Commit 9d405911 authored by Jakub Kicinski's avatar Jakub Kicinski
Browse files

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

Tariq Toukan says:

====================
mlx5-next updates 2026-01-13

* 'mlx5-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux:
  net/mlx5: Add IFC bits for extended ETS rate limit bandwidth value
  net/mlx5: Add support for querying bond speed
  net/mlx5: Handle port and vport speed change events in MPESW
  net/mlx5: Propagate LAG effective max_tx_speed to vports
  net/mlx5: Add max_tx_speed and its CAP bit to IFC
====================

Link: https://patch.msgid.link/1768299471-1603093-1-git-send-email-tariqt@nvidia.com


Signed-off-by: default avatarJakub Kicinski <kuba@kernel.org>
parents c65182ef 49e41f3e

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

+215 −0
Original line number Diff line number Diff line
@@ -233,14 +233,25 @@ static void mlx5_ldev_free(struct kref *ref) 
{

	struct mlx5_lag *ldev = container_of(ref, struct mlx5_lag, ref);

	struct net *net;

	int i;



	if (ldev->nb.notifier_call) {

		net = read_pnet(&ldev->net);

		unregister_netdevice_notifier_net(net, &ldev->nb);

	}



	mlx5_ldev_for_each(i, 0, ldev) {

		if (ldev->pf[i].dev &&

		    ldev->pf[i].port_change_nb.nb.notifier_call) {

			struct mlx5_nb *nb = &ldev->pf[i].port_change_nb;



			mlx5_eq_notifier_unregister(ldev->pf[i].dev, nb);

		}

	}



	mlx5_lag_mp_cleanup(ldev);

	cancel_delayed_work_sync(&ldev->bond_work);

	cancel_work_sync(&ldev->speed_update_work);

	destroy_workqueue(ldev->wq);

	mutex_destroy(&ldev->lock);

	kfree(ldev);
@@ -274,6 +285,7 @@ static struct mlx5_lag *mlx5_lag_dev_alloc(struct mlx5_core_dev *dev) 
	kref_init(&ldev->ref);

	mutex_init(&ldev->lock);

	INIT_DELAYED_WORK(&ldev->bond_work, mlx5_do_bond_work);

	INIT_WORK(&ldev->speed_update_work, mlx5_mpesw_speed_update_work);



	ldev->nb.notifier_call = mlx5_lag_netdev_event;

	write_pnet(&ldev->net, mlx5_core_net(dev));
@@ -996,6 +1008,137 @@ static bool mlx5_lag_should_disable_lag(struct mlx5_lag *ldev, bool do_bond) 
	       ldev->mode != MLX5_LAG_MODE_MPESW;

}



#ifdef CONFIG_MLX5_ESWITCH

static int

mlx5_lag_sum_devices_speed(struct mlx5_lag *ldev, u32 *sum_speed,

			   int (*get_speed)(struct mlx5_core_dev *, u32 *))

{

	struct mlx5_core_dev *pf_mdev;

	int pf_idx;

	u32 speed;

	int ret;



	*sum_speed = 0;

	mlx5_ldev_for_each(pf_idx, 0, ldev) {

		pf_mdev = ldev->pf[pf_idx].dev;

		if (!pf_mdev)

			continue;



		ret = get_speed(pf_mdev, &speed);

		if (ret) {

			mlx5_core_dbg(pf_mdev,

				      "Failed to get device speed using %ps. Device %s speed is not available (err=%d)\n",

				      get_speed, dev_name(pf_mdev->device),

				      ret);

			return ret;

		}



		*sum_speed += speed;

	}



	return 0;

}



static int mlx5_lag_sum_devices_max_speed(struct mlx5_lag *ldev, u32 *max_speed)

{

	return mlx5_lag_sum_devices_speed(ldev, max_speed,

					  mlx5_port_max_linkspeed);

}



static int mlx5_lag_sum_devices_oper_speed(struct mlx5_lag *ldev,

					   u32 *oper_speed)

{

	return mlx5_lag_sum_devices_speed(ldev, oper_speed,

					  mlx5_port_oper_linkspeed);

}



static void mlx5_lag_modify_device_vports_speed(struct mlx5_core_dev *mdev,

						u32 speed)

{

	u16 op_mod = MLX5_VPORT_STATE_OP_MOD_ESW_VPORT;

	struct mlx5_eswitch *esw = mdev->priv.eswitch;

	struct mlx5_vport *vport;

	unsigned long i;

	int ret;



	if (!esw)

		return;



	if (!MLX5_CAP_ESW(mdev, esw_vport_state_max_tx_speed))

		return;



	mlx5_esw_for_each_vport(esw, i, vport) {

		if (!vport)

			continue;



		if (vport->vport == MLX5_VPORT_UPLINK)

			continue;



		ret = mlx5_modify_vport_max_tx_speed(mdev, op_mod,

						     vport->vport, true, speed);

		if (ret)

			mlx5_core_dbg(mdev,

				      "Failed to set vport %d speed %d, err=%d\n",

				      vport->vport, speed, ret);

	}

}



void mlx5_lag_set_vports_agg_speed(struct mlx5_lag *ldev)

{

	struct mlx5_core_dev *mdev;

	u32 speed;

	int pf_idx;



	if (ldev->mode == MLX5_LAG_MODE_MPESW) {

		if (mlx5_lag_sum_devices_oper_speed(ldev, &speed))

			return;

	} else {

		speed = ldev->tracker.bond_speed_mbps;

		if (speed == SPEED_UNKNOWN)

			return;

	}



	/* If speed is not set, use the sum of max speeds of all PFs */

	if (!speed && mlx5_lag_sum_devices_max_speed(ldev, &speed))

		return;



	speed = speed / MLX5_MAX_TX_SPEED_UNIT;



	mlx5_ldev_for_each(pf_idx, 0, ldev) {

		mdev = ldev->pf[pf_idx].dev;

		if (!mdev)

			continue;



		mlx5_lag_modify_device_vports_speed(mdev, speed);

	}

}



void mlx5_lag_reset_vports_speed(struct mlx5_lag *ldev)

{

	struct mlx5_core_dev *mdev;

	u32 speed;

	int pf_idx;

	int ret;



	mlx5_ldev_for_each(pf_idx, 0, ldev) {

		mdev = ldev->pf[pf_idx].dev;

		if (!mdev)

			continue;



		ret = mlx5_port_oper_linkspeed(mdev, &speed);

		if (ret) {

			mlx5_core_dbg(mdev,

				      "Failed to reset vports speed for device %s. Oper speed is not available (err=%d)\n",

				      dev_name(mdev->device), ret);

			continue;

		}



		speed = speed / MLX5_MAX_TX_SPEED_UNIT;

		mlx5_lag_modify_device_vports_speed(mdev, speed);

	}

}

#endif



static void mlx5_do_bond(struct mlx5_lag *ldev)

{

	int idx = mlx5_lag_get_dev_index_by_seq(ldev, MLX5_LAG_P1);
@@ -1083,9 +1226,12 @@ static void mlx5_do_bond(struct mlx5_lag *ldev) 
						     ndev);

			dev_put(ndev);

		}

		mlx5_lag_set_vports_agg_speed(ldev);

	} else if (mlx5_lag_should_modify_lag(ldev, do_bond)) {

		mlx5_modify_lag(ldev, &tracker);

		mlx5_lag_set_vports_agg_speed(ldev);

	} else if (mlx5_lag_should_disable_lag(ldev, do_bond)) {

		mlx5_lag_reset_vports_speed(ldev);

		mlx5_disable_lag(ldev);

	}

}
@@ -1286,6 +1432,65 @@ static int mlx5_handle_changeinfodata_event(struct mlx5_lag *ldev, 
	return 1;

}



static void mlx5_lag_update_tracker_speed(struct lag_tracker *tracker,

					  struct net_device *ndev)

{

	struct ethtool_link_ksettings lksettings;

	struct net_device *bond_dev;

	int err;



	if (netif_is_lag_master(ndev))

		bond_dev = ndev;

	else

		bond_dev = netdev_master_upper_dev_get(ndev);



	if (!bond_dev) {

		tracker->bond_speed_mbps = SPEED_UNKNOWN;

		return;

	}



	err = __ethtool_get_link_ksettings(bond_dev, &lksettings);

	if (err) {

		netdev_dbg(bond_dev,

			   "Failed to get speed for bond dev %s, err=%d\n",

			   bond_dev->name, err);

		tracker->bond_speed_mbps = SPEED_UNKNOWN;

		return;

	}



	if (lksettings.base.speed == SPEED_UNKNOWN)

		tracker->bond_speed_mbps = 0;

	else

		tracker->bond_speed_mbps = lksettings.base.speed;

}



/* Returns speed in Mbps. */

int mlx5_lag_query_bond_speed(struct mlx5_core_dev *mdev, u32 *speed)

{

	struct mlx5_lag *ldev;

	unsigned long flags;

	int ret = 0;



	spin_lock_irqsave(&lag_lock, flags);

	ldev = mlx5_lag_dev(mdev);

	if (!ldev) {

		ret = -ENODEV;

		goto unlock;

	}



	*speed = ldev->tracker.bond_speed_mbps;



	if (*speed == SPEED_UNKNOWN) {

		mlx5_core_dbg(mdev, "Bond speed is unknown\n");

		ret = -EINVAL;

	}



unlock:

	spin_unlock_irqrestore(&lag_lock, flags);

	return ret;

}

EXPORT_SYMBOL_GPL(mlx5_lag_query_bond_speed);



/* this handler is always registered to netdev events */

static int mlx5_lag_netdev_event(struct notifier_block *this,

				 unsigned long event, void *ptr)
@@ -1317,6 +1522,9 @@ static int mlx5_lag_netdev_event(struct notifier_block *this, 
		break;

	}



	if (changed)

		mlx5_lag_update_tracker_speed(&tracker, ndev);



	ldev->tracker = tracker;



	if (changed)
@@ -1362,6 +1570,10 @@ static void mlx5_ldev_add_mdev(struct mlx5_lag *ldev, 


	ldev->pf[fn].dev = dev;

	dev->priv.lag = ldev;



	MLX5_NB_INIT(&ldev->pf[fn].port_change_nb,

		     mlx5_lag_mpesw_port_change_event, PORT_CHANGE);

	mlx5_eq_notifier_register(dev, &ldev->pf[fn].port_change_nb);

}



static void mlx5_ldev_remove_mdev(struct mlx5_lag *ldev,
@@ -1373,6 +1585,9 @@ static void mlx5_ldev_remove_mdev(struct mlx5_lag *ldev, 
	if (ldev->pf[fn].dev != dev)

		return;



	if (ldev->pf[fn].port_change_nb.nb.notifier_call)

		mlx5_eq_notifier_unregister(dev, &ldev->pf[fn].port_change_nb);



	ldev->pf[fn].dev = NULL;

	dev->priv.lag = NULL;

}

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

+11 −0
Original line number Diff line number Diff line
@@ -39,6 +39,7 @@ struct lag_func { 
	struct mlx5_core_dev *dev;

	struct net_device    *netdev;

	bool has_drop;

	struct mlx5_nb port_change_nb;

};



/* Used for collection of netdev event info. */
@@ -48,6 +49,7 @@ struct lag_tracker { 
	unsigned int is_bonded:1;

	unsigned int has_inactive:1;

	enum netdev_lag_hash hash_type;

	u32 bond_speed_mbps;

};



/* LAG data of a ConnectX card.
@@ -66,6 +68,7 @@ struct mlx5_lag { 
	struct lag_tracker        tracker;

	struct workqueue_struct   *wq;

	struct delayed_work       bond_work;

	struct work_struct        speed_update_work;

	struct notifier_block     nb;

	possible_net_t net;

	struct lag_mp             lag_mp;
@@ -116,6 +119,14 @@ int mlx5_deactivate_lag(struct mlx5_lag *ldev); 
void mlx5_lag_add_devices(struct mlx5_lag *ldev);

struct mlx5_devcom_comp_dev *mlx5_lag_get_devcom_comp(struct mlx5_lag *ldev);



#ifdef CONFIG_MLX5_ESWITCH

void mlx5_lag_set_vports_agg_speed(struct mlx5_lag *ldev);

void mlx5_lag_reset_vports_speed(struct mlx5_lag *ldev);

#else

static inline void mlx5_lag_set_vports_agg_speed(struct mlx5_lag *ldev) {}

static inline void mlx5_lag_reset_vports_speed(struct mlx5_lag *ldev) {}

#endif



static inline bool mlx5_lag_is_supported(struct mlx5_core_dev *dev)

{

	if (!MLX5_CAP_GEN(dev, vport_group_manager) ||

drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.c

+39 −0
Original line number Diff line number Diff line
@@ -110,6 +110,8 @@ static int enable_mpesw(struct mlx5_lag *ldev) 
			goto err_rescan_drivers;

	}



	mlx5_lag_set_vports_agg_speed(ldev);



	return 0;



err_rescan_drivers:
@@ -223,3 +225,40 @@ bool mlx5_lag_is_mpesw(struct mlx5_core_dev *dev) 
	return ldev && ldev->mode == MLX5_LAG_MODE_MPESW;

}

EXPORT_SYMBOL(mlx5_lag_is_mpesw);



void mlx5_mpesw_speed_update_work(struct work_struct *work)

{

	struct mlx5_lag *ldev = container_of(work, struct mlx5_lag,

					     speed_update_work);



	mutex_lock(&ldev->lock);

	if (ldev->mode == MLX5_LAG_MODE_MPESW) {

		if (ldev->mode_changes_in_progress)

			queue_work(ldev->wq, &ldev->speed_update_work);

		else

			mlx5_lag_set_vports_agg_speed(ldev);

	}



	mutex_unlock(&ldev->lock);

}



int mlx5_lag_mpesw_port_change_event(struct notifier_block *nb,

				     unsigned long event, void *data)

{

	struct mlx5_nb *mlx5_nb = container_of(nb, struct mlx5_nb, nb);

	struct lag_func *lag_func = container_of(mlx5_nb,

						 struct lag_func,

						 port_change_nb);

	struct mlx5_core_dev *dev = lag_func->dev;

	struct mlx5_lag *ldev = dev->priv.lag;

	struct mlx5_eqe *eqe = data;



	if (!ldev)

		return NOTIFY_DONE;



	if (eqe->sub_type == MLX5_PORT_CHANGE_SUBTYPE_DOWN ||

	    eqe->sub_type == MLX5_PORT_CHANGE_SUBTYPE_ACTIVE)

		queue_work(ldev->wq, &ldev->speed_update_work);



	return NOTIFY_OK;

}

drivers/net/ethernet/mellanox/mlx5/core/lag/mpesw.h

+14 −0
Original line number Diff line number Diff line
@@ -32,4 +32,18 @@ bool mlx5_lag_is_mpesw(struct mlx5_core_dev *dev); 
void mlx5_lag_mpesw_disable(struct mlx5_core_dev *dev);

int mlx5_lag_mpesw_enable(struct mlx5_core_dev *dev);



#ifdef CONFIG_MLX5_ESWITCH

void mlx5_mpesw_speed_update_work(struct work_struct *work);

int mlx5_lag_mpesw_port_change_event(struct notifier_block *nb,

				     unsigned long event, void *data);

#else

static inline void mlx5_mpesw_speed_update_work(struct work_struct *work) {}

static inline int mlx5_lag_mpesw_port_change_event(struct notifier_block *nb,

						   unsigned long event,

						   void *data)

{

	return NOTIFY_DONE;

}

#endif /* CONFIG_MLX5_ESWITCH */



#endif /* __MLX5_LAG_MPESW_H__ */

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

+1 −0
Original line number Diff line number Diff line
@@ -381,6 +381,7 @@ const struct mlx5_link_info *mlx5_port_ptys2info(struct mlx5_core_dev *mdev, 
u32 mlx5_port_info2linkmodes(struct mlx5_core_dev *mdev,

			     struct mlx5_link_info *info,

			     bool force_legacy);

int mlx5_port_oper_linkspeed(struct mlx5_core_dev *mdev, u32 *speed);

int mlx5_port_max_linkspeed(struct mlx5_core_dev *mdev, u32 *speed);



#define MLX5_PPS_CAP(mdev) (MLX5_CAP_GEN((mdev), pps) &&		\