[linux-block.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"

#include "mib.h"

/* path manager command handlers */

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

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

	lockdep_assert_held(&msk->pm.lock);

	if (add_addr &
	    (echo ? BIT(MPTCP_ADD_ADDR_ECHO) : BIT(MPTCP_ADD_ADDR_SIGNAL))) {
		pr_warn("addr_signal error, add_addr=%d, echo=%d", add_addr, echo);
		return -EINVAL;
	}

	if (echo) {
		msk->pm.remote = *addr;
		add_addr |= BIT(MPTCP_ADD_ADDR_ECHO);
	} else {
		msk->pm.local = *addr;
		add_addr |= BIT(MPTCP_ADD_ADDR_SIGNAL);
	}
	WRITE_ONCE(msk->pm.addr_signal, add_addr);
	return 0;
}

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

	pr_debug("msk=%p, rm_list_nr=%d", msk, rm_list->nr);

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

	msk->pm.rm_list_tx = *rm_list;
	rm_addr |= BIT(MPTCP_RM_ADDR_SIGNAL);
	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
	mptcp_pm_nl_addr_send_ack(msk);
	return 0;
}

int mptcp_pm_remove_subflow(struct mptcp_sock *msk, const struct mptcp_rm_list *rm_list)
{
	pr_debug("msk=%p, rm_list_nr=%d", msk, rm_list->nr);

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

/* path manager event handlers */

void mptcp_pm_new_connection(struct mptcp_sock *msk, const struct sock *ssk, 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);
	mptcp_event(MPTCP_EVENT_CREATED, msk, ssk, GFP_ATOMIC);
}

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

	if (mptcp_pm_is_userspace(msk))
		return mptcp_userspace_pm_active(msk);

	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, const struct sock *ssk, gfp_t gfp)
{
	struct mptcp_pm_data *pm = &msk->pm;
	bool announce = false;

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

	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);

	if ((msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)) == 0)
		announce = true;

	msk->pm.status |= BIT(MPTCP_PM_ALREADY_ESTABLISHED);
	spin_unlock_bh(&pm->lock);

	if (announce)
		mptcp_event(MPTCP_EVENT_ESTABLISHED, msk, ssk, gfp);
}

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_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_check_next(struct mptcp_sock *msk, const struct sock *ssk,
				 const struct mptcp_subflow_context *subflow)
{
	struct mptcp_pm_data *pm = &msk->pm;
	bool update_subflows;

	update_subflows = (subflow->request_join || subflow->mp_join) &&
			  mptcp_pm_is_kernel(msk);
	if (!READ_ONCE(pm->work_pending) && !update_subflows)
		return;

	spin_lock_bh(&pm->lock);
	if (update_subflows)
		__mptcp_pm_close_subflow(msk);

	/* Even if this subflow is not really established, tell the PM to try
	 * to pick the next ones, if possible.
	 */
	if (mptcp_pm_nl_check_work_pending(msk))
		mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);

	spin_unlock_bh(&pm->lock);
}

void mptcp_pm_add_addr_received(const struct sock *ssk,
				const struct mptcp_addr_info *addr)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
	struct mptcp_pm_data *pm = &msk->pm;

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

	mptcp_event_addr_announced(ssk, addr);

	spin_lock_bh(&pm->lock);

	if (mptcp_pm_is_userspace(msk)) {
		if (mptcp_userspace_pm_active(msk)) {
			mptcp_pm_announce_addr(msk, addr, true);
			mptcp_pm_add_addr_send_ack(msk);
		} else {
			__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_ADDADDRDROP);
		}
	} else if (!READ_ONCE(pm->accept_addr)) {
		mptcp_pm_announce_addr(msk, addr, true);
		mptcp_pm_add_addr_send_ack(msk);
	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
		pm->remote = *addr;
	} else {
		__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_ADDADDRDROP);
	}

	spin_unlock_bh(&pm->lock);
}

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

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

	spin_lock_bh(&pm->lock);

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

	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,
			       const struct mptcp_rm_list *rm_list)
{
	struct mptcp_pm_data *pm = &msk->pm;
	u8 i;

	pr_debug("msk=%p remote_ids_nr=%d", msk, rm_list->nr);

	for (i = 0; i < rm_list->nr; i++)
		mptcp_event_addr_removed(msk, rm_list->ids[i]);

	spin_lock_bh(&pm->lock);
	if (mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED))
		pm->rm_list_rx = *rm_list;
	else
		__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_RMADDRDROP);
	spin_unlock_bh(&pm->lock);
}

void mptcp_pm_mp_prio_received(struct sock *ssk, u8 bkup)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	struct sock *sk = subflow->conn;
	struct mptcp_sock *msk;

	pr_debug("subflow->backup=%d, bkup=%d\n", subflow->backup, bkup);
	msk = mptcp_sk(sk);
	if (subflow->backup != bkup) {
		subflow->backup = bkup;
		mptcp_data_lock(sk);
		if (!sock_owned_by_user(sk))
			msk->last_snd = NULL;
		else
			__set_bit(MPTCP_RESET_SCHEDULER,  &msk->cb_flags);
		mptcp_data_unlock(sk);
	}

	mptcp_event(MPTCP_EVENT_SUB_PRIORITY, msk, ssk, GFP_ATOMIC);
}

void mptcp_pm_mp_fail_received(struct sock *sk, u64 fail_seq)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
	struct sock *s = (struct sock *)msk;

	pr_debug("fail_seq=%llu", fail_seq);

	if (!READ_ONCE(msk->allow_infinite_fallback))
		return;

	if (!READ_ONCE(subflow->mp_fail_response_expect)) {
		pr_debug("send MP_FAIL response and infinite map");

		subflow->send_mp_fail = 1;
		subflow->send_infinite_map = 1;
	} else if (!sock_flag(sk, SOCK_DEAD)) {
		pr_debug("MP_FAIL response received");

		sk_stop_timer(s, &s->sk_timer);
	}
}

/* path manager helpers */

bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb,
			      unsigned int opt_size, unsigned int remaining,
			      struct mptcp_addr_info *addr, bool *echo,
			      bool *drop_other_suboptions)
{
	int ret = false;
	u8 add_addr;
	u8 family;
	bool port;

	spin_lock_bh(&msk->pm.lock);

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

	/* always drop every other options for pure ack ADD_ADDR; this is a
	 * plain dup-ack from TCP perspective. The other MPTCP-relevant info,
	 * if any, will be carried by the 'original' TCP ack
	 */
	if (skb && skb_is_tcp_pure_ack(skb)) {
		remaining += opt_size;
		*drop_other_suboptions = true;
	}

	*echo = mptcp_pm_should_add_signal_echo(msk);
	port = !!(*echo ? msk->pm.remote.port : msk->pm.local.port);

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

	if (*echo) {
		*addr = msk->pm.remote;
		add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_ECHO);
	} else {
		*addr = msk->pm.local;
		add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_SIGNAL);
	}
	WRITE_ONCE(msk->pm.addr_signal, add_addr);
	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,
			     struct mptcp_rm_list *rm_list)
{
	int ret = false, len;
	u8 rm_addr;

	spin_lock_bh(&msk->pm.lock);

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

	rm_addr = msk->pm.addr_signal & ~BIT(MPTCP_RM_ADDR_SIGNAL);
	len = mptcp_rm_addr_len(&msk->pm.rm_list_tx);
	if (len < 0) {
		WRITE_ONCE(msk->pm.addr_signal, rm_addr);
		goto out_unlock;
	}
	if (remaining < len)
		goto out_unlock;

	*rm_list = msk->pm.rm_list_tx;
	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
	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_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
{
	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	u32 rcv_tstamp = READ_ONCE(tcp_sk(ssk)->rcv_tstamp);

	/* keep track of rtx periods with no progress */
	if (!subflow->stale_count) {
		subflow->stale_rcv_tstamp = rcv_tstamp;
		subflow->stale_count++;
	} else if (subflow->stale_rcv_tstamp == rcv_tstamp) {
		if (subflow->stale_count < U8_MAX)
			subflow->stale_count++;
		mptcp_pm_nl_subflow_chk_stale(msk, ssk);
	} else {
		subflow->stale_count = 0;
		mptcp_subflow_set_active(subflow);
	}
}

void mptcp_pm_data_reset(struct mptcp_sock *msk)
{
	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
	struct mptcp_pm_data *pm = &msk->pm;

	pm->add_addr_signaled = 0;
	pm->add_addr_accepted = 0;
	pm->local_addr_used = 0;
	pm->subflows = 0;
	pm->rm_list_tx.nr = 0;
	pm->rm_list_rx.nr = 0;
	WRITE_ONCE(pm->pm_type, pm_type);

	if (pm_type == MPTCP_PM_TYPE_KERNEL) {
		bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);

		/* pm->work_pending must be only be set to 'true' when
		 * pm->pm_type is set to MPTCP_PM_TYPE_KERNEL
		 */
		WRITE_ONCE(pm->work_pending,
			   (!!mptcp_pm_get_local_addr_max(msk) &&
			    subflows_allowed) ||
			   !!mptcp_pm_get_add_addr_signal_max(msk));
		WRITE_ONCE(pm->accept_addr,
			   !!mptcp_pm_get_add_addr_accept_max(msk) &&
			   subflows_allowed);
		WRITE_ONCE(pm->accept_subflow, subflows_allowed);
	} else {
		WRITE_ONCE(pm->work_pending, 0);
		WRITE_ONCE(pm->accept_addr, 0);
		WRITE_ONCE(pm->accept_subflow, 0);
	}

	WRITE_ONCE(pm->addr_signal, 0);
	WRITE_ONCE(pm->remote_deny_join_id0, false);
	pm->status = 0;
	bitmap_fill(msk->pm.id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
}

void mptcp_pm_data_init(struct mptcp_sock *msk)
{
	spin_lock_init(&msk->pm.lock);
	INIT_LIST_HEAD(&msk->pm.anno_list);
	INIT_LIST_HEAD(&msk->pm.userspace_pm_local_addr_list);
	mptcp_pm_data_reset(msk);
}

void __init mptcp_pm_init(void)
{
	mptcp_pm_nl_init();
}
Commit	Line	Data
	1	// SPDX-License-Identifier: GPL-2.0
	2	/* Multipath TCP
	3	*
	4	* Copyright (c) 2019, Intel Corporation.
	5	*/
	6	#define pr_fmt(fmt) "MPTCP: " fmt
	7
	8	#include <linux/kernel.h>
	9	#include <net/tcp.h>
	10	#include <net/mptcp.h>
	11	#include "protocol.h"
	12
	13	#include "mib.h"
	14
	15	/* path manager command handlers */
	16
	17	int mptcp_pm_announce_addr(struct mptcp_sock *msk,
	18	const struct mptcp_addr_info *addr,
	19	bool echo)
	20	{
	21	u8 add_addr = READ_ONCE(msk->pm.addr_signal);
	22
	23	pr_debug("msk=%p, local_id=%d, echo=%d", msk, addr->id, echo);
	24
	25	lockdep_assert_held(&msk->pm.lock);
	26
	27	if (add_addr &
	28	(echo ? BIT(MPTCP_ADD_ADDR_ECHO) : BIT(MPTCP_ADD_ADDR_SIGNAL))) {
	29	pr_warn("addr_signal error, add_addr=%d, echo=%d", add_addr, echo);
	30	return -EINVAL;
	31	}
	32
	33	if (echo) {
	34	msk->pm.remote = *addr;
	35	add_addr \|= BIT(MPTCP_ADD_ADDR_ECHO);
	36	} else {
	37	msk->pm.local = *addr;
	38	add_addr \|= BIT(MPTCP_ADD_ADDR_SIGNAL);
	39	}
	40	WRITE_ONCE(msk->pm.addr_signal, add_addr);
	41	return 0;
	42	}
	43
	44	int mptcp_pm_remove_addr(struct mptcp_sock msk, const struct mptcp_rm_list rm_list)
	45	{
	46	u8 rm_addr = READ_ONCE(msk->pm.addr_signal);
	47
	48	pr_debug("msk=%p, rm_list_nr=%d", msk, rm_list->nr);
	49
	50	if (rm_addr) {
	51	pr_warn("addr_signal error, rm_addr=%d", rm_addr);
	52	return -EINVAL;
	53	}
	54
	55	msk->pm.rm_list_tx = *rm_list;
	56	rm_addr \|= BIT(MPTCP_RM_ADDR_SIGNAL);
	57	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
	58	mptcp_pm_nl_addr_send_ack(msk);
	59	return 0;
	60	}
	61
	62	int mptcp_pm_remove_subflow(struct mptcp_sock msk, const struct mptcp_rm_list rm_list)
	63	{
	64	pr_debug("msk=%p, rm_list_nr=%d", msk, rm_list->nr);
	65
	66	spin_lock_bh(&msk->pm.lock);
	67	mptcp_pm_nl_rm_subflow_received(msk, rm_list);
	68	spin_unlock_bh(&msk->pm.lock);
	69	return 0;
	70	}
	71
	72	/* path manager event handlers */
	73
	74	void mptcp_pm_new_connection(struct mptcp_sock msk, const struct sock ssk, int server_side)
	75	{
	76	struct mptcp_pm_data *pm = &msk->pm;
	77
	78	pr_debug("msk=%p, token=%u side=%d", msk, msk->token, server_side);
	79
	80	WRITE_ONCE(pm->server_side, server_side);
	81	mptcp_event(MPTCP_EVENT_CREATED, msk, ssk, GFP_ATOMIC);
	82	}
	83
	84	bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
	85	{
	86	struct mptcp_pm_data *pm = &msk->pm;
	87	unsigned int subflows_max;
	88	int ret = 0;
	89
	90	if (mptcp_pm_is_userspace(msk))
	91	return mptcp_userspace_pm_active(msk);
	92
	93	subflows_max = mptcp_pm_get_subflows_max(msk);
	94
	95	pr_debug("msk=%p subflows=%d max=%d allow=%d", msk, pm->subflows,
	96	subflows_max, READ_ONCE(pm->accept_subflow));
	97
	98	/* try to avoid acquiring the lock below */
	99	if (!READ_ONCE(pm->accept_subflow))
	100	return false;
	101
	102	spin_lock_bh(&pm->lock);
	103	if (READ_ONCE(pm->accept_subflow)) {
	104	ret = pm->subflows < subflows_max;
	105	if (ret && ++pm->subflows == subflows_max)
	106	WRITE_ONCE(pm->accept_subflow, false);
	107	}
	108	spin_unlock_bh(&pm->lock);
	109
	110	return ret;
	111	}
	112
	113	/* return true if the new status bit is currently cleared, that is, this event
	114	* can be server, eventually by an already scheduled work
	115	*/
	116	static bool mptcp_pm_schedule_work(struct mptcp_sock *msk,
	117	enum mptcp_pm_status new_status)
	118	{
	119	pr_debug("msk=%p status=%x new=%lx", msk, msk->pm.status,
	120	BIT(new_status));
	121	if (msk->pm.status & BIT(new_status))
	122	return false;
	123
	124	msk->pm.status \|= BIT(new_status);
	125	mptcp_schedule_work((struct sock *)msk);
	126	return true;
	127	}
	128
	129	void mptcp_pm_fully_established(struct mptcp_sock msk, const struct sock ssk, gfp_t gfp)
	130	{
	131	struct mptcp_pm_data *pm = &msk->pm;
	132	bool announce = false;
	133
	134	pr_debug("msk=%p", msk);
	135
	136	spin_lock_bh(&pm->lock);
	137
	138	/* mptcp_pm_fully_established() can be invoked by multiple
	139	* racing paths - accept() and check_fully_established()
	140	* be sure to serve this event only once.
	141	*/
	142	if (READ_ONCE(pm->work_pending) &&
	143	!(msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)))
	144	mptcp_pm_schedule_work(msk, MPTCP_PM_ESTABLISHED);
	145
	146	if ((msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)) == 0)
	147	announce = true;
	148
	149	msk->pm.status \|= BIT(MPTCP_PM_ALREADY_ESTABLISHED);
	150	spin_unlock_bh(&pm->lock);
	151
	152	if (announce)
	153	mptcp_event(MPTCP_EVENT_ESTABLISHED, msk, ssk, gfp);
	154	}
	155
	156	void mptcp_pm_connection_closed(struct mptcp_sock *msk)
	157	{
	158	pr_debug("msk=%p", msk);
	159	}
	160
	161	void mptcp_pm_subflow_established(struct mptcp_sock *msk)
	162	{
	163	struct mptcp_pm_data *pm = &msk->pm;
	164
	165	pr_debug("msk=%p", msk);
	166
	167	if (!READ_ONCE(pm->work_pending))
	168	return;
	169
	170	spin_lock_bh(&pm->lock);
	171
	172	if (READ_ONCE(pm->work_pending))
	173	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	174
	175	spin_unlock_bh(&pm->lock);
	176	}
	177
	178	void mptcp_pm_subflow_check_next(struct mptcp_sock msk, const struct sock ssk,
	179	const struct mptcp_subflow_context *subflow)
	180	{
	181	struct mptcp_pm_data *pm = &msk->pm;
	182	bool update_subflows;
	183
	184	update_subflows = (subflow->request_join \|\| subflow->mp_join) &&
	185	mptcp_pm_is_kernel(msk);
	186	if (!READ_ONCE(pm->work_pending) && !update_subflows)
	187	return;
	188
	189	spin_lock_bh(&pm->lock);
	190	if (update_subflows)
	191	__mptcp_pm_close_subflow(msk);
	192
	193	/* Even if this subflow is not really established, tell the PM to try
	194	* to pick the next ones, if possible.
	195	*/
	196	if (mptcp_pm_nl_check_work_pending(msk))
	197	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	198
	199	spin_unlock_bh(&pm->lock);
	200	}
	201
	202	void mptcp_pm_add_addr_received(const struct sock *ssk,
	203	const struct mptcp_addr_info *addr)
	204	{
	205	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	206	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
	207	struct mptcp_pm_data *pm = &msk->pm;
	208
	209	pr_debug("msk=%p remote_id=%d accept=%d", msk, addr->id,
	210	READ_ONCE(pm->accept_addr));
	211
	212	mptcp_event_addr_announced(ssk, addr);
	213
	214	spin_lock_bh(&pm->lock);
	215
	216	if (mptcp_pm_is_userspace(msk)) {
	217	if (mptcp_userspace_pm_active(msk)) {
	218	mptcp_pm_announce_addr(msk, addr, true);
	219	mptcp_pm_add_addr_send_ack(msk);
	220	} else {
	221	__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_ADDADDRDROP);
	222	}
	223	} else if (!READ_ONCE(pm->accept_addr)) {
	224	mptcp_pm_announce_addr(msk, addr, true);
	225	mptcp_pm_add_addr_send_ack(msk);
	226	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
	227	pm->remote = *addr;
	228	} else {
	229	__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_ADDADDRDROP);
	230	}
	231
	232	spin_unlock_bh(&pm->lock);
	233	}
	234
	235	void mptcp_pm_add_addr_echoed(struct mptcp_sock *msk,
	236	const struct mptcp_addr_info *addr)
	237	{
	238	struct mptcp_pm_data *pm = &msk->pm;
	239
	240	pr_debug("msk=%p", msk);
	241
	242	spin_lock_bh(&pm->lock);
	243
	244	if (mptcp_lookup_anno_list_by_saddr(msk, addr) && READ_ONCE(pm->work_pending))
	245	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	246
	247	spin_unlock_bh(&pm->lock);
	248	}
	249
	250	void mptcp_pm_add_addr_send_ack(struct mptcp_sock *msk)
	251	{
	252	if (!mptcp_pm_should_add_signal(msk))
	253	return;
	254
	255	mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_SEND_ACK);
	256	}
	257
	258	void mptcp_pm_rm_addr_received(struct mptcp_sock *msk,
	259	const struct mptcp_rm_list *rm_list)
	260	{
	261	struct mptcp_pm_data *pm = &msk->pm;
	262	u8 i;
	263
	264	pr_debug("msk=%p remote_ids_nr=%d", msk, rm_list->nr);
	265
	266	for (i = 0; i < rm_list->nr; i++)
	267	mptcp_event_addr_removed(msk, rm_list->ids[i]);
	268
	269	spin_lock_bh(&pm->lock);
	270	if (mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED))
	271	pm->rm_list_rx = *rm_list;
	272	else
	273	__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_RMADDRDROP);
	274	spin_unlock_bh(&pm->lock);
	275	}
	276
	277	void mptcp_pm_mp_prio_received(struct sock *ssk, u8 bkup)
	278	{
	279	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	280	struct sock *sk = subflow->conn;
	281	struct mptcp_sock *msk;
	282
	283	pr_debug("subflow->backup=%d, bkup=%d\n", subflow->backup, bkup);
	284	msk = mptcp_sk(sk);
	285	if (subflow->backup != bkup) {
	286	subflow->backup = bkup;
	287	mptcp_data_lock(sk);
	288	if (!sock_owned_by_user(sk))
	289	msk->last_snd = NULL;
	290	else
	291	__set_bit(MPTCP_RESET_SCHEDULER, &msk->cb_flags);
	292	mptcp_data_unlock(sk);
	293	}
	294
	295	mptcp_event(MPTCP_EVENT_SUB_PRIORITY, msk, ssk, GFP_ATOMIC);
	296	}
	297
	298	void mptcp_pm_mp_fail_received(struct sock *sk, u64 fail_seq)
	299	{
	300	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	301	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
	302	struct sock s = (struct sock )msk;
	303
	304	pr_debug("fail_seq=%llu", fail_seq);
	305
	306	if (!READ_ONCE(msk->allow_infinite_fallback))
	307	return;
	308
	309	if (!READ_ONCE(subflow->mp_fail_response_expect)) {
	310	pr_debug("send MP_FAIL response and infinite map");
	311
	312	subflow->send_mp_fail = 1;
	313	subflow->send_infinite_map = 1;
	314	} else if (!sock_flag(sk, SOCK_DEAD)) {
	315	pr_debug("MP_FAIL response received");
	316
	317	sk_stop_timer(s, &s->sk_timer);
	318	}
	319	}
	320
	321	/* path manager helpers */
	322
	323	bool mptcp_pm_add_addr_signal(struct mptcp_sock msk, const struct sk_buff skb,
	324	unsigned int opt_size, unsigned int remaining,
	325	struct mptcp_addr_info addr, bool echo,
	326	bool *drop_other_suboptions)
	327	{
	328	int ret = false;
	329	u8 add_addr;
	330	u8 family;
	331	bool port;
	332
	333	spin_lock_bh(&msk->pm.lock);
	334
	335	/* double check after the lock is acquired */
	336	if (!mptcp_pm_should_add_signal(msk))
	337	goto out_unlock;
	338
	339	/* always drop every other options for pure ack ADD_ADDR; this is a
	340	* plain dup-ack from TCP perspective. The other MPTCP-relevant info,
	341	* if any, will be carried by the 'original' TCP ack
	342	*/
	343	if (skb && skb_is_tcp_pure_ack(skb)) {
	344	remaining += opt_size;
	345	*drop_other_suboptions = true;
	346	}
	347
	348	*echo = mptcp_pm_should_add_signal_echo(msk);
	349	port = !!(*echo ? msk->pm.remote.port : msk->pm.local.port);
	350
	351	family = *echo ? msk->pm.remote.family : msk->pm.local.family;
	352	if (remaining < mptcp_add_addr_len(family, *echo, port))
	353	goto out_unlock;
	354
	355	if (*echo) {
	356	*addr = msk->pm.remote;
	357	add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_ECHO);
	358	} else {
	359	*addr = msk->pm.local;
	360	add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_SIGNAL);
	361	}
	362	WRITE_ONCE(msk->pm.addr_signal, add_addr);
	363	ret = true;
	364
	365	out_unlock:
	366	spin_unlock_bh(&msk->pm.lock);
	367	return ret;
	368	}
	369
	370	bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
	371	struct mptcp_rm_list *rm_list)
	372	{
	373	int ret = false, len;
	374	u8 rm_addr;
	375
	376	spin_lock_bh(&msk->pm.lock);
	377
	378	/* double check after the lock is acquired */
	379	if (!mptcp_pm_should_rm_signal(msk))
	380	goto out_unlock;
	381
	382	rm_addr = msk->pm.addr_signal & ~BIT(MPTCP_RM_ADDR_SIGNAL);
	383	len = mptcp_rm_addr_len(&msk->pm.rm_list_tx);
	384	if (len < 0) {
	385	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
	386	goto out_unlock;
	387	}
	388	if (remaining < len)
	389	goto out_unlock;
	390
	391	*rm_list = msk->pm.rm_list_tx;
	392	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
	393	ret = true;
	394
	395	out_unlock:
	396	spin_unlock_bh(&msk->pm.lock);
	397	return ret;
	398	}
	399
	400	int mptcp_pm_get_local_id(struct mptcp_sock msk, struct sock_common skc)
	401	{
	402	return mptcp_pm_nl_get_local_id(msk, skc);
	403	}
	404
	405	void mptcp_pm_subflow_chk_stale(const struct mptcp_sock msk, struct sock ssk)
	406	{
	407	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	408	u32 rcv_tstamp = READ_ONCE(tcp_sk(ssk)->rcv_tstamp);
	409
	410	/* keep track of rtx periods with no progress */
	411	if (!subflow->stale_count) {
	412	subflow->stale_rcv_tstamp = rcv_tstamp;
	413	subflow->stale_count++;
	414	} else if (subflow->stale_rcv_tstamp == rcv_tstamp) {
	415	if (subflow->stale_count < U8_MAX)
	416	subflow->stale_count++;
	417	mptcp_pm_nl_subflow_chk_stale(msk, ssk);
	418	} else {
	419	subflow->stale_count = 0;
	420	mptcp_subflow_set_active(subflow);
	421	}
	422	}
	423
	424	void mptcp_pm_data_reset(struct mptcp_sock *msk)
	425	{
	426	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
	427	struct mptcp_pm_data *pm = &msk->pm;
	428
	429	pm->add_addr_signaled = 0;
	430	pm->add_addr_accepted = 0;
	431	pm->local_addr_used = 0;
	432	pm->subflows = 0;
	433	pm->rm_list_tx.nr = 0;
	434	pm->rm_list_rx.nr = 0;
	435	WRITE_ONCE(pm->pm_type, pm_type);
	436
	437	if (pm_type == MPTCP_PM_TYPE_KERNEL) {
	438	bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
	439
	440	/* pm->work_pending must be only be set to 'true' when
	441	* pm->pm_type is set to MPTCP_PM_TYPE_KERNEL
	442	*/
	443	WRITE_ONCE(pm->work_pending,
	444	(!!mptcp_pm_get_local_addr_max(msk) &&
	445	subflows_allowed) \|\|
	446	!!mptcp_pm_get_add_addr_signal_max(msk));
	447	WRITE_ONCE(pm->accept_addr,
	448	!!mptcp_pm_get_add_addr_accept_max(msk) &&
	449	subflows_allowed);
	450	WRITE_ONCE(pm->accept_subflow, subflows_allowed);
	451	} else {
	452	WRITE_ONCE(pm->work_pending, 0);
	453	WRITE_ONCE(pm->accept_addr, 0);
	454	WRITE_ONCE(pm->accept_subflow, 0);
	455	}
	456
	457	WRITE_ONCE(pm->addr_signal, 0);
	458	WRITE_ONCE(pm->remote_deny_join_id0, false);
	459	pm->status = 0;
	460	bitmap_fill(msk->pm.id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
	461	}
	462
	463	void mptcp_pm_data_init(struct mptcp_sock *msk)
	464	{
	465	spin_lock_init(&msk->pm.lock);
	466	INIT_LIST_HEAD(&msk->pm.anno_list);
	467	INIT_LIST_HEAD(&msk->pm.userspace_pm_local_addr_list);
	468	mptcp_pm_data_reset(msk);
	469	}
	470
	471	void __init mptcp_pm_init(void)
	472	{
	473	mptcp_pm_nl_init();
	474	}