[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"

static struct workqueue_struct *pm_wq;

/* path manager command handlers */

int mptcp_pm_announce_addr(struct mptcp_sock *msk,
			   const struct mptcp_addr_info *addr)
{
	pr_debug("msk=%p, local_id=%d", msk, addr->id);

	msk->pm.local = *addr;
	WRITE_ONCE(msk->pm.addr_signal, true);
	return 0;
}

int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id)
{
	return -ENOTSUPP;
}

int mptcp_pm_remove_subflow(struct mptcp_sock *msk, u8 remote_id)
{
	return -ENOTSUPP;
}

/* 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;
	int ret;

	pr_debug("msk=%p subflows=%d max=%d allow=%d", msk, pm->subflows,
		 pm->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);
	ret = pm->subflows < pm->subflows_max;
	if (ret && ++pm->subflows == pm->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);
	if (queue_work(pm_wq, &msk->pm.work))
		sock_hold((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);

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

	/* avoid acquiring the lock if there is no room for fouther addresses */
	if (!READ_ONCE(pm->accept_addr))
		return;

	spin_lock_bh(&pm->lock);

	/* be sure there is something to signal re-checking under PM lock */
	if (READ_ONCE(pm->accept_addr) &&
	    mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED))
		pm->remote = *addr;

	spin_unlock_bh(&pm->lock);
}

/* path manager helpers */

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

	spin_lock_bh(&msk->pm.lock);

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

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

	*saddr = msk->pm.local;
	WRITE_ONCE(msk->pm.addr_signal, false);
	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);
}

static void pm_worker(struct work_struct *work)
{
	struct mptcp_pm_data *pm = container_of(work, struct mptcp_pm_data,
						work);
	struct mptcp_sock *msk = container_of(pm, struct mptcp_sock, pm);
	struct sock *sk = (struct sock *)msk;

	lock_sock(sk);
	spin_lock_bh(&msk->pm.lock);

	pr_debug("msk=%p status=%x", msk, pm->status);
	if (pm->status & BIT(MPTCP_PM_ADD_ADDR_RECEIVED)) {
		pm->status &= ~BIT(MPTCP_PM_ADD_ADDR_RECEIVED);
		mptcp_pm_nl_add_addr_received(msk);
	}
	if (pm->status & BIT(MPTCP_PM_ESTABLISHED)) {
		pm->status &= ~BIT(MPTCP_PM_ESTABLISHED);
		mptcp_pm_nl_fully_established(msk);
	}
	if (pm->status & BIT(MPTCP_PM_SUBFLOW_ESTABLISHED)) {
		pm->status &= ~BIT(MPTCP_PM_SUBFLOW_ESTABLISHED);
		mptcp_pm_nl_subflow_established(msk);
	}

	spin_unlock_bh(&msk->pm.lock);
	release_sock(sk);
	sock_put(sk);
}

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;
	WRITE_ONCE(msk->pm.work_pending, false);
	WRITE_ONCE(msk->pm.addr_signal, false);
	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_WORK(&msk->pm.work, pm_worker);

	mptcp_pm_nl_data_init(msk);
}

void mptcp_pm_close(struct mptcp_sock *msk)
{
	if (cancel_work_sync(&msk->pm.work))
		sock_put((struct sock *)msk);
}

void mptcp_pm_init(void)
{
	pm_wq = alloc_workqueue("pm_wq", WQ_UNBOUND | WQ_MEM_RECLAIM, 8);
	if (!pm_wq)
		panic("Failed to allocate workqueue");

	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
	13	static struct workqueue_struct *pm_wq;
	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	{
926bdeab PK	20	pr_debug("msk=%p, local_id=%d", msk, addr->id);
	21
	22	msk->pm.local = *addr;
	23	WRITE_ONCE(msk->pm.addr_signal, true);
	24	return 0;
1b1c7a0e PK	25	}
	26
	27	int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id)
	28	{
	29	return -ENOTSUPP;
	30	}
	31
	32	int mptcp_pm_remove_subflow(struct mptcp_sock *msk, u8 remote_id)
	33	{
	34	return -ENOTSUPP;
	35	}
	36
	37	/* path manager event handlers */
	38
	39	void mptcp_pm_new_connection(struct mptcp_sock *msk, int server_side)
	40	{
	41	struct mptcp_pm_data *pm = &msk->pm;
	42
	43	pr_debug("msk=%p, token=%u side=%d", msk, msk->token, server_side);
	44
	45	WRITE_ONCE(pm->server_side, server_side);
	46	}
	47
	48	bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
	49	{
926bdeab PK	50	struct mptcp_pm_data *pm = &msk->pm;
	51	int ret;
	52
	53	pr_debug("msk=%p subflows=%d max=%d allow=%d", msk, pm->subflows,
	54	pm->subflows_max, READ_ONCE(pm->accept_subflow));
	55
	56	/* try to avoid acquiring the lock below */
	57	if (!READ_ONCE(pm->accept_subflow))
	58	return false;
	59
	60	spin_lock_bh(&pm->lock);
	61	ret = pm->subflows < pm->subflows_max;
	62	if (ret && ++pm->subflows == pm->subflows_max)
	63	WRITE_ONCE(pm->accept_subflow, false);
	64	spin_unlock_bh(&pm->lock);
	65
	66	return ret;
	67	}
	68
	69	/* return true if the new status bit is currently cleared, that is, this event
	70	* can be server, eventually by an already scheduled work
	71	*/
	72	static bool mptcp_pm_schedule_work(struct mptcp_sock *msk,
	73	enum mptcp_pm_status new_status)
	74	{
	75	pr_debug("msk=%p status=%x new=%lx", msk, msk->pm.status,
	76	BIT(new_status));
	77	if (msk->pm.status & BIT(new_status))
	78	return false;
	79
	80	msk->pm.status \|= BIT(new_status);
	81	if (queue_work(pm_wq, &msk->pm.work))
	82	sock_hold((struct sock *)msk);
	83	return true;
1b1c7a0e PK	84	}
	85
	86	void mptcp_pm_fully_established(struct mptcp_sock *msk)
	87	{
926bdeab PK	88	struct mptcp_pm_data *pm = &msk->pm;
926bdeab PK	89
1b1c7a0e	90	pr_debug("msk=%p", msk);
926bdeab PK	91
	92	/* try to avoid acquiring the lock below */
	93	if (!READ_ONCE(pm->work_pending))
	94	return;
	95
	96	spin_lock_bh(&pm->lock);
	97
	98	if (READ_ONCE(pm->work_pending))
	99	mptcp_pm_schedule_work(msk, MPTCP_PM_ESTABLISHED);
	100
	101	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	102	}
	103
	104	void mptcp_pm_connection_closed(struct mptcp_sock *msk)
	105	{
	106	pr_debug("msk=%p", msk);
	107	}
	108
	109	void mptcp_pm_subflow_established(struct mptcp_sock *msk,
	110	struct mptcp_subflow_context *subflow)
	111	{
926bdeab PK	112	struct mptcp_pm_data *pm = &msk->pm;
926bdeab PK	113
1b1c7a0e	114	pr_debug("msk=%p", msk);
926bdeab PK	115
	116	if (!READ_ONCE(pm->work_pending))
	117	return;
	118
	119	spin_lock_bh(&pm->lock);
	120
	121	if (READ_ONCE(pm->work_pending))
	122	mptcp_pm_schedule_work(msk, MPTCP_PM_SUBFLOW_ESTABLISHED);
	123
	124	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	125	}
	126
	127	void mptcp_pm_subflow_closed(struct mptcp_sock *msk, u8 id)
	128	{
	129	pr_debug("msk=%p", msk);
	130	}
	131
	132	void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
	133	const struct mptcp_addr_info *addr)
	134	{
926bdeab PK	135	struct mptcp_pm_data *pm = &msk->pm;
	136
	137	pr_debug("msk=%p remote_id=%d accept=%d", msk, addr->id,
	138	READ_ONCE(pm->accept_addr));
	139
	140	/* avoid acquiring the lock if there is no room for fouther addresses */
	141	if (!READ_ONCE(pm->accept_addr))
	142	return;
	143
	144	spin_lock_bh(&pm->lock);
	145
	146	/* be sure there is something to signal re-checking under PM lock */
	147	if (READ_ONCE(pm->accept_addr) &&
	148	mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED))
	149	pm->remote = *addr;
	150
	151	spin_unlock_bh(&pm->lock);
1b1c7a0e PK	152	}
	153
	154	/* path manager helpers */
	155
	156	bool mptcp_pm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
	157	struct mptcp_addr_info *saddr)
	158	{
926bdeab PK	159	int ret = false;
	160
	161	spin_lock_bh(&msk->pm.lock);
	162
	163	/* double check after the lock is acquired */
	164	if (!mptcp_pm_should_signal(msk))
	165	goto out_unlock;
	166
	167	if (remaining < mptcp_add_addr_len(msk->pm.local.family))
	168	goto out_unlock;
	169
	170	*saddr = msk->pm.local;
	171	WRITE_ONCE(msk->pm.addr_signal, false);
	172	ret = true;
	173
	174	out_unlock:
	175	spin_unlock_bh(&msk->pm.lock);
	176	return ret;
1b1c7a0e PK	177	}
	178
	179	int mptcp_pm_get_local_id(struct mptcp_sock msk, struct sock_common skc)
	180	{
01cacb00	181	return mptcp_pm_nl_get_local_id(msk, skc);
1b1c7a0e PK	182	}
	183
	184	static void pm_worker(struct work_struct *work)
	185	{
926bdeab PK	186	struct mptcp_pm_data *pm = container_of(work, struct mptcp_pm_data,
	187	work);
	188	struct mptcp_sock *msk = container_of(pm, struct mptcp_sock, pm);
	189	struct sock sk = (struct sock )msk;
	190
	191	lock_sock(sk);
	192	spin_lock_bh(&msk->pm.lock);
	193
	194	pr_debug("msk=%p status=%x", msk, pm->status);
	195	if (pm->status & BIT(MPTCP_PM_ADD_ADDR_RECEIVED)) {
	196	pm->status &= ~BIT(MPTCP_PM_ADD_ADDR_RECEIVED);
01cacb00	197	mptcp_pm_nl_add_addr_received(msk);
926bdeab PK	198	}
	199	if (pm->status & BIT(MPTCP_PM_ESTABLISHED)) {
	200	pm->status &= ~BIT(MPTCP_PM_ESTABLISHED);
01cacb00	201	mptcp_pm_nl_fully_established(msk);
926bdeab PK	202	}
	203	if (pm->status & BIT(MPTCP_PM_SUBFLOW_ESTABLISHED)) {
	204	pm->status &= ~BIT(MPTCP_PM_SUBFLOW_ESTABLISHED);
01cacb00	205	mptcp_pm_nl_subflow_established(msk);
926bdeab PK	206	}
	207
	208	spin_unlock_bh(&msk->pm.lock);
	209	release_sock(sk);
	210	sock_put(sk);
1b1c7a0e PK	211	}
	212
	213	void mptcp_pm_data_init(struct mptcp_sock *msk)
	214	{
	215	msk->pm.add_addr_signaled = 0;
	216	msk->pm.add_addr_accepted = 0;
	217	msk->pm.local_addr_used = 0;
	218	msk->pm.subflows = 0;
	219	WRITE_ONCE(msk->pm.work_pending, false);
	220	WRITE_ONCE(msk->pm.addr_signal, false);
	221	WRITE_ONCE(msk->pm.accept_addr, false);
	222	WRITE_ONCE(msk->pm.accept_subflow, false);
	223	msk->pm.status = 0;
	224
	225	spin_lock_init(&msk->pm.lock);
	226	INIT_WORK(&msk->pm.work, pm_worker);
01cacb00 PA	227
01cacb00 PA	228	mptcp_pm_nl_data_init(msk);
1b1c7a0e PK	229	}
1b1c7a0e PK	230
926bdeab PK	231	void mptcp_pm_close(struct mptcp_sock *msk)
	232	{
	233	if (cancel_work_sync(&msk->pm.work))
	234	sock_put((struct sock *)msk);
	235	}
	236
1b1c7a0e PK	237	void mptcp_pm_init(void)
	238	{
	239	pm_wq = alloc_workqueue("pm_wq", WQ_UNBOUND \| WQ_MEM_RECLAIM, 8);
	240	if (!pm_wq)
	241	panic("Failed to allocate workqueue");
01cacb00 PA	242
01cacb00 PA	243	mptcp_pm_nl_init();
1b1c7a0e	244	}