Merge tag 'mlx5-updates-2023-12-13' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2023-12-13

Preparation for mlx5e socket direct feature.

Socket direct will allow multiple PF devices attached to different
NUMA nodes but sharing the same physical port.

The following series is a small refactoring series in preparation
to support socket direct in the following submission.

Highlights:
- Define required device registers and bits related to socket direct
- Flow steering re-arrangements
- Generalize TX objects (TISs) and store them in a common object, will
be useful in the next series for per function object management.
- Decouple raw CQ objects from their parent netdev priv
- Prepare devcom for Socket Direct device group discovery.

Please see the individual patches for more information.
====================

Signed-off-by: David S. Miller <[e-mail]>

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

···
 #define MLX5E_HW2SW_MTU(params, hwmtu) ((hwmtu) - ((params)->hard_mtu))
 #define MLX5E_SW2HW_MTU(params, swmtu) ((swmtu) + ((params)->hard_mtu))
 
-#define MLX5E_MAX_NUM_TC	8
 #define MLX5E_MAX_NUM_MQPRIO_CH_TC TC_QOPT_MAX_QUEUE
 
 #define MLX5_RX_HEADROOM NET_SKB_PAD
···
 	/* control */
 	struct net_device         *netdev;
 	struct mlx5_core_dev      *mdev;
-	struct mlx5e_priv         *priv;
+	struct workqueue_struct   *workqueue;
 	struct mlx5_wq_ctrl        wq_ctrl;
 } ____cacheline_aligned_in_smp;
 
···
 	/* data path */
 	struct mlx5e_rq            rq;
 	struct mlx5e_xdpsq         rq_xdpsq;
-	struct mlx5e_txqsq         sq[MLX5E_MAX_NUM_TC];
+	struct mlx5e_txqsq         sq[MLX5_MAX_NUM_TC];
 	struct mlx5e_icosq         icosq;   /* internal control operations */
 	struct mlx5e_txqsq __rcu * __rcu *qos_sqs;
 	bool                       xdp;
···
 
 struct mlx5e_channel_stats {
 	struct mlx5e_ch_stats ch;
-	struct mlx5e_sq_stats sq[MLX5E_MAX_NUM_TC];
+	struct mlx5e_sq_stats sq[MLX5_MAX_NUM_TC];
 	struct mlx5e_rq_stats rq;
 	struct mlx5e_rq_stats xskrq;
 	struct mlx5e_xdpsq_stats rq_xdpsq;
···
 
 struct mlx5e_ptp_stats {
 	struct mlx5e_ch_stats ch;
-	struct mlx5e_sq_stats sq[MLX5E_MAX_NUM_TC];
-	struct mlx5e_ptp_cq_stats cq[MLX5E_MAX_NUM_TC];
+	struct mlx5e_sq_stats sq[MLX5_MAX_NUM_TC];
+	struct mlx5e_ptp_cq_stats cq[MLX5_MAX_NUM_TC];
 	struct mlx5e_rq_stats rq;
 } ____cacheline_aligned_in_smp;
 
···
 	struct mlx5e_rq            drop_rq;
 
 	struct mlx5e_channels      channels;
-	u32                        tisn[MLX5_MAX_PORTS][MLX5E_MAX_NUM_TC];
 	struct mlx5e_rx_res       *rx_res;
 	u32                       *tx_rates;
 
···
 	void	(*update_stats)(struct mlx5e_priv *priv);
 	void	(*update_carrier)(struct mlx5e_priv *priv);
 	int	(*max_nch_limit)(struct mlx5_core_dev *mdev);
+	u32	(*get_tisn)(struct mlx5_core_dev *mdev, struct mlx5e_priv *priv,
+			    u8 lag_port, u8 tc);
 	unsigned int (*stats_grps_num)(struct mlx5e_priv *priv);
 	mlx5e_stats_grp_t *stats_grps;
 	const struct mlx5e_rx_handlers *rx_handlers;
 	int	max_tc;
 	u32     features;
 };
+
+u32 mlx5e_profile_get_tisn(struct mlx5_core_dev *mdev,
+			   struct mlx5e_priv *priv,
+			   const struct mlx5e_profile *profile,
+			   u8 lag_port, u8 tc);
 
 #define mlx5e_profile_feature_cap(profile, feature)	\
 	((profile)->features & BIT(MLX5E_PROFILE_FEATURE_##feature))
···
 void mlx5e_close_xdpsq(struct mlx5e_xdpsq *sq);
 
 struct mlx5e_create_cq_param {
+	struct net_device *netdev;
+	struct workqueue_struct *wq;
 	struct napi_struct *napi;
 	struct mlx5e_ch_stats *ch_stats;
 	int node;
···
 };
 
 struct mlx5e_cq_param;
-int mlx5e_open_cq(struct mlx5e_priv *priv, struct dim_cq_moder moder,
+int mlx5e_open_cq(struct mlx5_core_dev *mdev, struct dim_cq_moder moder,
 		  struct mlx5e_cq_param *param, struct mlx5e_create_cq_param *ccp,
 		  struct mlx5e_cq *cq);
 void mlx5e_close_cq(struct mlx5e_cq *cq);
···
 int mlx5e_create_tis(struct mlx5_core_dev *mdev, void *in, u32 *tisn);
 void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn);
 
-int mlx5e_create_tises(struct mlx5e_priv *priv);
-void mlx5e_destroy_tises(struct mlx5e_priv *priv);
 int mlx5e_update_nic_rx(struct mlx5e_priv *priv);
 void mlx5e_update_carrier(struct mlx5e_priv *priv);
 int mlx5e_close(struct net_device *netdev);

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

···
 	return true;
 }
 
-void mlx5e_monitor_counter_arm(struct mlx5e_priv *priv)
+static void mlx5e_monitor_counter_arm(struct mlx5e_priv *priv)
 {
 	u32 in[MLX5_ST_SZ_DW(arm_monitor_counter_in)] = {};
 

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

···
 int  mlx5e_monitor_counter_supported(struct mlx5e_priv *priv);
 void mlx5e_monitor_counter_init(struct mlx5e_priv *priv);
 void mlx5e_monitor_counter_cleanup(struct mlx5e_priv *priv);
-void mlx5e_monitor_counter_arm(struct mlx5e_priv *priv);
 
 #endif /* __MLX5_MONITOR_H__ */

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

···
 void mlx5e_build_create_cq_param(struct mlx5e_create_cq_param *ccp, struct mlx5e_channel *c)
 {
 	*ccp = (struct mlx5e_create_cq_param) {
+		.netdev = c->netdev,
+		.wq = c->priv->wq,
 		.napi = &c->napi,
 		.ch_stats = c->stats,
 		.node = cpu_to_node(c->cpu),

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

···
 
 	for (tc = 0; tc < num_tc; tc++) {
 		int txq_ix = ix_base + tc;
+		u32 tisn;
 
-		err = mlx5e_ptp_open_txqsq(c, c->priv->tisn[c->lag_port][tc], txq_ix,
-					   cparams, tc, &c->ptpsq[tc]);
+		tisn = mlx5e_profile_get_tisn(c->mdev, c->priv, c->priv->profile,
+					      c->lag_port, tc);
+		err = mlx5e_ptp_open_txqsq(c, tisn, txq_ix, cparams, tc, &c->ptpsq[tc]);
 		if (err)
 			goto close_txqsq;
 	}
···
 
 	num_tc = mlx5e_get_dcb_num_tc(params);
 
+	ccp.netdev   = c->netdev;
+	ccp.wq       = c->priv->wq;
 	ccp.node     = dev_to_node(mlx5_core_dma_dev(c->mdev));
 	ccp.ch_stats = c->stats;
 	ccp.napi     = &c->napi;
···
 	for (tc = 0; tc < num_tc; tc++) {
 		struct mlx5e_cq *cq = &c->ptpsq[tc].txqsq.cq;
 
-		err = mlx5e_open_cq(c->priv, ptp_moder, cq_param, &ccp, cq);
+		err = mlx5e_open_cq(c->mdev, ptp_moder, cq_param, &ccp, cq);
 		if (err)
 			goto out_err_txqsq_cq;
 	}
···
 		struct mlx5e_cq *cq = &c->ptpsq[tc].ts_cq;
 		struct mlx5e_ptpsq *ptpsq = &c->ptpsq[tc];
 
-		err = mlx5e_open_cq(c->priv, ptp_moder, cq_param, &ccp, cq);
+		err = mlx5e_open_cq(c->mdev, ptp_moder, cq_param, &ccp, cq);
 		if (err)
 			goto out_err_ts_cq;
 
···
 	struct mlx5e_cq_param *cq_param;
 	struct mlx5e_cq *cq = &c->rq.cq;
 
+	ccp.netdev   = c->netdev;
+	ccp.wq       = c->priv->wq;
 	ccp.node     = dev_to_node(mlx5_core_dma_dev(c->mdev));
 	ccp.ch_stats = c->stats;
 	ccp.napi     = &c->napi;
···
 
 	cq_param = &cparams->rq_param.cqp;
 
-	return mlx5e_open_cq(c->priv, ptp_moder, cq_param, &ccp, cq);
+	return mlx5e_open_cq(c->mdev, ptp_moder, cq_param, &ccp, cq);
 }
 
 static void mlx5e_ptp_close_tx_cqs(struct mlx5e_ptp *c)

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

···
 
 struct mlx5e_ptp {
 	/* data path */
-	struct mlx5e_ptpsq         ptpsq[MLX5E_MAX_NUM_TC];
+	struct mlx5e_ptpsq         ptpsq[MLX5_MAX_NUM_TC];
 	struct mlx5e_rq            rq;
 	struct napi_struct         napi;
 	struct device             *pdev;

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

···
 	struct mlx5e_params *params;
 	struct mlx5e_channel *c;
 	struct mlx5e_txqsq *sq;
+	u32 tisn;
 
 	params = &chs->params;
 
···
 	memset(&param_cq, 0, sizeof(param_cq));
 	mlx5e_build_sq_param(priv->mdev, params, &param_sq);
 	mlx5e_build_tx_cq_param(priv->mdev, params, &param_cq);
-	err = mlx5e_open_cq(priv, params->tx_cq_moderation, &param_cq, &ccp, &sq->cq);
+	err = mlx5e_open_cq(c->mdev, params->tx_cq_moderation, &param_cq, &ccp, &sq->cq);
 	if (err)
 		goto err_free_sq;
-	err = mlx5e_open_txqsq(c, priv->tisn[c->lag_port][0], txq_ix, params,
-			       &param_sq, sq, 0, hw_id,
+
+	tisn = mlx5e_profile_get_tisn(c->mdev, c->priv, c->priv->profile,
+				      c->lag_port, 0);
+	err = mlx5e_open_txqsq(c, tisn, txq_ix, params, &param_sq, sq, 0, hw_id,
 			       priv->htb_qos_sq_stats[node_qid]);
 	if (err)
 		goto err_close_cq;

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

···
 
 	node = dev_to_node(mdev->device);
 
+	ccp.netdev   = priv->netdev;
+	ccp.wq       = priv->wq;
 	ccp.node     = node;
 	ccp.ch_stats = t->stats;
 	ccp.napi     = &t->napi;
 	ccp.ix       = 0;
-	err = mlx5e_open_cq(priv, trap_moder, &rq_param->cqp, &ccp, &rq->cq);
+	err = mlx5e_open_cq(priv->mdev, trap_moder, &rq_param->cqp, &ccp, &rq->cq);
 	if (err)
 		return err;
 

drivers/net/ethernet/mellanox/mlx5/core/en/xsk/setup.c

···
 
 	mlx5e_build_xsk_cparam(priv->mdev, params, xsk, priv->q_counter, cparam);
 
-	err = mlx5e_open_cq(c->priv, params->rx_cq_moderation, &cparam->rq.cqp, &ccp,
+	err = mlx5e_open_cq(c->mdev, params->rx_cq_moderation, &cparam->rq.cqp, &ccp,
 			    &c->xskrq.cq);
 	if (unlikely(err))
 		goto err_free_cparam;
···
 	if (unlikely(err))
 		goto err_close_rx_cq;
 
-	err = mlx5e_open_cq(c->priv, params->tx_cq_moderation, &cparam->xdp_sq.cqp, &ccp,
+	err = mlx5e_open_cq(c->mdev, params->tx_cq_moderation, &cparam->xdp_sq.cqp, &ccp,
 			    &c->xsksq.cq);
 	if (unlikely(err))
 		goto err_close_rq;

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

···
 	return err;
 }
 
+int mlx5e_create_tis(struct mlx5_core_dev *mdev, void *in, u32 *tisn)
+{
+	void *tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
+
+	MLX5_SET(tisc, tisc, transport_domain, mdev->mlx5e_res.hw_objs.td.tdn);
+
+	if (mlx5_lag_is_lacp_owner(mdev))
+		MLX5_SET(tisc, tisc, strict_lag_tx_port_affinity, 1);
+
+	return mlx5_core_create_tis(mdev, in, tisn);
+}
+
+void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn)
+{
+	mlx5_core_destroy_tis(mdev, tisn);
+}
+
+static void mlx5e_destroy_tises(struct mlx5_core_dev *mdev, u32 tisn[MLX5_MAX_PORTS][MLX5_MAX_NUM_TC])
+{
+	int tc, i;
+
+	for (i = 0; i < MLX5_MAX_PORTS; i++)
+		for (tc = 0; tc < MLX5_MAX_NUM_TC; tc++)
+			mlx5e_destroy_tis(mdev, tisn[i][tc]);
+}
+
+static bool mlx5_lag_should_assign_affinity(struct mlx5_core_dev *mdev)
+{
+	return MLX5_CAP_GEN(mdev, lag_tx_port_affinity) && mlx5e_get_num_lag_ports(mdev) > 1;
+}
+
+static int mlx5e_create_tises(struct mlx5_core_dev *mdev, u32 tisn[MLX5_MAX_PORTS][MLX5_MAX_NUM_TC])
+{
+	int tc, i;
+	int err;
+
+	for (i = 0; i < MLX5_MAX_PORTS; i++) {
+		for (tc = 0; tc < MLX5_MAX_NUM_TC; tc++) {
+			u32 in[MLX5_ST_SZ_DW(create_tis_in)] = {};
+			void *tisc;
+
+			tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
+
+			MLX5_SET(tisc, tisc, prio, tc << 1);
+
+			if (mlx5_lag_should_assign_affinity(mdev))
+				MLX5_SET(tisc, tisc, lag_tx_port_affinity, i + 1);
+
+			err = mlx5e_create_tis(mdev, in, &tisn[i][tc]);
+			if (err)
+				goto err_close_tises;
+		}
+	}
+
+	return 0;
+
+err_close_tises:
+	for (; i >= 0; i--) {
+		for (tc--; tc >= 0; tc--)
+			mlx5e_destroy_tis(mdev, tisn[i][tc]);
+		tc = MLX5_MAX_NUM_TC;
+	}
+
+	return err;
+}
+
 int mlx5e_create_mdev_resources(struct mlx5_core_dev *mdev)
 {
 	struct mlx5e_hw_objs *res = &mdev->mlx5e_res.hw_objs;
···
 		goto err_destroy_mkey;
 	}
 
+	err = mlx5e_create_tises(mdev, res->tisn);
+	if (err) {
+		mlx5_core_err(mdev, "alloc tises failed, %d\n", err);
+		goto err_destroy_bfreg;
+	}
 	INIT_LIST_HEAD(&res->td.tirs_list);
 	mutex_init(&res->td.list_lock);
 
···
 
 	return 0;
 
+err_destroy_bfreg:
+	mlx5_free_bfreg(mdev, &res->bfreg);
 err_destroy_mkey:
 	mlx5_core_destroy_mkey(mdev, res->mkey);
 err_dealloc_transport_domain:
···
 
 	mlx5_crypto_dek_cleanup(mdev->mlx5e_res.dek_priv);
 	mdev->mlx5e_res.dek_priv = NULL;
+	mlx5e_destroy_tises(mdev, res->tisn);
 	mlx5_free_bfreg(mdev, &res->bfreg);
 	mlx5_core_destroy_mkey(mdev, res->mkey);
 	mlx5_core_dealloc_transport_domain(mdev, res->td.tdn);

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

···
 	mlx5e_free_rq(rq);
 }
 
+u32 mlx5e_profile_get_tisn(struct mlx5_core_dev *mdev,
+			   struct mlx5e_priv *priv,
+			   const struct mlx5e_profile *profile,
+			   u8 lag_port, u8 tc)
+{
+	if (profile->get_tisn)
+		return profile->get_tisn(mdev, priv, lag_port, tc);
+
+	return mdev->mlx5e_res.hw_objs.tisn[lag_port][tc];
+}
+
 static void mlx5e_free_xdpsq_db(struct mlx5e_xdpsq *sq)
 {
 	kvfree(sq->db.xdpi_fifo.xi);
···
 		return err;
 
 	csp.tis_lst_sz      = 1;
-	csp.tisn            = c->priv->tisn[c->lag_port][0]; /* tc = 0 */
+	csp.tisn            = mlx5e_profile_get_tisn(c->mdev, c->priv, c->priv->profile,
+						     c->lag_port, 0); /* tc = 0 */
 	csp.cqn             = sq->cq.mcq.cqn;
 	csp.wq_ctrl         = &sq->wq_ctrl;
 	csp.min_inline_mode = sq->min_inline_mode;
···
 	mlx5e_free_xdpsq(sq);
 }
 
-static int mlx5e_alloc_cq_common(struct mlx5e_priv *priv,
+static int mlx5e_alloc_cq_common(struct mlx5_core_dev *mdev,
+				 struct net_device *netdev,
+				 struct workqueue_struct *workqueue,
 				 struct mlx5e_cq_param *param,
 				 struct mlx5e_cq *cq)
 {
-	struct mlx5_core_dev *mdev = priv->mdev;
 	struct mlx5_core_cq *mcq = &cq->mcq;
 	int err;
 	u32 i;
···
 	}
 
 	cq->mdev = mdev;
-	cq->netdev = priv->netdev;
-	cq->priv = priv;
+	cq->netdev = netdev;
+	cq->workqueue = workqueue;
 
 	return 0;
 }
 
-static int mlx5e_alloc_cq(struct mlx5e_priv *priv,
+static int mlx5e_alloc_cq(struct mlx5_core_dev *mdev,
 			  struct mlx5e_cq_param *param,
 			  struct mlx5e_create_cq_param *ccp,
 			  struct mlx5e_cq *cq)
···
 	param->wq.db_numa_node  = ccp->node;
 	param->eq_ix            = ccp->ix;
 
-	err = mlx5e_alloc_cq_common(priv, param, cq);
+	err = mlx5e_alloc_cq_common(mdev, ccp->netdev, ccp->wq, param, cq);
 
 	cq->napi     = ccp->napi;
 	cq->ch_stats = ccp->ch_stats;
···
 	mlx5_core_destroy_cq(cq->mdev, &cq->mcq);
 }
 
-int mlx5e_open_cq(struct mlx5e_priv *priv, struct dim_cq_moder moder,
+int mlx5e_open_cq(struct mlx5_core_dev *mdev, struct dim_cq_moder moder,
 		  struct mlx5e_cq_param *param, struct mlx5e_create_cq_param *ccp,
 		  struct mlx5e_cq *cq)
 {
-	struct mlx5_core_dev *mdev = priv->mdev;
 	int err;
 
-	err = mlx5e_alloc_cq(priv, param, ccp, cq);
+	err = mlx5e_alloc_cq(mdev, param, ccp, cq);
 	if (err)
 		return err;
 
···
 	int tc;
 
 	for (tc = 0; tc < c->num_tc; tc++) {
-		err = mlx5e_open_cq(c->priv, params->tx_cq_moderation, &cparam->txq_sq.cqp,
+		err = mlx5e_open_cq(c->mdev, params->tx_cq_moderation, &cparam->txq_sq.cqp,
 				    ccp, &c->sq[tc].cq);
 		if (err)
 			goto err_close_tx_cqs;
···
 	for (tc = 0; tc < mlx5e_get_dcb_num_tc(params); tc++) {
 		int txq_ix = c->ix + tc * params->num_channels;
 		u32 qos_queue_group_id;
+		u32 tisn;
 
+		tisn = mlx5e_profile_get_tisn(c->mdev, c->priv, c->priv->profile,
+					      c->lag_port, tc);
 		err = mlx5e_txq_get_qos_node_hw_id(params, txq_ix, &qos_queue_group_id);
 		if (err)
 			goto err_close_sqs;
 
-		err = mlx5e_open_txqsq(c, c->priv->tisn[c->lag_port][tc], txq_ix,
+		err = mlx5e_open_txqsq(c, tisn, txq_ix,
 				       params, &cparam->txq_sq, &c->sq[tc], tc,
 				       qos_queue_group_id,
 				       &c->priv->channel_stats[c->ix]->sq[tc]);
···
 
 	mlx5e_build_create_cq_param(&ccp, c);
 
-	err = mlx5e_open_cq(c->priv, icocq_moder, &cparam->async_icosq.cqp, &ccp,
+	err = mlx5e_open_cq(c->mdev, icocq_moder, &cparam->async_icosq.cqp, &ccp,
 			    &c->async_icosq.cq);
 	if (err)
 		return err;
 
-	err = mlx5e_open_cq(c->priv, icocq_moder, &cparam->icosq.cqp, &ccp,
+	err = mlx5e_open_cq(c->mdev, icocq_moder, &cparam->icosq.cqp, &ccp,
 			    &c->icosq.cq);
 	if (err)
 		goto err_close_async_icosq_cq;
···
 	if (err)
 		goto err_close_icosq_cq;
 
-	err = mlx5e_open_cq(c->priv, params->tx_cq_moderation, &cparam->xdp_sq.cqp, &ccp,
+	err = mlx5e_open_cq(c->mdev, params->tx_cq_moderation, &cparam->xdp_sq.cqp, &ccp,
 			    &c->xdpsq.cq);
 	if (err)
 		goto err_close_tx_cqs;
 
-	err = mlx5e_open_cq(c->priv, params->rx_cq_moderation, &cparam->rq.cqp, &ccp,
+	err = mlx5e_open_cq(c->mdev, params->rx_cq_moderation, &cparam->rq.cqp, &ccp,
 			    &c->rq.cq);
 	if (err)
 		goto err_close_xdp_tx_cqs;
 
-	err = c->xdp ? mlx5e_open_cq(c->priv, params->tx_cq_moderation, &cparam->xdp_sq.cqp,
+	err = c->xdp ? mlx5e_open_cq(c->mdev, params->tx_cq_moderation, &cparam->xdp_sq.cqp,
 				     &ccp, &c->rq_xdpsq.cq) : 0;
 	if (err)
 		goto err_close_rx_cq;
···
 	param->wq.buf_numa_node = dev_to_node(mlx5_core_dma_dev(mdev));
 	param->wq.db_numa_node  = dev_to_node(mlx5_core_dma_dev(mdev));
 
-	return mlx5e_alloc_cq_common(priv, param, cq);
+	return mlx5e_alloc_cq_common(priv->mdev, priv->netdev, priv->wq, param, cq);
 }
 
 int mlx5e_open_drop_rq(struct mlx5e_priv *priv,
···
 	mlx5e_free_cq(&drop_rq->cq);
 }
 
-int mlx5e_create_tis(struct mlx5_core_dev *mdev, void *in, u32 *tisn)
-{
-	void *tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
-
-	MLX5_SET(tisc, tisc, transport_domain, mdev->mlx5e_res.hw_objs.td.tdn);
-
-	if (MLX5_GET(tisc, tisc, tls_en))
-		MLX5_SET(tisc, tisc, pd, mdev->mlx5e_res.hw_objs.pdn);
-
-	if (mlx5_lag_is_lacp_owner(mdev))
-		MLX5_SET(tisc, tisc, strict_lag_tx_port_affinity, 1);
-
-	return mlx5_core_create_tis(mdev, in, tisn);
-}
-
-void mlx5e_destroy_tis(struct mlx5_core_dev *mdev, u32 tisn)
-{
-	mlx5_core_destroy_tis(mdev, tisn);
-}
-
-void mlx5e_destroy_tises(struct mlx5e_priv *priv)
-{
-	int tc, i;
-
-	for (i = 0; i < mlx5e_get_num_lag_ports(priv->mdev); i++)
-		for (tc = 0; tc < priv->profile->max_tc; tc++)
-			mlx5e_destroy_tis(priv->mdev, priv->tisn[i][tc]);
-}
-
-static bool mlx5e_lag_should_assign_affinity(struct mlx5_core_dev *mdev)
-{
-	return MLX5_CAP_GEN(mdev, lag_tx_port_affinity) && mlx5e_get_num_lag_ports(mdev) > 1;
-}
-
-int mlx5e_create_tises(struct mlx5e_priv *priv)
-{
-	int tc, i;
-	int err;
-
-	for (i = 0; i < mlx5e_get_num_lag_ports(priv->mdev); i++) {
-		for (tc = 0; tc < priv->profile->max_tc; tc++) {
-			u32 in[MLX5_ST_SZ_DW(create_tis_in)] = {};
-			void *tisc;
-
-			tisc = MLX5_ADDR_OF(create_tis_in, in, ctx);
-
-			MLX5_SET(tisc, tisc, prio, tc << 1);
-
-			if (mlx5e_lag_should_assign_affinity(priv->mdev))
-				MLX5_SET(tisc, tisc, lag_tx_port_affinity, i + 1);
-
-			err = mlx5e_create_tis(priv->mdev, in, &priv->tisn[i][tc]);
-			if (err)
-				goto err_close_tises;
-		}
-	}
-
-	return 0;
-
-err_close_tises:
-	for (; i >= 0; i--) {
-		for (tc--; tc >= 0; tc--)
-			mlx5e_destroy_tis(priv->mdev, priv->tisn[i][tc]);
-		tc = priv->profile->max_tc;
-	}
-
-	return err;
-}
-
 static void mlx5e_cleanup_nic_tx(struct mlx5e_priv *priv)
 {
 	if (priv->mqprio_rl) {
···
 		priv->mqprio_rl = NULL;
 	}
 	mlx5e_accel_cleanup_tx(priv);
-	mlx5e_destroy_tises(priv);
 }
 
 static int mlx5e_modify_channels_vsd(struct mlx5e_channels *chs, bool vsd)
···
 
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;
 
-	if (tc && tc != MLX5E_MAX_NUM_TC)
+	if (tc && tc != MLX5_MAX_NUM_TC)
 		return -EINVAL;
 
 	new_params = priv->channels.params;
···
 {
 	int err;
 
-	err = mlx5e_create_tises(priv);
-	if (err) {
-		mlx5_core_warn(priv->mdev, "create tises failed, %d\n", err);
-		return err;
-	}
-
 	err = mlx5e_accel_init_tx(priv);
 	if (err)
-		goto err_destroy_tises;
+		return err;
 
 	mlx5e_set_mqprio_rl(priv);
 	mlx5e_dcbnl_initialize(priv);
 	return 0;
-
-err_destroy_tises:
-	mlx5e_destroy_tises(priv);
-	return err;
 }
 
 static void mlx5e_nic_enable(struct mlx5e_priv *priv)
···
 	.update_stats	   = mlx5e_stats_update_ndo_stats,
 	.update_carrier	   = mlx5e_update_carrier,
 	.rx_handlers       = &mlx5e_rx_handlers_nic,
-	.max_tc		   = MLX5E_MAX_NUM_TC,
+	.max_tc		   = MLX5_MAX_NUM_TC,
 	.stats_grps	   = mlx5e_nic_stats_grps,
 	.stats_grps_num	   = mlx5e_nic_stats_grps_num,
 	.features          = BIT(MLX5E_PROFILE_FEATURE_PTP_RX) |
···
 	return 0;
 }
 
-static int mlx5e_suspend(struct auxiliary_device *adev, pm_message_t state)
+static int _mlx5e_suspend(struct auxiliary_device *adev)
 {
 	struct mlx5e_dev *mlx5e_dev = auxiliary_get_drvdata(adev);
 	struct mlx5e_priv *priv = mlx5e_dev->priv;
···
 	return 0;
 }
 
-static int mlx5e_probe(struct auxiliary_device *adev,
-		       const struct auxiliary_device_id *id)
+static int mlx5e_suspend(struct auxiliary_device *adev, pm_message_t state)
+{
+	return _mlx5e_suspend(adev);
+}
+
+static int _mlx5e_probe(struct auxiliary_device *adev)
 {
 	struct mlx5_adev *edev = container_of(adev, struct mlx5_adev, adev);
 	const struct mlx5e_profile *profile = &mlx5e_nic_profile;
 	struct mlx5_core_dev *mdev = edev->mdev;
 	struct mlx5e_dev *mlx5e_dev;
 	struct net_device *netdev;
-	pm_message_t state = {};
 	struct mlx5e_priv *priv;
 	int err;
 
···
 	return 0;
 
 err_resume:
-	mlx5e_suspend(adev, state);
+	_mlx5e_suspend(adev);
 err_profile_cleanup:
 	profile->cleanup(priv);
 err_destroy_netdev:
···
 	return err;
 }
 
+static int mlx5e_probe(struct auxiliary_device *adev,
+		       const struct auxiliary_device_id *id)
+{
+	return _mlx5e_probe(adev);
+}
+
 static void mlx5e_remove(struct auxiliary_device *adev)
 {
 	struct mlx5e_dev *mlx5e_dev = auxiliary_get_drvdata(adev);
 	struct mlx5e_priv *priv = mlx5e_dev->priv;
-	pm_message_t state = {};
 
 	mlx5_core_uplink_netdev_set(priv->mdev, NULL);
 	mlx5e_dcbnl_delete_app(priv);
 	unregister_netdev(priv->netdev);
-	mlx5e_suspend(adev, state);
+	_mlx5e_suspend(adev);
 	priv->profile->cleanup(priv);
 	mlx5e_destroy_netdev(priv);
 	mlx5e_devlink_port_unregister(mlx5e_dev);

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

···
 	struct mlx5e_rep_priv *rpriv = priv->ppriv;
 	int err;
 
-	err = mlx5e_create_tises(priv);
-	if (err) {
-		mlx5_core_warn(priv->mdev, "create tises failed, %d\n", err);
-		return err;
-	}
-
 	err = mlx5e_rep_neigh_init(rpriv);
 	if (err)
 		goto err_neigh_init;
···
 err_init_tx:
 	mlx5e_rep_neigh_cleanup(rpriv);
 err_neigh_init:
-	mlx5e_destroy_tises(priv);
 	return err;
 }
 
···
 		mlx5e_cleanup_uplink_rep_tx(rpriv);
 
 	mlx5e_rep_neigh_cleanup(rpriv);
-	mlx5e_destroy_tises(priv);
 }
 
 static void mlx5e_rep_enable(struct mlx5e_priv *priv)
···
 	.update_stats           = mlx5e_stats_update_ndo_stats,
 	.update_carrier	        = mlx5e_update_carrier,
 	.rx_handlers            = &mlx5e_rx_handlers_rep,
-	.max_tc			= MLX5E_MAX_NUM_TC,
+	.max_tc			= MLX5_MAX_NUM_TC,
 	.stats_grps		= mlx5e_ul_rep_stats_grps,
 	.stats_grps_num		= mlx5e_ul_rep_stats_grps_num,
 };

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

···
 						     (struct mlx5_err_cqe *)cqe);
 				mlx5_wq_cyc_wqe_dump(&sq->wq, ci, wi->num_wqebbs);
 				if (!test_and_set_bit(MLX5E_SQ_STATE_RECOVERING, &sq->state))
-					queue_work(cq->priv->wq, &sq->recover_work);
+					queue_work(cq->workqueue, &sq->recover_work);
 				break;
 			}
 

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

···
 				mlx5e_dump_error_cqe(&sq->cq, sq->sqn,
 						     (struct mlx5_err_cqe *)cqe);
 				mlx5_wq_cyc_wqe_dump(&sq->wq, ci, wi->num_wqebbs);
-				queue_work(cq->priv->wq, &sq->recover_work);
+				queue_work(cq->workqueue, &sq->recover_work);
 			}
 			stats->cqe_err++;
 		}

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

···
 		return mlx5_fs_cmd_get_stub_cmds();
 	}
 }
+
+int mlx5_fs_cmd_set_l2table_entry_silent(struct mlx5_core_dev *dev, u8 silent_mode)
+{
+	u32 in[MLX5_ST_SZ_DW(set_l2_table_entry_in)] = {};
+
+	if (silent_mode && !MLX5_CAP_GEN(dev, silent_mode))
+		return -EOPNOTSUPP;
+
+	MLX5_SET(set_l2_table_entry_in, in, opcode, MLX5_CMD_OP_SET_L2_TABLE_ENTRY);
+	MLX5_SET(set_l2_table_entry_in, in, silent_mode_valid, 1);
+	MLX5_SET(set_l2_table_entry_in, in, silent_mode, silent_mode);
+
+	return mlx5_cmd_exec_in(dev, set_l2_table_entry, in);
+}
+
+int mlx5_fs_cmd_set_tx_flow_table_root(struct mlx5_core_dev *dev, u32 ft_id, bool disconnect)
+{
+	u32 out[MLX5_ST_SZ_DW(set_flow_table_root_out)] = {};
+	u32 in[MLX5_ST_SZ_DW(set_flow_table_root_in)] = {};
+
+	if (disconnect && MLX5_CAP_FLOWTABLE_NIC_TX(dev, reset_root_to_default))
+		return -EOPNOTSUPP;
+
+	MLX5_SET(set_flow_table_root_in, in, opcode,
+		 MLX5_CMD_OP_SET_FLOW_TABLE_ROOT);
+	MLX5_SET(set_flow_table_root_in, in, table_type,
+		 FS_FT_NIC_TX);
+	if (disconnect)
+		MLX5_SET(set_flow_table_root_in, in, op_mod, 1);
+	else
+		MLX5_SET(set_flow_table_root_in, in, table_id, ft_id);
+
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
+}

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

···
 const struct mlx5_flow_cmds *mlx5_fs_cmd_get_default(enum fs_flow_table_type type);
 const struct mlx5_flow_cmds *mlx5_fs_cmd_get_fw_cmds(void);
 
+int mlx5_fs_cmd_set_l2table_entry_silent(struct mlx5_core_dev *dev, u8 silent_mode);
+int mlx5_fs_cmd_set_tx_flow_table_root(struct mlx5_core_dev *dev, u32 ft_id, bool disconnect);
 #endif

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

···
 		return err;
 	}
 
-	err = mlx5i_create_tis(priv->mdev, ipriv->qpn, &priv->tisn[0][0]);
+	err = mlx5i_create_tis(priv->mdev, ipriv->qpn, &ipriv->tisn);
 	if (err) {
 		mlx5_core_warn(priv->mdev, "create tis failed, %d\n", err);
 		goto err_destroy_underlay_qp;
···
 {
 	struct mlx5i_priv *ipriv = priv->ppriv;
 
-	mlx5e_destroy_tis(priv->mdev, priv->tisn[0][0]);
+	mlx5e_destroy_tis(priv->mdev, ipriv->tisn);
 	mlx5i_destroy_underlay_qp(priv->mdev, ipriv->qpn);
 }
 
···
 	return ARRAY_SIZE(mlx5i_stats_grps);
 }
 
+u32 mlx5i_get_tisn(struct mlx5_core_dev *mdev, struct mlx5e_priv *priv, u8 lag_port, u8 tc)
+{
+	struct mlx5i_priv *ipriv = priv->ppriv;
+
+	if (WARN(lag_port || tc,
+		 "IPoIB unexpected non-zero value: lag_port (%u), tc (%u)\n",
+		 lag_port, tc))
+		return 0;
+
+	return ipriv->tisn;
+}
+
 static const struct mlx5e_profile mlx5i_nic_profile = {
 	.init		   = mlx5i_init,
 	.cleanup	   = mlx5i_cleanup,
···
 	.max_tc		   = MLX5I_MAX_NUM_TC,
 	.stats_grps        = mlx5i_stats_grps,
 	.stats_grps_num    = mlx5i_stats_grps_num,
+	.get_tisn          = mlx5i_get_tisn,
 };
 
 /* mlx5i netdev NDos */
···
 	*params = (struct rdma_netdev_alloc_params){
 		.sizeof_priv = sizeof(struct mlx5i_priv) +
 			       sizeof(struct mlx5e_priv),
-		.txqs = nch * MLX5E_MAX_NUM_TC,
+		.txqs = nch * MLX5_MAX_NUM_TC,
 		.rxqs = nch,
 		.param = mdev,
 		.initialize_rdma_netdev = mlx5_rdma_setup_rn,

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

···
 struct mlx5i_priv {
 	struct rdma_netdev rn; /* keep this first */
 	u32 qpn;
+	u32 tisn;
 	bool   sub_interface;
 	u32    num_sub_interfaces;
 	u32    qkey;
···
 };
 
 int mlx5i_create_tis(struct mlx5_core_dev *mdev, u32 underlay_qpn, u32 *tisn);
+u32 mlx5i_get_tisn(struct mlx5_core_dev *mdev, struct mlx5e_priv *priv, u8 lag_port, u8 tc);
 
 /* Underlay QP create/destroy functions */
 int mlx5i_create_underlay_qp(struct mlx5e_priv *priv);

drivers/net/ethernet/mellanox/mlx5/core/ipoib/ipoib_vlan.c

···
 		goto err_unint_underlay_qp;
 	}
 
-	err = mlx5i_create_tis(mdev, ipriv->qpn, &epriv->tisn[0][0]);
+	err = mlx5i_create_tis(mdev, ipriv->qpn, &ipriv->tisn);
 	if (err) {
 		mlx5_core_warn(mdev, "create child tis failed, %d\n", err);
 		goto err_remove_rx_uderlay_qp;
···
 err_close_channels:
 	mlx5e_close_channels(&epriv->channels);
 err_clear_state_opened_flag:
-	mlx5e_destroy_tis(mdev, epriv->tisn[0][0]);
+	mlx5e_destroy_tis(mdev, ipriv->tisn);
 err_remove_rx_uderlay_qp:
 	mlx5_fs_remove_rx_underlay_qpn(mdev, ipriv->qpn);
 err_unint_underlay_qp:
···
 	mlx5i_uninit_underlay_qp(priv);
 	mlx5e_deactivate_priv_channels(priv);
 	mlx5e_close_channels(&priv->channels);
-	mlx5e_destroy_tis(mdev, priv->tisn[0][0]);
+	mlx5e_destroy_tis(mdev, ipriv->tisn);
 unlock:
 	mutex_unlock(&priv->state_lock);
 	return 0;
···
 	.update_stats	   = NULL,
 	.rx_handlers       = &mlx5i_rx_handlers,
 	.max_tc		   = MLX5I_MAX_NUM_TC,
+	.get_tisn          = mlx5i_get_tisn,
 };
 
 const struct mlx5e_profile *mlx5i_pkey_get_profile(void)

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

···
 		devcom_free_comp_dev(devcom);
 }
 
+int mlx5_devcom_comp_get_size(struct mlx5_devcom_comp_dev *devcom)
+{
+	struct mlx5_devcom_comp *comp = devcom->comp;
+
+	return kref_read(&comp->ref);
+}
+
 int mlx5_devcom_send_event(struct mlx5_devcom_comp_dev *devcom,
 			   int event, int rollback_event,
 			   void *event_data)

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

···
 int mlx5_devcom_send_event(struct mlx5_devcom_comp_dev *devcom,
 			   int event, int rollback_event,
 			   void *event_data);
+int mlx5_devcom_comp_get_size(struct mlx5_devcom_comp_dev *devcom);
 
 void mlx5_devcom_comp_set_ready(struct mlx5_devcom_comp_dev *devcom, bool ready);
 bool mlx5_devcom_comp_is_ready(struct mlx5_devcom_comp_dev *devcom);

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

···
 			u8 access_reg_group);
 int mlx5_query_qcam_reg(struct mlx5_core_dev *mdev, u32 *qcam,
 			u8 feature_group, u8 access_reg_group);
+int mlx5_query_mpir_reg(struct mlx5_core_dev *dev, u32 *mpir);
 
 void mlx5_lag_add_netdev(struct mlx5_core_dev *dev, struct net_device *netdev);
 void mlx5_lag_remove_netdev(struct mlx5_core_dev *dev, struct net_device *netdev);

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

···
 	*speed = max_speed;
 	return 0;
 }
+
+int mlx5_query_mpir_reg(struct mlx5_core_dev *dev, u32 *mpir)
+{
+	u32 in[MLX5_ST_SZ_DW(mpir_reg)] = {};
+	int sz = MLX5_ST_SZ_BYTES(mpir_reg);
+
+	MLX5_SET(mpir_reg, in, local_port, 1);
+
+	return mlx5_core_access_reg(dev, in, sz, mpir, sz, MLX5_REG_MPIR, 0, 0);
+}

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

···
 				   bool ignore_flow_level,
 				   u32 flow_source)
 {
-	struct mlx5dr_cmd_flow_destination_hw_info tmp_hw_dest;
 	struct mlx5dr_cmd_flow_destination_hw_info *hw_dests;
 	struct mlx5dr_action **ref_actions;
 	struct mlx5dr_action *action;
···
 	 * one that done in the TX.
 	 * So, if one of the ft target is wire, put it at the end of the dest list.
 	 */
-	if (is_ft_wire && num_dst_ft > 1) {
-		tmp_hw_dest = hw_dests[last_dest];
-		hw_dests[last_dest] = hw_dests[num_of_dests - 1];
-		hw_dests[num_of_dests - 1] = tmp_hw_dest;
-	}
+	if (is_ft_wire && num_dst_ft > 1)
+		swap(hw_dests[last_dest], hw_dests[num_of_dests - 1]);
 
 	action = dr_action_create_generic(DR_ACTION_TYP_FT);
 	if (!action)

include/linux/mlx5/driver.h

···
 	MLX5_REG_MTPPSE		 = 0x9054,
 	MLX5_REG_MTUTC		 = 0x9055,
 	MLX5_REG_MPEGC		 = 0x9056,
+	MLX5_REG_MPIR		 = 0x9059,
 	MLX5_REG_MCQS		 = 0x9060,
 	MLX5_REG_MCQI		 = 0x9061,
 	MLX5_REG_MCC		 = 0x9062,
···
 		struct mlx5_td             td;
 		u32			   mkey;
 		struct mlx5_sq_bfreg       bfreg;
+#define MLX5_MAX_NUM_TC 8
+		u32                        tisn[MLX5_MAX_PORTS][MLX5_MAX_NUM_TC];
 	} hw_objs;
 	struct net_device *uplink_netdev;
 	struct mutex uplink_netdev_lock;

include/linux/mlx5/mlx5_ifc.h

···
 	u8         flow_table_modify[0x1];
 	u8         reformat[0x1];
 	u8         decap[0x1];
-	u8         reserved_at_9[0x1];
+	u8         reset_root_to_default[0x1];
 	u8         pop_vlan[0x1];
 	u8         push_vlan[0x1];
 	u8         reserved_at_c[0x1];
···
 	u8         disable_local_lb_uc[0x1];
 	u8         disable_local_lb_mc[0x1];
 	u8         log_min_hairpin_wq_data_sz[0x5];
-	u8         reserved_at_3e8[0x2];
+	u8         reserved_at_3e8[0x1];
+	u8         silent_mode[0x1];
 	u8         vhca_state[0x1];
 	u8         log_max_vlan_list[0x5];
 	u8         reserved_at_3f0[0x3];
···
 
 	u8         reserved_at_460[0x1];
 	u8         ats[0x1];
-	u8         reserved_at_462[0x1];
+	u8         cross_vhca_rqt[0x1];
 	u8         log_max_uctx[0x5];
 	u8         reserved_at_468[0x1];
 	u8         crypto[0x1];
···
 
 enum {
 	MLX5_CROSS_VHCA_OBJ_TO_OBJ_SUPPORTED_LOCAL_FLOW_TABLE_TO_REMOTE_FLOW_TABLE_MISS  = 0x80000,
+	MLX5_CROSS_VHCA_OBJ_TO_OBJ_SUPPORTED_LOCAL_FLOW_TABLE_ROOT_TO_REMOTE_FLOW_TABLE  = (1ULL << 20),
 };
 
 enum {
···
 	u8	   reserved_at_260[0x120];
 	u8	   reserved_at_380[0x10];
 	u8	   ec_vf_vport_base[0x10];
-	u8	   reserved_at_3a0[0x460];
+
+	u8	   reserved_at_3a0[0x10];
+	u8	   max_rqt_vhca_id[0x10];
+
+	u8	   reserved_at_3c0[0x440];
 };
 
 enum mlx5_ifc_flow_destination_type {
···
 struct mlx5_ifc_rq_num_bits {
 	u8         reserved_at_0[0x8];
 	u8         rq_num[0x18];
+};
+
+struct mlx5_ifc_rq_vhca_bits {
+	u8         reserved_at_0[0x8];
+	u8         rq_num[0x18];
+	u8         reserved_at_20[0x10];
+	u8         rq_vhca_id[0x10];
 };
 
 struct mlx5_ifc_mac_address_layout_bits {
···
 
 	u8    reserved_at_e0[0x6a0];
 
-	struct mlx5_ifc_rq_num_bits rq_num[];
+	union {
+		DECLARE_FLEX_ARRAY(struct mlx5_ifc_rq_num_bits, rq_num);
+		DECLARE_FLEX_ARRAY(struct mlx5_ifc_rq_vhca_bits, rq_vhca);
+	};
 };
 
 enum {
···
 
 	u8         reserved_at_c0[0x20];
 
-	u8         reserved_at_e0[0x13];
+	u8         reserved_at_e0[0x10];
+	u8         silent_mode_valid[0x1];
+	u8         silent_mode[0x1];
+	u8         reserved_at_f2[0x1];
 	u8         vlan_valid[0x1];
 	u8         vlan[0xc];
 
···
 	u8         tx_lossy_overflow_oper[0x2];
 
 	u8         reserved_at_60[0x100];
+};
+
+struct mlx5_ifc_mpir_reg_bits {
+	u8         sdm[0x1];
+	u8         reserved_at_1[0x1b];
+	u8         host_buses[0x4];
+
+	u8         reserved_at_20[0x20];
+
+	u8         local_port[0x8];
+	u8         reserved_at_28[0x15];
+	u8         sd_group[0x3];
+
+	u8         reserved_at_60[0x20];
 };
 
 enum {