[net-next,08/10] net/mlx5: qos: Support cross-esw tx scheduling

Commit Message

Tariq Toukan Feb. 13, 2025, 6:01 p.m. UTC

From: Cosmin Ratiu <[e-mail]>

Up to now, rate groups could only contain vports from the same E-Switch.
This patch relaxes that restriction if the device supports it
(HCA_CAP.esw_cross_esw_sched == true) and the right conditions are met:
- Link Aggregation (LAG) is enabled.
- The E-Switches use the same qos domain.

This also enables the use of the previously added shared esw qos
domains.

Signed-off-by: Cosmin Ratiu <[e-mail]>
Reviewed-by: Carolina Jubran <[e-mail]>
Signed-off-by: Tariq Toukan <[e-mail]>
---
 .../net/ethernet/mellanox/mlx5/core/esw/qos.c | 59 +++++++++++++++----
 1 file changed, 49 insertions(+), 10 deletions(-)

Patch

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c
index 6a469f214187..e6dcfe348a7e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/qos.c
@@ -147,7 +147,9 @@  struct mlx5_esw_sched_node {
 	enum sched_node_type type;
 	/* The eswitch this node belongs to. */
 	struct mlx5_eswitch *esw;
-	/* The children nodes of this node, empty list for leaf nodes. */
+	/* The children nodes of this node, empty list for leaf nodes.
+	 * Can be from multiple E-Switches.
+	 */
 	struct list_head children;
 	/* Valid only if this node is associated with a vport. */
 	struct mlx5_vport *vport;
@@ -398,6 +400,7 @@  static int esw_qos_vport_create_sched_element(struct mlx5_esw_sched_node *vport_
 {
 	u32 sched_ctx[MLX5_ST_SZ_DW(scheduling_context)] = {};
 	struct mlx5_core_dev *dev = vport_node->esw->dev;
+	struct mlx5_vport *vport = vport_node->vport;
 	void *attr;
 
 	if (!mlx5_qos_element_type_supported(dev,
@@ -408,7 +411,13 @@  static int esw_qos_vport_create_sched_element(struct mlx5_esw_sched_node *vport_
 	MLX5_SET(scheduling_context, sched_ctx, element_type,
 		 SCHEDULING_CONTEXT_ELEMENT_TYPE_VPORT);
 	attr = MLX5_ADDR_OF(scheduling_context, sched_ctx, element_attributes);
-	MLX5_SET(vport_element, attr, vport_number, vport_node->vport->vport);
+	MLX5_SET(vport_element, attr, vport_number, vport->vport);
+	if (vport->dev != dev) {
+		/* The port is assigned to a node on another eswitch. */
+		MLX5_SET(vport_element, attr, eswitch_owner_vhca_id_valid, true);
+		MLX5_SET(vport_element, attr, eswitch_owner_vhca_id,
+			 MLX5_CAP_GEN(vport->dev, vhca_id));
+	}
 	MLX5_SET(scheduling_context, sched_ctx, parent_element_id, vport_node->parent->ix);
 	MLX5_SET(scheduling_context, sched_ctx, max_average_bw, vport_node->max_rate);
 
@@ -887,10 +896,16 @@  static int esw_qos_devlink_rate_to_mbps(struct mlx5_core_dev *mdev, const char *
 
 int mlx5_esw_qos_init(struct mlx5_eswitch *esw)
 {
-	if (esw->qos.domain)
-		return 0;  /* Nothing to change. */
+	bool use_shared_domain = esw->mode == MLX5_ESWITCH_OFFLOADS &&
+		MLX5_CAP_QOS(esw->dev, esw_cross_esw_sched);
+
+	if (esw->qos.domain) {
+		if (esw->qos.domain->shared == use_shared_domain)
+			return 0;  /* Nothing to change. */
+		esw_qos_domain_release(esw);
+	}
 
-	return esw_qos_domain_init(esw, false);
+	return esw_qos_domain_init(esw, use_shared_domain);
 }
 
 void mlx5_esw_qos_cleanup(struct mlx5_eswitch *esw)
@@ -1021,16 +1036,40 @@  int mlx5_esw_devlink_rate_node_del(struct devlink_rate *rate_node, void *priv,
 	return 0;
 }
 
-int mlx5_esw_qos_vport_update_parent(struct mlx5_vport *vport, struct mlx5_esw_sched_node *parent,
-				     struct netlink_ext_ack *extack)
+static int mlx5_esw_validate_cross_esw_scheduling(struct mlx5_eswitch *esw,
+						  struct mlx5_esw_sched_node *parent,
+						  struct netlink_ext_ack *extack)
 {
-	struct mlx5_eswitch *esw = vport->dev->priv.eswitch;
-	int err = 0;
+	if (!parent || esw == parent->esw)
+		return 0;
 
-	if (parent && parent->esw != esw) {
+	if (!MLX5_CAP_QOS(esw->dev, esw_cross_esw_sched)) {
 		NL_SET_ERR_MSG_MOD(extack, "Cross E-Switch scheduling is not supported");
 		return -EOPNOTSUPP;
 	}
+	if (esw->qos.domain != parent->esw->qos.domain) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Cannot add vport to a parent belonging to a different qos domain");
+		return -EOPNOTSUPP;
+	}
+	if (!mlx5_lag_is_active(esw->dev)) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Cross E-Switch scheduling requires LAG to be activated");
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
+int mlx5_esw_qos_vport_update_parent(struct mlx5_vport *vport, struct mlx5_esw_sched_node *parent,
+				     struct netlink_ext_ack *extack)
+{
+	struct mlx5_eswitch *esw = vport->dev->priv.eswitch;
+	int err;
+
+	err = mlx5_esw_validate_cross_esw_scheduling(esw, parent, extack);
+	if (err)
+		return err;
 
 	esw_qos_lock(esw);
 	if (!vport->qos.sched_node && parent)