Merge branch 'mlx5-next' of https://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux

Leon Romanovsky says:

====================
This PR is collected from
https://lore.kernel.org/all/cover.1695296682.git.leon@kernel.org

This series from Patrisious extends mlx5 to support IPsec packet offload
in multiport devices (MPV, see [1] for more details).

These devices have single flow steering logic and two netdev interfaces,
which require extra logic to manage IPsec configurations as they performed
on netdevs.

[1] https://lore.kernel.org/linux-rdma/20180104152544.28919-1-leon@kernel.org/

* 'mlx5-next' of https://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux:
  net/mlx5: Handle IPsec steering upon master unbind/bind
  net/mlx5: Configure IPsec steering for ingress RoCEv2 MPV traffic
  net/mlx5: Configure IPsec steering for egress RoCEv2 MPV traffic
  net/mlx5: Add create alias flow table function to ipsec roce
  net/mlx5: Implement alias object allow and create functions
  net/mlx5: Add alias flow table bits
  net/mlx5: Store devcom pointer inside IPsec RoCE
  net/mlx5: Register mlx5e priv to devcom in MPV mode
  RDMA/mlx5: Send events from IB driver about device affiliation state
  net/mlx5: Introduce ifc bits for migration in a chunk mode

====================

Link: https://lore.kernel.org/r/20231002083832.19746-1-leon@kernel.orgSigned-off-by: Jakub Kicinski <kuba@kernel.org>

drivers/infiniband/hw/mlx5/main.c

@@ -24,6 +24,7 @@
 #include <linux/mlx5/vport.h>
 #include <linux/mlx5/fs.h>
 #include <linux/mlx5/eswitch.h>
+#include <linux/mlx5/driver.h>
 #include <linux/list.h>
 #include <rdma/ib_smi.h>
 #include <rdma/ib_umem_odp.h>
@@ -3175,6 +3176,13 @@ static void mlx5_ib_unbind_slave_port(struct mlx5_ib_dev *ibdev,
 
 	lockdep_assert_held(&mlx5_ib_multiport_mutex);
 
+	mlx5_core_mp_event_replay(ibdev->mdev,
+				  MLX5_DRIVER_EVENT_AFFILIATION_REMOVED,
+				  NULL);
+	mlx5_core_mp_event_replay(mpi->mdev,
+				  MLX5_DRIVER_EVENT_AFFILIATION_REMOVED,
+				  NULL);
+
 	mlx5_ib_cleanup_cong_debugfs(ibdev, port_num);
 
 	spin_lock(&port->mp.mpi_lock);
@@ -3226,6 +3234,7 @@ static bool mlx5_ib_bind_slave_port(struct mlx5_ib_dev *ibdev,
 				    struct mlx5_ib_multiport_info *mpi)
 {
 	u32 port_num = mlx5_core_native_port_num(mpi->mdev) - 1;
+	u64 key;
 	int err;
 
 	lockdep_assert_held(&mlx5_ib_multiport_mutex);
@@ -3254,6 +3263,14 @@ static bool mlx5_ib_bind_slave_port(struct mlx5_ib_dev *ibdev,
 
 	mlx5_ib_init_cong_debugfs(ibdev, port_num);
 
+	key = ibdev->ib_dev.index;
+	mlx5_core_mp_event_replay(mpi->mdev,
+				  MLX5_DRIVER_EVENT_AFFILIATION_DONE,
+				  &key);
+	mlx5_core_mp_event_replay(ibdev->mdev,
+				  MLX5_DRIVER_EVENT_AFFILIATION_DONE,
+				  &key);
+
 	return true;
 
 unbind:

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

@@ -525,6 +525,7 @@ static int mlx5_internal_err_ret_value(struct mlx5_core_dev *dev, u16 op,
 	case MLX5_CMD_OP_SAVE_VHCA_STATE:
 	case MLX5_CMD_OP_LOAD_VHCA_STATE:
 	case MLX5_CMD_OP_SYNC_CRYPTO:
+	case MLX5_CMD_OP_ALLOW_OTHER_VHCA_ACCESS:
 		*status = MLX5_DRIVER_STATUS_ABORTED;
 		*synd = MLX5_DRIVER_SYND;
 		return -ENOLINK;
@@ -728,6 +729,7 @@ const char *mlx5_command_str(int command)
 	MLX5_COMMAND_STR_CASE(SAVE_VHCA_STATE);
 	MLX5_COMMAND_STR_CASE(LOAD_VHCA_STATE);
 	MLX5_COMMAND_STR_CASE(SYNC_CRYPTO);
+	MLX5_COMMAND_STR_CASE(ALLOW_OTHER_VHCA_ACCESS);
 	default: return "unknown command opcode";
 	}
 }
@@ -2090,6 +2092,74 @@ int mlx5_cmd_exec_cb(struct mlx5_async_ctx *ctx, void *in, int in_size,
 }
 EXPORT_SYMBOL(mlx5_cmd_exec_cb);
 
+int mlx5_cmd_allow_other_vhca_access(struct mlx5_core_dev *dev,
+				     struct mlx5_cmd_allow_other_vhca_access_attr *attr)
+{
+	u32 out[MLX5_ST_SZ_DW(allow_other_vhca_access_out)] = {};
+	u32 in[MLX5_ST_SZ_DW(allow_other_vhca_access_in)] = {};
+	void *key;
+
+	MLX5_SET(allow_other_vhca_access_in,
+		 in, opcode, MLX5_CMD_OP_ALLOW_OTHER_VHCA_ACCESS);
+	MLX5_SET(allow_other_vhca_access_in,
+		 in, object_type_to_be_accessed, attr->obj_type);
+	MLX5_SET(allow_other_vhca_access_in,
+		 in, object_id_to_be_accessed, attr->obj_id);
+
+	key = MLX5_ADDR_OF(allow_other_vhca_access_in, in, access_key);
+	memcpy(key, attr->access_key, sizeof(attr->access_key));
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
+int mlx5_cmd_alias_obj_create(struct mlx5_core_dev *dev,
+			      struct mlx5_cmd_alias_obj_create_attr *alias_attr,
+			      u32 *obj_id)
+{
+	u32 out[MLX5_ST_SZ_DW(general_obj_out_cmd_hdr)] = {};
+	u32 in[MLX5_ST_SZ_DW(create_alias_obj_in)] = {};
+	void *param;
+	void *attr;
+	void *key;
+	int ret;
+
+	attr = MLX5_ADDR_OF(create_alias_obj_in, in, hdr);
+	MLX5_SET(general_obj_in_cmd_hdr,
+		 attr, opcode, MLX5_CMD_OP_CREATE_GENERAL_OBJECT);
+	MLX5_SET(general_obj_in_cmd_hdr,
+		 attr, obj_type, alias_attr->obj_type);
+	param = MLX5_ADDR_OF(general_obj_in_cmd_hdr, in, op_param);
+	MLX5_SET(general_obj_create_param, param, alias_object, 1);
+
+	attr = MLX5_ADDR_OF(create_alias_obj_in, in, alias_ctx);
+	MLX5_SET(alias_context, attr, vhca_id_to_be_accessed, alias_attr->vhca_id);
+	MLX5_SET(alias_context, attr, object_id_to_be_accessed, alias_attr->obj_id);
+
+	key = MLX5_ADDR_OF(alias_context, attr, access_key);
+	memcpy(key, alias_attr->access_key, sizeof(alias_attr->access_key));
+
+	ret = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+	if (ret)
+		return ret;
+
+	*obj_id = MLX5_GET(general_obj_out_cmd_hdr, out, obj_id);
+
+	return 0;
+}
+
+int mlx5_cmd_alias_obj_destroy(struct mlx5_core_dev *dev, u32 obj_id,
+			       u16 obj_type)
+{
+	u32 out[MLX5_ST_SZ_DW(general_obj_out_cmd_hdr)] = {};
+	u32 in[MLX5_ST_SZ_DW(general_obj_in_cmd_hdr)] = {};
+
+	MLX5_SET(general_obj_in_cmd_hdr, in, opcode, MLX5_CMD_OP_DESTROY_GENERAL_OBJECT);
+	MLX5_SET(general_obj_in_cmd_hdr, in, obj_type, obj_type);
+	MLX5_SET(general_obj_in_cmd_hdr, in, obj_id, obj_id);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}
+
 static void destroy_msg_cache(struct mlx5_core_dev *dev)
 {
 	struct cmd_msg_cache *ch;

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

@@ -168,6 +168,13 @@ struct page_pool;
 #define mlx5e_state_dereference(priv, p) \
 	rcu_dereference_protected((p), lockdep_is_held(&(priv)->state_lock))
 
+enum mlx5e_devcom_events {
+	MPV_DEVCOM_MASTER_UP,
+	MPV_DEVCOM_MASTER_DOWN,
+	MPV_DEVCOM_IPSEC_MASTER_UP,
+	MPV_DEVCOM_IPSEC_MASTER_DOWN,
+};
+
 static inline u8 mlx5e_get_num_lag_ports(struct mlx5_core_dev *mdev)
 {
 	if (mlx5_lag_is_lacp_owner(mdev))
@@ -936,6 +943,7 @@ struct mlx5e_priv {
 	struct mlx5e_htb          *htb;
 	struct mlx5e_mqprio_rl    *mqprio_rl;
 	struct dentry             *dfs_root;
+	struct mlx5_devcom_comp_dev *devcom;
 };
 
 struct mlx5e_dev {

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c

@@ -850,6 +850,7 @@ void mlx5e_ipsec_init(struct mlx5e_priv *priv)
 
 	xa_init_flags(&ipsec->sadb, XA_FLAGS_ALLOC);
 	ipsec->mdev = priv->mdev;
+	init_completion(&ipsec->comp);
 	ipsec->wq = alloc_workqueue("mlx5e_ipsec: %s", WQ_UNBOUND, 0,
 				    priv->netdev->name);
 	if (!ipsec->wq)
@@ -870,7 +871,7 @@ void mlx5e_ipsec_init(struct mlx5e_priv *priv)
 	}
 
 	ipsec->is_uplink_rep = mlx5e_is_uplink_rep(priv);
-	ret = mlx5e_accel_ipsec_fs_init(ipsec);
+	ret = mlx5e_accel_ipsec_fs_init(ipsec, &priv->devcom);
 	if (ret)
 		goto err_fs_init;
 

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h

@@ -38,6 +38,7 @@
 #include <net/xfrm.h>
 #include <linux/idr.h>
 #include "lib/aso.h"
+#include "lib/devcom.h"
 
 #define MLX5E_IPSEC_SADB_RX_BITS 10
 #define MLX5E_IPSEC_ESN_SCOPE_MID 0x80000000L
@@ -221,12 +222,20 @@ struct mlx5e_ipsec_tx_create_attr {
 	enum mlx5_flow_namespace_type chains_ns;
 };
 
+struct mlx5e_ipsec_mpv_work {
+	int event;
+	struct work_struct work;
+	struct mlx5e_priv *slave_priv;
+	struct mlx5e_priv *master_priv;
+};
+
 struct mlx5e_ipsec {
 	struct mlx5_core_dev *mdev;
 	struct xarray sadb;
 	struct mlx5e_ipsec_sw_stats sw_stats;
 	struct mlx5e_ipsec_hw_stats hw_stats;
 	struct workqueue_struct *wq;
+	struct completion comp;
 	struct mlx5e_flow_steering *fs;
 	struct mlx5e_ipsec_rx *rx_ipv4;
 	struct mlx5e_ipsec_rx *rx_ipv6;
@@ -238,6 +247,7 @@ struct mlx5e_ipsec {
 	struct notifier_block netevent_nb;
 	struct mlx5_ipsec_fs *roce;
 	u8 is_uplink_rep: 1;
+	struct mlx5e_ipsec_mpv_work mpv_work;
 };
 
 struct mlx5e_ipsec_esn_state {
@@ -302,7 +312,7 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv);
 void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv);
 
 void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec);
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_ipsec *ipsec);
+int mlx5e_accel_ipsec_fs_init(struct mlx5e_ipsec *ipsec, struct mlx5_devcom_comp_dev **devcom);
 int mlx5e_accel_ipsec_fs_add_rule(struct mlx5e_ipsec_sa_entry *sa_entry);
 void mlx5e_accel_ipsec_fs_del_rule(struct mlx5e_ipsec_sa_entry *sa_entry);
 int mlx5e_accel_ipsec_fs_add_pol(struct mlx5e_ipsec_pol_entry *pol_entry);
@@ -328,6 +338,10 @@ void mlx5e_accel_ipsec_fs_read_stats(struct mlx5e_priv *priv,
 
 void mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry,
 					struct mlx5_accel_esp_xfrm_attrs *attrs);
+void mlx5e_ipsec_handle_mpv_event(int event, struct mlx5e_priv *slave_priv,
+				  struct mlx5e_priv *master_priv);
+void mlx5e_ipsec_send_event(struct mlx5e_priv *priv, int event);
+
 static inline struct mlx5_core_dev *
 mlx5e_ipsec_sa2dev(struct mlx5e_ipsec_sa_entry *sa_entry)
 {
@@ -363,6 +377,15 @@ static inline u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev)
 {
 	return 0;
 }
+
+static inline void mlx5e_ipsec_handle_mpv_event(int event, struct mlx5e_priv *slave_priv,
+						struct mlx5e_priv *master_priv)
+{
+}
+
+static inline void mlx5e_ipsec_send_event(struct mlx5e_priv *priv, int event)
+{
+}
 #endif
 
 #endif	/* __MLX5E_IPSEC_H__ */

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c

@@ -229,6 +229,83 @@ static int ipsec_miss_create(struct mlx5_core_dev *mdev,
 	return err;
 }
 
+static void handle_ipsec_rx_bringup(struct mlx5e_ipsec *ipsec, u32 family)
+{
+	struct mlx5e_ipsec_rx *rx = ipsec_rx(ipsec, family, XFRM_DEV_OFFLOAD_PACKET);
+	struct mlx5_flow_namespace *ns = mlx5e_fs_get_ns(ipsec->fs, false);
+	struct mlx5_flow_destination old_dest, new_dest;
+
+	old_dest = mlx5_ttc_get_default_dest(mlx5e_fs_get_ttc(ipsec->fs, false),
+					     family2tt(family));
+
+	mlx5_ipsec_fs_roce_rx_create(ipsec->mdev, ipsec->roce, ns, &old_dest, family,
+				     MLX5E_ACCEL_FS_ESP_FT_ROCE_LEVEL, MLX5E_NIC_PRIO);
+
+	new_dest.ft = mlx5_ipsec_fs_roce_ft_get(ipsec->roce, family);
+	new_dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	mlx5_modify_rule_destination(rx->status.rule, &new_dest, &old_dest);
+	mlx5_modify_rule_destination(rx->sa.rule, &new_dest, &old_dest);
+}
+
+static void handle_ipsec_rx_cleanup(struct mlx5e_ipsec *ipsec, u32 family)
+{
+	struct mlx5e_ipsec_rx *rx = ipsec_rx(ipsec, family, XFRM_DEV_OFFLOAD_PACKET);
+	struct mlx5_flow_destination old_dest, new_dest;
+
+	old_dest.ft = mlx5_ipsec_fs_roce_ft_get(ipsec->roce, family);
+	old_dest.type = MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE;
+	new_dest = mlx5_ttc_get_default_dest(mlx5e_fs_get_ttc(ipsec->fs, false),
+					     family2tt(family));
+	mlx5_modify_rule_destination(rx->sa.rule, &new_dest, &old_dest);
+	mlx5_modify_rule_destination(rx->status.rule, &new_dest, &old_dest);
+
+	mlx5_ipsec_fs_roce_rx_destroy(ipsec->roce, family, ipsec->mdev);
+}
+
+static void ipsec_mpv_work_handler(struct work_struct *_work)
+{
+	struct mlx5e_ipsec_mpv_work *work = container_of(_work, struct mlx5e_ipsec_mpv_work, work);
+	struct mlx5e_ipsec *ipsec = work->slave_priv->ipsec;
+
+	switch (work->event) {
+	case MPV_DEVCOM_IPSEC_MASTER_UP:
+		mutex_lock(&ipsec->tx->ft.mutex);
+		if (ipsec->tx->ft.refcnt)
+			mlx5_ipsec_fs_roce_tx_create(ipsec->mdev, ipsec->roce, ipsec->tx->ft.pol,
+						     true);
+		mutex_unlock(&ipsec->tx->ft.mutex);
+
+		mutex_lock(&ipsec->rx_ipv4->ft.mutex);
+		if (ipsec->rx_ipv4->ft.refcnt)
+			handle_ipsec_rx_bringup(ipsec, AF_INET);
+		mutex_unlock(&ipsec->rx_ipv4->ft.mutex);
+
+		mutex_lock(&ipsec->rx_ipv6->ft.mutex);
+		if (ipsec->rx_ipv6->ft.refcnt)
+			handle_ipsec_rx_bringup(ipsec, AF_INET6);
+		mutex_unlock(&ipsec->rx_ipv6->ft.mutex);
+		break;
+	case MPV_DEVCOM_IPSEC_MASTER_DOWN:
+		mutex_lock(&ipsec->tx->ft.mutex);
+		if (ipsec->tx->ft.refcnt)
+			mlx5_ipsec_fs_roce_tx_destroy(ipsec->roce, ipsec->mdev);
+		mutex_unlock(&ipsec->tx->ft.mutex);
+
+		mutex_lock(&ipsec->rx_ipv4->ft.mutex);
+		if (ipsec->rx_ipv4->ft.refcnt)
+			handle_ipsec_rx_cleanup(ipsec, AF_INET);
+		mutex_unlock(&ipsec->rx_ipv4->ft.mutex);
+
+		mutex_lock(&ipsec->rx_ipv6->ft.mutex);
+		if (ipsec->rx_ipv6->ft.refcnt)
+			handle_ipsec_rx_cleanup(ipsec, AF_INET6);
+		mutex_unlock(&ipsec->rx_ipv6->ft.mutex);
+		break;
+	}
+
+	complete(&work->master_priv->ipsec->comp);
+}
+
 static void ipsec_rx_ft_disconnect(struct mlx5e_ipsec *ipsec, u32 family)
 {
 	struct mlx5_ttc_table *ttc = mlx5e_fs_get_ttc(ipsec->fs, false);
@@ -264,7 +341,7 @@ static void rx_destroy(struct mlx5_core_dev *mdev, struct mlx5e_ipsec *ipsec,
 	}
 	mlx5_destroy_flow_table(rx->ft.status);
 
-	mlx5_ipsec_fs_roce_rx_destroy(ipsec->roce, family);
+	mlx5_ipsec_fs_roce_rx_destroy(ipsec->roce, family, mdev);
 }
 
 static void ipsec_rx_create_attr_set(struct mlx5e_ipsec *ipsec,
@@ -422,7 +499,7 @@ static int rx_create(struct mlx5_core_dev *mdev, struct mlx5e_ipsec *ipsec,
 err_add:
 	mlx5_destroy_flow_table(rx->ft.status);
 err_fs_ft_status:
-	mlx5_ipsec_fs_roce_rx_destroy(ipsec->roce, family);
+	mlx5_ipsec_fs_roce_rx_destroy(ipsec->roce, family, mdev);
 	return err;
 }
 
@@ -562,7 +639,7 @@ static int ipsec_counter_rule_tx(struct mlx5_core_dev *mdev, struct mlx5e_ipsec_
 static void tx_destroy(struct mlx5e_ipsec *ipsec, struct mlx5e_ipsec_tx *tx,
 		       struct mlx5_ipsec_fs *roce)
 {
-	mlx5_ipsec_fs_roce_tx_destroy(roce);
+	mlx5_ipsec_fs_roce_tx_destroy(roce, ipsec->mdev);
 	if (tx->chains) {
 		ipsec_chains_destroy(tx->chains);
 	} else {
@@ -665,7 +742,7 @@ static int tx_create(struct mlx5e_ipsec *ipsec, struct mlx5e_ipsec_tx *tx,
 	}
 
 connect_roce:
-	err = mlx5_ipsec_fs_roce_tx_create(mdev, roce, tx->ft.pol);
+	err = mlx5_ipsec_fs_roce_tx_create(mdev, roce, tx->ft.pol, false);
 	if (err)
 		goto err_roce;
 	return 0;
@@ -1888,7 +1965,8 @@ void mlx5e_accel_ipsec_fs_cleanup(struct mlx5e_ipsec *ipsec)
 	}
 }
 
-int mlx5e_accel_ipsec_fs_init(struct mlx5e_ipsec *ipsec)
+int mlx5e_accel_ipsec_fs_init(struct mlx5e_ipsec *ipsec,
+			      struct mlx5_devcom_comp_dev **devcom)
 {
 	struct mlx5_core_dev *mdev = ipsec->mdev;
 	struct mlx5_flow_namespace *ns, *ns_esw;
@@ -1940,7 +2018,9 @@ int mlx5e_accel_ipsec_fs_init(struct mlx5e_ipsec *ipsec)
 		ipsec->tx_esw->ns = ns_esw;
 		xa_init_flags(&ipsec->rx_esw->ipsec_obj_id_map, XA_FLAGS_ALLOC1);
 	} else if (mlx5_ipsec_device_caps(mdev) & MLX5_IPSEC_CAP_ROCE) {
-		ipsec->roce = mlx5_ipsec_fs_roce_init(mdev);
+		ipsec->roce = mlx5_ipsec_fs_roce_init(mdev, devcom);
+	} else {
+		mlx5_core_warn(mdev, "IPsec was initialized without RoCE support\n");
 	}
 
 	return 0;
@@ -1987,3 +2067,33 @@ bool mlx5e_ipsec_fs_tunnel_enabled(struct mlx5e_ipsec_sa_entry *sa_entry)
 
 	return rx->allow_tunnel_mode;
 }
+
+void mlx5e_ipsec_handle_mpv_event(int event, struct mlx5e_priv *slave_priv,
+				  struct mlx5e_priv *master_priv)
+{
+	struct mlx5e_ipsec_mpv_work *work;
+
+	reinit_completion(&master_priv->ipsec->comp);
+
+	if (!slave_priv->ipsec) {
+		complete(&master_priv->ipsec->comp);
+		return;
+	}
+
+	work = &slave_priv->ipsec->mpv_work;
+
+	INIT_WORK(&work->work, ipsec_mpv_work_handler);
+	work->event = event;
+	work->slave_priv = slave_priv;
+	work->master_priv = master_priv;
+	queue_work(slave_priv->ipsec->wq, &work->work);
+}
+
+void mlx5e_ipsec_send_event(struct mlx5e_priv *priv, int event)
+{
+	if (!priv->ipsec)
+		return; /* IPsec not supported */
+
+	mlx5_devcom_send_event(priv->devcom, event, event, priv);
+	wait_for_completion(&priv->ipsec->comp);
+}

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_offload.c

@@ -5,6 +5,7 @@
 #include "en.h"
 #include "ipsec.h"
 #include "lib/crypto.h"
+#include "lib/ipsec_fs_roce.h"
 
 enum {
 	MLX5_IPSEC_ASO_REMOVE_FLOW_PKT_CNT_OFFSET,
@@ -63,7 +64,7 @@ u32 mlx5_ipsec_device_caps(struct mlx5_core_dev *mdev)
 			caps |= MLX5_IPSEC_CAP_ESPINUDP;
 	}
 
-	if (mlx5_get_roce_state(mdev) &&
+	if (mlx5_get_roce_state(mdev) && mlx5_ipsec_fs_is_mpv_roce_supported(mdev) &&
 	    MLX5_CAP_GEN_2(mdev, flow_table_type_2_type) & MLX5_FT_NIC_RX_2_NIC_RX_RDMA &&
 	    MLX5_CAP_GEN_2(mdev, flow_table_type_2_type) & MLX5_FT_NIC_TX_RDMA_2_NIC_TX)
 		caps |= MLX5_IPSEC_CAP_ROCE;

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

@@ -69,6 +69,7 @@
 #include "en/htb.h"
 #include "qos.h"
 #include "en/trap.h"
+#include "lib/devcom.h"
 
 bool mlx5e_check_fragmented_striding_rq_cap(struct mlx5_core_dev *mdev, u8 page_shift,
 					    enum mlx5e_mpwrq_umr_mode umr_mode)
@@ -178,6 +179,61 @@ static void mlx5e_disable_async_events(struct mlx5e_priv *priv)
 	mlx5_notifier_unregister(priv->mdev, &priv->events_nb);
 }
 
+static int mlx5e_devcom_event_mpv(int event, void *my_data, void *event_data)
+{
+	struct mlx5e_priv *slave_priv = my_data;
+
+	switch (event) {
+	case MPV_DEVCOM_MASTER_UP:
+		mlx5_devcom_comp_set_ready(slave_priv->devcom, true);
+		break;
+	case MPV_DEVCOM_MASTER_DOWN:
+		/* no need for comp set ready false since we unregister after
+		 * and it hurts cleanup flow.
+		 */
+		break;
+	case MPV_DEVCOM_IPSEC_MASTER_UP:
+	case MPV_DEVCOM_IPSEC_MASTER_DOWN:
+		mlx5e_ipsec_handle_mpv_event(event, my_data, event_data);
+		break;
+	}
+
+	return 0;
+}
+
+static int mlx5e_devcom_init_mpv(struct mlx5e_priv *priv, u64 *data)
+{
+	priv->devcom = mlx5_devcom_register_component(priv->mdev->priv.devc,
+						      MLX5_DEVCOM_MPV,
+						      *data,
+						      mlx5e_devcom_event_mpv,
+						      priv);
+	if (IS_ERR_OR_NULL(priv->devcom))
+		return -EOPNOTSUPP;
+
+	if (mlx5_core_is_mp_master(priv->mdev)) {
+		mlx5_devcom_send_event(priv->devcom, MPV_DEVCOM_MASTER_UP,
+				       MPV_DEVCOM_MASTER_UP, priv);
+		mlx5e_ipsec_send_event(priv, MPV_DEVCOM_IPSEC_MASTER_UP);
+	}
+
+	return 0;
+}
+
+static void mlx5e_devcom_cleanup_mpv(struct mlx5e_priv *priv)
+{
+	if (IS_ERR_OR_NULL(priv->devcom))
+		return;
+
+	if (mlx5_core_is_mp_master(priv->mdev)) {
+		mlx5_devcom_send_event(priv->devcom, MPV_DEVCOM_MASTER_DOWN,
+				       MPV_DEVCOM_MASTER_DOWN, priv);
+		mlx5e_ipsec_send_event(priv, MPV_DEVCOM_IPSEC_MASTER_DOWN);
+	}
+
+	mlx5_devcom_unregister_component(priv->devcom);
+}
+
 static int blocking_event(struct notifier_block *nb, unsigned long event, void *data)
 {
 	struct mlx5e_priv *priv = container_of(nb, struct mlx5e_priv, blocking_events_nb);
@@ -192,6 +248,13 @@ static int blocking_event(struct notifier_block *nb, unsigned long event, void *
 			return NOTIFY_BAD;
 		}
 		break;
+	case MLX5_DRIVER_EVENT_AFFILIATION_DONE:
+		if (mlx5e_devcom_init_mpv(priv, data))
+			return NOTIFY_BAD;
+		break;
+	case MLX5_DRIVER_EVENT_AFFILIATION_REMOVED:
+		mlx5e_devcom_cleanup_mpv(priv);
+		break;
 	default:
 		return NOTIFY_DONE;
 	}

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

@@ -114,9 +114,9 @@
 #define ETHTOOL_NUM_PRIOS 11
 #define ETHTOOL_MIN_LEVEL (KERNEL_MIN_LEVEL + ETHTOOL_NUM_PRIOS)
 /* Promiscuous, Vlan, mac, ttc, inner ttc, {UDP/ANY/aRFS/accel/{esp, esp_err}}, IPsec policy,
- * IPsec RoCE policy
+ * {IPsec RoCE MPV,Alias table},IPsec RoCE policy
  */
-#define KERNEL_NIC_PRIO_NUM_LEVELS 9
+#define KERNEL_NIC_PRIO_NUM_LEVELS 11
 #define KERNEL_NIC_NUM_PRIOS 1
 /* One more level for tc */
 #define KERNEL_MIN_LEVEL (KERNEL_NIC_PRIO_NUM_LEVELS + 1)
@@ -137,7 +137,7 @@
 #define LAG_MIN_LEVEL (OFFLOADS_MIN_LEVEL + KERNEL_RX_MACSEC_MIN_LEVEL + 1)
 
 #define KERNEL_TX_IPSEC_NUM_PRIOS  1
-#define KERNEL_TX_IPSEC_NUM_LEVELS 3
+#define KERNEL_TX_IPSEC_NUM_LEVELS 4
 #define KERNEL_TX_IPSEC_MIN_LEVEL        (KERNEL_TX_IPSEC_NUM_LEVELS)
 
 #define KERNEL_TX_MACSEC_NUM_PRIOS  1
@@ -231,7 +231,7 @@ enum {
 };
 
 #define RDMA_RX_IPSEC_NUM_PRIOS 1
-#define RDMA_RX_IPSEC_NUM_LEVELS 2
+#define RDMA_RX_IPSEC_NUM_LEVELS 4
 #define RDMA_RX_IPSEC_MIN_LEVEL  (RDMA_RX_IPSEC_NUM_LEVELS)
 
 #define RDMA_RX_BYPASS_MIN_LEVEL MLX5_BY_PASS_NUM_REGULAR_PRIOS
@@ -288,7 +288,7 @@ enum {
 #define RDMA_TX_BYPASS_MIN_LEVEL MLX5_BY_PASS_NUM_PRIOS
 #define RDMA_TX_COUNTERS_MIN_LEVEL (RDMA_TX_BYPASS_MIN_LEVEL + 1)
 
-#define RDMA_TX_IPSEC_NUM_PRIOS 1
+#define RDMA_TX_IPSEC_NUM_PRIOS 2
 #define RDMA_TX_IPSEC_PRIO_NUM_LEVELS 1
 #define RDMA_TX_IPSEC_MIN_LEVEL  (RDMA_TX_COUNTERS_MIN_LEVEL + RDMA_TX_IPSEC_NUM_PRIOS)
 

drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.h

@@ -8,6 +8,7 @@
 
 enum mlx5_devcom_component {
 	MLX5_DEVCOM_ESW_OFFLOADS,
+	MLX5_DEVCOM_MPV,
 	MLX5_DEVCOM_NUM_COMPONENTS,
 };
 

drivers/net/ethernet/mellanox/mlx5/core/lib/ipsec_fs_roce.h

@@ -4,22 +4,28 @@
 #ifndef __MLX5_LIB_IPSEC_H__
 #define __MLX5_LIB_IPSEC_H__
 
+#include "lib/devcom.h"
+
 struct mlx5_ipsec_fs;
 
 struct mlx5_flow_table *
 mlx5_ipsec_fs_roce_ft_get(struct mlx5_ipsec_fs *ipsec_roce, u32 family);
 void mlx5_ipsec_fs_roce_rx_destroy(struct mlx5_ipsec_fs *ipsec_roce,
-				   u32 family);
+				   u32 family, struct mlx5_core_dev *mdev);
 int mlx5_ipsec_fs_roce_rx_create(struct mlx5_core_dev *mdev,
 				 struct mlx5_ipsec_fs *ipsec_roce,
 				 struct mlx5_flow_namespace *ns,
 				 struct mlx5_flow_destination *default_dst,
 				 u32 family, u32 level, u32 prio);
-void mlx5_ipsec_fs_roce_tx_destroy(struct mlx5_ipsec_fs *ipsec_roce);
+void mlx5_ipsec_fs_roce_tx_destroy(struct mlx5_ipsec_fs *ipsec_roce,
+				   struct mlx5_core_dev *mdev);
 int mlx5_ipsec_fs_roce_tx_create(struct mlx5_core_dev *mdev,
 				 struct mlx5_ipsec_fs *ipsec_roce,
-				 struct mlx5_flow_table *pol_ft);
+				 struct mlx5_flow_table *pol_ft,
+				 bool from_event);
 void mlx5_ipsec_fs_roce_cleanup(struct mlx5_ipsec_fs *ipsec_roce);
-struct mlx5_ipsec_fs *mlx5_ipsec_fs_roce_init(struct mlx5_core_dev *mdev);
+struct mlx5_ipsec_fs *mlx5_ipsec_fs_roce_init(struct mlx5_core_dev *mdev,
+					      struct mlx5_devcom_comp_dev **devcom);
+bool mlx5_ipsec_fs_is_mpv_roce_supported(struct mlx5_core_dev *mdev);
 
 #endif /* __MLX5_LIB_IPSEC_H__ */

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

@@ -361,6 +361,12 @@ void mlx5_core_uplink_netdev_event_replay(struct mlx5_core_dev *dev)
 }
 EXPORT_SYMBOL(mlx5_core_uplink_netdev_event_replay);
 
+void mlx5_core_mp_event_replay(struct mlx5_core_dev *dev, u32 event, void *data)
+{
+	mlx5_blocking_notifier_call_chain(dev, event, data);
+}
+EXPORT_SYMBOL(mlx5_core_mp_event_replay);
+
 int mlx5_core_get_caps_mode(struct mlx5_core_dev *dev, enum mlx5_cap_type cap_type,
 			    enum mlx5_cap_mode cap_mode)
 {

drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h

@@ -97,6 +97,22 @@ do {								\
 			     __func__, __LINE__, current->pid,	\
 			     ##__VA_ARGS__)
 
+#define ACCESS_KEY_LEN  32
+#define FT_ID_FT_TYPE_OFFSET 24
+
+struct mlx5_cmd_allow_other_vhca_access_attr {
+	u16 obj_type;
+	u32 obj_id;
+	u8 access_key[ACCESS_KEY_LEN];
+};
+
+struct mlx5_cmd_alias_obj_create_attr {
+	u32 obj_id;
+	u16 vhca_id;
+	u16 obj_type;
+	u8 access_key[ACCESS_KEY_LEN];
+};
+
 static inline void mlx5_printk(struct mlx5_core_dev *dev, int level, const char *format, ...)
 {
 	struct device *device = dev->device;
@@ -343,6 +359,12 @@ bool mlx5_eth_supported(struct mlx5_core_dev *dev);
 bool mlx5_rdma_supported(struct mlx5_core_dev *dev);
 bool mlx5_vnet_supported(struct mlx5_core_dev *dev);
 bool mlx5_same_hw_devs(struct mlx5_core_dev *dev, struct mlx5_core_dev *peer_dev);
+int mlx5_cmd_allow_other_vhca_access(struct mlx5_core_dev *dev,
+				     struct mlx5_cmd_allow_other_vhca_access_attr *attr);
+int mlx5_cmd_alias_obj_create(struct mlx5_core_dev *dev,
+			      struct mlx5_cmd_alias_obj_create_attr *alias_attr,
+			      u32 *obj_id);
+int mlx5_cmd_alias_obj_destroy(struct mlx5_core_dev *dev, u32 obj_id, u16 obj_type);
 
 static inline u16 mlx5_core_ec_vf_vport_base(const struct mlx5_core_dev *dev)
 {

include/linux/mlx5/device.h

@@ -367,6 +367,8 @@ enum mlx5_driver_event {
 	MLX5_DRIVER_EVENT_MACSEC_SA_ADDED,
 	MLX5_DRIVER_EVENT_MACSEC_SA_DELETED,
 	MLX5_DRIVER_EVENT_SF_PEER_DEVLINK,
+	MLX5_DRIVER_EVENT_AFFILIATION_DONE,
+	MLX5_DRIVER_EVENT_AFFILIATION_REMOVED,
 };
 
 enum {

include/linux/mlx5/driver.h

@@ -1029,6 +1029,8 @@ bool mlx5_cmd_is_down(struct mlx5_core_dev *dev);
 void mlx5_core_uplink_netdev_set(struct mlx5_core_dev *mdev, struct net_device *netdev);
 void mlx5_core_uplink_netdev_event_replay(struct mlx5_core_dev *mdev);
 
+void mlx5_core_mp_event_replay(struct mlx5_core_dev *dev, u32 event, void *data);
+
 void mlx5_health_cleanup(struct mlx5_core_dev *dev);
 int mlx5_health_init(struct mlx5_core_dev *dev);
 void mlx5_start_health_poll(struct mlx5_core_dev *dev);

include/linux/mlx5/mlx5_ifc.h

@@ -312,6 +312,7 @@ enum {
 	MLX5_CMD_OP_QUERY_VHCA_STATE              = 0xb0d,
 	MLX5_CMD_OP_MODIFY_VHCA_STATE             = 0xb0e,
 	MLX5_CMD_OP_SYNC_CRYPTO                   = 0xb12,
+	MLX5_CMD_OP_ALLOW_OTHER_VHCA_ACCESS       = 0xb16,
 	MLX5_CMD_OP_MAX
 };
 
@@ -1934,6 +1935,14 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         match_definer_format_supported[0x40];
 };
 
+enum {
+	MLX5_CROSS_VHCA_OBJ_TO_OBJ_SUPPORTED_LOCAL_FLOW_TABLE_TO_REMOTE_FLOW_TABLE_MISS  = 0x80000,
+};
+
+enum {
+	MLX5_ALLOWED_OBJ_FOR_OTHER_VHCA_ACCESS_FLOW_TABLE       = 0x200,
+};
+
 struct mlx5_ifc_cmd_hca_cap_2_bits {
 	u8	   reserved_at_0[0x80];
 
@@ -1948,9 +1957,15 @@ struct mlx5_ifc_cmd_hca_cap_2_bits {
 	u8	   reserved_at_c0[0x8];
 	u8	   migration_multi_load[0x1];
 	u8	   migration_tracking_state[0x1];
-	u8	   reserved_at_ca[0x16];
+	u8	   reserved_at_ca[0x6];
+	u8	   migration_in_chunks[0x1];
+	u8	   reserved_at_d1[0xf];
+
+	u8	   cross_vhca_object_to_object_supported[0x20];
+
+	u8	   allowed_object_for_other_vhca_access[0x40];
 
-	u8	   reserved_at_e0[0xc0];
+	u8	   reserved_at_140[0x60];
 
 	u8	   flow_table_type_2_type[0x8];
 	u8	   reserved_at_1a8[0x3];
@@ -6369,6 +6384,28 @@ struct mlx5_ifc_general_obj_out_cmd_hdr_bits {
 	u8         reserved_at_60[0x20];
 };
 
+struct mlx5_ifc_allow_other_vhca_access_in_bits {
+	u8 opcode[0x10];
+	u8 uid[0x10];
+	u8 reserved_at_20[0x10];
+	u8 op_mod[0x10];
+	u8 reserved_at_40[0x50];
+	u8 object_type_to_be_accessed[0x10];
+	u8 object_id_to_be_accessed[0x20];
+	u8 reserved_at_c0[0x40];
+	union {
+		u8 access_key_raw[0x100];
+		u8 access_key[8][0x20];
+	};
+};
+
+struct mlx5_ifc_allow_other_vhca_access_out_bits {
+	u8 status[0x8];
+	u8 reserved_at_8[0x18];
+	u8 syndrome[0x20];
+	u8 reserved_at_40[0x40];
+};
+
 struct mlx5_ifc_modify_header_arg_bits {
 	u8         reserved_at_0[0x80];
 
@@ -6391,6 +6428,24 @@ struct mlx5_ifc_create_match_definer_out_bits {
 	struct mlx5_ifc_general_obj_out_cmd_hdr_bits general_obj_out_cmd_hdr;
 };
 
+struct mlx5_ifc_alias_context_bits {
+	u8 vhca_id_to_be_accessed[0x10];
+	u8 reserved_at_10[0xd];
+	u8 status[0x3];
+	u8 object_id_to_be_accessed[0x20];
+	u8 reserved_at_40[0x40];
+	union {
+		u8 access_key_raw[0x100];
+		u8 access_key[8][0x20];
+	};
+	u8 metadata[0x80];
+};
+
+struct mlx5_ifc_create_alias_obj_in_bits {
+	struct mlx5_ifc_general_obj_in_cmd_hdr_bits hdr;
+	struct mlx5_ifc_alias_context_bits alias_ctx;
+};
+
 enum {
 	MLX5_QUERY_FLOW_GROUP_OUT_MATCH_CRITERIA_ENABLE_OUTER_HEADERS    = 0x0,
 	MLX5_QUERY_FLOW_GROUP_OUT_MATCH_CRITERIA_ENABLE_MISC_PARAMETERS  = 0x1,
@@ -11920,6 +11975,7 @@ enum {
 	MLX5_GENERAL_OBJECT_TYPES_FLOW_METER_ASO = 0x24,
 	MLX5_GENERAL_OBJECT_TYPES_MACSEC = 0x27,
 	MLX5_GENERAL_OBJECT_TYPES_INT_KEK = 0x47,
+	MLX5_GENERAL_OBJECT_TYPES_FLOW_TABLE_ALIAS = 0xff15,
 };
 
 enum {
@@ -12395,7 +12451,8 @@ struct mlx5_ifc_query_vhca_migration_state_in_bits {
 	u8         op_mod[0x10];
 
 	u8         incremental[0x1];
-	u8         reserved_at_41[0xf];
+	u8         chunk[0x1];
+	u8         reserved_at_42[0xe];
 	u8         vhca_id[0x10];
 
 	u8         reserved_at_60[0x20];
@@ -12411,7 +12468,11 @@ struct mlx5_ifc_query_vhca_migration_state_out_bits {
 
 	u8         required_umem_size[0x20];
 
-	u8         reserved_at_a0[0x160];
+	u8         reserved_at_a0[0x20];
+
+	u8         remaining_total_size[0x40];
+
+	u8         reserved_at_100[0x100];
 };
 
 struct mlx5_ifc_save_vhca_state_in_bits {
@@ -12443,7 +12504,7 @@ struct mlx5_ifc_save_vhca_state_out_bits {
 
 	u8         actual_image_size[0x20];
 
-	u8         reserved_at_60[0x20];
+	u8         next_required_umem_size[0x20];
 };
 
 struct mlx5_ifc_load_vhca_state_in_bits {