[mirror_ubuntu-jammy-kernel.git] / net / mptcp / pm.c

// SPDX-License-Identifier: GPL-2.0
/* Multipath TCP
 *
 * Copyright (c) 2019, Intel Corporation.
 */
#define pr_fmt(fmt) "MPTCP: " fmt

#include <linux/kernel.h>
#include <net/tcp.h>
#include <net/mptcp.h>
#include "protocol.h"

/* path manager command handlers */

int mptcp_pm_announce_addr(struct mptcp_sock *msk,
			   const struct mptcp_addr_info *addr,
			   bool echo, bool port)
{
	u8 add_addr = READ_ONCE(msk->pm.addr_signal);

	pr_debug("msk=%p, local_id=%d", msk, addr->id);

	if (add_addr) {
		pr_warn("addr_signal error, add_addr=%d", add_addr);
		return -EINVAL;
	}

	msk->pm.local = *addr;
	add_addr |= BIT(MPTCP_ADD_ADDR_SIGNAL);
	if (echo)
		add_addr |= BIT(MPTCP_ADD_ADDR_ECHO);
	if (addr->family == AF_INET6)
		add_addr |= BIT(MPTCP_ADD_ADDR_IPV6);
	if (port)
		add_addr |= BIT(MPTCP_ADD_ADDR_PORT);
	WRITE_ONCE(msk->pm.addr_signal, add_addr);
	return 0;
}

int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id)
{
	u8 rm_addr = READ_ONCE(msk->pm.addr_signal);

	pr_debug("msk=%p, local_id=%d", msk, local_id);

	if (rm_addr) {
		pr_warn("addr_signal error, rm_addr=%d", rm_addr);
		return -EINVAL;
	}

	msk->pm.rm_id = local_id;
	rm_addr |= BIT(MPTCP_RM_ADDR_SIGNAL);
	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
	return 0;
}

int mptcp_pm_remove_subflow(struct mptcp_sock *msk, u8 local_id)
{
	pr_debug("msk=%p, local_id=%d", msk, local_id);

	spin_lock_bh(&msk->pm.lock);
	mptcp_pm_nl_rm_subflow_received(msk, local_id);
	spin_unlock_bh(&msk->pm.lock);
	return 0;
}

/* path manager event handlers */

void mptcp_pm_new_connection(struct mptcp_sock *msk, int server_side)
{
	struct mptcp_pm_data *pm = &msk->pm;

	pr_debug("msk=%p, token=%u side=%d", msk, msk->token, server_side);

	WRITE_ONCE(pm->server_side, server_side);
}

bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
{
	struct mptcp_pm_data *pm = &msk->pm;
	unsigned int subflows_max;
	int ret = 0;

	subflows_max = mptcp_pm_get_subflows_max(msk);

	pr_debug("msk=%p subflows=%d max=%d allow=%d", msk, pm->subflows,
		 subflows_max, READ_ONCE(pm->accept_subflow));

	/* try to avoid acquiring the lock below */
	if (!READ_ONCE(pm->accept_subflow))
		return false;

	spin_lock_bh(&pm->lock);
	if (READ_ONCE(pm->accept_subflow)) {
		ret = pm->subflows < subflows_max;
		if (ret && ++pm->subflows == subflows_max)
			WRITE_ONCE(pm->accept_subflow, false);
	}
	spin_unlock_bh(&pm->lock);

	return ret;
}

/* return true if the new status bit is currently cleared, that is, this event
 * can be server, eventually by an already scheduled work
 */
static bool mptcp_pm_schedule_work(struct mptcp_sock *msk,
				   enum mptcp_pm_status new_status)
{
	pr_debug("msk=%p status=%x new=%lx", msk, msk->pm.status,
		 BIT(new_status));
	if (msk->pm.status & BIT(new_status))
		return false;

	msk->pm.status |= BIT(new_status);
	mptcp_schedule_work((struct sock *)msk);
	return true;
}

void mptcp_pm_fully_established(struct mptcp_sock *msk)
{
	struct mptcp_pm_data *pm = &msk->pm;

	pr_debug("msk=%p", msk);

	/* try to avoid acquiring the lock below */
	if (!READ_ONCE(pm->work_pending))
		return;

	spin_lock_bh(&pm->lock);

	/* mptcp_pm_fully_established() can be invoked by multiple
	 * racing paths - accept() and check_fully_established()
	 * be sure to serve this event only once.
	 */
	if (READ_ONCE(pm->work_pending) &&
	    !(msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)))
		mptcp_pm_schedule_work(msk, MPTCP_PM_ESTABLISHED);
	msk->pm.status |= BIT(MPTCP_PM_ALREADY_ESTABLISHED);

	spin_unlock_bh(&pm->lock);
}

void mptcp_pm_connection_closed(struct mptcp_sock *msk)
{
	pr_debug("msk=%p", msk);
}

void mptcp_pm_subflow_established(struct mptcp_sock *msk,
				  struct mptcp_subflow_context *subflow)
{
	struct mptcp_pm_data *pm = &msk->pm;

	pr_debug("msk=%p", msk);

	if (!READ_ONCE(pm->work_pending))
		return;

	spin_lock_bh(&pm->lock);

	if (READ_ONCE(pm->work_pending))
		mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);

	spin_unlock_bh(&pm->lock);
}

void mptcp_pm_subflow_closed(struct mptcp_sock *msk, u8 id)
{
	pr_debug("msk=%p", msk);
}

void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
				const struct mptcp_addr_info *addr)
{
	struct mptcp_pm_data *pm = &msk->pm;

	pr_debug("msk=%p remote_id=%d accept=%d", msk, addr->id,
		 READ_ONCE(pm->accept_addr));

	spin_lock_bh(&pm->lock);

	if (!READ_ONCE(pm->accept_addr)) {
		mptcp_pm_announce_addr(msk, addr, true, addr->port);
		mptcp_pm_add_addr_send_ack(msk);
	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
		pm->remote = *addr;
	}

	spin_unlock_bh(&pm->lock);
}

void mptcp_pm_add_addr_send_ack(struct mptcp_sock *msk)
{
	if (!mptcp_pm_should_add_signal(msk))
		return;

	mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_SEND_ACK);
}

void mptcp_pm_rm_addr_received(struct mptcp_sock *msk, u8 rm_id)
{
	struct mptcp_pm_data *pm = &msk->pm;

	pr_debug("msk=%p remote_id=%d", msk, rm_id);

	spin_lock_bh(&pm->lock);
	mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED);
	pm->rm_id = rm_id;
	spin_unlock_bh(&pm->lock);
}

void mptcp_pm_mp_prio_received(struct sock *sk, u8 bkup)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);

	pr_debug("subflow->backup=%d, bkup=%d\n", subflow->backup, bkup);
	subflow->backup = bkup;
}

/* path manager helpers */

bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
			      struct mptcp_addr_info *saddr, bool *echo, bool *port)
{
	int ret = false;

	spin_lock_bh(&msk->pm.lock);

	/* double check after the lock is acquired */
	if (!mptcp_pm_should_add_signal(msk))
		goto out_unlock;

	*echo = mptcp_pm_should_add_signal_echo(msk);
	*port = mptcp_pm_should_add_signal_port(msk);

	if (remaining < mptcp_add_addr_len(msk->pm.local.family, *echo, *port))
		goto out_unlock;

	*saddr = msk->pm.local;
	WRITE_ONCE(msk->pm.addr_signal, 0);
	ret = true;

out_unlock:
	spin_unlock_bh(&msk->pm.lock);
	return ret;
}

bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
			     u8 *rm_id)
{
	int ret = false;

	spin_lock_bh(&msk->pm.lock);

	/* double check after the lock is acquired */
	if (!mptcp_pm_should_rm_signal(msk))
		goto out_unlock;

	if (remaining < TCPOLEN_MPTCP_RM_ADDR_BASE)
		goto out_unlock;

	*rm_id = msk->pm.rm_id;
	WRITE_ONCE(msk->pm.addr_signal, 0);
	ret = true;

out_unlock:
	spin_unlock_bh(&msk->pm.lock);
	return ret;
}

int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
{
	return mptcp_pm_nl_get_local_id(msk, skc);
}

void mptcp_pm_data_init(struct mptcp_sock *msk)
{
	msk->pm.add_addr_signaled = 0;
	msk->pm.add_addr_accepted = 0;
	msk->pm.local_addr_used = 0;
	msk->pm.subflows = 0;
	msk->pm.rm_id = 0;
	WRITE_ONCE(msk->pm.work_pending, false);
	WRITE_ONCE(msk->pm.addr_signal, 0);
	WRITE_ONCE(msk->pm.accept_addr, false);
	WRITE_ONCE(msk->pm.accept_subflow, false);
	msk->pm.status = 0;

	spin_lock_init(&msk->pm.lock);
	INIT_LIST_HEAD(&msk->pm.anno_list);

	mptcp_pm_nl_data_init(msk);
}

void __init mptcp_pm_init(void)
{
	mptcp_pm_nl_init();
}
Commit	Line	Data
1b1c7a0e PK	1	// SPDX-License-Identifier: GPL-2.0
	2	/* Multipath TCP
	3	*
	4	* Copyright (c) 2019, Intel Corporation.
	5	*/
c85adced GT	6	#define pr_fmt(fmt) "MPTCP: " fmt
c85adced GT	7
1b1c7a0e PK	8	#include <linux/kernel.h>
	9	#include <net/tcp.h>
	10	#include <net/mptcp.h>
	11	#include "protocol.h"
	12
1b1c7a0e PK	13	/* path manager command handlers */
	14
	15	int mptcp_pm_announce_addr(struct mptcp_sock *msk,
6a6c05a8	16	const struct mptcp_addr_info *addr,
0f5c9e3f	17	bool echo, bool port)
1b1c7a0e	18	{
13ad9f01	19	u8 add_addr = READ_ONCE(msk->pm.addr_signal);
d91d322a	20
926bdeab PK	21	pr_debug("msk=%p, local_id=%d", msk, addr->id);
926bdeab PK	22
42842a42 GT	23	if (add_addr) {
	24	pr_warn("addr_signal error, add_addr=%d", add_addr);
	25	return -EINVAL;
	26	}
	27
926bdeab	28	msk->pm.local = *addr;
d91d322a GT	29	add_addr \|= BIT(MPTCP_ADD_ADDR_SIGNAL);
	30	if (echo)
	31	add_addr \|= BIT(MPTCP_ADD_ADDR_ECHO);
84dfe367 GT	32	if (addr->family == AF_INET6)
84dfe367 GT	33	add_addr \|= BIT(MPTCP_ADD_ADDR_IPV6);
0f5c9e3f GT	34	if (port)
0f5c9e3f GT	35	add_addr \|= BIT(MPTCP_ADD_ADDR_PORT);
13ad9f01	36	WRITE_ONCE(msk->pm.addr_signal, add_addr);
926bdeab	37	return 0;
1b1c7a0e PK	38	}
	39
	40	int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id)
	41	{
13ad9f01	42	u8 rm_addr = READ_ONCE(msk->pm.addr_signal);
42842a42	43
b6c08380 GT	44	pr_debug("msk=%p, local_id=%d", msk, local_id);
b6c08380 GT	45
42842a42 GT	46	if (rm_addr) {
	47	pr_warn("addr_signal error, rm_addr=%d", rm_addr);
	48	return -EINVAL;
	49	}
	50
b6c08380	51	msk->pm.rm_id = local_id;
42842a42	52	rm_addr \|= BIT(MPTCP_RM_ADDR_SIGNAL);
13ad9f01	53	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
b6c08380	54	return 0;
1b1c7a0e PK	55	}
1b1c7a0e PK	56
0ee4261a	57	int mptcp_pm_remove_subflow(struct mptcp_sock *msk, u8 local_id)
1b1c7a0e	58	{
0ee4261a GT	59	pr_debug("msk=%p, local_id=%d", msk, local_id);
	60
	61	spin_lock_bh(&msk->pm.lock);
	62	mptcp_pm_nl_rm_subflow_received(msk, local_id);
	63	spin_unlock_bh(&msk->pm.lock);
	64	return 0;
1b1c7a0e PK	65	}
	66
	67	/* path manager event handlers */
	68
	69	void mptcp_pm_new_connection(struct mptcp_sock *msk, int server_side)
	70	{
	71	struct mptcp_pm_data *pm = &msk->pm;
	72
	73	pr_debug("msk=%p, token=%u side=%d", msk, msk->token, server_side);
	74
	75	WRITE_ONCE(pm->server_side, server_side);
	76	}
	77
	78	bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
	79	{
926bdeab	80	struct mptcp_pm_data *pm = &msk->pm;
a914e586	81	unsigned int subflows_max;
f58f065a	82	int ret = 0;
926bdeab	83
a914e586 GT	84	subflows_max = mptcp_pm_get_subflows_max(msk);
a914e586 GT	85
926bdeab	86	pr_debug("msk=%p subflows=%d max=%d allow=%d", msk, pm->subflows,
a914e586	87	subflows_max, READ_ONCE(pm->accept_subflow));
926bdeab PK	88
	89	/* try to avoid acquiring the lock below */
	90	if (!READ_ONCE(pm->accept_subflow))
	91	return false;
	92
	93	spin_lock_bh(&pm->lock);
f58f065a	94	if (READ_ONCE(pm->accept_subflow)) {
a914e586 GT	95	ret = pm->subflows < subflows_max;
a914e586 GT	96	if (ret && ++pm->subflows == subflows_max)
f58f065a GT	97	WRITE_ONCE(pm->accept_subflow, false);
f58f065a GT	98	}
926bdeab PK	99	spin_unlock_bh(&pm->lock);
	100
	101	return ret;
	102	}
	103
	104	/* return true if the new status bit is currently cleared, that is, this event
	105	* can be server, eventually by an already scheduled work
	106	*/
	107	static bool mptcp_pm_schedule_work(struct mptcp_sock *msk,
	108	enum mptcp_pm_status new_status)
	109	{
	110	pr_debug("msk=%p status=%x new=%lx", msk, msk->pm.status,
	111	BIT(new_status));
	112	if (msk->pm.status & BIT(new_status))
	113	return false;
	114
	115	msk->pm.status \|= BIT(new_status);
ba8f48f7	116	mptcp_schedule_work((struct sock *)msk);
926bdeab	117	return true;
1b1c7a0e PK	118	}
	119
	120	void mptcp_pm_fully_established(struct mptcp_sock *msk)
	121	{
926bdeab PK	122	struct mptcp_pm_data *pm = &msk->pm;
926bdeab PK	123
1b1c7a0e	124	pr_debug("msk=%p", msk);
926bdeab PK	125
	126	/* try to avoid acquiring the lock below */
	127	if (!READ_ONCE(pm->work_pending))
	128	return;
	129
	130	spin_lock_bh(&pm->lock);
	131
5b950ff4 PA	132	/* mptcp_pm_fully_established() can be invoked by multiple
	133	* racing paths - accept() and check_fully_established()
	134	* be sure to serve this event only once.
	135	*/
	136	if (READ_ONCE(pm->work_pending) &&
	137	!(msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)))
926bdeab	138	mptcp_pm_schedule_work(msk, MPTCP_PM_ESTABLISHED);
5b950ff4	139	msk->pm.status \|= BIT(MPTCP_PM_ALREADY_ESTABLISHED);
926bdeab PK	140
926bdeab PK	141	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	142	}
	143
	144	void mptcp_pm_connection_closed(struct mptcp_sock *msk)
	145	{
	146	pr_debug("msk=%p", msk);
	147	}
	148
	149	void mptcp_pm_subflow_established(struct mptcp_sock *msk,
	150	struct mptcp_subflow_context *subflow)
	151	{
926bdeab PK	152	struct mptcp_pm_data *pm = &msk->pm;
926bdeab PK	153
1b1c7a0e	154	pr_debug("msk=%p", msk);
926bdeab PK	155
	156	if (!READ_ONCE(pm->work_pending))
	157	return;
	158
	159	spin_lock_bh(&pm->lock);
	160
	161	if (READ_ONCE(pm->work_pending))
	162	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	163
	164	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	165	}
	166
	167	void mptcp_pm_subflow_closed(struct mptcp_sock *msk, u8 id)
	168	{
	169	pr_debug("msk=%p", msk);
	170	}
	171
	172	void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
	173	const struct mptcp_addr_info *addr)
	174	{
926bdeab PK	175	struct mptcp_pm_data *pm = &msk->pm;
	176
	177	pr_debug("msk=%p remote_id=%d accept=%d", msk, addr->id,
	178	READ_ONCE(pm->accept_addr));
	179
926bdeab PK	180	spin_lock_bh(&pm->lock);
926bdeab PK	181
84dfe367	182	if (!READ_ONCE(pm->accept_addr)) {
0f5c9e3f	183	mptcp_pm_announce_addr(msk, addr, true, addr->port);
84dfe367 GT	184	mptcp_pm_add_addr_send_ack(msk);
84dfe367 GT	185	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
926bdeab	186	pm->remote = *addr;
84dfe367	187	}
926bdeab PK	188
926bdeab PK	189	spin_unlock_bh(&pm->lock);
84dfe367 GT	190	}
	191
	192	void mptcp_pm_add_addr_send_ack(struct mptcp_sock *msk)
	193	{
b5a7acd3	194	if (!mptcp_pm_should_add_signal(msk))
84dfe367 GT	195	return;
	196
	197	mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_SEND_ACK);
1b1c7a0e PK	198	}
1b1c7a0e PK	199
d0876b22 GT	200	void mptcp_pm_rm_addr_received(struct mptcp_sock *msk, u8 rm_id)
	201	{
	202	struct mptcp_pm_data *pm = &msk->pm;
	203
	204	pr_debug("msk=%p remote_id=%d", msk, rm_id);
	205
	206	spin_lock_bh(&pm->lock);
	207	mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED);
	208	pm->rm_id = rm_id;
	209	spin_unlock_bh(&pm->lock);
	210	}
	211
40453a5c GT	212	void mptcp_pm_mp_prio_received(struct sock *sk, u8 bkup)
	213	{
	214	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	215
	216	pr_debug("subflow->backup=%d, bkup=%d\n", subflow->backup, bkup);
	217	subflow->backup = bkup;
	218	}
	219
1b1c7a0e PK	220	/* path manager helpers */
1b1c7a0e PK	221
f643b803	222	bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
4a2777a8	223	struct mptcp_addr_info saddr, bool echo, bool *port)
1b1c7a0e	224	{
926bdeab PK	225	int ret = false;
	226
	227	spin_lock_bh(&msk->pm.lock);
	228
	229	/* double check after the lock is acquired */
f643b803	230	if (!mptcp_pm_should_add_signal(msk))
926bdeab PK	231	goto out_unlock;
926bdeab PK	232
d91d322a	233	*echo = mptcp_pm_should_add_signal_echo(msk);
4a2777a8	234	*port = mptcp_pm_should_add_signal_port(msk);
456afe01	235
4a2777a8	236	if (remaining < mptcp_add_addr_len(msk->pm.local.family, echo, port))
926bdeab PK	237	goto out_unlock;
	238
	239	*saddr = msk->pm.local;
13ad9f01	240	WRITE_ONCE(msk->pm.addr_signal, 0);
926bdeab PK	241	ret = true;
	242
	243	out_unlock:
	244	spin_unlock_bh(&msk->pm.lock);
	245	return ret;
1b1c7a0e PK	246	}
1b1c7a0e PK	247
5cb104ae GT	248	bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
	249	u8 *rm_id)
	250	{
	251	int ret = false;
	252
	253	spin_lock_bh(&msk->pm.lock);
	254
	255	/* double check after the lock is acquired */
	256	if (!mptcp_pm_should_rm_signal(msk))
	257	goto out_unlock;
	258
	259	if (remaining < TCPOLEN_MPTCP_RM_ADDR_BASE)
	260	goto out_unlock;
	261
	262	*rm_id = msk->pm.rm_id;
13ad9f01	263	WRITE_ONCE(msk->pm.addr_signal, 0);
5cb104ae GT	264	ret = true;
	265
	266	out_unlock:
	267	spin_unlock_bh(&msk->pm.lock);
	268	return ret;
	269	}
	270
1b1c7a0e PK	271	int mptcp_pm_get_local_id(struct mptcp_sock msk, struct sock_common skc)
1b1c7a0e PK	272	{
01cacb00	273	return mptcp_pm_nl_get_local_id(msk, skc);
1b1c7a0e PK	274	}
1b1c7a0e PK	275
1b1c7a0e PK	276	void mptcp_pm_data_init(struct mptcp_sock *msk)
	277	{
	278	msk->pm.add_addr_signaled = 0;
	279	msk->pm.add_addr_accepted = 0;
	280	msk->pm.local_addr_used = 0;
	281	msk->pm.subflows = 0;
5cb104ae	282	msk->pm.rm_id = 0;
1b1c7a0e	283	WRITE_ONCE(msk->pm.work_pending, false);
13ad9f01	284	WRITE_ONCE(msk->pm.addr_signal, 0);
1b1c7a0e PK	285	WRITE_ONCE(msk->pm.accept_addr, false);
	286	WRITE_ONCE(msk->pm.accept_subflow, false);
	287	msk->pm.status = 0;
	288
	289	spin_lock_init(&msk->pm.lock);
b6c08380	290	INIT_LIST_HEAD(&msk->pm.anno_list);
01cacb00 PA	291
01cacb00 PA	292	mptcp_pm_nl_data_init(msk);
1b1c7a0e PK	293	}
1b1c7a0e PK	294
d39dceca	295	void __init mptcp_pm_init(void)
1b1c7a0e	296	{
01cacb00	297	mptcp_pm_nl_init();
1b1c7a0e	298	}