Release 4.11 net/ipv4/tcp_cong.c

Directory: net/ipv4
/*
 * Pluggable TCP congestion control support and newReno
 * congestion control.
 * Based on ideas from I/O scheduler support and Web100.
 *
 * Copyright (C) 2005 Stephen Hemminger <shemminger@osdl.org>
 */


#define pr_fmt(fmt) "TCP: " fmt

#include <linux/module.h>
#include <linux/mm.h>
#include <linux/types.h>
#include <linux/list.h>
#include <linux/gfp.h>
#include <linux/jhash.h>
#include <net/tcp.h>

static DEFINE_SPINLOCK(tcp_cong_list_lock);
static LIST_HEAD(tcp_cong_list);

/* Simple linear search, don't expect many entries! */


static struct tcp_congestion_ops *tcp_ca_find(const char *name)
{
	struct tcp_congestion_ops *e;

	list_for_each_entry_rcu(e, &tcp_cong_list, list) {
		if (strcmp(e->name, name) == 0)
			return e;
	}

	return NULL;
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 47 100.00% 2 100.00%
Total 47 100.00% 2 100.00%

/* Must be called with rcu lock held */


static const struct tcp_congestion_ops *__tcp_ca_find_autoload(const char *name)
{
	const struct tcp_congestion_ops *ca = tcp_ca_find(name);
#ifdef CONFIG_MODULES
	if (!ca && capable(CAP_NET_ADMIN)) {
		rcu_read_unlock();
		request_module("tcp_%s", name);
		rcu_read_lock();
		ca = tcp_ca_find(name);
	}
#endif
	return ca;
}

Contributors
Person Tokens Prop Commits CommitProp
Daniel Borkmann 65 100.00% 1 100.00%
Total 65 100.00% 1 100.00%

/* Simple linear search, not much in here. */


struct tcp_congestion_ops *tcp_ca_find_key(u32 key)
{
	struct tcp_congestion_ops *e;

	list_for_each_entry_rcu(e, &tcp_cong_list, list) {
		if (e->key == key)
			return e;
	}

	return NULL;
}

Contributors
Person Tokens Prop Commits CommitProp
Daniel Borkmann 39 100.00% 1 100.00%
Total 39 100.00% 1 100.00%

/*
 * Attach new congestion control algorithm to the list
 * of available options.
 */


int tcp_register_congestion_control(struct tcp_congestion_ops *ca)
{
	int ret = 0;

	/* all algorithms must implement these */
	if (!ca->ssthresh || !ca->undo_cwnd ||
	    !(ca->cong_avoid || ca->cong_control)) {
		pr_err("%s does not implement required ops\n", ca->name);
		return -EINVAL;
	}

	ca->key = jhash(ca->name, sizeof(ca->name), strlen(ca->name));

	spin_lock(&tcp_cong_list_lock);
	if (ca->key == TCP_CA_UNSPEC || tcp_ca_find_key(ca->key)) {
		pr_notice("%s already registered or non-unique key\n",
			  ca->name);
		ret = -EEXIST;
	} else {
		list_add_tail_rcu(&ca->list, &tcp_cong_list);
		pr_debug("%s registered\n", ca->name);
	}
	spin_unlock(&tcp_cong_list_lock);

	return ret;
}
Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 98 66.22% 3 37.50%
Daniel Borkmann 34 22.97% 1 12.50%
Florian Westphal 6 4.05% 1 12.50%
Yuchung Cheng 6 4.05% 1 12.50%
Joe Perches 4 2.70% 2 25.00%
Total 148 100.00% 8 100.00%


EXPORT_SYMBOL_GPL(tcp_register_congestion_control);

/*
 * Remove congestion control algorithm, called from
 * the module's remove function.  Module ref counts are used
 * to ensure that this can't be done till all sockets using
 * that method are closed.
 */


void tcp_unregister_congestion_control(struct tcp_congestion_ops *ca)
{
	spin_lock(&tcp_cong_list_lock);
	list_del_rcu(&ca->list);
	spin_unlock(&tcp_cong_list_lock);

	/* Wait for outstanding readers to complete before the
         * module gets removed entirely.
         *
         * A try_module_get() should fail by now as our module is
         * in "going" state since no refs are held anymore and
         * module_exit() handler being called.
         */
	synchronize_rcu();
}
Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 30 88.24% 1 50.00%
Daniel Borkmann 4 11.76% 1 50.00%
Total 34 100.00% 2 100.00%


EXPORT_SYMBOL_GPL(tcp_unregister_congestion_control);



u32 tcp_ca_get_key_by_name(const char *name, bool *ecn_ca)
{
	const struct tcp_congestion_ops *ca;
	u32 key = TCP_CA_UNSPEC;

	might_sleep();

	rcu_read_lock();
	ca = __tcp_ca_find_autoload(name);
	if (ca) {
		key = ca->key;
		*ecn_ca = ca->flags & TCP_CONG_NEEDS_ECN;
	}
	rcu_read_unlock();

	return key;
}
Contributors
Person Tokens Prop Commits CommitProp
Daniel Borkmann 65 100.00% 2 100.00%
Total 65 100.00% 2 100.00%


EXPORT_SYMBOL_GPL(tcp_ca_get_key_by_name);



char *tcp_ca_get_name_by_key(u32 key, char *buffer)
{
	const struct tcp_congestion_ops *ca;
	char *ret = NULL;

	rcu_read_lock();
	ca = tcp_ca_find_key(key);
	if (ca)
		ret = strncpy(buffer, ca->name,
			      TCP_CA_NAME_MAX);
	rcu_read_unlock();

	return ret;
}
Contributors
Person Tokens Prop Commits CommitProp
Daniel Borkmann 58 100.00% 1 100.00%
Total 58 100.00% 1 100.00%


EXPORT_SYMBOL_GPL(tcp_ca_get_name_by_key);

/* Assign choice of congestion control. */


void tcp_assign_congestion_control(struct sock *sk)
{
	struct inet_connection_sock *icsk = inet_csk(sk);
	struct tcp_congestion_ops *ca;

	rcu_read_lock();
	list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
		if (likely(try_module_get(ca->owner))) {
			icsk->icsk_ca_ops = ca;
			goto out;
		}
		/* Fallback to next available. The last really
                 * guaranteed fallback is Reno from this list.
                 */
	}
out:
	rcu_read_unlock();
	memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));

	if (ca->flags & TCP_CONG_NEEDS_ECN)
		INET_ECN_xmit(sk);
	else
		INET_ECN_dontxmit(sk);
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 43 42.16% 2 40.00%
Florian Westphal 26 25.49% 1 20.00%
Eric Dumazet 19 18.63% 1 20.00%
Arnaldo Carvalho de Melo 14 13.73% 1 20.00%
Total 102 100.00% 5 100.00%



void tcp_init_congestion_control(struct sock *sk)
{
	const struct inet_connection_sock *icsk = inet_csk(sk);

	if (icsk->icsk_ca_ops->init)
		icsk->icsk_ca_ops->init(sk);
	if (tcp_ca_needs_ecn(sk))
		INET_ECN_xmit(sk);
	else
		INET_ECN_dontxmit(sk);
}

Contributors
Person Tokens Prop Commits CommitProp
Florian Westphal 20 35.71% 1 25.00%
Eric Dumazet 18 32.14% 1 25.00%
Stephen Hemminger 13 23.21% 1 25.00%
Arnaldo Carvalho de Melo 5 8.93% 1 25.00%
Total 56 100.00% 4 100.00%



static void tcp_reinit_congestion_control(struct sock *sk,
					  const struct tcp_congestion_ops *ca)
{
	struct inet_connection_sock *icsk = inet_csk(sk);

	tcp_cleanup_congestion_control(sk);
	icsk->icsk_ca_ops = ca;
	icsk->icsk_ca_setsockopt = 1;
	memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));

	if (sk->sk_state != TCP_CLOSE)
		tcp_init_congestion_control(sk);
}

Contributors
Person Tokens Prop Commits CommitProp
Daniel Borkmann 42 57.53% 1 20.00%
Florian Westphal 16 21.92% 1 20.00%
Wei Wang 8 10.96% 1 20.00%
Neal Cardwell 6 8.22% 1 20.00%
Eric Dumazet 1 1.37% 1 20.00%
Total 73 100.00% 5 100.00%

/* Manage refcounts on socket close. */


void tcp_cleanup_congestion_control(struct sock *sk)
{
	struct inet_connection_sock *icsk = inet_csk(sk);

	if (icsk->icsk_ca_ops->release)
		icsk->icsk_ca_ops->release(sk);
	module_put(icsk->icsk_ca_ops->owner);
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 27 58.70% 1 50.00%
Arnaldo Carvalho de Melo 19 41.30% 1 50.00%
Total 46 100.00% 2 100.00%

/* Used by sysctl to change default congestion control */


int tcp_set_default_congestion_control(const char *name)
{
	struct tcp_congestion_ops *ca;
	int ret = -ENOENT;

	spin_lock(&tcp_cong_list_lock);
	ca = tcp_ca_find(name);
#ifdef CONFIG_MODULES
	if (!ca && capable(CAP_NET_ADMIN)) {
		spin_unlock(&tcp_cong_list_lock);

		request_module("tcp_%s", name);
		spin_lock(&tcp_cong_list_lock);
		ca = tcp_ca_find(name);
	}
#endif

	if (ca) {
		ca->flags |= TCP_CONG_NON_RESTRICTED;	/* default is always allowed */
		list_move(&ca->list, &tcp_cong_list);
		ret = 0;
	}
	spin_unlock(&tcp_cong_list_lock);

	return ret;
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 112 98.25% 4 66.67%
Eric Paris 1 0.88% 1 16.67%
Johannes Berg 1 0.88% 1 16.67%
Total 114 100.00% 6 100.00%

/* Set default value from kernel configuration at bootup */


static int __init tcp_congestion_default(void)
{
	return tcp_set_default_congestion_control(CONFIG_DEFAULT_TCP_CONG);
}
Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 15 100.00% 1 100.00%
Total 15 100.00% 1 100.00%


late_initcall(tcp_congestion_default);

/* Build string with list of available congestion control values */


void tcp_get_available_congestion_control(char *buf, size_t maxlen)
{
	struct tcp_congestion_ops *ca;
	size_t offs = 0;

	rcu_read_lock();
	list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
		offs += snprintf(buf + offs, maxlen - offs,
				 "%s%s",
				 offs == 0 ? "" : " ", ca->name);
	}
	rcu_read_unlock();
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 65 100.00% 1 100.00%
Total 65 100.00% 1 100.00%

/* Get current default congestion control */


void tcp_get_default_congestion_control(char *name)
{
	struct tcp_congestion_ops *ca;
	/* We will always have reno... */
	BUG_ON(list_empty(&tcp_cong_list));

	rcu_read_lock();
	ca = list_entry(tcp_cong_list.next, struct tcp_congestion_ops, list);
	strncpy(name, ca->name, TCP_CA_NAME_MAX);
	rcu_read_unlock();
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 55 100.00% 1 100.00%
Total 55 100.00% 1 100.00%

/* Built list of non-restricted congestion control values */


void tcp_get_allowed_congestion_control(char *buf, size_t maxlen)
{
	struct tcp_congestion_ops *ca;
	size_t offs = 0;

	*buf = '\0';
	rcu_read_lock();
	list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
		if (!(ca->flags & TCP_CONG_NON_RESTRICTED))
			continue;
		offs += snprintf(buf + offs, maxlen - offs,
				 "%s%s",
				 offs == 0 ? "" : " ", ca->name);
	}
	rcu_read_unlock();
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 82 100.00% 2 100.00%
Total 82 100.00% 2 100.00%

/* Change list of non-restricted congestion control */


int tcp_set_allowed_congestion_control(char *val)
{
	struct tcp_congestion_ops *ca;
	char *saved_clone, *clone, *name;
	int ret = 0;

	saved_clone = clone = kstrdup(val, GFP_USER);
	if (!clone)
		return -ENOMEM;

	spin_lock(&tcp_cong_list_lock);
	/* pass 1 check for bad entries */
	while ((name = strsep(&clone, " ")) && *name) {
		ca = tcp_ca_find(name);
		if (!ca) {
			ret = -ENOENT;
			goto out;
		}
	}

	/* pass 2 clear old values */
	list_for_each_entry_rcu(ca, &tcp_cong_list, list)
		ca->flags &= ~TCP_CONG_NON_RESTRICTED;

	/* pass 3 mark as allowed */
	while ((name = strsep(&val, " ")) && *name) {
		ca = tcp_ca_find(name);
		WARN_ON(!ca);
		if (ca)
			ca->flags |= TCP_CONG_NON_RESTRICTED;
	}
out:
	spin_unlock(&tcp_cong_list_lock);
	kfree(saved_clone);

	return ret;
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 162 94.19% 2 66.67%
Julia Lawall 10 5.81% 1 33.33%
Total 172 100.00% 3 100.00%

/* Change congestion control for socket */


int tcp_set_congestion_control(struct sock *sk, const char *name)
{
	struct inet_connection_sock *icsk = inet_csk(sk);
	const struct tcp_congestion_ops *ca;
	int err = 0;

	if (icsk->icsk_ca_dst_locked)
		return -EPERM;

	rcu_read_lock();
	ca = __tcp_ca_find_autoload(name);
	/* No change asking for existing value */
	if (ca == icsk->icsk_ca_ops) {
		icsk->icsk_ca_setsockopt = 1;
		goto out;
	}
	if (!ca)
		err = -ENOENT;
	else if (!((ca->flags & TCP_CONG_NON_RESTRICTED) ||
		   ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)))
		err = -EPERM;
	else if (!try_module_get(ca->owner))
		err = -EBUSY;
	else
		tcp_reinit_congestion_control(sk, ca);
 out:
	rcu_read_unlock();
	return err;
}

Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 103 69.13% 3 37.50%
Arnaldo Carvalho de Melo 15 10.07% 1 12.50%
Daniel Borkmann 15 10.07% 2 25.00%
Eric W. Biedermann 8 5.37% 1 12.50%
Neal Cardwell 8 5.37% 1 12.50%
Total 149 100.00% 8 100.00%

/* Slow start is used when congestion window is no greater than the slow start
 * threshold. We base on RFC2581 and also handle stretch ACKs properly.
 * We do not implement RFC3465 Appropriate Byte Counting (ABC) per se but
 * something better;) a packet is only considered (s)acked in its entirety to
 * defend the ACK attacks described in the RFC. Slow start processes a stretch
 * ACK of degree N as if N acks of degree 1 are received back to back except
 * ABC caps N to 2. Slow start exits when cwnd grows over ssthresh and
 * returns the leftover acks to adjust cwnd in congestion avoidance mode.
 */


u32 tcp_slow_start(struct tcp_sock *tp, u32 acked)
{
	u32 cwnd = min(tp->snd_cwnd + acked, tp->snd_ssthresh);

	acked -= cwnd - tp->snd_cwnd;
	tp->snd_cwnd = min(cwnd, tp->snd_cwnd_clamp);

	return acked;
}
Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 15 28.30% 1 14.29%
Yuchung Cheng 13 24.53% 2 28.57%
Neal Cardwell 12 22.64% 1 14.29%
Eric Dumazet 11 20.75% 2 28.57%
John Heffner 2 3.77% 1 14.29%
Total 53 100.00% 7 100.00%


EXPORT_SYMBOL_GPL(tcp_slow_start);

/* In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd (or alternative w),
 * for every packet that was ACKed.
 */


void tcp_cong_avoid_ai(struct tcp_sock *tp, u32 w, u32 acked)
{
	/* If credits accumulated at a higher w, apply them gently now. */
	if (tp->snd_cwnd_cnt >= w) {
		tp->snd_cwnd_cnt = 0;
		tp->snd_cwnd++;
	}

	tp->snd_cwnd_cnt += acked;
	if (tp->snd_cwnd_cnt >= w) {
		u32 delta = tp->snd_cwnd_cnt / w;

		tp->snd_cwnd_cnt -= delta * w;
		tp->snd_cwnd += delta;
	}
	tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_cwnd_clamp);
}
Contributors
Person Tokens Prop Commits CommitProp
Neal Cardwell 56 60.87% 3 75.00%
Ilpo Järvinen 36 39.13% 1 25.00%
Total 92 100.00% 4 100.00%


EXPORT_SYMBOL_GPL(tcp_cong_avoid_ai);

/*
 * TCP Reno congestion control
 * This is special case used for fallback as well.
 */
/* This is Jacobson's slow start and congestion avoidance.
 * SIGCOMM '88, p. 328.
 */


void tcp_reno_cong_avoid(struct sock *sk, u32 ack, u32 acked)
{
	struct tcp_sock *tp = tcp_sk(sk);

	if (!tcp_is_cwnd_limited(sk))
		return;

	/* In "safe" area, increase. */
	if (tcp_in_slow_start(tp)) {
		acked = tcp_slow_start(tp, acked);
		if (!acked)
			return;
	}
	/* In dangerous area, increase slowly. */
	tcp_cong_avoid_ai(tp, tp->snd_cwnd, acked);
}
Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 38 52.78% 4 40.00%
Neal Cardwell 12 16.67% 2 20.00%
Arnaldo Carvalho de Melo 12 16.67% 1 10.00%
Yuchung Cheng 6 8.33% 2 20.00%
Ilpo Järvinen 4 5.56% 1 10.00%
Total 72 100.00% 10 100.00%


EXPORT_SYMBOL_GPL(tcp_reno_cong_avoid);

/* Slow start threshold is half the congestion window (min 2) */


u32 tcp_reno_ssthresh(struct sock *sk)
{
	const struct tcp_sock *tp = tcp_sk(sk);

	return max(tp->snd_cwnd >> 1U, 2U);
}
Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 20 60.61% 1 50.00%
Arnaldo Carvalho de Melo 13 39.39% 1 50.00%
Total 33 100.00% 2 100.00%


EXPORT_SYMBOL_GPL(tcp_reno_ssthresh);



u32 tcp_reno_undo_cwnd(struct sock *sk)
{
	const struct tcp_sock *tp = tcp_sk(sk);

	return max(tp->snd_cwnd, tp->snd_ssthresh << 1);
}
Contributors
Person Tokens Prop Commits CommitProp
Florian Westphal 35 100.00% 1 100.00%
Total 35 100.00% 1 100.00%


EXPORT_SYMBOL_GPL(tcp_reno_undo_cwnd);


struct tcp_congestion_ops tcp_reno = {
	.flags		= TCP_CONG_NON_RESTRICTED,
	.name		= "reno",
	.owner		= THIS_MODULE,
	.ssthresh	= tcp_reno_ssthresh,
	.cong_avoid	= tcp_reno_cong_avoid,
	.undo_cwnd	= tcp_reno_undo_cwnd,
};
Overall Contributors
Person Tokens Prop Commits CommitProp
Stephen Hemminger 1027 56.55% 14 31.11%
Daniel Borkmann 337 18.56% 3 6.67%
Florian Westphal 113 6.22% 3 6.67%
Neal Cardwell 95 5.23% 5 11.11%
Arnaldo Carvalho de Melo 78 4.30% 1 2.22%
Eric Dumazet 49 2.70% 3 6.67%
Ilpo Järvinen 45 2.48% 1 2.22%
Yuchung Cheng 26 1.43% 4 8.89%
Joe Perches 11 0.61% 2 4.44%
Julia Lawall 10 0.55% 1 2.22%
Wei Wang 8 0.44% 1 2.22%
Eric W. Biedermann 8 0.44% 1 2.22%
Tejun Heo 3 0.17% 1 2.22%
John Heffner 2 0.11% 1 2.22%
Fabian Frederick 1 0.06% 1 2.22%
Robert P. J. Day 1 0.06% 1 2.22%
Eric Paris 1 0.06% 1 2.22%
Johannes Berg 1 0.06% 1 2.22%
Total 1816 100.00% 45 100.00%
Directory: net/ipv4

Information contained on this website is for historical information purposes only and does not indicate or represent copyright ownership.
Created with cregit.
Person	Tokens	Prop	Commits	CommitProp
Stephen Hemminger	47	100.00%	2	100.00%
Total	47	100.00%	2	100.00%
cregit-Linux how code gets into the kernel

Release 4.11 net/ipv4/tcp_cong.c

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Contributors

Overall Contributors
