codel: Controlled Delay AQM

An implementation of CoDel AQM, from Kathleen Nichols and Van Jacobson.

http://queue.acm.org/detail.cfm?id=2209336

This AQM main input is no longer queue size in bytes or packets, but the
delay packets stay in (FIFO) queue.

As we don't have infinite memory, we still can drop packets in enqueue()
in case of massive load, but mean of CoDel is to drop packets in
dequeue(), using a control law based on two simple parameters :

target : target sojourn time (default 5ms)
interval : width of moving time window (default 100ms)

Based on initial work from Dave Taht.

Refactored to help future codel inclusion as a plugin for other linux
qdisc (FQ_CODEL, ...), like RED.

include/net/codel.h contains codel algorithm as close as possible than
Kathleen reference.

net/sched/sch_codel.c contains the linux qdisc specific glue.

Separate structures permit a memory efficient implementation of fq_codel
(to be sent as a separate work) : Each flow has its own struct
codel_vars.

timestamps are taken at enqueue() time with 1024 ns precision, allowing
a range of 2199 seconds in queue, and 100Gb links support. iproute2 uses
usec as base unit.

Selected packets are dropped, unless ECN is enabled and packets can get
ECN mark instead.

Tested from 2Mb to 10Gb speeds with no particular problems, on ixgbe and
tg3 drivers (BQL enabled).

Usage: tc qdisc ... codel [ limit PACKETS ] [ target TIME ]
                          [ interval TIME ] [ ecn ]

qdisc codel 10: parent 1:1 limit 2000p target 3.0ms interval 60.0ms ecn
 Sent 13347099587 bytes 8815805 pkt (dropped 0, overlimits 0 requeues 0)
 rate 202365Kbit 16708pps backlog 113550b 75p requeues 0
  count 116 lastcount 98 ldelay 4.3ms dropping drop_next 816us
  maxpacket 1514 ecn_mark 84399 drop_overlimit 0

CoDel must be seen as a base module, and should be used keeping in mind
there is still a FIFO queue. So a typical setup will probably need a
hierarchy of several qdiscs and packet classifiers to be able to meet
whatever constraints a user might have.

One possible example would be to use fq_codel, which combines Fair
Queueing and CoDel, in replacement of sfq / sfq_red.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Dave Taht <dave.taht@bufferbloat.net>
Cc: Kathleen Nichols <nichols@pollere.com>
Cc: Van Jacobson <van@pollere.net>
Cc: Tom Herbert <therbert@google.com>
Cc: Matt Mathis <mattmathis@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

5 files changed, 645 insertions, 0 deletions

diff --git a/include/linux/pkt_sched.h b/include/linux/pkt_sched.h
index ffe975c3f1d8..cde56c22bdab 100644
--- a/include/linux/pkt_sched.h
+++ b/include/linux/pkt_sched.h
@@ -655,4 +655,30 @@ struct tc_qfq_stats {
 	__u32 lmax;
 };
 
+/* CODEL */
+
+enum {
+	TCA_CODEL_UNSPEC,
+	TCA_CODEL_TARGET,
+	TCA_CODEL_LIMIT,
+	TCA_CODEL_INTERVAL,
+	TCA_CODEL_ECN,
+	__TCA_CODEL_MAX
+};
+
+#define TCA_CODEL_MAX	(__TCA_CODEL_MAX - 1)
+
+struct tc_codel_xstats {
+	__u32	maxpacket; /* largest packet we've seen so far */
+	__u32	count;	   /* how many drops we've done since the last time we
+			    * entered dropping state
+			    */
+	__u32	lastcount; /* count at entry to dropping state */
+	__u32	ldelay;    /* in-queue delay seen by most recently dequeued packet */
+	__s32	drop_next; /* time to drop next packet */
+	__u32	drop_overlimit; /* number of time max qdisc packet limit was hit */
+	__u32	ecn_mark;  /* number of packets we ECN marked instead of dropped */
+	__u32	dropping;  /* are we in dropping state ? */
+};
+
 #endif
diff --git a/include/net/codel.h b/include/net/codel.h
new file mode 100644
index 000000000000..bce2cefa8c94
--- /dev/null
+++ b/include/net/codel.h
@@ -0,0 +1,332 @@
+#ifndef __NET_SCHED_CODEL_H
+#define __NET_SCHED_CODEL_H
+
+/*
+ * Codel - The Controlled-Delay Active Queue Management algorithm
+ *
+ *  Copyright (C) 2011-2012 Kathleen Nichols <nichols@pollere.com>
+ *  Copyright (C) 2011-2012 Van Jacobson <van@pollere.net>
+ *  Copyright (C) 2012 Michael D. Taht <dave.taht@bufferbloat.net>
+ *  Copyright (C) 2012 Eric Dumazet <edumazet@google.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions, and the following disclaimer,
+ *    without modification.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. The names of the authors may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission.
+ *
+ * Alternatively, provided that this notice is retained in full, this
+ * software may be distributed under the terms of the GNU General
+ * Public License ("GPL") version 2, in which case the provisions of the
+ * GPL apply INSTEAD OF those given above.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
+ * DAMAGE.
+ *
+ */
+
+#include <linux/types.h>
+#include <linux/ktime.h>
+#include <linux/skbuff.h>
+#include <net/pkt_sched.h>
+#include <net/inet_ecn.h>
+
+/* Controlling Queue Delay (CoDel) algorithm
+ * =========================================
+ * Source : Kathleen Nichols and Van Jacobson
+ * http://queue.acm.org/detail.cfm?id=2209336
+ *
+ * Implemented on linux by Dave Taht and Eric Dumazet
+ */
+
+
+/* CoDel uses a 1024 nsec clock, encoded in u32
+ * This gives a range of 2199 seconds, because of signed compares
+ */
+typedef u32 codel_time_t;
+typedef s32 codel_tdiff_t;
+#define CODEL_SHIFT 10
+#define MS2TIME(a) ((a * NSEC_PER_MSEC) >> CODEL_SHIFT)
+
+static inline codel_time_t codel_get_time(void)
+{
+	u64 ns = ktime_to_ns(ktime_get());
+
+	return ns >> CODEL_SHIFT;
+}
+
+#define codel_time_after(a, b)		((s32)(a) - (s32)(b) > 0)
+#define codel_time_after_eq(a, b)	((s32)(a) - (s32)(b) >= 0)
+#define codel_time_before(a, b)		((s32)(a) - (s32)(b) < 0)
+#define codel_time_before_eq(a, b)	((s32)(a) - (s32)(b) <= 0)
+
+/* Qdiscs using codel plugin must use codel_skb_cb in their own cb[] */
+struct codel_skb_cb {
+	codel_time_t enqueue_time;
+};
+
+static struct codel_skb_cb *get_codel_cb(const struct sk_buff *skb)
+{
+	qdisc_cb_private_validate(skb, sizeof(struct codel_skb_cb));
+	return (struct codel_skb_cb *)qdisc_skb_cb(skb)->data;
+}
+
+static codel_time_t codel_get_enqueue_time(const struct sk_buff *skb)
+{
+	return get_codel_cb(skb)->enqueue_time;
+}
+
+static void codel_set_enqueue_time(struct sk_buff *skb)
+{
+	get_codel_cb(skb)->enqueue_time = codel_get_time();
+}
+
+static inline u32 codel_time_to_us(codel_time_t val)
+{
+	u64 valns = ((u64)val << CODEL_SHIFT);
+
+	do_div(valns, NSEC_PER_USEC);
+	return (u32)valns;
+}
+
+/**
+ * struct codel_params - contains codel parameters
+ * @target:	target queue size (in time units)
+ * @interval:	width of moving time window
+ * @ecn:	is Explicit Congestion Notification enabled
+ */
+struct codel_params {
+	codel_time_t	target;
+	codel_time_t	interval;
+	bool		ecn;
+};
+
+/**
+ * struct codel_vars - contains codel variables
+ * @count:		how many drops we've done since the last time we
+ *			entered dropping state
+ * @lastcount:		count at entry to dropping state
+ * @dropping:		set to true if in dropping state
+ * @first_above_time:	when we went (or will go) continuously above target
+ *			for interval
+ * @drop_next:		time to drop next packet, or when we dropped last
+ * @ldelay:		sojourn time of last dequeued packet
+ */
+struct codel_vars {
+	u32		count;
+	u32		lastcount;
+	bool		dropping;
+	codel_time_t	first_above_time;
+	codel_time_t	drop_next;
+	codel_time_t	ldelay;
+};
+
+/**
+ * struct codel_stats - contains codel shared variables and stats
+ * @maxpacket:	largest packet we've seen so far
+ * @drop_count:	temp count of dropped packets in dequeue()
+ * ecn_mark:	number of packets we ECN marked instead of dropping
+ */
+struct codel_stats {
+	u32		maxpacket;
+	u32		drop_count;
+	u32		ecn_mark;
+};
+
+static void codel_params_init(struct codel_params *params)
+{
+	params->interval = MS2TIME(100);
+	params->target = MS2TIME(5);
+	params->ecn = false;
+}
+
+static void codel_vars_init(struct codel_vars *vars)
+{
+	vars->drop_next = 0;
+	vars->first_above_time = 0;
+	vars->dropping = false; /* exit dropping state */
+	vars->count = 0;
+	vars->lastcount = 0;
+}
+
+static void codel_stats_init(struct codel_stats *stats)
+{
+	stats->maxpacket = 256;
+}
+
+/* return interval/sqrt(x) with good precision
+ * relies on int_sqrt(unsigned long x) kernel implementation
+ */
+static u32 codel_inv_sqrt(u32 _interval, u32 _x)
+{
+	u64 interval = _interval;
+	unsigned long x = _x;
+
+	/* Scale operands for max precision */
+
+#if BITS_PER_LONG == 64
+	x <<= 32; /* On 64bit arches, we can prescale x by 32bits */
+	interval <<= 16;
+#endif
+
+	while (x < (1UL << (BITS_PER_LONG - 2))) {
+		x <<= 2;
+		interval <<= 1;
+	}
+	do_div(interval, int_sqrt(x));
+	return (u32)interval;
+}
+
+static codel_time_t codel_control_law(codel_time_t t,
+				      codel_time_t interval,
+				      u32 count)
+{
+	return t + codel_inv_sqrt(interval, count);
+}
+
+
+static bool codel_should_drop(struct sk_buff *skb,
+			      unsigned int *backlog,
+			      struct codel_vars *vars,
+			      struct codel_params *params,
+			      struct codel_stats *stats,
+			      codel_time_t now)
+{
+	bool ok_to_drop;
+
+	if (!skb) {
+		vars->first_above_time = 0;
+		return false;
+	}
+
+	vars->ldelay = now - codel_get_enqueue_time(skb);
+	*backlog -= qdisc_pkt_len(skb);
+
+	if (unlikely(qdisc_pkt_len(skb) > stats->maxpacket))
+		stats->maxpacket = qdisc_pkt_len(skb);
+
+	if (codel_time_before(vars->ldelay, params->target) ||
+	    *backlog <= stats->maxpacket) {
+		/* went below - stay below for at least interval */
+		vars->first_above_time = 0;
+		return false;
+	}
+	ok_to_drop = false;
+	if (vars->first_above_time == 0) {
+		/* just went above from below. If we stay above
+		 * for at least interval we'll say it's ok to drop
+		 */
+		vars->first_above_time = now + params->interval;
+	} else if (codel_time_after(now, vars->first_above_time)) {
+		ok_to_drop = true;
+	}
+	return ok_to_drop;
+}
+
+typedef struct sk_buff * (*codel_skb_dequeue_t)(struct codel_vars *vars,
+						struct Qdisc *sch);
+
+static struct sk_buff *codel_dequeue(struct Qdisc *sch,
+				     struct codel_params *params,
+				     struct codel_vars *vars,
+				     struct codel_stats *stats,
+				     codel_skb_dequeue_t dequeue_func,
+				     u32 *backlog)
+{
+	struct sk_buff *skb = dequeue_func(vars, sch);
+	codel_time_t now;
+	bool drop;
+
+	if (!skb) {
+		vars->dropping = false;
+		return skb;
+	}
+	now = codel_get_time();
+	drop = codel_should_drop(skb, backlog, vars, params, stats, now);
+	if (vars->dropping) {
+		if (!drop) {
+			/* sojourn time below target - leave dropping state */
+			vars->dropping = false;
+		} else if (codel_time_after_eq(now, vars->drop_next)) {
+			/* It's time for the next drop. Drop the current
+			 * packet and dequeue the next. The dequeue might
+			 * take us out of dropping state.
+			 * If not, schedule the next drop.
+			 * A large backlog might result in drop rates so high
+			 * that the next drop should happen now,
+			 * hence the while loop.
+			 */
+			while (vars->dropping &&
+			       codel_time_after_eq(now, vars->drop_next)) {
+				if (++vars->count == 0) /* avoid zero divides */
+					vars->count = ~0U;
+				if (params->ecn && INET_ECN_set_ce(skb)) {
+					stats->ecn_mark++;
+					vars->drop_next =
+						codel_control_law(vars->drop_next,
+								  params->interval,
+								  vars->count);
+					goto end;
+				}
+				qdisc_drop(skb, sch);
+				stats->drop_count++;
+				skb = dequeue_func(vars, sch);
+				if (!codel_should_drop(skb, backlog,
+						       vars, params, stats, now)) {
+					/* leave dropping state */
+					vars->dropping = false;
+				} else {
+					/* and schedule the next drop */
+					vars->drop_next =
+						codel_control_law(vars->drop_next,
+								  params->interval,
+								  vars->count);
+				}
+			}
+		}
+	} else if (drop) {
+		if (params->ecn && INET_ECN_set_ce(skb)) {
+			stats->ecn_mark++;
+		} else {
+			qdisc_drop(skb, sch);
+			stats->drop_count++;
+
+			skb = dequeue_func(vars, sch);
+			drop = codel_should_drop(skb, backlog, vars, params,
+						 stats, now);
+		}
+		vars->dropping = true;
+		/* if min went above target close to when we last went below it
+		 * assume that the drop rate that controlled the queue on the
+		 * last cycle is a good starting point to control it now.
+		 */
+		if (codel_time_before(now - vars->drop_next,
+				      16 * params->interval)) {
+			vars->count = (vars->count - vars->lastcount) | 1;
+		} else {
+			vars->count = 1;
+		}
+		vars->lastcount = vars->count;
+		vars->drop_next = codel_control_law(now, params->interval,
+						    vars->count);
+	}
+end:
+	return skb;
+}
+#endif
diff --git a/net/sched/Kconfig b/net/sched/Kconfig
index 75b58f81d53d..fadd2522053d 100644
--- a/net/sched/Kconfig
+++ b/net/sched/Kconfig
@@ -250,6 +250,17 @@ config NET_SCH_QFQ
 
 	  If unsure, say N.
 
+config NET_SCH_CODEL
+	tristate "Controlled Delay AQM (CODEL)"
+	help
+	  Say Y here if you want to use the Controlled Delay (CODEL)
+	  packet scheduling algorithm.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called sch_codel.
+
+	  If unsure, say N.
+
 config NET_SCH_INGRESS
 	tristate "Ingress Qdisc"
 	depends on NET_CLS_ACT
diff --git a/net/sched/Makefile b/net/sched/Makefile
index 8cdf4e2b51d3..30fab03b8516 100644
--- a/net/sched/Makefile
+++ b/net/sched/Makefile
@@ -37,6 +37,7 @@ obj-$(CONFIG_NET_SCH_PLUG)	+= sch_plug.o
 obj-$(CONFIG_NET_SCH_MQPRIO)	+= sch_mqprio.o
 obj-$(CONFIG_NET_SCH_CHOKE)	+= sch_choke.o
 obj-$(CONFIG_NET_SCH_QFQ)	+= sch_qfq.o
+obj-$(CONFIG_NET_SCH_CODEL)	+= sch_codel.o
 
 obj-$(CONFIG_NET_CLS_U32)	+= cls_u32.o
 obj-$(CONFIG_NET_CLS_ROUTE4)	+= cls_route.o
diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c
new file mode 100644
index 000000000000..b4a1a81e757e
--- /dev/null
+++ b/net/sched/sch_codel.c
@@ -0,0 +1,275 @@
+/*
+ * Codel - The Controlled-Delay Active Queue Management algorithm
+ *
+ *  Copyright (C) 2011-2012 Kathleen Nichols <nichols@pollere.com>
+ *  Copyright (C) 2011-2012 Van Jacobson <van@pollere.net>
+ *
+ *  Implemented on linux by :
+ *  Copyright (C) 2012 Michael D. Taht <dave.taht@bufferbloat.net>
+ *  Copyright (C) 2012 Eric Dumazet <edumazet@google.com>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions, and the following disclaimer,
+ *    without modification.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. The names of the authors may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission.
+ *
+ * Alternatively, provided that this notice is retained in full, this
+ * software may be distributed under the terms of the GNU General
+ * Public License ("GPL") version 2, in which case the provisions of the
+ * GPL apply INSTEAD OF those given above.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
+ * DAMAGE.
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/skbuff.h>
+#include <net/pkt_sched.h>
+#include <net/codel.h>
+
+
+#define DEFAULT_CODEL_LIMIT 1000
+
+struct codel_sched_data {
+	struct codel_params	params;
+	struct codel_vars	vars;
+	struct codel_stats	stats;
+	u32			drop_overlimit;
+};
+
+/* This is the specific function called from codel_dequeue()
+ * to dequeue a packet from queue. Note: backlog is handled in
+ * codel, we dont need to reduce it here.
+ */
+static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch)
+{
+	struct sk_buff *skb = __skb_dequeue(&sch->q);
+
+	prefetch(&skb->end); /* we'll need skb_shinfo() */
+	return skb;
+}
+
+static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch)
+{
+	struct codel_sched_data *q = qdisc_priv(sch);
+	struct sk_buff *skb;
+
+	skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats,
+			    dequeue, &sch->qstats.backlog);
+	/* We cant call qdisc_tree_decrease_qlen() if our qlen is 0,
+	 * or HTB crashes. Defer it for next round.
+	 */
+	if (q->stats.drop_count && sch->q.qlen) {
+		qdisc_tree_decrease_qlen(sch, q->stats.drop_count);
+		q->stats.drop_count = 0;
+	}
+	if (skb)
+		qdisc_bstats_update(sch, skb);
+	return skb;
+}
+
+static int codel_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
+{
+	struct codel_sched_data *q;
+
+	if (likely(qdisc_qlen(sch) < sch->limit)) {
+		codel_set_enqueue_time(skb);
+		return qdisc_enqueue_tail(skb, sch);
+	}
+	q = qdisc_priv(sch);
+	q->drop_overlimit++;
+	return qdisc_drop(skb, sch);
+}
+
+static const struct nla_policy codel_policy[TCA_CODEL_MAX + 1] = {
+	[TCA_CODEL_TARGET]	= { .type = NLA_U32 },
+	[TCA_CODEL_LIMIT]	= { .type = NLA_U32 },
+	[TCA_CODEL_INTERVAL]	= { .type = NLA_U32 },
+	[TCA_CODEL_ECN]		= { .type = NLA_U32 },
+};
+
+static int codel_change(struct Qdisc *sch, struct nlattr *opt)
+{
+	struct codel_sched_data *q = qdisc_priv(sch);
+	struct nlattr *tb[TCA_CODEL_MAX + 1];
+	unsigned int qlen;
+	int err;
+
+	if (!opt)
+		return -EINVAL;
+
+	err = nla_parse_nested(tb, TCA_CODEL_MAX, opt, codel_policy);
+	if (err < 0)
+		return err;
+
+	sch_tree_lock(sch);
+
+	if (tb[TCA_CODEL_TARGET]) {
+		u32 target = nla_get_u32(tb[TCA_CODEL_TARGET]);
+
+		q->params.target = ((u64)target * NSEC_PER_USEC) >> CODEL_SHIFT;
+	}
+
+	if (tb[TCA_CODEL_INTERVAL]) {
+		u32 interval = nla_get_u32(tb[TCA_CODEL_INTERVAL]);
+
+		q->params.interval = ((u64)interval * NSEC_PER_USEC) >> CODEL_SHIFT;
+	}
+
+	if (tb[TCA_CODEL_LIMIT])
+		sch->limit = nla_get_u32(tb[TCA_CODEL_LIMIT]);
+
+	if (tb[TCA_CODEL_ECN])
+		q->params.ecn = !!nla_get_u32(tb[TCA_CODEL_ECN]);
+
+	qlen = sch->q.qlen;
+	while (sch->q.qlen > sch->limit) {
+		struct sk_buff *skb = __skb_dequeue(&sch->q);
+
+		sch->qstats.backlog -= qdisc_pkt_len(skb);
+		qdisc_drop(skb, sch);
+	}
+	qdisc_tree_decrease_qlen(sch, qlen - sch->q.qlen);
+
+	sch_tree_unlock(sch);
+	return 0;
+}
+
+static int codel_init(struct Qdisc *sch, struct nlattr *opt)
+{
+	struct codel_sched_data *q = qdisc_priv(sch);
+
+	sch->limit = DEFAULT_CODEL_LIMIT;
+
+	codel_params_init(&q->params);
+	codel_vars_init(&q->vars);
+	codel_stats_init(&q->stats);
+
+	if (opt) {
+		int err = codel_change(sch, opt);
+
+		if (err)
+			return err;
+	}
+
+	if (sch->limit >= 1)
+		sch->flags |= TCQ_F_CAN_BYPASS;
+	else
+		sch->flags &= ~TCQ_F_CAN_BYPASS;
+
+	return 0;
+}
+
+static int codel_dump(struct Qdisc *sch, struct sk_buff *skb)
+{
+	struct codel_sched_data *q = qdisc_priv(sch);
+	struct nlattr *opts;
+
+	opts = nla_nest_start(skb, TCA_OPTIONS);
+	if (opts == NULL)
+		goto nla_put_failure;
+
+	if (nla_put_u32(skb, TCA_CODEL_TARGET,
+			codel_time_to_us(q->params.target)) ||
+	    nla_put_u32(skb, TCA_CODEL_LIMIT,
+			sch->limit) ||
+	    nla_put_u32(skb, TCA_CODEL_INTERVAL,
+			codel_time_to_us(q->params.interval)) ||
+	    nla_put_u32(skb, TCA_CODEL_ECN,
+			q->params.ecn))
+		goto nla_put_failure;
+
+	return nla_nest_end(skb, opts);
+
+nla_put_failure:
+	nla_nest_cancel(skb, opts);
+	return -1;
+}
+
+static int codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
+{
+	const struct codel_sched_data *q = qdisc_priv(sch);
+	struct tc_codel_xstats st = {
+		.maxpacket	= q->stats.maxpacket,
+		.count		= q->vars.count,
+		.lastcount	= q->vars.lastcount,
+		.drop_overlimit = q->drop_overlimit,
+		.ldelay		= codel_time_to_us(q->vars.ldelay),
+		.dropping	= q->vars.dropping,
+		.ecn_mark	= q->stats.ecn_mark,
+	};
+
+	if (q->vars.dropping) {
+		codel_tdiff_t delta = q->vars.drop_next - codel_get_time();
+
+		if (delta >= 0)
+			st.drop_next = codel_time_to_us(delta);
+		else
+			st.drop_next = -codel_time_to_us(-delta);
+	}
+
+	return gnet_stats_copy_app(d, &st, sizeof(st));
+}
+
+static void codel_reset(struct Qdisc *sch)
+{
+	struct codel_sched_data *q = qdisc_priv(sch);
+
+	qdisc_reset_queue(sch);
+	codel_vars_init(&q->vars);
+}
+
+static struct Qdisc_ops codel_qdisc_ops __read_mostly = {
+	.id		=	"codel",
+	.priv_size	=	sizeof(struct codel_sched_data),
+
+	.enqueue	=	codel_qdisc_enqueue,
+	.dequeue	=	codel_qdisc_dequeue,
+	.peek		=	qdisc_peek_dequeued,
+	.init		=	codel_init,
+	.reset		=	codel_reset,
+	.change 	=	codel_change,
+	.dump		=	codel_dump,
+	.dump_stats	=	codel_dump_stats,
+	.owner		=	THIS_MODULE,
+};
+
+static int __init codel_module_init(void)
+{
+	return register_qdisc(&codel_qdisc_ops);
+}
+
+static void __exit codel_module_exit(void)
+{
+	unregister_qdisc(&codel_qdisc_ops);
+}
+
+module_init(codel_module_init)
+module_exit(codel_module_exit)
+
+MODULE_DESCRIPTION("Controlled Delay queue discipline");
+MODULE_AUTHOR("Dave Taht");
+MODULE_AUTHOR("Eric Dumazet");
+MODULE_LICENSE("Dual BSD/GPL");