Merge tag 'mlx5-updates-2021-04-19' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2021-04-19

This patchset provides some updates to mlx5e and mlx5 SW steering drivers:

1) Tariq and Vladyslav they both provide some trivial update to mlx5e netdev.

The next 12 patches in the patchset are focused toward mlx5 SW steering:
2) 3 trivial cleanup patches

3) Dynamic Flex parser support:
   Flex parser is a HW parser that can support protocols that are not
    natively supported by the HCA, such as Geneve (TLV options) and GTP-U.
    There are 8 such parsers, and each of them can be assigned to parse a
    specific set of protocols.

4) Enable matching on Geneve TLV options

5) Use Flex parser for MPLS over UDP/GRE

6) Enable matching on tunnel GTP-U and GTP-U first extension
   header using

7) Improved QoS for SW steering internal QPair for a better insertion rate
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/mellanox/mlx5/core/en/devlink.c

@@ -55,12 +55,17 @@ void mlx5e_devlink_port_unregister(struct mlx5e_priv *priv)
 {
 	struct devlink_port *dl_port = mlx5e_devlink_get_dl_port(priv);
 
+	if (dl_port->registered)
 		devlink_port_unregister(dl_port);
 }
 
 struct devlink_port *mlx5e_get_devlink_port(struct net_device *dev)
 {
 	struct mlx5e_priv *priv = netdev_priv(dev);
+	struct devlink_port *port;
 
-	return mlx5e_devlink_get_dl_port(priv);
+	port = mlx5e_devlink_get_dl_port(priv);
+	if (port->registered)
+		return port;
+	return NULL;
 }

drivers/net/ethernet/mellanox/mlx5/core/en/params.c

@@ -6,6 +6,7 @@
 #include "en/port.h"
 #include "en_accel/en_accel.h"
 #include "accel/ipsec.h"
+#include "fpga/ipsec.h"
 
 static bool mlx5e_rx_is_xdp(struct mlx5e_params *params,
 			    struct mlx5e_xsk_param *xsk)
@@ -89,30 +90,39 @@ bool mlx5e_rx_is_linear_skb(struct mlx5e_params *params,
 	return !params->lro_en && linear_frag_sz <= PAGE_SIZE;
 }
 
-#define MLX5_MAX_MPWQE_LOG_WQE_STRIDE_SZ ((BIT(__mlx5_bit_sz(wq, log_wqe_stride_size)) - 1) + \
-					  MLX5_MPWQE_LOG_STRIDE_SZ_BASE)
-bool mlx5e_rx_mpwqe_is_linear_skb(struct mlx5_core_dev *mdev,
-				  struct mlx5e_params *params,
-				  struct mlx5e_xsk_param *xsk)
+bool mlx5e_verify_rx_mpwqe_strides(struct mlx5_core_dev *mdev,
+				   u8 log_stride_sz, u8 log_num_strides)
 {
-	u32 linear_frag_sz = mlx5e_rx_get_linear_frag_sz(params, xsk);
-	s8 signed_log_num_strides_param;
-	u8 log_num_strides;
+	if (log_stride_sz + log_num_strides != MLX5_MPWRQ_LOG_WQE_SZ)
+		return false;
 
-	if (!mlx5e_rx_is_linear_skb(params, xsk))
+	if (log_stride_sz < MLX5_MPWQE_LOG_STRIDE_SZ_BASE ||
+	    log_stride_sz > MLX5_MPWQE_LOG_STRIDE_SZ_MAX)
 		return false;
 
-	if (order_base_2(linear_frag_sz) > MLX5_MAX_MPWQE_LOG_WQE_STRIDE_SZ)
+	if (log_num_strides > MLX5_MPWQE_LOG_NUM_STRIDES_MAX)
 		return false;
 
 	if (MLX5_CAP_GEN(mdev, ext_stride_num_range))
-		return true;
+		return log_num_strides >= MLX5_MPWQE_LOG_NUM_STRIDES_EXT_BASE;
+
+	return log_num_strides >= MLX5_MPWQE_LOG_NUM_STRIDES_BASE;
+}
+
+bool mlx5e_rx_mpwqe_is_linear_skb(struct mlx5_core_dev *mdev,
+				  struct mlx5e_params *params,
+				  struct mlx5e_xsk_param *xsk)
+{
+	s8 log_num_strides;
+	u8 log_stride_sz;
+
+	if (!mlx5e_rx_is_linear_skb(params, xsk))
+		return false;
 
-	log_num_strides = MLX5_MPWRQ_LOG_WQE_SZ - order_base_2(linear_frag_sz);
-	signed_log_num_strides_param =
-		(s8)log_num_strides - MLX5_MPWQE_LOG_NUM_STRIDES_BASE;
+	log_stride_sz = order_base_2(mlx5e_rx_get_linear_frag_sz(params, xsk));
+	log_num_strides = MLX5_MPWRQ_LOG_WQE_SZ - log_stride_sz;
 
-	return signed_log_num_strides_param >= 0;
+	return mlx5e_verify_rx_mpwqe_strides(mdev, log_stride_sz, log_num_strides);
 }
 
 u8 mlx5e_mpwqe_get_log_rq_size(struct mlx5e_params *params,
@@ -282,7 +292,7 @@ bool mlx5e_striding_rq_possible(struct mlx5_core_dev *mdev,
 	if (!mlx5e_check_fragmented_striding_rq_cap(mdev))
 		return false;
 
-	if (MLX5_IPSEC_DEV(mdev))
+	if (mlx5_fpga_is_ipsec_device(mdev))
 		return false;
 
 	if (params->xdp_prog) {
@@ -364,7 +374,7 @@ static void mlx5e_build_rq_frags_info(struct mlx5_core_dev *mdev,
 	u32 buf_size = 0;
 	int i;
 
-	if (MLX5_IPSEC_DEV(mdev))
+	if (mlx5_fpga_is_ipsec_device(mdev))
 		byte_count += MLX5E_METADATA_ETHER_LEN;
 
 	if (mlx5e_rx_is_linear_skb(params, xsk)) {
@@ -461,7 +471,7 @@ static void mlx5e_build_rx_cq_param(struct mlx5_core_dev *mdev,
 	param->cq_period_mode = params->rx_cq_moderation.cq_period_mode;
 }
 
-void mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
+int mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
 			 struct mlx5e_params *params,
 			 struct mlx5e_xsk_param *xsk,
 			 u16 q_counter,
@@ -472,15 +482,25 @@ void mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
 	int ndsegs = 1;
 
 	switch (params->rq_wq_type) {
-	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
+	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ: {
+		u8 log_wqe_num_of_strides = mlx5e_mpwqe_get_log_num_strides(mdev, params, xsk);
+		u8 log_wqe_stride_size = mlx5e_mpwqe_get_log_stride_size(mdev, params, xsk);
+
+		if (!mlx5e_verify_rx_mpwqe_strides(mdev, log_wqe_stride_size,
+						   log_wqe_num_of_strides)) {
+			mlx5_core_err(mdev,
+				      "Bad RX MPWQE params: log_stride_size %u, log_num_strides %u\n",
+				      log_wqe_stride_size, log_wqe_num_of_strides);
+			return -EINVAL;
+		}
+
 		MLX5_SET(wq, wq, log_wqe_num_of_strides,
-			 mlx5e_mpwqe_get_log_num_strides(mdev, params, xsk) -
-			 MLX5_MPWQE_LOG_NUM_STRIDES_BASE);
+			 log_wqe_num_of_strides - MLX5_MPWQE_LOG_NUM_STRIDES_BASE);
 		MLX5_SET(wq, wq, log_wqe_stride_size,
-			 mlx5e_mpwqe_get_log_stride_size(mdev, params, xsk) -
-			 MLX5_MPWQE_LOG_STRIDE_SZ_BASE);
+			 log_wqe_stride_size - MLX5_MPWQE_LOG_STRIDE_SZ_BASE);
 		MLX5_SET(wq, wq, log_wq_sz, mlx5e_mpwqe_get_log_rq_size(params, xsk));
 		break;
+	}
 	default: /* MLX5_WQ_TYPE_CYCLIC */
 		MLX5_SET(wq, wq, log_wq_sz, params->log_rq_mtu_frames);
 		mlx5e_build_rq_frags_info(mdev, params, xsk, &param->frags_info);
@@ -498,6 +518,8 @@ void mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
 
 	param->wq.buf_numa_node = dev_to_node(mlx5_core_dma_dev(mdev));
 	mlx5e_build_rx_cq_param(mdev, params, xsk, &param->cqp);
+
+	return 0;
 }
 
 void mlx5e_build_drop_rq_param(struct mlx5_core_dev *mdev,
@@ -642,14 +664,17 @@ void mlx5e_build_xdpsq_param(struct mlx5_core_dev *mdev,
 	mlx5e_build_tx_cq_param(mdev, params, &param->cqp);
 }
 
-void mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
+int mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
 			      struct mlx5e_params *params,
 			      u16 q_counter,
 			      struct mlx5e_channel_param *cparam)
 {
 	u8 icosq_log_wq_sz, async_icosq_log_wq_sz;
+	int err;
 
-	mlx5e_build_rq_param(mdev, params, NULL, q_counter, &cparam->rq);
+	err = mlx5e_build_rq_param(mdev, params, NULL, q_counter, &cparam->rq);
+	if (err)
+		return err;
 
 	icosq_log_wq_sz = mlx5e_build_icosq_log_wq_sz(params, &cparam->rq);
 	async_icosq_log_wq_sz = mlx5e_build_async_icosq_log_wq_sz(mdev);
@@ -658,4 +683,6 @@ void mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
 	mlx5e_build_xdpsq_param(mdev, params, &cparam->xdp_sq);
 	mlx5e_build_icosq_param(mdev, icosq_log_wq_sz, &cparam->icosq);
 	mlx5e_build_async_icosq_param(mdev, async_icosq_log_wq_sz, &cparam->async_icosq);
+
+	return 0;
 }

drivers/net/ethernet/mellanox/mlx5/core/en/params.h

@@ -96,6 +96,8 @@ void mlx5e_build_rq_params(struct mlx5_core_dev *mdev, struct mlx5e_params *para
 void mlx5e_set_rq_type(struct mlx5_core_dev *mdev, struct mlx5e_params *params);
 void mlx5e_init_rq_type_params(struct mlx5_core_dev *mdev, struct mlx5e_params *params);
 
+bool mlx5e_verify_rx_mpwqe_strides(struct mlx5_core_dev *mdev,
+				   u8 log_stride_sz, u8 log_num_strides);
 u16 mlx5e_get_linear_rq_headroom(struct mlx5e_params *params,
 				 struct mlx5e_xsk_param *xsk);
 u32 mlx5e_rx_get_min_frag_sz(struct mlx5e_params *params,
@@ -122,7 +124,7 @@ u16 mlx5e_get_rq_headroom(struct mlx5_core_dev *mdev,
 /* Build queue parameters */
 
 void mlx5e_build_create_cq_param(struct mlx5e_create_cq_param *ccp, struct mlx5e_channel *c);
-void mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
+int mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
 			 struct mlx5e_params *params,
 			 struct mlx5e_xsk_param *xsk,
 			 u16 q_counter,
@@ -141,7 +143,7 @@ void mlx5e_build_tx_cq_param(struct mlx5_core_dev *mdev,
 void mlx5e_build_xdpsq_param(struct mlx5_core_dev *mdev,
 			     struct mlx5e_params *params,
 			     struct mlx5e_sq_param *param);
-void mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
+int mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
 			      struct mlx5e_params *params,
 			      u16 q_counter,
 			      struct mlx5e_channel_param *cparam);

drivers/net/ethernet/mellanox/mlx5/core/en_main.c

@@ -2086,7 +2086,10 @@ int mlx5e_open_channels(struct mlx5e_priv *priv,
 	if (!chs->c || !cparam)
 		goto err_free;
 
-	mlx5e_build_channel_param(priv->mdev, &chs->params, priv->q_counter, cparam);
+	err = mlx5e_build_channel_param(priv->mdev, &chs->params, priv->q_counter, cparam);
+	if (err)
+		goto err_free;
+
 	for (i = 0; i < chs->num; i++) {
 		struct xsk_buff_pool *xsk_pool = NULL;
 
@@ -4886,6 +4889,7 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
 			  struct net_device *netdev)
 {
 	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct devlink_port *dl_port;
 	int err;
 
 	mlx5e_build_nic_params(priv, &priv->xsk, netdev->mtu);
@@ -4901,6 +4905,8 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
 	if (err)
 		mlx5_core_err(mdev, "TLS initialization failed, %d\n", err);
 
+	dl_port = mlx5e_devlink_get_dl_port(priv);
+	if (dl_port->registered)
 		mlx5e_health_create_reporters(priv);
 
 	return 0;
@@ -4908,6 +4914,9 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
 
 static void mlx5e_nic_cleanup(struct mlx5e_priv *priv)
 {
+	struct devlink_port *dl_port = mlx5e_devlink_get_dl_port(priv);
+
+	if (dl_port->registered)
 		mlx5e_health_destroy_reporters(priv);
 	mlx5e_tls_cleanup(priv);
 	mlx5e_ipsec_cleanup(priv);

drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.c

@@ -83,14 +83,16 @@ mlx5_eswitch_termtbl_create(struct mlx5_core_dev *dev,
 	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");
+		esw_warn(dev, "Failed to create termination table (error %d)\n",
+			 IS_ERR(tt->termtbl));
 		return -EOPNOTSUPP;
 	}
 
 	tt->rule = mlx5_add_flow_rules(tt->termtbl, NULL, flow_act,
 				       &tt->dest, 1);
 	if (IS_ERR(tt->rule)) {
-		esw_warn(dev, "Failed to create termination table rule\n");
+		esw_warn(dev, "Failed to create termination table rule (error %d)\n",
+			 IS_ERR(tt->rule));
 		goto add_flow_err;
 	}
 	return 0;
@@ -140,10 +142,9 @@ mlx5_eswitch_termtbl_get_create(struct mlx5_eswitch *esw,
 	memcpy(&tt->flow_act, flow_act, sizeof(*flow_act));
 
 	err = mlx5_eswitch_termtbl_create(esw->dev, tt, flow_act);
-	if (err) {
-		esw_warn(esw->dev, "Failed to create termination table\n");
+	if (err)
 		goto tt_create_err;
-	}
+
 	hash_add(esw->offloads.termtbl_tbl, &tt->termtbl_hlist, hash_key);
 tt_add_ref:
 	tt->ref_count++;
@@ -282,7 +283,8 @@ mlx5_eswitch_add_termtbl_rule(struct mlx5_eswitch *esw,
 		tt = mlx5_eswitch_termtbl_get_create(esw, &term_tbl_act,
 						     &dest[i], attr);
 		if (IS_ERR(tt)) {
-			esw_warn(esw->dev, "Failed to create termination table\n");
+			esw_warn(esw->dev, "Failed to get termination table (error %d)\n",
+				 IS_ERR(tt));
 			goto revert_changes;
 		}
 		attr->dests[num_vport_dests].termtbl = tt;

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_cmd.c

@@ -85,15 +85,53 @@ int mlx5dr_cmd_query_esw_caps(struct mlx5_core_dev *mdev,
 	return 0;
 }
 
+static int dr_cmd_query_nic_vport_roce_en(struct mlx5_core_dev *mdev,
+					  u16 vport, bool *roce_en)
+{
+	u32 out[MLX5_ST_SZ_DW(query_nic_vport_context_out)] = {};
+	u32 in[MLX5_ST_SZ_DW(query_nic_vport_context_in)] = {};
+	int err;
+
+	MLX5_SET(query_nic_vport_context_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_NIC_VPORT_CONTEXT);
+	MLX5_SET(query_nic_vport_context_in, in, vport_number, vport);
+	MLX5_SET(query_nic_vport_context_in, in, other_vport, !!vport);
+
+	err = mlx5_cmd_exec(mdev, in, sizeof(in), out, sizeof(out));
+	if (err)
+		return err;
+
+	*roce_en = MLX5_GET(query_nic_vport_context_out, out,
+			    nic_vport_context.roce_en);
+	return 0;
+}
+
 int mlx5dr_cmd_query_device(struct mlx5_core_dev *mdev,
 			    struct mlx5dr_cmd_caps *caps)
 {
+	bool roce_en;
+	int err;
+
 	caps->prio_tag_required	= MLX5_CAP_GEN(mdev, prio_tag_required);
 	caps->eswitch_manager	= MLX5_CAP_GEN(mdev, eswitch_manager);
 	caps->gvmi		= MLX5_CAP_GEN(mdev, vhca_id);
 	caps->flex_protocols	= MLX5_CAP_GEN(mdev, flex_parser_protocols);
 	caps->sw_format_ver	= MLX5_CAP_GEN(mdev, steering_format_version);
 
+	if (MLX5_CAP_GEN(mdev, roce)) {
+		err = dr_cmd_query_nic_vport_roce_en(mdev, 0, &roce_en);
+		if (err)
+			return err;
+
+		caps->roce_caps.roce_en = roce_en;
+		caps->roce_caps.fl_rc_qp_when_roce_disabled =
+			MLX5_CAP_ROCE(mdev, fl_rc_qp_when_roce_disabled);
+		caps->roce_caps.fl_rc_qp_when_roce_enabled =
+			MLX5_CAP_ROCE(mdev, fl_rc_qp_when_roce_enabled);
+	}
+
+	caps->isolate_vl_tc = MLX5_CAP_GEN(mdev, isolate_vl_tc_new);
+
 	if (caps->flex_protocols & MLX5_FLEX_PARSER_ICMP_V4_ENABLED) {
 		caps->flex_parser_id_icmp_dw0 = MLX5_CAP_GEN(mdev, flex_parser_id_icmp_dw0);
 		caps->flex_parser_id_icmp_dw1 = MLX5_CAP_GEN(mdev, flex_parser_id_icmp_dw1);
@@ -106,6 +144,34 @@ int mlx5dr_cmd_query_device(struct mlx5_core_dev *mdev,
 			MLX5_CAP_GEN(mdev, flex_parser_id_icmpv6_dw1);
 	}
 
+	if (caps->flex_protocols & MLX5_FLEX_PARSER_GENEVE_TLV_OPTION_0_ENABLED)
+		caps->flex_parser_id_geneve_tlv_option_0 =
+			MLX5_CAP_GEN(mdev, flex_parser_id_geneve_tlv_option_0);
+
+	if (caps->flex_protocols & MLX5_FLEX_PARSER_MPLS_OVER_GRE_ENABLED)
+		caps->flex_parser_id_mpls_over_gre =
+			MLX5_CAP_GEN(mdev, flex_parser_id_outer_first_mpls_over_gre);
+
+	if (caps->flex_protocols & mlx5_FLEX_PARSER_MPLS_OVER_UDP_ENABLED)
+		caps->flex_parser_id_mpls_over_udp =
+			MLX5_CAP_GEN(mdev, flex_parser_id_outer_first_mpls_over_udp_label);
+
+	if (caps->flex_protocols & MLX5_FLEX_PARSER_GTPU_DW_0_ENABLED)
+		caps->flex_parser_id_gtpu_dw_0 =
+			MLX5_CAP_GEN(mdev, flex_parser_id_gtpu_dw_0);
+
+	if (caps->flex_protocols & MLX5_FLEX_PARSER_GTPU_TEID_ENABLED)
+		caps->flex_parser_id_gtpu_teid =
+			MLX5_CAP_GEN(mdev, flex_parser_id_gtpu_teid);
+
+	if (caps->flex_protocols & MLX5_FLEX_PARSER_GTPU_DW_2_ENABLED)
+		caps->flex_parser_id_gtpu_dw_2 =
+			MLX5_CAP_GEN(mdev, flex_parser_id_gtpu_dw_2);
+
+	if (caps->flex_protocols & MLX5_FLEX_PARSER_GTPU_FIRST_EXT_DW_0_ENABLED)
+		caps->flex_parser_id_gtpu_first_ext_dw_0 =
+			MLX5_CAP_GEN(mdev, flex_parser_id_gtpu_first_ext_dw_0);
+
 	caps->nic_rx_drop_address =
 		MLX5_CAP64_FLOWTABLE(mdev, sw_steering_nic_rx_action_drop_icm_address);
 	caps->nic_tx_drop_address =

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_rule.c

@@ -952,6 +952,17 @@ static bool dr_rule_verify(struct mlx5dr_matcher *matcher,
 			return false;
 		}
 	}
+
+	if (match_criteria & DR_MATCHER_CRITERIA_MISC4) {
+		s_idx = offsetof(struct mlx5dr_match_param, misc4);
+		e_idx = min(s_idx + sizeof(param->misc4), value_size);
+
+		if (!dr_rule_cmp_value_to_mask(mask_p, param_p, s_idx, e_idx)) {
+			mlx5dr_err(matcher->tbl->dmn,
+				   "Rule misc4 parameters contains a value not specified by mask\n");
+			return false;
+		}
+	}
 	return true;
 }
 

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_send.c

@@ -32,6 +32,7 @@ struct dr_qp_rtr_attr {
 	u8 min_rnr_timer;
 	u8 sgid_index;
 	u16 udp_src_port;
+	u8 fl:1;
 };
 
 struct dr_qp_rts_attr {
@@ -45,6 +46,7 @@ struct dr_qp_init_attr {
 	u32 pdn;
 	u32 max_send_wr;
 	struct mlx5_uars_page *uar;
+	u8 isolate_vl_tc:1;
 };
 
 static int dr_parse_cqe(struct mlx5dr_cq *dr_cq, struct mlx5_cqe64 *cqe64)
@@ -157,6 +159,7 @@ static struct mlx5dr_qp *dr_create_rc_qp(struct mlx5_core_dev *mdev,
 	qpc = MLX5_ADDR_OF(create_qp_in, in, qpc);
 	MLX5_SET(qpc, qpc, st, MLX5_QP_ST_RC);
 	MLX5_SET(qpc, qpc, pm_state, MLX5_QP_PM_MIGRATED);
+	MLX5_SET(qpc, qpc, isolate_vl_tc, attr->isolate_vl_tc);
 	MLX5_SET(qpc, qpc, pd, attr->pdn);
 	MLX5_SET(qpc, qpc, uar_page, attr->uar->index);
 	MLX5_SET(qpc, qpc, log_page_size,
@@ -213,7 +216,7 @@ static void dr_destroy_qp(struct mlx5_core_dev *mdev,
 static void dr_cmd_notify_hw(struct mlx5dr_qp *dr_qp, void *ctrl)
 {
 	dma_wmb();
-	*dr_qp->wq.sq.db = cpu_to_be32(dr_qp->sq.pc & 0xfffff);
+	*dr_qp->wq.sq.db = cpu_to_be32(dr_qp->sq.pc & 0xffff);
 
 	/* After wmb() the hw aware of new work */
 	wmb();
@@ -223,7 +226,7 @@ static void dr_cmd_notify_hw(struct mlx5dr_qp *dr_qp, void *ctrl)
 
 static void dr_rdma_segments(struct mlx5dr_qp *dr_qp, u64 remote_addr,
 			     u32 rkey, struct dr_data_seg *data_seg,
-			     u32 opcode, int nreq)
+			     u32 opcode, bool notify_hw)
 {
 	struct mlx5_wqe_raddr_seg *wq_raddr;
 	struct mlx5_wqe_ctrl_seg *wq_ctrl;
@@ -255,16 +258,16 @@ static void dr_rdma_segments(struct mlx5dr_qp *dr_qp, u64 remote_addr,
 
 	dr_qp->sq.wqe_head[idx] = dr_qp->sq.pc++;
 
-	if (nreq)
+	if (notify_hw)
 		dr_cmd_notify_hw(dr_qp, wq_ctrl);
 }
 
 static void dr_post_send(struct mlx5dr_qp *dr_qp, struct postsend_info *send_info)
 {
 	dr_rdma_segments(dr_qp, send_info->remote_addr, send_info->rkey,
-			 &send_info->write, MLX5_OPCODE_RDMA_WRITE, 0);
+			 &send_info->write, MLX5_OPCODE_RDMA_WRITE, false);
 	dr_rdma_segments(dr_qp, send_info->remote_addr, send_info->rkey,
-			 &send_info->read, MLX5_OPCODE_RDMA_READ, 1);
+			 &send_info->read, MLX5_OPCODE_RDMA_READ, true);
 }
 
 /**
@@ -650,6 +653,7 @@ static int dr_cmd_modify_qp_init2rtr(struct mlx5_core_dev *mdev,
 			 attr->udp_src_port);
 
 	MLX5_SET(qpc, qpc, primary_address_path.vhca_port_num, attr->port_num);
+	MLX5_SET(qpc, qpc, primary_address_path.fl, attr->fl);
 	MLX5_SET(qpc, qpc, min_rnr_nak, 1);
 
 	MLX5_SET(init2rtr_qp_in, in, opcode, MLX5_CMD_OP_INIT2RTR_QP);
@@ -658,6 +662,19 @@ static int dr_cmd_modify_qp_init2rtr(struct mlx5_core_dev *mdev,
 	return mlx5_cmd_exec_in(mdev, init2rtr_qp, in);
 }
 
+static bool dr_send_allow_fl(struct mlx5dr_cmd_caps *caps)
+{
+	/* Check whether RC RoCE QP creation with force loopback is allowed.
+	 * There are two separate capability bits for this:
+	 *  - force loopback when RoCE is enabled
+	 *  - force loopback when RoCE is disabled
+	 */
+	return ((caps->roce_caps.roce_en &&
+		 caps->roce_caps.fl_rc_qp_when_roce_enabled) ||
+		(!caps->roce_caps.roce_en &&
+		 caps->roce_caps.fl_rc_qp_when_roce_disabled));
+}
+
 static int dr_prepare_qp_to_rts(struct mlx5dr_domain *dmn)
 {
 	struct mlx5dr_qp *dr_qp = dmn->send_ring->qp;
@@ -676,17 +693,26 @@ static int dr_prepare_qp_to_rts(struct mlx5dr_domain *dmn)
 	}
 
 	/* RTR */
-	ret = mlx5dr_cmd_query_gid(dmn->mdev, port, gid_index, &rtr_attr.dgid_attr);
-	if (ret)
-		return ret;
-
 	rtr_attr.mtu		= mtu;
 	rtr_attr.qp_num		= dr_qp->qpn;
 	rtr_attr.min_rnr_timer	= 12;
 	rtr_attr.port_num	= port;
-	rtr_attr.sgid_index	= gid_index;
 	rtr_attr.udp_src_port	= dmn->info.caps.roce_min_src_udp;
 
+	/* If QP creation with force loopback is allowed, then there
+	 * is no need for GID index when creating the QP.
+	 * Otherwise we query GID attributes and use GID index.
+	 */
+	rtr_attr.fl = dr_send_allow_fl(&dmn->info.caps);
+	if (!rtr_attr.fl) {
+		ret = mlx5dr_cmd_query_gid(dmn->mdev, port, gid_index,
+					   &rtr_attr.dgid_attr);
+		if (ret)
+			return ret;
+
+		rtr_attr.sgid_index = gid_index;
+	}
+
 	ret = dr_cmd_modify_qp_init2rtr(dmn->mdev, dr_qp, &rtr_attr);
 	if (ret) {
 		mlx5dr_err(dmn, "Failed modify QP init2rtr\n");
@@ -900,6 +926,11 @@ int mlx5dr_send_ring_alloc(struct mlx5dr_domain *dmn)
 	init_attr.pdn = dmn->pdn;
 	init_attr.uar = dmn->uar;
 	init_attr.max_send_wr = QUEUE_SIZE;
+
+	/* Isolated VL is applicable only if force loopback is supported */
+	if (dr_send_allow_fl(&dmn->info.caps))
+		init_attr.isolate_vl_tc = dmn->info.caps.isolate_vl_tc;
+
 	spin_lock_init(&dmn->send_ring->lock);
 
 	dmn->send_ring->qp = dr_create_rc_qp(dmn->mdev, &init_attr);

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c

@@ -852,6 +852,35 @@ static void dr_ste_copy_mask_misc3(char *mask, struct mlx5dr_match_misc3 *spec)
 	spec->icmpv4_code = MLX5_GET(fte_match_set_misc3, mask, icmp_code);
 	spec->icmpv6_type = MLX5_GET(fte_match_set_misc3, mask, icmpv6_type);
 	spec->icmpv6_code = MLX5_GET(fte_match_set_misc3, mask, icmpv6_code);
+	spec->geneve_tlv_option_0_data =
+		MLX5_GET(fte_match_set_misc3, mask, geneve_tlv_option_0_data);
+	spec->gtpu_msg_flags = MLX5_GET(fte_match_set_misc3, mask, gtpu_msg_flags);
+	spec->gtpu_msg_type = MLX5_GET(fte_match_set_misc3, mask, gtpu_msg_type);
+	spec->gtpu_teid = MLX5_GET(fte_match_set_misc3, mask, gtpu_teid);
+	spec->gtpu_dw_0 = MLX5_GET(fte_match_set_misc3, mask, gtpu_dw_0);
+	spec->gtpu_dw_2 = MLX5_GET(fte_match_set_misc3, mask, gtpu_dw_2);
+	spec->gtpu_first_ext_dw_0 =
+		MLX5_GET(fte_match_set_misc3, mask, gtpu_first_ext_dw_0);
+}
+
+static void dr_ste_copy_mask_misc4(char *mask, struct mlx5dr_match_misc4 *spec)
+{
+	spec->prog_sample_field_id_0 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_id_0);
+	spec->prog_sample_field_value_0 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_value_0);
+	spec->prog_sample_field_id_1 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_id_1);
+	spec->prog_sample_field_value_1 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_value_1);
+	spec->prog_sample_field_id_2 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_id_2);
+	spec->prog_sample_field_value_2 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_value_2);
+	spec->prog_sample_field_id_3 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_id_3);
+	spec->prog_sample_field_value_3 =
+		MLX5_GET(fte_match_set_misc4, mask, prog_sample_field_value_3);
 }
 
 void mlx5dr_ste_copy_param(u8 match_criteria,
@@ -925,6 +954,20 @@ void mlx5dr_ste_copy_param(u8 match_criteria,
 		}
 		dr_ste_copy_mask_misc3(buff, &set_param->misc3);
 	}
+
+	param_location += sizeof(struct mlx5dr_match_misc3);
+
+	if (match_criteria & DR_MATCHER_CRITERIA_MISC4) {
+		if (mask->match_sz < param_location +
+		    sizeof(struct mlx5dr_match_misc4)) {
+			memcpy(tail_param, data + param_location,
+			       mask->match_sz - param_location);
+			buff = tail_param;
+		} else {
+			buff = data + param_location;
+		}
+		dr_ste_copy_mask_misc4(buff, &set_param->misc4);
+	}
 }
 
 void mlx5dr_ste_build_eth_l2_src_dst(struct mlx5dr_ste_ctx *ste_ctx,
@@ -1051,17 +1094,31 @@ void mlx5dr_ste_build_tnl_gre(struct mlx5dr_ste_ctx *ste_ctx,
 	ste_ctx->build_tnl_gre_init(sb, mask);
 }
 
-void mlx5dr_ste_build_tnl_mpls(struct mlx5dr_ste_ctx *ste_ctx,
+void mlx5dr_ste_build_tnl_mpls_over_gre(struct mlx5dr_ste_ctx *ste_ctx,
+					struct mlx5dr_ste_build *sb,
+					struct mlx5dr_match_param *mask,
+					struct mlx5dr_cmd_caps *caps,
+					bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->inner = inner;
+	sb->caps = caps;
+	return ste_ctx->build_tnl_mpls_over_gre_init(sb, mask);
+}
+
+void mlx5dr_ste_build_tnl_mpls_over_udp(struct mlx5dr_ste_ctx *ste_ctx,
 					struct mlx5dr_ste_build *sb,
 					struct mlx5dr_match_param *mask,
+					struct mlx5dr_cmd_caps *caps,
 					bool inner, bool rx)
 {
 	sb->rx = rx;
 	sb->inner = inner;
-	ste_ctx->build_tnl_mpls_init(sb, mask);
+	sb->caps = caps;
+	return ste_ctx->build_tnl_mpls_over_udp_init(sb, mask);
 }
 
-int mlx5dr_ste_build_icmp(struct mlx5dr_ste_ctx *ste_ctx,
+void mlx5dr_ste_build_icmp(struct mlx5dr_ste_ctx *ste_ctx,
 			   struct mlx5dr_ste_build *sb,
 			   struct mlx5dr_match_param *mask,
 			   struct mlx5dr_cmd_caps *caps,
@@ -1070,7 +1127,7 @@ int mlx5dr_ste_build_icmp(struct mlx5dr_ste_ctx *ste_ctx,
 	sb->rx = rx;
 	sb->inner = inner;
 	sb->caps = caps;
-	return ste_ctx->build_icmp_init(sb, mask);
+	ste_ctx->build_icmp_init(sb, mask);
 }
 
 void mlx5dr_ste_build_general_purpose(struct mlx5dr_ste_ctx *ste_ctx,
@@ -1113,6 +1170,52 @@ void mlx5dr_ste_build_tnl_geneve(struct mlx5dr_ste_ctx *ste_ctx,
 	ste_ctx->build_tnl_geneve_init(sb, mask);
 }
 
+void mlx5dr_ste_build_tnl_geneve_tlv_opt(struct mlx5dr_ste_ctx *ste_ctx,
+					 struct mlx5dr_ste_build *sb,
+					 struct mlx5dr_match_param *mask,
+					 struct mlx5dr_cmd_caps *caps,
+					 bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->caps = caps;
+	sb->inner = inner;
+	ste_ctx->build_tnl_geneve_tlv_opt_init(sb, mask);
+}
+
+void mlx5dr_ste_build_tnl_gtpu(struct mlx5dr_ste_ctx *ste_ctx,
+			       struct mlx5dr_ste_build *sb,
+			       struct mlx5dr_match_param *mask,
+			       bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->inner = inner;
+	ste_ctx->build_tnl_gtpu_init(sb, mask);
+}
+
+void mlx5dr_ste_build_tnl_gtpu_flex_parser_0(struct mlx5dr_ste_ctx *ste_ctx,
+					     struct mlx5dr_ste_build *sb,
+					     struct mlx5dr_match_param *mask,
+					     struct mlx5dr_cmd_caps *caps,
+					     bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->caps = caps;
+	sb->inner = inner;
+	ste_ctx->build_tnl_gtpu_flex_parser_0_init(sb, mask);
+}
+
+void mlx5dr_ste_build_tnl_gtpu_flex_parser_1(struct mlx5dr_ste_ctx *ste_ctx,
+					     struct mlx5dr_ste_build *sb,
+					     struct mlx5dr_match_param *mask,
+					     struct mlx5dr_cmd_caps *caps,
+					     bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->caps = caps;
+	sb->inner = inner;
+	ste_ctx->build_tnl_gtpu_flex_parser_1_init(sb, mask);
+}
+
 void mlx5dr_ste_build_register_0(struct mlx5dr_ste_ctx *ste_ctx,
 				 struct mlx5dr_ste_build *sb,
 				 struct mlx5dr_match_param *mask,
@@ -1148,6 +1251,26 @@ void mlx5dr_ste_build_src_gvmi_qpn(struct mlx5dr_ste_ctx *ste_ctx,
 	ste_ctx->build_src_gvmi_qpn_init(sb, mask);
 }
 
+void mlx5dr_ste_build_flex_parser_0(struct mlx5dr_ste_ctx *ste_ctx,
+				    struct mlx5dr_ste_build *sb,
+				    struct mlx5dr_match_param *mask,
+				    bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->inner = inner;
+	ste_ctx->build_flex_parser_0_init(sb, mask);
+}
+
+void mlx5dr_ste_build_flex_parser_1(struct mlx5dr_ste_ctx *ste_ctx,
+				    struct mlx5dr_ste_build *sb,
+				    struct mlx5dr_match_param *mask,
+				    bool inner, bool rx)
+{
+	sb->rx = rx;
+	sb->inner = inner;
+	ste_ctx->build_flex_parser_1_init(sb, mask);
+}
+
 static struct mlx5dr_ste_ctx *mlx5dr_ste_ctx_arr[] = {
 	[MLX5_STEERING_FORMAT_CONNECTX_5] = &ste_ctx_v0,
 	[MLX5_STEERING_FORMAT_CONNECTX_6DX] = &ste_ctx_v1,

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.h

@@ -62,6 +62,13 @@
 		       in_out##_first_mpls_ttl); \
 } while (0)
 
+#define DR_STE_SET_FLEX_PARSER_FIELD(tag, fname, caps, spec) do { \
+	u8 parser_id = (caps)->flex_parser_id_##fname; \
+	u8 *parser_ptr = dr_ste_calc_flex_parser_offset(tag, parser_id); \
+	*(__be32 *)parser_ptr = cpu_to_be32((spec)->fname);\
+	(spec)->fname = 0;\
+} while (0)
+
 #define DR_STE_IS_OUTER_MPLS_OVER_GRE_SET(_misc) (\
 	(_misc)->outer_first_mpls_over_gre_label || \
 	(_misc)->outer_first_mpls_over_gre_exp || \
@@ -86,8 +93,22 @@ enum dr_ste_action_modify_type_l4 {
 	DR_STE_ACTION_MDFY_TYPE_L4_UDP	= 0x2,
 };
 
+enum {
+	HDR_MPLS_OFFSET_LABEL	= 12,
+	HDR_MPLS_OFFSET_EXP	= 9,
+	HDR_MPLS_OFFSET_S_BOS	= 8,
+	HDR_MPLS_OFFSET_TTL	= 0,
+};
+
 u16 mlx5dr_ste_conv_bit_to_byte_mask(u8 *bit_mask);
 
+static inline u8 *
+dr_ste_calc_flex_parser_offset(u8 *tag, u8 parser_id)
+{
+	/* Calculate tag byte offset based on flex parser id */
+	return tag + 4 * (3 - (parser_id % 4));
+}
+
 #define DR_STE_CTX_BUILDER(fname) \
 	((*build_##fname##_init)(struct mlx5dr_ste_build *sb, \
 				 struct mlx5dr_match_param *mask))
@@ -106,14 +127,22 @@ struct mlx5dr_ste_ctx {
 	void DR_STE_CTX_BUILDER(mpls);
 	void DR_STE_CTX_BUILDER(tnl_gre);
 	void DR_STE_CTX_BUILDER(tnl_mpls);
-	int  DR_STE_CTX_BUILDER(icmp);
+	void DR_STE_CTX_BUILDER(tnl_mpls_over_gre);
+	void DR_STE_CTX_BUILDER(tnl_mpls_over_udp);
+	void DR_STE_CTX_BUILDER(icmp);
 	void DR_STE_CTX_BUILDER(general_purpose);
 	void DR_STE_CTX_BUILDER(eth_l4_misc);
 	void DR_STE_CTX_BUILDER(tnl_vxlan_gpe);
 	void DR_STE_CTX_BUILDER(tnl_geneve);
+	void DR_STE_CTX_BUILDER(tnl_geneve_tlv_opt);
 	void DR_STE_CTX_BUILDER(register_0);
 	void DR_STE_CTX_BUILDER(register_1);
 	void DR_STE_CTX_BUILDER(src_gvmi_qpn);
+	void DR_STE_CTX_BUILDER(flex_parser_0);
+	void DR_STE_CTX_BUILDER(flex_parser_1);
+	void DR_STE_CTX_BUILDER(tnl_gtpu);
+	void DR_STE_CTX_BUILDER(tnl_gtpu_flex_parser_0);
+	void DR_STE_CTX_BUILDER(tnl_gtpu_flex_parser_1);
 
 	/* Getters and Setters */
 	void (*ste_init)(u8 *hw_ste_p, u16 lu_type,

drivers/net/ethernet/mellanox/mlx5/core/steering/dr_types.h

@@ -12,17 +12,30 @@
 #include "mlx5_ifc_dr.h"
 #include "mlx5dr.h"
 
-#define DR_RULE_MAX_STES 17
+#define DR_RULE_MAX_STES 18
 #define DR_ACTION_MAX_STES 5
 #define WIRE_PORT 0xFFFF
 #define DR_STE_SVLAN 0x1
 #define DR_STE_CVLAN 0x2
 #define DR_SZ_MATCH_PARAM (MLX5_ST_SZ_DW_MATCH_PARAM * 4)
+#define DR_NUM_OF_FLEX_PARSERS 8
+#define DR_STE_MAX_FLEX_0_ID 3
+#define DR_STE_MAX_FLEX_1_ID 7
 
 #define mlx5dr_err(dmn, arg...) mlx5_core_err((dmn)->mdev, ##arg)
 #define mlx5dr_info(dmn, arg...) mlx5_core_info((dmn)->mdev, ##arg)
 #define mlx5dr_dbg(dmn, arg...) mlx5_core_dbg((dmn)->mdev, ##arg)
 
+static inline bool dr_is_flex_parser_0_id(u8 parser_id)
+{
+	return parser_id <= DR_STE_MAX_FLEX_0_ID;
+}
+
+static inline bool dr_is_flex_parser_1_id(u8 parser_id)
+{
+	return parser_id > DR_STE_MAX_FLEX_0_ID;
+}
+
 enum mlx5dr_icm_chunk_size {
 	DR_CHUNK_SIZE_1,
 	DR_CHUNK_SIZE_MIN = DR_CHUNK_SIZE_1, /* keep updated when changing */
@@ -87,7 +100,8 @@ enum mlx5dr_matcher_criteria {
 	DR_MATCHER_CRITERIA_INNER = 1 << 2,
 	DR_MATCHER_CRITERIA_MISC2 = 1 << 3,
 	DR_MATCHER_CRITERIA_MISC3 = 1 << 4,
-	DR_MATCHER_CRITERIA_MAX = 1 << 5,
+	DR_MATCHER_CRITERIA_MISC4 = 1 << 5,
+	DR_MATCHER_CRITERIA_MAX = 1 << 6,
 };
 
 enum mlx5dr_action_type {
@@ -389,7 +403,17 @@ void mlx5dr_ste_build_tnl_mpls(struct mlx5dr_ste_ctx *ste_ctx,
 			       struct mlx5dr_ste_build *sb,
 			       struct mlx5dr_match_param *mask,
 			       bool inner, bool rx);
-int mlx5dr_ste_build_icmp(struct mlx5dr_ste_ctx *ste_ctx,
+void mlx5dr_ste_build_tnl_mpls_over_gre(struct mlx5dr_ste_ctx *ste_ctx,
+					struct mlx5dr_ste_build *sb,
+					struct mlx5dr_match_param *mask,
+					struct mlx5dr_cmd_caps *caps,
+					bool inner, bool rx);
+void mlx5dr_ste_build_tnl_mpls_over_udp(struct mlx5dr_ste_ctx *ste_ctx,
+					struct mlx5dr_ste_build *sb,
+					struct mlx5dr_match_param *mask,
+					struct mlx5dr_cmd_caps *caps,
+					bool inner, bool rx);
+void mlx5dr_ste_build_icmp(struct mlx5dr_ste_ctx *ste_ctx,
 			   struct mlx5dr_ste_build *sb,
 			   struct mlx5dr_match_param *mask,
 			   struct mlx5dr_cmd_caps *caps,
@@ -402,6 +426,25 @@ void mlx5dr_ste_build_tnl_geneve(struct mlx5dr_ste_ctx *ste_ctx,
 				 struct mlx5dr_ste_build *sb,
 				 struct mlx5dr_match_param *mask,
 				 bool inner, bool rx);
+void mlx5dr_ste_build_tnl_geneve_tlv_opt(struct mlx5dr_ste_ctx *ste_ctx,
+					 struct mlx5dr_ste_build *sb,
+					 struct mlx5dr_match_param *mask,
+					 struct mlx5dr_cmd_caps *caps,
+					 bool inner, bool rx);
+void mlx5dr_ste_build_tnl_gtpu(struct mlx5dr_ste_ctx *ste_ctx,
+			       struct mlx5dr_ste_build *sb,
+			       struct mlx5dr_match_param *mask,
+			       bool inner, bool rx);
+void mlx5dr_ste_build_tnl_gtpu_flex_parser_0(struct mlx5dr_ste_ctx *ste_ctx,
+					     struct mlx5dr_ste_build *sb,
+					     struct mlx5dr_match_param *mask,
+					     struct mlx5dr_cmd_caps *caps,
+					     bool inner, bool rx);
+void mlx5dr_ste_build_tnl_gtpu_flex_parser_1(struct mlx5dr_ste_ctx *ste_ctx,
+					     struct mlx5dr_ste_build *sb,
+					     struct mlx5dr_match_param *mask,
+					     struct mlx5dr_cmd_caps *caps,
+					     bool inner, bool rx);
 void mlx5dr_ste_build_general_purpose(struct mlx5dr_ste_ctx *ste_ctx,
 				      struct mlx5dr_ste_build *sb,
 				      struct mlx5dr_match_param *mask,
@@ -419,6 +462,14 @@ void mlx5dr_ste_build_src_gvmi_qpn(struct mlx5dr_ste_ctx *ste_ctx,
 				   struct mlx5dr_match_param *mask,
 				   struct mlx5dr_domain *dmn,
 				   bool inner, bool rx);
+void mlx5dr_ste_build_flex_parser_0(struct mlx5dr_ste_ctx *ste_ctx,
+				    struct mlx5dr_ste_build *sb,
+				    struct mlx5dr_match_param *mask,
+				    bool inner, bool rx);
+void mlx5dr_ste_build_flex_parser_1(struct mlx5dr_ste_ctx *ste_ctx,
+				    struct mlx5dr_ste_build *sb,
+				    struct mlx5dr_match_param *mask,
+				    bool inner, bool rx);
 void mlx5dr_ste_build_empty_always_hit(struct mlx5dr_ste_build *sb, bool rx);
 
 /* Actions utils */
@@ -646,7 +697,24 @@ struct mlx5dr_match_misc3 {
 	u8 icmpv6_type;
 	u8 icmpv4_code;
 	u8 icmpv4_type;
-	u8 reserved_auto3[0x1c];
+	u32 geneve_tlv_option_0_data;
+	u8 gtpu_msg_flags;
+	u8 gtpu_msg_type;
+	u32 gtpu_teid;
+	u32 gtpu_dw_2;
+	u32 gtpu_first_ext_dw_0;
+	u32 gtpu_dw_0;
+};
+
+struct mlx5dr_match_misc4 {
+	u32 prog_sample_field_value_0;
+	u32 prog_sample_field_id_0;
+	u32 prog_sample_field_value_1;
+	u32 prog_sample_field_id_1;
+	u32 prog_sample_field_value_2;
+	u32 prog_sample_field_id_2;
+	u32 prog_sample_field_value_3;
+	u32 prog_sample_field_id_3;
 };
 
 struct mlx5dr_match_param {
@@ -655,6 +723,7 @@ struct mlx5dr_match_param {
 	struct mlx5dr_match_spec inner;
 	struct mlx5dr_match_misc2 misc2;
 	struct mlx5dr_match_misc3 misc3;
+	struct mlx5dr_match_misc4 misc4;
 };
 
 #define DR_MASK_IS_ICMPV4_SET(_misc3) ((_misc3)->icmpv4_type || \
@@ -678,6 +747,12 @@ struct mlx5dr_cmd_vport_cap {
 	u32 num;
 };
 
+struct mlx5dr_roce_cap {
+	u8 roce_en:1;
+	u8 fl_rc_qp_when_roce_disabled:1;
+	u8 fl_rc_qp_when_roce_enabled:1;
+};
+
 struct mlx5dr_cmd_caps {
 	u16 gvmi;
 	u64 nic_rx_drop_address;
@@ -692,6 +767,13 @@ struct mlx5dr_cmd_caps {
 	u8 flex_parser_id_icmp_dw1;
 	u8 flex_parser_id_icmpv6_dw0;
 	u8 flex_parser_id_icmpv6_dw1;
+	u8 flex_parser_id_geneve_tlv_option_0;
+	u8 flex_parser_id_mpls_over_gre;
+	u8 flex_parser_id_mpls_over_udp;
+	u8 flex_parser_id_gtpu_dw_0;
+	u8 flex_parser_id_gtpu_teid;
+	u8 flex_parser_id_gtpu_dw_2;
+	u8 flex_parser_id_gtpu_first_ext_dw_0;
 	u8 max_ft_level;
 	u16 roce_min_src_udp;
 	u8 num_esw_ports;
@@ -707,6 +789,8 @@ struct mlx5dr_cmd_caps {
 	struct mlx5dr_esw_caps esw_caps;
 	struct mlx5dr_cmd_vport_cap *vports_caps;
 	bool prio_tag_required;
+	struct mlx5dr_roce_cap roce_caps;
+	u8 isolate_vl_tc:1;
 };
 
 struct mlx5dr_domain_rx_tx {
@@ -1081,6 +1165,7 @@ struct mlx5dr_cmd_qp_create_attr {
 	u32 sq_wqe_cnt;
 	u32 rq_wqe_cnt;
 	u32 rq_wqe_shift;
+	u8 isolate_vl_tc:1;
 };
 
 int mlx5dr_cmd_query_gid(struct mlx5_core_dev *mdev, u8 vhca_port_num,

drivers/net/ethernet/mellanox/mlx5/core/steering/mlx5_ifc_dr.h

@@ -434,10 +434,7 @@ struct mlx5_ifc_ste_gre_bits {
 };
 
 struct mlx5_ifc_ste_flex_parser_0_bits {
-	u8         parser_3_label[0x14];
-	u8         parser_3_exp[0x3];
-	u8         parser_3_s_bos[0x1];
-	u8         parser_3_ttl[0x8];
+	u8         flex_parser_3[0x20];
 
 	u8         flex_parser_2[0x20];
 
@@ -488,6 +485,17 @@ struct mlx5_ifc_ste_flex_parser_tnl_geneve_bits {
 	u8         reserved_at_40[0x40];
 };
 
+struct mlx5_ifc_ste_flex_parser_tnl_gtpu_bits {
+	u8	   reserved_at_0[0x5];
+	u8	   gtpu_msg_flags[0x3];
+	u8	   gtpu_msg_type[0x8];
+	u8	   reserved_at_10[0x10];
+
+	u8	   gtpu_teid[0x20];
+
+	u8	   reserved_at_40[0x40];
+};
+
 struct mlx5_ifc_ste_general_purpose_bits {
 	u8         general_purpose_lookup_field[0x20];
 

include/linux/mlx5/device.h

@@ -911,8 +911,11 @@ static inline u16 get_cqe_flow_tag(struct mlx5_cqe64 *cqe)
 	return be32_to_cpu(cqe->sop_drop_qpn) & 0xFFF;
 }
 
-#define MLX5_MPWQE_LOG_NUM_STRIDES_BASE	(9)
-#define MLX5_MPWQE_LOG_STRIDE_SZ_BASE	(6)
+#define MLX5_MPWQE_LOG_NUM_STRIDES_EXT_BASE	3
+#define MLX5_MPWQE_LOG_NUM_STRIDES_BASE		9
+#define MLX5_MPWQE_LOG_NUM_STRIDES_MAX		16
+#define MLX5_MPWQE_LOG_STRIDE_SZ_BASE		6
+#define MLX5_MPWQE_LOG_STRIDE_SZ_MAX		13
 
 struct mpwrq_cqe_bc {
 	__be16	filler_consumed_strides;

include/linux/mlx5/mlx5_ifc.h

@@ -622,7 +622,19 @@ struct mlx5_ifc_fte_match_set_misc3_bits {
 
 	u8         geneve_tlv_option_0_data[0x20];
 
-	u8         reserved_at_140[0xc0];
+	u8	   gtpu_teid[0x20];
+
+	u8	   gtpu_msg_type[0x8];
+	u8	   gtpu_msg_flags[0x8];
+	u8	   reserved_at_170[0x10];
+
+	u8	   gtpu_dw_2[0x20];
+
+	u8	   gtpu_first_ext_dw_0[0x20];
+
+	u8	   gtpu_dw_0[0x20];
+
+	u8	   reserved_at_1e0[0x20];
 };
 
 struct mlx5_ifc_fte_match_set_misc4_bits {
@@ -949,7 +961,9 @@ struct mlx5_ifc_roce_cap_bits {
 	u8         roce_apm[0x1];
 	u8         reserved_at_1[0x3];
 	u8         sw_r_roce_src_udp_port[0x1];
-	u8         reserved_at_5[0x19];
+	u8         fl_rc_qp_when_roce_disabled[0x1];
+	u8         fl_rc_qp_when_roce_enabled[0x1];
+	u8         reserved_at_7[0x17];
 	u8	   qp_ts_format[0x2];
 
 	u8         reserved_at_20[0x60];
@@ -1237,9 +1251,17 @@ enum {
 
 enum {
 	MLX5_FLEX_PARSER_GENEVE_ENABLED		= 1 << 3,
+	MLX5_FLEX_PARSER_MPLS_OVER_GRE_ENABLED	= 1 << 4,
+	mlx5_FLEX_PARSER_MPLS_OVER_UDP_ENABLED	= 1 << 5,
 	MLX5_FLEX_PARSER_VXLAN_GPE_ENABLED	= 1 << 7,
 	MLX5_FLEX_PARSER_ICMP_V4_ENABLED	= 1 << 8,
 	MLX5_FLEX_PARSER_ICMP_V6_ENABLED	= 1 << 9,
+	MLX5_FLEX_PARSER_GENEVE_TLV_OPTION_0_ENABLED = 1 << 10,
+	MLX5_FLEX_PARSER_GTPU_ENABLED		= 1 << 11,
+	MLX5_FLEX_PARSER_GTPU_DW_2_ENABLED	= 1 << 16,
+	MLX5_FLEX_PARSER_GTPU_FIRST_EXT_DW_0_ENABLED = 1 << 17,
+	MLX5_FLEX_PARSER_GTPU_DW_0_ENABLED	= 1 << 18,
+	MLX5_FLEX_PARSER_GTPU_TEID_ENABLED	= 1 << 19,
 };
 
 enum {
@@ -1297,7 +1319,9 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         log_max_srq_sz[0x8];
 	u8         log_max_qp_sz[0x8];
 	u8         event_cap[0x1];
-	u8         reserved_at_91[0x7];
+	u8         reserved_at_91[0x2];
+	u8         isolate_vl_tc_new[0x1];
+	u8         reserved_at_94[0x4];
 	u8         prio_tag_required[0x1];
 	u8         reserved_at_99[0x2];
 	u8         log_max_qp[0x5];
@@ -1637,7 +1661,9 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         cqe_compression_timeout[0x10];
 	u8         cqe_compression_max_num[0x10];
 
-	u8         reserved_at_5e0[0x10];
+	u8         reserved_at_5e0[0x8];
+	u8         flex_parser_id_gtpu_dw_0[0x4];
+	u8         reserved_at_5ec[0x4];
 	u8         tag_matching[0x1];
 	u8         rndv_offload_rc[0x1];
 	u8         rndv_offload_dc[0x1];
@@ -1648,7 +1674,8 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8	   affiliate_nic_vport_criteria[0x8];
 	u8	   native_port_num[0x8];
 	u8	   num_vhca_ports[0x8];
-	u8	   reserved_at_618[0x6];
+	u8         flex_parser_id_gtpu_teid[0x4];
+	u8         reserved_at_61c[0x2];
 	u8	   sw_owner_id[0x1];
 	u8         reserved_at_61f[0x1];
 
@@ -1683,7 +1710,8 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8	   reserved_at_6e0[0x10];
 	u8	   sf_base_id[0x10];
 
-	u8	   reserved_at_700[0x8];
+	u8         flex_parser_id_gtpu_dw_2[0x4];
+	u8         flex_parser_id_gtpu_first_ext_dw_0[0x4];
 	u8	   num_total_dynamic_vf_msix[0x18];
 	u8	   reserved_at_720[0x14];
 	u8	   dynamic_msix_table_size[0xc];
@@ -2918,7 +2946,8 @@ struct mlx5_ifc_qpc_bits {
 	u8         state[0x4];
 	u8         lag_tx_port_affinity[0x4];
 	u8         st[0x8];
-	u8         reserved_at_10[0x3];
+	u8         reserved_at_10[0x2];
+	u8	   isolate_vl_tc[0x1];
 	u8         pm_state[0x2];
 	u8         reserved_at_15[0x1];
 	u8         req_e2e_credit_mode[0x2];