Merge branch 'octeontx2-pf-round-robin-sched'

Hariprasad Kelam says:

====================
octeontx2-pf: support Round Robin scheduling

octeontx2 and CN10K silicons support Round Robin scheduling. When multiple
traffic flows reach transmit level with the same priority, with Round Robin
scheduling traffic flow with the highest quantum value is picked. With this
support, the user can add multiple classes with the same priority and
different quantum in htb offload.

This series of patches adds support for the same.

Patch1: implement transmit schedular allocation algorithm as preparation
        for support round robin scheduling.

Patch2: Allow quantum parameter in HTB offload mode.

Patch3: extends octeontx2 htb offload support for Round Robin scheduling

Patch4: extend QOS documentation for Round Robin scheduling

Hariprasad Kelam (1):
  docs: octeontx2: extend documentation for Round Robin scheduling

Naveen Mamindlapalli (3):
  octeontx2-pf: implement transmit schedular allocation algorithm
  sch_htb: Allow HTB quantum parameter in offload mode
  octeontx2-pf: htb offload support for Round Robin scheduling
---
v4 * update classid values in documentation.

v3 * 1. update QOS documentation for round robin scheduling
     2. added out of bound checks for quantum parameter

v2 * change data type of otx2_index_used to reduce size of structure
     otx2_qos_cfg
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

8 files changed, 388 insertions, 43 deletions

diff --git a/Documentation/networking/device_drivers/ethernet/marvell/octeontx2.rst b/Documentation/networking/device_drivers/ethernet/marvell/octeontx2.rst
index bfd233cfac35..1e196cb9ce25 100644
--- a/Documentation/networking/device_drivers/ethernet/marvell/octeontx2.rst
+++ b/Documentation/networking/device_drivers/ethernet/marvell/octeontx2.rst
@@ -332,3 +332,11 @@ Setup HTB offload
         # tc class add dev <interface> parent 1: classid 1:1 htb rate 10Gbit prio 1
 
         # tc class add dev <interface> parent 1: classid 1:2 htb rate 10Gbit prio 7
+
+4. Create tc classes with same priorities and different quantum::
+
+        # tc class add dev <interface> parent 1: classid 1:1 htb rate 10Gbit prio 2 quantum 409600
+
+        # tc class add dev <interface> parent 1: classid 1:2 htb rate 10Gbit prio 2 quantum 188416
+
+        # tc class add dev <interface> parent 1: classid 1:3 htb rate 10Gbit prio 2 quantum 32768
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c
index 77c8f650f7ac..8cdd92dd9762 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c
@@ -774,6 +774,7 @@ int otx2_txsch_alloc(struct otx2_nic *pfvf)
 				rsp->schq_list[lvl][schq];
 
 	pfvf->hw.txschq_link_cfg_lvl = rsp->link_cfg_lvl;
+	pfvf->hw.txschq_aggr_lvl_rr_prio = rsp->aggr_lvl_rr_prio;
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
index ba8091131ec0..37d792f18809 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h
@@ -224,6 +224,7 @@ struct otx2_hw {
 
 	/* NIX */
 	u8			txschq_link_cfg_lvl;
+	u8			txschq_aggr_lvl_rr_prio;
 	u16			txschq_list[NIX_TXSCH_LVL_CNT][MAX_TXSCHQ_PER_FUNC];
 	u16			matchall_ipolicer;
 	u32			dwrr_mtu;
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/qos.c b/drivers/net/ethernet/marvell/octeontx2/nic/qos.c
index d3a76c5ccda8..1e77bbf5d22a 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/qos.c
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/qos.c
@@ -19,6 +19,9 @@
 #define OTX2_QOS_CLASS_NONE		0
 #define OTX2_QOS_DEFAULT_PRIO		0xF
 #define OTX2_QOS_INVALID_SQ		0xFFFF
+#define OTX2_QOS_INVALID_TXSCHQ_IDX	0xFFFF
+#define CN10K_MAX_RR_WEIGHT		GENMASK_ULL(13, 0)
+#define OTX2_MAX_RR_QUANTUM		GENMASK_ULL(23, 0)
 
 static void otx2_qos_update_tx_netdev_queues(struct otx2_nic *pfvf)
 {
@@ -65,11 +68,24 @@ static void otx2_qos_get_regaddr(struct otx2_qos_node *node,
 	}
 }
 
+static int otx2_qos_quantum_to_dwrr_weight(struct otx2_nic *pfvf, u32 quantum)
+{
+	u32 weight;
+
+	weight = quantum / pfvf->hw.dwrr_mtu;
+	if (quantum % pfvf->hw.dwrr_mtu)
+		weight += 1;
+
+	return weight;
+}
+
 static void otx2_config_sched_shaping(struct otx2_nic *pfvf,
 				      struct otx2_qos_node *node,
 				      struct nix_txschq_config *cfg,
 				      int *num_regs)
 {
+	u32 rr_weight;
+	u32 quantum;
 	u64 maxrate;
 
 	otx2_qos_get_regaddr(node, cfg, *num_regs);
@@ -86,8 +102,17 @@ static void otx2_config_sched_shaping(struct otx2_nic *pfvf,
 		return;
 	}
 
-	/* configure priority  */
-	cfg->regval[*num_regs] = (node->schq - node->parent->prio_anchor) << 24;
+	/* configure priority/quantum  */
+	if (node->is_static) {
+		cfg->regval[*num_regs] =
+			(node->schq - node->parent->prio_anchor) << 24;
+	} else {
+		quantum = node->quantum ?
+			  node->quantum : pfvf->tx_max_pktlen;
+		rr_weight = otx2_qos_quantum_to_dwrr_weight(pfvf, quantum);
+		cfg->regval[*num_regs] = node->parent->child_dwrr_prio << 24 |
+					 rr_weight;
+	}
 	(*num_regs)++;
 
 	/* configure PIR */
@@ -195,9 +220,8 @@ static int otx2_qos_txschq_set_parent_topology(struct otx2_nic *pfvf,
 		cfg->reg[0] = NIX_AF_TL1X_TOPOLOGY(parent->schq);
 
 	cfg->regval[0] = (u64)parent->prio_anchor << 32;
-	if (parent->level == NIX_TXSCH_LVL_TL1)
-		cfg->regval[0] |= (u64)TXSCH_TL1_DFLT_RR_PRIO << 1;
-
+	cfg->regval[0] |= ((parent->child_dwrr_prio != OTX2_QOS_DEFAULT_PRIO) ?
+			    parent->child_dwrr_prio : 0)  << 1;
 	cfg->num_regs++;
 
 	rc = otx2_sync_mbox_msg(&pfvf->mbox);
@@ -315,9 +339,14 @@ static void otx2_qos_fill_cfg_tl(struct otx2_qos_node *parent,
 
 	list_for_each_entry(node, &parent->child_list, list) {
 		otx2_qos_fill_cfg_tl(node, cfg);
-		cfg->schq_contig[node->level]++;
 		otx2_qos_fill_cfg_schq(node, cfg);
 	}
+
+	/* Assign the required number of transmit schedular queues under the
+	 * given class
+	 */
+	cfg->schq_contig[parent->level - 1] += parent->child_dwrr_cnt +
+					       parent->max_static_prio + 1;
 }
 
 static void otx2_qos_prepare_txschq_cfg(struct otx2_nic *pfvf,
@@ -378,10 +407,12 @@ otx2_qos_alloc_root(struct otx2_nic *pfvf)
 		return ERR_PTR(-ENOMEM);
 
 	node->parent = NULL;
-	if (!is_otx2_vf(pfvf->pcifunc))
+	if (!is_otx2_vf(pfvf->pcifunc)) {
 		node->level = NIX_TXSCH_LVL_TL1;
-	else
+	} else {
 		node->level = NIX_TXSCH_LVL_TL2;
+		node->child_dwrr_prio = OTX2_QOS_DEFAULT_PRIO;
+	}
 
 	WRITE_ONCE(node->qid, OTX2_QOS_QID_INNER);
 	node->classid = OTX2_QOS_ROOT_CLASSID;
@@ -401,9 +432,13 @@ static int otx2_qos_add_child_node(struct otx2_qos_node *parent,
 	struct otx2_qos_node *tmp_node;
 	struct list_head *tmp;
 
+	if (node->prio > parent->max_static_prio)
+		parent->max_static_prio = node->prio;
+
 	for (tmp = head->next; tmp != head; tmp = tmp->next) {
 		tmp_node = list_entry(tmp, struct otx2_qos_node, list);
-		if (tmp_node->prio == node->prio)
+		if (tmp_node->prio == node->prio &&
+		    tmp_node->is_static)
 			return -EEXIST;
 		if (tmp_node->prio > node->prio) {
 			list_add_tail(&node->list, tmp);
@@ -434,6 +469,10 @@ static int otx2_qos_alloc_txschq_node(struct otx2_nic *pfvf,
 		txschq_node->rate = 0;
 		txschq_node->ceil = 0;
 		txschq_node->prio = 0;
+		txschq_node->quantum = 0;
+		txschq_node->is_static = true;
+		txschq_node->child_dwrr_prio = OTX2_QOS_DEFAULT_PRIO;
+		txschq_node->txschq_idx = OTX2_QOS_INVALID_TXSCHQ_IDX;
 
 		mutex_lock(&pfvf->qos.qos_lock);
 		list_add_tail(&txschq_node->list, &node->child_schq_list);
@@ -459,7 +498,7 @@ static struct otx2_qos_node *
 otx2_qos_sw_create_leaf_node(struct otx2_nic *pfvf,
 			     struct otx2_qos_node *parent,
 			     u16 classid, u32 prio, u64 rate, u64 ceil,
-			     u16 qid)
+			     u32 quantum, u16 qid, bool static_cfg)
 {
 	struct otx2_qos_node *node;
 	int err;
@@ -476,6 +515,10 @@ otx2_qos_sw_create_leaf_node(struct otx2_nic *pfvf,
 	node->rate = otx2_convert_rate(rate);
 	node->ceil = otx2_convert_rate(ceil);
 	node->prio = prio;
+	node->quantum = quantum;
+	node->is_static = static_cfg;
+	node->child_dwrr_prio = OTX2_QOS_DEFAULT_PRIO;
+	node->txschq_idx = OTX2_QOS_INVALID_TXSCHQ_IDX;
 
 	__set_bit(qid, pfvf->qos.qos_sq_bmap);
 
@@ -622,12 +665,28 @@ static int otx2_qos_txschq_alloc(struct otx2_nic *pfvf,
 	}
 
 	pfvf->qos.link_cfg_lvl = rsp->link_cfg_lvl;
+	pfvf->hw.txschq_aggr_lvl_rr_prio = rsp->aggr_lvl_rr_prio;
 
 out:
 	mutex_unlock(&mbox->lock);
 	return rc;
 }
 
+static void otx2_qos_free_unused_txschq(struct otx2_nic *pfvf,
+					struct otx2_qos_cfg *cfg)
+{
+	int lvl, idx, schq;
+
+	for (lvl = 0; lvl < NIX_TXSCH_LVL_CNT; lvl++) {
+		for (idx = 0; idx < cfg->schq_contig[lvl]; idx++) {
+			if (!cfg->schq_index_used[lvl][idx]) {
+				schq = cfg->schq_contig_list[lvl][idx];
+				otx2_txschq_free_one(pfvf, lvl, schq);
+			}
+		}
+	}
+}
+
 static void otx2_qos_txschq_fill_cfg_schq(struct otx2_nic *pfvf,
 					  struct otx2_qos_node *node,
 					  struct otx2_qos_cfg *cfg)
@@ -652,9 +711,11 @@ static void otx2_qos_txschq_fill_cfg_tl(struct otx2_nic *pfvf,
 	list_for_each_entry(tmp, &node->child_list, list) {
 		otx2_qos_txschq_fill_cfg_tl(pfvf, tmp, cfg);
 		cnt = cfg->static_node_pos[tmp->level];
-		tmp->schq = cfg->schq_contig_list[tmp->level][cnt];
+		tmp->schq = cfg->schq_contig_list[tmp->level][tmp->txschq_idx];
+		cfg->schq_index_used[tmp->level][tmp->txschq_idx] = true;
 		if (cnt == 0)
-			node->prio_anchor = tmp->schq;
+			node->prio_anchor =
+				cfg->schq_contig_list[tmp->level][0];
 		cfg->static_node_pos[tmp->level]++;
 		otx2_qos_txschq_fill_cfg_schq(pfvf, tmp, cfg);
 	}
@@ -667,9 +728,87 @@ static void otx2_qos_txschq_fill_cfg(struct otx2_nic *pfvf,
 	mutex_lock(&pfvf->qos.qos_lock);
 	otx2_qos_txschq_fill_cfg_tl(pfvf, node, cfg);
 	otx2_qos_txschq_fill_cfg_schq(pfvf, node, cfg);
+	otx2_qos_free_unused_txschq(pfvf, cfg);
 	mutex_unlock(&pfvf->qos.qos_lock);
 }
 
+static void __otx2_qos_assign_base_idx_tl(struct otx2_nic *pfvf,
+					  struct otx2_qos_node *tmp,
+					  unsigned long *child_idx_bmap,
+					  int child_cnt)
+{
+	int idx;
+
+	if (tmp->txschq_idx != OTX2_QOS_INVALID_TXSCHQ_IDX)
+		return;
+
+	/* assign static nodes 1:1 prio mapping first, then remaining nodes */
+	for (idx = 0; idx < child_cnt; idx++) {
+		if (tmp->is_static && tmp->prio == idx &&
+		    !test_bit(idx, child_idx_bmap)) {
+			tmp->txschq_idx = idx;
+			set_bit(idx, child_idx_bmap);
+			return;
+		} else if (!tmp->is_static && idx >= tmp->prio &&
+			   !test_bit(idx, child_idx_bmap)) {
+			tmp->txschq_idx = idx;
+			set_bit(idx, child_idx_bmap);
+			return;
+		}
+	}
+}
+
+static int otx2_qos_assign_base_idx_tl(struct otx2_nic *pfvf,
+				       struct otx2_qos_node *node)
+{
+	unsigned long *child_idx_bmap;
+	struct otx2_qos_node *tmp;
+	int child_cnt;
+
+	list_for_each_entry(tmp, &node->child_list, list)
+		tmp->txschq_idx = OTX2_QOS_INVALID_TXSCHQ_IDX;
+
+	/* allocate child index array */
+	child_cnt = node->child_dwrr_cnt + node->max_static_prio + 1;
+	child_idx_bmap = kcalloc(BITS_TO_LONGS(child_cnt),
+				 sizeof(unsigned long),
+				 GFP_KERNEL);
+	if (!child_idx_bmap)
+		return -ENOMEM;
+
+	list_for_each_entry(tmp, &node->child_list, list)
+		otx2_qos_assign_base_idx_tl(pfvf, tmp);
+
+	/* assign base index of static priority children first */
+	list_for_each_entry(tmp, &node->child_list, list) {
+		if (!tmp->is_static)
+			continue;
+		__otx2_qos_assign_base_idx_tl(pfvf, tmp, child_idx_bmap,
+					      child_cnt);
+	}
+
+	/* assign base index of dwrr priority children */
+	list_for_each_entry(tmp, &node->child_list, list)
+		__otx2_qos_assign_base_idx_tl(pfvf, tmp, child_idx_bmap,
+					      child_cnt);
+
+	kfree(child_idx_bmap);
+
+	return 0;
+}
+
+static int otx2_qos_assign_base_idx(struct otx2_nic *pfvf,
+				    struct otx2_qos_node *node)
+{
+	int ret = 0;
+
+	mutex_lock(&pfvf->qos.qos_lock);
+	ret = otx2_qos_assign_base_idx_tl(pfvf, node);
+	mutex_unlock(&pfvf->qos.qos_lock);
+
+	return ret;
+}
+
 static int otx2_qos_txschq_push_cfg_schq(struct otx2_nic *pfvf,
 					 struct otx2_qos_node *node,
 					 struct otx2_qos_cfg *cfg)
@@ -761,8 +900,10 @@ static void otx2_qos_free_cfg(struct otx2_nic *pfvf, struct otx2_qos_cfg *cfg)
 
 	for (lvl = 0; lvl < NIX_TXSCH_LVL_CNT; lvl++) {
 		for (idx = 0; idx < cfg->schq_contig[lvl]; idx++) {
-			schq = cfg->schq_contig_list[lvl][idx];
-			otx2_txschq_free_one(pfvf, lvl, schq);
+			if (cfg->schq_index_used[lvl][idx]) {
+				schq = cfg->schq_contig_list[lvl][idx];
+				otx2_txschq_free_one(pfvf, lvl, schq);
+			}
 		}
 	}
 }
@@ -838,6 +979,10 @@ static int otx2_qos_push_txschq_cfg(struct otx2_nic *pfvf,
 	if (ret)
 		return -ENOSPC;
 
+	ret = otx2_qos_assign_base_idx(pfvf, node);
+	if (ret)
+		return -ENOMEM;
+
 	if (!(pfvf->netdev->flags & IFF_UP)) {
 		otx2_qos_txschq_fill_cfg(pfvf, node, cfg);
 		return 0;
@@ -894,6 +1039,13 @@ static int otx2_qos_root_add(struct otx2_nic *pfvf, u16 htb_maj_id, u16 htb_defc
 		goto free_root_node;
 	}
 
+	/* Update TL1 RR PRIO */
+	if (root->level == NIX_TXSCH_LVL_TL1) {
+		root->child_dwrr_prio = pfvf->hw.txschq_aggr_lvl_rr_prio;
+		netdev_dbg(pfvf->netdev,
+			   "TL1 DWRR Priority %d\n", root->child_dwrr_prio);
+	}
+
 	if (!(pfvf->netdev->flags & IFF_UP) ||
 	    root->level == NIX_TXSCH_LVL_TL1) {
 		root->schq = new_cfg->schq_list[root->level][0];
@@ -940,37 +1092,126 @@ static int otx2_qos_root_destroy(struct otx2_nic *pfvf)
 	return 0;
 }
 
+static int otx2_qos_validate_quantum(struct otx2_nic *pfvf, u32 quantum)
+{
+	u32 rr_weight = otx2_qos_quantum_to_dwrr_weight(pfvf, quantum);
+	int err = 0;
+
+	/* Max Round robin weight supported by octeontx2 and CN10K
+	 * is different. Validate accordingly
+	 */
+	if (is_dev_otx2(pfvf->pdev))
+		err = (rr_weight > OTX2_MAX_RR_QUANTUM) ? -EINVAL : 0;
+	else if	(rr_weight > CN10K_MAX_RR_WEIGHT)
+		err = -EINVAL;
+
+	return err;
+}
+
+static int otx2_qos_validate_dwrr_cfg(struct otx2_qos_node *parent,
+				      struct netlink_ext_ack *extack,
+				      struct otx2_nic *pfvf,
+				      u64 prio, u64 quantum)
+{
+	int err;
+
+	err = otx2_qos_validate_quantum(pfvf, quantum);
+	if (err) {
+		NL_SET_ERR_MSG_MOD(extack, "Unsupported quantum value");
+		return err;
+	}
+
+	if (parent->child_dwrr_prio == OTX2_QOS_DEFAULT_PRIO) {
+		parent->child_dwrr_prio = prio;
+	} else if (prio != parent->child_dwrr_prio) {
+		NL_SET_ERR_MSG_MOD(extack, "Only one DWRR group is allowed");
+		return -EOPNOTSUPP;
+	}
+
+	return 0;
+}
+
 static int otx2_qos_validate_configuration(struct otx2_qos_node *parent,
 					   struct netlink_ext_ack *extack,
 					   struct otx2_nic *pfvf,
-					   u64 prio)
+					   u64 prio, bool static_cfg)
 {
-	if (test_bit(prio, parent->prio_bmap)) {
-		NL_SET_ERR_MSG_MOD(extack,
-				   "Static priority child with same priority exists");
+	if (prio == parent->child_dwrr_prio && static_cfg) {
+		NL_SET_ERR_MSG_MOD(extack, "DWRR child group with same priority exists");
 		return -EEXIST;
 	}
 
-	if (prio == TXSCH_TL1_DFLT_RR_PRIO) {
+	if (static_cfg && test_bit(prio, parent->prio_bmap)) {
 		NL_SET_ERR_MSG_MOD(extack,
-				   "Priority is reserved for Round Robin");
-		return -EINVAL;
+				   "Static priority child with same priority exists");
+		return -EEXIST;
 	}
 
 	return 0;
 }
 
+static void otx2_reset_dwrr_prio(struct otx2_qos_node *parent, u64 prio)
+{
+	/* For PF, root node dwrr priority is static */
+	if (parent->level == NIX_TXSCH_LVL_TL1)
+		return;
+
+	if (parent->child_dwrr_prio != OTX2_QOS_DEFAULT_PRIO) {
+		parent->child_dwrr_prio = OTX2_QOS_DEFAULT_PRIO;
+		clear_bit(prio, parent->prio_bmap);
+	}
+}
+
+static bool is_qos_node_dwrr(struct otx2_qos_node *parent,
+			     struct otx2_nic *pfvf,
+			     u64 prio)
+{
+	struct otx2_qos_node *node;
+	bool ret = false;
+
+	if (parent->child_dwrr_prio == prio)
+		return true;
+
+	mutex_lock(&pfvf->qos.qos_lock);
+	list_for_each_entry(node, &parent->child_list, list) {
+		if (prio == node->prio) {
+			if (parent->child_dwrr_prio != OTX2_QOS_DEFAULT_PRIO &&
+			    parent->child_dwrr_prio != prio)
+				continue;
+
+			if (otx2_qos_validate_quantum(pfvf, node->quantum)) {
+				netdev_err(pfvf->netdev,
+					   "Unsupported quantum value for existing classid=0x%x quantum=%d prio=%d",
+					    node->classid, node->quantum,
+					    node->prio);
+				break;
+			}
+			/* mark old node as dwrr */
+			node->is_static = false;
+			parent->child_dwrr_cnt++;
+			parent->child_static_cnt--;
+			ret = true;
+			break;
+		}
+	}
+	mutex_unlock(&pfvf->qos.qos_lock);
+
+	return ret;
+}
+
 static int otx2_qos_leaf_alloc_queue(struct otx2_nic *pfvf, u16 classid,
 				     u32 parent_classid, u64 rate, u64 ceil,
-				     u64 prio, struct netlink_ext_ack *extack)
+				     u64 prio, u32 quantum,
+				     struct netlink_ext_ack *extack)
 {
 	struct otx2_qos_cfg *old_cfg, *new_cfg;
 	struct otx2_qos_node *node, *parent;
 	int qid, ret, err;
+	bool static_cfg;
 
 	netdev_dbg(pfvf->netdev,
-		   "TC_HTB_LEAF_ALLOC_QUEUE: classid=0x%x parent_classid=0x%x rate=%lld ceil=%lld prio=%lld\n",
-		   classid, parent_classid, rate, ceil, prio);
+		   "TC_HTB_LEAF_ALLOC_QUEUE: classid=0x%x parent_classid=0x%x rate=%lld ceil=%lld prio=%lld quantum=%d\n",
+		   classid, parent_classid, rate, ceil, prio, quantum);
 
 	if (prio > OTX2_QOS_MAX_PRIO) {
 		NL_SET_ERR_MSG_MOD(extack, "Valid priority range 0 to 7");
@@ -978,6 +1219,12 @@ static int otx2_qos_leaf_alloc_queue(struct otx2_nic *pfvf, u16 classid,
 		goto out;
 	}
 
+	if (!quantum || quantum > INT_MAX) {
+		NL_SET_ERR_MSG_MOD(extack, "Invalid quantum, range 1 - 2147483647 bytes");
+		ret = -EOPNOTSUPP;
+		goto out;
+	}
+
 	/* get parent node */
 	parent = otx2_sw_node_find(pfvf, parent_classid);
 	if (!parent) {
@@ -991,10 +1238,24 @@ static int otx2_qos_leaf_alloc_queue(struct otx2_nic *pfvf, u16 classid,
 		goto out;
 	}
 
-	ret = otx2_qos_validate_configuration(parent, extack, pfvf, prio);
+	static_cfg = !is_qos_node_dwrr(parent, pfvf, prio);
+	ret = otx2_qos_validate_configuration(parent, extack, pfvf, prio,
+					      static_cfg);
 	if (ret)
 		goto out;
 
+	if (!static_cfg) {
+		ret = otx2_qos_validate_dwrr_cfg(parent, extack, pfvf, prio,
+						 quantum);
+		if (ret)
+			goto out;
+	}
+
+	if (static_cfg)
+		parent->child_static_cnt++;
+	else
+		parent->child_dwrr_cnt++;
+
 	set_bit(prio, parent->prio_bmap);
 
 	/* read current txschq configuration */
@@ -1019,7 +1280,7 @@ static int otx2_qos_leaf_alloc_queue(struct otx2_nic *pfvf, u16 classid,
 
 	/* allocate and initialize a new child node */
 	node = otx2_qos_sw_create_leaf_node(pfvf, parent, classid, prio, rate,
-					    ceil, qid);
+					    ceil, quantum, qid, static_cfg);
 	if (IS_ERR(node)) {
 		NL_SET_ERR_MSG_MOD(extack, "Unable to allocate leaf node");
 		ret = PTR_ERR(node);
@@ -1067,6 +1328,11 @@ free_node:
 free_old_cfg:
 	kfree(old_cfg);
 reset_prio:
+	if (static_cfg)
+		parent->child_static_cnt--;
+	else
+		parent->child_dwrr_cnt--;
+
 	clear_bit(prio, parent->prio_bmap);
 out:
 	return ret;
@@ -1074,10 +1340,11 @@ out:
 
 static int otx2_qos_leaf_to_inner(struct otx2_nic *pfvf, u16 classid,
 				  u16 child_classid, u64 rate, u64 ceil, u64 prio,
-				  struct netlink_ext_ack *extack)
+				  u32 quantum, struct netlink_ext_ack *extack)
 {
 	struct otx2_qos_cfg *old_cfg, *new_cfg;
 	struct otx2_qos_node *node, *child;
+	bool static_cfg;
 	int ret, err;
 	u16 qid;
 
@@ -1091,6 +1358,12 @@ static int otx2_qos_leaf_to_inner(struct otx2_nic *pfvf, u16 classid,
 		goto out;
 	}
 
+	if (!quantum || quantum > INT_MAX) {
+		NL_SET_ERR_MSG_MOD(extack, "Invalid quantum, range 1 - 2147483647 bytes");
+		ret = -EOPNOTSUPP;
+		goto out;
+	}
+
 	/* find node related to classid */
 	node = otx2_sw_node_find(pfvf, classid);
 	if (!node) {
@@ -1105,6 +1378,19 @@ static int otx2_qos_leaf_to_inner(struct otx2_nic *pfvf, u16 classid,
 		goto out;
 	}
 
+	static_cfg = !is_qos_node_dwrr(node, pfvf, prio);
+	if (!static_cfg) {
+		ret = otx2_qos_validate_dwrr_cfg(node, extack, pfvf, prio,
+						 quantum);
+		if (ret)
+			goto out;
+	}
+
+	if (static_cfg)
+		node->child_static_cnt++;
+	else
+		node->child_dwrr_cnt++;
+
 	set_bit(prio, node->prio_bmap);
 
 	/* store the qid to assign to leaf node */
@@ -1127,7 +1413,8 @@ static int otx2_qos_leaf_to_inner(struct otx2_nic *pfvf, u16 classid,
 
 	/* allocate and initialize a new child node */
 	child = otx2_qos_sw_create_leaf_node(pfvf, node, child_classid,
-					     prio, rate, ceil, qid);
+					     prio, rate, ceil, quantum,
+					     qid, static_cfg);
 	if (IS_ERR(child)) {
 		NL_SET_ERR_MSG_MOD(extack, "Unable to allocate leaf node");
 		ret = PTR_ERR(child);
@@ -1178,6 +1465,10 @@ free_node:
 free_old_cfg:
 	kfree(old_cfg);
 reset_prio:
+	if (static_cfg)
+		node->child_static_cnt--;
+	else
+		node->child_dwrr_cnt--;
 	clear_bit(prio, node->prio_bmap);
 out:
 	return ret;
@@ -1187,6 +1478,7 @@ static int otx2_qos_leaf_del(struct otx2_nic *pfvf, u16 *classid,
 			     struct netlink_ext_ack *extack)
 {
 	struct otx2_qos_node *node, *parent;
+	int dwrr_del_node = false;
 	u64 prio;
 	u16 qid;
 
@@ -1202,12 +1494,27 @@ static int otx2_qos_leaf_del(struct otx2_nic *pfvf, u16 *classid,
 	prio   = node->prio;
 	qid    = node->qid;
 
+	if (!node->is_static)
+		dwrr_del_node = true;
+
 	otx2_qos_disable_sq(pfvf, node->qid);
 
 	otx2_qos_destroy_node(pfvf, node);
 	pfvf->qos.qid_to_sqmap[qid] = OTX2_QOS_INVALID_SQ;
 
-	clear_bit(prio, parent->prio_bmap);
+	if (dwrr_del_node) {
+		parent->child_dwrr_cnt--;
+	} else {
+		parent->child_static_cnt--;
+		clear_bit(prio, parent->prio_bmap);
+	}
+
+	/* Reset DWRR priority if all dwrr nodes are deleted */
+	if (!parent->child_dwrr_cnt)
+		otx2_reset_dwrr_prio(parent, prio);
+
+	if (!parent->child_static_cnt)
+		parent->max_static_prio = 0;
 
 	return 0;
 }
@@ -1217,6 +1524,7 @@ static int otx2_qos_leaf_del_last(struct otx2_nic *pfvf, u16 classid, bool force
 {
 	struct otx2_qos_node *node, *parent;
 	struct otx2_qos_cfg *new_cfg;
+	int dwrr_del_node = false;
 	u64 prio;
 	int err;
 	u16 qid;
@@ -1241,11 +1549,26 @@ static int otx2_qos_leaf_del_last(struct otx2_nic *pfvf, u16 classid, bool force
 		return -ENOENT;
 	}
 
+	if (!node->is_static)
+		dwrr_del_node = true;
+
 	/* destroy the leaf node */
 	otx2_qos_destroy_node(pfvf, node);
 	pfvf->qos.qid_to_sqmap[qid] = OTX2_QOS_INVALID_SQ;
 
-	clear_bit(prio, parent->prio_bmap);
+	if (dwrr_del_node) {
+		parent->child_dwrr_cnt--;
+	} else {
+		parent->child_static_cnt--;
+		clear_bit(prio, parent->prio_bmap);
+	}
+
+	/* Reset DWRR priority if all dwrr nodes are deleted */
+	if (!parent->child_dwrr_cnt)
+		otx2_reset_dwrr_prio(parent, prio);
+
+	if (!parent->child_static_cnt)
+		parent->max_static_prio = 0;
 
 	/* create downstream txschq entries to parent */
 	err = otx2_qos_alloc_txschq_node(pfvf, parent);
@@ -1298,10 +1621,12 @@ void otx2_qos_config_txschq(struct otx2_nic *pfvf)
 	if (!root)
 		return;
 
-	err = otx2_qos_txschq_config(pfvf, root);
-	if (err) {
-		netdev_err(pfvf->netdev, "Error update txschq configuration\n");
-		goto root_destroy;
+	if (root->level != NIX_TXSCH_LVL_TL1) {
+		err = otx2_qos_txschq_config(pfvf, root);
+		if (err) {
+			netdev_err(pfvf->netdev, "Error update txschq configuration\n");
+			goto root_destroy;
+		}
 	}
 
 	err = otx2_qos_txschq_push_cfg_tl(pfvf, root, NULL);
@@ -1334,7 +1659,8 @@ int otx2_setup_tc_htb(struct net_device *ndev, struct tc_htb_qopt_offload *htb)
 		res = otx2_qos_leaf_alloc_queue(pfvf, htb->classid,
 						htb->parent_classid,
 						htb->rate, htb->ceil,
-						htb->prio, htb->extack);
+						htb->prio, htb->quantum,
+						htb->extack);
 		if (res < 0)
 			return res;
 		htb->qid = res;
@@ -1343,7 +1669,7 @@ int otx2_setup_tc_htb(struct net_device *ndev, struct tc_htb_qopt_offload *htb)
 		return otx2_qos_leaf_to_inner(pfvf, htb->parent_classid,
 					      htb->classid, htb->rate,
 					      htb->ceil, htb->prio,
-					      htb->extack);
+					      htb->quantum, htb->extack);
 	case TC_HTB_LEAF_DEL:
 		return otx2_qos_leaf_del(pfvf, &htb->classid, htb->extack);
 	case TC_HTB_LEAF_DEL_LAST:
diff --git a/drivers/net/ethernet/marvell/octeontx2/nic/qos.h b/drivers/net/ethernet/marvell/octeontx2/nic/qos.h
index 19773284be27..221bd0438f60 100644
--- a/drivers/net/ethernet/marvell/octeontx2/nic/qos.h
+++ b/drivers/net/ethernet/marvell/octeontx2/nic/qos.h
@@ -35,6 +35,7 @@ struct otx2_qos_cfg {
 	int dwrr_node_pos[NIX_TXSCH_LVL_CNT];
 	u16 schq_contig_list[NIX_TXSCH_LVL_CNT][MAX_TXSCHQ_PER_FUNC];
 	u16 schq_list[NIX_TXSCH_LVL_CNT][MAX_TXSCHQ_PER_FUNC];
+	bool schq_index_used[NIX_TXSCH_LVL_CNT][MAX_TXSCHQ_PER_FUNC];
 };
 
 struct otx2_qos {
@@ -59,10 +60,18 @@ struct otx2_qos_node {
 	u64 ceil;
 	u32 classid;
 	u32 prio;
-	u16 schq; /* hw txschq */
+	u32 quantum;
+	/* hw txschq */
+	u16 schq;
 	u16 qid;
 	u16 prio_anchor;
+	u16 max_static_prio;
+	u16 child_dwrr_cnt;
+	u16 child_static_cnt;
+	u16 child_dwrr_prio;
+	u16 txschq_idx;			/* txschq allocation index */
 	u8 level;
+	bool is_static;
 };
 
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/qos.c b/drivers/net/ethernet/mellanox/mlx5/core/en/qos.c
index 1874c2f0587f..244bc15a42ab 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/qos.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/qos.c
@@ -379,9 +379,9 @@ int mlx5e_htb_setup_tc(struct mlx5e_priv *priv, struct tc_htb_qopt_offload *htb_
 	if (!htb && htb_qopt->command != TC_HTB_CREATE)
 		return -EINVAL;
 
-	if (htb_qopt->prio) {
+	if (htb_qopt->prio || htb_qopt->quantum) {
 		NL_SET_ERR_MSG_MOD(htb_qopt->extack,
-				   "prio parameter is not supported by device with HTB offload enabled.");
+				   "prio and quantum parameters are not supported by device with HTB offload enabled.");
 		return -EOPNOTSUPP;
 	}
 
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index a2ea45c7b53e..139cd09828af 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -866,6 +866,7 @@ struct tc_htb_qopt_offload {
 	u32 parent_classid;
 	u16 classid;
 	u16 qid;
+	u32 quantum;
 	u64 rate;
 	u64 ceil;
 	u8 prio;
diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c
index 325c29041c7d..333800a7d4eb 100644
--- a/net/sched/sch_htb.c
+++ b/net/sched/sch_htb.c
@@ -1810,10 +1810,6 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 			NL_SET_ERR_MSG(extack, "HTB offload doesn't support the mpu parameter");
 			goto failure;
 		}
-		if (hopt->quantum) {
-			NL_SET_ERR_MSG(extack, "HTB offload doesn't support the quantum parameter");
-			goto failure;
-		}
 	}
 
 	/* Keeping backward compatible with rate_table based iproute2 tc */
@@ -1910,6 +1906,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 				.rate = max_t(u64, hopt->rate.rate, rate64),
 				.ceil = max_t(u64, hopt->ceil.rate, ceil64),
 				.prio = hopt->prio,
+				.quantum = hopt->quantum,
 				.extack = extack,
 			};
 			err = htb_offload(dev, &offload_opt);
@@ -1931,6 +1928,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 				.rate = max_t(u64, hopt->rate.rate, rate64),
 				.ceil = max_t(u64, hopt->ceil.rate, ceil64),
 				.prio = hopt->prio,
+				.quantum = hopt->quantum,
 				.extack = extack,
 			};
 			err = htb_offload(dev, &offload_opt);
@@ -2017,6 +2015,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 				.rate = max_t(u64, hopt->rate.rate, rate64),
 				.ceil = max_t(u64, hopt->ceil.rate, ceil64),
 				.prio = hopt->prio,
+				.quantum = hopt->quantum,
 				.extack = extack,
 			};
 			err = htb_offload(dev, &offload_opt);