[linux-block.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"

#define SMC_CLOSE_WAIT_LISTEN_CLCSOCK_TIME	(5 * HZ)

/* 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);
}

/* 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:
		sk->sk_state = SMC_PEERABORTWAIT;
		release_sock(sk);
		cancel_delayed_work_sync(&smc->conn.tx_work);
		lock_sock(sk);
		sk->sk_state = SMC_CLOSED;
		sock_put(sk); /* passive closing */
		break;
	case SMC_APPCLOSEWAIT1:
	case SMC_APPCLOSEWAIT2:
		release_sock(sk);
		cancel_delayed_work_sync(&smc->conn.tx_work);
		lock_sock(sk);
		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_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_CLOSED;
		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) {
			rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
			/* wake up kernel_accept of smc_tcp_listen_worker */
			smc->clcsock->sk->sk_data_ready(smc->clcsock->sk);
		}
		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;
		} 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;
		/* fall through */
		/* 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
127f4970 UB	23	#define SMC_CLOSE_WAIT_LISTEN_CLCSOCK_TIME (5 * HZ)
127f4970 UB	24
fd57770d KG	25	/* release the clcsock that is assigned to the smc_sock */
	26	void smc_clcsock_release(struct smc_sock *smc)
	27	{
	28	struct socket *tcp;
	29
	30	if (smc->listen_smc && current_work() != &smc->smc_listen_work)
	31	cancel_work_sync(&smc->smc_listen_work);
	32	mutex_lock(&smc->clcsock_release_lock);
	33	if (smc->clcsock) {
	34	tcp = smc->clcsock;
	35	smc->clcsock = NULL;
	36	sock_release(tcp);
	37	}
	38	mutex_unlock(&smc->clcsock_release_lock);
	39	}
	40
b38d7324 UB	41	static void smc_close_cleanup_listen(struct sock *parent)
	42	{
	43	struct sock *sk;
	44
	45	/* Close non-accepted connections */
	46	while ((sk = smc_accept_dequeue(parent, NULL)))
	47	smc_close_non_accepted(sk);
	48	}
	49
b38d7324 UB	50	/* wait for sndbuf data being transmitted */
	51	static void smc_close_stream_wait(struct smc_sock *smc, long timeout)
	52	{
	53	DEFINE_WAIT_FUNC(wait, woken_wake_function);
	54	struct sock *sk = &smc->sk;
	55
	56	if (!timeout)
	57	return;
	58
	59	if (!smc_tx_prepared_sends(&smc->conn))
	60	return;
	61
	62	smc->wait_close_tx_prepared = 1;
	63	add_wait_queue(sk_sleep(sk), &wait);
	64	while (!signal_pending(current) && timeout) {
	65	int rc;
	66
	67	rc = sk_wait_event(sk, &timeout,
	68	!smc_tx_prepared_sends(&smc->conn) \|\|
d18963cf	69	sk->sk_err == ECONNABORTED \|\|
b2900980 UB	70	sk->sk_err == ECONNRESET \|\|
b2900980 UB	71	smc->conn.killed,
b38d7324 UB	72	&wait);
	73	if (rc)
	74	break;
	75	}
	76	remove_wait_queue(sk_sleep(sk), &wait);
	77	smc->wait_close_tx_prepared = 0;
	78	}
	79
	80	void smc_close_wake_tx_prepared(struct smc_sock *smc)
	81	{
	82	if (smc->wait_close_tx_prepared)
	83	/* wake up socket closing */
	84	smc->sk.sk_state_change(&smc->sk);
	85	}
	86
	87	static int smc_close_wr(struct smc_connection *conn)
	88	{
	89	conn->local_tx_ctrl.conn_state_flags.peer_done_writing = 1;
	90
	91	return smc_cdc_get_slot_and_msg_send(conn);
	92	}
	93
	94	static int smc_close_final(struct smc_connection *conn)
	95	{
	96	if (atomic_read(&conn->bytes_to_rcv))
	97	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	98	else
	99	conn->local_tx_ctrl.conn_state_flags.peer_conn_closed = 1;
b2900980 UB	100	if (conn->killed)
b2900980 UB	101	return -EPIPE;
b38d7324 UB	102
	103	return smc_cdc_get_slot_and_msg_send(conn);
	104	}
	105
83179760	106	int smc_close_abort(struct smc_connection *conn)
b38d7324 UB	107	{
	108	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	109
	110	return smc_cdc_get_slot_and_msg_send(conn);
	111	}
	112
	113	/* terminate smc socket abnormally - active abort
732720fa	114	* link group is terminated, i.e. RDMA communication no longer possible
b38d7324	115	*/
81cf4f47	116	void smc_close_active_abort(struct smc_sock *smc)
b38d7324	117	{
3163c507	118	struct sock *sk = &smc->sk;
81cf4f47	119	bool release_clcsock = false;
3163c507	120
dd65d87a UB	121	if (sk->sk_state != SMC_INIT && smc->clcsock && smc->clcsock->sk) {
dd65d87a UB	122	sk->sk_err = ECONNABORTED;
83179760 UB	123	if (smc->clcsock && smc->clcsock->sk)
83179760 UB	124	tcp_abort(smc->clcsock->sk, ECONNABORTED);
b38d7324	125	}
3163c507	126	switch (sk->sk_state) {
46c28dbd	127	case SMC_ACTIVE:
3163c507	128	sk->sk_state = SMC_PEERABORTWAIT;
611b63a1 UB	129	release_sock(sk);
	130	cancel_delayed_work_sync(&smc->conn.tx_work);
	131	lock_sock(sk);
d18963cf	132	sk->sk_state = SMC_CLOSED;
51f1de79	133	sock_put(sk); /* passive closing */
b38d7324 UB	134	break;
	135	case SMC_APPCLOSEWAIT1:
	136	case SMC_APPCLOSEWAIT2:
611b63a1 UB	137	release_sock(sk);
	138	cancel_delayed_work_sync(&smc->conn.tx_work);
	139	lock_sock(sk);
d18963cf	140	sk->sk_state = SMC_CLOSED;
81cf4f47	141	sock_put(sk); /* postponed passive closing */
b38d7324 UB	142	break;
	143	case SMC_PEERCLOSEWAIT1:
	144	case SMC_PEERCLOSEWAIT2:
d18963cf UB	145	case SMC_PEERFINCLOSEWAIT:
d18963cf UB	146	sk->sk_state = SMC_CLOSED;
81cf4f47 UB	147	smc_conn_free(&smc->conn);
81cf4f47 UB	148	release_clcsock = true;
51f1de79	149	sock_put(sk); /* passive closing */
b38d7324 UB	150	break;
	151	case SMC_PROCESSABORT:
	152	case SMC_APPFINCLOSEWAIT:
3163c507	153	sk->sk_state = SMC_CLOSED;
b38d7324	154	break;
dd65d87a	155	case SMC_INIT:
b38d7324 UB	156	case SMC_PEERABORTWAIT:
	157	case SMC_CLOSED:
	158	break;
	159	}
	160
3163c507 UB	161	sock_set_flag(sk, SOCK_DEAD);
3163c507 UB	162	sk->sk_state_change(sk);
81cf4f47 UB	163
	164	if (release_clcsock) {
	165	release_sock(sk);
	166	smc_clcsock_release(smc);
	167	lock_sock(sk);
	168	}
b38d7324 UB	169	}
b38d7324 UB	170
a98bf8c0 UB	171	static inline bool smc_close_sent_any_close(struct smc_connection *conn)
	172	{
	173	return conn->local_tx_ctrl.conn_state_flags.peer_conn_abort \|\|
	174	conn->local_tx_ctrl.conn_state_flags.peer_conn_closed;
	175	}
	176
b38d7324 UB	177	int smc_close_active(struct smc_sock *smc)
	178	{
	179	struct smc_cdc_conn_state_flags *txflags =
	180	&smc->conn.local_tx_ctrl.conn_state_flags;
b38d7324 UB	181	struct smc_connection *conn = &smc->conn;
	182	struct sock *sk = &smc->sk;
	183	int old_state;
8c96feee	184	long timeout;
b38d7324 UB	185	int rc = 0;
b38d7324 UB	186
8c96feee UB	187	timeout = current->flags & PF_EXITING ?
	188	0 : sock_flag(sk, SOCK_LINGER) ?
	189	sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;
b38d7324	190
b38d7324	191	old_state = sk->sk_state;
bbb96bf2 UB	192	again:
bbb96bf2 UB	193	switch (sk->sk_state) {
b38d7324 UB	194	case SMC_INIT:
b38d7324 UB	195	sk->sk_state = SMC_CLOSED;
b38d7324 UB	196	break;
	197	case SMC_LISTEN:
	198	sk->sk_state = SMC_CLOSED;
	199	sk->sk_state_change(sk); /* wake up accept */
	200	if (smc->clcsock && smc->clcsock->sk) {
	201	rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
	202	/* wake up kernel_accept of smc_tcp_listen_worker */
	203	smc->clcsock->sk->sk_data_ready(smc->clcsock->sk);
	204	}
b38d7324	205	smc_close_cleanup_listen(sk);
3d502067 UB	206	release_sock(sk);
	207	flush_work(&smc->tcp_listen_work);
	208	lock_sock(sk);
b38d7324 UB	209	break;
	210	case SMC_ACTIVE:
	211	smc_close_stream_wait(smc, timeout);
	212	release_sock(sk);
18e537cd	213	cancel_delayed_work_sync(&conn->tx_work);
b38d7324 UB	214	lock_sock(sk);
	215	if (sk->sk_state == SMC_ACTIVE) {
	216	/* send close request */
	217	rc = smc_close_final(conn);
	218	sk->sk_state = SMC_PEERCLOSEWAIT1;
	219	} else {
	220	/* peer event has changed the state */
	221	goto again;
	222	}
	223	break;
	224	case SMC_APPFINCLOSEWAIT:
	225	/* socket already shutdown wr or both (active close) */
	226	if (txflags->peer_done_writing &&
a98bf8c0	227	!smc_close_sent_any_close(conn)) {
b38d7324 UB	228	/* just shutdown wr done, send close request */
	229	rc = smc_close_final(conn);
	230	}
	231	sk->sk_state = SMC_CLOSED;
b38d7324 UB	232	break;
	233	case SMC_APPCLOSEWAIT1:
	234	case SMC_APPCLOSEWAIT2:
	235	if (!smc_cdc_rxed_any_close(conn))
	236	smc_close_stream_wait(smc, timeout);
	237	release_sock(sk);
18e537cd	238	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	239	lock_sock(sk);
bbb96bf2 UB	240	if (sk->sk_state != SMC_APPCLOSEWAIT1 &&
	241	sk->sk_state != SMC_APPCLOSEWAIT2)
	242	goto again;
	243	/* confirm close from peer */
	244	rc = smc_close_final(conn);
51f1de79	245	if (smc_cdc_rxed_any_close(conn)) {
b38d7324 UB	246	/* peer has closed the socket already */
b38d7324 UB	247	sk->sk_state = SMC_CLOSED;
51f1de79 UB	248	sock_put(sk); /* postponed passive closing */
51f1de79 UB	249	} else {
b38d7324 UB	250	/* peer has just issued a shutdown write */
b38d7324 UB	251	sk->sk_state = SMC_PEERFINCLOSEWAIT;
51f1de79	252	}
b38d7324 UB	253	break;
	254	case SMC_PEERCLOSEWAIT1:
	255	case SMC_PEERCLOSEWAIT2:
a98bf8c0 UB	256	if (txflags->peer_done_writing &&
	257	!smc_close_sent_any_close(conn)) {
	258	/* just shutdown wr done, send close request */
	259	rc = smc_close_final(conn);
	260	}
	261	/* peer sending PeerConnectionClosed will cause transition */
	262	break;
b38d7324 UB	263	case SMC_PEERFINCLOSEWAIT:
	264	/* peer sending PeerConnectionClosed will cause transition */
	265	break;
	266	case SMC_PROCESSABORT:
d18963cf	267	rc = smc_close_abort(conn);
b38d7324	268	sk->sk_state = SMC_CLOSED;
b38d7324 UB	269	break;
b38d7324 UB	270	case SMC_PEERABORTWAIT:
d18963cf UB	271	sk->sk_state = SMC_CLOSED;
d18963cf UB	272	break;
b38d7324 UB	273	case SMC_CLOSED:
	274	/* nothing to do, add tracing in future patch */
	275	break;
	276	}
	277
	278	if (old_state != sk->sk_state)
3163c507	279	sk->sk_state_change(sk);
b38d7324 UB	280	return rc;
	281	}
	282
	283	static void smc_close_passive_abort_received(struct smc_sock *smc)
	284	{
	285	struct smc_cdc_conn_state_flags *txflags =
	286	&smc->conn.local_tx_ctrl.conn_state_flags;
	287	struct sock *sk = &smc->sk;
	288
	289	switch (sk->sk_state) {
51f1de79	290	case SMC_INIT:
b38d7324	291	case SMC_ACTIVE:
b38d7324	292	case SMC_APPCLOSEWAIT1:
51f1de79 UB	293	sk->sk_state = SMC_PROCESSABORT;
	294	sock_put(sk); /* passive closing */
	295	break;
	296	case SMC_APPFINCLOSEWAIT:
b38d7324 UB	297	sk->sk_state = SMC_PROCESSABORT;
	298	break;
	299	case SMC_PEERCLOSEWAIT1:
	300	case SMC_PEERCLOSEWAIT2:
	301	if (txflags->peer_done_writing &&
51f1de79	302	!smc_close_sent_any_close(&smc->conn))
b38d7324	303	/* just shutdown, but not yet closed locally */
b38d7324	304	sk->sk_state = SMC_PROCESSABORT;
51f1de79	305	else
b38d7324	306	sk->sk_state = SMC_CLOSED;
51f1de79	307	sock_put(sk); /* passive closing */
b38d7324	308	break;
51f1de79	309	case SMC_APPCLOSEWAIT2:
b38d7324	310	case SMC_PEERFINCLOSEWAIT:
51f1de79 UB	311	sk->sk_state = SMC_CLOSED;
	312	sock_put(sk); /* passive closing */
	313	break;
b38d7324 UB	314	case SMC_PEERABORTWAIT:
	315	sk->sk_state = SMC_CLOSED;
	316	break;
b38d7324 UB	317	case SMC_PROCESSABORT:
	318	/* nothing to do, add tracing in future patch */
	319	break;
	320	}
	321	}
	322
732720fa UB	323	/* Either some kind of closing has been received: peer_conn_closed,
	324	* peer_conn_abort, or peer_done_writing
	325	* or the link group of the connection terminates abnormally.
b38d7324	326	*/
46c28dbd	327	static void smc_close_passive_work(struct work_struct *work)
b38d7324	328	{
46c28dbd UB	329	struct smc_connection *conn = container_of(work,
	330	struct smc_connection,
	331	close_work);
	332	struct smc_sock *smc = container_of(conn, struct smc_sock, conn);
	333	struct smc_cdc_conn_state_flags *rxflags;
fd57770d	334	bool release_clcsock = false;
b38d7324 UB	335	struct sock *sk = &smc->sk;
	336	int old_state;
	337
3163c507	338	lock_sock(sk);
b38d7324 UB	339	old_state = sk->sk_state;
b38d7324 UB	340
3163c507	341	rxflags = &conn->local_rx_ctrl.conn_state_flags;
b38d7324	342	if (rxflags->peer_conn_abort) {
732720fa	343	/* peer has not received all data */
b38d7324	344	smc_close_passive_abort_received(smc);
611b63a1 UB	345	release_sock(&smc->sk);
	346	cancel_delayed_work_sync(&conn->tx_work);
	347	lock_sock(&smc->sk);
b38d7324 UB	348	goto wakeup;
	349	}
	350
	351	switch (sk->sk_state) {
	352	case SMC_INIT:
84b799a2	353	sk->sk_state = SMC_APPCLOSEWAIT1;
b38d7324 UB	354	break;
	355	case SMC_ACTIVE:
	356	sk->sk_state = SMC_APPCLOSEWAIT1;
51f1de79 UB	357	/* postpone sock_put() for passive closing to cover
	358	* received SEND_SHUTDOWN as well
	359	*/
b38d7324 UB	360	break;
	361	case SMC_PEERCLOSEWAIT1:
	362	if (rxflags->peer_done_writing)
	363	sk->sk_state = SMC_PEERCLOSEWAIT2;
7f6b437e GS	364	/* fall through */
7f6b437e GS	365	/* to check for closing */
b38d7324	366	case SMC_PEERCLOSEWAIT2:
3163c507	367	if (!smc_cdc_rxed_any_close(conn))
b38d7324 UB	368	break;
b38d7324 UB	369	if (sock_flag(sk, SOCK_DEAD) &&
a98bf8c0	370	smc_close_sent_any_close(conn)) {
b38d7324 UB	371	/* smc_release has already been called locally */
	372	sk->sk_state = SMC_CLOSED;
	373	} else {
	374	/* just shutdown, but not yet closed locally */
	375	sk->sk_state = SMC_APPFINCLOSEWAIT;
	376	}
51f1de79	377	sock_put(sk); /* passive closing */
b38d7324	378	break;
5ac92a00	379	case SMC_PEERFINCLOSEWAIT:
51f1de79	380	if (smc_cdc_rxed_any_close(conn)) {
5ac92a00	381	sk->sk_state = SMC_CLOSED;
51f1de79 UB	382	sock_put(sk); /* passive closing */
51f1de79 UB	383	}
5ac92a00	384	break;
b38d7324 UB	385	case SMC_APPCLOSEWAIT1:
b38d7324 UB	386	case SMC_APPCLOSEWAIT2:
51f1de79 UB	387	/* postpone sock_put() for passive closing to cover
	388	* received SEND_SHUTDOWN as well
	389	*/
	390	break;
b38d7324 UB	391	case SMC_APPFINCLOSEWAIT:
	392	case SMC_PEERABORTWAIT:
	393	case SMC_PROCESSABORT:
	394	case SMC_CLOSED:
	395	/* nothing to do, add tracing in future patch */
	396	break;
	397	}
	398
	399	wakeup:
b38d7324 UB	400	sk->sk_data_ready(sk); /* wakeup blocked rcvbuf consumers */
	401	sk->sk_write_space(sk); /* wakeup blocked sndbuf producers */
	402
a98bf8c0 UB	403	if (old_state != sk->sk_state) {
	404	sk->sk_state_change(sk);
	405	if ((sk->sk_state == SMC_CLOSED) &&
b03faa1f	406	(sock_flag(sk, SOCK_DEAD) \|\| !sk->sk_socket)) {
3163c507	407	smc_conn_free(conn);
fd57770d KG	408	if (smc->clcsock)
fd57770d KG	409	release_clcsock = true;
b03faa1f	410	}
b38d7324	411	}
3163c507	412	release_sock(sk);
fd57770d KG	413	if (release_clcsock)
fd57770d KG	414	smc_clcsock_release(smc);
51f1de79	415	sock_put(sk); /* sock_hold done by schedulers of close_work */
b38d7324 UB	416	}
	417
	418	int smc_close_shutdown_write(struct smc_sock *smc)
	419	{
	420	struct smc_connection *conn = &smc->conn;
b38d7324 UB	421	struct sock *sk = &smc->sk;
b38d7324 UB	422	int old_state;
8c96feee	423	long timeout;
b38d7324 UB	424	int rc = 0;
b38d7324 UB	425
8c96feee UB	426	timeout = current->flags & PF_EXITING ?
	427	0 : sock_flag(sk, SOCK_LINGER) ?
	428	sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;
b38d7324	429
b38d7324	430	old_state = sk->sk_state;
bbb96bf2 UB	431	again:
bbb96bf2 UB	432	switch (sk->sk_state) {
b38d7324 UB	433	case SMC_ACTIVE:
	434	smc_close_stream_wait(smc, timeout);
	435	release_sock(sk);
18e537cd	436	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	437	lock_sock(sk);
bbb96bf2 UB	438	if (sk->sk_state != SMC_ACTIVE)
bbb96bf2 UB	439	goto again;
b38d7324 UB	440	/* send close wr request */
b38d7324 UB	441	rc = smc_close_wr(conn);
bbb96bf2	442	sk->sk_state = SMC_PEERCLOSEWAIT1;
b38d7324 UB	443	break;
	444	case SMC_APPCLOSEWAIT1:
	445	/* passive close */
	446	if (!smc_cdc_rxed_any_close(conn))
	447	smc_close_stream_wait(smc, timeout);
	448	release_sock(sk);
18e537cd	449	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	450	lock_sock(sk);
bbb96bf2 UB	451	if (sk->sk_state != SMC_APPCLOSEWAIT1)
bbb96bf2 UB	452	goto again;
b38d7324 UB	453	/* confirm close from peer */
	454	rc = smc_close_wr(conn);
	455	sk->sk_state = SMC_APPCLOSEWAIT2;
	456	break;
	457	case SMC_APPCLOSEWAIT2:
	458	case SMC_PEERFINCLOSEWAIT:
	459	case SMC_PEERCLOSEWAIT1:
	460	case SMC_PEERCLOSEWAIT2:
	461	case SMC_APPFINCLOSEWAIT:
	462	case SMC_PROCESSABORT:
	463	case SMC_PEERABORTWAIT:
	464	/* nothing to do, add tracing in future patch */
	465	break;
	466	}
	467
	468	if (old_state != sk->sk_state)
3163c507	469	sk->sk_state_change(sk);
b38d7324 UB	470	return rc;
b38d7324 UB	471	}
46c28dbd UB	472
	473	/* Initialize close properties on connection establishment. */
	474	void smc_close_init(struct smc_sock *smc)
	475	{
	476	INIT_WORK(&smc->conn.close_work, smc_close_passive_work);
	477	}