diff --git a/net/sched/act_api.c b/net/sched/act_api.c
index 585829f..3258da3 100644
--- a/net/sched/act_api.c
+++ b/net/sched/act_api.c
@@ -501,7 +501,7 @@ int tcf_idr_create(struct tc_action_net *tn, u32 index, struct nlattr *est,
 	if (est) {
 		err = gen_new_estimator(&p->tcfa_bstats, p->cpu_bstats,
 					&p->tcfa_rate_est,
-					&p->tcfa_lock, NULL, est);
+					&p->tcfa_lock, false, est);
 		if (err)
 			goto err4;
 	}
@@ -1173,9 +1173,10 @@ int tcf_action_copy_stats(struct sk_buff *skb, struct tc_action *p,
 	if (err < 0)
 		goto errout;
 
-	if (gnet_stats_copy_basic(NULL, &d, p->cpu_bstats, &p->tcfa_bstats) < 0 ||
-	    gnet_stats_copy_basic_hw(NULL, &d, p->cpu_bstats_hw,
-				     &p->tcfa_bstats_hw) < 0 ||
+	if (gnet_stats_copy_basic(&d, p->cpu_bstats,
+				  &p->tcfa_bstats, false) < 0 ||
+	    gnet_stats_copy_basic_hw(&d, p->cpu_bstats_hw,
+				     &p->tcfa_bstats_hw, false) < 0 ||
 	    gnet_stats_copy_rate_est(&d, &p->tcfa_rate_est) < 0 ||
 	    gnet_stats_copy_queue(&d, p->cpu_qstats,
 				  &p->tcfa_qstats,
diff --git a/net/sched/act_police.c b/net/sched/act_police.c
index c938380..9e77ba8 100644
--- a/net/sched/act_police.c
+++ b/net/sched/act_police.c
@@ -125,7 +125,7 @@ static int tcf_police_init(struct net *net, struct nlattr *nla,
 					    police->common.cpu_bstats,
 					    &police->tcf_rate_est,
 					    &police->tcf_lock,
-					    NULL, est);
+					    false, est);
 		if (err)
 			goto failure;
 	} else if (tb[TCA_POLICE_AVRATE] &&
diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index 70f006c..efcd0b5 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -943,8 +943,7 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid,
 		cpu_qstats = q->cpu_qstats;
 	}
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(q),
-				  &d, cpu_bstats, &q->bstats) < 0 ||
+	if (gnet_stats_copy_basic(&d, cpu_bstats, &q->bstats, true) < 0 ||
 	    gnet_stats_copy_rate_est(&d, &q->rate_est) < 0 ||
 	    gnet_stats_copy_queue(&d, cpu_qstats, &q->qstats, qlen) < 0)
 		goto nla_put_failure;
@@ -1265,26 +1264,17 @@ static struct Qdisc *qdisc_create(struct net_device *dev,
 		rcu_assign_pointer(sch->stab, stab);
 	}
 	if (tca[TCA_RATE]) {
-		seqcount_t *running;
-
 		err = -EOPNOTSUPP;
 		if (sch->flags & TCQ_F_MQROOT) {
 			NL_SET_ERR_MSG(extack, "Cannot attach rate estimator to a multi-queue root qdisc");
 			goto err_out4;
 		}
 
-		if (sch->parent != TC_H_ROOT &&
-		    !(sch->flags & TCQ_F_INGRESS) &&
-		    (!p || !(p->flags & TCQ_F_MQROOT)))
-			running = qdisc_root_sleeping_running(sch);
-		else
-			running = &sch->running;
-
 		err = gen_new_estimator(&sch->bstats,
 					sch->cpu_bstats,
 					&sch->rate_est,
 					NULL,
-					running,
+					true,
 					tca[TCA_RATE]);
 		if (err) {
 			NL_SET_ERR_MSG(extack, "Failed to generate new estimator");
@@ -1360,7 +1350,7 @@ static int qdisc_change(struct Qdisc *sch, struct nlattr **tca,
 				      sch->cpu_bstats,
 				      &sch->rate_est,
 				      NULL,
-				      qdisc_root_sleeping_running(sch),
+				      true,
 				      tca[TCA_RATE]);
 	}
 out:
diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c
index fbfe4ce..4c8e994 100644
--- a/net/sched/sch_atm.c
+++ b/net/sched/sch_atm.c
@@ -653,8 +653,7 @@ atm_tc_dump_class_stats(struct Qdisc *sch, unsigned long arg,
 {
 	struct atm_flow_data *flow = (struct atm_flow_data *)arg;
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, NULL, &flow->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, NULL, &flow->bstats, true) < 0 ||
 	    gnet_stats_copy_queue(d, NULL, &flow->qstats, flow->q->q.qlen) < 0)
 		return -1;
 
diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c
index f0b1282f..02d9f0d 100644
--- a/net/sched/sch_cbq.c
+++ b/net/sched/sch_cbq.c
@@ -1383,8 +1383,7 @@ cbq_dump_class_stats(struct Qdisc *sch, unsigned long arg,
 	if (cl->undertime != PSCHED_PASTPERFECT)
 		cl->xstats.undertime = cl->undertime - q->now;
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, NULL, &cl->bstats) < 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;
@@ -1518,7 +1517,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t
 			err = gen_replace_estimator(&cl->bstats, NULL,
 						    &cl->rate_est,
 						    NULL,
-						    qdisc_root_sleeping_running(sch),
+						    true,
 						    tca[TCA_RATE]);
 			if (err) {
 				NL_SET_ERR_MSG(extack, "Failed to replace specified rate estimator");
@@ -1619,9 +1618,7 @@ cbq_change_class(struct Qdisc *sch, u32 classid, u32 parentid, struct nlattr **t
 
 	if (tca[TCA_RATE]) {
 		err = gen_new_estimator(&cl->bstats, NULL, &cl->rate_est,
-					NULL,
-					qdisc_root_sleeping_running(sch),
-					tca[TCA_RATE]);
+					NULL, true, tca[TCA_RATE]);
 		if (err) {
 			NL_SET_ERR_MSG(extack, "Couldn't create new estimator");
 			tcf_block_put(cl->block);
diff --git a/net/sched/sch_drr.c b/net/sched/sch_drr.c
index 7243617..18e4f7a 100644
--- a/net/sched/sch_drr.c
+++ b/net/sched/sch_drr.c
@@ -85,8 +85,7 @@ static int drr_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 		if (tca[TCA_RATE]) {
 			err = gen_replace_estimator(&cl->bstats, NULL,
 						    &cl->rate_est,
-						    NULL,
-						    qdisc_root_sleeping_running(sch),
+						    NULL, true,
 						    tca[TCA_RATE]);
 			if (err) {
 				NL_SET_ERR_MSG(extack, "Failed to replace estimator");
@@ -119,9 +118,7 @@ static int drr_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 
 	if (tca[TCA_RATE]) {
 		err = gen_replace_estimator(&cl->bstats, NULL, &cl->rate_est,
-					    NULL,
-					    qdisc_root_sleeping_running(sch),
-					    tca[TCA_RATE]);
+					    NULL, true, tca[TCA_RATE]);
 		if (err) {
 			NL_SET_ERR_MSG(extack, "Failed to replace estimator");
 			qdisc_put(cl->qdisc);
@@ -268,8 +265,7 @@ static int drr_dump_class_stats(struct Qdisc *sch, unsigned long arg,
 	if (qlen)
 		xstats.deficit = cl->deficit;
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, NULL, &cl->bstats) < 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, cl_q->cpu_qstats, &cl_q->qstats, qlen) < 0)
 		return -1;
diff --git a/net/sched/sch_ets.c b/net/sched/sch_ets.c
index af56d15..0eae9ff 100644
--- a/net/sched/sch_ets.c
+++ b/net/sched/sch_ets.c
@@ -325,8 +325,7 @@ static int ets_class_dump_stats(struct Qdisc *sch, unsigned long arg,
 	struct ets_class *cl = ets_class_from_arg(sch, arg);
 	struct Qdisc *cl_q = cl->qdisc;
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, NULL, &cl_q->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, NULL, &cl_q->bstats, true) < 0 ||
 	    qdisc_qstats_copy(d, cl_q) < 0)
 		return -1;
 
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 989186e..b0ff0df 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -304,8 +304,8 @@ static struct sk_buff *dequeue_skb(struct Qdisc *q, bool *validate,
 
 /*
  * Transmit possibly several skbs, and handle the return status as
- * required. Owning running seqcount bit guarantees that
- * only one CPU can execute this function.
+ * required. Owning qdisc running bit guarantees that only one CPU
+ * can execute this function.
  *
  * Returns to the caller:
  *				false  - hardware queue frozen backoff
@@ -606,7 +606,6 @@ struct Qdisc noop_qdisc = {
 	.ops		=	&noop_qdisc_ops,
 	.q.lock		=	__SPIN_LOCK_UNLOCKED(noop_qdisc.q.lock),
 	.dev_queue	=	&noop_netdev_queue,
-	.running	=	SEQCNT_ZERO(noop_qdisc.running),
 	.busylock	=	__SPIN_LOCK_UNLOCKED(noop_qdisc.busylock),
 	.gso_skb = {
 		.next = (struct sk_buff *)&noop_qdisc.gso_skb,
@@ -867,7 +866,6 @@ struct Qdisc_ops pfifo_fast_ops __read_mostly = {
 EXPORT_SYMBOL(pfifo_fast_ops);
 
 static struct lock_class_key qdisc_tx_busylock;
-static struct lock_class_key qdisc_running_key;
 
 struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
 			  const struct Qdisc_ops *ops,
@@ -917,10 +915,6 @@ struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue,
 	lockdep_set_class(&sch->seqlock,
 			  dev->qdisc_tx_busylock ?: &qdisc_tx_busylock);
 
-	seqcount_init(&sch->running);
-	lockdep_set_class(&sch->running,
-			  dev->qdisc_running_key ?: &qdisc_running_key);
-
 	sch->ops = ops;
 	sch->flags = ops->static_flags;
 	sch->enqueue = ops->enqueue;
diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c
index 181c290..d3979a6 100644
--- a/net/sched/sch_hfsc.c
+++ b/net/sched/sch_hfsc.c
@@ -965,7 +965,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 			err = gen_replace_estimator(&cl->bstats, NULL,
 						    &cl->rate_est,
 						    NULL,
-						    qdisc_root_sleeping_running(sch),
+						    true,
 						    tca[TCA_RATE]);
 			if (err)
 				return err;
@@ -1033,9 +1033,7 @@ hfsc_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 
 	if (tca[TCA_RATE]) {
 		err = gen_new_estimator(&cl->bstats, NULL, &cl->rate_est,
-					NULL,
-					qdisc_root_sleeping_running(sch),
-					tca[TCA_RATE]);
+					NULL, true, tca[TCA_RATE]);
 		if (err) {
 			tcf_block_put(cl->block);
 			kfree(cl);
@@ -1328,7 +1326,7 @@ hfsc_dump_class_stats(struct Qdisc *sch, unsigned long arg,
 	xstats.work    = cl->cl_total;
 	xstats.rtwork  = cl->cl_cumul;
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch), d, NULL, &cl->bstats) < 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;
diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c
index adceb9e..cf1d45d 100644
--- a/net/sched/sch_htb.c
+++ b/net/sched/sch_htb.c
@@ -1368,8 +1368,7 @@ htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d)
 		}
 	}
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, NULL, &cl->bstats) < 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, &qs, qlen) < 0)
 		return -1;
@@ -1865,7 +1864,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 			err = gen_new_estimator(&cl->bstats, NULL,
 						&cl->rate_est,
 						NULL,
-						qdisc_root_sleeping_running(sch),
+						true,
 						tca[TCA_RATE] ? : &est.nla);
 			if (err)
 				goto err_block_put;
@@ -1991,7 +1990,7 @@ static int htb_change_class(struct Qdisc *sch, u32 classid,
 			err = gen_replace_estimator(&cl->bstats, NULL,
 						    &cl->rate_est,
 						    NULL,
-						    qdisc_root_sleeping_running(sch),
+						    true,
 						    tca[TCA_RATE]);
 			if (err)
 				return err;
diff --git a/net/sched/sch_mq.c b/net/sched/sch_mq.c
index cedd0b3..83d2e54 100644
--- a/net/sched/sch_mq.c
+++ b/net/sched/sch_mq.c
@@ -144,8 +144,8 @@ static int mq_dump(struct Qdisc *sch, struct sk_buff *skb)
 		qdisc = netdev_get_tx_queue(dev, ntx)->qdisc_sleeping;
 		spin_lock_bh(qdisc_lock(qdisc));
 
-		gnet_stats_add_basic(NULL, &sch->bstats, qdisc->cpu_bstats,
-				     &qdisc->bstats);
+		gnet_stats_add_basic(&sch->bstats, qdisc->cpu_bstats,
+				     &qdisc->bstats, false);
 		gnet_stats_add_queue(&sch->qstats, qdisc->cpu_qstats,
 				     &qdisc->qstats);
 		sch->q.qlen += qdisc_qlen(qdisc);
@@ -231,8 +231,7 @@ static int mq_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 	struct netdev_queue *dev_queue = mq_queue_get(sch, cl);
 
 	sch = dev_queue->qdisc_sleeping;
-	if (gnet_stats_copy_basic(&sch->running, d, sch->cpu_bstats,
-				  &sch->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, sch->cpu_bstats, &sch->bstats, true) < 0 ||
 	    qdisc_qstats_copy(d, sch) < 0)
 		return -1;
 	return 0;
diff --git a/net/sched/sch_mqprio.c b/net/sched/sch_mqprio.c
index 3f7f756f..b29f345 100644
--- a/net/sched/sch_mqprio.c
+++ b/net/sched/sch_mqprio.c
@@ -402,8 +402,8 @@ static int mqprio_dump(struct Qdisc *sch, struct sk_buff *skb)
 		qdisc = netdev_get_tx_queue(dev, ntx)->qdisc_sleeping;
 		spin_lock_bh(qdisc_lock(qdisc));
 
-		gnet_stats_add_basic(NULL, &sch->bstats, qdisc->cpu_bstats,
-				     &qdisc->bstats);
+		gnet_stats_add_basic(&sch->bstats, qdisc->cpu_bstats,
+				     &qdisc->bstats, false);
 		gnet_stats_add_queue(&sch->qstats, qdisc->cpu_qstats,
 				     &qdisc->qstats);
 		sch->q.qlen += qdisc_qlen(qdisc);
@@ -519,8 +519,8 @@ static int mqprio_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 
 			spin_lock_bh(qdisc_lock(qdisc));
 
-			gnet_stats_add_basic(NULL, &bstats, qdisc->cpu_bstats,
-					     &qdisc->bstats);
+			gnet_stats_add_basic(&bstats, qdisc->cpu_bstats,
+					     &qdisc->bstats, false);
 			gnet_stats_add_queue(&qstats, qdisc->cpu_qstats,
 					     &qdisc->qstats);
 			sch->q.qlen += qdisc_qlen(qdisc);
@@ -532,15 +532,15 @@ static int mqprio_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 		/* Reclaim root sleeping lock before completing stats */
 		if (d->lock)
 			spin_lock_bh(d->lock);
-		if (gnet_stats_copy_basic(NULL, d, NULL, &bstats) < 0 ||
+		if (gnet_stats_copy_basic(d, NULL, &bstats, false) < 0 ||
 		    gnet_stats_copy_queue(d, NULL, &qstats, qlen) < 0)
 			return -1;
 	} else {
 		struct netdev_queue *dev_queue = mqprio_queue_get(sch, cl);
 
 		sch = dev_queue->qdisc_sleeping;
-		if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch), d,
-					  sch->cpu_bstats, &sch->bstats) < 0 ||
+		if (gnet_stats_copy_basic(d, sch->cpu_bstats,
+					  &sch->bstats, true) < 0 ||
 		    qdisc_qstats_copy(d, sch) < 0)
 			return -1;
 	}
diff --git a/net/sched/sch_multiq.c b/net/sched/sch_multiq.c
index e282e73..cd8ab90 100644
--- a/net/sched/sch_multiq.c
+++ b/net/sched/sch_multiq.c
@@ -338,8 +338,7 @@ static int multiq_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 	struct Qdisc *cl_q;
 
 	cl_q = q->queues[cl - 1];
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, cl_q->cpu_bstats, &cl_q->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, cl_q->cpu_bstats, &cl_q->bstats, true) < 0 ||
 	    qdisc_qstats_copy(d, cl_q) < 0)
 		return -1;
 
diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c
index 03fdf31..3b8d719 100644
--- a/net/sched/sch_prio.c
+++ b/net/sched/sch_prio.c
@@ -361,8 +361,8 @@ static int prio_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 	struct Qdisc *cl_q;
 
 	cl_q = q->queues[cl - 1];
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, cl_q->cpu_bstats, &cl_q->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, cl_q->cpu_bstats,
+				  &cl_q->bstats, true) < 0 ||
 	    qdisc_qstats_copy(d, cl_q) < 0)
 		return -1;
 
diff --git a/net/sched/sch_qfq.c b/net/sched/sch_qfq.c
index a35200f..0b7f9ba 100644
--- a/net/sched/sch_qfq.c
+++ b/net/sched/sch_qfq.c
@@ -451,7 +451,7 @@ static int qfq_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 			err = gen_replace_estimator(&cl->bstats, NULL,
 						    &cl->rate_est,
 						    NULL,
-						    qdisc_root_sleeping_running(sch),
+						    true,
 						    tca[TCA_RATE]);
 			if (err)
 				return err;
@@ -478,7 +478,7 @@ static int qfq_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
 		err = gen_new_estimator(&cl->bstats, NULL,
 					&cl->rate_est,
 					NULL,
-					qdisc_root_sleeping_running(sch),
+					true,
 					tca[TCA_RATE]);
 		if (err)
 			goto destroy_class;
@@ -640,8 +640,7 @@ static int qfq_dump_class_stats(struct Qdisc *sch, unsigned long arg,
 	xstats.weight = cl->agg->class_weight;
 	xstats.lmax = cl->agg->lmax;
 
-	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
-				  d, NULL, &cl->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, NULL, &cl->bstats, true) < 0 ||
 	    gnet_stats_copy_rate_est(d, &cl->rate_est) < 0 ||
 	    qdisc_qstats_copy(d, cl->qdisc) < 0)
 		return -1;
diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c
index b9fd18d..9ab068f 100644
--- a/net/sched/sch_taprio.c
+++ b/net/sched/sch_taprio.c
@@ -1977,7 +1977,7 @@ static int taprio_dump_class_stats(struct Qdisc *sch, unsigned long cl,
 	struct netdev_queue *dev_queue = taprio_queue_get(sch, cl);
 
 	sch = dev_queue->qdisc_sleeping;
-	if (gnet_stats_copy_basic(&sch->running, d, NULL, &sch->bstats) < 0 ||
+	if (gnet_stats_copy_basic(d, NULL, &sch->bstats, true) < 0 ||
 	    qdisc_qstats_copy(d, sch) < 0)
 		return -1;
 	return 0;