[linux-2.6-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))) {
		MPTCP_INC_STATS(sock_net((struct sock *)msk),
				echo ? MPTCP_MIB_ECHOADDTXDROP : MPTCP_MIB_ADDADDRTXDROP);
		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) {
		MPTCP_ADD_STATS(sock_net((struct sock *)msk),
				MPTCP_MIB_RMADDRTXDROP, rm_list->nr);
		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)) {
		if (mptcp_userspace_pm_active(msk)) {
			spin_lock_bh(&pm->lock);
			pm->subflows++;
			spin_unlock_bh(&pm->lock);
			return true;
		}
		return false;
	}

	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)
{
	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_ATOMIC);
}

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 mptcp_subflow_context *subflow)
{
	struct mptcp_pm_data *pm = &msk->pm;
	bool update_subflows;

	update_subflows = subflow->request_join || subflow->mp_join;
	if (mptcp_pm_is_userspace(msk)) {
		if (update_subflows) {
			spin_lock_bh(&pm->lock);
			pm->subflows--;
			spin_unlock_bh(&pm->lock);
		}
		return;
	}

	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_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);

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

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

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

		subflow->send_mp_fail = 1;
		subflow->send_infinite_map = 1;
		tcp_send_ack(sk);
	} else {
		pr_debug("MP_FAIL response received");
		WRITE_ONCE(subflow->fail_tout, 0);
	}
}

/* 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)
{
	struct mptcp_addr_info skc_local;
	struct mptcp_addr_info msk_local;

	if (WARN_ON_ONCE(!msk))
		return -1;

	/* The 0 ID mapping is defined by the first subflow, copied into the msk
	 * addr
	 */
	mptcp_local_address((struct sock_common *)msk, &msk_local);
	mptcp_local_address((struct sock_common *)skc, &skc_local);
	if (mptcp_addresses_equal(&msk_local, &skc_local, false))
		return 0;

	if (mptcp_pm_is_userspace(msk))
		return mptcp_userspace_pm_get_local_id(msk, &skc_local);
	return mptcp_pm_nl_get_local_id(msk, &skc_local);
}

int mptcp_pm_get_flags_and_ifindex_by_id(struct mptcp_sock *msk, unsigned int id,
					 u8 *flags, int *ifindex)
{
	*flags = 0;
	*ifindex = 0;

	if (!id)
		return 0;

	if (mptcp_pm_is_userspace(msk))
		return mptcp_userspace_pm_get_flags_and_ifindex_by_id(msk, id, flags, ifindex);
	return mptcp_pm_nl_get_flags_and_ifindex_by_id(msk, id, flags, ifindex);
}

int mptcp_pm_set_flags(struct net *net, struct nlattr *token,
		       struct mptcp_pm_addr_entry *loc,
		       struct mptcp_pm_addr_entry *rem, u8 bkup)
{
	if (token)
		return mptcp_userspace_pm_set_flags(net, token, loc, rem, bkup);
	return mptcp_pm_nl_set_flags(net, loc, bkup);
}

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);
	}
}

/* if sk is ipv4 or ipv6_only allows only same-family local and remote addresses,
 * otherwise allow any matching local/remote pair
 */
bool mptcp_pm_addr_families_match(const struct sock *sk,
				  const struct mptcp_addr_info *loc,
				  const struct mptcp_addr_info *rem)
{
	bool mptcp_is_v4 = sk->sk_family == AF_INET;

#if IS_ENABLED(CONFIG_MPTCP_IPV6)
	bool loc_is_v4 = loc->family == AF_INET || ipv6_addr_v4mapped(&loc->addr6);
	bool rem_is_v4 = rem->family == AF_INET || ipv6_addr_v4mapped(&rem->addr6);

	if (mptcp_is_v4)
		return loc_is_v4 && rem_is_v4;

	if (ipv6_only_sock(sk))
		return !loc_is_v4 && !rem_is_v4;

	return loc_is_v4 == rem_is_v4;
#else
	return mptcp_is_v4 && loc->family == AF_INET && rem->family == AF_INET;
#endif
}

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
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
fc1b4e3b PA	13	#include "mib.h"
fc1b4e3b PA	14
1b1c7a0e PK	15	/* path manager command handlers */
	16
	17	int mptcp_pm_announce_addr(struct mptcp_sock *msk,
6a6c05a8	18	const struct mptcp_addr_info *addr,
f7efc777	19	bool echo)
1b1c7a0e	20	{
13ad9f01	21	u8 add_addr = READ_ONCE(msk->pm.addr_signal);
d91d322a	22
18fc1a92	23	pr_debug("msk=%p, local_id=%d, echo=%d", msk, addr->id, echo);
926bdeab	24
3abc05d9 FW	25	lockdep_assert_held(&msk->pm.lock);
3abc05d9 FW	26
18fc1a92 YL	27	if (add_addr &
18fc1a92 YL	28	(echo ? BIT(MPTCP_ADD_ADDR_ECHO) : BIT(MPTCP_ADD_ADDR_SIGNAL))) {
45b1a122 PA	29	MPTCP_INC_STATS(sock_net((struct sock *)msk),
45b1a122 PA	30	echo ? MPTCP_MIB_ECHOADDTXDROP : MPTCP_MIB_ADDADDRTXDROP);
42842a42 GT	31	return -EINVAL;
	32	}
	33
18fc1a92 YL	34	if (echo) {
18fc1a92 YL	35	msk->pm.remote = *addr;
d91d322a	36	add_addr \|= BIT(MPTCP_ADD_ADDR_ECHO);
18fc1a92 YL	37	} else {
	38	msk->pm.local = *addr;
	39	add_addr \|= BIT(MPTCP_ADD_ADDR_SIGNAL);
	40	}
13ad9f01	41	WRITE_ONCE(msk->pm.addr_signal, add_addr);
926bdeab	42	return 0;
1b1c7a0e PK	43	}
1b1c7a0e PK	44
cbde2787	45	int mptcp_pm_remove_addr(struct mptcp_sock msk, const struct mptcp_rm_list rm_list)
1b1c7a0e	46	{
13ad9f01	47	u8 rm_addr = READ_ONCE(msk->pm.addr_signal);
42842a42	48
cbde2787	49	pr_debug("msk=%p, rm_list_nr=%d", msk, rm_list->nr);
b6c08380	50
42842a42	51	if (rm_addr) {
45b1a122 PA	52	MPTCP_ADD_STATS(sock_net((struct sock *)msk),
45b1a122 PA	53	MPTCP_MIB_RMADDRTXDROP, rm_list->nr);
42842a42 GT	54	return -EINVAL;
	55	}
	56
cbde2787	57	msk->pm.rm_list_tx = *rm_list;
42842a42	58	rm_addr \|= BIT(MPTCP_RM_ADDR_SIGNAL);
13ad9f01	59	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
b46a0238	60	mptcp_pm_nl_addr_send_ack(msk);
b6c08380	61	return 0;
1b1c7a0e PK	62	}
1b1c7a0e PK	63
ddd14bb8	64	int mptcp_pm_remove_subflow(struct mptcp_sock msk, const struct mptcp_rm_list rm_list)
1b1c7a0e	65	{
ddd14bb8	66	pr_debug("msk=%p, rm_list_nr=%d", msk, rm_list->nr);
0ee4261a GT	67
0ee4261a GT	68	spin_lock_bh(&msk->pm.lock);
ddd14bb8	69	mptcp_pm_nl_rm_subflow_received(msk, rm_list);
0ee4261a GT	70	spin_unlock_bh(&msk->pm.lock);
0ee4261a GT	71	return 0;
1b1c7a0e PK	72	}
	73
	74	/* path manager event handlers */
	75
6c714f1b	76	void mptcp_pm_new_connection(struct mptcp_sock msk, const struct sock ssk, int server_side)
1b1c7a0e PK	77	{
	78	struct mptcp_pm_data *pm = &msk->pm;
	79
	80	pr_debug("msk=%p, token=%u side=%d", msk, msk->token, server_side);
	81
	82	WRITE_ONCE(pm->server_side, server_side);
b911c97c	83	mptcp_event(MPTCP_EVENT_CREATED, msk, ssk, GFP_ATOMIC);
1b1c7a0e PK	84	}
	85
	86	bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
	87	{
926bdeab	88	struct mptcp_pm_data *pm = &msk->pm;
a914e586	89	unsigned int subflows_max;
f58f065a	90	int ret = 0;
926bdeab	91
77e4b94a GT	92	if (mptcp_pm_is_userspace(msk)) {
	93	if (mptcp_userspace_pm_active(msk)) {
	94	spin_lock_bh(&pm->lock);
	95	pm->subflows++;
	96	spin_unlock_bh(&pm->lock);
	97	return true;
	98	}
	99	return false;
	100	}
4d25247d	101
a914e586 GT	102	subflows_max = mptcp_pm_get_subflows_max(msk);
a914e586 GT	103
926bdeab	104	pr_debug("msk=%p subflows=%d max=%d allow=%d", msk, pm->subflows,
a914e586	105	subflows_max, READ_ONCE(pm->accept_subflow));
926bdeab PK	106
	107	/* try to avoid acquiring the lock below */
	108	if (!READ_ONCE(pm->accept_subflow))
	109	return false;
	110
	111	spin_lock_bh(&pm->lock);
f58f065a	112	if (READ_ONCE(pm->accept_subflow)) {
a914e586 GT	113	ret = pm->subflows < subflows_max;
a914e586 GT	114	if (ret && ++pm->subflows == subflows_max)
f58f065a GT	115	WRITE_ONCE(pm->accept_subflow, false);
f58f065a GT	116	}
926bdeab PK	117	spin_unlock_bh(&pm->lock);
	118
	119	return ret;
	120	}
	121
	122	/* return true if the new status bit is currently cleared, that is, this event
	123	* can be server, eventually by an already scheduled work
	124	*/
	125	static bool mptcp_pm_schedule_work(struct mptcp_sock *msk,
	126	enum mptcp_pm_status new_status)
	127	{
	128	pr_debug("msk=%p status=%x new=%lx", msk, msk->pm.status,
	129	BIT(new_status));
	130	if (msk->pm.status & BIT(new_status))
	131	return false;
	132
	133	msk->pm.status \|= BIT(new_status);
ba8f48f7	134	mptcp_schedule_work((struct sock *)msk);
926bdeab	135	return true;
1b1c7a0e PK	136	}
1b1c7a0e PK	137
7a486c44	138	void mptcp_pm_fully_established(struct mptcp_sock msk, const struct sock ssk)
1b1c7a0e	139	{
926bdeab	140	struct mptcp_pm_data *pm = &msk->pm;
b911c97c	141	bool announce = false;
926bdeab	142
1b1c7a0e	143	pr_debug("msk=%p", msk);
926bdeab	144
926bdeab PK	145	spin_lock_bh(&pm->lock);
926bdeab PK	146
5b950ff4 PA	147	/* mptcp_pm_fully_established() can be invoked by multiple
	148	* racing paths - accept() and check_fully_established()
	149	* be sure to serve this event only once.
	150	*/
	151	if (READ_ONCE(pm->work_pending) &&
	152	!(msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)))
926bdeab PK	153	mptcp_pm_schedule_work(msk, MPTCP_PM_ESTABLISHED);
926bdeab PK	154
b911c97c FW	155	if ((msk->pm.status & BIT(MPTCP_PM_ALREADY_ESTABLISHED)) == 0)
	156	announce = true;
	157
	158	msk->pm.status \|= BIT(MPTCP_PM_ALREADY_ESTABLISHED);
926bdeab	159	spin_unlock_bh(&pm->lock);
b911c97c FW	160
b911c97c FW	161	if (announce)
7a486c44	162	mptcp_event(MPTCP_EVENT_ESTABLISHED, msk, ssk, GFP_ATOMIC);
1b1c7a0e PK	163	}
	164
	165	void mptcp_pm_connection_closed(struct mptcp_sock *msk)
	166	{
	167	pr_debug("msk=%p", msk);
	168	}
	169
62535200	170	void mptcp_pm_subflow_established(struct mptcp_sock *msk)
1b1c7a0e	171	{
926bdeab PK	172	struct mptcp_pm_data *pm = &msk->pm;
926bdeab PK	173
1b1c7a0e	174	pr_debug("msk=%p", msk);
926bdeab PK	175
	176	if (!READ_ONCE(pm->work_pending))
	177	return;
	178
	179	spin_lock_bh(&pm->lock);
	180
	181	if (READ_ONCE(pm->work_pending))
	182	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	183
	184	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	185	}
1b1c7a0e PK	186
74cbb0c6	187	void mptcp_pm_subflow_check_next(struct mptcp_sock *msk,
a88c9e49	188	const struct mptcp_subflow_context *subflow)
1b1c7a0e	189	{
a88c9e49 PA	190	struct mptcp_pm_data *pm = &msk->pm;
	191	bool update_subflows;
	192
77e4b94a GT	193	update_subflows = subflow->request_join \|\| subflow->mp_join;
	194	if (mptcp_pm_is_userspace(msk)) {
	195	if (update_subflows) {
	196	spin_lock_bh(&pm->lock);
	197	pm->subflows--;
	198	spin_unlock_bh(&pm->lock);
	199	}
	200	return;
	201	}
	202
a88c9e49 PA	203	if (!READ_ONCE(pm->work_pending) && !update_subflows)
	204	return;
	205
	206	spin_lock_bh(&pm->lock);
	207	if (update_subflows)
95d68651	208	__mptcp_pm_close_subflow(msk);
a88c9e49 PA	209
	210	/* Even if this subflow is not really established, tell the PM to try
	211	* to pick the next ones, if possible.
	212	*/
	213	if (mptcp_pm_nl_check_work_pending(msk))
	214	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	215
	216	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	217	}
1b1c7a0e PK	218
d1ace2d9	219	void mptcp_pm_add_addr_received(const struct sock *ssk,
1b1c7a0e PK	220	const struct mptcp_addr_info *addr)
1b1c7a0e PK	221	{
d1ace2d9 KM	222	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
d1ace2d9 KM	223	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
926bdeab PK	224	struct mptcp_pm_data *pm = &msk->pm;
	225
	226	pr_debug("msk=%p remote_id=%d accept=%d", msk, addr->id,
	227	READ_ONCE(pm->accept_addr));
	228
d1ace2d9	229	mptcp_event_addr_announced(ssk, addr);
b911c97c	230
926bdeab PK	231	spin_lock_bh(&pm->lock);
926bdeab PK	232
4d25247d KM	233	if (mptcp_pm_is_userspace(msk)) {
	234	if (mptcp_userspace_pm_active(msk)) {
	235	mptcp_pm_announce_addr(msk, addr, true);
	236	mptcp_pm_add_addr_send_ack(msk);
	237	} else {
	238	__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_ADDADDRDROP);
	239	}
	240	} else if (!READ_ONCE(pm->accept_addr)) {
f7efc777	241	mptcp_pm_announce_addr(msk, addr, true);
84dfe367 GT	242	mptcp_pm_add_addr_send_ack(msk);
84dfe367 GT	243	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
926bdeab	244	pm->remote = *addr;
f73c1194 PA	245	} else {
f73c1194 PA	246	__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_ADDADDRDROP);
84dfe367	247	}
926bdeab PK	248
926bdeab PK	249	spin_unlock_bh(&pm->lock);
84dfe367 GT	250	}
84dfe367 GT	251
557963c3	252	void mptcp_pm_add_addr_echoed(struct mptcp_sock *msk,
90d93088	253	const struct mptcp_addr_info *addr)
557963c3 GT	254	{
	255	struct mptcp_pm_data *pm = &msk->pm;
	256
	257	pr_debug("msk=%p", msk);
	258
	259	spin_lock_bh(&pm->lock);
	260
	261	if (mptcp_lookup_anno_list_by_saddr(msk, addr) && READ_ONCE(pm->work_pending))
	262	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	263
	264	spin_unlock_bh(&pm->lock);
	265	}
	266
84dfe367 GT	267	void mptcp_pm_add_addr_send_ack(struct mptcp_sock *msk)
84dfe367 GT	268	{
b5a7acd3	269	if (!mptcp_pm_should_add_signal(msk))
84dfe367 GT	270	return;
	271
	272	mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_SEND_ACK);
1b1c7a0e PK	273	}
1b1c7a0e PK	274
5c4a824d GT	275	void mptcp_pm_rm_addr_received(struct mptcp_sock *msk,
5c4a824d GT	276	const struct mptcp_rm_list *rm_list)
d0876b22 GT	277	{
d0876b22 GT	278	struct mptcp_pm_data *pm = &msk->pm;
5c4a824d	279	u8 i;
d0876b22	280
5c4a824d	281	pr_debug("msk=%p remote_ids_nr=%d", msk, rm_list->nr);
d0876b22	282
5c4a824d GT	283	for (i = 0; i < rm_list->nr; i++)
5c4a824d GT	284	mptcp_event_addr_removed(msk, rm_list->ids[i]);
b911c97c	285
d0876b22	286	spin_lock_bh(&pm->lock);
f73c1194 PA	287	if (mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED))
	288	pm->rm_list_rx = *rm_list;
	289	else
	290	__MPTCP_INC_STATS(sock_net((struct sock *)msk), MPTCP_MIB_RMADDRDROP);
d0876b22 GT	291	spin_unlock_bh(&pm->lock);
	292	}
	293
43f5b111	294	void mptcp_pm_mp_prio_received(struct sock *ssk, u8 bkup)
40453a5c	295	{
43f5b111 PA	296	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	297	struct sock *sk = subflow->conn;
	298	struct mptcp_sock *msk;
40453a5c GT	299
40453a5c GT	300	pr_debug("subflow->backup=%d, bkup=%d\n", subflow->backup, bkup);
43f5b111	301	msk = mptcp_sk(sk);
ebc1e08f	302	if (subflow->backup != bkup)
43f5b111	303	subflow->backup = bkup;
b911c97c	304
43f5b111	305	mptcp_event(MPTCP_EVENT_SUB_PRIORITY, msk, ssk, GFP_ATOMIC);
40453a5c GT	306	}
40453a5c GT	307
5580d41b GT	308	void mptcp_pm_mp_fail_received(struct sock *sk, u64 fail_seq)
5580d41b GT	309	{
1e39e5a3 GT	310	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
	311	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
	312
5580d41b	313	pr_debug("fail_seq=%llu", fail_seq);
1e39e5a3	314
7b16871f	315	if (!READ_ONCE(msk->allow_infinite_fallback))
9c81be0d GT	316	return;
9c81be0d GT	317
76a13b31	318	if (!subflow->fail_tout) {
9c81be0d GT	319	pr_debug("send MP_FAIL response and infinite map");
	320
	321	subflow->send_mp_fail = 1;
1e39e5a3	322	subflow->send_infinite_map = 1;
76a13b31 GT	323	tcp_send_ack(sk);
76a13b31 GT	324	} else {
49fa1919	325	pr_debug("MP_FAIL response received");
76a13b31	326	WRITE_ONCE(subflow->fail_tout, 0);
9c81be0d	327	}
5580d41b GT	328	}
5580d41b GT	329
1b1c7a0e PK	330	/* path manager helpers */
1b1c7a0e PK	331
90d93088	332	bool mptcp_pm_add_addr_signal(struct mptcp_sock msk, const struct sk_buff skb,
1f5e9e2f	333	unsigned int opt_size, unsigned int remaining,
f462a446	334	struct mptcp_addr_info addr, bool echo,
af7939f3	335	bool *drop_other_suboptions)
1b1c7a0e	336	{
926bdeab	337	int ret = false;
119c0220	338	u8 add_addr;
f462a446	339	u8 family;
af7939f3	340	bool port;
926bdeab PK	341
	342	spin_lock_bh(&msk->pm.lock);
	343
	344	/* double check after the lock is acquired */
f643b803	345	if (!mptcp_pm_should_add_signal(msk))
926bdeab PK	346	goto out_unlock;
926bdeab PK	347
1f5e9e2f YL	348	/* always drop every other options for pure ack ADD_ADDR; this is a
	349	* plain dup-ack from TCP perspective. The other MPTCP-relevant info,
	350	* if any, will be carried by the 'original' TCP ack
	351	*/
	352	if (skb && skb_is_tcp_pure_ack(skb)) {
	353	remaining += opt_size;
	354	*drop_other_suboptions = true;
	355	}
	356
d91d322a	357	*echo = mptcp_pm_should_add_signal_echo(msk);
af7939f3	358	port = !!(*echo ? msk->pm.remote.port : msk->pm.local.port);
456afe01	359
f462a446	360	family = *echo ? msk->pm.remote.family : msk->pm.local.family;
af7939f3	361	if (remaining < mptcp_add_addr_len(family, *echo, port))
926bdeab PK	362	goto out_unlock;
926bdeab PK	363
f462a446 YL	364	if (*echo) {
f462a446 YL	365	*addr = msk->pm.remote;
119c0220	366	add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_ECHO);
f462a446 YL	367	} else {
f462a446 YL	368	*addr = msk->pm.local;
119c0220	369	add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_SIGNAL);
f462a446	370	}
119c0220	371	WRITE_ONCE(msk->pm.addr_signal, add_addr);
926bdeab PK	372	ret = true;
	373
	374	out_unlock:
	375	spin_unlock_bh(&msk->pm.lock);
	376	return ret;
1b1c7a0e PK	377	}
1b1c7a0e PK	378
5cb104ae	379	bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
6445e17a	380	struct mptcp_rm_list *rm_list)
5cb104ae	381	{
cbde2787	382	int ret = false, len;
119c0220	383	u8 rm_addr;
5cb104ae GT	384
	385	spin_lock_bh(&msk->pm.lock);
	386
	387	/* double check after the lock is acquired */
	388	if (!mptcp_pm_should_rm_signal(msk))
	389	goto out_unlock;
	390
119c0220	391	rm_addr = msk->pm.addr_signal & ~BIT(MPTCP_RM_ADDR_SIGNAL);
cbde2787 GT	392	len = mptcp_rm_addr_len(&msk->pm.rm_list_tx);
cbde2787 GT	393	if (len < 0) {
119c0220	394	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
cbde2787 GT	395	goto out_unlock;
	396	}
	397	if (remaining < len)
5cb104ae GT	398	goto out_unlock;
5cb104ae GT	399
cbde2787	400	*rm_list = msk->pm.rm_list_tx;
119c0220	401	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
5cb104ae GT	402	ret = true;
	403
	404	out_unlock:
	405	spin_unlock_bh(&msk->pm.lock);
	406	return ret;
	407	}
	408
1b1c7a0e PK	409	int mptcp_pm_get_local_id(struct mptcp_sock msk, struct sock_common skc)
1b1c7a0e PK	410	{
9bbec87e GT	411	struct mptcp_addr_info skc_local;
	412	struct mptcp_addr_info msk_local;
	413
	414	if (WARN_ON_ONCE(!msk))
	415	return -1;
	416
	417	/* The 0 ID mapping is defined by the first subflow, copied into the msk
	418	* addr
	419	*/
	420	mptcp_local_address((struct sock_common *)msk, &msk_local);
	421	mptcp_local_address((struct sock_common *)skc, &skc_local);
	422	if (mptcp_addresses_equal(&msk_local, &skc_local, false))
	423	return 0;
	424
	425	if (mptcp_pm_is_userspace(msk))
	426	return mptcp_userspace_pm_get_local_id(msk, &skc_local);
	427	return mptcp_pm_nl_get_local_id(msk, &skc_local);
1b1c7a0e PK	428	}
1b1c7a0e PK	429
f40be0db GT	430	int mptcp_pm_get_flags_and_ifindex_by_id(struct mptcp_sock *msk, unsigned int id,
	431	u8 flags, int ifindex)
	432	{
	433	*flags = 0;
	434	*ifindex = 0;
	435
	436	if (!id)
	437	return 0;
	438
	439	if (mptcp_pm_is_userspace(msk))
	440	return mptcp_userspace_pm_get_flags_and_ifindex_by_id(msk, id, flags, ifindex);
	441	return mptcp_pm_nl_get_flags_and_ifindex_by_id(msk, id, flags, ifindex);
	442	}
	443
6ba7ce89 GT	444	int mptcp_pm_set_flags(struct net net, struct nlattr token,
	445	struct mptcp_pm_addr_entry *loc,
	446	struct mptcp_pm_addr_entry *rem, u8 bkup)
	447	{
	448	if (token)
	449	return mptcp_userspace_pm_set_flags(net, token, loc, rem, bkup);
	450	return mptcp_pm_nl_set_flags(net, loc, bkup);
	451	}
	452
71b7dec2 PA	453	void mptcp_pm_subflow_chk_stale(const struct mptcp_sock msk, struct sock ssk)
	454	{
	455	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
	456	u32 rcv_tstamp = READ_ONCE(tcp_sk(ssk)->rcv_tstamp);
	457
	458	/* keep track of rtx periods with no progress */
	459	if (!subflow->stale_count) {
	460	subflow->stale_rcv_tstamp = rcv_tstamp;
	461	subflow->stale_count++;
	462	} else if (subflow->stale_rcv_tstamp == rcv_tstamp) {
	463	if (subflow->stale_count < U8_MAX)
	464	subflow->stale_count++;
ff5a0b42	465	mptcp_pm_nl_subflow_chk_stale(msk, ssk);
71b7dec2 PA	466	} else {
71b7dec2 PA	467	subflow->stale_count = 0;
ff5a0b42	468	mptcp_subflow_set_active(subflow);
71b7dec2 PA	469	}
	470	}
	471
fb00ee4f MB	472	/* if sk is ipv4 or ipv6_only allows only same-family local and remote addresses,
	473	* otherwise allow any matching local/remote pair
	474	*/
	475	bool mptcp_pm_addr_families_match(const struct sock *sk,
	476	const struct mptcp_addr_info *loc,
	477	const struct mptcp_addr_info *rem)
	478	{
	479	bool mptcp_is_v4 = sk->sk_family == AF_INET;
	480
	481	#if IS_ENABLED(CONFIG_MPTCP_IPV6)
	482	bool loc_is_v4 = loc->family == AF_INET \|\| ipv6_addr_v4mapped(&loc->addr6);
	483	bool rem_is_v4 = rem->family == AF_INET \|\| ipv6_addr_v4mapped(&rem->addr6);
	484
	485	if (mptcp_is_v4)
	486	return loc_is_v4 && rem_is_v4;
	487
	488	if (ipv6_only_sock(sk))
	489	return !loc_is_v4 && !rem_is_v4;
	490
	491	return loc_is_v4 == rem_is_v4;
	492	#else
	493	return mptcp_is_v4 && loc->family == AF_INET && rem->family == AF_INET;
	494	#endif
	495	}
	496
b29fcfb5	497	void mptcp_pm_data_reset(struct mptcp_sock *msk)
1b1c7a0e	498	{
6bb63ccc	499	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
9273b9d5	500	struct mptcp_pm_data *pm = &msk->pm;
1b1c7a0e	501
9273b9d5 MM	502	pm->add_addr_signaled = 0;
	503	pm->add_addr_accepted = 0;
	504	pm->local_addr_used = 0;
	505	pm->subflows = 0;
	506	pm->rm_list_tx.nr = 0;
	507	pm->rm_list_rx.nr = 0;
6bb63ccc MM	508	WRITE_ONCE(pm->pm_type, pm_type);
	509
	510	if (pm_type == MPTCP_PM_TYPE_KERNEL) {
	511	bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
	512
	513	/* pm->work_pending must be only be set to 'true' when
	514	* pm->pm_type is set to MPTCP_PM_TYPE_KERNEL
	515	*/
	516	WRITE_ONCE(pm->work_pending,
	517	(!!mptcp_pm_get_local_addr_max(msk) &&
	518	subflows_allowed) \|\|
	519	!!mptcp_pm_get_add_addr_signal_max(msk));
	520	WRITE_ONCE(pm->accept_addr,
	521	!!mptcp_pm_get_add_addr_accept_max(msk) &&
	522	subflows_allowed);
	523	WRITE_ONCE(pm->accept_subflow, subflows_allowed);
	524	} else {
	525	WRITE_ONCE(pm->work_pending, 0);
	526	WRITE_ONCE(pm->accept_addr, 0);
	527	WRITE_ONCE(pm->accept_subflow, 0);
	528	}
	529
9273b9d5	530	WRITE_ONCE(pm->addr_signal, 0);
9273b9d5 MM	531	WRITE_ONCE(pm->remote_deny_join_id0, false);
	532	pm->status = 0;
	533	bitmap_fill(msk->pm.id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
b29fcfb5 PA	534	}
	535
	536	void mptcp_pm_data_init(struct mptcp_sock *msk)
	537	{
1b1c7a0e	538	spin_lock_init(&msk->pm.lock);
b6c08380	539	INIT_LIST_HEAD(&msk->pm.anno_list);
4638de5a	540	INIT_LIST_HEAD(&msk->pm.userspace_pm_local_addr_list);
b29fcfb5	541	mptcp_pm_data_reset(msk);
1b1c7a0e PK	542	}
1b1c7a0e PK	543
d39dceca	544	void __init mptcp_pm_init(void)
1b1c7a0e	545	{
01cacb00	546	mptcp_pm_nl_init();
1b1c7a0e	547	}