Merge branch 'net-sched-bulk-dequeue'

Eric Dumazet says:

====================
net_sched: bulk dequeue and deferred drops

First patch adds an additional parameter to ->enqueue() qdisc method
so that drops can be done outside of critical section
(after locks are released).

Then fq_codel can have a small optimization to reduce number of cache
lines misses during a drop event
(possibly accumulating hundreds of packets to be freed).

A small htb change exports the backlog in class dumps.

Final patch adds bulk dequeue to qdiscs that were lacking this feature.

This series brings a nice qdisc performance increase (more than 80 %
in some cases).
====================

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

29 files changed, 247 insertions, 134 deletions

diff --git a/include/net/codel_qdisc.h b/include/net/codel_qdisc.h
index 8144d9cd2908..098630f83a55 100644
--- a/include/net/codel_qdisc.h
+++ b/include/net/codel_qdisc.h
@@ -52,6 +52,7 @@
 /* Qdiscs using codel plugin must use codel_skb_cb in their own cb[] */
 struct codel_skb_cb {
 	codel_time_t enqueue_time;
+	unsigned int mem_usage;
 };
 
 static struct codel_skb_cb *get_codel_cb(const struct sk_buff *skb)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index 4f7cee8344c4..909aff2db2b3 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -37,8 +37,10 @@ struct qdisc_size_table {
 };
 
 struct Qdisc {
-	int 			(*enqueue)(struct sk_buff *skb, struct Qdisc *dev);
-	struct sk_buff *	(*dequeue)(struct Qdisc *dev);
+	int 			(*enqueue)(struct sk_buff *skb,
+					   struct Qdisc *sch,
+					   struct sk_buff **to_free);
+	struct sk_buff *	(*dequeue)(struct Qdisc *sch);
 	unsigned int		flags;
 #define TCQ_F_BUILTIN		1
 #define TCQ_F_INGRESS		2
@@ -73,13 +75,14 @@ struct Qdisc {
 	/*
 	 * For performance sake on SMP, we put highly modified fields at the end
 	 */
-	struct Qdisc		*next_sched ____cacheline_aligned_in_smp;
-	struct sk_buff		*gso_skb;
-	unsigned long		state;
+	struct sk_buff		*gso_skb ____cacheline_aligned_in_smp;
 	struct sk_buff_head	q;
 	struct gnet_stats_basic_packed bstats;
 	seqcount_t		running;
 	struct gnet_stats_queue	qstats;
+	unsigned long		state;
+	struct Qdisc            *next_sched;
+	struct sk_buff		*skb_bad_txq;
 	struct rcu_head		rcu_head;
 	int			padded;
 	atomic_t		refcnt;
@@ -160,7 +163,9 @@ struct Qdisc_ops {
 	char			id[IFNAMSIZ];
 	int			priv_size;
 
-	int 			(*enqueue)(struct sk_buff *, struct Qdisc *);
+	int 			(*enqueue)(struct sk_buff *skb,
+					   struct Qdisc *sch,
+					   struct sk_buff **to_free);
 	struct sk_buff *	(*dequeue)(struct Qdisc *);
 	struct sk_buff *	(*peek)(struct Qdisc *);
 
@@ -498,10 +503,11 @@ static inline void qdisc_calculate_pkt_len(struct sk_buff *skb,
 #endif
 }
 
-static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+				struct sk_buff **to_free)
 {
 	qdisc_calculate_pkt_len(skb, sch);
-	return sch->enqueue(skb, sch);
+	return sch->enqueue(skb, sch, to_free);
 }
 
 static inline bool qdisc_is_percpu_stats(const struct Qdisc *q)
@@ -626,24 +632,36 @@ static inline struct sk_buff *qdisc_dequeue_head(struct Qdisc *sch)
 	return __qdisc_dequeue_head(sch, &sch->q);
 }
 
+/* Instead of calling kfree_skb() while root qdisc lock is held,
+ * queue the skb for future freeing at end of __dev_xmit_skb()
+ */
+static inline void __qdisc_drop(struct sk_buff *skb, struct sk_buff **to_free)
+{
+	skb->next = *to_free;
+	*to_free = skb;
+}
+
 static inline unsigned int __qdisc_queue_drop_head(struct Qdisc *sch,
-					      struct sk_buff_head *list)
+						   struct sk_buff_head *list,
+						   struct sk_buff **to_free)
 {
 	struct sk_buff *skb = __skb_dequeue(list);
 
 	if (likely(skb != NULL)) {
 		unsigned int len = qdisc_pkt_len(skb);
+
 		qdisc_qstats_backlog_dec(sch, skb);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return len;
 	}
 
 	return 0;
 }
 
-static inline unsigned int qdisc_queue_drop_head(struct Qdisc *sch)
+static inline unsigned int qdisc_queue_drop_head(struct Qdisc *sch,
+						 struct sk_buff **to_free)
 {
-	return __qdisc_queue_drop_head(sch, &sch->q);
+	return __qdisc_queue_drop_head(sch, &sch->q, to_free);
 }
 
 static inline struct sk_buff *qdisc_peek_head(struct Qdisc *sch)
@@ -724,9 +742,11 @@ static inline void rtnl_qdisc_drop(struct sk_buff *skb, struct Qdisc *sch)
 	qdisc_qstats_drop(sch);
 }
 
-static inline int qdisc_drop(struct sk_buff *skb, struct Qdisc *sch)
+
+static inline int qdisc_drop(struct sk_buff *skb, struct Qdisc *sch,
+			     struct sk_buff **to_free)
 {
-	kfree_skb(skb);
+	__qdisc_drop(skb, to_free);
 	qdisc_qstats_drop(sch);
 
 	return NET_XMIT_DROP;
diff --git a/net/core/dev.c b/net/core/dev.c
index d40593b3b9fb..aba10d2a8bc3 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3070,6 +3070,7 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
 				 struct netdev_queue *txq)
 {
 	spinlock_t *root_lock = qdisc_lock(q);
+	struct sk_buff *to_free = NULL;
 	bool contended;
 	int rc;
 
@@ -3086,7 +3087,7 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
 
 	spin_lock(root_lock);
 	if (unlikely(test_bit(__QDISC_STATE_DEACTIVATED, &q->state))) {
-		kfree_skb(skb);
+		__qdisc_drop(skb, &to_free);
 		rc = NET_XMIT_DROP;
 	} else if ((q->flags & TCQ_F_CAN_BYPASS) && !qdisc_qlen(q) &&
 		   qdisc_run_begin(q)) {
@@ -3109,7 +3110,7 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
 
 		rc = NET_XMIT_SUCCESS;
 	} else {
-		rc = q->enqueue(skb, q) & NET_XMIT_MASK;
+		rc = q->enqueue(skb, q, &to_free) & NET_XMIT_MASK;
 		if (qdisc_run_begin(q)) {
 			if (unlikely(contended)) {
 				spin_unlock(&q->busylock);
@@ -3119,6 +3120,8 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
 		}
 	}
 	spin_unlock(root_lock);
+	if (unlikely(to_free))
+		kfree_skb_list(to_free);
 	if (unlikely(contended))
 		spin_unlock(&q->busylock);
 	return rc;
diff --git a/net/sched/sch_atm.c b/net/sched/sch_atm.c
index e04ea6994d1c..481e4f12aeb4 100644
--- a/net/sched/sch_atm.c
+++ b/net/sched/sch_atm.c
@@ -357,7 +357,8 @@ static struct tcf_proto __rcu **atm_tc_find_tcf(struct Qdisc *sch,
 
 /* --------------------------- Qdisc operations ---------------------------- */
 
-static int atm_tc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int atm_tc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			  struct sk_buff **to_free)
 {
 	struct atm_qdisc_data *p = qdisc_priv(sch);
 	struct atm_flow_data *flow;
@@ -398,10 +399,10 @@ done:
 		switch (result) {
 		case TC_ACT_QUEUED:
 		case TC_ACT_STOLEN:
-			kfree_skb(skb);
+			__qdisc_drop(skb, to_free);
 			return NET_XMIT_SUCCESS | __NET_XMIT_STOLEN;
 		case TC_ACT_SHOT:
-			kfree_skb(skb);
+			__qdisc_drop(skb, to_free);
 			goto drop;
 		case TC_ACT_RECLASSIFY:
 			if (flow->excess)
@@ -413,7 +414,7 @@ done:
 #endif
 	}
 
-	ret = qdisc_enqueue(skb, flow->q);
+	ret = qdisc_enqueue(skb, flow->q, to_free);
 	if (ret != NET_XMIT_SUCCESS) {
 drop: __maybe_unused
 		if (net_xmit_drop_count(ret)) {
diff --git a/net/sched/sch_blackhole.c b/net/sched/sch_blackhole.c
index 3fee70d9814f..c98a61e980ba 100644
--- a/net/sched/sch_blackhole.c
+++ b/net/sched/sch_blackhole.c
@@ -17,9 +17,10 @@
 #include <linux/skbuff.h>
 #include <net/pkt_sched.h>
 
-static int blackhole_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int blackhole_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			     struct sk_buff **to_free)
 {
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	return NET_XMIT_SUCCESS;
 }
 
diff --git a/net/sched/sch_cbq.c b/net/sched/sch_cbq.c
index a29fd811d7b9..beb554aa8cfb 100644
--- a/net/sched/sch_cbq.c
+++ b/net/sched/sch_cbq.c
@@ -358,7 +358,8 @@ cbq_mark_toplevel(struct cbq_sched_data *q, struct cbq_class *cl)
 }
 
 static int
-cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+	    struct sk_buff **to_free)
 {
 	struct cbq_sched_data *q = qdisc_priv(sch);
 	int uninitialized_var(ret);
@@ -370,11 +371,11 @@ cbq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (cl == NULL) {
 		if (ret & __NET_XMIT_BYPASS)
 			qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return ret;
 	}
 
-	ret = qdisc_enqueue(skb, cl->q);
+	ret = qdisc_enqueue(skb, cl->q, to_free);
 	if (ret == NET_XMIT_SUCCESS) {
 		sch->q.qlen++;
 		cbq_mark_toplevel(q, cl);
diff --git a/net/sched/sch_choke.c b/net/sched/sch_choke.c
index 789b69ee9e51..3b6d5bd69101 100644
--- a/net/sched/sch_choke.c
+++ b/net/sched/sch_choke.c
@@ -115,7 +115,8 @@ static void choke_zap_tail_holes(struct choke_sched_data *q)
 }
 
 /* Drop packet from queue array by creating a "hole" */
-static void choke_drop_by_idx(struct Qdisc *sch, unsigned int idx)
+static void choke_drop_by_idx(struct Qdisc *sch, unsigned int idx,
+			      struct sk_buff **to_free)
 {
 	struct choke_sched_data *q = qdisc_priv(sch);
 	struct sk_buff *skb = q->tab[idx];
@@ -129,7 +130,7 @@ static void choke_drop_by_idx(struct Qdisc *sch, unsigned int idx)
 
 	qdisc_qstats_backlog_dec(sch, skb);
 	qdisc_tree_reduce_backlog(sch, 1, qdisc_pkt_len(skb));
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	--sch->q.qlen;
 }
 
@@ -261,7 +262,8 @@ static bool choke_match_random(const struct choke_sched_data *q,
 	return choke_match_flow(oskb, nskb);
 }
 
-static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			 struct sk_buff **to_free)
 {
 	int ret = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
 	struct choke_sched_data *q = qdisc_priv(sch);
@@ -288,7 +290,7 @@ static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		/* Draw a packet at random from queue and compare flow */
 		if (choke_match_random(q, skb, &idx)) {
 			q->stats.matched++;
-			choke_drop_by_idx(sch, idx);
+			choke_drop_by_idx(sch, idx, to_free);
 			goto congestion_drop;
 		}
 
@@ -331,16 +333,16 @@ static int choke_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	}
 
 	q->stats.pdrop++;
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 
 congestion_drop:
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	return NET_XMIT_CN;
 
 other_drop:
 	if (ret & __NET_XMIT_BYPASS)
 		qdisc_qstats_drop(sch);
-	kfree_skb(skb);
+	__qdisc_drop(skb, to_free);
 	return ret;
 }
 
diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c
index c5bc424e3b3c..4002df3c7d9f 100644
--- a/net/sched/sch_codel.c
+++ b/net/sched/sch_codel.c
@@ -82,7 +82,8 @@ static void drop_func(struct sk_buff *skb, void *ctx)
 {
 	struct Qdisc *sch = ctx;
 
-	qdisc_drop(skb, sch);
+	kfree_skb(skb);
+	qdisc_qstats_drop(sch);
 }
 
 static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch)
@@ -107,7 +108,8 @@ static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch)
 	return skb;
 }
 
-static int codel_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int codel_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			       struct sk_buff **to_free)
 {
 	struct codel_sched_data *q;
 
@@ -117,7 +119,7 @@ static int codel_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	}
 	q = qdisc_priv(sch);
 	q->drop_overlimit++;
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 }
 
 static const struct nla_policy codel_policy[TCA_CODEL_MAX + 1] = {
diff --git a/net/sched/sch_drr.c b/net/sched/sch_drr.c
index 22609e4e845f..8af5c59eef84 100644
--- a/net/sched/sch_drr.c
+++ b/net/sched/sch_drr.c
@@ -350,7 +350,8 @@ static struct drr_class *drr_classify(struct sk_buff *skb, struct Qdisc *sch,
 	return NULL;
 }
 
-static int drr_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int drr_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	struct drr_sched *q = qdisc_priv(sch);
 	struct drr_class *cl;
@@ -360,11 +361,11 @@ static int drr_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (cl == NULL) {
 		if (err & __NET_XMIT_BYPASS)
 			qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return err;
 	}
 
-	err = qdisc_enqueue(skb, cl->qdisc);
+	err = qdisc_enqueue(skb, cl->qdisc, to_free);
 	if (unlikely(err != NET_XMIT_SUCCESS)) {
 		if (net_xmit_drop_count(err)) {
 			cl->qstats.drops++;
diff --git a/net/sched/sch_dsmark.c b/net/sched/sch_dsmark.c
index b9ba5f658528..1308bbf460f7 100644
--- a/net/sched/sch_dsmark.c
+++ b/net/sched/sch_dsmark.c
@@ -191,7 +191,8 @@ static inline struct tcf_proto __rcu **dsmark_find_tcf(struct Qdisc *sch,
 
 /* --------------------------- Qdisc operations ---------------------------- */
 
-static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			  struct sk_buff **to_free)
 {
 	struct dsmark_qdisc_data *p = qdisc_priv(sch);
 	int err;
@@ -234,7 +235,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 #ifdef CONFIG_NET_CLS_ACT
 		case TC_ACT_QUEUED:
 		case TC_ACT_STOLEN:
-			kfree_skb(skb);
+			__qdisc_drop(skb, to_free);
 			return NET_XMIT_SUCCESS | __NET_XMIT_STOLEN;
 
 		case TC_ACT_SHOT:
@@ -251,7 +252,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		}
 	}
 
-	err = qdisc_enqueue(skb, p->q);
+	err = qdisc_enqueue(skb, p->q, to_free);
 	if (err != NET_XMIT_SUCCESS) {
 		if (net_xmit_drop_count(err))
 			qdisc_qstats_drop(sch);
@@ -264,7 +265,7 @@ static int dsmark_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	return NET_XMIT_SUCCESS;
 
 drop:
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
 }
 
diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c
index dea70e3ef0ba..6ea0db427f91 100644
--- a/net/sched/sch_fifo.c
+++ b/net/sched/sch_fifo.c
@@ -19,29 +19,32 @@
 
 /* 1 band FIFO pseudo-"scheduler" */
 
-static int bfifo_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int bfifo_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			 struct sk_buff **to_free)
 {
 	if (likely(sch->qstats.backlog + qdisc_pkt_len(skb) <= sch->limit))
 		return qdisc_enqueue_tail(skb, sch);
 
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 }
 
-static int pfifo_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int pfifo_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			 struct sk_buff **to_free)
 {
 	if (likely(skb_queue_len(&sch->q) < sch->limit))
 		return qdisc_enqueue_tail(skb, sch);
 
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 }
 
-static int pfifo_tail_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int pfifo_tail_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			      struct sk_buff **to_free)
 {
 	if (likely(skb_queue_len(&sch->q) < sch->limit))
 		return qdisc_enqueue_tail(skb, sch);
 
 	/* queue full, remove one skb to fulfill the limit */
-	__qdisc_queue_drop_head(sch, &sch->q);
+	__qdisc_queue_drop_head(sch, &sch->q, to_free);
 	qdisc_qstats_drop(sch);
 	qdisc_enqueue_tail(skb, sch);
 
diff --git a/net/sched/sch_fq.c b/net/sched/sch_fq.c
index 6eb06674f778..e5458b99e09c 100644
--- a/net/sched/sch_fq.c
+++ b/net/sched/sch_fq.c
@@ -368,18 +368,19 @@ static void flow_queue_add(struct fq_flow *flow, struct sk_buff *skb)
 	}
 }
 
-static int fq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int fq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		      struct sk_buff **to_free)
 {
 	struct fq_sched_data *q = qdisc_priv(sch);
 	struct fq_flow *f;
 
 	if (unlikely(sch->q.qlen >= sch->limit))
-		return qdisc_drop(skb, sch);
+		return qdisc_drop(skb, sch, to_free);
 
 	f = fq_classify(skb, q);
 	if (unlikely(f->qlen >= q->flow_plimit && f != &q->internal)) {
 		q->stat_flows_plimit++;
-		return qdisc_drop(skb, sch);
+		return qdisc_drop(skb, sch, to_free);
 	}
 
 	f->qlen++;
diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c
index 2dc0a849515a..a5ea0e9b6be4 100644
--- a/net/sched/sch_fq_codel.c
+++ b/net/sched/sch_fq_codel.c
@@ -139,7 +139,8 @@ static inline void flow_queue_add(struct fq_codel_flow *flow,
 	skb->next = NULL;
 }
 
-static unsigned int fq_codel_drop(struct Qdisc *sch, unsigned int max_packets)
+static unsigned int fq_codel_drop(struct Qdisc *sch, unsigned int max_packets,
+				  struct sk_buff **to_free)
 {
 	struct fq_codel_sched_data *q = qdisc_priv(sch);
 	struct sk_buff *skb;
@@ -171,8 +172,8 @@ static unsigned int fq_codel_drop(struct Qdisc *sch, unsigned int max_packets)
 	do {
 		skb = dequeue_head(flow);
 		len += qdisc_pkt_len(skb);
-		mem += skb->truesize;
-		kfree_skb(skb);
+		mem += get_codel_cb(skb)->mem_usage;
+		__qdisc_drop(skb, to_free);
 	} while (++i < max_packets && len < threshold);
 
 	flow->dropped += i;
@@ -184,7 +185,8 @@ static unsigned int fq_codel_drop(struct Qdisc *sch, unsigned int max_packets)
 	return idx;
 }
 
-static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			    struct sk_buff **to_free)
 {
 	struct fq_codel_sched_data *q = qdisc_priv(sch);
 	unsigned int idx, prev_backlog, prev_qlen;
@@ -197,7 +199,7 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (idx == 0) {
 		if (ret & __NET_XMIT_BYPASS)
 			qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return ret;
 	}
 	idx--;
@@ -214,7 +216,8 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		flow->deficit = q->quantum;
 		flow->dropped = 0;
 	}
-	q->memory_usage += skb->truesize;
+	get_codel_cb(skb)->mem_usage = skb->truesize;
+	q->memory_usage += get_codel_cb(skb)->mem_usage;
 	memory_limited = q->memory_usage > q->memory_limit;
 	if (++sch->q.qlen <= sch->limit && !memory_limited)
 		return NET_XMIT_SUCCESS;
@@ -229,7 +232,7 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	 * So instead of dropping a single packet, drop half of its backlog
 	 * with a 64 packets limit to not add a too big cpu spike here.
 	 */
-	ret = fq_codel_drop(sch, q->drop_batch_size);
+	ret = fq_codel_drop(sch, q->drop_batch_size, to_free);
 
 	prev_qlen -= sch->q.qlen;
 	prev_backlog -= sch->qstats.backlog;
@@ -265,7 +268,7 @@ static struct sk_buff *dequeue_func(struct codel_vars *vars, void *ctx)
 	if (flow->head) {
 		skb = dequeue_head(flow);
 		q->backlogs[flow - q->flows] -= qdisc_pkt_len(skb);
-		q->memory_usage -= skb->truesize;
+		q->memory_usage -= get_codel_cb(skb)->mem_usage;
 		sch->q.qlen--;
 		sch->qstats.backlog -= qdisc_pkt_len(skb);
 	}
@@ -276,7 +279,8 @@ static void drop_func(struct sk_buff *skb, void *ctx)
 {
 	struct Qdisc *sch = ctx;
 
-	qdisc_drop(skb, sch);
+	kfree_skb(skb);
+	qdisc_qstats_drop(sch);
 }
 
 static struct sk_buff *fq_codel_dequeue(struct Qdisc *sch)
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 773b632e1e33..e95b67cd5718 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -77,6 +77,34 @@ static void try_bulk_dequeue_skb(struct Qdisc *q,
 	skb->next = NULL;
 }
 
+/* This variant of try_bulk_dequeue_skb() makes sure
+ * all skbs in the chain are for the same txq
+ */
+static void try_bulk_dequeue_skb_slow(struct Qdisc *q,
+				      struct sk_buff *skb,
+				      int *packets)
+{
+	int mapping = skb_get_queue_mapping(skb);
+	struct sk_buff *nskb;
+	int cnt = 0;
+
+	do {
+		nskb = q->dequeue(q);
+		if (!nskb)
+			break;
+		if (unlikely(skb_get_queue_mapping(nskb) != mapping)) {
+			q->skb_bad_txq = nskb;
+			qdisc_qstats_backlog_inc(q, nskb);
+			q->q.qlen++;
+			break;
+		}
+		skb->next = nskb;
+		skb = nskb;
+	} while (++cnt < 8);
+	(*packets) += cnt;
+	skb->next = NULL;
+}
+
 /* Note that dequeue_skb can possibly return a SKB list (via skb->next).
  * A requeued skb (via q->gso_skb) can also be a SKB list.
  */
@@ -87,8 +115,9 @@ static struct sk_buff *dequeue_skb(struct Qdisc *q, bool *validate,
 	const struct netdev_queue *txq = q->dev_queue;
 
 	*packets = 1;
-	*validate = true;
 	if (unlikely(skb)) {
+		/* skb in gso_skb were already validated */
+		*validate = false;
 		/* check the reason of requeuing without tx lock first */
 		txq = skb_get_tx_queue(txq->dev, skb);
 		if (!netif_xmit_frozen_or_stopped(txq)) {
@@ -97,15 +126,30 @@ static struct sk_buff *dequeue_skb(struct Qdisc *q, bool *validate,
 			q->q.qlen--;
 		} else
 			skb = NULL;
-		/* skb in gso_skb were already validated */
-		*validate = false;
-	} else {
-		if (!(q->flags & TCQ_F_ONETXQUEUE) ||
-		    !netif_xmit_frozen_or_stopped(txq)) {
-			skb = q->dequeue(q);
-			if (skb && qdisc_may_bulk(q))
-				try_bulk_dequeue_skb(q, skb, txq, packets);
+		return skb;
+	}
+	*validate = true;
+	skb = q->skb_bad_txq;
+	if (unlikely(skb)) {
+		/* check the reason of requeuing without tx lock first */
+		txq = skb_get_tx_queue(txq->dev, skb);
+		if (!netif_xmit_frozen_or_stopped(txq)) {
+			q->skb_bad_txq = NULL;
+			qdisc_qstats_backlog_dec(q, skb);
+			q->q.qlen--;
+			goto bulk;
 		}
+		return NULL;
+	}
+	if (!(q->flags & TCQ_F_ONETXQUEUE) ||
+	    !netif_xmit_frozen_or_stopped(txq))
+		skb = q->dequeue(q);
+	if (skb) {
+bulk:
+		if (qdisc_may_bulk(q))
+			try_bulk_dequeue_skb(q, skb, txq, packets);
+		else
+			try_bulk_dequeue_skb_slow(q, skb, packets);
 	}
 	return skb;
 }
@@ -348,9 +392,10 @@ EXPORT_SYMBOL(netif_carrier_off);
    cheaper.
  */
 
-static int noop_enqueue(struct sk_buff *skb, struct Qdisc *qdisc)
+static int noop_enqueue(struct sk_buff *skb, struct Qdisc *qdisc,
+			struct sk_buff **to_free)
 {
-	kfree_skb(skb);
+	__qdisc_drop(skb, to_free);
 	return NET_XMIT_CN;
 }
 
@@ -439,7 +484,8 @@ static inline struct sk_buff_head *band2list(struct pfifo_fast_priv *priv,
 	return priv->q + band;
 }
 
-static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc *qdisc)
+static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc *qdisc,
+			      struct sk_buff **to_free)
 {
 	if (skb_queue_len(&qdisc->q) < qdisc_dev(qdisc)->tx_queue_len) {
 		int band = prio2band[skb->priority & TC_PRIO_MAX];
@@ -451,7 +497,7 @@ static int pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc *qdisc)
 		return __qdisc_enqueue_tail(skb, qdisc, list);
 	}
 
-	return qdisc_drop(skb, qdisc);
+	return qdisc_drop(skb, qdisc, to_free);
 }
 
 static struct sk_buff *pfifo_fast_dequeue(struct Qdisc *qdisc)
@@ -622,11 +668,14 @@ void qdisc_reset(struct Qdisc *qdisc)
 	if (ops->reset)
 		ops->reset(qdisc);
 
+	kfree_skb(qdisc->skb_bad_txq);
+	qdisc->skb_bad_txq = NULL;
+
 	if (qdisc->gso_skb) {
 		kfree_skb_list(qdisc->gso_skb);
 		qdisc->gso_skb = NULL;
-		qdisc->q.qlen = 0;
 	}
+	qdisc->q.qlen = 0;
 }
 EXPORT_SYMBOL(qdisc_reset);
 
@@ -665,6 +714,7 @@ void qdisc_destroy(struct Qdisc *qdisc)
 	dev_put(qdisc_dev(qdisc));
 
 	kfree_skb_list(qdisc->gso_skb);
+	kfree_skb(qdisc->skb_bad_txq);
 	/*
 	 * gen_estimator est_timer() might access qdisc->q.lock,
 	 * wait a RCU grace period before freeing qdisc.
diff --git a/net/sched/sch_gred.c b/net/sched/sch_gred.c
index b5fb63c7be02..c78a093c551a 100644
--- a/net/sched/sch_gred.c
+++ b/net/sched/sch_gred.c
@@ -149,7 +149,8 @@ static inline int gred_use_harddrop(struct gred_sched *t)
 	return t->red_flags & TC_RED_HARDDROP;
 }
 
-static int gred_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int gred_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			struct sk_buff **to_free)
 {
 	struct gred_sched_data *q = NULL;
 	struct gred_sched *t = qdisc_priv(sch);
@@ -237,10 +238,10 @@ static int gred_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 
 	q->stats.pdrop++;
 drop:
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 
 congestion_drop:
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	return NET_XMIT_CN;
 }
 
diff --git a/net/sched/sch_hfsc.c b/net/sched/sch_hfsc.c
index bd08c363a26d..8cb5eff7b79c 100644
--- a/net/sched/sch_hfsc.c
+++ b/net/sched/sch_hfsc.c
@@ -1572,7 +1572,7 @@ 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)
 {
 	struct hfsc_class *cl;
 	int uninitialized_var(err);
@@ -1581,11 +1581,11 @@ hfsc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (cl == NULL) {
 		if (err & __NET_XMIT_BYPASS)
 			qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return err;
 	}
 
-	err = qdisc_enqueue(skb, cl->qdisc);
+	err = qdisc_enqueue(skb, cl->qdisc, to_free);
 	if (unlikely(err != NET_XMIT_SUCCESS)) {
 		if (net_xmit_drop_count(err)) {
 			cl->qstats.drops++;
diff --git a/net/sched/sch_hhf.c b/net/sched/sch_hhf.c
index c44593b8e65a..e3d0458af17b 100644
--- a/net/sched/sch_hhf.c
+++ b/net/sched/sch_hhf.c
@@ -345,7 +345,7 @@ static void bucket_add(struct wdrr_bucket *bucket, struct sk_buff *skb)
 	skb->next = NULL;
 }
 
-static unsigned int hhf_drop(struct Qdisc *sch)
+static unsigned int hhf_drop(struct Qdisc *sch, struct sk_buff **to_free)
 {
 	struct hhf_sched_data *q = qdisc_priv(sch);
 	struct wdrr_bucket *bucket;
@@ -359,16 +359,16 @@ static unsigned int hhf_drop(struct Qdisc *sch)
 		struct sk_buff *skb = dequeue_head(bucket);
 
 		sch->q.qlen--;
-		qdisc_qstats_drop(sch);
 		qdisc_qstats_backlog_dec(sch, skb);
-		kfree_skb(skb);
+		qdisc_drop(skb, sch, to_free);
 	}
 
 	/* Return id of the bucket from which the packet was dropped. */
 	return bucket - q->buckets;
 }
 
-static int hhf_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int hhf_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	struct hhf_sched_data *q = qdisc_priv(sch);
 	enum wdrr_bucket_idx idx;
@@ -406,7 +406,7 @@ static int hhf_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	/* Return Congestion Notification only if we dropped a packet from this
 	 * bucket.
 	 */
-	if (hhf_drop(sch) == idx)
+	if (hhf_drop(sch, to_free) == idx)
 		return NET_XMIT_CN;
 
 	/* As we dropped a packet, better let upper stack know this. */
diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c
index a454605ab5cb..ba098f2654b4 100644
--- a/net/sched/sch_htb.c
+++ b/net/sched/sch_htb.c
@@ -117,7 +117,6 @@ struct htb_class {
 	 * Written often fields
 	 */
 	struct gnet_stats_basic_packed bstats;
-	struct gnet_stats_queue	qstats;
 	struct tc_htb_xstats	xstats;	/* our special stats */
 
 	/* token bucket parameters */
@@ -140,6 +139,8 @@ struct htb_class {
 	enum htb_cmode		cmode;		/* current mode of the class */
 	struct rb_node		pq_node;	/* node for event queue */
 	struct rb_node		node[TC_HTB_NUMPRIO];	/* node for self or feed tree */
+
+	unsigned int drops ____cacheline_aligned_in_smp;
 };
 
 struct htb_level {
@@ -569,7 +570,8 @@ static inline void htb_deactivate(struct htb_sched *q, struct htb_class *cl)
 	list_del_init(&cl->un.leaf.drop_list);
 }
 
-static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	int uninitialized_var(ret);
 	struct htb_sched *q = qdisc_priv(sch);
@@ -581,19 +583,20 @@ static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 			__skb_queue_tail(&q->direct_queue, skb);
 			q->direct_pkts++;
 		} else {
-			return qdisc_drop(skb, sch);
+			return qdisc_drop(skb, sch, to_free);
 		}
 #ifdef CONFIG_NET_CLS_ACT
 	} else if (!cl) {
 		if (ret & __NET_XMIT_BYPASS)
 			qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return ret;
 #endif
-	} else if ((ret = qdisc_enqueue(skb, cl->un.leaf.q)) != NET_XMIT_SUCCESS) {
+	} else if ((ret = qdisc_enqueue(skb, cl->un.leaf.q,
+					to_free)) != NET_XMIT_SUCCESS) {
 		if (net_xmit_drop_count(ret)) {
 			qdisc_qstats_drop(sch);
-			cl->qstats.drops++;
+			cl->drops++;
 		}
 		return ret;
 	} else {
@@ -1108,17 +1111,22 @@ static int
 htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d)
 {
 	struct htb_class *cl = (struct htb_class *)arg;
+	struct gnet_stats_queue qs = {
+		.drops = cl->drops,
+	};
 	__u32 qlen = 0;
 
-	if (!cl->level && cl->un.leaf.q)
+	if (!cl->level && cl->un.leaf.q) {
 		qlen = cl->un.leaf.q->q.qlen;
+		qs.backlog = cl->un.leaf.q->qstats.backlog;
+	}
 	cl->xstats.tokens = PSCHED_NS2TICKS(cl->tokens);
 	cl->xstats.ctokens = PSCHED_NS2TICKS(cl->ctokens);
 
 	if (gnet_stats_copy_basic(qdisc_root_sleeping_running(sch),
 				  d, NULL, &cl->bstats) < 0 ||
 	    gnet_stats_copy_rate_est(d, NULL, &cl->rate_est) < 0 ||
-	    gnet_stats_copy_queue(d, NULL, &cl->qstats, qlen) < 0)
+	    gnet_stats_copy_queue(d, NULL, &qs, qlen) < 0)
 		return -1;
 
 	return gnet_stats_copy_app(d, &cl->xstats, sizeof(cl->xstats));
diff --git a/net/sched/sch_multiq.c b/net/sched/sch_multiq.c
index 5ea93305d705..9ffbb025b37e 100644
--- a/net/sched/sch_multiq.c
+++ b/net/sched/sch_multiq.c
@@ -65,7 +65,8 @@ multiq_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr)
 }
 
 static int
-multiq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+multiq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+	       struct sk_buff **to_free)
 {
 	struct Qdisc *qdisc;
 	int ret;
@@ -76,12 +77,12 @@ multiq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 
 		if (ret & __NET_XMIT_BYPASS)
 			qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return ret;
 	}
 #endif
 
-	ret = qdisc_enqueue(skb, qdisc);
+	ret = qdisc_enqueue(skb, qdisc, to_free);
 	if (ret == NET_XMIT_SUCCESS) {
 		sch->q.qlen++;
 		return NET_XMIT_SUCCESS;
diff --git a/net/sched/sch_netem.c b/net/sched/sch_netem.c
index e271967439bf..ccca8ca4c722 100644
--- a/net/sched/sch_netem.c
+++ b/net/sched/sch_netem.c
@@ -397,7 +397,8 @@ static void tfifo_enqueue(struct sk_buff *nskb, struct Qdisc *sch)
  * when we statistically choose to corrupt one, we instead segment it, returning
  * the first packet to be corrupted, and re-enqueue the remaining frames
  */
-static struct sk_buff *netem_segment(struct sk_buff *skb, struct Qdisc *sch)
+static struct sk_buff *netem_segment(struct sk_buff *skb, struct Qdisc *sch,
+				     struct sk_buff **to_free)
 {
 	struct sk_buff *segs;
 	netdev_features_t features = netif_skb_features(skb);
@@ -405,7 +406,7 @@ static struct sk_buff *netem_segment(struct sk_buff *skb, struct Qdisc *sch)
 	segs = skb_gso_segment(skb, features & ~NETIF_F_GSO_MASK);
 
 	if (IS_ERR_OR_NULL(segs)) {
-		qdisc_drop(skb, sch);
+		qdisc_drop(skb, sch, to_free);
 		return NULL;
 	}
 	consume_skb(skb);
@@ -418,7 +419,8 @@ static struct sk_buff *netem_segment(struct sk_buff *skb, struct Qdisc *sch)
  * 	NET_XMIT_DROP: queue length didn't change.
  *      NET_XMIT_SUCCESS: one skb was queued.
  */
-static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			 struct sk_buff **to_free)
 {
 	struct netem_sched_data *q = qdisc_priv(sch);
 	/* We don't fill cb now as skb_unshare() may invalidate it */
@@ -443,7 +445,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	}
 	if (count == 0) {
 		qdisc_qstats_drop(sch);
-		kfree_skb(skb);
+		__qdisc_drop(skb, to_free);
 		return NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
 	}
 
@@ -463,7 +465,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		u32 dupsave = q->duplicate; /* prevent duplicating a dup... */
 
 		q->duplicate = 0;
-		rootq->enqueue(skb2, rootq);
+		rootq->enqueue(skb2, rootq, to_free);
 		q->duplicate = dupsave;
 	}
 
@@ -475,7 +477,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	 */
 	if (q->corrupt && q->corrupt >= get_crandom(&q->corrupt_cor)) {
 		if (skb_is_gso(skb)) {
-			segs = netem_segment(skb, sch);
+			segs = netem_segment(skb, sch, to_free);
 			if (!segs)
 				return NET_XMIT_DROP;
 		} else {
@@ -488,7 +490,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		if (!(skb = skb_unshare(skb, GFP_ATOMIC)) ||
 		    (skb->ip_summed == CHECKSUM_PARTIAL &&
 		     skb_checksum_help(skb))) {
-			rc = qdisc_drop(skb, sch);
+			rc = qdisc_drop(skb, sch, to_free);
 			goto finish_segs;
 		}
 
@@ -497,7 +499,7 @@ static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	}
 
 	if (unlikely(skb_queue_len(&sch->q) >= sch->limit))
-		return qdisc_drop(skb, sch);
+		return qdisc_drop(skb, sch, to_free);
 
 	qdisc_qstats_backlog_inc(sch, skb);
 
@@ -557,7 +559,7 @@ finish_segs:
 			segs->next = NULL;
 			qdisc_skb_cb(segs)->pkt_len = segs->len;
 			last_len = segs->len;
-			rc = qdisc_enqueue(segs, sch);
+			rc = qdisc_enqueue(segs, sch, to_free);
 			if (rc != NET_XMIT_SUCCESS) {
 				if (net_xmit_drop_count(rc))
 					qdisc_qstats_drop(sch);
@@ -615,8 +617,11 @@ deliver:
 #endif
 
 			if (q->qdisc) {
-				int err = qdisc_enqueue(skb, q->qdisc);
+				struct sk_buff *to_free = NULL;
+				int err;
 
+				err = qdisc_enqueue(skb, q->qdisc, &to_free);
+				kfree_skb_list(to_free);
 				if (unlikely(err != NET_XMIT_SUCCESS)) {
 					if (net_xmit_drop_count(err)) {
 						qdisc_qstats_drop(sch);
diff --git a/net/sched/sch_pie.c b/net/sched/sch_pie.c
index 912a46a5d02e..a570b0bb254c 100644
--- a/net/sched/sch_pie.c
+++ b/net/sched/sch_pie.c
@@ -134,7 +134,8 @@ static bool drop_early(struct Qdisc *sch, u32 packet_size)
 	return false;
 }
 
-static int pie_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int pie_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			     struct sk_buff **to_free)
 {
 	struct pie_sched_data *q = qdisc_priv(sch);
 	bool enqueue = false;
@@ -166,7 +167,7 @@ static int pie_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 
 out:
 	q->stats.dropped++;
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 }
 
 static const struct nla_policy pie_policy[TCA_PIE_MAX + 1] = {
diff --git a/net/sched/sch_plug.c b/net/sched/sch_plug.c
index a12cd37680f8..1c6cbab3e7b9 100644
--- a/net/sched/sch_plug.c
+++ b/net/sched/sch_plug.c
@@ -88,7 +88,8 @@ struct plug_sched_data {
 	u32 pkts_to_release;
 };
 
-static int plug_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int plug_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+			struct sk_buff **to_free)
 {
 	struct plug_sched_data *q = qdisc_priv(sch);
 
@@ -98,7 +99,7 @@ static int plug_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		return qdisc_enqueue_tail(skb, sch);
 	}
 
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 }
 
 static struct sk_buff *plug_dequeue(struct Qdisc *sch)
diff --git a/net/sched/sch_prio.c b/net/sched/sch_prio.c
index de492682caee..f4d443aeae54 100644
--- a/net/sched/sch_prio.c
+++ b/net/sched/sch_prio.c
@@ -67,7 +67,7 @@ prio_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr)
 }
 
 static int
-prio_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+prio_enqueue(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff **to_free)
 {
 	struct Qdisc *qdisc;
 	int ret;
@@ -83,7 +83,7 @@ prio_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	}
 #endif
 
-	ret = qdisc_enqueue(skb, qdisc);
+	ret = qdisc_enqueue(skb, qdisc, to_free);
 	if (ret == NET_XMIT_SUCCESS) {
 		qdisc_qstats_backlog_inc(sch, skb);
 		sch->q.qlen++;
diff --git a/net/sched/sch_qfq.c b/net/sched/sch_qfq.c
index 0427fa8b23f2..f27ffee106f6 100644
--- a/net/sched/sch_qfq.c
+++ b/net/sched/sch_qfq.c
@@ -1217,7 +1217,8 @@ static struct qfq_aggregate *qfq_choose_next_agg(struct qfq_sched *q)
 	return agg;
 }
 
-static int qfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int qfq_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	struct qfq_sched *q = qdisc_priv(sch);
 	struct qfq_class *cl;
@@ -1240,11 +1241,11 @@ static int qfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 				     qdisc_pkt_len(skb));
 		if (err) {
 			cl->qstats.drops++;
-			return qdisc_drop(skb, sch);
+			return qdisc_drop(skb, sch, to_free);
 		}
 	}
 
-	err = qdisc_enqueue(skb, cl->qdisc);
+	err = qdisc_enqueue(skb, cl->qdisc, to_free);
 	if (unlikely(err != NET_XMIT_SUCCESS)) {
 		pr_debug("qfq_enqueue: enqueue failed %d\n", err);
 		if (net_xmit_drop_count(err)) {
diff --git a/net/sched/sch_red.c b/net/sched/sch_red.c
index a0d57530335e..249b2a18acbd 100644
--- a/net/sched/sch_red.c
+++ b/net/sched/sch_red.c
@@ -56,7 +56,8 @@ static inline int red_use_harddrop(struct red_sched_data *q)
 	return q->flags & TC_RED_HARDDROP;
 }
 
-static int red_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int red_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	struct red_sched_data *q = qdisc_priv(sch);
 	struct Qdisc *child = q->qdisc;
@@ -95,7 +96,7 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		break;
 	}
 
-	ret = qdisc_enqueue(skb, child);
+	ret = qdisc_enqueue(skb, child, to_free);
 	if (likely(ret == NET_XMIT_SUCCESS)) {
 		qdisc_qstats_backlog_inc(sch, skb);
 		sch->q.qlen++;
@@ -106,7 +107,7 @@ static int red_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	return ret;
 
 congestion_drop:
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	return NET_XMIT_CN;
 }
 
diff --git a/net/sched/sch_sfb.c b/net/sched/sch_sfb.c
index c69611640fa5..add3cc7d37ec 100644
--- a/net/sched/sch_sfb.c
+++ b/net/sched/sch_sfb.c
@@ -275,7 +275,8 @@ static bool sfb_classify(struct sk_buff *skb, struct tcf_proto *fl,
 	return false;
 }
 
-static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 
 	struct sfb_sched_data *q = qdisc_priv(sch);
@@ -397,7 +398,7 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	}
 
 enqueue:
-	ret = qdisc_enqueue(skb, child);
+	ret = qdisc_enqueue(skb, child, to_free);
 	if (likely(ret == NET_XMIT_SUCCESS)) {
 		sch->q.qlen++;
 		increment_qlen(skb, q);
@@ -408,7 +409,7 @@ enqueue:
 	return ret;
 
 drop:
-	qdisc_drop(skb, sch);
+	qdisc_drop(skb, sch, to_free);
 	return NET_XMIT_CN;
 other_drop:
 	if (ret & __NET_XMIT_BYPASS)
diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c
index 57d118b41cad..7f195ed4d568 100644
--- a/net/sched/sch_sfq.c
+++ b/net/sched/sch_sfq.c
@@ -343,7 +343,7 @@ static int sfq_headdrop(const struct sfq_sched_data *q)
 }
 
 static int
-sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff **to_free)
 {
 	struct sfq_sched_data *q = qdisc_priv(sch);
 	unsigned int hash, dropped;
@@ -367,7 +367,7 @@ sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (x == SFQ_EMPTY_SLOT) {
 		x = q->dep[0].next; /* get a free slot */
 		if (x >= SFQ_MAX_FLOWS)
-			return qdisc_drop(skb, sch);
+			return qdisc_drop(skb, sch, to_free);
 		q->ht[hash] = x;
 		slot = &q->slots[x];
 		slot->hash = hash;
@@ -424,14 +424,14 @@ sfq_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (slot->qlen >= q->maxdepth) {
 congestion_drop:
 		if (!sfq_headdrop(q))
-			return qdisc_drop(skb, sch);
+			return qdisc_drop(skb, sch, to_free);
 
 		/* We know we have at least one packet in queue */
 		head = slot_dequeue_head(slot);
 		delta = qdisc_pkt_len(head) - qdisc_pkt_len(skb);
 		sch->qstats.backlog -= delta;
 		slot->backlog -= delta;
-		qdisc_drop(head, sch);
+		qdisc_drop(head, sch, to_free);
 
 		slot_queue_add(slot, skb);
 		return NET_XMIT_CN;
diff --git a/net/sched/sch_tbf.c b/net/sched/sch_tbf.c
index c12df84d1078..303355c449ab 100644
--- a/net/sched/sch_tbf.c
+++ b/net/sched/sch_tbf.c
@@ -155,7 +155,8 @@ static unsigned int skb_gso_mac_seglen(const struct sk_buff *skb)
 /* GSO packet is too big, segment it so that tbf can transmit
  * each segment in time
  */
-static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch)
+static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	struct tbf_sched_data *q = qdisc_priv(sch);
 	struct sk_buff *segs, *nskb;
@@ -166,7 +167,7 @@ static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch)
 	segs = skb_gso_segment(skb, features & ~NETIF_F_GSO_MASK);
 
 	if (IS_ERR_OR_NULL(segs))
-		return qdisc_drop(skb, sch);
+		return qdisc_drop(skb, sch, to_free);
 
 	nb = 0;
 	while (segs) {
@@ -174,7 +175,7 @@ static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch)
 		segs->next = NULL;
 		qdisc_skb_cb(segs)->pkt_len = segs->len;
 		len += segs->len;
-		ret = qdisc_enqueue(segs, q->qdisc);
+		ret = qdisc_enqueue(segs, q->qdisc, to_free);
 		if (ret != NET_XMIT_SUCCESS) {
 			if (net_xmit_drop_count(ret))
 				qdisc_qstats_drop(sch);
@@ -190,17 +191,18 @@ static int tbf_segment(struct sk_buff *skb, struct Qdisc *sch)
 	return nb > 0 ? NET_XMIT_SUCCESS : NET_XMIT_DROP;
 }
 
-static int tbf_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+static int tbf_enqueue(struct sk_buff *skb, struct Qdisc *sch,
+		       struct sk_buff **to_free)
 {
 	struct tbf_sched_data *q = qdisc_priv(sch);
 	int ret;
 
 	if (qdisc_pkt_len(skb) > q->max_size) {
 		if (skb_is_gso(skb) && skb_gso_mac_seglen(skb) <= q->max_size)
-			return tbf_segment(skb, sch);
-		return qdisc_drop(skb, sch);
+			return tbf_segment(skb, sch, to_free);
+		return qdisc_drop(skb, sch, to_free);
 	}
-	ret = qdisc_enqueue(skb, q->qdisc);
+	ret = qdisc_enqueue(skb, q->qdisc, to_free);
 	if (ret != NET_XMIT_SUCCESS) {
 		if (net_xmit_drop_count(ret))
 			qdisc_qstats_drop(sch);
diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c
index e02687185a59..2cd9b4478b92 100644
--- a/net/sched/sch_teql.c
+++ b/net/sched/sch_teql.c
@@ -77,7 +77,7 @@ struct teql_sched_data {
 /* "teql*" qdisc routines */
 
 static int
-teql_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+teql_enqueue(struct sk_buff *skb, struct Qdisc *sch, struct sk_buff **to_free)
 {
 	struct net_device *dev = qdisc_dev(sch);
 	struct teql_sched_data *q = qdisc_priv(sch);
@@ -87,7 +87,7 @@ teql_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		return NET_XMIT_SUCCESS;
 	}
 
-	return qdisc_drop(skb, sch);
+	return qdisc_drop(skb, sch, to_free);
 }
 
 static struct sk_buff *