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

Pull networking updates from David Miller:

 1) Support offloading wireless authentication to userspace via
    NL80211_CMD_EXTERNAL_AUTH, from Srinivas Dasari.

 2) A lot of work on network namespace setup/teardown from Kirill Tkhai.
    Setup and cleanup of namespaces now all run asynchronously and thus
    performance is significantly increased.

 3) Add rx/tx timestamping support to mv88e6xxx driver, from Brandon
    Streiff.

 4) Support zerocopy on RDS sockets, from Sowmini Varadhan.

 5) Use denser instruction encoding in x86 eBPF JIT, from Daniel
    Borkmann.

 6) Support hw offload of vlan filtering in mvpp2 dreiver, from Maxime
    Chevallier.

 7) Support grafting of child qdiscs in mlxsw driver, from Nogah
    Frankel.

 8) Add packet forwarding tests to selftests, from Ido Schimmel.

 9) Deal with sub-optimal GSO packets better in BBR congestion control,
    from Eric Dumazet.

10) Support 5-tuple hashing in ipv6 multipath routing, from David Ahern.

11) Add path MTU tests to selftests, from Stefano Brivio.

12) Various bits of IPSEC offloading support for mlx5, from Aviad
    Yehezkel, Yossi Kuperman, and Saeed Mahameed.

13) Support RSS spreading on ntuple filters in SFC driver, from Edward
    Cree.

14) Lots of sockmap work from John Fastabend. Applications can use eBPF
    to filter sendmsg and sendpage operations.

15) In-kernel receive TLS support, from Dave Watson.

16) Add XDP support to ixgbevf, this is significant because it should
    allow optimized XDP usage in various cloud environments. From Tony
    Nguyen.

17) Add new Intel E800 series "ice" ethernet driver, from Anirudh
    Venkataramanan et al.

18) IP fragmentation match offload support in nfp driver, from Pieter
    Jansen van Vuuren.

19) Support XDP redirect in i40e driver, from Björn Töpel.

20) Add BPF_RAW_TRACEPOINT program type for accessing the arguments of
    tracepoints in their raw form, from Alexei Starovoitov.

21) Lots of striding RQ improvements to mlx5 driver with many
    performance improvements, from Tariq Toukan.

22) Use rhashtable for inet frag reassembly, from Eric Dumazet.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1678 commits)
  net: mvneta: improve suspend/resume
  net: mvneta: split rxq/txq init and txq deinit into SW and HW parts
  ipv6: frags: fix /proc/sys/net/ipv6/ip6frag_low_thresh
  net: bgmac: Fix endian access in bgmac_dma_tx_ring_free()
  net: bgmac: Correctly annotate register space
  route: check sysctl_fib_multipath_use_neigh earlier than hash
  fix typo in command value in drivers/net/phy/mdio-bitbang.
  sky2: Increase D3 delay to sky2 stops working after suspend
  net/mlx5e: Set EQE based as default TX interrupt moderation mode
  ibmvnic: Disable irqs before exiting reset from closed state
  net: sched: do not emit messages while holding spinlock
  vlan: also check phy_driver ts_info for vlan's real device
  Bluetooth: Mark expected switch fall-throughs
  Bluetooth: Set HCI_QUIRK_SIMULTANEOUS_DISCOVERY for BTUSB_QCA_ROME
  Bluetooth: btrsi: remove unused including <linux/version.h>
  Bluetooth: hci_bcm: Remove DMI quirk for the MINIX Z83-4
  sh_eth: kill useless check in __sh_eth_get_regs()
  sh_eth: add sh_eth_cpu_data::no_xdfar flag
  ipv6: factorize sk_wmem_alloc updates done by __ip6_append_data()
  ipv4: factorize sk_wmem_alloc updates done by __ip_append_data()
  ...

1 files changed, 301 insertions, 93 deletions

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 7f8bda3a2005..071fd9a7b919 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -57,7 +57,9 @@
 #include <linux/in.h>
 #include <linux/etherdevice.h>
 #include "mlx5_ib.h"
+#include "ib_rep.h"
 #include "cmd.h"
+#include <linux/mlx5/fs_helpers.h>
 
 #define DRIVER_NAME "mlx5_ib"
 #define DRIVER_VERSION "5.0-0"
@@ -130,7 +132,7 @@ static int get_port_state(struct ib_device *ibdev,
 	int ret;
 
 	memset(&attr, 0, sizeof(attr));
-	ret = mlx5_ib_query_port(ibdev, port_num, &attr);
+	ret = ibdev->query_port(ibdev, port_num, &attr);
 	if (!ret)
 		*state = attr.state;
 	return ret;
@@ -154,10 +156,19 @@ static int mlx5_netdev_event(struct notifier_block *this,
 	case NETDEV_REGISTER:
 	case NETDEV_UNREGISTER:
 		write_lock(&roce->netdev_lock);
-
-		if (ndev->dev.parent == &mdev->pdev->dev)
-			roce->netdev = (event == NETDEV_UNREGISTER) ?
+		if (ibdev->rep) {
+			struct mlx5_eswitch *esw = ibdev->mdev->priv.eswitch;
+			struct net_device *rep_ndev;
+
+			rep_ndev = mlx5_ib_get_rep_netdev(esw,
+							  ibdev->rep->vport);
+			if (rep_ndev == ndev)
+				roce->netdev = (event == NETDEV_UNREGISTER) ?
 					NULL : ndev;
+		} else if (ndev->dev.parent == &ibdev->mdev->pdev->dev) {
+			roce->netdev = (event == NETDEV_UNREGISTER) ?
+				NULL : ndev;
+		}
 		write_unlock(&roce->netdev_lock);
 		break;
 
@@ -1272,6 +1283,22 @@ int mlx5_ib_query_port(struct ib_device *ibdev, u8 port,
 	return ret;
 }
 
+static int mlx5_ib_rep_query_port(struct ib_device *ibdev, u8 port,
+				  struct ib_port_attr *props)
+{
+	int ret;
+
+	/* Only link layer == ethernet is valid for representors */
+	ret = mlx5_query_port_roce(ibdev, port, props);
+	if (ret || !props)
+		return ret;
+
+	/* We don't support GIDS */
+	props->gid_tbl_len = 0;
+
+	return ret;
+}
+
 static int mlx5_ib_query_gid(struct ib_device *ibdev, u8 port, int index,
 			     union ib_gid *gid)
 {
@@ -2290,11 +2317,9 @@ static void set_tos(void *outer_c, void *outer_v, u8 mask, u8 val)
 		   offsetof(typeof(filter), field) -\
 		   sizeof(filter.field))
 
-#define IPV4_VERSION 4
-#define IPV6_VERSION 6
 static int parse_flow_attr(struct mlx5_core_dev *mdev, u32 *match_c,
 			   u32 *match_v, const union ib_flow_spec *ib_spec,
-			   u32 *tag_id, bool *is_drop)
+			   struct mlx5_flow_act *action)
 {
 	void *misc_params_c = MLX5_ADDR_OF(fte_match_param, match_c,
 					   misc_parameters);
@@ -2377,7 +2402,7 @@ static int parse_flow_attr(struct mlx5_core_dev *mdev, u32 *match_c,
 			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
 				 ip_version, 0xf);
 			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
-				 ip_version, IPV4_VERSION);
+				 ip_version, MLX5_FS_IPV4_VERSION);
 		} else {
 			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
 				 ethertype, 0xffff);
@@ -2416,7 +2441,7 @@ static int parse_flow_attr(struct mlx5_core_dev *mdev, u32 *match_c,
 			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
 				 ip_version, 0xf);
 			MLX5_SET(fte_match_set_lyr_2_4, headers_v,
-				 ip_version, IPV6_VERSION);
+				 ip_version, MLX5_FS_IPV6_VERSION);
 		} else {
 			MLX5_SET(fte_match_set_lyr_2_4, headers_c,
 				 ethertype, 0xffff);
@@ -2512,13 +2537,14 @@ static int parse_flow_attr(struct mlx5_core_dev *mdev, u32 *match_c,
 		if (ib_spec->flow_tag.tag_id >= BIT(24))
 			return -EINVAL;
 
-		*tag_id = ib_spec->flow_tag.tag_id;
+		action->flow_tag = ib_spec->flow_tag.tag_id;
+		action->has_flow_tag = true;
 		break;
 	case IB_FLOW_SPEC_ACTION_DROP:
 		if (FIELDS_NOT_SUPPORTED(ib_spec->drop,
 					 LAST_DROP_FIELD))
 			return -EOPNOTSUPP;
-		*is_drop = true;
+		action->action |= MLX5_FLOW_CONTEXT_ACTION_DROP;
 		break;
 	default:
 		return -EINVAL;
@@ -2635,7 +2661,7 @@ static int mlx5_ib_destroy_flow(struct ib_flow *flow_id)
 							  ibflow);
 	struct mlx5_ib_flow_handler *iter, *tmp;
 
-	mutex_lock(&dev->flow_db.lock);
+	mutex_lock(&dev->flow_db->lock);
 
 	list_for_each_entry_safe(iter, tmp, &handler->list, list) {
 		mlx5_del_flow_rules(iter->rule);
@@ -2646,7 +2672,7 @@ static int mlx5_ib_destroy_flow(struct ib_flow *flow_id)
 
 	mlx5_del_flow_rules(handler->rule);
 	put_flow_table(dev, handler->prio, true);
-	mutex_unlock(&dev->flow_db.lock);
+	mutex_unlock(&dev->flow_db->lock);
 
 	kfree(handler);
 
@@ -2695,7 +2721,7 @@ static struct mlx5_ib_flow_prio *get_flow_table(struct mlx5_ib_dev *dev,
 					     MLX5_FLOW_NAMESPACE_BYPASS);
 		num_entries = MLX5_FS_MAX_ENTRIES;
 		num_groups = MLX5_FS_MAX_TYPES;
-		prio = &dev->flow_db.prios[priority];
+		prio = &dev->flow_db->prios[priority];
 	} else if (flow_attr->type == IB_FLOW_ATTR_ALL_DEFAULT ||
 		   flow_attr->type == IB_FLOW_ATTR_MC_DEFAULT) {
 		ns = mlx5_get_flow_namespace(dev->mdev,
@@ -2703,7 +2729,7 @@ static struct mlx5_ib_flow_prio *get_flow_table(struct mlx5_ib_dev *dev,
 		build_leftovers_ft_param(&priority,
 					 &num_entries,
 					 &num_groups);
-		prio = &dev->flow_db.prios[MLX5_IB_FLOW_LEFTOVERS_PRIO];
+		prio = &dev->flow_db->prios[MLX5_IB_FLOW_LEFTOVERS_PRIO];
 	} else if (flow_attr->type == IB_FLOW_ATTR_SNIFFER) {
 		if (!MLX5_CAP_FLOWTABLE(dev->mdev,
 					allow_sniffer_and_nic_rx_shared_tir))
@@ -2713,7 +2739,7 @@ static struct mlx5_ib_flow_prio *get_flow_table(struct mlx5_ib_dev *dev,
 					     MLX5_FLOW_NAMESPACE_SNIFFER_RX :
 					     MLX5_FLOW_NAMESPACE_SNIFFER_TX);
 
-		prio = &dev->flow_db.sniffer[ft_type];
+		prio = &dev->flow_db->sniffer[ft_type];
 		priority = 0;
 		num_entries = 1;
 		num_groups = 1;
@@ -2771,13 +2797,11 @@ static struct mlx5_ib_flow_handler *_create_flow_rule(struct mlx5_ib_dev *dev,
 {
 	struct mlx5_flow_table	*ft = ft_prio->flow_table;
 	struct mlx5_ib_flow_handler *handler;
-	struct mlx5_flow_act flow_act = {0};
+	struct mlx5_flow_act flow_act = {.flow_tag = MLX5_FS_DEFAULT_FLOW_TAG};
 	struct mlx5_flow_spec *spec;
 	struct mlx5_flow_destination *rule_dst = dst;
 	const void *ib_flow = (const void *)flow_attr + sizeof(*flow_attr);
 	unsigned int spec_index;
-	u32 flow_tag = MLX5_FS_DEFAULT_FLOW_TAG;
-	bool is_drop = false;
 	int err = 0;
 	int dest_num = 1;
 
@@ -2796,7 +2820,7 @@ static struct mlx5_ib_flow_handler *_create_flow_rule(struct mlx5_ib_dev *dev,
 	for (spec_index = 0; spec_index < flow_attr->num_of_specs; spec_index++) {
 		err = parse_flow_attr(dev->mdev, spec->match_criteria,
 				      spec->match_value,
-				      ib_flow, &flow_tag, &is_drop);
+				      ib_flow, &flow_act);
 		if (err < 0)
 			goto free;
 
@@ -2806,9 +2830,20 @@ static struct mlx5_ib_flow_handler *_create_flow_rule(struct mlx5_ib_dev *dev,
 	if (!flow_is_multicast_only(flow_attr))
 		set_underlay_qp(dev, spec, underlay_qpn);
 
+	if (dev->rep) {
+		void *misc;
+
+		misc = MLX5_ADDR_OF(fte_match_param, spec->match_value,
+				    misc_parameters);
+		MLX5_SET(fte_match_set_misc, misc, source_port,
+			 dev->rep->vport);
+		misc = MLX5_ADDR_OF(fte_match_param, spec->match_criteria,
+				    misc_parameters);
+		MLX5_SET_TO_ONES(fte_match_set_misc, misc, source_port);
+	}
+
 	spec->match_criteria_enable = get_match_criteria_enable(spec->match_criteria);
-	if (is_drop) {
-		flow_act.action = MLX5_FLOW_CONTEXT_ACTION_DROP;
+	if (flow_act.action & MLX5_FLOW_CONTEXT_ACTION_DROP) {
 		rule_dst = NULL;
 		dest_num = 0;
 	} else {
@@ -2816,15 +2851,14 @@ static struct mlx5_ib_flow_handler *_create_flow_rule(struct mlx5_ib_dev *dev,
 		    MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO;
 	}
 
-	if (flow_tag != MLX5_FS_DEFAULT_FLOW_TAG &&
+	if (flow_act.has_flow_tag &&
 	    (flow_attr->type == IB_FLOW_ATTR_ALL_DEFAULT ||
 	     flow_attr->type == IB_FLOW_ATTR_MC_DEFAULT)) {
 		mlx5_ib_warn(dev, "Flow tag %u and attribute type %x isn't allowed in leftovers\n",
-			     flow_tag, flow_attr->type);
+			     flow_act.flow_tag, flow_attr->type);
 		err = -EINVAL;
 		goto free;
 	}
-	flow_act.flow_tag = flow_tag;
 	handler->rule = mlx5_add_flow_rules(ft, spec,
 					    &flow_act,
 					    rule_dst, dest_num);
@@ -3003,7 +3037,7 @@ static struct ib_flow *mlx5_ib_create_flow(struct ib_qp *qp,
 	if (!dst)
 		return ERR_PTR(-ENOMEM);
 
-	mutex_lock(&dev->flow_db.lock);
+	mutex_lock(&dev->flow_db->lock);
 
 	ft_prio = get_flow_table(dev, flow_attr, MLX5_IB_FT_RX);
 	if (IS_ERR(ft_prio)) {
@@ -3052,7 +3086,7 @@ static struct ib_flow *mlx5_ib_create_flow(struct ib_qp *qp,
 		goto destroy_ft;
 	}
 
-	mutex_unlock(&dev->flow_db.lock);
+	mutex_unlock(&dev->flow_db->lock);
 	kfree(dst);
 
 	return &handler->ibflow;
@@ -3062,7 +3096,7 @@ destroy_ft:
 	if (ft_prio_tx)
 		put_flow_table(dev, ft_prio_tx, false);
 unlock:
-	mutex_unlock(&dev->flow_db.lock);
+	mutex_unlock(&dev->flow_db->lock);
 	kfree(dst);
 	kfree(handler);
 	return ERR_PTR(err);
@@ -3775,6 +3809,25 @@ static int mlx5_port_immutable(struct ib_device *ibdev, u8 port_num,
 	return 0;
 }
 
+static int mlx5_port_rep_immutable(struct ib_device *ibdev, u8 port_num,
+				   struct ib_port_immutable *immutable)
+{
+	struct ib_port_attr attr;
+	int err;
+
+	immutable->core_cap_flags = RDMA_CORE_PORT_RAW_PACKET;
+
+	err = ib_query_port(ibdev, port_num, &attr);
+	if (err)
+		return err;
+
+	immutable->pkey_tbl_len = attr.pkey_tbl_len;
+	immutable->gid_tbl_len = attr.gid_tbl_len;
+	immutable->core_cap_flags = RDMA_CORE_PORT_RAW_PACKET;
+
+	return 0;
+}
+
 static void get_dev_fw_str(struct ib_device *ibdev, char *str)
 {
 	struct mlx5_ib_dev *dev =
@@ -3805,7 +3858,7 @@ static int mlx5_eth_lag_init(struct mlx5_ib_dev *dev)
 		goto err_destroy_vport_lag;
 	}
 
-	dev->flow_db.lag_demux_ft = ft;
+	dev->flow_db->lag_demux_ft = ft;
 	return 0;
 
 err_destroy_vport_lag:
@@ -3817,9 +3870,9 @@ static void mlx5_eth_lag_cleanup(struct mlx5_ib_dev *dev)
 {
 	struct mlx5_core_dev *mdev = dev->mdev;
 
-	if (dev->flow_db.lag_demux_ft) {
-		mlx5_destroy_flow_table(dev->flow_db.lag_demux_ft);
-		dev->flow_db.lag_demux_ft = NULL;
+	if (dev->flow_db->lag_demux_ft) {
+		mlx5_destroy_flow_table(dev->flow_db->lag_demux_ft);
+		dev->flow_db->lag_demux_ft = NULL;
 
 		mlx5_cmd_destroy_vport_lag(mdev);
 	}
@@ -3851,14 +3904,10 @@ static int mlx5_enable_eth(struct mlx5_ib_dev *dev, u8 port_num)
 {
 	int err;
 
-	err = mlx5_add_netdev_notifier(dev, port_num);
-	if (err)
-		return err;
-
 	if (MLX5_CAP_GEN(dev->mdev, roce)) {
 		err = mlx5_nic_vport_enable_roce(dev->mdev);
 		if (err)
-			goto err_unregister_netdevice_notifier;
+			return err;
 	}
 
 	err = mlx5_eth_lag_init(dev);
@@ -3871,8 +3920,6 @@ err_disable_roce:
 	if (MLX5_CAP_GEN(dev->mdev, roce))
 		mlx5_nic_vport_disable_roce(dev->mdev);
 
-err_unregister_netdevice_notifier:
-	mlx5_remove_netdev_notifier(dev, port_num);
 	return err;
 }
 
@@ -4506,7 +4553,7 @@ static void mlx5_ib_cleanup_multiport_master(struct mlx5_ib_dev *dev)
 	mlx5_nic_vport_disable_roce(dev->mdev);
 }
 
-static void mlx5_ib_stage_init_cleanup(struct mlx5_ib_dev *dev)
+void mlx5_ib_stage_init_cleanup(struct mlx5_ib_dev *dev)
 {
 	mlx5_ib_cleanup_multiport_master(dev);
 #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
@@ -4515,7 +4562,7 @@ static void mlx5_ib_stage_init_cleanup(struct mlx5_ib_dev *dev)
 	kfree(dev->port);
 }
 
-static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
 {
 	struct mlx5_core_dev *mdev = dev->mdev;
 	const char *name;
@@ -4537,8 +4584,6 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
 		goto err_free_port;
 
 	if (!mlx5_core_mp_enabled(mdev)) {
-		int i;
-
 		for (i = 1; i <= dev->num_ports; i++) {
 			err = get_port_caps(dev, i);
 			if (err)
@@ -4567,7 +4612,6 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
 		dev->mdev->priv.eq_table.num_comp_vectors;
 	dev->ib_dev.dev.parent		= &mdev->pdev->dev;
 
-	mutex_init(&dev->flow_db.lock);
 	mutex_init(&dev->cap_mask_mutex);
 	INIT_LIST_HEAD(&dev->qp_list);
 	spin_lock_init(&dev->reset_flow_resource_lock);
@@ -4588,7 +4632,38 @@ err_free_port:
 	return -ENOMEM;
 }
 
-static int mlx5_ib_stage_caps_init(struct mlx5_ib_dev *dev)
+static int mlx5_ib_stage_flow_db_init(struct mlx5_ib_dev *dev)
+{
+	dev->flow_db = kzalloc(sizeof(*dev->flow_db), GFP_KERNEL);
+
+	if (!dev->flow_db)
+		return -ENOMEM;
+
+	mutex_init(&dev->flow_db->lock);
+
+	return 0;
+}
+
+int mlx5_ib_stage_rep_flow_db_init(struct mlx5_ib_dev *dev)
+{
+	struct mlx5_ib_dev *nic_dev;
+
+	nic_dev = mlx5_ib_get_uplink_ibdev(dev->mdev->priv.eswitch);
+
+	if (!nic_dev)
+		return -EINVAL;
+
+	dev->flow_db = nic_dev->flow_db;
+
+	return 0;
+}
+
+static void mlx5_ib_stage_flow_db_cleanup(struct mlx5_ib_dev *dev)
+{
+	kfree(dev->flow_db);
+}
+
+int mlx5_ib_stage_caps_init(struct mlx5_ib_dev *dev)
 {
 	struct mlx5_core_dev *mdev = dev->mdev;
 	int err;
@@ -4629,7 +4704,6 @@ static int mlx5_ib_stage_caps_init(struct mlx5_ib_dev *dev)
 		(1ull << IB_USER_VERBS_EX_CMD_MODIFY_CQ);
 
 	dev->ib_dev.query_device	= mlx5_ib_query_device;
-	dev->ib_dev.query_port		= mlx5_ib_query_port;
 	dev->ib_dev.get_link_layer	= mlx5_ib_port_link_layer;
 	dev->ib_dev.query_gid		= mlx5_ib_query_gid;
 	dev->ib_dev.add_gid		= mlx5_ib_add_gid;
@@ -4672,7 +4746,6 @@ static int mlx5_ib_stage_caps_init(struct mlx5_ib_dev *dev)
 	dev->ib_dev.alloc_mr		= mlx5_ib_alloc_mr;
 	dev->ib_dev.map_mr_sg		= mlx5_ib_map_mr_sg;
 	dev->ib_dev.check_mr_status	= mlx5_ib_check_mr_status;
-	dev->ib_dev.get_port_immutable  = mlx5_port_immutable;
 	dev->ib_dev.get_dev_fw_str      = get_dev_fw_str;
 	dev->ib_dev.get_vector_affinity	= mlx5_ib_get_vector_affinity;
 	if (MLX5_CAP_GEN(mdev, ipoib_enhanced_offloads))
@@ -4723,6 +4796,80 @@ static int mlx5_ib_stage_caps_init(struct mlx5_ib_dev *dev)
 	return 0;
 }
 
+static int mlx5_ib_stage_non_default_cb(struct mlx5_ib_dev *dev)
+{
+	dev->ib_dev.get_port_immutable  = mlx5_port_immutable;
+	dev->ib_dev.query_port		= mlx5_ib_query_port;
+
+	return 0;
+}
+
+int mlx5_ib_stage_rep_non_default_cb(struct mlx5_ib_dev *dev)
+{
+	dev->ib_dev.get_port_immutable  = mlx5_port_rep_immutable;
+	dev->ib_dev.query_port		= mlx5_ib_rep_query_port;
+
+	return 0;
+}
+
+static int mlx5_ib_stage_common_roce_init(struct mlx5_ib_dev *dev,
+					  u8 port_num)
+{
+	int i;
+
+	for (i = 0; i < dev->num_ports; i++) {
+		dev->roce[i].dev = dev;
+		dev->roce[i].native_port_num = i + 1;
+		dev->roce[i].last_port_state = IB_PORT_DOWN;
+	}
+
+	dev->ib_dev.get_netdev	= mlx5_ib_get_netdev;
+	dev->ib_dev.create_wq	 = mlx5_ib_create_wq;
+	dev->ib_dev.modify_wq	 = mlx5_ib_modify_wq;
+	dev->ib_dev.destroy_wq	 = mlx5_ib_destroy_wq;
+	dev->ib_dev.create_rwq_ind_table = mlx5_ib_create_rwq_ind_table;
+	dev->ib_dev.destroy_rwq_ind_table = mlx5_ib_destroy_rwq_ind_table;
+
+	dev->ib_dev.uverbs_ex_cmd_mask |=
+			(1ull << IB_USER_VERBS_EX_CMD_CREATE_WQ) |
+			(1ull << IB_USER_VERBS_EX_CMD_MODIFY_WQ) |
+			(1ull << IB_USER_VERBS_EX_CMD_DESTROY_WQ) |
+			(1ull << IB_USER_VERBS_EX_CMD_CREATE_RWQ_IND_TBL) |
+			(1ull << IB_USER_VERBS_EX_CMD_DESTROY_RWQ_IND_TBL);
+
+	return mlx5_add_netdev_notifier(dev, port_num);
+}
+
+static void mlx5_ib_stage_common_roce_cleanup(struct mlx5_ib_dev *dev)
+{
+	u8 port_num = mlx5_core_native_port_num(dev->mdev) - 1;
+
+	mlx5_remove_netdev_notifier(dev, port_num);
+}
+
+int mlx5_ib_stage_rep_roce_init(struct mlx5_ib_dev *dev)
+{
+	struct mlx5_core_dev *mdev = dev->mdev;
+	enum rdma_link_layer ll;
+	int port_type_cap;
+	int err = 0;
+	u8 port_num;
+
+	port_num = mlx5_core_native_port_num(dev->mdev) - 1;
+	port_type_cap = MLX5_CAP_GEN(mdev, port_type);
+	ll = mlx5_port_type_cap_to_rdma_ll(port_type_cap);
+
+	if (ll == IB_LINK_LAYER_ETHERNET)
+		err = mlx5_ib_stage_common_roce_init(dev, port_num);
+
+	return err;
+}
+
+void mlx5_ib_stage_rep_roce_cleanup(struct mlx5_ib_dev *dev)
+{
+	mlx5_ib_stage_common_roce_cleanup(dev);
+}
+
 static int mlx5_ib_stage_roce_init(struct mlx5_ib_dev *dev)
 {
 	struct mlx5_core_dev *mdev = dev->mdev;
@@ -4730,37 +4877,26 @@ static int mlx5_ib_stage_roce_init(struct mlx5_ib_dev *dev)
 	int port_type_cap;
 	u8 port_num;
 	int err;
-	int i;
 
 	port_num = mlx5_core_native_port_num(dev->mdev) - 1;
 	port_type_cap = MLX5_CAP_GEN(mdev, port_type);
 	ll = mlx5_port_type_cap_to_rdma_ll(port_type_cap);
 
 	if (ll == IB_LINK_LAYER_ETHERNET) {
-		for (i = 0; i < dev->num_ports; i++) {
-			dev->roce[i].dev = dev;
-			dev->roce[i].native_port_num = i + 1;
-			dev->roce[i].last_port_state = IB_PORT_DOWN;
-		}
+		err = mlx5_ib_stage_common_roce_init(dev, port_num);
+		if (err)
+			return err;
 
-		dev->ib_dev.get_netdev	= mlx5_ib_get_netdev;
-		dev->ib_dev.create_wq	 = mlx5_ib_create_wq;
-		dev->ib_dev.modify_wq	 = mlx5_ib_modify_wq;
-		dev->ib_dev.destroy_wq	 = mlx5_ib_destroy_wq;
-		dev->ib_dev.create_rwq_ind_table = mlx5_ib_create_rwq_ind_table;
-		dev->ib_dev.destroy_rwq_ind_table = mlx5_ib_destroy_rwq_ind_table;
-		dev->ib_dev.uverbs_ex_cmd_mask |=
-			(1ull << IB_USER_VERBS_EX_CMD_CREATE_WQ) |
-			(1ull << IB_USER_VERBS_EX_CMD_MODIFY_WQ) |
-			(1ull << IB_USER_VERBS_EX_CMD_DESTROY_WQ) |
-			(1ull << IB_USER_VERBS_EX_CMD_CREATE_RWQ_IND_TBL) |
-			(1ull << IB_USER_VERBS_EX_CMD_DESTROY_RWQ_IND_TBL);
 		err = mlx5_enable_eth(dev, port_num);
 		if (err)
-			return err;
+			goto cleanup;
 	}
 
 	return 0;
+cleanup:
+	mlx5_ib_stage_common_roce_cleanup(dev);
+
+	return err;
 }
 
 static void mlx5_ib_stage_roce_cleanup(struct mlx5_ib_dev *dev)
@@ -4776,16 +4912,16 @@ static void mlx5_ib_stage_roce_cleanup(struct mlx5_ib_dev *dev)
 
 	if (ll == IB_LINK_LAYER_ETHERNET) {
 		mlx5_disable_eth(dev);
-		mlx5_remove_netdev_notifier(dev, port_num);
+		mlx5_ib_stage_common_roce_cleanup(dev);
 	}
 }
 
-static int mlx5_ib_stage_dev_res_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_dev_res_init(struct mlx5_ib_dev *dev)
 {
 	return create_dev_resources(&dev->devr);
 }
 
-static void mlx5_ib_stage_dev_res_cleanup(struct mlx5_ib_dev *dev)
+void mlx5_ib_stage_dev_res_cleanup(struct mlx5_ib_dev *dev)
 {
 	destroy_dev_resources(&dev->devr);
 }
@@ -4797,7 +4933,7 @@ static int mlx5_ib_stage_odp_init(struct mlx5_ib_dev *dev)
 	return mlx5_ib_odp_init_one(dev);
 }
 
-static int mlx5_ib_stage_counters_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_counters_init(struct mlx5_ib_dev *dev)
 {
 	if (MLX5_CAP_GEN(dev->mdev, max_qp_cnt)) {
 		dev->ib_dev.get_hw_stats	= mlx5_ib_get_hw_stats;
@@ -4809,7 +4945,7 @@ static int mlx5_ib_stage_counters_init(struct mlx5_ib_dev *dev)
 	return 0;
 }
 
-static void mlx5_ib_stage_counters_cleanup(struct mlx5_ib_dev *dev)
+void mlx5_ib_stage_counters_cleanup(struct mlx5_ib_dev *dev)
 {
 	if (MLX5_CAP_GEN(dev->mdev, max_qp_cnt))
 		mlx5_ib_dealloc_counters(dev);
@@ -4840,7 +4976,7 @@ static void mlx5_ib_stage_uar_cleanup(struct mlx5_ib_dev *dev)
 	mlx5_put_uars_page(dev->mdev, dev->mdev->priv.uar);
 }
 
-static int mlx5_ib_stage_bfrag_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_bfrag_init(struct mlx5_ib_dev *dev)
 {
 	int err;
 
@@ -4855,28 +4991,28 @@ static int mlx5_ib_stage_bfrag_init(struct mlx5_ib_dev *dev)
 	return err;
 }
 
-static void mlx5_ib_stage_bfrag_cleanup(struct mlx5_ib_dev *dev)
+void mlx5_ib_stage_bfrag_cleanup(struct mlx5_ib_dev *dev)
 {
 	mlx5_free_bfreg(dev->mdev, &dev->fp_bfreg);
 	mlx5_free_bfreg(dev->mdev, &dev->bfreg);
 }
 
-static int mlx5_ib_stage_ib_reg_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_ib_reg_init(struct mlx5_ib_dev *dev)
 {
 	return ib_register_device(&dev->ib_dev, NULL);
 }
 
-static void mlx5_ib_stage_pre_ib_reg_umr_cleanup(struct mlx5_ib_dev *dev)
+void mlx5_ib_stage_pre_ib_reg_umr_cleanup(struct mlx5_ib_dev *dev)
 {
 	destroy_umrc_res(dev);
 }
 
-static void mlx5_ib_stage_ib_reg_cleanup(struct mlx5_ib_dev *dev)
+void mlx5_ib_stage_ib_reg_cleanup(struct mlx5_ib_dev *dev)
 {
 	ib_unregister_device(&dev->ib_dev);
 }
 
-static int mlx5_ib_stage_post_ib_reg_umr_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_post_ib_reg_umr_init(struct mlx5_ib_dev *dev)
 {
 	return create_umr_res(dev);
 }
@@ -4893,7 +5029,7 @@ static void mlx5_ib_stage_delay_drop_cleanup(struct mlx5_ib_dev *dev)
 	cancel_delay_drop(dev);
 }
 
-static int mlx5_ib_stage_class_attr_init(struct mlx5_ib_dev *dev)
+int mlx5_ib_stage_class_attr_init(struct mlx5_ib_dev *dev)
 {
 	int err;
 	int i;
@@ -4908,9 +5044,21 @@ static int mlx5_ib_stage_class_attr_init(struct mlx5_ib_dev *dev)
 	return 0;
 }
 
-static void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
-			     const struct mlx5_ib_profile *profile,
-			     int stage)
+static int mlx5_ib_stage_rep_reg_init(struct mlx5_ib_dev *dev)
+{
+	mlx5_ib_register_vport_reps(dev);
+
+	return 0;
+}
+
+static void mlx5_ib_stage_rep_reg_cleanup(struct mlx5_ib_dev *dev)
+{
+	mlx5_ib_unregister_vport_reps(dev);
+}
+
+void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
+		      const struct mlx5_ib_profile *profile,
+		      int stage)
 {
 	/* Number of stages to cleanup */
 	while (stage) {
@@ -4924,23 +5072,14 @@ static void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
 
 static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev, u8 port_num);
 
-static void *__mlx5_ib_add(struct mlx5_core_dev *mdev,
-			   const struct mlx5_ib_profile *profile)
+void *__mlx5_ib_add(struct mlx5_ib_dev *dev,
+		    const struct mlx5_ib_profile *profile)
 {
-	struct mlx5_ib_dev *dev;
 	int err;
 	int i;
 
 	printk_once(KERN_INFO "%s", mlx5_version);
 
-	dev = (struct mlx5_ib_dev *)ib_alloc_device(sizeof(*dev));
-	if (!dev)
-		return NULL;
-
-	dev->mdev = mdev;
-	dev->num_ports = max(MLX5_CAP_GEN(mdev, num_ports),
-			     MLX5_CAP_GEN(mdev, num_vhca_ports));
-
 	for (i = 0; i < MLX5_IB_STAGE_MAX; i++) {
 		if (profile->stage[i].init) {
 			err = profile->stage[i].init(dev);
@@ -4964,9 +5103,15 @@ static const struct mlx5_ib_profile pf_profile = {
 	STAGE_CREATE(MLX5_IB_STAGE_INIT,
 		     mlx5_ib_stage_init_init,
 		     mlx5_ib_stage_init_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_FLOW_DB,
+		     mlx5_ib_stage_flow_db_init,
+		     mlx5_ib_stage_flow_db_cleanup),
 	STAGE_CREATE(MLX5_IB_STAGE_CAPS,
 		     mlx5_ib_stage_caps_init,
 		     NULL),
+	STAGE_CREATE(MLX5_IB_STAGE_NON_DEFAULT_CB,
+		     mlx5_ib_stage_non_default_cb,
+		     NULL),
 	STAGE_CREATE(MLX5_IB_STAGE_ROCE,
 		     mlx5_ib_stage_roce_init,
 		     mlx5_ib_stage_roce_cleanup),
@@ -5005,6 +5150,51 @@ static const struct mlx5_ib_profile pf_profile = {
 		     NULL),
 };
 
+static const struct mlx5_ib_profile nic_rep_profile = {
+	STAGE_CREATE(MLX5_IB_STAGE_INIT,
+		     mlx5_ib_stage_init_init,
+		     mlx5_ib_stage_init_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_FLOW_DB,
+		     mlx5_ib_stage_flow_db_init,
+		     mlx5_ib_stage_flow_db_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_CAPS,
+		     mlx5_ib_stage_caps_init,
+		     NULL),
+	STAGE_CREATE(MLX5_IB_STAGE_NON_DEFAULT_CB,
+		     mlx5_ib_stage_rep_non_default_cb,
+		     NULL),
+	STAGE_CREATE(MLX5_IB_STAGE_ROCE,
+		     mlx5_ib_stage_rep_roce_init,
+		     mlx5_ib_stage_rep_roce_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_DEVICE_RESOURCES,
+		     mlx5_ib_stage_dev_res_init,
+		     mlx5_ib_stage_dev_res_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_COUNTERS,
+		     mlx5_ib_stage_counters_init,
+		     mlx5_ib_stage_counters_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_UAR,
+		     mlx5_ib_stage_uar_init,
+		     mlx5_ib_stage_uar_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_BFREG,
+		     mlx5_ib_stage_bfrag_init,
+		     mlx5_ib_stage_bfrag_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_PRE_IB_REG_UMR,
+		     NULL,
+		     mlx5_ib_stage_pre_ib_reg_umr_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_IB_REG,
+		     mlx5_ib_stage_ib_reg_init,
+		     mlx5_ib_stage_ib_reg_cleanup),
+	STAGE_CREATE(MLX5_IB_STAGE_POST_IB_REG_UMR,
+		     mlx5_ib_stage_post_ib_reg_umr_init,
+		     NULL),
+	STAGE_CREATE(MLX5_IB_STAGE_CLASS_ATTR,
+		     mlx5_ib_stage_class_attr_init,
+		     NULL),
+	STAGE_CREATE(MLX5_IB_STAGE_REP_REG,
+		     mlx5_ib_stage_rep_reg_init,
+		     mlx5_ib_stage_rep_reg_cleanup),
+};
+
 static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev, u8 port_num)
 {
 	struct mlx5_ib_multiport_info *mpi;
@@ -5050,8 +5240,11 @@ static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev, u8 port_num)
 static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 {
 	enum rdma_link_layer ll;
+	struct mlx5_ib_dev *dev;
 	int port_type_cap;
 
+	printk_once(KERN_INFO "%s", mlx5_version);
+
 	port_type_cap = MLX5_CAP_GEN(mdev, port_type);
 	ll = mlx5_port_type_cap_to_rdma_ll(port_type_cap);
 
@@ -5061,7 +5254,22 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 		return mlx5_ib_add_slave_port(mdev, port_num);
 	}
 
-	return __mlx5_ib_add(mdev, &pf_profile);
+	dev = (struct mlx5_ib_dev *)ib_alloc_device(sizeof(*dev));
+	if (!dev)
+		return NULL;
+
+	dev->mdev = mdev;
+	dev->num_ports = max(MLX5_CAP_GEN(mdev, num_ports),
+			     MLX5_CAP_GEN(mdev, num_vhca_ports));
+
+	if (MLX5_VPORT_MANAGER(mdev) &&
+	    mlx5_ib_eswitch_mode(mdev->priv.eswitch) == SRIOV_OFFLOADS) {
+		dev->rep = mlx5_ib_vport_rep(mdev->priv.eswitch, 0);
+
+		return __mlx5_ib_add(dev, &nic_rep_profile);
+	}
+
+	return __mlx5_ib_add(dev, &pf_profile);
 }
 
 static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)