path: root/net/sched/sch_pie.c

/* Copyright (C) 2013 Cisco Systems, Inc, 2013.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version 2
 * of the License.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * Author: Vijay Subramanian <vijaynsu@cisco.com>
 * Author: Mythili Prabhu <mysuryan@cisco.com>
 *
 * ECN support is added by Naeem Khademi <naeemk@ifi.uio.no>
 * University of Oslo, Norway.
 */

#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/inet_ecn.h>

#define QUEUE_THRESHOLD 10000
#define DQCOUNT_INVALID -1
#define MAX_PROB  0xffffffff
#define PIE_SCALE 8

/* parameters used */
struct pie_params {
	psched_time_t target;	/* user specified target delay in pschedtime */
	u32 tupdate;		/* timer frequency (in jiffies) */
	u32 limit;		/* number of packets that can be enqueued */
	u32 alpha;		/* alpha and beta are between -4 and 4 */
	u32 beta;		/* and are used for shift relative to 1 */
	bool ecn;		/* true if ecn is enabled */
	bool bytemode;		/* to scale drop early prob based on pkt size */
};

/* variables used */
struct pie_vars {
	u32 prob;		/* probability but scaled by u32 limit. */
	psched_time_t burst_time;
	psched_time_t qdelay;
	psched_time_t qdelay_old;
	u64 dq_count;		/* measured in bytes */
	psched_time_t dq_tstamp;	/* drain rate */
	u32 avg_dq_rate;	/* bytes per pschedtime tick,scaled */
	u32 qlen_old;		/* in bytes */
};

/* statistics gathering */
struct pie_stats {
	u32 packets_in;		/* total number of packets enqueued */
	u32 dropped;		/* packets dropped due to pie_action */
	u32 overlimit;		/* dropped due to lack of space in queue */
	u32 maxq;		/* maximum queue size */
	u32 ecn_mark;		/* packets marked with ECN */
};

/* private data for the Qdisc */
struct pie_sched_data {
	struct pie_params params;
	struct pie_vars vars;
	struct pie_stats stats;
	struct timer_list adapt_timer;
};

static void pie_params_init(struct pie_params *params)
{
	params->alpha = 2;
	params->beta = 20;
	params->tupdate = usecs_to_jiffies(30 * USEC_PER_MSEC);	/* 30 ms */
	params->limit = 1000;	/* default of 1000 packets */
	params->target = PSCHED_NS2TICKS(20 * NSEC_PER_MSEC);	/* 20 ms */
	params->ecn = false;
	params->bytemode = false;
}

static void pie_vars_init(struct pie_vars *vars)
{
	vars->dq_count = DQCOUNT_INVALID;
	vars->avg_dq_rate = 0;
	/* default of 100 ms in pschedtime */
	vars->burst_time = PSCHED_NS2TICKS(100 * NSEC_PER_MSEC);
}

static bool drop_early(struct Qdisc *sch, u32 packet_size)
{
	struct pie_sched_data *q = qdisc_priv(sch);
	u32 rnd;
	u32 local_prob = q->vars.prob;
	u32 mtu = psched_mtu(qdisc_dev(sch));

	/* If there is still burst allowance left skip random early drop */
	if (q->vars.burst_time > 0)
		return false;

	/* If current delay is less than half of target, and
	 * if drop prob is low already, disable early_drop
	 */
	if ((q->vars.qdelay < q->params.target / 2)
	    && (q->vars.prob < MAX_PROB / 5))
		return false;

	/* If we have fewer than 2 mtu-sized packets, disable drop_early,
	 * similar to min_th in RED
	 */
	if (sch->qstats.backlog < 2 * mtu)
		return false;

	/* If bytemode is turned on, use packet size to compute new
	 * probablity. Smaller packets will have lower drop prob in this case
	 */
	if (q->params.bytemode && packet_size <= mtu)
		local_prob = (local_prob / mtu) * packet_size;
	else
		local_prob = q->vars.prob;

	rnd = prandom_u32();
	if (rnd < local_prob)
		return true;

	return false;
}

static int pie_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
	struct pie_sched_data *q = qdisc_priv(sch);
	bool enqueue = false;

	if (unlikely(qdisc_qlen(sch) >= sch->limit)) {
		q->stats.overlimit++;
		goto out;
	}

	if (!drop_early(sch, skb->len)) {
		enqueue = true;
	} else if (q->params.ecn && (q->vars.prob <= MAX_PROB / 10) &&
		   INET_ECN_set_ce(skb)) {
		/* If packet is ecn capable, mark it if drop probability
		 * is lower than 10%, else drop it.
		 */
		q->stats.ecn_mark++;
		enqueue = true;
	}

	/* we can enqueue the packet */
	if (enqueue) {
		q->stats.packets_in++;
		if (qdisc_qlen(sch) > q->stats.maxq)
			q->stats.maxq = qdisc_qlen(sch);

		return qdisc_enqueue_tail(skb, sch);
	}

out:
	q->stats.dropped++;
	return qdisc_drop(skb, sch);
}

static const struct nla_policy pie_policy[TCA_PIE_MAX + 1] = {
	[TCA_PIE_TARGET] = {.type = NLA_U32},
	[TCA_PIE_LIMIT] = {.type = NLA_U32},
	[TCA_PIE_TUPDATE] = {.type = NLA_U32},
	[TCA_PIE_ALPHA] = {.type = NLA_U32},
	[TCA_PIE_BETA] = {.type = NLA_U32},
	[TCA_PIE_ECN] = {.type = NLA_U32},
	[TCA_PIE_BYTEMODE] = {.type = NLA_U32},
};

static int pie_change(struct Qdisc *sch, struct nlattr *opt)
{
	struct pie_sched_data *q = qdisc_priv(sch);
	struct nlattr *tb[TCA_PIE_MAX + 1];
	unsigned int qlen;
	int err;

	if (!opt)
		return -EINVAL;

	err = nla_parse_nested(tb, TCA_PIE_MAX, opt, pie_policy);
	if (err < 0)
		return err;

	sch_tree_lock(sch);

	/* convert from microseconds to pschedtime */
	if (tb[TCA_PIE_TARGET]) {
		/* target is in us */
		u32 target = nla_get_u32(tb[TCA_PIE_TARGET]);

		/* convert to pschedtime */
		q->params.target = PSCHED_NS2TICKS((u64)target * NSEC_PER_USEC);
	}

	/* tupdate is in jiffies */
	if (tb[TCA_PIE_TUPDATE])
		q->params.tupdate = usecs_to_jiffies(nla_get_u32(tb[TCA_PIE_TUPDATE]));

	if (tb[TCA_PIE_LIMIT]) {
		u32 limit = nla_get_u32(tb[TCA_PIE_LIMIT]);

		q->params.limit = limit;
		sch->limit = limit;
	}

	if (tb[TCA_PIE_ALPHA])
		q->params.alpha = nla_get_u32(tb[TCA_PIE_ALPHA]);

	if (tb[TCA_PIE_BETA])
		q->params.beta = nla_get_u32(tb[TCA_PIE_BETA]);

	if (tb[TCA_PIE_ECN])
		q->params.ecn = nla_get_u32(tb[TCA_PIE_ECN]);

	if (tb[TCA_PIE_BYTEMODE])
		q->params.bytemode = nla_get_u32(tb[TCA_PIE_BYTEMODE]);

	/* Drop excess packets if new limit is lower */
	qlen = sch->q.qlen;
	while (sch->q.qlen > sch->limit) {
		struct sk_buff *skb = __skb_dequeue(&sch->q);

		sch->qstats.backlog -= qdisc_pkt_len(