1 files changed, 219 insertions, 268 deletions

diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c
index c4075610502c..d8fd35da32a7 100644
--- a/net/sched/sch_hfsc.c
+++ b/net/sched/sch_hfsc.c
@@ -110,14 +110,13 @@ enum hfsc_class_flags {
 
 struct hfsc_class {
 	struct Qdisc_class_common cl_common;
-	unsigned int	refcnt;		/* usage count */
 
-	struct gnet_stats_basic_packed bstats;
+	struct gnet_stats_basic_sync bstats;
 	struct gnet_stats_queue qstats;
-	struct gnet_stats_rate_est64 rate_est;
+	struct net_rate_estimator __rcu *rate_est;
+	struct tcf_proto __rcu *filter_list; /* filter list */
+	struct tcf_block *block;
 	unsigned int	level;		/* class level in hierarchy */
-	struct tcf_proto *filter_list;	/* filter list */
-	unsigned int	filter_cnt;	/* filter count */
 
 	struct hfsc_sched *sched;	/* scheduler data */
 	struct hfsc_class *cl_parent;	/* parent class */
@@ -130,7 +129,6 @@ struct hfsc_class {
 	struct rb_node vt_node;		/* parent's vt_tree member */
 	struct rb_root cf_tree;		/* active children sorted by cl_f */
 	struct rb_node cf_node;		/* parent's cf_heap member */
-	struct list_head dlist;		/* drop list member */
 
 	u64	cl_total;		/* total work in bytes */
 	u64	cl_cumul;		/* cumulative work in bytes done by
@@ -143,8 +141,6 @@ struct hfsc_class {
 					   link-sharing, max(myf, cfmin) */
 	u64	cl_myf;			/* my fit-time (calculated from this
 					   class's own upperlimit curve) */
-	u64	cl_myfadj;		/* my fit-time adjustment (to cancel
-					   history dependence) */
 	u64	cl_cfmin;		/* earliest children's fit-time (used
 					   with cl_myf to obtain cl_f) */
 	u64	cl_cvtmin;		/* minimal virtual time among the
@@ -152,11 +148,8 @@ struct hfsc_class {
 					   (monotonic within a period) */
 	u64	cl_vtadj;		/* intra-period cumulative vt
 					   adjustment */
-	u64	cl_vtoff;		/* inter-period cumulative vt offset */
-	u64	cl_cvtmax;		/* max child's vt in the last period */
-	u64	cl_cvtoff;		/* cumulative cvtmax of all periods */
-	u64	cl_pcvtoff;		/* parent's cvtoff at initialization
-					   time */
+	u64	cl_cvtoff;		/* largest virtual time seen among
+					   the children */
 
 	struct internal_sc cl_rsc;	/* internal real-time service curve */
 	struct internal_sc cl_fsc;	/* internal fair service curve */
@@ -166,10 +159,10 @@ struct hfsc_class {
 	struct runtime_sc cl_virtual;	/* virtual curve */
 	struct runtime_sc cl_ulimit;	/* upperlimit curve */
 
-	unsigned long	cl_flags;	/* which curves are valid */
-	unsigned long	cl_vtperiod;	/* vt period sequence number */
-	unsigned long	cl_parentperiod;/* parent's vt period sequence number*/
-	unsigned long	cl_nactive;	/* number of active children */
+	u8		cl_flags;	/* which curves are valid */
+	u32		cl_vtperiod;	/* vt period sequence number */
+	u32		cl_parentperiod;/* parent's vt period sequence number*/
+	u32		cl_nactive;	/* number of active children */
 };
 
 struct hfsc_sched {
@@ -177,13 +170,16 @@ struct hfsc_sched {
 	struct hfsc_class root;			/* root class */
 	struct Qdisc_class_hash clhash;		/* class hash */
 	struct rb_root eligible;		/* eligible tree */
-	struct list_head droplist;		/* active leaf class list (for
-						   dropping) */
 	struct qdisc_watchdog watchdog;		/* watchdog timer */
 };
 
 #define	HT_INFINITY	0xffffffffffffffffULL	/* infinite time value */
 
+static bool cl_in_el_or_vttree(struct hfsc_class *cl)
+{
+	return ((cl->cl_flags & HFSC_FSC) && cl->cl_nactive) ||
+		((cl->cl_flags & HFSC_RSC) && !RB_EMPTY_NODE(&cl->el_node));
+}
 
 /*
  * eligible tree holds backlogged classes being sorted by their eligible times.
@@ -212,7 +208,10 @@ eltree_insert(struct hfsc_class *cl)
 static inline void
 eltree_remove(struct hfsc_class *cl)
 {
-	rb_erase(&cl->el_node, &cl->sched->eligible);
+	if (!RB_EMPTY_NODE(&cl->el_node)) {
+		rb_erase(&cl->el_node, &cl->sched->eligible);
+		RB_CLEAR_NODE(&cl->el_node);
+	}
 }
 
 static inline void
@@ -704,28 +703,16 @@ init_vf(struct hfsc_class *cl, unsigned int len)
 			} else {
 				/*
 				 * first child for a new parent backlog period.
-				 * add parent's cvtmax to cvtoff to make a new
-				 * vt (vtoff + vt) larger than the vt in the
-				 * last period for all children.
+				 * initialize cl_vt to the highest value seen
+				 * among the siblings. this is analogous to
+				 * what cur_time would provide in realtime case.
 				 */
-				vt = cl->cl_parent->cl_cvtmax;
-				cl->cl_parent->cl_cvtoff += vt;
-				cl->cl_parent->cl_cvtmax = 0;
+				cl->cl_vt = cl->cl_parent->cl_cvtoff;
 				cl->cl_parent->cl_cvtmin = 0;
-				cl->cl_vt = 0;
 			}
 
-			cl->cl_vtoff = cl->cl_parent->cl_cvtoff -
-							cl->cl_pcvtoff;
-
 			/* update the virtual curve */
-			vt = cl->cl_vt + cl->cl_vtoff;
-			rtsc_min(&cl->cl_virtual, &cl->cl_fsc, vt,
-						      cl->cl_total);
-			if (cl->cl_virtual.x == vt) {
-				cl->cl_virtual.x -= cl->cl_vtoff;
-				cl->cl_vtoff = 0;
-			}
+			rtsc_min(&cl->cl_virtual, &cl->cl_fsc, cl->cl_vt, cl->cl_total);
 			cl->cl_vtadj = 0;
 
 			cl->cl_vtperiod++;  /* increment vt period */
@@ -748,7 +735,6 @@ init_vf(struct hfsc_class *cl, unsigned int len)
 				/* compute myf */
 				cl->cl_myf = rtsc_y2x(&cl->cl_ulimit,
 						      cl->cl_total);
-				cl->cl_myfadj = 0;
 			}
 		}
 
@@ -781,12 +767,25 @@ update_vf(struct hfsc_class *cl, unsigned int len, u64 cur_time)
 		else
 			go_passive = 0;
 
+		/* update vt */
+		cl->cl_vt = rtsc_y2x(&cl->cl_virtual, cl->cl_total) + cl->cl_vtadj;
+
+		/*
+		 * if vt of the class is smaller than cvtmin,
+		 * the class was skipped in the past due to non-fit.
+		 * if so, we need to adjust vtadj.
+		 */
+		if (cl->cl_vt < cl->cl_parent->cl_cvtmin) {
+			cl->cl_vtadj += cl->cl_parent->cl_cvtmin - cl->cl_vt;
+			cl->cl_vt = cl->cl_parent->cl_cvtmin;
+		}
+
 		if (go_passive) {
 			/* no more active child, going passive */
 
-			/* update cvtmax of the parent class */
-			if (cl->cl_vt > cl->cl_parent->cl_cvtmax)
-				cl->cl_parent->cl_cvtmax = cl->cl_vt;
+			/* update cvtoff of the parent class */
+			if (cl->cl_vt > cl->cl_parent->cl_cvtoff)
+				cl->cl_parent->cl_cvtoff = cl->cl_vt;
 
 			/* remove this class from the vt tree */
 			vttree_remove(cl);
@@ -797,29 +796,15 @@ update_vf(struct hfsc_class *cl, unsigned int len, u64 cur_time)
 			continue;
 		}
 
-		/*
-		 * update vt and f
-		 */
-		cl->cl_vt = rtsc_y2x(&cl->cl_virtual, cl->cl_total)
-			    - cl->cl_vtoff + cl->cl_vtadj;
-
-		/*
-		 * if vt of the class is smaller than cvtmin,
-		 * the class was skipped in the past due to non-fit.
-		 * if so, we need to adjust vtadj.
-		 */
-		if (cl->cl_vt < cl->cl_parent->cl_cvtmin) {
-			cl->cl_vtadj += cl->cl_parent->cl_cvtmin - cl->cl_vt;
-			cl->cl_vt = cl->cl_parent->cl_cvtmin;
-		}
-
 		/* update the vt tree */
 		vttree_update(cl);
 
+		/* update f */
 		if (cl->cl_flags & HFSC_USC) {
+			cl->cl_myf = rtsc_y2x(&cl->cl_ulimit, cl->cl_total);
+#if 0
 			cl->cl_myf = cl->cl_myfadj + rtsc_y2x(&cl->cl_ulimit,
 							      cl->cl_total);
-#if 0
 			/*
 			 * This code causes classes to stay way under their
 			 * limit when multiple classes are used at gigabit
@@ -851,56 +836,6 @@ update_vf(struct hfsc_class *cl, unsigned int len, u64 cur_time)
 }
 
 static void
-set_active(struct hfsc_class *cl, unsigned int len)
-{
-	if (cl->cl_flags & HFSC_RSC)
-		init_ed(cl, len);
-	if (cl->cl_flags & HFSC_FSC)
-		init_vf(cl, len);
-
-	list_add_tail(&cl->dlist, &cl->sched->droplist);
-}
-
-static void
-set_passive(struct hfsc_class *cl)
-{
-	if (cl->cl_flags & HFSC_RSC)
-		eltree_remove(cl);
-
-	list_del(&cl->dlist);
-
-	/*
-	 * vttree is now handled in update_vf() so that update_vf(cl, 0, 0)
-	 * needs to be called explicitly to remove a class from vttree.
-	 */
-}
-
-static unsigned int
-qdisc_peek_len(struct Qdisc *sch)
-{
-	struct sk_buff *skb;
-	unsigned int len;
-
-	skb = sch->ops->peek(sch);
-	if (skb == NULL) {
-		qdisc_warn_nonwc("qdisc_peek_len", sch);
-		return 0;
-	}
-	len = qdisc_pkt_len(skb);
-
-	return len;
-}
-
-static void
-hfsc_purge_queue(struct Qdisc *sch, struct hfsc_class *cl)
-{
-	unsigned int len = cl->qdisc->q.qlen;
-
-	qdisc_reset(cl->qdisc);
-	qdisc_tree_decrease_qlen(cl->qdisc, len);
-}
-
-static void
 hfsc_adjust_levels(struct hfsc_class *cl)
 {
 	struct hfsc_class *p;
@@ -959,6 +894,14 @@ hfsc_change_usc(struct hfsc_class *cl, struct tc_service_curve *usc,
 	cl->cl_flags |= HFSC_USC;
 }
 
+static void
+hfsc_upgrade_rt(struct hfsc_class *cl)
+{
+	cl->cl_fsc = cl->cl_rsc;
+	rtsc_init(&cl->cl_virtual, &cl->cl_fsc, cl->cl_vt, cl->cl_total);
+	cl->cl_flags |= HFSC_FSC;
+}
+
 static const struct nla_policy hfsc_policy[TCA_HFSC_MAX + 1] = {
 	[TCA_HFSC_RSC]	= { .len = sizeof(struct tc_service_curve) },
 	[TCA_HFSC_FSC]	= { .len = sizeof(struct tc_service_curve) },
@@ -967,7 +910,8 @@ static const struct nla_policy hfsc_policy[TCA_HFSC_MAX + 1] = {
 
 static int
 hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
-		  struct nlattr **tca, unsigned long *arg)
+		  struct nlattr **tca, unsigned long *arg,
+		  struct netlink_ext_ack *extack)
 {
 	struct hfsc_sched *q = qdisc_priv(sch);
 	struct hfsc_class *cl = (struct hfsc_class *)*arg;
@@ -981,7 +925,8 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 	if (opt == NULL)
 		return -EINVAL;
 
-	err = nla_parse_nested(tb, TCA_HFSC_MAX, opt, hfsc_policy);
+	err = nla_parse_nested_deprecated(tb, TCA_HFSC_MAX, opt, hfsc_policy,
+					  NULL);
 	if (err < 0)
 		return err;
 
@@ -1004,6 +949,9 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 	}
 
 	if (cl != NULL) {
+		int old_flags;
+		int len = 0;
+
 		if (parentid) {
 			if (cl->cl_parent &&
 			    cl->cl_parent->cl_common.classid != parentid)
@@ -1014,14 +962,18 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 		cur_time = psched_get_time();
 
 		if (tca[TCA_RATE]) {
-			err = gen_replace_estimator(&cl->bstats, &cl->rate_est,
-					      qdisc_root_sleeping_lock(sch),
-					      tca[TCA_RATE]);
+			err = gen_replace_estimator(&cl->bstats, NULL,
+						    &cl->rate_est,
+						    NULL,
+						    true,
+						    tca[TCA_RATE]);
 			if (err)
 				return err;
 		}
 
 		sch_tree_lock(sch);
+		old_flags = cl->cl_flags;
+
 		if (rsc != NULL)
 			hfsc_change_rsc(cl, rsc, cur_time);
 		if (fsc != NULL)
@@ -1029,11 +981,26 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 		if (usc != NULL)
 			hfsc_change_usc(cl, usc, cur_time);
 
+		if (cl->qdisc->q.qlen != 0)
+			len = qdisc_peek_len(cl->qdisc);
+		/* Check queue length again since some qdisc implementations
+		 * (e.g., netem/codel) might empty the queue during the peek
+		 * operation.
+		 */
 		if (cl->qdisc->q.qlen != 0) {
-			if (cl->cl_flags & HFSC_RSC)
-				update_ed(cl, qdisc_peek_len(cl->qdisc));
-			if (cl->cl_flags & HFSC_FSC)
-				update_vf(cl, 0, cur_time);
+			if (cl->cl_flags & HFSC_RSC) {
+				if (old_flags & HFSC_RSC)
+					update_ed(cl, len);
+				else
+					init_ed(cl, len);
+			}
+
+			if (cl->cl_flags & HFSC_FSC) {
+				if (old_flags & HFSC_FSC)
+					update_vf(cl, 0, cur_time);
+				else
+					init_vf(cl, len);
+			}
 		}
 		sch_tree_unlock(sch);
 
@@ -1062,11 +1029,19 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 	if (cl == NULL)
 		return -ENOBUFS;
 
+	RB_CLEAR_NODE(&cl->el_node);
+
+	err = tcf_block_get(&cl->block, &cl->filter_list, sch, extack);
+	if (err) {
+		kfree(cl);
+		return err;
+	}
+
 	if (tca[TCA_RATE]) {
-		err = gen_new_estimator(&cl->bstats, &cl->rate_est,
-					qdisc_root_sleeping_lock(sch),
-					tca[TCA_RATE]);
+		err = gen_new_estimator(&cl->bstats, NULL, &cl->rate_est,
+					NULL, true, tca[TCA_RATE]);
 		if (err) {
+			tcf_block_put(cl->block);
 			kfree(cl);
 			return err;
 		}
@@ -1080,24 +1055,30 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 		hfsc_change_usc(cl, usc, 0);
 
 	cl->cl_common.classid = classid;
-	cl->refcnt    = 1;
 	cl->sched     = q;
 	cl->cl_parent = parent;
-	cl->qdisc = qdisc_create_dflt(sch->dev_queue,
-				      &pfifo_qdisc_ops, classid);
+	cl->qdisc = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops,
+				      classid, NULL);
 	if (cl->qdisc == NULL)
 		cl->qdisc = &noop_qdisc;
+	else
+		qdisc_hash_add(cl->qdisc, true);
 	INIT_LIST_HEAD(&cl->children);
 	cl->vt_tree = RB_ROOT;
 	cl->cf_tree = RB_ROOT;
 
 	sch_tree_lock(sch);
+	/* Check if the inner class is a misconfigured 'rt' */
+	if (!(parent->cl_flags & HFSC_FSC) && parent != &q->root) {
+		NL_SET_ERR_MSG(extack,
+			       "Forced curve change on parent 'rt' to 'sc'");
+		hfsc_upgrade_rt(parent);
+	}
 	qdisc_class_hash_insert(&q->clhash, &cl->cl_common);
 	list_add_tail(&cl->siblings, &parent->children);
 	if (parent->level == 0)
-		hfsc_purge_queue(sch, parent);
+		qdisc_purge_queue(parent->qdisc);
 	hfsc_adjust_levels(parent);
-	cl->cl_pcvtoff = parent->cl_cvtoff;
 	sch_tree_unlock(sch);
 
 	qdisc_class_hash_grow(sch, &q->clhash);
@@ -1111,37 +1092,37 @@ hfsc_destroy_class(struct Qdisc *sch, struct hfsc_class *cl)
 {
 	struct hfsc_sched *q = qdisc_priv(sch);
 
-	tcf_destroy_chain(&cl->filter_list);
-	qdisc_destroy(cl->qdisc);
-	gen_kill_estimator(&cl->bstats, &cl->rate_est);
+	tcf_block_put(cl->block);
+	qdisc_put(cl->qdisc);
+	gen_kill_estimator(&cl->rate_est);
 	if (cl != &q->root)
 		kfree(cl);
 }
 
 static int
-hfsc_delete_class(struct Qdisc *sch, unsigned long arg)
+hfsc_delete_class(struct Qdisc *sch, unsigned long arg,
+		  struct netlink_ext_ack *extack)
 {
 	struct hfsc_sched *q = qdisc_priv(sch);
 	struct hfsc_class *cl = (struct hfsc_class *)arg;
 
-	if (cl->level > 0 || cl->filter_cnt > 0 || cl == &q->root)
+	if (cl->level > 0 || qdisc_class_in_use(&cl->cl_common) ||
+	    cl == &q->root) {
+		NL_SET_ERR_MSG(extack, "HFSC class in use");
 		return -EBUSY;
+	}
 
 	sch_tree_lock(sch);
 
 	list_del(&cl->siblings);
 	hfsc_adjust_levels(cl->cl_parent);
 
-	hfsc_purge_queue(sch, cl);
+	qdisc_purge_queue(cl->qdisc);
 	qdisc_class_hash_remove(&q->clhash, &cl->cl_common);
 
-	BUG_ON(--cl->refcnt == 0);
-	/*
-	 * This shouldn't happen: we "hold" one cops->get() when called
-	 * from tc_ctl_tclass; the destroy method is done from cops->put().
-	 */
-
 	sch_tree_unlock(sch);
+
+	hfsc_destroy_class(sch, cl);
 	return 0;
 }
 
@@ -1161,13 +1142,15 @@ hfsc_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr)
 
 	*qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
 	head = &q->root;
-	tcf = q->root.filter_list;
-	while (tcf && (result = tc_classify(skb, tcf, &res)) >= 0) {
+	tcf = rcu_dereference_bh(q->root.filter_list);
+	while (tcf && (result = tcf_classify(skb, NULL, tcf, &res, false)) >= 0) {
 #ifdef CONFIG_NET_CLS_ACT
 		switch (result) {
 		case TC_ACT_QUEUED:
 		case TC_ACT_STOLEN:
+		case TC_ACT_TRAP:
 			*qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN;
+			fallthrough;
 		case TC_ACT_SHOT:
 			return NULL;
 		}
@@ -1185,12 +1168,13 @@ hfsc_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr)
 			return cl; /* hit leaf class */
 
 		/* apply inner filter chain */
-		tcf = cl->filter_list;
+		tcf = rcu_dereference_bh(cl->filter_list);
 		head = cl;
 	}
 
 	/* classification failed, try default class */
-	cl = hfsc_find_class(TC_H_MAKE(TC_H_MAJ(sch->handle), q->defcls), sch);
+	cl = hfsc_find_class(TC_H_MAKE(TC_H_MAJ(sch->handle),
+				       READ_ONCE(q->defcls)), sch);
 	if (cl == NULL || cl->level > 0)
 		return NULL;
 
@@ -1199,7 +1183,7 @@ hfsc_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr)
 
 static int
 hfsc_graft_class(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
-		 struct Qdisc **old)
+		 struct Qdisc **old, struct netlink_ext_ack *extack)
 {
 	struct hfsc_class *cl = (struct hfsc_class *)arg;
 
@@ -1207,16 +1191,12 @@ hfsc_graft_class(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
 		return -EINVAL;
 	if (new == NULL) {
 		new = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops,
-					cl->cl_common.classid);
+					cl->cl_common.classid, NULL);
 		if (new == NULL)
 			new = &noop_qdisc;
 	}
 
-	sch_tree_lock(sch);
-	hfsc_purge_queue(sch, cl);
-	*old = cl->qdisc;
-	cl->qdisc = new;
-	sch_tree_unlock(sch);
+	*old = qdisc_replace(sch, new, &cl->qdisc);
 	return 0;
 }
 
@@ -1236,30 +1216,19 @@ hfsc_qlen_notify(struct Qdisc *sch, unsigned long arg)
 {
 	struct hfsc_class *cl = (struct hfsc_class *)arg;
 
-	if (cl->qdisc->q.qlen == 0) {
+	/* vttree is now handled in update_vf() so that update_vf(cl, 0, 0)
+	 * needs to be called explicitly to remove a class from vttree.
+	 */
+	if (cl->cl_nactive)
 		update_vf(cl, 0, 0);
-		set_passive(cl);
-	}
+	if (cl->cl_flags & HFSC_RSC)
+		eltree_remove(cl);
 }
 
 static unsigned long
-hfsc_get_class(struct Qdisc *sch, u32 classid)
-{
-	struct hfsc_class *cl = hfsc_find_class(classid, sch);
-
-	if (cl != NULL)
-		cl->refcnt++;
-
-	return (unsigned long)cl;
-}
-
-static void
-hfsc_put_class(struct Qdisc *sch, unsigned long arg)
+hfsc_search_class(struct Qdisc *sch, u32 classid)
 {
-	struct hfsc_class *cl = (struct hfsc_class *)arg;
-
-	if (--cl->refcnt == 0)
-		hfsc_destroy_class(sch, cl);
+	return (unsigned long)hfsc_find_class(classid, sch);
 }
 
 static unsigned long
@@ -1271,7 +1240,7 @@ hfsc_bind_tcf(struct Qdisc *sch, unsigned long parent, u32 classid)
 	if (cl != NULL) {
 		if (p != NULL && p->level <= cl->level)
 			return 0;
-		cl->filter_cnt++;
+		qdisc_class_get(&cl->cl_common);
 	}
 
 	return (unsigned long)cl;
@@ -1282,11 +1251,11 @@ hfsc_unbind_tcf(struct Qdisc *sch, unsigned long arg)
 {
 	struct hfsc_class *cl = (struct hfsc_class *)arg;
 
-	cl->filter_cnt--;
+	qdisc_class_put(&cl->cl_common);
 }
 
-static struct tcf_proto **
-hfsc_tcf_chain(struct Qdisc *sch, unsigned long arg)
+static struct tcf_block *hfsc_tcf_block(struct Qdisc *sch, unsigned long arg,
+					struct netlink_ext_ack *extack)
 {
 	struct hfsc_sched *q = qdisc_priv(sch);
 	struct hfsc_class *cl = (struct hfsc_class *)arg;
@@ -1294,7 +1263,7 @@ hfsc_tcf_chain(struct Qdisc *sch, unsigned long arg)
 	if (cl == NULL)
 		cl = &q->root;
 
-	return &cl->filter_list;
+	return cl->block;
 }
 
 static int
@@ -1348,13 +1317,12 @@ hfsc_dump_class(struct Qdisc *sch, unsigned long arg, struct sk_buff *skb,
 	if (cl->level == 0)
 		tcm->tcm_info = cl->qdisc->handle;
 
-	nest = nla_nest_start(skb, TCA_OPTIONS);
+	nest = nla_nest_start_noflag(skb, TCA_OPTIONS);
 	if (nest == NULL)
 		goto nla_put_failure;
 	if (hfsc_dump_curves(skb, cl) < 0)
 		goto nla_put_failure;
-	nla_nest_end(skb, nest);
-	return skb->len;
+	return nla_nest_end(skb, nest);
 
  nla_put_failure:
 	nla_nest_cancel(skb, nest);
@@ -1367,17 +1335,17 @@ hfsc_dump_class_stats(struct Qdisc *sch, unsigned long arg,
 {
 	struct hfsc_class *cl = (struct hfsc_class *)arg;
 	struct tc_hfsc_stats xstats;
+	__u32 qlen;
 
-	cl->qstats.qlen = cl->qdisc->q.qlen;
-	cl->qstats.backlog = cl->qdisc->qstats.backlog;
+	qdisc_qstats_qlen_backlog(cl->qdisc, &qlen, &cl->qstats.backlog);
 	xstats.level   = cl->level;
 	xstats.period  = cl->cl_vtperiod;
 	xstats.work    = cl->cl_total;
 	xstats.rtwork  = cl->cl_cumul;
 
-	if (gnet_stats_copy_basic(d, &cl->bstats) < 0 ||
-	    gnet_stats_copy_rate_est(d, &cl->bstats, &cl->rate_est) < 0 ||
-	    gnet_stats_copy_queue(d, &cl->qstats) < 0)
+	if (gnet_stats_copy_basic(d, NULL, &cl->bstats, true) < 0 ||
+	    gnet_stats_copy_rate_est(d, &cl->rate_est) < 0 ||
+	    gnet_stats_copy_queue(d, NULL, &cl->qstats, qlen) < 0)
 		return -1;
 
 	return gnet_stats_copy_app(d, &xstats, sizeof(xstats));
@@ -1398,15 +1366,8 @@ hfsc_walk(struct Qdisc *sch, struct qdisc_walker *arg)
 	for (i = 0; i < q->clhash.hashsize; i++) {
 		hlist_for_each_entry(cl, &q->clhash.hash[i],
 				     cl_common.hnode) {
-			if (arg->count < arg->skip) {
-				arg->count++;
-				continue;
-			}
-			if (arg->fn(sch, (unsigned long)cl, arg) < 0) {
-				arg->stop = 1;
+			if (!tc_qdisc_stats_dump(sch, (unsigned long)cl, arg))
 				return;
-			}
-			arg->count++;
 		}
 	}
 }
@@ -1425,18 +1386,21 @@ hfsc_schedule_watchdog(struct Qdisc *sch)
 		if (next_time == 0 || next_time > q->root.cl_cfmin)
 			next_time = q->root.cl_cfmin;
 	}
-	WARN_ON(next_time == 0);
-	qdisc_watchdog_schedule(&q->watchdog, next_time);
+	if (next_time)
+		qdisc_watchdog_schedule(&q->watchdog, next_time);
 }
 
 static int
-hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt)
+hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt,
+		struct netlink_ext_ack *extack)
 {
 	struct hfsc_sched *q = qdisc_priv(sch);
 	struct tc_hfsc_qopt *qopt;
 	int err;
 
-	if (opt == NULL || nla_len(opt) < sizeof(*qopt))
+	qdisc_watchdog_init(&q->watchdog, sch);
+
+	if (!opt || nla_len(opt) < sizeof(*qopt))
 		return -EINVAL;
 	qopt = nla_data(opt);
 
@@ -1445,15 +1409,20 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt)
 	if (err < 0)
 		return err;
 	q->eligible = RB_ROOT;
-	INIT_LIST_HEAD(&q->droplist);
 
+	err = tcf_block_get(&q->root.block, &q->root.filter_list, sch, extack);
+	if (err)
+		return err;
+
+	gnet_stats_basic_sync_init(&q->root.bstats);
 	q->root.cl_common.classid = sch->handle;
-	q->root.refcnt  = 1;
 	q->root.sched   = q;
 	q->root.qdisc = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops,
-					  sch->handle);
+					  sch->handle, NULL);
 	if (q->root.qdisc == NULL)
 		q->root.qdisc = &noop_qdisc;
+	else
+		qdisc_hash_add(q->root.qdisc, true);
 	INIT_LIST_HEAD(&q->root.children);
 	q->root.vt_tree = RB_ROOT;
 	q->root.cf_tree = RB_ROOT;
@@ -1461,24 +1430,21 @@ hfsc_init_qdisc(struct Qdisc *sch, struct nlattr *opt)
 	qdisc_class_hash_insert(&q->clhash, &q->root.cl_common);
 	qdisc_class_hash_grow(sch, &q->clhash);
 
-	qdisc_watchdog_init(&q->watchdog, sch);
-
 	return 0;
 }
 
 static int
-hfsc_change_qdisc(struct Qdisc *sch, struct nlattr *opt)
+hfsc_change_qdisc(struct Qdisc *sch, struct nlattr *opt,
+		  struct netlink_ext_ack *extack)
 {
 	struct hfsc_sched *q = qdisc_priv(sch);
 	struct tc_hfsc_qopt *qopt;
 
-	if (opt == NULL || nla_len(opt) < sizeof(*qopt))
+	if (nla_len(opt) < sizeof(*qopt))
 		return -EINVAL;
 	qopt = nla_data(opt);
 
-	sch_tree_lock(sch);
-	q->defcls = qopt->defcls;
-	sch_tree_unlock(sch);
+	WRITE_ONCE(q->defcls, qopt->defcls);
 
 	return 0;
 }
@@ -1492,16 +1458,12 @@ hfsc_reset_class(struct hfsc_class *cl)
 	cl->cl_e            = 0;
 	cl->cl_vt           = 0;
 	cl->cl_vtadj        = 0;
-	cl->cl_vtoff        = 0;
 	cl->cl_cvtmin       = 0;
-	cl->cl_cvtmax       = 0;
 	cl->cl_cvtoff       = 0;
-	cl->cl_pcvtoff      = 0;
 	cl->cl_vtperiod     = 0;
 	cl->cl_parentperiod = 0;
 	cl->cl_f            = 0;
 	cl->cl_myf          = 0;
-	cl->cl_myfadj       = 0;
 	cl->cl_cfmin        = 0;
 	cl->cl_nactive      = 0;
 
@@ -1529,9 +1491,7 @@ hfsc_reset_qdisc(struct Qdisc *sch)
 			hfsc_reset_class(cl);
 	}
 	q->eligible = RB_ROOT;
-	INIT_LIST_HEAD(&q->droplist);
 	qdisc_watchdog_cancel(&q->watchdog);
-	sch->q.qlen = 0;
 }
 
 static void
@@ -1543,8 +1503,10 @@ hfsc_destroy_qdisc(struct Qdisc *sch)
 	unsigned int i;
 
 	for (i = 0; i < q->clhash.hashsize; i++) {
-		hlist_for_each_entry(cl, &q->clhash.hash[i], cl_common.hnode)
-			tcf_destroy_chain(&cl->filter_list);
+		hlist_for_each_entry(cl, &q->clhash.hash[i], cl_common.hnode) {
+			tcf_block_put(cl->block);
+			cl->block = NULL;
+		}
 	}
 	for (i = 0; i < q->clhash.hashsize; i++) {
 		hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i],
@@ -1561,16 +1523,8 @@ hfsc_dump_qdisc(struct Qdisc *sch, struct sk_buff *skb)
 	struct hfsc_sched *q = qdisc_priv(sch);
 	unsigned char *b = skb_tail_pointer(skb);
 	struct tc_hfsc_qopt qopt;
-	struct hfsc_class *cl;
-	unsigned int i;
-
-	sch->qstats.backlog = 0;
-	for (i = 0; i < q->clhash.hashsize; i++) {
-		hlist_for_each_entry(cl, &q->clhash.hash[i], cl_common.hnode)
-			sch->qstats.backlog += cl->qdisc->qstats.backlog;
-	}
 
-	qopt.defcls = q->defcls;
+	qopt.defcls = READ_ONCE(q->defcls);
 	if (nla_put(skb, TCA_OPTIONS, sizeof(qopt), &qopt))
 		goto nla_put_failure;
 	return skb->len;
@@ -1581,33 +1535,49 @@ hfsc_dump_qdisc(struct Qdisc *sch, struct sk_buff *skb)
 }
 
 static int
-hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff **to_free)
 {
+	unsigned int len = qdisc_pkt_len(skb);
 	struct hfsc_class *cl;
-	int uninitialized_var(err);
+	int err;
+	bool first;
 
 	cl = hfsc_classify(skb, sch, &err);
 	if (cl == NULL) {
 		if (err & __NET_XMIT_BYPASS)
-			sch->qstats.drops++;
-		kfree_skb(skb);
+			qdisc_qstats_drop(sch);
+		__qdisc_drop(skb, to_free);
 		return err;
 	}
 
-	err = qdisc_enqueue(skb, cl->qdisc);
+	first = !cl->qdisc->q.qlen;
+	err = qdisc_enqueue(skb, cl->qdisc, to_free);
 	if (unlikely(err != NET_XMIT_SUCCESS)) {
 		if (net_xmit_drop_count(err)) {
 			cl->qstats.drops++;
-			sch->qstats.drops++;
+			qdisc_qstats_drop(sch);
 		}
 		return err;
 	}
 
-	if (cl->qdisc->q.qlen == 1)
-		set_active(cl, qdisc_pkt_len(skb));
-
+	sch->qstats.backlog += len;
 	sch->q.qlen++;
 
+	if (first && !cl_in_el_or_vttree(cl)) {
+		if (cl->cl_flags & HFSC_RSC)
+			init_ed(cl, len);
+		if (cl->cl_flags & HFSC_FSC)
+			init_vf(cl, len);
+		/*
+		 * If this is the first packet, isolate the head so an eventual
+		 * head drop before the first dequeue operation has no chance
+		 * to invalidate the deadline.
+		 */
+		if (cl->cl_flags & HFSC_RSC)
+			cl->qdisc->ops->peek(cl->qdisc);
+
+	}
+
 	return NET_XMIT_SUCCESS;
 }
 
@@ -1641,7 +1611,7 @@ hfsc_dequeue(struct Qdisc *sch)
 		 */
 		cl = vttree_get_minvt(&q->root, cur_time);
 		if (cl == NULL) {
-			sch->qstats.overlimits++;
+			qdisc_qstats_overlimit(sch);
 			hfsc_schedule_watchdog(sch);
 			return NULL;
 		}
@@ -1658,63 +1628,43 @@ hfsc_dequeue(struct Qdisc *sch)
 	if (realtime)
 		cl->cl_cumul += qdisc_pkt_len(skb);
 
-	if (cl->qdisc->q.qlen != 0) {
-		if (cl->cl_flags & HFSC_RSC) {
+	if (cl->cl_flags & HFSC_RSC) {
+		if (cl->qdisc->q.qlen != 0) {
 			/* update ed */
 			next_len = qdisc_peek_len(cl->qdisc);
-			if (realtime)
-				update_ed(cl, next_len);
-			else
-				update_d(cl, next_len);
+			/* Check queue length again since some qdisc implementations
+			 * (e.g., netem/codel) might empty the queue during the peek
+			 * operation.
+			 */
+			if (cl->qdisc->q.qlen != 0) {
+				if (realtime)
+					update_ed(cl, next_len);
+				else
+					update_d(cl, next_len);
+			}
+		} else {
+			/* the class becomes passive */
+			eltree_remove(cl);
 		}
-	} else {
-		/* the class becomes passive */
-		set_passive(cl);
 	}
 
-	qdisc_unthrottled(sch);
 	qdisc_bstats_update(sch, skb);
+	qdisc_qstats_backlog_dec(sch, skb);
 	sch->q.qlen--;
 
 	return skb;
 }
 
-static unsigned int
-hfsc_drop(struct Qdisc *sch)
-{
-	struct hfsc_sched *q = qdisc_priv(sch);
-	struct hfsc_class *cl;
-	unsigned int len;
-
-	list_for_each_entry(cl, &q->droplist, dlist) {
-		if (cl->qdisc->ops->drop != NULL &&
-		    (len = cl->qdisc->ops->drop(cl->qdisc)) > 0) {
-			if (cl->qdisc->q.qlen == 0) {
-				update_vf(cl, 0, 0);
-				set_passive(cl);
-			} else {
-				list_move_tail(&cl->dlist, &q->droplist);
-			}
-			cl->qstats.drops++;
-			sch->qstats.drops++;
-			sch->q.qlen--;
-			return len;
-		}
-	}
-	return 0;
-}
-
 static const struct Qdisc_class_ops hfsc_class_ops = {
 	.change		= hfsc_change_class,
 	.delete		= hfsc_delete_class,
 	.graft		= hfsc_graft_class,
 	.leaf		= hfsc_class_leaf,
 	.qlen_notify	= hfsc_qlen_notify,
-	.get		= hfsc_get_class,
-	.put		= hfsc_put_class,
+	.find		= hfsc_search_class,
 	.bind_tcf	= hfsc_bind_tcf,
 	.unbind_tcf	= hfsc_unbind_tcf,
-	.tcf_chain	= hfsc_tcf_chain,
+	.tcf_block	= hfsc_tcf_block,
 	.dump		= hfsc_dump_class,
 	.dump_stats	= hfsc_dump_class_stats,
 	.walk		= hfsc_walk
@@ -1730,11 +1680,11 @@ static struct Qdisc_ops hfsc_qdisc_ops __read_mostly = {
 	.enqueue	= hfsc_enqueue,
 	.dequeue	= hfsc_dequeue,
 	.peek		= qdisc_peek_dequeued,
-	.drop		= hfsc_drop,
 	.cl_ops		= &hfsc_class_ops,
 	.priv_size	= sizeof(struct hfsc_sched),
 	.owner		= THIS_MODULE
 };
+MODULE_ALIAS_NET_SCH("hfsc");
 
 static int __init
 hfsc_init(void)
@@ -1749,5 +1699,6 @@ hfsc_cleanup(void)
 }
 
 MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Hierarchical Fair Service Curve scheduler");
 module_init(hfsc_init);
 module_exit(hfsc_cleanup);