[people/ms/linux.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;

	/* Send out corked data remaining in sndbuf */
	smc_tx_pending(&smc->conn);

	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;
	int rc1 = 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) {
			write_lock_bh(&smc->clcsock->sk->sk_callback_lock);
			smc_clcsock_restore_cb(&smc->clcsock->sk->sk_data_ready,
					       &smc->clcsk_data_ready);
			smc->clcsock->sk->sk_user_data = NULL;
			write_unlock_bh(&smc->clcsock->sk->sk_callback_lock);
			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) {
				rc1 = kernel_sock_shutdown(smc->clcsock,
							   SHUT_RDWR);
				rc = rc ? rc : rc1;
			}
		} 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(sk);
		cancel_delayed_work_sync(&conn->tx_work);
		lock_sock(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
906b3d64 WG	60	/* Send out corked data remaining in sndbuf */
	61	smc_tx_pending(&smc->conn);
	62
b38d7324 UB	63	smc->wait_close_tx_prepared = 1;
	64	add_wait_queue(sk_sleep(sk), &wait);
	65	while (!signal_pending(current) && timeout) {
	66	int rc;
	67
	68	rc = sk_wait_event(sk, &timeout,
	69	!smc_tx_prepared_sends(&smc->conn) \|\|
d18963cf	70	sk->sk_err == ECONNABORTED \|\|
b2900980 UB	71	sk->sk_err == ECONNRESET \|\|
b2900980 UB	72	smc->conn.killed,
b38d7324 UB	73	&wait);
	74	if (rc)
	75	break;
	76	}
	77	remove_wait_queue(sk_sleep(sk), &wait);
	78	smc->wait_close_tx_prepared = 0;
	79	}
	80
	81	void smc_close_wake_tx_prepared(struct smc_sock *smc)
	82	{
	83	if (smc->wait_close_tx_prepared)
	84	/* wake up socket closing */
	85	smc->sk.sk_state_change(&smc->sk);
	86	}
	87
	88	static int smc_close_wr(struct smc_connection *conn)
	89	{
	90	conn->local_tx_ctrl.conn_state_flags.peer_done_writing = 1;
	91
	92	return smc_cdc_get_slot_and_msg_send(conn);
	93	}
	94
	95	static int smc_close_final(struct smc_connection *conn)
	96	{
	97	if (atomic_read(&conn->bytes_to_rcv))
	98	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	99	else
	100	conn->local_tx_ctrl.conn_state_flags.peer_conn_closed = 1;
b2900980 UB	101	if (conn->killed)
b2900980 UB	102	return -EPIPE;
b38d7324 UB	103
	104	return smc_cdc_get_slot_and_msg_send(conn);
	105	}
	106
83179760	107	int smc_close_abort(struct smc_connection *conn)
b38d7324 UB	108	{
	109	conn->local_tx_ctrl.conn_state_flags.peer_conn_abort = 1;
	110
	111	return smc_cdc_get_slot_and_msg_send(conn);
	112	}
	113
42bfba9e UB	114	static void smc_close_cancel_work(struct smc_sock *smc)
	115	{
	116	struct sock *sk = &smc->sk;
	117
	118	release_sock(sk);
	119	cancel_work_sync(&smc->conn.close_work);
	120	cancel_delayed_work_sync(&smc->conn.tx_work);
	121	lock_sock(sk);
42bfba9e UB	122	}
42bfba9e UB	123
b38d7324	124	/* terminate smc socket abnormally - active abort
732720fa	125	* link group is terminated, i.e. RDMA communication no longer possible
b38d7324	126	*/
81cf4f47	127	void smc_close_active_abort(struct smc_sock *smc)
b38d7324	128	{
3163c507	129	struct sock *sk = &smc->sk;
81cf4f47	130	bool release_clcsock = false;
3163c507	131
dd65d87a UB	132	if (sk->sk_state != SMC_INIT && smc->clcsock && smc->clcsock->sk) {
dd65d87a UB	133	sk->sk_err = ECONNABORTED;
83179760 UB	134	if (smc->clcsock && smc->clcsock->sk)
83179760 UB	135	tcp_abort(smc->clcsock->sk, ECONNABORTED);
b38d7324	136	}
3163c507	137	switch (sk->sk_state) {
46c28dbd	138	case SMC_ACTIVE:
b38d7324 UB	139	case SMC_APPCLOSEWAIT1:
b38d7324 UB	140	case SMC_APPCLOSEWAIT2:
5fb8642a	141	sk->sk_state = SMC_PEERABORTWAIT;
42bfba9e	142	smc_close_cancel_work(smc);
5fb8642a UB	143	if (sk->sk_state != SMC_PEERABORTWAIT)
5fb8642a UB	144	break;
d18963cf	145	sk->sk_state = SMC_CLOSED;
5fb8642a	146	sock_put(sk); /* (postponed) passive closing */
b38d7324 UB	147	break;
	148	case SMC_PEERCLOSEWAIT1:
	149	case SMC_PEERCLOSEWAIT2:
d18963cf	150	case SMC_PEERFINCLOSEWAIT:
42bfba9e UB	151	sk->sk_state = SMC_PEERABORTWAIT;
42bfba9e UB	152	smc_close_cancel_work(smc);
5fb8642a UB	153	if (sk->sk_state != SMC_PEERABORTWAIT)
5fb8642a UB	154	break;
d18963cf	155	sk->sk_state = SMC_CLOSED;
81cf4f47 UB	156	smc_conn_free(&smc->conn);
81cf4f47 UB	157	release_clcsock = true;
51f1de79	158	sock_put(sk); /* passive closing */
b38d7324 UB	159	break;
	160	case SMC_PROCESSABORT:
	161	case SMC_APPFINCLOSEWAIT:
42bfba9e UB	162	sk->sk_state = SMC_PEERABORTWAIT;
42bfba9e UB	163	smc_close_cancel_work(smc);
5fb8642a UB	164	if (sk->sk_state != SMC_PEERABORTWAIT)
5fb8642a UB	165	break;
3163c507	166	sk->sk_state = SMC_CLOSED;
42bfba9e UB	167	smc_conn_free(&smc->conn);
42bfba9e UB	168	release_clcsock = true;
b38d7324	169	break;
dd65d87a	170	case SMC_INIT:
b38d7324 UB	171	case SMC_PEERABORTWAIT:
	172	case SMC_CLOSED:
	173	break;
	174	}
	175
3163c507 UB	176	sock_set_flag(sk, SOCK_DEAD);
3163c507 UB	177	sk->sk_state_change(sk);
81cf4f47 UB	178
	179	if (release_clcsock) {
	180	release_sock(sk);
	181	smc_clcsock_release(smc);
	182	lock_sock(sk);
	183	}
b38d7324 UB	184	}
b38d7324 UB	185
a98bf8c0 UB	186	static inline bool smc_close_sent_any_close(struct smc_connection *conn)
	187	{
	188	return conn->local_tx_ctrl.conn_state_flags.peer_conn_abort \|\|
	189	conn->local_tx_ctrl.conn_state_flags.peer_conn_closed;
	190	}
	191
b38d7324 UB	192	int smc_close_active(struct smc_sock *smc)
	193	{
	194	struct smc_cdc_conn_state_flags *txflags =
	195	&smc->conn.local_tx_ctrl.conn_state_flags;
b38d7324 UB	196	struct smc_connection *conn = &smc->conn;
	197	struct sock *sk = &smc->sk;
	198	int old_state;
8c96feee	199	long timeout;
b38d7324	200	int rc = 0;
00e158fb	201	int rc1 = 0;
b38d7324	202
8c96feee UB	203	timeout = current->flags & PF_EXITING ?
	204	0 : sock_flag(sk, SOCK_LINGER) ?
	205	sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;
b38d7324	206
b38d7324	207	old_state = sk->sk_state;
bbb96bf2 UB	208	again:
bbb96bf2 UB	209	switch (sk->sk_state) {
b38d7324 UB	210	case SMC_INIT:
b38d7324 UB	211	sk->sk_state = SMC_CLOSED;
b38d7324 UB	212	break;
	213	case SMC_LISTEN:
	214	sk->sk_state = SMC_CLOSED;
	215	sk->sk_state_change(sk); /* wake up accept */
ddcc9b7f	216	if (smc->clcsock && smc->clcsock->sk) {
0558226c	217	write_lock_bh(&smc->clcsock->sk->sk_callback_lock);
97b9af7a WG	218	smc_clcsock_restore_cb(&smc->clcsock->sk->sk_data_ready,
97b9af7a WG	219	&smc->clcsk_data_ready);
ddcc9b7f	220	smc->clcsock->sk->sk_user_data = NULL;
0558226c	221	write_unlock_bh(&smc->clcsock->sk->sk_callback_lock);
b38d7324	222	rc = kernel_sock_shutdown(smc->clcsock, SHUT_RDWR);
ddcc9b7f	223	}
b38d7324	224	smc_close_cleanup_listen(sk);
3d502067 UB	225	release_sock(sk);
	226	flush_work(&smc->tcp_listen_work);
	227	lock_sock(sk);
b38d7324 UB	228	break;
	229	case SMC_ACTIVE:
	230	smc_close_stream_wait(smc, timeout);
	231	release_sock(sk);
18e537cd	232	cancel_delayed_work_sync(&conn->tx_work);
b38d7324 UB	233	lock_sock(sk);
	234	if (sk->sk_state == SMC_ACTIVE) {
	235	/* send close request */
	236	rc = smc_close_final(conn);
	237	sk->sk_state = SMC_PEERCLOSEWAIT1;
606a63c9 TL	238
	239	/* actively shutdown clcsock before peer close it,
	240	* prevent peer from entering TIME_WAIT state.
	241	*/
00e158fb TL	242	if (smc->clcsock && smc->clcsock->sk) {
	243	rc1 = kernel_sock_shutdown(smc->clcsock,
	244	SHUT_RDWR);
	245	rc = rc ? rc : rc1;
	246	}
b38d7324 UB	247	} else {
	248	/* peer event has changed the state */
	249	goto again;
	250	}
	251	break;
	252	case SMC_APPFINCLOSEWAIT:
	253	/* socket already shutdown wr or both (active close) */
	254	if (txflags->peer_done_writing &&
a98bf8c0	255	!smc_close_sent_any_close(conn)) {
b38d7324 UB	256	/* just shutdown wr done, send close request */
	257	rc = smc_close_final(conn);
	258	}
	259	sk->sk_state = SMC_CLOSED;
b38d7324 UB	260	break;
	261	case SMC_APPCLOSEWAIT1:
	262	case SMC_APPCLOSEWAIT2:
	263	if (!smc_cdc_rxed_any_close(conn))
	264	smc_close_stream_wait(smc, timeout);
	265	release_sock(sk);
18e537cd	266	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	267	lock_sock(sk);
bbb96bf2 UB	268	if (sk->sk_state != SMC_APPCLOSEWAIT1 &&
	269	sk->sk_state != SMC_APPCLOSEWAIT2)
	270	goto again;
	271	/* confirm close from peer */
	272	rc = smc_close_final(conn);
51f1de79	273	if (smc_cdc_rxed_any_close(conn)) {
b38d7324 UB	274	/* peer has closed the socket already */
b38d7324 UB	275	sk->sk_state = SMC_CLOSED;
51f1de79 UB	276	sock_put(sk); /* postponed passive closing */
51f1de79 UB	277	} else {
b38d7324 UB	278	/* peer has just issued a shutdown write */
b38d7324 UB	279	sk->sk_state = SMC_PEERFINCLOSEWAIT;
51f1de79	280	}
b38d7324 UB	281	break;
	282	case SMC_PEERCLOSEWAIT1:
	283	case SMC_PEERCLOSEWAIT2:
a98bf8c0 UB	284	if (txflags->peer_done_writing &&
	285	!smc_close_sent_any_close(conn)) {
	286	/* just shutdown wr done, send close request */
	287	rc = smc_close_final(conn);
	288	}
	289	/* peer sending PeerConnectionClosed will cause transition */
	290	break;
b38d7324 UB	291	case SMC_PEERFINCLOSEWAIT:
	292	/* peer sending PeerConnectionClosed will cause transition */
	293	break;
	294	case SMC_PROCESSABORT:
d18963cf	295	rc = smc_close_abort(conn);
b38d7324	296	sk->sk_state = SMC_CLOSED;
b38d7324 UB	297	break;
b38d7324 UB	298	case SMC_PEERABORTWAIT:
d18963cf UB	299	sk->sk_state = SMC_CLOSED;
d18963cf UB	300	break;
b38d7324 UB	301	case SMC_CLOSED:
	302	/* nothing to do, add tracing in future patch */
	303	break;
	304	}
	305
	306	if (old_state != sk->sk_state)
3163c507	307	sk->sk_state_change(sk);
b38d7324 UB	308	return rc;
	309	}
	310
	311	static void smc_close_passive_abort_received(struct smc_sock *smc)
	312	{
	313	struct smc_cdc_conn_state_flags *txflags =
	314	&smc->conn.local_tx_ctrl.conn_state_flags;
	315	struct sock *sk = &smc->sk;
	316
	317	switch (sk->sk_state) {
51f1de79	318	case SMC_INIT:
b38d7324	319	case SMC_ACTIVE:
b38d7324	320	case SMC_APPCLOSEWAIT1:
51f1de79 UB	321	sk->sk_state = SMC_PROCESSABORT;
	322	sock_put(sk); /* passive closing */
	323	break;
	324	case SMC_APPFINCLOSEWAIT:
b38d7324 UB	325	sk->sk_state = SMC_PROCESSABORT;
	326	break;
	327	case SMC_PEERCLOSEWAIT1:
	328	case SMC_PEERCLOSEWAIT2:
	329	if (txflags->peer_done_writing &&
51f1de79	330	!smc_close_sent_any_close(&smc->conn))
b38d7324	331	/* just shutdown, but not yet closed locally */
b38d7324	332	sk->sk_state = SMC_PROCESSABORT;
51f1de79	333	else
b38d7324	334	sk->sk_state = SMC_CLOSED;
51f1de79	335	sock_put(sk); /* passive closing */
b38d7324	336	break;
51f1de79	337	case SMC_APPCLOSEWAIT2:
b38d7324	338	case SMC_PEERFINCLOSEWAIT:
51f1de79 UB	339	sk->sk_state = SMC_CLOSED;
	340	sock_put(sk); /* passive closing */
	341	break;
b38d7324 UB	342	case SMC_PEERABORTWAIT:
	343	sk->sk_state = SMC_CLOSED;
	344	break;
b38d7324 UB	345	case SMC_PROCESSABORT:
	346	/* nothing to do, add tracing in future patch */
	347	break;
	348	}
	349	}
	350
732720fa UB	351	/* Either some kind of closing has been received: peer_conn_closed,
	352	* peer_conn_abort, or peer_done_writing
	353	* or the link group of the connection terminates abnormally.
b38d7324	354	*/
46c28dbd	355	static void smc_close_passive_work(struct work_struct *work)
b38d7324	356	{
46c28dbd UB	357	struct smc_connection *conn = container_of(work,
	358	struct smc_connection,
	359	close_work);
	360	struct smc_sock *smc = container_of(conn, struct smc_sock, conn);
	361	struct smc_cdc_conn_state_flags *rxflags;
fd57770d	362	bool release_clcsock = false;
b38d7324 UB	363	struct sock *sk = &smc->sk;
	364	int old_state;
	365
3163c507	366	lock_sock(sk);
b38d7324 UB	367	old_state = sk->sk_state;
b38d7324 UB	368
3163c507	369	rxflags = &conn->local_rx_ctrl.conn_state_flags;
b38d7324	370	if (rxflags->peer_conn_abort) {
732720fa	371	/* peer has not received all data */
b38d7324	372	smc_close_passive_abort_received(smc);
45c3ff7a	373	release_sock(sk);
611b63a1	374	cancel_delayed_work_sync(&conn->tx_work);
45c3ff7a	375	lock_sock(sk);
b38d7324 UB	376	goto wakeup;
	377	}
	378
	379	switch (sk->sk_state) {
	380	case SMC_INIT:
84b799a2	381	sk->sk_state = SMC_APPCLOSEWAIT1;
b38d7324 UB	382	break;
	383	case SMC_ACTIVE:
	384	sk->sk_state = SMC_APPCLOSEWAIT1;
51f1de79 UB	385	/* postpone sock_put() for passive closing to cover
	386	* received SEND_SHUTDOWN as well
	387	*/
b38d7324 UB	388	break;
	389	case SMC_PEERCLOSEWAIT1:
	390	if (rxflags->peer_done_writing)
	391	sk->sk_state = SMC_PEERCLOSEWAIT2;
df561f66	392	fallthrough;
7f6b437e	393	/* to check for closing */
b38d7324	394	case SMC_PEERCLOSEWAIT2:
3163c507	395	if (!smc_cdc_rxed_any_close(conn))
b38d7324 UB	396	break;
b38d7324 UB	397	if (sock_flag(sk, SOCK_DEAD) &&
a98bf8c0	398	smc_close_sent_any_close(conn)) {
b38d7324 UB	399	/* smc_release has already been called locally */
	400	sk->sk_state = SMC_CLOSED;
	401	} else {
	402	/* just shutdown, but not yet closed locally */
	403	sk->sk_state = SMC_APPFINCLOSEWAIT;
	404	}
51f1de79	405	sock_put(sk); /* passive closing */
b38d7324	406	break;
5ac92a00	407	case SMC_PEERFINCLOSEWAIT:
51f1de79	408	if (smc_cdc_rxed_any_close(conn)) {
5ac92a00	409	sk->sk_state = SMC_CLOSED;
51f1de79 UB	410	sock_put(sk); /* passive closing */
51f1de79 UB	411	}
5ac92a00	412	break;
b38d7324 UB	413	case SMC_APPCLOSEWAIT1:
b38d7324 UB	414	case SMC_APPCLOSEWAIT2:
51f1de79 UB	415	/* postpone sock_put() for passive closing to cover
	416	* received SEND_SHUTDOWN as well
	417	*/
	418	break;
b38d7324 UB	419	case SMC_APPFINCLOSEWAIT:
	420	case SMC_PEERABORTWAIT:
	421	case SMC_PROCESSABORT:
	422	case SMC_CLOSED:
	423	/* nothing to do, add tracing in future patch */
	424	break;
	425	}
	426
	427	wakeup:
b38d7324 UB	428	sk->sk_data_ready(sk); /* wakeup blocked rcvbuf consumers */
	429	sk->sk_write_space(sk); /* wakeup blocked sndbuf producers */
	430
a98bf8c0 UB	431	if (old_state != sk->sk_state) {
	432	sk->sk_state_change(sk);
	433	if ((sk->sk_state == SMC_CLOSED) &&
b03faa1f	434	(sock_flag(sk, SOCK_DEAD) \|\| !sk->sk_socket)) {
3163c507	435	smc_conn_free(conn);
fd57770d KG	436	if (smc->clcsock)
fd57770d KG	437	release_clcsock = true;
b03faa1f	438	}
b38d7324	439	}
3163c507	440	release_sock(sk);
fd57770d KG	441	if (release_clcsock)
fd57770d KG	442	smc_clcsock_release(smc);
51f1de79	443	sock_put(sk); /* sock_hold done by schedulers of close_work */
b38d7324 UB	444	}
	445
	446	int smc_close_shutdown_write(struct smc_sock *smc)
	447	{
	448	struct smc_connection *conn = &smc->conn;
b38d7324 UB	449	struct sock *sk = &smc->sk;
b38d7324 UB	450	int old_state;
8c96feee	451	long timeout;
b38d7324 UB	452	int rc = 0;
b38d7324 UB	453
8c96feee UB	454	timeout = current->flags & PF_EXITING ?
	455	0 : sock_flag(sk, SOCK_LINGER) ?
	456	sk->sk_lingertime : SMC_MAX_STREAM_WAIT_TIMEOUT;
b38d7324	457
b38d7324	458	old_state = sk->sk_state;
bbb96bf2 UB	459	again:
bbb96bf2 UB	460	switch (sk->sk_state) {
b38d7324 UB	461	case SMC_ACTIVE:
	462	smc_close_stream_wait(smc, timeout);
	463	release_sock(sk);
18e537cd	464	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	465	lock_sock(sk);
bbb96bf2 UB	466	if (sk->sk_state != SMC_ACTIVE)
bbb96bf2 UB	467	goto again;
b38d7324 UB	468	/* send close wr request */
b38d7324 UB	469	rc = smc_close_wr(conn);
bbb96bf2	470	sk->sk_state = SMC_PEERCLOSEWAIT1;
b38d7324 UB	471	break;
	472	case SMC_APPCLOSEWAIT1:
	473	/* passive close */
	474	if (!smc_cdc_rxed_any_close(conn))
	475	smc_close_stream_wait(smc, timeout);
	476	release_sock(sk);
18e537cd	477	cancel_delayed_work_sync(&conn->tx_work);
b38d7324	478	lock_sock(sk);
bbb96bf2 UB	479	if (sk->sk_state != SMC_APPCLOSEWAIT1)
bbb96bf2 UB	480	goto again;
b38d7324 UB	481	/* confirm close from peer */
	482	rc = smc_close_wr(conn);
	483	sk->sk_state = SMC_APPCLOSEWAIT2;
	484	break;
	485	case SMC_APPCLOSEWAIT2:
	486	case SMC_PEERFINCLOSEWAIT:
	487	case SMC_PEERCLOSEWAIT1:
	488	case SMC_PEERCLOSEWAIT2:
	489	case SMC_APPFINCLOSEWAIT:
	490	case SMC_PROCESSABORT:
	491	case SMC_PEERABORTWAIT:
	492	/* nothing to do, add tracing in future patch */
	493	break;
	494	}
	495
	496	if (old_state != sk->sk_state)
3163c507	497	sk->sk_state_change(sk);
b38d7324 UB	498	return rc;
b38d7324 UB	499	}
46c28dbd UB	500
	501	/* Initialize close properties on connection establishment. */
	502	void smc_close_init(struct smc_sock *smc)
	503	{
	504	INIT_WORK(&smc->conn.close_work, smc_close_passive_work);
	505	}