[mirror_ubuntu-jammy-kernel.git] / net / smc / smc_close.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  Shared Memory Communications over RDMA (SMC-R) and RoCE
 *
 *  Socket Closing - normal and abnormal
 *
 *  Copyright IBM Corp. 2016
 *
 *  Author(s):  Ursula Braun <ubraun@linux.vnet.ibm.com>
 */

#include <linux/workqueue.h>
#include <linux/sched/signal.h>

#include <net/sock.h>
#include <net/tcp.h>

#include "smc.h"
#include "smc_tx.h"
#include "smc_cdc.h"
#include "smc_close.h"

/* release the clcsock that is assigned to the smc_sock */
void smc_clcsock_release(struct smc_sock *smc)
{
	struct socket *tcp;

	if (smc->listen_smc && current_work() != &smc->smc_listen_work)
		cancel_work_sync(&smc->smc_listen_work);
	mutex_lock(&smc->clcsock_release_lock);
	if (smc->clcsock) {
		tcp = smc->clcsock;
		smc->clcsock = NULL;
		sock_release(tcp);
	}
	mutex_unlock(&smc->clcsock_release_lock);
}

static void smc_close_cleanup_listen(struct sock *parent)
{
	struct sock *sk;

	/* Close non-accepted connections */
	while ((sk = smc_accept_dequeue(parent, NULL)))
		smc_close_non_accepted(sk);
}

/* wait for sndbuf data being transmitted */
static void smc_close_stream_wait(struct smc_sock *smc, long timeout)
{
	DEFINE_WAIT_FUNC(wait, woken_wake_function);
	struct sock *sk = &smc->sk;

	if (!timeout)
		return;

	if (!smc_tx_prepared_sends(&smc->conn))
		return;

	smc->wait_close_tx_prepared = 1;
	add_wait_queue(sk_sleep(sk), &wait);
	while (!signal_pending(current) && timeout) {
		int rc;

		rc = sk_wait_event(sk, &timeout,
				   !smc_tx_prepared_sends(&smc->conn) ||
				   sk->sk_err == ECONNABORTED ||
				   sk->sk_err == ECONNRESET ||
				   smc->conn.killed,
				   &wait);
		if (rc)
			break;
	}
	remove_wait_queue(sk_sleep(sk), &wait);
	smc->wait_close_tx_prepared = 0;
}

void smc_close_wake_tx_prepared(struct smc_sock *smc)
{
	if (smc->wait_close_tx_prepared)
		/* wake up socket closing */
		smc->sk.sk_state_change(&smc->sk);
}

static int smc_close_wr(struct smc_connection *conn)
{
	conn->local_tx_ctrl.conn_state_flags.peer_done_writing = 1;

	return smc_cdc_get_slot_and_msg_send(conn);
}

static int smc_close_final(struct smc_connection *conn)
{
	if (atomic_read(&conn->bytes_to_rcv))
		conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	else
		conn->local_tx_ctrl.conn_state_flags.peer_conn_closed = 1;
	if (conn->killed)
		return -EPIPE;

	return smc_cdc_get_slot_and_msg_send(conn);
}

int smc_close_abort(struct smc_connection *conn)
{
	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;

	return smc_cdc_get_slot_and_msg_send(conn);
}

static void smc_close_cancel_work(struct smc_sock *smc)
{
	struct sock *sk = &smc->sk;

	release_sock(sk);
	cancel_work_sync(&smc->conn.close_work);
	cancel_delayed_work_sync(&smc->conn.tx_work);
	lock_sock(sk);
}

/* terminate smc socket abnormally - active abort
 * link group is terminated, i.e. RDMA communication no longer possible
 */
void smc_close_active_abort(struct smc_sock *smc)
{
	struct sock *sk = &smc->sk;
	bool release_clcsock = false;

	if (sk->sk_state != SMC_INIT && smc->clcsock && smc->clcsock->sk) {
		sk->sk_err = ECONNABORTED;
		if (smc->clcsock && smc->clcsock->sk)
			tcp_abort(smc->clcsock->sk, ECONNABORTED);
	}
	switch (sk->sk_state) {
	case SMC_ACTIVE:
	case SMC_APPCLOSEWAIT1:
	case SMC_APPCLOSEWAIT2:
		sk->sk_state = SMC_PEERABORTWAIT;
		smc_close_cancel_work(smc);
		if (sk->sk_state != SMC_PEERABORTWAIT)
			break;
		sk->sk_state = SMC_CLOSED;
		sock_put(sk); /* (postponed) passive closing */
		break;
	case SMC_PEERCLOSEWAIT1:
	case SMC_PEERCLOSEWAIT2:
	case SMC_PEERFINCLOSEWAIT:
		sk->sk_state = SMC_PEERABORTWAIT;
		smc_close_cancel_work(smc);
		if (sk->sk_state != SMC_PEERABORTWAIT)
			break;
		sk->sk_state = SMC_CLOSED;
		smc_conn_free(&smc->conn);
		release_clcsock = true;
		sock_put(sk); /* passive closing */
		break;
	case SMC_PROCESSABORT:
	case SMC_APPFINCLOSEWAIT:
		sk->sk_state = SMC_PEERABORTWAIT;
		smc_close_cancel_work(smc);
		if (sk->sk_state != SMC_PEERABORTWAIT)
			break;
		sk->sk_state = SMC_CLOSED;
		smc_conn_free(&smc->conn);
		release_clcsock = true;
		break;
	case SMC_INIT:
	case SMC_PEERABORTWAIT:
	case SMC_CLOSED:
		break;
	}

	sock_set_flag(sk, SOCK_DEAD);
	sk->sk_state_change(sk);

	if (release_clcsock) {
		release_sock(sk);
		smc_clcsock_release(smc);
		lock_sock(sk);
	}
}

static inline bool smc_close_sent_any_close(struct smc_connection *conn)
{
	return conn->local_tx_ctrl.conn_state_flags.peer_conn_abort ||
	       conn->local_tx_ctrl.conn_state_flags.peer_conn_closed;
}

int smc_close_active(struct smc_sock *smc)
{
	struct smc_cdc_conn_state_flags *txflags =
		&smc->conn.local_tx_ctrl.conn_state_flags;
	struct smc_connection *conn = &smc->conn;
	struct sock *sk = &smc->sk;
	int old_state;
	long timeout;
	int rc = 0;

	timeout = current->flags & PF_EXITING ?
		  0 : sock_flag(sk, SOCK_LINGER) ?
		      sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;

	old_state = sk->sk_state;
again:
	switch (sk->sk_state) {
	case SMC_INIT:
		sk->sk_state = SMC_CLOSED;
		break;
	case SMC_LISTEN:
		sk->sk_state = SMC_CLOSED;
		sk->sk_state_change(sk); /* wake up accept */
		if (smc->clcsock && smc->clcsock->sk) {
			smc->clcsock->sk->sk_data_ready = smc->clcsk_data_ready;
			smc->clcsock->sk->sk_user_data = NULL;
			rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
		}
		smc_close_cleanup_listen(sk);
		release_sock(sk);
		flush_work(&smc->tcp_listen_work);
		lock_sock(sk);
		break;
	case SMC_ACTIVE:
		smc_close_stream_wait(smc, timeout);
		release_sock(sk);
		cancel_delayed_work_sync(&conn->tx_work);
		lock_sock(sk);
		if (sk->sk_state == SMC_ACTIVE) {
			/* send close request */
			rc = smc_close_final(conn);
			sk->sk_state = SMC_PEERCLOSEWAIT1;

			/* actively shutdown clcsock before peer close it,
			 * prevent peer from entering TIME_WAIT state.
			 */
			if (smc->clcsock && smc->clcsock->sk)
				rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
		} else {
			/* peer event has changed the state */
			goto again;
		}
		break;
	case SMC_APPFINCLOSEWAIT:
		/* socket already shutdown wr or both (active close) */
		if (txflags->peer_done_writing &&
		    !smc_close_sent_any_close(conn)) {
			/* just shutdown wr done, send close request */
			rc = smc_close_final(conn);
		}
		sk->sk_state = SMC_CLOSED;
		break;
	case SMC_APPCLOSEWAIT1:
	case SMC_APPCLOSEWAIT2:
		if (!smc_cdc_rxed_any_close(conn))
			smc_close_stream_wait(smc, timeout);
		release_sock(sk);
		cancel_delayed_work_sync(&conn->tx_work);
		lock_sock(sk);
		if (sk->sk_state != SMC_APPCLOSEWAIT1 &&
		    sk->sk_state != SMC_APPCLOSEWAIT2)
			goto again;
		/* confirm close from peer */
		rc = smc_close_final(conn);
		if (smc_cdc_rxed_any_close(conn)) {
			/* peer has closed the socket already */
			sk->sk_state = SMC_CLOSED;
			sock_put(sk); /* postponed passive closing */
		} else {
			/* peer has just issued a shutdown write */
			sk->sk_state = SMC_PEERFINCLOSEWAIT;
		}
		break;
	case SMC_PEERCLOSEWAIT1:
	case SMC_PEERCLOSEWAIT2:
		if (txflags->peer_done_writing &&
		    !smc_close_sent_any_close(conn)) {
			/* just shutdown wr done, send close request */
			rc = smc_close_final(conn);
		}
		/* peer sending PeerConnectionClosed will cause transition */
		break;
	case SMC_PEERFINCLOSEWAIT:
		/* peer sending PeerConnectionClosed will cause transition */
		break;
	case SMC_PROCESSABORT:
		rc = smc_close_abort(conn);
		sk->sk_state = SMC_CLOSED;
		break;
	case SMC_PEERABORTWAIT:
		sk->sk_state = SMC_CLOSED;
		break;
	case SMC_CLOSED:
		/* nothing to do, add tracing in future patch */
		break;
	}

	if (old_state != sk->sk_state)
		sk->sk_state_change(sk);
	return rc;
}

static void smc_close_passive_abort_received(struct smc_sock *smc)
{
	struct smc_cdc_conn_state_flags *txflags =
		&smc->conn.local_tx_ctrl.conn_state_flags;
	struct sock *sk = &smc->sk;

	switch (sk->sk_state) {
	case SMC_INIT:
	case SMC_ACTIVE:
	case SMC_APPCLOSEWAIT1:
		sk->sk_state = SMC_PROCESSABORT;
		sock_put(sk); /* passive closing */
		break;
	case SMC_APPFINCLOSEWAIT:
		sk->sk_state = SMC_PROCESSABORT;
		break;
	case SMC_PEERCLOSEWAIT1:
	case SMC_PEERCLOSEWAIT2:
		if (txflags->peer_done_writing &&
		    !smc_close_sent_any_close(&smc->conn))
			/* just shutdown, but not yet closed locally */
			sk->sk_state = SMC_PROCESSABORT;
		else
			sk->sk_state = SMC_CLOSED;
		sock_put(sk); /* passive closing */
		break;
	case SMC_APPCLOSEWAIT2:
	case SMC_PEERFINCLOSEWAIT:
		sk->sk_state = SMC_CLOSED;
		sock_put(sk); /* passive closing */
		break;
	case SMC_PEERABORTWAIT:
		sk->sk_state = SMC_CLOSED;
		break;
	case SMC_PROCESSABORT:
	/* nothing to do, add tracing in future patch */
		break;
	}
}

/* Either some kind of closing has been received: peer_conn_closed,
 * peer_conn_abort, or peer_done_writing
 * or the link group of the connection terminates abnormally.
 */
static void smc_close_passive_work(struct work_struct *work)
{
	struct smc_connection *conn = container_of(work,
						   struct smc_connection,
						   close_work);
	struct smc_sock *smc = container_of(conn, struct smc_sock, conn);
	struct smc_cdc_conn_state_flags *rxflags;
	bool release_clcsock = false;
	struct sock *sk = &smc->sk;
	int old_state;

	lock_sock(sk);
	old_state = sk->sk_state;

	rxflags = &conn->local_rx_ctrl.conn_state_flags;
	if (rxflags->peer_conn_abort) {
		/* peer has not received all data */
		smc_close_passive_abort_received(smc);
		release_sock(&smc->sk);
		cancel_delayed_work_sync(&conn->tx_work);
		lock_sock(&smc->sk);
		goto wakeup;
	}

	switch (sk->sk_state) {
	case SMC_INIT:
		sk->sk_state = SMC_APPCLOSEWAIT1;
		break;
	case SMC_ACTIVE:
		sk->sk_state = SMC_APPCLOSEWAIT1;
		/* postpone sock_put() for passive closing to cover
		 * received SEND_SHUTDOWN as well
		 */
		break;
	case SMC_PEERCLOSEWAIT1:
		if (rxflags->peer_done_writing)
			sk->sk_state = SMC_PEERCLOSEWAIT2;
		fallthrough;
		/* to check for closing */
	case SMC_PEERCLOSEWAIT2:
		if (!smc_cdc_rxed_any_close(conn))
			break;
		if (sock_flag(sk, SOCK_DEAD) &&
		    smc_close_sent_any_close(conn)) {
			/* smc_release has already been called locally */
			sk->sk_state = SMC_CLOSED;
		} else {
			/* just shutdown, but not yet closed locally */
			sk->sk_state = SMC_APPFINCLOSEWAIT;
		}
		sock_put(sk); /* passive closing */
		break;
	case SMC_PEERFINCLOSEWAIT:
		if (smc_cdc_rxed_any_close(conn)) {
			sk->sk_state = SMC_CLOSED;
			sock_put(sk); /* passive closing */
		}
		break;
	case SMC_APPCLOSEWAIT1:
	case SMC_APPCLOSEWAIT2:
		/* postpone sock_put() for passive closing to cover
		 * received SEND_SHUTDOWN as well
		 */
		break;
	case SMC_APPFINCLOSEWAIT:
	case SMC_PEERABORTWAIT:
	case SMC_PROCESSABORT:
	case SMC_CLOSED:
		/* nothing to do, add tracing in future patch */
		break;
	}

wakeup:
	sk->sk_data_ready(sk); /* wakeup blocked rcvbuf consumers */
	sk->sk_write_space(sk); /* wakeup blocked sndbuf producers */

	if (old_state != sk->sk_state) {
		sk->sk_state_change(sk);
		if ((sk->sk_state == SMC_CLOSED) &&
		    (sock_flag(sk, SOCK_DEAD) || !sk->sk_socket)) {
			smc_conn_free(conn);
			if (smc->clcsock)
				release_clcsock = true;
		}
	}
	release_sock(sk);
	if (release_clcsock)
		smc_clcsock_release(smc);
	sock_put(sk); /* sock_hold done by schedulers of close_work */
}

int smc_close_shutdown_write(struct smc_sock *smc)
{
	struct smc_connection *conn = &smc->conn;
	struct sock *sk = &smc->sk;
	int old_state;
	long timeout;
	int rc = 0;

	timeout = current->flags & PF_EXITING ?
		  0 : sock_flag(sk, SOCK_LINGER) ?
		      sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;

	old_state = sk->sk_state;
again:
	switch (sk->sk_state) {
	case SMC_ACTIVE:
		smc_close_stream_wait(smc, timeout);
		release_sock(sk);
		cancel_delayed_work_sync(&conn->tx_work);
		lock_sock(sk);
		if (sk->sk_state != SMC_ACTIVE)
			goto again;
		/* send close wr request */
		rc = smc_close_wr(conn);
		sk->sk_state = SMC_PEERCLOSEWAIT1;
		break;
	case SMC_APPCLOSEWAIT1:
		/* passive close */
		if (!smc_cdc_rxed_any_close(conn))
			smc_close_stream_wait(smc, timeout);
		release_sock(sk);
		cancel_delayed_work_sync(&conn->tx_work);
		lock_sock(sk);
		if (sk->sk_state != SMC_APPCLOSEWAIT1)
			goto again;
		/* confirm close from peer */
		rc = smc_close_wr(conn);
		sk->sk_state = SMC_APPCLOSEWAIT2;
		break;
	case SMC_APPCLOSEWAIT2:
	case SMC_PEERFINCLOSEWAIT:
	case SMC_PEERCLOSEWAIT1:
	case SMC_PEERCLOSEWAIT2:
	case SMC_APPFINCLOSEWAIT:
	case SMC_PROCESSABORT:
	case SMC_PEERABORTWAIT:
		/* nothing to do, add tracing in future patch */
		break;
	}

	if (old_state != sk->sk_state)
		sk->sk_state_change(sk);
	return rc;
}

/* Initialize close properties on connection establishment. */
void smc_close_init(struct smc_sock *smc)
{
	INIT_WORK(&smc->conn.close_work, smc_close_passive_work);
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
b38d7324 UB	2	/*
	3	* Shared Memory Communications over RDMA (SMC-R) and RoCE
	4	*
	5	* Socket Closing - normal and abnormal
	6	*
	7	* Copyright IBM Corp. 2016
	8	*
	9	* Author(s): Ursula Braun <ubraun@linux.vnet.ibm.com>
	10	*/
	11
	12	#include <linux/workqueue.h>
c3edc401 IM	13	#include <linux/sched/signal.h>
c3edc401 IM	14
b38d7324	15	#include <net/sock.h>
83179760	16	#include <net/tcp.h>
b38d7324 UB	17
	18	#include "smc.h"
	19	#include "smc_tx.h"
	20	#include "smc_cdc.h"
	21	#include "smc_close.h"
	22
fd57770d KG	23	/* release the clcsock that is assigned to the smc_sock */
	24	void smc_clcsock_release(struct smc_sock *smc)
	25	{
	26	struct socket *tcp;
	27
	28	if (smc->listen_smc && current_work() != &smc->smc_listen_work)
	29	cancel_work_sync(&smc->smc_listen_work);
	30	mutex_lock(&smc->clcsock_release_lock);
	31	if (smc->clcsock) {
	32	tcp = smc->clcsock;
	33	smc->clcsock = NULL;
	34	sock_release(tcp);
	35	}
	36	mutex_unlock(&smc->clcsock_release_lock);
	37	}
	38
b38d7324 UB	39	static void smc_close_cleanup_listen(struct sock *parent)
	40	{
	41	struct sock *sk;
	42
	43	/* Close non-accepted connections */
	44	while ((sk = smc_accept_dequeue(parent, NULL)))
	45	smc_close_non_accepted(sk);
	46	}
	47
b38d7324 UB	48	/* wait for sndbuf data being transmitted */
	49	static void smc_close_stream_wait(struct smc_sock *smc, long timeout)
	50	{
	51	DEFINE_WAIT_FUNC(wait, woken_wake_function);
	52	struct sock *sk = &smc->sk;
	53
	54	if (!timeout)
	55	return;
	56
	57	if (!smc_tx_prepared_sends(&smc->conn))
	58	return;
	59
	60	smc->wait_close_tx_prepared = 1;
	61	add_wait_queue(sk_sleep(sk), &wait);
	62	while (!signal_pending(current) && timeout) {
	63	int rc;
	64
	65	rc = sk_wait_event(sk, &timeout,
	66	!smc_tx_prepared_sends(&smc->conn) \|\|
d18963cf	67	sk->sk_err == ECONNABORTED \|\|
b2900980 UB	68	sk->sk_err == ECONNRESET \|\|
b2900980 UB	69	smc->conn.killed,
b38d7324 UB	70	&wait);
	71	if (rc)
	72	break;
	73	}
	74	remove_wait_queue(sk_sleep(sk), &wait);
	75	smc->wait_close_tx_prepared = 0;
	76	}
	77
	78	void smc_close_wake_tx_prepared(struct smc_sock *smc)
	79	{
	80	if (smc->wait_close_tx_prepared)
	81	/* wake up socket closing */
	82	smc->sk.sk_state_change(&smc->sk);
	83	}
	84
	85	static int smc_close_wr(struct smc_connection *conn)
	86	{
	87	conn->local_tx_ctrl.conn_state_flags.peer_done_writing = 1;
	88
	89	return smc_cdc_get_slot_and_msg_send(conn);
	90	}
	91
	92	static int smc_close_final(struct smc_connection *conn)
	93	{
	94	if (atomic_read(&conn->bytes_to_rcv))
	95	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	96	else
	97	conn->local_tx_ctrl.conn_state_flags.peer_conn_closed = 1;
b2900980 UB	98	if (conn->killed)
b2900980 UB	99	return -EPIPE;
b38d7324 UB	100
	101	return smc_cdc_get_slot_and_msg_send(conn);
	102	}
	103
83179760	104	int smc_close_abort(struct smc_connection *conn)
b38d7324 UB	105	{
	106	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	107
	108	return smc_cdc_get_slot_and_msg_send(conn);
	109	}
	110
42bfba9e UB	111	static void smc_close_cancel_work(struct smc_sock *smc)
	112	{
	113	struct sock *sk = &smc->sk;
	114
	115	release_sock(sk);
	116	cancel_work_sync(&smc->conn.close_work);
	117	cancel_delayed_work_sync(&smc->conn.tx_work);
	118	lock_sock(sk);
42bfba9e UB	119	}
42bfba9e UB	120
b38d7324	121	/* terminate smc socket abnormally - active abort
732720fa	122	* link group is terminated, i.e. RDMA communication no longer possible
b38d7324	123	*/
81cf4f47	124	void smc_close_active_abort(struct smc_sock *smc)
b38d7324	125	{
3163c507	126	struct sock *sk = &smc->sk;
81cf4f47	127	bool release_clcsock = false;
3163c507	128
dd65d87a UB	129	if (sk->sk_state != SMC_INIT && smc->clcsock && smc->clcsock->sk) {
dd65d87a UB	130	sk->sk_err = ECONNABORTED;
83179760 UB	131	if (smc->clcsock && smc->clcsock->sk)
83179760 UB	132	tcp_abort(smc->clcsock->sk, ECONNABORTED);
b38d7324	133	}
3163c507	134	switch (sk->sk_state) {
46c28dbd	135	case SMC_ACTIVE:
b38d7324 UB	136	case SMC_APPCLOSEWAIT1:
b38d7324 UB	137	case SMC_APPCLOSEWAIT2:
5fb8642a	138	sk->sk_state = SMC_PEERABORTWAIT;
42bfba9e	139	smc_close_cancel_work(smc);
5fb8642a UB	140	if (sk->sk_state != SMC_PEERABORTWAIT)
5fb8642a UB	141	break;
d18963cf	142	sk->sk_state = SMC_CLOSED;
5fb8642a	143	sock_put(sk); /* (postponed) passive closing */
b38d7324 UB	144	break;
	145	case SMC_PEERCLOSEWAIT1:
	146	case SMC_PEERCLOSEWAIT2:
d18963cf	147	case SMC_PEERFINCLOSEWAIT:
42bfba9e UB	148	sk->sk_state = SMC_PEERABORTWAIT;
42bfba9e UB	149	smc_close_cancel_work(smc);
5fb8642a UB	150	if (sk->sk_state != SMC_PEERABORTWAIT)
5fb8642a UB	151	break;
d18963cf	152	sk->sk_state = SMC_CLOSED;
81cf4f47 UB	153	smc_conn_free(&smc->conn);
81cf4f47 UB	154	release_clcsock = true;
51f1de79	155	sock_put(sk); /* passive closing */
b38d7324 UB	156	break;
	157	case SMC_PROCESSABORT:
	158	case SMC_APPFINCLOSEWAIT:
42bfba9e UB	159	sk->sk_state = SMC_PEERABORTWAIT;
42bfba9e UB	160	smc_close_cancel_work(smc);
5fb8642a UB	161	if (sk->sk_state != SMC_PEERABORTWAIT)
5fb8642a UB	162	break;
3163c507	163	sk->sk_state = SMC_CLOSED;
42bfba9e UB	164	smc_conn_free(&smc->conn);
42bfba9e UB	165	release_clcsock = true;
b38d7324	166	break;
dd65d87a	167	case SMC_INIT:
b38d7324 UB	168	case SMC_PEERABORTWAIT:
	169	case SMC_CLOSED:
	170	break;
	171	}
	172
3163c507 UB	173	sock_set_flag(sk, SOCK_DEAD);
3163c507 UB	174	sk->sk_state_change(sk);
81cf4f47 UB	175
	176	if (release_clcsock) {
	177	release_sock(sk);
	178	smc_clcsock_release(smc);
	179	lock_sock(sk);
	180	}
b38d7324 UB	181	}
b38d7324 UB	182
a98bf8c0 UB	183	static inline bool smc_close_sent_any_close(struct smc_connection *conn)
	184	{
	185	return conn->local_tx_ctrl.conn_state_flags.peer_conn_abort \|\|
	186	conn->local_tx_ctrl.conn_state_flags.peer_conn_closed;
	187	}
	188
b38d7324 UB	189	int smc_close_active(struct smc_sock *smc)
	190	{
	191	struct smc_cdc_conn_state_flags *txflags =
	192	&smc->conn.local_tx_ctrl.conn_state_flags;
b38d7324 UB	193	struct smc_connection *conn = &smc->conn;
	194	struct sock *sk = &smc->sk;
	195	int old_state;
8c96feee	196	long timeout;
b38d7324 UB	197	int rc = 0;
b38d7324 UB	198
8c96feee UB	199	timeout = current->flags & PF_EXITING ?
	200	0 : sock_flag(sk, SOCK_LINGER) ?
	201	sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;
b38d7324	202
b38d7324	203	old_state = sk->sk_state;
bbb96bf2 UB	204	again:
bbb96bf2 UB	205	switch (sk->sk_state) {
b38d7324 UB	206	case SMC_INIT:
b38d7324 UB	207	sk->sk_state = SMC_CLOSED;
b38d7324 UB	208	break;
	209	case SMC_LISTEN:
	210	sk->sk_state = SMC_CLOSED;
	211	sk->sk_state_change(sk); /* wake up accept */
ddcc9b7f KG	212	if (smc->clcsock && smc->clcsock->sk) {
	213	smc->clcsock->sk->sk_data_ready = smc->clcsk_data_ready;
	214	smc->clcsock->sk->sk_user_data = NULL;
b38d7324	215	rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
ddcc9b7f	216	}
b38d7324	217	smc_close_cleanup_listen(sk);
3d502067 UB	218	release_sock(sk);
	219	flush_work(&smc->tcp_listen_work);
	220	lock_sock(sk);
b38d7324 UB	221	break;
	222	case SMC_ACTIVE:
	223	smc_close_stream_wait(smc, timeout);
	224	release_sock(sk);
18e537cd	225	cancel_delayed_work_sync(&conn->tx_work);
b38d7324 UB	226	lock_sock(sk);
	227	if (sk->sk_state == SMC_ACTIVE) {
	228	/* send close request */
	229	rc = smc_close_final(conn);
	230	sk->sk_state = SMC_PEERCLOSEWAIT1;
e24fce04 TL	231
	232	/* actively shutdown clcsock before peer close it,
	233	* prevent peer from entering TIME_WAIT state.
	234	*/
	235	if (smc->clcsock && smc->clcsock->sk)
	236	rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
b38d7324 UB	237	} else {
	238	/* peer event has changed the state */
	239	goto again;
	240	}
	241	break;
	242	case SMC_APPFINCLOSEWAIT:
	243	/* socket already shutdown wr or both (active close) */
	244	if (txflags->peer_done_writing &&
a98bf8c0	245	!smc_close_sent_any_close(conn)) {
b38d7324 UB	246	/* just shutdown wr done, send close request */
	247	rc = smc_close_final(conn);
	248	}
	249	sk->sk_state = SMC_CLOSED;
b38d7324 UB	250	break;
	251	case SMC_APPCLOSEWAIT1:
	252	case SMC_APPCLOSEWAIT2:
	253	if (!smc_cdc_rxed_any_close(conn))
	254	smc_close_stream_wait(smc, timeout);
	255	release_sock(sk);
18e537cd	256	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	257	lock_sock(sk);
bbb96bf2 UB	258	if (sk->sk_state != SMC_APPCLOSEWAIT1 &&
	259	sk->sk_state != SMC_APPCLOSEWAIT2)
	260	goto again;
	261	/* confirm close from peer */
	262	rc = smc_close_final(conn);
51f1de79	263	if (smc_cdc_rxed_any_close(conn)) {
b38d7324 UB	264	/* peer has closed the socket already */
b38d7324 UB	265	sk->sk_state = SMC_CLOSED;
51f1de79 UB	266	sock_put(sk); /* postponed passive closing */
51f1de79 UB	267	} else {
b38d7324 UB	268	/* peer has just issued a shutdown write */
b38d7324 UB	269	sk->sk_state = SMC_PEERFINCLOSEWAIT;
51f1de79	270	}
b38d7324 UB	271	break;
	272	case SMC_PEERCLOSEWAIT1:
	273	case SMC_PEERCLOSEWAIT2:
a98bf8c0 UB	274	if (txflags->peer_done_writing &&
	275	!smc_close_sent_any_close(conn)) {
	276	/* just shutdown wr done, send close request */
	277	rc = smc_close_final(conn);
	278	}
	279	/* peer sending PeerConnectionClosed will cause transition */
	280	break;
b38d7324 UB	281	case SMC_PEERFINCLOSEWAIT:
	282	/* peer sending PeerConnectionClosed will cause transition */
	283	break;
	284	case SMC_PROCESSABORT:
d18963cf	285	rc = smc_close_abort(conn);
b38d7324	286	sk->sk_state = SMC_CLOSED;
b38d7324 UB	287	break;
b38d7324 UB	288	case SMC_PEERABORTWAIT:
d18963cf UB	289	sk->sk_state = SMC_CLOSED;
d18963cf UB	290	break;
b38d7324 UB	291	case SMC_CLOSED:
	292	/* nothing to do, add tracing in future patch */
	293	break;
	294	}
	295
	296	if (old_state != sk->sk_state)
3163c507	297	sk->sk_state_change(sk);
b38d7324 UB	298	return rc;
	299	}
	300
	301	static void smc_close_passive_abort_received(struct smc_sock *smc)
	302	{
	303	struct smc_cdc_conn_state_flags *txflags =
	304	&smc->conn.local_tx_ctrl.conn_state_flags;
	305	struct sock *sk = &smc->sk;
	306
	307	switch (sk->sk_state) {
51f1de79	308	case SMC_INIT:
b38d7324	309	case SMC_ACTIVE:
b38d7324	310	case SMC_APPCLOSEWAIT1:
51f1de79 UB	311	sk->sk_state = SMC_PROCESSABORT;
	312	sock_put(sk); /* passive closing */
	313	break;
	314	case SMC_APPFINCLOSEWAIT:
b38d7324 UB	315	sk->sk_state = SMC_PROCESSABORT;
	316	break;
	317	case SMC_PEERCLOSEWAIT1:
	318	case SMC_PEERCLOSEWAIT2:
	319	if (txflags->peer_done_writing &&
51f1de79	320	!smc_close_sent_any_close(&smc->conn))
b38d7324	321	/* just shutdown, but not yet closed locally */
b38d7324	322	sk->sk_state = SMC_PROCESSABORT;
51f1de79	323	else
b38d7324	324	sk->sk_state = SMC_CLOSED;
51f1de79	325	sock_put(sk); /* passive closing */
b38d7324	326	break;
51f1de79	327	case SMC_APPCLOSEWAIT2:
b38d7324	328	case SMC_PEERFINCLOSEWAIT:
51f1de79 UB	329	sk->sk_state = SMC_CLOSED;
	330	sock_put(sk); /* passive closing */
	331	break;
b38d7324 UB	332	case SMC_PEERABORTWAIT:
	333	sk->sk_state = SMC_CLOSED;
	334	break;
b38d7324 UB	335	case SMC_PROCESSABORT:
	336	/* nothing to do, add tracing in future patch */
	337	break;
	338	}
	339	}
	340
732720fa UB	341	/* Either some kind of closing has been received: peer_conn_closed,
	342	* peer_conn_abort, or peer_done_writing
	343	* or the link group of the connection terminates abnormally.
b38d7324	344	*/
46c28dbd	345	static void smc_close_passive_work(struct work_struct *work)
b38d7324	346	{
46c28dbd UB	347	struct smc_connection *conn = container_of(work,
	348	struct smc_connection,
	349	close_work);
	350	struct smc_sock *smc = container_of(conn, struct smc_sock, conn);
	351	struct smc_cdc_conn_state_flags *rxflags;
fd57770d	352	bool release_clcsock = false;
b38d7324 UB	353	struct sock *sk = &smc->sk;
	354	int old_state;
	355
3163c507	356	lock_sock(sk);
b38d7324 UB	357	old_state = sk->sk_state;
b38d7324 UB	358
3163c507	359	rxflags = &conn->local_rx_ctrl.conn_state_flags;
b38d7324	360	if (rxflags->peer_conn_abort) {
732720fa	361	/* peer has not received all data */
b38d7324	362	smc_close_passive_abort_received(smc);
611b63a1 UB	363	release_sock(&smc->sk);
	364	cancel_delayed_work_sync(&conn->tx_work);
	365	lock_sock(&smc->sk);
b38d7324 UB	366	goto wakeup;
	367	}
	368
	369	switch (sk->sk_state) {
	370	case SMC_INIT:
84b799a2	371	sk->sk_state = SMC_APPCLOSEWAIT1;
b38d7324 UB	372	break;
	373	case SMC_ACTIVE:
	374	sk->sk_state = SMC_APPCLOSEWAIT1;
51f1de79 UB	375	/* postpone sock_put() for passive closing to cover
	376	* received SEND_SHUTDOWN as well
	377	*/
b38d7324 UB	378	break;
	379	case SMC_PEERCLOSEWAIT1:
	380	if (rxflags->peer_done_writing)
	381	sk->sk_state = SMC_PEERCLOSEWAIT2;
df561f66	382	fallthrough;
7f6b437e	383	/* to check for closing */
b38d7324	384	case SMC_PEERCLOSEWAIT2:
3163c507	385	if (!smc_cdc_rxed_any_close(conn))
b38d7324 UB	386	break;
b38d7324 UB	387	if (sock_flag(sk, SOCK_DEAD) &&
a98bf8c0	388	smc_close_sent_any_close(conn)) {
b38d7324 UB	389	/* smc_release has already been called locally */
	390	sk->sk_state = SMC_CLOSED;
	391	} else {
	392	/* just shutdown, but not yet closed locally */
	393	sk->sk_state = SMC_APPFINCLOSEWAIT;
	394	}
51f1de79	395	sock_put(sk); /* passive closing */
b38d7324	396	break;
5ac92a00	397	case SMC_PEERFINCLOSEWAIT:
51f1de79	398	if (smc_cdc_rxed_any_close(conn)) {
5ac92a00	399	sk->sk_state = SMC_CLOSED;
51f1de79 UB	400	sock_put(sk); /* passive closing */
51f1de79 UB	401	}
5ac92a00	402	break;
b38d7324 UB	403	case SMC_APPCLOSEWAIT1:
b38d7324 UB	404	case SMC_APPCLOSEWAIT2:
51f1de79 UB	405	/* postpone sock_put() for passive closing to cover
	406	* received SEND_SHUTDOWN as well
	407	*/
	408	break;
b38d7324 UB	409	case SMC_APPFINCLOSEWAIT:
	410	case SMC_PEERABORTWAIT:
	411	case SMC_PROCESSABORT:
	412	case SMC_CLOSED:
	413	/* nothing to do, add tracing in future patch */
	414	break;
	415	}
	416
	417	wakeup:
b38d7324 UB	418	sk->sk_data_ready(sk); /* wakeup blocked rcvbuf consumers */
	419	sk->sk_write_space(sk); /* wakeup blocked sndbuf producers */
	420
a98bf8c0 UB	421	if (old_state != sk->sk_state) {
	422	sk->sk_state_change(sk);
	423	if ((sk->sk_state == SMC_CLOSED) &&
b03faa1f	424	(sock_flag(sk, SOCK_DEAD) \|\| !sk->sk_socket)) {
3163c507	425	smc_conn_free(conn);
fd57770d KG	426	if (smc->clcsock)
fd57770d KG	427	release_clcsock = true;
b03faa1f	428	}
b38d7324	429	}
3163c507	430	release_sock(sk);
fd57770d KG	431	if (release_clcsock)
fd57770d KG	432	smc_clcsock_release(smc);
51f1de79	433	sock_put(sk); /* sock_hold done by schedulers of close_work */
b38d7324 UB	434	}
	435
	436	int smc_close_shutdown_write(struct smc_sock *smc)
	437	{
	438	struct smc_connection *conn = &smc->conn;
b38d7324 UB	439	struct sock *sk = &smc->sk;
b38d7324 UB	440	int old_state;
8c96feee	441	long timeout;
b38d7324 UB	442	int rc = 0;
b38d7324 UB	443
8c96feee UB	444	timeout = current->flags & PF_EXITING ?
	445	0 : sock_flag(sk, SOCK_LINGER) ?
	446	sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;
b38d7324	447
b38d7324	448	old_state = sk->sk_state;
bbb96bf2 UB	449	again:
bbb96bf2 UB	450	switch (sk->sk_state) {
b38d7324 UB	451	case SMC_ACTIVE:
	452	smc_close_stream_wait(smc, timeout);
	453	release_sock(sk);
18e537cd	454	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	455	lock_sock(sk);
bbb96bf2 UB	456	if (sk->sk_state != SMC_ACTIVE)
bbb96bf2 UB	457	goto again;
b38d7324 UB	458	/* send close wr request */
b38d7324 UB	459	rc = smc_close_wr(conn);
bbb96bf2	460	sk->sk_state = SMC_PEERCLOSEWAIT1;
b38d7324 UB	461	break;
	462	case SMC_APPCLOSEWAIT1:
	463	/* passive close */
	464	if (!smc_cdc_rxed_any_close(conn))
	465	smc_close_stream_wait(smc, timeout);
	466	release_sock(sk);
18e537cd	467	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	468	lock_sock(sk);
bbb96bf2 UB	469	if (sk->sk_state != SMC_APPCLOSEWAIT1)
bbb96bf2 UB	470	goto again;
b38d7324 UB	471	/* confirm close from peer */
	472	rc = smc_close_wr(conn);
	473	sk->sk_state = SMC_APPCLOSEWAIT2;
	474	break;
	475	case SMC_APPCLOSEWAIT2:
	476	case SMC_PEERFINCLOSEWAIT:
	477	case SMC_PEERCLOSEWAIT1:
	478	case SMC_PEERCLOSEWAIT2:
	479	case SMC_APPFINCLOSEWAIT:
	480	case SMC_PROCESSABORT:
	481	case SMC_PEERABORTWAIT:
	482	/* nothing to do, add tracing in future patch */
	483	break;
	484	}
	485
	486	if (old_state != sk->sk_state)
3163c507	487	sk->sk_state_change(sk);
b38d7324 UB	488	return rc;
b38d7324 UB	489	}
46c28dbd UB	490
	491	/* Initialize close properties on connection establishment. */
	492	void smc_close_init(struct smc_sock *smc)
	493	{
	494	INIT_WORK(&smc->conn.close_work, smc_close_passive_work);
	495	}