break;
+	case AF_INET6:
+		mlxsw_sp_nexthop6_group_offload_refresh(mlxsw_sp, nh_grp);
+		break;
+	}
+}
+
 static void
 mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 			       struct mlxsw_sp_nexthop_group *nh_grp)
@@ -3447,6 +3478,8 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 		goto set_trap;
 	}
 
+	mlxsw_sp_nexthop_group_offload_refresh(mlxsw_sp, nh_grp);
+
 	if (!old_adj_index_valid) {
 		/* The trap was set for fib entries, so we have to call
 		 * fib entry update to unset it and use adjacency index.
@@ -3468,9 +3501,6 @@ mlxsw_sp_nexthop_group_refresh(struct mlxsw_sp *mlxsw_sp,
 		goto set_trap;
 	}
 
-	/* Offload state within the group changed, so update the flags. */
-	mlxsw_sp_nexthop_fib_entries_refresh(nh_grp);
-
 	return;
 
 set_trap:
@@ -3483,6 +3513,7 @@ set_trap:
 	err = mlxsw_sp_nexthop_fib_entries_update(mlxsw_sp, nh_grp);
 	if (err)
 		dev_warn(mlxsw_sp->bus_info->dev, "Failed to set traps for fib entries.\n");
+	mlxsw_sp_nexthop_group_offload_refresh(mlxsw_sp, nh_grp);
 	if (old_adj_index_valid)
 		mlxsw_sp_kvdl_free(mlxsw_sp, MLXSW_SP_KVDL_ENTRY_TYPE_ADJ,
 				   nh_grp->ecmp_size, nh_grp->adj_index);
@@ -3845,7 +3876,7 @@ static void mlxsw_sp_nexthop4_event(struct mlxsw_sp *mlxsw_sp,
 
 	key.fib_nh = fib_nh;
 	nh = mlxsw_sp_nexthop_lookup(mlxsw_sp, key);
-	if (WARN_ON_ONCE(!nh))
+	if (!nh)
 		return;
 
 	switch (event) {
@@ -4065,131 +4096,128 @@ mlxsw_sp_rt6_nexthop(struct mlxsw_sp_nexthop_group *nh_grp,
 }
 
 static void
-mlxsw_sp_fib4_entry_offload_set(struct mlxsw_sp_fib_entry *fib_entry)
+mlxsw_sp_fib4_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
+				 struct mlxsw_sp_fib_entry *fib_entry)
 {
-	struct mlxsw_sp_nexthop_group *nh_grp = fib_entry->nh_group;
-	int i;
-
-	if (fib_entry->type == MLXSW_SP_FIB_ENTRY_TYPE_LOCAL ||
-	    fib_entry->type == MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE ||
-	    fib_entry->type == MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP ||
-	    fib_entry->type == MLXSW_SP_FIB_ENTRY_TYPE_NVE_DECAP) {
-		nh_grp->nexthops->key.fib_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
-		return;
-	}
-
-	for (i = 0; i < nh_grp->count; i++) {
-		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
+	struct fib_info *fi = mlxsw_sp_nexthop4_group_fi(fib_entry->nh_group);
+	u32 *p_dst = (u32 *) fib_entry->fib_node->key.addr;
+	int dst_len = fib_entry->fib_node->key.prefix_len;
+	struct mlxsw_sp_fib4_entry *fib4_entry;
+	struct fib_rt_info fri;
+	bool should_offload;
 
-		if (nh->offloaded)
-			nh->key.fib_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
-		else
-			nh->key.fib_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
-	}
+	should_offload = mlxsw_sp_fib_entry_should_offload(fib_entry);
+	fib4_entry = container_of(fib_entry, struct mlxsw_sp_fib4_entry,
+				  common);
+	fri.fi = fi;
+	fri.tb_id = fib4_entry->tb_id;
+	fri.dst = cpu_to_be32(*p_dst);
+	fri.dst_len = dst_len;
+	fri.tos = fib4_entry->tos;
+	fri.type = fib4_entry->type;
+	fri.offload = should_offload;
+	fri.trap = !should_offload;
+	fib_alias_hw_flags_set(mlxsw_sp_net(mlxsw_sp), &fri);
 }
 
 static void
-mlxsw_sp_fib4_entry_offload_unset(struct mlxsw_sp_fib_entry *fib_entry)
+mlxsw_sp_fib4_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
+				   struct mlxsw_sp_fib_entry *fib_entry)
 {
-	struct mlxsw_sp_nexthop_group *nh_grp = fib_entry->nh_group;
-	int i;
-
-	if (!list_is_singular(&nh_grp->fib_list))
-		return;
-
-	for (i = 0; i < nh_grp->count; i++) {
-		struct mlxsw_sp_nexthop *nh = &nh_grp->nexthops[i];
+	struct fib_info *fi = mlxsw_sp_nexthop4_group_fi(fib_entry->nh_group);
+	u32 *p_dst = (u32 *) fib_entry->fib_node->key.addr;
+	int dst_len = fib_entry->fib_node->key.prefix_len;
+	struct mlxsw_sp_fib4_entry *fib4_entry;
+	struct fib_rt_info fri;
 
-		nh->key.fib_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
-	}
+	fib4_entry = container_of(fib_entry, struct mlxsw_sp_fib4_entry,
+				  common);
+	fri.fi = fi;
+	fri.tb_id = fib4_entry->tb_id;
+	fri.dst = cpu_to_be32(*p_dst);
+	fri.dst_len = dst_len;
+	fri.tos = fib4_entry->tos;
+	fri.type = fib4_entry->type;
+	fri.offload = false;
+	fri.trap = false;
+	fib_alias_hw_flags_set(mlxsw_sp_net(mlxsw_sp), &fri);
 }
 
 static void
-mlxsw_sp_fib6_entry_offload_set(struct mlxsw_sp_fib_entry *fib_entry)
+mlxsw_sp_fib6_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
+				 struct mlxsw_sp_fib_entry *fib_entry)
 {
 	struct mlxsw_sp_fib6_entry *fib6_entry;
 	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
+	bool should_offload;
+
+	should_offload = mlxsw_sp_fib_entry_should_offload(fib_entry);
 
+	/* In IPv6 a multipath route is represented using multiple routes, so
+	 * we need to set the flags on all of them.
+	 */
 	fib6_entry = container_of(fib_entry, struct mlxsw_sp_fib6_entry,
 				  common);
-
-	if (fib_entry->type == MLXSW_SP_FIB_ENTRY_TYPE_LOCAL ||
-	    fib_entry->type == MLXSW_SP_FIB_ENTRY_TYPE_BLACKHOLE) {
-		list_first_entry(&fib6_entry->rt6_list, struct mlxsw_sp_rt6,
-				 list)->rt->fib6_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
-		return;
-	}
-
-	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
-		struct mlxsw_sp_nexthop_group *nh_grp = fib_entry->nh_group;
-		struct fib6_nh *fib6_nh = mlxsw_sp_rt6->rt->fib6_nh;
-		struct mlxsw_sp_nexthop *nh;
-
-		nh = mlxsw_sp_rt6_nexthop(nh_grp, mlxsw_sp_rt6);
-		if (nh && nh->offloaded)
-			fib6_nh->fib_nh_flags |= RTNH_F_OFFLOAD;
-		else
-			fib6_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
-	}
+	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list)
+		fib6_info_hw_flags_set(mlxsw_sp_rt6->rt, should_offload,
+				       !should_offload);
 }
 
 static void
-mlxsw_sp_fib6_entry_offload_unset(struct mlxsw_sp_fib_entry *fib_entry)
+mlxsw_sp_fib6_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
+				   struct mlxsw_sp_fib_entry *fib_entry)
 {
 	struct mlxsw_sp_fib6_entry *fib6_entry;
 	struct mlxsw_sp_rt6 *mlxsw_sp_rt6;
 
 	fib6_entry = container_of(fib_entry, struct mlxsw_sp_fib6_entry,
 				  common);
-	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list) {
-		struct fib6_info *rt = mlxsw_sp_rt6->rt;
-
-		rt->fib6_nh->fib_nh_flags &= ~RTNH_F_OFFLOAD;
-	}
+	list_for_each_entry(mlxsw_sp_rt6, &fib6_entry->rt6_list, list)
+		fib6_info_hw_flags_set(mlxsw_sp_rt6->rt, false, false);
 }
 
-static void mlxsw_sp_fib_entry_offload_set(struct mlxsw_sp_fib_entry *fib_entry)
+static void
+mlxsw_sp_fib_entry_hw_flags_set(struct mlxsw_sp *mlxsw_sp,
+				struct mlxsw_sp_fib_entry *fib_entry)
 {
 	switch (fib_entry->fib_node->fib->proto) {
 	case MLXSW_SP_L3_PROTO_IPV4:
-		mlxsw_sp_fib4_entry_offload_set(fib_entry);
+		mlxsw_sp_fib4_entry_hw_flags_set(mlxsw_sp, fib_entry);
 		break;
 	case MLXSW_SP_L3_PROTO_IPV6:
-		mlxsw_sp_fib6_entry_offload_set(fib_entry);
+		mlxsw_sp_fib6_entry_hw_flags_set(mlxsw_sp, fib_entry);
 		break;
 	}
 }
 
 static void
-mlxsw_sp_fib_entry_offload_unset(struct mlxsw_sp_fib_entry *fib_entry)
+mlxsw_sp_fib_entry_hw_flags_clear(struct mlxsw_sp *mlxsw_sp,
+				  struct mlxsw_sp_fib_entry *fib_entry)
 {
 	switch (fib_entry->fib_node->fib->proto) {
 	case MLXSW_SP_L3_PROTO_IPV4:
-		mlxsw_sp_fib4_entry_offload_unset(fib_entry);
+		mlxsw_sp_fib4_entry_hw_flags_clear(mlxsw_sp, fib_entry);
 		break;
 	case MLXSW_SP_L3_PROTO_IPV6:
-		mlxsw_sp_fib6_entry_offload_unset(fib_entry);
+		mlxsw_sp_fib6_entry_hw_flags_clear(mlxsw_sp, fib_entry);
 		break;
 	}
 }
 
 static void
-mlxsw_sp_fib_entry_offload_refresh(struct mlxsw_sp_fib_entry *fib_entry,
-				   enum mlxsw_reg_ralue_op op, int err)
+mlxsw_sp_fib_entry_hw_flags_refresh(struct mlxsw_sp *mlxsw_sp,
+				    struct mlxsw_sp_fib_entry *fib_entry,
+				    enum mlxsw_reg_ralue_op op)
 {
 	switch (op) {
-	case MLXSW_REG_RALUE_OP_WRITE_DELETE:
-		return mlxsw_sp_fib_entry_offload_unset(fib_entry);
 	case MLXSW_REG_RALUE_OP_WRITE_WRITE:
-		if (err)
-			return;
-		if (mlxsw_sp_fib_entry_should_offload(fib_entry))
-			mlxsw_sp_fib_entry_offload_set(fib_entry);
-		else
-			mlxsw_sp_fib_entry_offload_unset(fib_entry);
-		return;
+		mlxsw_sp_fib_entry_hw_flags_set(mlxsw_sp, fib_entry);
+		break;
+	case MLXSW_REG_RALUE_OP_WRITE_DELETE:
+		mlxsw_sp_fib_entry_hw_flags_clear(mlxsw_sp, fib_entry);
+		break;
 	default:
-		return;
+		break;
 	}
 }
 
@@ -4416,7 +4444,10 @@ static int mlxsw_sp_fib_entry_op(struct mlxsw_sp *mlxsw_sp,
 {
 	int err = __mlxsw_sp_fib_entry_op(mlxsw_sp, fib_entry, op);
 
-	mlxsw_sp_fib_entry_offload_refresh(fib_entry, op, err);
+	if (err)
+		return err;
+
+	mlxsw_sp_fib_entry_hw_flags_refresh(mlxsw_sp, fib_entry, op);
 
 	return err;
 }
@@ -4491,6 +4522,19 @@ mlxsw_sp_fib4_entry_type_set(struct mlxsw_sp *mlxsw_sp,
 	}
 }
 
+static void
+mlxsw_sp_fib4_entry_type_unset(struct mlxsw_sp *mlxsw_sp,
+			       struct mlxsw_sp_fib_entry *fib_entry)
+{
+	switch (fib_entry->type) {
+	case MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP:
+		mlxsw_sp_fib_entry_decap_fini(mlxsw_sp, fib_entry);
+		break;
+	default:
+		break;
+	}
+}
+
 static struct mlxsw_sp_fib4_entry *
 mlxsw_sp_fib4_entry_create(struct mlxsw_sp *mlxsw_sp,
 			   struct mlxsw_sp_fib_node *fib_node,
@@ -4523,6 +4567,7 @@ mlxsw_sp_fib4_entry_create(struct mlxsw_sp *mlxsw_sp,
 	return fib4_entry;
 
 err_nexthop4_group_get:
+	mlxsw_sp_fib4_entry_type_unset(mlxsw_sp, fib_entry);
 err_fib4_entry_type_set:
 	kfree(fib4_entry);
 	return ERR_PTR(err);
@@ -4532,6 +4577,7 @@ static void mlxsw_sp_fib4_entry_destroy(struct mlxsw_sp *mlxsw_sp,
 					struct mlxsw_sp_fib4_entry *fib4_entry)
 {
 	mlxsw_sp_nexthop4_group_put(mlxsw_sp, &fib4_entry->common);
+	mlxsw_sp_fib4_entry_type_unset(mlxsw_sp, &fib4_entry->common);
 	kfree(fib4_entry);
 }
 
@@ -4555,15 +4601,14 @@ mlxsw_sp_fib4_entry_lookup(struct mlxsw_sp *mlxsw_sp,
 	if (!fib_node)
 		return NULL;
 
-	list_for_each_entry(fib4_entry, &fib_node->entry_list, common.list) {
-		if (fib4_entry->tb_id == fen_info->tb_id &&
-		    fib4_entry->tos == fen_info->tos &&
-		    fib4_entry->type == fen_info->type &&
-		    mlxsw_sp_nexthop4_group_fi(fib4_entry->common.nh_group) ==
-		    fen_info->fi) {
-			return fib4_entry;
-		}
-	}
+	fib4_entry = container_of(fib_node->fib_entry,
+				  struct mlxsw_sp_fib4_entry, common);
+	if (fib4_entry->tb_id == fen_info->tb_id &&
+	    fib4_entry->tos == fen_info->tos &&
+	    fib4_entry->type == fen_info->type &&
+	    mlxsw_sp_nexthop4_group_fi(fib4_entry->common.nh_group) ==
+	    fen_info->fi)
+		return fib4_entry;
 
 	return NULL;
 }
@@ -4611,7 +4656,6 @@ mlxsw_sp_fib_node_create(struct mlxsw_sp_fib *fib, const void *addr,
 	if (!fib_node)
 		return NULL;
 
-	INIT_LIST_HEAD(&fib_node->entry_list);
 	list_add(&fib_node->list, &fib->node_list);
 	memcpy(fib_node->key.addr, addr, addr_len);
 	fib_node->key.prefix_len = prefix_len;
@@ -4622,18 +4666,9 @@ mlxsw_sp_fib_node_create(struct mlxsw_sp_fib *fib, const void *addr,
 static void mlxsw_sp_fib_node_destroy(struct mlxsw_sp_fib_node *fib_node)
 {
 	list_del(&fib_node->list);
-	WARN_ON(!list_empty(&fib_node->entry_list));
 	kfree(fib_node);
 }
 
-static bool
-mlxsw_sp_fib_node_entry_is_first(const struct mlxsw_sp_fib_node *fib_node,
-				 const struct mlxsw_sp_fib_entry *fib_entry)
-{
-	return list_first_entry(&fib_node->entry_list,
-				struct mlxsw_sp_fib_entry, list) == fib_entry;
-}
-
 static int mlxsw_sp_fib_lpm_tree_link(struct mlxsw_sp *mlxsw_sp,
 				      struct mlxsw_sp_fib_node *fib_node)
 {
@@ -4773,200 +4808,48 @@ static void mlxsw_sp_fib_node_put(struct mlxsw_sp *mlxsw_sp,
 {
 	struct mlxsw_sp_vr *vr = fib_node->fib->vr;
 
-	if (!list_empty(&fib_node->entry_list))
+	if (fib_node->fib_entry)
 		return;
 	mlxsw_sp_fib_node_fini(mlxsw_sp, fib_node);
 	mlxsw_sp_fib_node_destroy(fib_node);
 	mlxsw_sp_vr_put(mlxsw_sp, vr);
 }
 
-static struct mlxsw_sp_fib4_entry *
-mlxsw_sp_fib4_node_entry_find(const struct mlxsw_sp_fib_node *fib_node,
-			      const struct mlxsw_sp_fib4_entry *new4_entry)
-{
-	struct mlxsw_sp_fib4_entry *fib4_entry;
-
-	list_for_each_entry(fib4_entry, &fib_node->entry_list, common.list) {
-		if (fib4_entry->tb_id > new4_entry->tb_id)
-			continue;
-		if (fib4_entry->tb_id != new4_entry->tb_id)
-			break;
-		if (fib4_entry->tos > new4_entry->tos)
-			continue;
-		if (fib4_entry->prio >= new4_entry->prio ||
-		    fib4_entry->tos < new4_entry->tos)
-			return fib4_entry;
-	}
-
-	return NULL;
-}
-
-static int
-mlxsw_sp_fib4_node_list_append(struct mlxsw_sp_fib4_entry *fib4_entry,
-			       struct mlxsw_sp_fib4_entry *new4_entry)
-{
-	struct mlxsw_sp_fib_node *fib_node;
-
-	if (WARN_ON(!fib4_entry))
-		return -EINVAL;
-
-	fib_node = fib4_entry->common.fib_node;
-	list_for_each_entry_from(fib4_entry, &fib_node->entry_list,
-				 common.list) {
-		if (fib4_entry->tb_id != new4_entry->tb_id ||
-		    fib4_entry->tos != new4_entry->tos ||
-		    fib4_entry->prio != new4_entry->prio)
-			break;
-	}
-
-	list_add_tail(&new4_entry->common.list, &fib4_entry->common.list);
-	return 0;
-}
-
-static int
-mlxsw_sp_fib4_node_list_insert(struct mlxsw_sp_fib4_entry *new4_entry,
-			       bool replace, bool append)
-{
-	struct mlxsw_sp_fib_node *fib_node = new4_entry->common.fib_node;
-	struct mlxsw_sp_fib4_entry *fib4_entry;
-
-	fib4_entry = mlxsw_sp_fib4_node_entry_find(fib_node, new4_entry);
-
-	if (append)
-		return mlxsw_sp_fib4_node_list_append(fib4_entry, new4_entry);
-	if (replace && WARN_ON(!fib4_entry))
-		return -EINVAL;
-
-	/* Insert new entry before replaced one, so that we can later
-	 * remove the second.
-	 */
-	if (fib4_entry) {
-		list_add_tail(&new4_entry->common.list,
-			      &fib4_entry->common.list);
-	} else {
-		struct mlxsw_sp_fib4_entry *last;
-
-		list_for_each_entry(last, &fib_node->entry_list, common.list) {
-			if (new4_entry->tb_id > last->tb_id)
-				break;
-			fib4_entry = last;
-		}
-
-		if (fib4_entry)
-			list_add(&new4_entry->common.list,
-				 &fib4_entry->common.list);
-		else
-			list_add(&new4_entry->common.list,
-				 &fib_node->entry_list);
-	}
-
-	return 0;
-}
-
-static void
-mlxsw_sp_fib4_node_list_remove(struct mlxsw_sp_fib4_entry *fib4_entry)
-{
-	list_del(&fib4_entry->common.list);
-}
-
-static int mlxsw_sp_fib_node_entry_add(struct mlxsw_sp *mlxsw_sp,
-				       struct mlxsw_sp_fib_entry *fib_entry)
-{
-	struct mlxsw_sp_fib_node *fib_node = fib_entry->fib_node;
-
-	if (!mlxsw_sp_fib_node_entry_is_first(fib_node, fib_entry))
-		return 0;
-
-	/* To prevent packet loss, overwrite the previously offloaded
-	 * entry.
-	 */
-	if (!list_is_singular(&fib_node->entry_list)) {
-		enum mlxsw_reg_ralue_op op = MLXSW_REG_RALUE_OP_WRITE_DELETE;
-		struct mlxsw_sp_fib_entry *n = list_next_entry(fib_entry, list);
-
-		mlxsw_sp_fib_entry_offload_refresh(n, op, 0);
-	}
-
-	return mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
-}
-
-static void mlxsw_sp_fib_node_entry_del(struct mlxsw_sp *mlxsw_sp,
+static int mlxsw_sp_fib_node_entry_link(struct mlxsw_sp *mlxsw_sp,
 					struct mlxsw_sp_fib_entry *fib_entry)
 {
 	struct mlxsw_sp_fib_node *fib_node = fib_entry->fib_node;
-
-	if (!mlxsw_sp_fib_node_entry_is_first(fib_node, fib_entry))
-		return;
-
-	/* Promote the next entry by overwriting the deleted entry */
-	if (!list_is_singular(&fib_node->entry_list)) {
-		struct mlxsw_sp_fib_entry *n = list_next_entry(fib_entry, list);
-		enum mlxsw_reg_ralue_op op = MLXSW_REG_RALUE_OP_WRITE_DELETE;
-
-		mlxsw_sp_fib_entry_update(mlxsw_sp, n);
-		mlxsw_sp_fib_entry_offload_refresh(fib_entry, op, 0);
-		return;
-	}
-
-	mlxsw_sp_fib_entry_del(mlxsw_sp, fib_entry);
-}
-
-static int mlxsw_sp_fib4_node_entry_link(struct mlxsw_sp *mlxsw_sp,
-					 struct mlxsw_sp_fib4_entry *fib4_entry,
-					 bool replace, bool append)
-{
 	int err;
 
-	err = mlxsw_sp_fib4_node_list_insert(fib4_entry, replace, append);
-	if (err)
-		return err;
+	fib_node->fib_entry = fib_entry;
 
-	err = mlxsw_sp_fib_node_entry_add(mlxsw_sp, &fib4_entry->common);
+	err = mlxsw_sp_fib_entry_update(mlxsw_sp, fib_entry);
 	if (err)
-		goto err_fib_node_entry_add;
+		goto err_fib_entry_update;
 
 	return 0;
 
-err_fib_node_entry_add:
-	mlxsw_sp_fib4_node_list_remove(fib4_entry);
+err_fib_entry_update:
+	fib_node->fib_entry = NULL;
 	return err;
 }
 
 static void
-mlxsw_sp_fib4_node_entry_unlink(struct mlxsw_sp *mlxsw_sp,
-				struct mlxsw_sp_fib4_entry *fib4_entry)
-{
-	mlxsw_sp_fib_node_entry_del(mlxsw_sp, &fib4_entry->common);
-	mlxsw_sp_fib4_node_list_remove(fib4_entry);
-
-	if (fib4_entry->common.type == MLXSW_SP_FIB_ENTRY_TYPE_IPIP_DECAP)
-		mlxsw_sp_fib_entry_decap_fini(mlxsw_sp, &fib4_entry->common);
-}
-
-static void mlxsw_sp_fib4_entry_replace(struct mlxsw_sp *mlxsw_sp,
-					struct mlxsw_sp_fib4_entry *fib4_entry,
-					bool replace)
+mlxsw_sp_fib_node_entry_unlink(struct mlxsw_sp *mlxsw_sp,
+			       struct mlxsw_sp_fib_entry *fib_entry)
 {
-	struct mlxsw_sp_fib_node *fib_node = fib4_entry->common.fib_node;
-	struct mlxsw_sp_fib4_entry *replaced;
-
-	if (!replace)
-		return;
-
-	/* We inserted the new entry before replaced one */
-	replaced = list_next_entry(fib4_entry, common.list);
+	struct mlxsw_sp_fib_node *fib_node = fib_entry->fib_node;
 
-	mlxsw_sp_fib4_node_entry_unlink(mlxsw_sp, replaced);
-	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, replaced);
-	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
+	mlxsw_sp_fib_entry_del(mlxsw_sp, fib_entry);
+	fib_node->fib_entry = NULL;
 }
 
 static int
-mlxsw_sp_router_fib4_add(struct mlxsw_sp *mlxsw_sp,
-			 const struct fib_entry_notifier_info *fen_info,
-			 bool replace, bool append)
+mlxsw_sp_router_fib4_replace(struct mlxsw_sp *mlxsw_sp,
+			     const struct fib_entry_notifier_info *fen_info)
 {
-	struct mlxsw_sp_fib4_entry *fib4_entry;
+	struct mlxsw_sp_fib4_entry *fib4_entry, *fib4_replaced;
+	struct mlxsw_sp_fib_entry *replaced;
 	struct mlxsw_sp_fib_node *fib_node;
 	int err;
 
@@ -4989,18 +4872,26 @@ mlxsw_sp_router_fib4_add(struct mlxsw_sp *mlxsw_sp,
 		goto err_fib4_entry_create;
 	}
 
-	err = mlxsw_sp_fib4_node_entry_link(mlxsw_sp, fib4_entry, replace,
-					    append);
+	replaced = fib_node->fib_entry;
+	err = mlxsw_sp_fib_node_entry_link(mlxsw_sp, &fib4_entry->common);
 	if (err) {
 		dev_warn(mlxsw_sp->bus_info->dev, "Failed to link FIB entry to node\n");
-		goto err_fib4_node_entry_link;
+		goto err_fib_node_entry_link;
 	}
 
-	mlxsw_sp_fib4_entry_replace(mlxsw_sp, fib4_entry, replace);
+	/* Nothing to replace */
+	if (!replaced)
+		return 0;
+
+	mlxsw_sp_fib_entry_hw_flags_clear(mlxsw_sp, replaced);
+	fib4_replaced = container_of(replaced, struct mlxsw_sp_fib4_entry,
+				     common);
+	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_replaced);
 
 	return 0;
 
-err_fib4_node_entry_link:
+err_fib_node_entry_link:
+	fib_node->fib_entry = replaced;
 	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
 err_fib4_entry_create:
 	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
@@ -5021,7 +4912,7 @@ static void mlxsw_sp_router_fib4_del(struct mlxsw_sp *mlxsw_sp,
 		return;
 	fib_node = fib4_entry->common.fib_node;
 
-	mlxsw_sp_fib4_node_entry_unlink(mlxsw_sp, fib4_entry);
+	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, &fib4_entry->common);
 	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
 	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
 }
@@ -5083,13 +4974,6 @@ static void mlxsw_sp_rt6_destroy(struct mlxsw_sp_rt6 *mlxsw_sp_rt6)
 	kfree(mlxsw_sp_rt6);
 }
 
-static bool mlxsw_sp_fib6_rt_can_mp(const struct fib6_info *rt)
-{
-	/* RTF_CACHE routes are ignored */
-	return !(rt->fib6_flags & RTF_ADDRCONF) &&
-		rt->fib6_nh->fib_nh_gw_family;
-}
-
 static struct fib6_info *
 mlxsw_sp_fib6_entry_rt(const struct mlxsw_sp_fib6_entry *fib6_entry)
 {
@@ -5097,37 +4981,6 @@ mlxsw_sp_fib6_entry_rt(const struct mlxsw_sp_fib6_entry *fib6_entry)
 				list)->rt;
 }
 
-static struct mlxsw_sp_fib6_entry *
-mlxsw_sp_fib6_node_mp_entry_find(const struct mlxsw_sp_fib_node *fib_node,
-				 const struct fib6_info *nrt, bool replace)
-{
-	struct mlxsw_sp_fib6_entry *fib6_entry;
-
-	if (!mlxsw_sp_fib6_rt_can_mp(nrt) || replace)
-		return NULL;
-
-	list_for_each_entry(fib6_entry, &fib_node->entry_list, common.list) {
-		struct fib6_info *rt = mlxsw_sp_fib6_entry_rt(fib6_entry);
-
-		/* RT6_TABLE_LOCAL and RT6_TABLE_MAIN share the same
-		 * virtual router.
-		 */
-		if (rt->fib6_table->tb6_id > nrt->fib6_table->tb6_id)
-			continue;
-		if (rt->fib6_table->tb6_id != nrt->fib6_table->tb6_id)
-			break;
-		if (rt->fib6_metric < nrt->fib6_metric)
-			continue;
-		if (rt->fib6_metric == nrt->fib6_metric &&
-		    mlxsw_sp_fib6_rt_can_mp(rt))
-			return fib6_entry;
-		if (rt->fib6_metric > nrt->fib6_metric)
-			break;
-	}
-
-	return NULL;
-}
-
 static struct mlxsw_sp_rt6 *
 mlxsw_sp_fib6_entry_rt_find(const struct mlxsw_sp_fib6_entry *fib6_entry,
 			    const struct fib6_info *rt)
@@ -5313,6 +5166,11 @@ static int mlxsw_sp_nexthop6_group_get(struct mlxsw_sp *mlxsw_sp,
 		      &nh_grp->fib_list);
 	fib6_entry->common.nh_group = nh_grp;
 
+	/* The route and the nexthop are described by the same struct, so we
+	 * need to the update the nexthop offload indication for the new route.
+	 */
+	__mlxsw_sp_nexthop6_group_offload_refresh(nh_grp, fib6_entry);
+
 	return 0;
 }
 
@@ -5345,16 +5203,16 @@ mlxsw_sp_nexthop6_group_update(struct mlxsw_sp *mlxsw_sp,
 	 * currently associated with it in the device's table is that
 	 * of the old group. Start using the new one instead.
 	 */
-	err = mlxsw_sp_fib_node_entry_add(mlxsw_sp, &fib6_entry->common);
+	err = mlxsw_sp_fib_entry_update(mlxsw_sp, &fib6_entry->common);
 	if (err)
-		goto err_fib_node_entry_add;
+		goto err_fib_entry_update;
 
 	if (list_empty(&old_nh_grp->fib_list))
 		mlxsw_sp_nexthop6_group_destroy(mlxsw_sp, old_nh_grp);
 
 	return 0;
 
-err_fib_node_entry_add:
+err_fib_entry_update:
 	mlxsw_sp_nexthop6_group_put(mlxsw_sp, &fib6_entry->common);
 err_nexthop6_group_get:
 	list_add_tail(&fib6_entry->common.nexthop_group_node,
@@ -5519,112 +5377,13 @@ static void mlxsw_sp_fib6_entry_destroy(struct mlxsw_sp *mlxsw_sp,
 }
 
 static struct mlxsw_sp_fib6_entry *
-mlxsw_sp_fib6_node_entry_find(const struct mlxsw_sp_fib_node *fib_node,
-			      const struct fib6_info *nrt, bool replace)
-{
-	struct mlxsw_sp_fib6_entry *fib6_entry, *fallback = NULL;
-
-	list_for_each_entry(fib6_entry, &fib_node->entry_list, common.list) {
-		struct fib6_info *rt = mlxsw_sp_fib6_entry_rt(fib6_entry);
-
-		if (rt->fib6_table->tb6_id > nrt->fib6_table->tb6_id)
-			continue;
-		if (rt->fib6_table->tb6_id != nrt->fib6_table->tb6_id)
-			break;
-		if (replace && rt->fib6_metric == nrt->fib6_metric) {
-			if (mlxsw_sp_fib6_rt_can_mp(rt) ==
-			    mlxsw_sp_fib6_rt_can_mp(nrt))
-				return fib6_entry;
-			if (mlxsw_sp_fib6_rt_can_mp(nrt))
-				fallback = fallback ?: fib6_entry;
-		}
-		if (rt->fib6_metric > nrt->fib6_metric)
-			return fallback ?: fib6_entry;
-	}
-
-	return fallback;
-}
-
-static int
-mlxsw_sp_fib6_node_list_insert(struct mlxsw_sp_fib6_entry *new6_entry,
-			       bool *p_replace)
-{
-	struct mlxsw_sp_fib_node *fib_node = new6_entry->common.fib_node;
-	struct fib6_info *nrt = mlxsw_sp_fib6_entry_rt(new6_entry);
-	struct mlxsw_sp_fib6_entry *fib6_entry;
-
-	fib6_entry = mlxsw_sp_fib6_node_entry_find(fib_node, nrt, *p_replace);
-
-	if (*p_replace && !fib6_entry)
-		*p_replace = false;
-
-	if (fib6_entry) {
-		list_add_tail(&new6_entry->common.list,
-			      &fib6_entry->common.list);
-	} else {
-		struct mlxsw_sp_fib6_entry *last;
-
-		list_for_each_entry(last, &fib_node->entry_list, common.list) {
-			struct fib6_info *rt = mlxsw_sp_fib6_entry_rt(last);
-
-			if (nrt->fib6_table->tb6_id > rt->fib6_table->tb6_id)
-				break;
-			fib6_entry = last;
-		}
-
-		if (fib6_entry)
-			list_add(&new6_entry->common.list,
-				 &fib6_entry->common.list);
-		else
-			list_add(&new6_entry->common.list,
-				 &fib_node->entry_list);
-	}
-
-	return 0;
-}
-
-static void
-mlxsw_sp_fib6_node_list_remove(struct mlxsw_sp_fib6_entry *fib6_entry)
-{
-	list_del(&fib6_entry->common.list);
-}
-
-static int mlxsw_sp_fib6_node_entry_link(struct mlxsw_sp *mlxsw_sp,
-					 struct mlxsw_sp_fib6_entry *fib6_entry,
-					 bool *p_replace)
-{
-	int err;
-
-	err = mlxsw_sp_fib6_node_list_insert(fib6_entry, p_replace);
-	if (err)
-		return err;
-
-	err = mlxsw_sp_fib_node_entry_add(mlxsw_sp, &fib6_entry->common);
-	if (err)
-		goto err_fib_node_entry_add;
-
-	return 0;
-
-err_fib_node_entry_add:
-	mlxsw_sp_fib6_node_list_remove(fib6_entry);
-	return err;
-}
-
-static void
-mlxsw_sp_fib6_node_entry_unlink(struct mlxsw_sp *mlxsw_sp,
-				struct mlxsw_sp_fib6_entry *fib6_entry)
-{
-	mlxsw_sp_fib_node_entry_del(mlxsw_sp, &fib6_entry->common);
-	mlxsw_sp_fib6_node_list_remove(fib6_entry);
-}
-
-static struct mlxsw_sp_fib6_entry *
 mlxsw_sp_fib6_entry_lookup(struct mlxsw_sp *mlxsw_sp,
 			   const struct fib6_info *rt)
 {
 	struct mlxsw_sp_fib6_entry *fib6_entry;
 	struct mlxsw_sp_fib_node *fib_node;
 	struct mlxsw_sp_fib *fib;
+	struct fib6_info *cmp_rt;
 	struct mlxsw_sp_vr *vr;
 
 	vr = mlxsw_sp_vr_find(mlxsw_sp, rt->fib6_table->tb6_id);
@@ -5638,40 +5397,23 @@ mlxsw_sp_fib6_entry_lookup(struct mlxsw_sp *mlxsw_sp,
 	if (!fib_node)
 		return NULL;
 
-	list_for_each_entry(fib6_entry, &fib_node->entry_list, common.list) {
-		struct fib6_info *iter_rt = mlxsw_sp_fib6_entry_rt(fib6_entry);
-
-		if (rt->fib6_table->tb6_id == iter_rt->fib6_table->tb6_id &&
-		    rt->fib6_metric == iter_rt->fib6_metric &&
-		    mlxsw_sp_fib6_entry_rt_find(fib6_entry, rt))
-			return fib6_entry;
-	}
+	fib6_entry = container_of(fib_node->fib_entry,
+				  struct mlxsw_sp_fib6_entry, common);
+	cmp_rt = mlxsw_sp_fib6_entry_rt(fib6_entry);
+	if (rt->fib6_table->tb6_id == cmp_rt->fib6_table->tb6_id &&
+	    rt->fib6_metric == cmp_rt->fib6_metric &&
+	    mlxsw_sp_fib6_entry_rt_find(fib6_entry, rt))
+		return fib6_entry;
 
 	return NULL;
 }
 
-static void mlxsw_sp_fib6_entry_replace(struct mlxsw_sp *mlxsw_sp,
-					struct mlxsw_sp_fib6_entry *fib6_entry,
-					bool replace)
-{
-	struct mlxsw_sp_fib_node *fib_node = fib6_entry->common.fib_node;
-	struct mlxsw_sp_fib6_entry *replaced;
-
-	if (!replace)
-		return;
-
-	replaced = list_next_entry(fib6_entry, common.list);
-
-	mlxsw_sp_fib6_node_entry_unlink(mlxsw_sp, replaced);
-	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, replaced);
-	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
-}
-
-static int mlxsw_sp_router_fib6_add(struct mlxsw_sp *mlxsw_sp,
-				    struct fib6_info **rt_arr,
-				    unsigned int nrt6, bool replace)
+static int mlxsw_sp_router_fib6_replace(struct mlxsw_sp *mlxsw_sp,
+					struct fib6_info **rt_arr,
+					unsigned int nrt6)
 {
-	struct mlxsw_sp_fib6_entry *fib6_entry;
+	struct mlxsw_sp_fib6_entry *fib6_entry, *fib6_replaced;
+	struct mlxsw_sp_fib_entry *replaced;
 	struct mlxsw_sp_fib_node *fib_node;
 	struct fib6_info *rt = rt_arr[0];
 	int err;
@@ -5693,18 +5435,6 @@ static int mlxsw_sp_router_fib6_add(struct mlxsw_sp *mlxsw_sp,
 	if (IS_ERR(fib_node))
 		return PTR_ERR(fib_node);
 
-	/* Before creating a new entry, try to append route to an existing
-	 * multipath entry.
-	 */
-	fib6_entry = mlxsw_sp_fib6_node_mp_entry_find(fib_node, rt, replace);
-	if (fib6_entry) {
-		err = mlxsw_sp_fib6_entry_nexthop_add(mlxsw_sp, fib6_entry,
-						      rt_arr, nrt6);
-		if (err)
-			goto err_fib6_entry_nexthop_add;
-		return 0;
-	}
-
 	fib6_entry = mlxsw_sp_fib6_entry_create(mlxsw_sp, fib_node, rt_arr,
 						nrt6);
 	if (IS_ERR(fib6_entry)) {
@@ -5712,17 +5442,70 @@ static int mlxsw_sp_router_fib6_add(struct mlxsw_sp *mlxsw_sp,
 		goto err_fib6_entry_create;
 	}
 
-	err = mlxsw_sp_fib6_node_entry_link(mlxsw_sp, fib6_entry, &replace);
+	replaced = fib_node->fib_entry;
+	err = mlxsw_sp_fib_node_entry_link(mlxsw_sp, &fib6_entry->common);
 	if (err)
-		goto err_fib6_node_entry_link;
+		goto err_fib_node_entry_link;
+
+	/* Nothing to replace */
+	if (!replaced)
+		return 0;
 
-	mlxsw_sp_fib6_entry_replace(mlxsw_sp, fib6_entry, replace);
+	mlxsw_sp_fib_entry_hw_flags_clear(mlxsw_sp, replaced);
+	fib6_replaced = container_of(replaced, struct mlxsw_sp_fib6_entry,
+				     common);
+	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_replaced);
 
 	return 0;
 
-err_fib6_node_entry_link:
+err_fib_node_entry_link:
+	fib_node->fib_entry = replaced;
 	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
 err_fib6_entry_create:
+	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
+	return err;
+}
+
+static int mlxsw_sp_router_fib6_append(struct mlxsw_sp *mlxsw_sp,
+				       struct fib6_info **rt_arr,
+				       unsigned int nrt6)
+{
+	struct mlxsw_sp_fib6_entry *fib6_entry;
+	struct mlxsw_sp_fib_node *fib_node;
+	struct fib6_info *rt = rt_arr[0];
+	int err;
+
+	if (mlxsw_sp->router->aborted)
+		return 0;
+
+	if (rt->fib6_src.plen)
+		return -EINVAL;
+
+	if (mlxsw_sp_fib6_rt_should_ignore(rt))
+		return 0;
+
+	fib_node = mlxsw_sp_fib_node_get(mlxsw_sp, rt->fib6_table->tb6_id,
+					 &rt->fib6_dst.addr,
+					 sizeof(rt->fib6_dst.addr),
+					 rt->fib6_dst.plen,
+					 MLXSW_SP_L3_PROTO_IPV6);
+	if (IS_ERR(fib_node))
+		return PTR_ERR(fib_node);
+
+	if (WARN_ON_ONCE(!fib_node->fib_entry)) {
+		mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
+		return -EINVAL;
+	}
+
+	fib6_entry = container_of(fib_node->fib_entry,
+				  struct mlxsw_sp_fib6_entry, common);
+	err = mlxsw_sp_fib6_entry_nexthop_add(mlxsw_sp, fib6_entry, rt_arr,
+					      nrt6);
+	if (err)
+		goto err_fib6_entry_nexthop_add;
+
+	return 0;
+
 err_fib6_entry_nexthop_add:
 	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
 	return err;
@@ -5762,7 +5545,7 @@ static void mlxsw_sp_router_fib6_del(struct mlxsw_sp *mlxsw_sp,
 
 	fib_node = fib6_entry->common.fib_node;
 
-	mlxsw_sp_fib6_node_entry_unlink(mlxsw_sp, fib6_entry);
+	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, &fib6_entry->common);
 	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
 	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
 }
@@ -5916,39 +5699,25 @@ static int mlxsw_sp_router_set_abort_trap(struct mlxsw_sp *mlxsw_sp)
 static void mlxsw_sp_fib4_node_flush(struct mlxsw_sp *mlxsw_sp,
 				     struct mlxsw_sp_fib_node *fib_node)
 {
-	struct mlxsw_sp_fib4_entry *fib4_entry, *tmp;
-
-	list_for_each_entry_safe(fib4_entry, tmp, &fib_node->entry_list,
-				 common.list) {
-		bool do_break = &tmp->common.list == &fib_node->entry_list;
+	struct mlxsw_sp_fib4_entry *fib4_entry;
 
-		mlxsw_sp_fib4_node_entry_unlink(mlxsw_sp, fib4_entry);
-		mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
-		mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
-		/* Break when entry list is empty and node was freed.
-		 * Otherwise, we'll access freed memory in the next
-		 * iteration.
-		 */
-		if (do_break)
-			break;
-	}
+	fib4_entry = container_of(fib_node->fib_entry,
+				  struct mlxsw_sp_fib4_entry, common);
+	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, fib_node->fib_entry);
+	mlxsw_sp_fib4_entry_destroy(mlxsw_sp, fib4_entry);
+	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
 }
 
 static void mlxsw_sp_fib6_node_flush(struct mlxsw_sp *mlxsw_sp,
 				     struct mlxsw_sp_fib_node *fib_node)
 {
-	struct mlxsw_sp_fib6_entry *fib6_entry, *tmp;
-
-	list_for_each_entry_safe(fib6_entry, tmp, &fib_node->entry_list,
-				 common.list) {
-		bool do_break = &tmp->common.list == &fib_node->entry_list;
+	struct mlxsw_sp_fib6_entry *fib6_entry;
 
-		mlxsw_sp_fib6_node_entry_unlink(mlxsw_sp, fib6_entry);
-		mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
-		mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
-		if (do_break)
-			break;
-	}
+	fib6_entry = container_of(fib_node->fib_entry,
+				  struct mlxsw_sp_fib6_entry, common);
+	mlxsw_sp_fib_node_entry_unlink(mlxsw_sp, fib_node->fib_entry);
+	mlxsw_sp_fib6_entry_destroy(mlxsw_sp, fib6_entry);
+	mlxsw_sp_fib_node_put(mlxsw_sp, fib_node);
 }
 
 static void mlxsw_sp_fib_node_flush(struct mlxsw_sp *mlxsw_sp,
@@ -6099,7 +5868,6 @@ static void mlxsw_sp_router_fib4_event_work(struct work_struct *work)
 	struct mlxsw_sp_fib_event_work *fib_work =
 		container_of(work, struct mlxsw_sp_fib_event_work, work);
 	struct mlxsw_sp *mlxsw_sp = fib_work->mlxsw_sp;
-	bool replace, append;
 	int err;
 
 	/* Protect internal structures from changes */
@@ -6107,13 +5875,9 @@ static void mlxsw_sp_router_fib4_event_work(struct work_struct *work)
 	mlxsw_sp_span_respin(mlxsw_sp);
 
 	switch (fib_work->event) {
-	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_APPEND: /* fall through */
-	case FIB_EVENT_ENTRY_ADD:
-		replace = fib_work->event == FIB_EVENT_ENTRY_REPLACE;
-		append = fib_work->event == FIB_EVENT_ENTRY_APPEND;
-		err = mlxsw_sp_router_fib4_add(mlxsw_sp, &fib_work->fen_info,
-					       replace, append);
+	case FIB_EVENT_ENTRY_REPLACE:
+		err = mlxsw_sp_router_fib4_replace(mlxsw_sp,
+						   &fib_work->fen_info);
 		if (err)
 			mlxsw_sp_router_fib_abort(mlxsw_sp);
 		fib_info_put(fib_work->fen_info.fi);
@@ -6138,20 +5902,24 @@ static void mlxsw_sp_router_fib6_event_work(struct work_struct *work)
 	struct mlxsw_sp_fib_event_work *fib_work =
 		container_of(work, struct mlxsw_sp_fib_event_work, work);
 	struct mlxsw_sp *mlxsw_sp = fib_work->mlxsw_sp;
-	bool replace;
 	int err;
 
 	rtnl_lock();
 	mlxsw_sp_span_respin(mlxsw_sp);
 
 	switch (fib_work->event) {
-	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_ADD:
-		replace = fib_work->event == FIB_EVENT_ENTRY_REPLACE;
-		err = mlxsw_sp_router_fib6_add(mlxsw_sp,
-					       fib_work->fib6_work.rt_arr,
-					       fib_work->fib6_work.nrt6,
-					       replace);
+	case FIB_EVENT_ENTRY_REPLACE:
+		err = mlxsw_sp_router_fib6_replace(mlxsw_sp,
+						   fib_work->fib6_work.rt_arr,
+						   fib_work->fib6_work.nrt6);
+		if (err)
+			mlxsw_sp_router_fib_abort(mlxsw_sp);
+		mlxsw_sp_router_fib6_work_fini(&fib_work->fib6_work);
+		break;
+	case FIB_EVENT_ENTRY_APPEND:
+		err = mlxsw_sp_router_fib6_append(mlxsw_sp,
+						  fib_work->fib6_work.rt_arr,
+						  fib_work->fib6_work.nrt6);
 		if (err)
 			mlxsw_sp_router_fib_abort(mlxsw_sp);
 		mlxsw_sp_router_fib6_work_fini(&fib_work->fib6_work);
@@ -6216,8 +5984,6 @@ static void mlxsw_sp_router_fib4_event(struct mlxsw_sp_fib_event_work *fib_work,
 
 	switch (fib_work->event) {
 	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_APPEND: /* fall through */
-	case FIB_EVENT_ENTRY_ADD: /* fall through */
 	case FIB_EVENT_ENTRY_DEL:
 		fen_info = container_of(info, struct fib_entry_notifier_info,
 					info);
@@ -6245,7 +6011,7 @@ static int mlxsw_sp_router_fib6_event(struct mlxsw_sp_fib_event_work *fib_work,
 
 	switch (fib_work->event) {
 	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_ADD: /* fall through */
+	case FIB_EVENT_ENTRY_APPEND: /* fall through */
 	case FIB_EVENT_ENTRY_DEL:
 		fen6_info = container_of(info, struct fib6_entry_notifier_info,
 					 info);
@@ -6348,9 +6114,9 @@ static int mlxsw_sp_router_fib_event(struct notifier_block *nb,
 		err = mlxsw_sp_router_fib_rule_event(event, info,
 						     router->mlxsw_sp);
 		return notifier_from_errno(err);
-	case FIB_EVENT_ENTRY_ADD:
+	case FIB_EVENT_ENTRY_ADD: /* fall through */
 	case FIB_EVENT_ENTRY_REPLACE: /* fall through */
-	case FIB_EVENT_ENTRY_APPEND:  /* fall through */
+	case FIB_EVENT_ENTRY_APPEND:
 		if (router->aborted) {
 			NL_SET_ERR_MSG_MOD(info->extack, "FIB offload was aborted. Not configuring route");
 			return notifier_from_errno(-EINVAL);
@@ -8025,8 +7791,18 @@ mlxsw_sp_ipip_config_tigcr(struct mlxsw_sp *mlxsw_sp)
 
 static int mlxsw_sp_ipips_init(struct mlxsw_sp *mlxsw_sp)
 {
+	int err;
+
 	mlxsw_sp->router->ipip_ops_arr = mlxsw_sp_ipip_ops_arr;
 	INIT_LIST_HEAD(&mlxsw_sp->router->ipip_list);
+
+	err = mlxsw_sp_ipip_ecn_encap_init(mlxsw_sp);
+	if (err)
+		return err;
+	err = mlxsw_sp_ipip_ecn_decap_init(mlxsw_sp);
+	if (err)
+		return err;
+
 	return mlxsw_sp_ipip_config_tigcr(mlxsw_sp);
 }
 
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.h
index cc1de91e8217..c9b94f435cdd 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.h
@@ -104,4 +104,7 @@ static inline bool mlxsw_sp_l3addr_eq(const union mlxsw_sp_l3addr *addr1,
 	return !memcmp(addr1, addr2, sizeof(*addr1));
 }
 
+int mlxsw_sp_ipip_ecn_encap_init(struct mlxsw_sp *mlxsw_sp);
+int mlxsw_sp_ipip_ecn_decap_init(struct mlxsw_sp *mlxsw_sp);
+
 #endif /* _MLXSW_ROUTER_H_*/
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c
index 200d324e6d99..0cdd7954a085 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c
@@ -748,33 +748,50 @@ static bool mlxsw_sp_span_is_egress_mirror(struct mlxsw_sp_port *port)
 	return false;
 }
 
-static int mlxsw_sp_span_mtu_to_buffsize(const struct mlxsw_sp *mlxsw_sp,
-					 int mtu)
+static int
+mlxsw_sp_span_port_buffsize_update(struct mlxsw_sp_port *mlxsw_sp_port, u16 mtu)
 {
-	return mlxsw_sp_bytes_cells(mlxsw_sp, mtu * 5 / 2) + 1;
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char sbib_pl[MLXSW_REG_SBIB_LEN];
+	u32 buffsize;
+	u32 speed;
+	int err;
+
+	err = mlxsw_sp_port_speed_get(mlxsw_sp_port, &speed);
+	if (err)
+		return err;
+	if (speed == SPEED_UNKNOWN)
+		speed = 0;
+
+	buffsize = mlxsw_sp_span_buffsize_get(mlxsw_sp, speed, mtu);
+	mlxsw_reg_sbib_pack(sbib_pl, mlxsw_sp_port->local_port, buffsize);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sbib), sbib_pl);
 }
 
 int mlxsw_sp_span_port_mtu_update(struct mlxsw_sp_port *port, u16 mtu)
 {
-	struct mlxsw_sp *mlxsw_sp = port->mlxsw_sp;
-	char sbib_pl[MLXSW_REG_SBIB_LEN];
-	int err;
-
 	/* If port is egress mirrored, the shared buffer size should be
 	 * updated according to the mtu value
 	 */
-	if (mlxsw_sp_span_is_egress_mirror(port)) {
-		u32 buffsize = mlxsw_sp_span_mtu_to_buffsize(mlxsw_sp, mtu);
+	if (mlxsw_sp_span_is_egress_mirror(port))
+		return mlxsw_sp_span_port_buffsize_update(port, mtu);
+	return 0;
+}
 
-		mlxsw_reg_sbib_pack(sbib_pl, port->local_port, buffsize);
-		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sbib), sbib_pl);
-		if (err) {
-			netdev_err(port->dev, "Could not update shared buffer for mirroring\n");
-			return err;
-		}
-	}
+void mlxsw_sp_span_speed_update_work(struct work_struct *work)
+{
+	struct delayed_work *dwork = to_delayed_work(work);
+	struct mlxsw_sp_port *mlxsw_sp_port;
 
-	return 0;
+	mlxsw_sp_port = container_of(dwork, struct mlxsw_sp_port,
+				     span.speed_update_dw);
+
+	/* If port is egress mirrored, the shared buffer size should be
+	 * updated according to the speed value.
+	 */
+	if (mlxsw_sp_span_is_egress_mirror(mlxsw_sp_port))
+		mlxsw_sp_span_port_buffsize_update(mlxsw_sp_port,
+						   mlxsw_sp_port->dev->mtu);
 }
 
 static struct mlxsw_sp_span_inspected_port *
@@ -836,15 +853,9 @@ mlxsw_sp_span_inspected_port_add(struct mlxsw_sp_port *port,
 
 	/* if it is an egress SPAN, bind a shared buffer to it */
 	if (type == MLXSW_SP_SPAN_EGRESS) {
-		u32 buffsize = mlxsw_sp_span_mtu_to_buffsize(mlxsw_sp,
-							     port->dev->mtu);
-
-		mlxsw_reg_sbib_pack(sbib_pl, port->local_port, buffsize);
-		err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(sbib), sbib_pl);
-		if (err) {
-			netdev_err(port->dev, "Could not create shared buffer for mirroring\n");
+		err = mlxsw_sp_span_port_buffsize_update(port, port->dev->mtu);
+		if (err)
 			return err;
-		}
 	}
 
 	if (bind) {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.h
index 5e04252f2a11..59724335525f 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.h
@@ -74,5 +74,6 @@ void mlxsw_sp_span_entry_invalidate(struct mlxsw_sp *mlxsw_sp,
 				    struct mlxsw_sp_span_entry *span_entry);
 
 int mlxsw_sp_span_port_mtu_update(struct mlxsw_sp_port *port, u16 mtu);
+void mlxsw_sp_span_speed_update_work(struct work_struct *work);
 
 #endif
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_trap.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_trap.c
index e0d7c49ffae0..60205aa3f6a5 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_trap.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_trap.c
@@ -9,6 +9,20 @@
 #include "reg.h"
 #include "spectrum.h"
 
+/* All driver-specific traps must be documented in
+ * Documentation/networking/devlink/mlxsw.rst
+ */
+enum {
+	DEVLINK_MLXSW_TRAP_ID_BASE = DEVLINK_TRAP_GENERIC_ID_MAX,
+	DEVLINK_MLXSW_TRAP_ID_IRIF_DISABLED,
+	DEVLINK_MLXSW_TRAP_ID_ERIF_DISABLED,
+};
+
+#define DEVLINK_MLXSW_TRAP_NAME_IRIF_DISABLED \
+	"irif_disabled"
+#define DEVLINK_MLXSW_TRAP_NAME_ERIF_DISABLED \
+	"erif_disabled"
+
 #define MLXSW_SP_TRAP_METADATA DEVLINK_TRAP_METADATA_TYPE_F_IN_PORT
 
 static void mlxsw_sp_rx_drop_listener(struct sk_buff *skb, u8 local_port,
@@ -21,6 +35,12 @@ static void mlxsw_sp_rx_exception_listener(struct sk_buff *skb, u8 local_port,
 			     DEVLINK_TRAP_GROUP_GENERIC(_group_id),	      \
 			     MLXSW_SP_TRAP_METADATA)
 
+#define MLXSW_SP_TRAP_DRIVER_DROP(_id, _group_id)			      \
+	DEVLINK_TRAP_DRIVER(DROP, DROP, DEVLINK_MLXSW_TRAP_ID_##_id,	      \
+			    DEVLINK_MLXSW_TRAP_NAME_##_id,		      \
+			    DEVLINK_TRAP_GROUP_GENERIC(_group_id),	      \
+			    MLXSW_SP_TRAP_METADATA)
+
 #define MLXSW_SP_TRAP_EXCEPTION(_id, _group_id)		      \
 	DEVLINK_TRAP_GENERIC(EXCEPTION, TRAP, _id,			      \
 			     DEVLINK_TRAP_GROUP_GENERIC(_group_id),	      \
@@ -58,6 +78,11 @@ static struct devlink_trap mlxsw_sp_traps_arr[] = {
 	MLXSW_SP_TRAP_EXCEPTION(UNRESOLVED_NEIGH, L3_DROPS),
 	MLXSW_SP_TRAP_EXCEPTION(IPV4_LPM_UNICAST_MISS, L3_DROPS),
 	MLXSW_SP_TRAP_EXCEPTION(IPV6_LPM_UNICAST_MISS, L3_DROPS),
+	MLXSW_SP_TRAP_DRIVER_DROP(IRIF_DISABLED, L3_DROPS),
+	MLXSW_SP_TRAP_DRIVER_DROP(ERIF_DISABLED, L3_DROPS),
+	MLXSW_SP_TRAP_DROP(NON_ROUTABLE, L3_DROPS),
+	MLXSW_SP_TRAP_EXCEPTION(DECAP_ERROR, TUNNEL_DROPS),
+	MLXSW_SP_TRAP_DROP(OVERLAY_SMAC_MC, TUNNEL_DROPS),
 };
 
 static struct mlxsw_listener mlxsw_sp_listeners_arr[] = {
@@ -90,6 +115,15 @@ static struct mlxsw_listener mlxsw_sp_listeners_arr[] = {
 			       TRAP_EXCEPTION_TO_CPU),
 	MLXSW_SP_RXL_EXCEPTION(DISCARD_ROUTER_LPM6, ROUTER_EXP,
 			       TRAP_EXCEPTION_TO_CPU),
+	MLXSW_SP_RXL_DISCARD(ROUTER_IRIF_EN, L3_DISCARDS),
+	MLXSW_SP_RXL_DISCARD(ROUTER_ERIF_EN, L3_DISCARDS),
+	MLXSW_SP_RXL_DISCARD(NON_ROUTABLE, L3_DISCARDS),
+	MLXSW_SP_RXL_EXCEPTION(DECAP_ECN0, ROUTER_EXP, TRAP_EXCEPTION_TO_CPU),
+	MLXSW_SP_RXL_EXCEPTION(IPIP_DECAP_ERROR, ROUTER_EXP,
+			       TRAP_EXCEPTION_TO_CPU),
+	MLXSW_SP_RXL_EXCEPTION(DISCARD_DEC_PKT, TUNNEL_DISCARDS,
+			       TRAP_EXCEPTION_TO_CPU),
+	MLXSW_SP_RXL_DISCARD(OVERLAY_SMAC_MC, TUNNEL_DISCARDS),
 };
 
 /* Mapping between hardware trap and devlink trap. Multiple hardware traps can
@@ -123,6 +157,13 @@ static u16 mlxsw_sp_listener_devlink_map[] = {
 	DEVLINK_TRAP_GENERIC_ID_UNRESOLVED_NEIGH,
 	DEVLINK_TRAP_GENERIC_ID_IPV4_LPM_UNICAST_MISS,
 	DEVLINK_TRAP_GENERIC_ID_IPV6_LPM_UNICAST_MISS,
+	DEVLINK_MLXSW_TRAP_ID_IRIF_DISABLED,
+	DEVLINK_MLXSW_TRAP_ID_ERIF_DISABLED,
+	DEVLINK_TRAP_GENERIC_ID_NON_ROUTABLE,
+	DEVLINK_TRAP_GENERIC_ID_DECAP_ERROR,
+	DEVLINK_TRAP_GENERIC_ID_DECAP_ERROR,
+	DEVLINK_TRAP_GENERIC_ID_DECAP_ERROR,
+	DEVLINK_TRAP_GENERIC_ID_OVERLAY_SMAC_MC,
 };
 
 static int mlxsw_sp_rx_listener(struct mlxsw_sp *mlxsw_sp, struct sk_buff *skb,
@@ -304,8 +345,9 @@ mlxsw_sp_trap_group_policer_init(struct mlxsw_sp *mlxsw_sp,
 	u32 rate;
 
 	switch (group->id) {
-	case DEVLINK_TRAP_GROUP_GENERIC_ID_L3_DROPS:/* fall through */
-	case DEVLINK_TRAP_GROUP_GENERIC_ID_L2_DROPS:
+	case DEVLINK_TRAP_GROUP_GENERIC_ID_L2_DROPS: /* fall through */
+	case DEVLINK_TRAP_GROUP_GENERIC_ID_L3_DROPS: /* fall through */
+	case DEVLINK_TRAP_GROUP_GENERIC_ID_TUNNEL_DROPS:
 		policer_id = MLXSW_SP_DISCARD_POLICER_ID;
 		ir_units = MLXSW_REG_QPCR_IR_UNITS_M;
 		is_bytes = false;
@@ -342,6 +384,12 @@ __mlxsw_sp_trap_group_init(struct mlxsw_sp *mlxsw_sp,
 		priority = 0;
 		tc = 1;
 		break;
+	case DEVLINK_TRAP_GROUP_GENERIC_ID_TUNNEL_DROPS:
+		group_id = MLXSW_REG_HTGT_TRAP_GROUP_SP_TUNNEL_DISCARDS;
+		policer_id = MLXSW_SP_DISCARD_POLICER_ID;
+		priority = 0;
+		tc = 1;
+		break;
 	default:
 		return -EINVAL;
 	}
diff --git a/drivers/net/ethernet/mellanox/mlxsw/trap.h b/drivers/net/ethernet/mellanox/mlxsw/trap.h
index 0c1c142bb6b0..12e1fa998d42 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/trap.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/trap.h
@@ -67,6 +67,7 @@ enum {
 	MLXSW_TRAP_ID_NVE_ENCAP_ARP = 0xBD,
 	MLXSW_TRAP_ID_ROUTER_ALERT_IPV4 = 0xD6,
 	MLXSW_TRAP_ID_ROUTER_ALERT_IPV6 = 0xD7,
+	MLXSW_TRAP_ID_DISCARD_NON_ROUTABLE = 0x11A,
 	MLXSW_TRAP_ID_DISCARD_ROUTER2 = 0x130,
 	MLXSW_TRAP_ID_DISCARD_ROUTER3 = 0x131,
 	MLXSW_TRAP_ID_DISCARD_ING_PACKET_SMAC_MC = 0x140,
@@ -80,12 +81,20 @@ enum {
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_UC_DIP_MC_DMAC = 0x161,
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_DIP_LB = 0x162,
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_SIP_MC = 0x163,
+	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_SIP_CLASS_E = 0x164,
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_SIP_LB = 0x165,
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_CORRUPTED_IP_HDR = 0x167,
+	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_MC_DMAC = 0x168,
+	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_SIP_DIP = 0x169,
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_IPV4_SIP_BC = 0x16A,
 	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_IPV4_DIP_LOCAL_NET = 0x16B,
+	MLXSW_TRAP_ID_DISCARD_ING_ROUTER_DIP_LINK_LOCAL = 0x16C,
+	MLXSW_TRAP_ID_DISCARD_ROUTER_IRIF_EN = 0x178,
+	MLXSW_TRAP_ID_DISCARD_ROUTER_ERIF_EN = 0x179,
 	MLXSW_TRAP_ID_DISCARD_ROUTER_LPM4 = 0x17B,
 	MLXSW_TRAP_ID_DISCARD_ROUTER_LPM6 = 0x17C,
+	MLXSW_TRAP_ID_DISCARD_DEC_PKT = 0x188,
+	MLXSW_TRAP_ID_DISCARD_OVERLAY_SMAC_MC = 0x190,
 	MLXSW_TRAP_ID_DISCARD_IPV6_MC_DIP_RESERVED_SCOPE = 0x1B0,
 	MLXSW_TRAP_ID_DISCARD_IPV6_MC_DIP_INTERFACE_LOCAL_SCOPE = 0x1B1,
 	MLXSW_TRAP_ID_ACL0 = 0x1C0,