xref: /openbmc/linux/net/tipc/socket.c (revision 5a170e9e)
1 /*
2  * net/tipc/socket.c: TIPC socket API
3  *
4  * Copyright (c) 2001-2007, 2012-2017, Ericsson AB
5  * Copyright (c) 2004-2008, 2010-2013, Wind River Systems
6  * All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions are met:
10  *
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 3. Neither the names of the copyright holders nor the names of its
17  *    contributors may be used to endorse or promote products derived from
18  *    this software without specific prior written permission.
19  *
20  * Alternatively, this software may be distributed under the terms of the
21  * GNU General Public License ("GPL") version 2 as published by the Free
22  * Software Foundation.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
25  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
28  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
29  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
30  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
31  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
32  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
33  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
34  * POSSIBILITY OF SUCH DAMAGE.
35  */
36 
37 #include <linux/rhashtable.h>
38 #include <linux/sched/signal.h>
39 
40 #include "core.h"
41 #include "name_table.h"
42 #include "node.h"
43 #include "link.h"
44 #include "name_distr.h"
45 #include "socket.h"
46 #include "bcast.h"
47 #include "netlink.h"
48 #include "group.h"
49 #include "trace.h"
50 
51 #define CONN_TIMEOUT_DEFAULT    8000    /* default connect timeout = 8s */
52 #define CONN_PROBING_INTV	msecs_to_jiffies(3600000)  /* [ms] => 1 h */
53 #define TIPC_FWD_MSG		1
54 #define TIPC_MAX_PORT		0xffffffff
55 #define TIPC_MIN_PORT		1
56 #define TIPC_ACK_RATE		4       /* ACK at 1/4 of of rcv window size */
57 
58 enum {
59 	TIPC_LISTEN = TCP_LISTEN,
60 	TIPC_ESTABLISHED = TCP_ESTABLISHED,
61 	TIPC_OPEN = TCP_CLOSE,
62 	TIPC_DISCONNECTING = TCP_CLOSE_WAIT,
63 	TIPC_CONNECTING = TCP_SYN_SENT,
64 };
65 
66 struct sockaddr_pair {
67 	struct sockaddr_tipc sock;
68 	struct sockaddr_tipc member;
69 };
70 
71 /**
72  * struct tipc_sock - TIPC socket structure
73  * @sk: socket - interacts with 'port' and with user via the socket API
74  * @conn_type: TIPC type used when connection was established
75  * @conn_instance: TIPC instance used when connection was established
76  * @published: non-zero if port has one or more associated names
77  * @max_pkt: maximum packet size "hint" used when building messages sent by port
78  * @portid: unique port identity in TIPC socket hash table
79  * @phdr: preformatted message header used when sending messages
80  * #cong_links: list of congested links
81  * @publications: list of publications for port
82  * @blocking_link: address of the congested link we are currently sleeping on
83  * @pub_count: total # of publications port has made during its lifetime
84  * @conn_timeout: the time we can wait for an unresponded setup request
85  * @dupl_rcvcnt: number of bytes counted twice, in both backlog and rcv queue
86  * @cong_link_cnt: number of congested links
87  * @snt_unacked: # messages sent by socket, and not yet acked by peer
88  * @rcv_unacked: # messages read by user, but not yet acked back to peer
89  * @peer: 'connected' peer for dgram/rdm
90  * @node: hash table node
91  * @mc_method: cookie for use between socket and broadcast layer
92  * @rcu: rcu struct for tipc_sock
93  */
94 struct tipc_sock {
95 	struct sock sk;
96 	u32 conn_type;
97 	u32 conn_instance;
98 	int published;
99 	u32 max_pkt;
100 	u32 portid;
101 	struct tipc_msg phdr;
102 	struct list_head cong_links;
103 	struct list_head publications;
104 	u32 pub_count;
105 	atomic_t dupl_rcvcnt;
106 	u16 conn_timeout;
107 	bool probe_unacked;
108 	u16 cong_link_cnt;
109 	u16 snt_unacked;
110 	u16 snd_win;
111 	u16 peer_caps;
112 	u16 rcv_unacked;
113 	u16 rcv_win;
114 	struct sockaddr_tipc peer;
115 	struct rhash_head node;
116 	struct tipc_mc_method mc_method;
117 	struct rcu_head rcu;
118 	struct tipc_group *group;
119 	bool group_is_open;
120 };
121 
122 static int tipc_sk_backlog_rcv(struct sock *sk, struct sk_buff *skb);
123 static void tipc_data_ready(struct sock *sk);
124 static void tipc_write_space(struct sock *sk);
125 static void tipc_sock_destruct(struct sock *sk);
126 static int tipc_release(struct socket *sock);
127 static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags,
128 		       bool kern);
129 static void tipc_sk_timeout(struct timer_list *t);
130 static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
131 			   struct tipc_name_seq const *seq);
132 static int tipc_sk_withdraw(struct tipc_sock *tsk, uint scope,
133 			    struct tipc_name_seq const *seq);
134 static int tipc_sk_leave(struct tipc_sock *tsk);
135 static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid);
136 static int tipc_sk_insert(struct tipc_sock *tsk);
137 static void tipc_sk_remove(struct tipc_sock *tsk);
138 static int __tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dsz);
139 static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dsz);
140 
141 static const struct proto_ops packet_ops;
142 static const struct proto_ops stream_ops;
143 static const struct proto_ops msg_ops;
144 static struct proto tipc_proto;
145 static const struct rhashtable_params tsk_rht_params;
146 
147 static u32 tsk_own_node(struct tipc_sock *tsk)
148 {
149 	return msg_prevnode(&tsk->phdr);
150 }
151 
152 static u32 tsk_peer_node(struct tipc_sock *tsk)
153 {
154 	return msg_destnode(&tsk->phdr);
155 }
156 
157 static u32 tsk_peer_port(struct tipc_sock *tsk)
158 {
159 	return msg_destport(&tsk->phdr);
160 }
161 
162 static  bool tsk_unreliable(struct tipc_sock *tsk)
163 {
164 	return msg_src_droppable(&tsk->phdr) != 0;
165 }
166 
167 static void tsk_set_unreliable(struct tipc_sock *tsk, bool unreliable)
168 {
169 	msg_set_src_droppable(&tsk->phdr, unreliable ? 1 : 0);
170 }
171 
172 static bool tsk_unreturnable(struct tipc_sock *tsk)
173 {
174 	return msg_dest_droppable(&tsk->phdr) != 0;
175 }
176 
177 static void tsk_set_unreturnable(struct tipc_sock *tsk, bool unreturnable)
178 {
179 	msg_set_dest_droppable(&tsk->phdr, unreturnable ? 1 : 0);
180 }
181 
182 static int tsk_importance(struct tipc_sock *tsk)
183 {
184 	return msg_importance(&tsk->phdr);
185 }
186 
187 static int tsk_set_importance(struct tipc_sock *tsk, int imp)
188 {
189 	if (imp > TIPC_CRITICAL_IMPORTANCE)
190 		return -EINVAL;
191 	msg_set_importance(&tsk->phdr, (u32)imp);
192 	return 0;
193 }
194 
195 static struct tipc_sock *tipc_sk(const struct sock *sk)
196 {
197 	return container_of(sk, struct tipc_sock, sk);
198 }
199 
200 static bool tsk_conn_cong(struct tipc_sock *tsk)
201 {
202 	return tsk->snt_unacked > tsk->snd_win;
203 }
204 
205 static u16 tsk_blocks(int len)
206 {
207 	return ((len / FLOWCTL_BLK_SZ) + 1);
208 }
209 
210 /* tsk_blocks(): translate a buffer size in bytes to number of
211  * advertisable blocks, taking into account the ratio truesize(len)/len
212  * We can trust that this ratio is always < 4 for len >= FLOWCTL_BLK_SZ
213  */
214 static u16 tsk_adv_blocks(int len)
215 {
216 	return len / FLOWCTL_BLK_SZ / 4;
217 }
218 
219 /* tsk_inc(): increment counter for sent or received data
220  * - If block based flow control is not supported by peer we
221  *   fall back to message based ditto, incrementing the counter
222  */
223 static u16 tsk_inc(struct tipc_sock *tsk, int msglen)
224 {
225 	if (likely(tsk->peer_caps & TIPC_BLOCK_FLOWCTL))
226 		return ((msglen / FLOWCTL_BLK_SZ) + 1);
227 	return 1;
228 }
229 
230 /**
231  * tsk_advance_rx_queue - discard first buffer in socket receive queue
232  *
233  * Caller must hold socket lock
234  */
235 static void tsk_advance_rx_queue(struct sock *sk)
236 {
237 	trace_tipc_sk_advance_rx(sk, NULL, TIPC_DUMP_SK_RCVQ, " ");
238 	kfree_skb(__skb_dequeue(&sk->sk_receive_queue));
239 }
240 
241 /* tipc_sk_respond() : send response message back to sender
242  */
243 static void tipc_sk_respond(struct sock *sk, struct sk_buff *skb, int err)
244 {
245 	u32 selector;
246 	u32 dnode;
247 	u32 onode = tipc_own_addr(sock_net(sk));
248 
249 	if (!tipc_msg_reverse(onode, &skb, err))
250 		return;
251 
252 	trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_NONE, "@sk_respond!");
253 	dnode = msg_destnode(buf_msg(skb));
254 	selector = msg_origport(buf_msg(skb));
255 	tipc_node_xmit_skb(sock_net(sk), skb, dnode, selector);
256 }
257 
258 /**
259  * tsk_rej_rx_queue - reject all buffers in socket receive queue
260  *
261  * Caller must hold socket lock
262  */
263 static void tsk_rej_rx_queue(struct sock *sk)
264 {
265 	struct sk_buff *skb;
266 
267 	while ((skb = __skb_dequeue(&sk->sk_receive_queue)))
268 		tipc_sk_respond(sk, skb, TIPC_ERR_NO_PORT);
269 }
270 
271 static bool tipc_sk_connected(struct sock *sk)
272 {
273 	return sk->sk_state == TIPC_ESTABLISHED;
274 }
275 
276 /* tipc_sk_type_connectionless - check if the socket is datagram socket
277  * @sk: socket
278  *
279  * Returns true if connection less, false otherwise
280  */
281 static bool tipc_sk_type_connectionless(struct sock *sk)
282 {
283 	return sk->sk_type == SOCK_RDM || sk->sk_type == SOCK_DGRAM;
284 }
285 
286 /* tsk_peer_msg - verify if message was sent by connected port's peer
287  *
288  * Handles cases where the node's network address has changed from
289  * the default of <0.0.0> to its configured setting.
290  */
291 static bool tsk_peer_msg(struct tipc_sock *tsk, struct tipc_msg *msg)
292 {
293 	struct sock *sk = &tsk->sk;
294 	u32 self = tipc_own_addr(sock_net(sk));
295 	u32 peer_port = tsk_peer_port(tsk);
296 	u32 orig_node, peer_node;
297 
298 	if (unlikely(!tipc_sk_connected(sk)))
299 		return false;
300 
301 	if (unlikely(msg_origport(msg) != peer_port))
302 		return false;
303 
304 	orig_node = msg_orignode(msg);
305 	peer_node = tsk_peer_node(tsk);
306 
307 	if (likely(orig_node == peer_node))
308 		return true;
309 
310 	if (!orig_node && peer_node == self)
311 		return true;
312 
313 	if (!peer_node && orig_node == self)
314 		return true;
315 
316 	return false;
317 }
318 
319 /* tipc_set_sk_state - set the sk_state of the socket
320  * @sk: socket
321  *
322  * Caller must hold socket lock
323  *
324  * Returns 0 on success, errno otherwise
325  */
326 static int tipc_set_sk_state(struct sock *sk, int state)
327 {
328 	int oldsk_state = sk->sk_state;
329 	int res = -EINVAL;
330 
331 	switch (state) {
332 	case TIPC_OPEN:
333 		res = 0;
334 		break;
335 	case TIPC_LISTEN:
336 	case TIPC_CONNECTING:
337 		if (oldsk_state == TIPC_OPEN)
338 			res = 0;
339 		break;
340 	case TIPC_ESTABLISHED:
341 		if (oldsk_state == TIPC_CONNECTING ||
342 		    oldsk_state == TIPC_OPEN)
343 			res = 0;
344 		break;
345 	case TIPC_DISCONNECTING:
346 		if (oldsk_state == TIPC_CONNECTING ||
347 		    oldsk_state == TIPC_ESTABLISHED)
348 			res = 0;
349 		break;
350 	}
351 
352 	if (!res)
353 		sk->sk_state = state;
354 
355 	return res;
356 }
357 
358 static int tipc_sk_sock_err(struct socket *sock, long *timeout)
359 {
360 	struct sock *sk = sock->sk;
361 	int err = sock_error(sk);
362 	int typ = sock->type;
363 
364 	if (err)
365 		return err;
366 	if (typ == SOCK_STREAM || typ == SOCK_SEQPACKET) {
367 		if (sk->sk_state == TIPC_DISCONNECTING)
368 			return -EPIPE;
369 		else if (!tipc_sk_connected(sk))
370 			return -ENOTCONN;
371 	}
372 	if (!*timeout)
373 		return -EAGAIN;
374 	if (signal_pending(current))
375 		return sock_intr_errno(*timeout);
376 
377 	return 0;
378 }
379 
380 #define tipc_wait_for_cond(sock_, timeo_, condition_)			       \
381 ({                                                                             \
382 	struct sock *sk_;						       \
383 	int rc_;							       \
384 									       \
385 	while ((rc_ = !(condition_))) {					       \
386 		DEFINE_WAIT_FUNC(wait_, woken_wake_function);	               \
387 		sk_ = (sock_)->sk;					       \
388 		rc_ = tipc_sk_sock_err((sock_), timeo_);		       \
389 		if (rc_)						       \
390 			break;						       \
391 		prepare_to_wait(sk_sleep(sk_), &wait_, TASK_INTERRUPTIBLE);    \
392 		release_sock(sk_);					       \
393 		*(timeo_) = wait_woken(&wait_, TASK_INTERRUPTIBLE, *(timeo_)); \
394 		sched_annotate_sleep();				               \
395 		lock_sock(sk_);						       \
396 		remove_wait_queue(sk_sleep(sk_), &wait_);		       \
397 	}								       \
398 	rc_;								       \
399 })
400 
401 /**
402  * tipc_sk_create - create a TIPC socket
403  * @net: network namespace (must be default network)
404  * @sock: pre-allocated socket structure
405  * @protocol: protocol indicator (must be 0)
406  * @kern: caused by kernel or by userspace?
407  *
408  * This routine creates additional data structures used by the TIPC socket,
409  * initializes them, and links them together.
410  *
411  * Returns 0 on success, errno otherwise
412  */
413 static int tipc_sk_create(struct net *net, struct socket *sock,
414 			  int protocol, int kern)
415 {
416 	const struct proto_ops *ops;
417 	struct sock *sk;
418 	struct tipc_sock *tsk;
419 	struct tipc_msg *msg;
420 
421 	/* Validate arguments */
422 	if (unlikely(protocol != 0))
423 		return -EPROTONOSUPPORT;
424 
425 	switch (sock->type) {
426 	case SOCK_STREAM:
427 		ops = &stream_ops;
428 		break;
429 	case SOCK_SEQPACKET:
430 		ops = &packet_ops;
431 		break;
432 	case SOCK_DGRAM:
433 	case SOCK_RDM:
434 		ops = &msg_ops;
435 		break;
436 	default:
437 		return -EPROTOTYPE;
438 	}
439 
440 	/* Allocate socket's protocol area */
441 	sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto, kern);
442 	if (sk == NULL)
443 		return -ENOMEM;
444 
445 	tsk = tipc_sk(sk);
446 	tsk->max_pkt = MAX_PKT_DEFAULT;
447 	INIT_LIST_HEAD(&tsk->publications);
448 	INIT_LIST_HEAD(&tsk->cong_links);
449 	msg = &tsk->phdr;
450 
451 	/* Finish initializing socket data structures */
452 	sock->ops = ops;
453 	sock_init_data(sock, sk);
454 	tipc_set_sk_state(sk, TIPC_OPEN);
455 	if (tipc_sk_insert(tsk)) {
456 		pr_warn("Socket create failed; port number exhausted\n");
457 		return -EINVAL;
458 	}
459 
460 	/* Ensure tsk is visible before we read own_addr. */
461 	smp_mb();
462 
463 	tipc_msg_init(tipc_own_addr(net), msg, TIPC_LOW_IMPORTANCE,
464 		      TIPC_NAMED_MSG, NAMED_H_SIZE, 0);
465 
466 	msg_set_origport(msg, tsk->portid);
467 	timer_setup(&sk->sk_timer, tipc_sk_timeout, 0);
468 	sk->sk_shutdown = 0;
469 	sk->sk_backlog_rcv = tipc_sk_backlog_rcv;
470 	sk->sk_rcvbuf = sysctl_tipc_rmem[1];
471 	sk->sk_data_ready = tipc_data_ready;
472 	sk->sk_write_space = tipc_write_space;
473 	sk->sk_destruct = tipc_sock_destruct;
474 	tsk->conn_timeout = CONN_TIMEOUT_DEFAULT;
475 	tsk->group_is_open = true;
476 	atomic_set(&tsk->dupl_rcvcnt, 0);
477 
478 	/* Start out with safe limits until we receive an advertised window */
479 	tsk->snd_win = tsk_adv_blocks(RCVBUF_MIN);
480 	tsk->rcv_win = tsk->snd_win;
481 
482 	if (tipc_sk_type_connectionless(sk)) {
483 		tsk_set_unreturnable(tsk, true);
484 		if (sock->type == SOCK_DGRAM)
485 			tsk_set_unreliable(tsk, true);
486 	}
487 
488 	trace_tipc_sk_create(sk, NULL, TIPC_DUMP_NONE, " ");
489 	return 0;
490 }
491 
492 static void tipc_sk_callback(struct rcu_head *head)
493 {
494 	struct tipc_sock *tsk = container_of(head, struct tipc_sock, rcu);
495 
496 	sock_put(&tsk->sk);
497 }
498 
499 /* Caller should hold socket lock for the socket. */
500 static void __tipc_shutdown(struct socket *sock, int error)
501 {
502 	struct sock *sk = sock->sk;
503 	struct tipc_sock *tsk = tipc_sk(sk);
504 	struct net *net = sock_net(sk);
505 	long timeout = CONN_TIMEOUT_DEFAULT;
506 	u32 dnode = tsk_peer_node(tsk);
507 	struct sk_buff *skb;
508 
509 	/* Avoid that hi-prio shutdown msgs bypass msgs in link wakeup queue */
510 	tipc_wait_for_cond(sock, &timeout, (!tsk->cong_link_cnt &&
511 					    !tsk_conn_cong(tsk)));
512 
513 	/* Remove any pending SYN message */
514 	__skb_queue_purge(&sk->sk_write_queue);
515 
516 	/* Reject all unreceived messages, except on an active connection
517 	 * (which disconnects locally & sends a 'FIN+' to peer).
518 	 */
519 	while ((skb = __skb_dequeue(&sk->sk_receive_queue)) != NULL) {
520 		if (TIPC_SKB_CB(skb)->bytes_read) {
521 			kfree_skb(skb);
522 			continue;
523 		}
524 		if (!tipc_sk_type_connectionless(sk) &&
525 		    sk->sk_state != TIPC_DISCONNECTING) {
526 			tipc_set_sk_state(sk, TIPC_DISCONNECTING);
527 			tipc_node_remove_conn(net, dnode, tsk->portid);
528 		}
529 		tipc_sk_respond(sk, skb, error);
530 	}
531 
532 	if (tipc_sk_type_connectionless(sk))
533 		return;
534 
535 	if (sk->sk_state != TIPC_DISCONNECTING) {
536 		skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE,
537 				      TIPC_CONN_MSG, SHORT_H_SIZE, 0, dnode,
538 				      tsk_own_node(tsk), tsk_peer_port(tsk),
539 				      tsk->portid, error);
540 		if (skb)
541 			tipc_node_xmit_skb(net, skb, dnode, tsk->portid);
542 		tipc_node_remove_conn(net, dnode, tsk->portid);
543 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
544 	}
545 }
546 
547 /**
548  * tipc_release - destroy a TIPC socket
549  * @sock: socket to destroy
550  *
551  * This routine cleans up any messages that are still queued on the socket.
552  * For DGRAM and RDM socket types, all queued messages are rejected.
553  * For SEQPACKET and STREAM socket types, the first message is rejected
554  * and any others are discarded.  (If the first message on a STREAM socket
555  * is partially-read, it is discarded and the next one is rejected instead.)
556  *
557  * NOTE: Rejected messages are not necessarily returned to the sender!  They
558  * are returned or discarded according to the "destination droppable" setting
559  * specified for the message by the sender.
560  *
561  * Returns 0 on success, errno otherwise
562  */
563 static int tipc_release(struct socket *sock)
564 {
565 	struct sock *sk = sock->sk;
566 	struct tipc_sock *tsk;
567 
568 	/*
569 	 * Exit if socket isn't fully initialized (occurs when a failed accept()
570 	 * releases a pre-allocated child socket that was never used)
571 	 */
572 	if (sk == NULL)
573 		return 0;
574 
575 	tsk = tipc_sk(sk);
576 	lock_sock(sk);
577 
578 	trace_tipc_sk_release(sk, NULL, TIPC_DUMP_ALL, " ");
579 	__tipc_shutdown(sock, TIPC_ERR_NO_PORT);
580 	sk->sk_shutdown = SHUTDOWN_MASK;
581 	tipc_sk_leave(tsk);
582 	tipc_sk_withdraw(tsk, 0, NULL);
583 	sk_stop_timer(sk, &sk->sk_timer);
584 	tipc_sk_remove(tsk);
585 
586 	sock_orphan(sk);
587 	/* Reject any messages that accumulated in backlog queue */
588 	release_sock(sk);
589 	tipc_dest_list_purge(&tsk->cong_links);
590 	tsk->cong_link_cnt = 0;
591 	call_rcu(&tsk->rcu, tipc_sk_callback);
592 	sock->sk = NULL;
593 
594 	return 0;
595 }
596 
597 /**
598  * tipc_bind - associate or disassocate TIPC name(s) with a socket
599  * @sock: socket structure
600  * @uaddr: socket address describing name(s) and desired operation
601  * @uaddr_len: size of socket address data structure
602  *
603  * Name and name sequence binding is indicated using a positive scope value;
604  * a negative scope value unbinds the specified name.  Specifying no name
605  * (i.e. a socket address length of 0) unbinds all names from the socket.
606  *
607  * Returns 0 on success, errno otherwise
608  *
609  * NOTE: This routine doesn't need to take the socket lock since it doesn't
610  *       access any non-constant socket information.
611  */
612 static int tipc_bind(struct socket *sock, struct sockaddr *uaddr,
613 		     int uaddr_len)
614 {
615 	struct sock *sk = sock->sk;
616 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr;
617 	struct tipc_sock *tsk = tipc_sk(sk);
618 	int res = -EINVAL;
619 
620 	lock_sock(sk);
621 	if (unlikely(!uaddr_len)) {
622 		res = tipc_sk_withdraw(tsk, 0, NULL);
623 		goto exit;
624 	}
625 	if (tsk->group) {
626 		res = -EACCES;
627 		goto exit;
628 	}
629 	if (uaddr_len < sizeof(struct sockaddr_tipc)) {
630 		res = -EINVAL;
631 		goto exit;
632 	}
633 	if (addr->family != AF_TIPC) {
634 		res = -EAFNOSUPPORT;
635 		goto exit;
636 	}
637 
638 	if (addr->addrtype == TIPC_ADDR_NAME)
639 		addr->addr.nameseq.upper = addr->addr.nameseq.lower;
640 	else if (addr->addrtype != TIPC_ADDR_NAMESEQ) {
641 		res = -EAFNOSUPPORT;
642 		goto exit;
643 	}
644 
645 	if ((addr->addr.nameseq.type < TIPC_RESERVED_TYPES) &&
646 	    (addr->addr.nameseq.type != TIPC_TOP_SRV) &&
647 	    (addr->addr.nameseq.type != TIPC_CFG_SRV)) {
648 		res = -EACCES;
649 		goto exit;
650 	}
651 
652 	res = (addr->scope >= 0) ?
653 		tipc_sk_publish(tsk, addr->scope, &addr->addr.nameseq) :
654 		tipc_sk_withdraw(tsk, -addr->scope, &addr->addr.nameseq);
655 exit:
656 	release_sock(sk);
657 	return res;
658 }
659 
660 /**
661  * tipc_getname - get port ID of socket or peer socket
662  * @sock: socket structure
663  * @uaddr: area for returned socket address
664  * @uaddr_len: area for returned length of socket address
665  * @peer: 0 = own ID, 1 = current peer ID, 2 = current/former peer ID
666  *
667  * Returns 0 on success, errno otherwise
668  *
669  * NOTE: This routine doesn't need to take the socket lock since it only
670  *       accesses socket information that is unchanging (or which changes in
671  *       a completely predictable manner).
672  */
673 static int tipc_getname(struct socket *sock, struct sockaddr *uaddr,
674 			int peer)
675 {
676 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr;
677 	struct sock *sk = sock->sk;
678 	struct tipc_sock *tsk = tipc_sk(sk);
679 
680 	memset(addr, 0, sizeof(*addr));
681 	if (peer) {
682 		if ((!tipc_sk_connected(sk)) &&
683 		    ((peer != 2) || (sk->sk_state != TIPC_DISCONNECTING)))
684 			return -ENOTCONN;
685 		addr->addr.id.ref = tsk_peer_port(tsk);
686 		addr->addr.id.node = tsk_peer_node(tsk);
687 	} else {
688 		addr->addr.id.ref = tsk->portid;
689 		addr->addr.id.node = tipc_own_addr(sock_net(sk));
690 	}
691 
692 	addr->addrtype = TIPC_ADDR_ID;
693 	addr->family = AF_TIPC;
694 	addr->scope = 0;
695 	addr->addr.name.domain = 0;
696 
697 	return sizeof(*addr);
698 }
699 
700 /**
701  * tipc_poll - read and possibly block on pollmask
702  * @file: file structure associated with the socket
703  * @sock: socket for which to calculate the poll bits
704  * @wait: ???
705  *
706  * Returns pollmask value
707  *
708  * COMMENTARY:
709  * It appears that the usual socket locking mechanisms are not useful here
710  * since the pollmask info is potentially out-of-date the moment this routine
711  * exits.  TCP and other protocols seem to rely on higher level poll routines
712  * to handle any preventable race conditions, so TIPC will do the same ...
713  *
714  * IMPORTANT: The fact that a read or write operation is indicated does NOT
715  * imply that the operation will succeed, merely that it should be performed
716  * and will not block.
717  */
718 static __poll_t tipc_poll(struct file *file, struct socket *sock,
719 			      poll_table *wait)
720 {
721 	struct sock *sk = sock->sk;
722 	struct tipc_sock *tsk = tipc_sk(sk);
723 	__poll_t revents = 0;
724 
725 	sock_poll_wait(file, sock, wait);
726 	trace_tipc_sk_poll(sk, NULL, TIPC_DUMP_ALL, " ");
727 
728 	if (sk->sk_shutdown & RCV_SHUTDOWN)
729 		revents |= EPOLLRDHUP | EPOLLIN | EPOLLRDNORM;
730 	if (sk->sk_shutdown == SHUTDOWN_MASK)
731 		revents |= EPOLLHUP;
732 
733 	switch (sk->sk_state) {
734 	case TIPC_ESTABLISHED:
735 	case TIPC_CONNECTING:
736 		if (!tsk->cong_link_cnt && !tsk_conn_cong(tsk))
737 			revents |= EPOLLOUT;
738 		/* fall thru' */
739 	case TIPC_LISTEN:
740 		if (!skb_queue_empty(&sk->sk_receive_queue))
741 			revents |= EPOLLIN | EPOLLRDNORM;
742 		break;
743 	case TIPC_OPEN:
744 		if (tsk->group_is_open && !tsk->cong_link_cnt)
745 			revents |= EPOLLOUT;
746 		if (!tipc_sk_type_connectionless(sk))
747 			break;
748 		if (skb_queue_empty(&sk->sk_receive_queue))
749 			break;
750 		revents |= EPOLLIN | EPOLLRDNORM;
751 		break;
752 	case TIPC_DISCONNECTING:
753 		revents = EPOLLIN | EPOLLRDNORM | EPOLLHUP;
754 		break;
755 	}
756 	return revents;
757 }
758 
759 /**
760  * tipc_sendmcast - send multicast message
761  * @sock: socket structure
762  * @seq: destination address
763  * @msg: message to send
764  * @dlen: length of data to send
765  * @timeout: timeout to wait for wakeup
766  *
767  * Called from function tipc_sendmsg(), which has done all sanity checks
768  * Returns the number of bytes sent on success, or errno
769  */
770 static int tipc_sendmcast(struct  socket *sock, struct tipc_name_seq *seq,
771 			  struct msghdr *msg, size_t dlen, long timeout)
772 {
773 	struct sock *sk = sock->sk;
774 	struct tipc_sock *tsk = tipc_sk(sk);
775 	struct tipc_msg *hdr = &tsk->phdr;
776 	struct net *net = sock_net(sk);
777 	int mtu = tipc_bcast_get_mtu(net);
778 	struct tipc_mc_method *method = &tsk->mc_method;
779 	struct sk_buff_head pkts;
780 	struct tipc_nlist dsts;
781 	int rc;
782 
783 	if (tsk->group)
784 		return -EACCES;
785 
786 	/* Block or return if any destination link is congested */
787 	rc = tipc_wait_for_cond(sock, &timeout, !tsk->cong_link_cnt);
788 	if (unlikely(rc))
789 		return rc;
790 
791 	/* Lookup destination nodes */
792 	tipc_nlist_init(&dsts, tipc_own_addr(net));
793 	tipc_nametbl_lookup_dst_nodes(net, seq->type, seq->lower,
794 				      seq->upper, &dsts);
795 	if (!dsts.local && !dsts.remote)
796 		return -EHOSTUNREACH;
797 
798 	/* Build message header */
799 	msg_set_type(hdr, TIPC_MCAST_MSG);
800 	msg_set_hdr_sz(hdr, MCAST_H_SIZE);
801 	msg_set_lookup_scope(hdr, TIPC_CLUSTER_SCOPE);
802 	msg_set_destport(hdr, 0);
803 	msg_set_destnode(hdr, 0);
804 	msg_set_nametype(hdr, seq->type);
805 	msg_set_namelower(hdr, seq->lower);
806 	msg_set_nameupper(hdr, seq->upper);
807 
808 	/* Build message as chain of buffers */
809 	skb_queue_head_init(&pkts);
810 	rc = tipc_msg_build(hdr, msg, 0, dlen, mtu, &pkts);
811 
812 	/* Send message if build was successful */
813 	if (unlikely(rc == dlen)) {
814 		trace_tipc_sk_sendmcast(sk, skb_peek(&pkts),
815 					TIPC_DUMP_SK_SNDQ, " ");
816 		rc = tipc_mcast_xmit(net, &pkts, method, &dsts,
817 				     &tsk->cong_link_cnt);
818 	}
819 
820 	tipc_nlist_purge(&dsts);
821 
822 	return rc ? rc : dlen;
823 }
824 
825 /**
826  * tipc_send_group_msg - send a message to a member in the group
827  * @net: network namespace
828  * @m: message to send
829  * @mb: group member
830  * @dnode: destination node
831  * @dport: destination port
832  * @dlen: total length of message data
833  */
834 static int tipc_send_group_msg(struct net *net, struct tipc_sock *tsk,
835 			       struct msghdr *m, struct tipc_member *mb,
836 			       u32 dnode, u32 dport, int dlen)
837 {
838 	u16 bc_snd_nxt = tipc_group_bc_snd_nxt(tsk->group);
839 	struct tipc_mc_method *method = &tsk->mc_method;
840 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
841 	struct tipc_msg *hdr = &tsk->phdr;
842 	struct sk_buff_head pkts;
843 	int mtu, rc;
844 
845 	/* Complete message header */
846 	msg_set_type(hdr, TIPC_GRP_UCAST_MSG);
847 	msg_set_hdr_sz(hdr, GROUP_H_SIZE);
848 	msg_set_destport(hdr, dport);
849 	msg_set_destnode(hdr, dnode);
850 	msg_set_grp_bc_seqno(hdr, bc_snd_nxt);
851 
852 	/* Build message as chain of buffers */
853 	skb_queue_head_init(&pkts);
854 	mtu = tipc_node_get_mtu(net, dnode, tsk->portid);
855 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
856 	if (unlikely(rc != dlen))
857 		return rc;
858 
859 	/* Send message */
860 	rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
861 	if (unlikely(rc == -ELINKCONG)) {
862 		tipc_dest_push(&tsk->cong_links, dnode, 0);
863 		tsk->cong_link_cnt++;
864 	}
865 
866 	/* Update send window */
867 	tipc_group_update_member(mb, blks);
868 
869 	/* A broadcast sent within next EXPIRE period must follow same path */
870 	method->rcast = true;
871 	method->mandatory = true;
872 	return dlen;
873 }
874 
875 /**
876  * tipc_send_group_unicast - send message to a member in the group
877  * @sock: socket structure
878  * @m: message to send
879  * @dlen: total length of message data
880  * @timeout: timeout to wait for wakeup
881  *
882  * Called from function tipc_sendmsg(), which has done all sanity checks
883  * Returns the number of bytes sent on success, or errno
884  */
885 static int tipc_send_group_unicast(struct socket *sock, struct msghdr *m,
886 				   int dlen, long timeout)
887 {
888 	struct sock *sk = sock->sk;
889 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
890 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
891 	struct tipc_sock *tsk = tipc_sk(sk);
892 	struct net *net = sock_net(sk);
893 	struct tipc_member *mb = NULL;
894 	u32 node, port;
895 	int rc;
896 
897 	node = dest->addr.id.node;
898 	port = dest->addr.id.ref;
899 	if (!port && !node)
900 		return -EHOSTUNREACH;
901 
902 	/* Block or return if destination link or member is congested */
903 	rc = tipc_wait_for_cond(sock, &timeout,
904 				!tipc_dest_find(&tsk->cong_links, node, 0) &&
905 				tsk->group &&
906 				!tipc_group_cong(tsk->group, node, port, blks,
907 						 &mb));
908 	if (unlikely(rc))
909 		return rc;
910 
911 	if (unlikely(!mb))
912 		return -EHOSTUNREACH;
913 
914 	rc = tipc_send_group_msg(net, tsk, m, mb, node, port, dlen);
915 
916 	return rc ? rc : dlen;
917 }
918 
919 /**
920  * tipc_send_group_anycast - send message to any member with given identity
921  * @sock: socket structure
922  * @m: message to send
923  * @dlen: total length of message data
924  * @timeout: timeout to wait for wakeup
925  *
926  * Called from function tipc_sendmsg(), which has done all sanity checks
927  * Returns the number of bytes sent on success, or errno
928  */
929 static int tipc_send_group_anycast(struct socket *sock, struct msghdr *m,
930 				   int dlen, long timeout)
931 {
932 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
933 	struct sock *sk = sock->sk;
934 	struct tipc_sock *tsk = tipc_sk(sk);
935 	struct list_head *cong_links = &tsk->cong_links;
936 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
937 	struct tipc_msg *hdr = &tsk->phdr;
938 	struct tipc_member *first = NULL;
939 	struct tipc_member *mbr = NULL;
940 	struct net *net = sock_net(sk);
941 	u32 node, port, exclude;
942 	struct list_head dsts;
943 	u32 type, inst, scope;
944 	int lookups = 0;
945 	int dstcnt, rc;
946 	bool cong;
947 
948 	INIT_LIST_HEAD(&dsts);
949 
950 	type = msg_nametype(hdr);
951 	inst = dest->addr.name.name.instance;
952 	scope = msg_lookup_scope(hdr);
953 
954 	while (++lookups < 4) {
955 		exclude = tipc_group_exclude(tsk->group);
956 
957 		first = NULL;
958 
959 		/* Look for a non-congested destination member, if any */
960 		while (1) {
961 			if (!tipc_nametbl_lookup(net, type, inst, scope, &dsts,
962 						 &dstcnt, exclude, false))
963 				return -EHOSTUNREACH;
964 			tipc_dest_pop(&dsts, &node, &port);
965 			cong = tipc_group_cong(tsk->group, node, port, blks,
966 					       &mbr);
967 			if (!cong)
968 				break;
969 			if (mbr == first)
970 				break;
971 			if (!first)
972 				first = mbr;
973 		}
974 
975 		/* Start over if destination was not in member list */
976 		if (unlikely(!mbr))
977 			continue;
978 
979 		if (likely(!cong && !tipc_dest_find(cong_links, node, 0)))
980 			break;
981 
982 		/* Block or return if destination link or member is congested */
983 		rc = tipc_wait_for_cond(sock, &timeout,
984 					!tipc_dest_find(cong_links, node, 0) &&
985 					tsk->group &&
986 					!tipc_group_cong(tsk->group, node, port,
987 							 blks, &mbr));
988 		if (unlikely(rc))
989 			return rc;
990 
991 		/* Send, unless destination disappeared while waiting */
992 		if (likely(mbr))
993 			break;
994 	}
995 
996 	if (unlikely(lookups >= 4))
997 		return -EHOSTUNREACH;
998 
999 	rc = tipc_send_group_msg(net, tsk, m, mbr, node, port, dlen);
1000 
1001 	return rc ? rc : dlen;
1002 }
1003 
1004 /**
1005  * tipc_send_group_bcast - send message to all members in communication group
1006  * @sk: socket structure
1007  * @m: message to send
1008  * @dlen: total length of message data
1009  * @timeout: timeout to wait for wakeup
1010  *
1011  * Called from function tipc_sendmsg(), which has done all sanity checks
1012  * Returns the number of bytes sent on success, or errno
1013  */
1014 static int tipc_send_group_bcast(struct socket *sock, struct msghdr *m,
1015 				 int dlen, long timeout)
1016 {
1017 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1018 	struct sock *sk = sock->sk;
1019 	struct net *net = sock_net(sk);
1020 	struct tipc_sock *tsk = tipc_sk(sk);
1021 	struct tipc_nlist *dsts;
1022 	struct tipc_mc_method *method = &tsk->mc_method;
1023 	bool ack = method->mandatory && method->rcast;
1024 	int blks = tsk_blocks(MCAST_H_SIZE + dlen);
1025 	struct tipc_msg *hdr = &tsk->phdr;
1026 	int mtu = tipc_bcast_get_mtu(net);
1027 	struct sk_buff_head pkts;
1028 	int rc = -EHOSTUNREACH;
1029 
1030 	/* Block or return if any destination link or member is congested */
1031 	rc = tipc_wait_for_cond(sock, &timeout,
1032 				!tsk->cong_link_cnt && tsk->group &&
1033 				!tipc_group_bc_cong(tsk->group, blks));
1034 	if (unlikely(rc))
1035 		return rc;
1036 
1037 	dsts = tipc_group_dests(tsk->group);
1038 	if (!dsts->local && !dsts->remote)
1039 		return -EHOSTUNREACH;
1040 
1041 	/* Complete message header */
1042 	if (dest) {
1043 		msg_set_type(hdr, TIPC_GRP_MCAST_MSG);
1044 		msg_set_nameinst(hdr, dest->addr.name.name.instance);
1045 	} else {
1046 		msg_set_type(hdr, TIPC_GRP_BCAST_MSG);
1047 		msg_set_nameinst(hdr, 0);
1048 	}
1049 	msg_set_hdr_sz(hdr, GROUP_H_SIZE);
1050 	msg_set_destport(hdr, 0);
1051 	msg_set_destnode(hdr, 0);
1052 	msg_set_grp_bc_seqno(hdr, tipc_group_bc_snd_nxt(tsk->group));
1053 
1054 	/* Avoid getting stuck with repeated forced replicasts */
1055 	msg_set_grp_bc_ack_req(hdr, ack);
1056 
1057 	/* Build message as chain of buffers */
1058 	skb_queue_head_init(&pkts);
1059 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
1060 	if (unlikely(rc != dlen))
1061 		return rc;
1062 
1063 	/* Send message */
1064 	rc = tipc_mcast_xmit(net, &pkts, method, dsts, &tsk->cong_link_cnt);
1065 	if (unlikely(rc))
1066 		return rc;
1067 
1068 	/* Update broadcast sequence number and send windows */
1069 	tipc_group_update_bc_members(tsk->group, blks, ack);
1070 
1071 	/* Broadcast link is now free to choose method for next broadcast */
1072 	method->mandatory = false;
1073 	method->expires = jiffies;
1074 
1075 	return dlen;
1076 }
1077 
1078 /**
1079  * tipc_send_group_mcast - send message to all members with given identity
1080  * @sock: socket structure
1081  * @m: message to send
1082  * @dlen: total length of message data
1083  * @timeout: timeout to wait for wakeup
1084  *
1085  * Called from function tipc_sendmsg(), which has done all sanity checks
1086  * Returns the number of bytes sent on success, or errno
1087  */
1088 static int tipc_send_group_mcast(struct socket *sock, struct msghdr *m,
1089 				 int dlen, long timeout)
1090 {
1091 	struct sock *sk = sock->sk;
1092 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1093 	struct tipc_sock *tsk = tipc_sk(sk);
1094 	struct tipc_group *grp = tsk->group;
1095 	struct tipc_msg *hdr = &tsk->phdr;
1096 	struct net *net = sock_net(sk);
1097 	u32 type, inst, scope, exclude;
1098 	struct list_head dsts;
1099 	u32 dstcnt;
1100 
1101 	INIT_LIST_HEAD(&dsts);
1102 
1103 	type = msg_nametype(hdr);
1104 	inst = dest->addr.name.name.instance;
1105 	scope = msg_lookup_scope(hdr);
1106 	exclude = tipc_group_exclude(grp);
1107 
1108 	if (!tipc_nametbl_lookup(net, type, inst, scope, &dsts,
1109 				 &dstcnt, exclude, true))
1110 		return -EHOSTUNREACH;
1111 
1112 	if (dstcnt == 1) {
1113 		tipc_dest_pop(&dsts, &dest->addr.id.node, &dest->addr.id.ref);
1114 		return tipc_send_group_unicast(sock, m, dlen, timeout);
1115 	}
1116 
1117 	tipc_dest_list_purge(&dsts);
1118 	return tipc_send_group_bcast(sock, m, dlen, timeout);
1119 }
1120 
1121 /**
1122  * tipc_sk_mcast_rcv - Deliver multicast messages to all destination sockets
1123  * @arrvq: queue with arriving messages, to be cloned after destination lookup
1124  * @inputq: queue with cloned messages, delivered to socket after dest lookup
1125  *
1126  * Multi-threaded: parallel calls with reference to same queues may occur
1127  */
1128 void tipc_sk_mcast_rcv(struct net *net, struct sk_buff_head *arrvq,
1129 		       struct sk_buff_head *inputq)
1130 {
1131 	u32 self = tipc_own_addr(net);
1132 	u32 type, lower, upper, scope;
1133 	struct sk_buff *skb, *_skb;
1134 	u32 portid, onode;
1135 	struct sk_buff_head tmpq;
1136 	struct list_head dports;
1137 	struct tipc_msg *hdr;
1138 	int user, mtyp, hlen;
1139 	bool exact;
1140 
1141 	__skb_queue_head_init(&tmpq);
1142 	INIT_LIST_HEAD(&dports);
1143 
1144 	skb = tipc_skb_peek(arrvq, &inputq->lock);
1145 	for (; skb; skb = tipc_skb_peek(arrvq, &inputq->lock)) {
1146 		hdr = buf_msg(skb);
1147 		user = msg_user(hdr);
1148 		mtyp = msg_type(hdr);
1149 		hlen = skb_headroom(skb) + msg_hdr_sz(hdr);
1150 		onode = msg_orignode(hdr);
1151 		type = msg_nametype(hdr);
1152 
1153 		if (mtyp == TIPC_GRP_UCAST_MSG || user == GROUP_PROTOCOL) {
1154 			spin_lock_bh(&inputq->lock);
1155 			if (skb_peek(arrvq) == skb) {
1156 				__skb_dequeue(arrvq);
1157 				__skb_queue_tail(inputq, skb);
1158 			}
1159 			kfree_skb(skb);
1160 			spin_unlock_bh(&inputq->lock);
1161 			continue;
1162 		}
1163 
1164 		/* Group messages require exact scope match */
1165 		if (msg_in_group(hdr)) {
1166 			lower = 0;
1167 			upper = ~0;
1168 			scope = msg_lookup_scope(hdr);
1169 			exact = true;
1170 		} else {
1171 			/* TIPC_NODE_SCOPE means "any scope" in this context */
1172 			if (onode == self)
1173 				scope = TIPC_NODE_SCOPE;
1174 			else
1175 				scope = TIPC_CLUSTER_SCOPE;
1176 			exact = false;
1177 			lower = msg_namelower(hdr);
1178 			upper = msg_nameupper(hdr);
1179 		}
1180 
1181 		/* Create destination port list: */
1182 		tipc_nametbl_mc_lookup(net, type, lower, upper,
1183 				       scope, exact, &dports);
1184 
1185 		/* Clone message per destination */
1186 		while (tipc_dest_pop(&dports, NULL, &portid)) {
1187 			_skb = __pskb_copy(skb, hlen, GFP_ATOMIC);
1188 			if (_skb) {
1189 				msg_set_destport(buf_msg(_skb), portid);
1190 				__skb_queue_tail(&tmpq, _skb);
1191 				continue;
1192 			}
1193 			pr_warn("Failed to clone mcast rcv buffer\n");
1194 		}
1195 		/* Append to inputq if not already done by other thread */
1196 		spin_lock_bh(&inputq->lock);
1197 		if (skb_peek(arrvq) == skb) {
1198 			skb_queue_splice_tail_init(&tmpq, inputq);
1199 			kfree_skb(__skb_dequeue(arrvq));
1200 		}
1201 		spin_unlock_bh(&inputq->lock);
1202 		__skb_queue_purge(&tmpq);
1203 		kfree_skb(skb);
1204 	}
1205 	tipc_sk_rcv(net, inputq);
1206 }
1207 
1208 /**
1209  * tipc_sk_conn_proto_rcv - receive a connection mng protocol message
1210  * @tsk: receiving socket
1211  * @skb: pointer to message buffer.
1212  */
1213 static void tipc_sk_conn_proto_rcv(struct tipc_sock *tsk, struct sk_buff *skb,
1214 				   struct sk_buff_head *inputq,
1215 				   struct sk_buff_head *xmitq)
1216 {
1217 	struct tipc_msg *hdr = buf_msg(skb);
1218 	u32 onode = tsk_own_node(tsk);
1219 	struct sock *sk = &tsk->sk;
1220 	int mtyp = msg_type(hdr);
1221 	bool conn_cong;
1222 
1223 	/* Ignore if connection cannot be validated: */
1224 	if (!tsk_peer_msg(tsk, hdr)) {
1225 		trace_tipc_sk_drop_msg(sk, skb, TIPC_DUMP_NONE, "@proto_rcv!");
1226 		goto exit;
1227 	}
1228 
1229 	if (unlikely(msg_errcode(hdr))) {
1230 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
1231 		tipc_node_remove_conn(sock_net(sk), tsk_peer_node(tsk),
1232 				      tsk_peer_port(tsk));
1233 		sk->sk_state_change(sk);
1234 
1235 		/* State change is ignored if socket already awake,
1236 		 * - convert msg to abort msg and add to inqueue
1237 		 */
1238 		msg_set_user(hdr, TIPC_CRITICAL_IMPORTANCE);
1239 		msg_set_type(hdr, TIPC_CONN_MSG);
1240 		msg_set_size(hdr, BASIC_H_SIZE);
1241 		msg_set_hdr_sz(hdr, BASIC_H_SIZE);
1242 		__skb_queue_tail(inputq, skb);
1243 		return;
1244 	}
1245 
1246 	tsk->probe_unacked = false;
1247 
1248 	if (mtyp == CONN_PROBE) {
1249 		msg_set_type(hdr, CONN_PROBE_REPLY);
1250 		if (tipc_msg_reverse(onode, &skb, TIPC_OK))
1251 			__skb_queue_tail(xmitq, skb);
1252 		return;
1253 	} else if (mtyp == CONN_ACK) {
1254 		conn_cong = tsk_conn_cong(tsk);
1255 		tsk->snt_unacked -= msg_conn_ack(hdr);
1256 		if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL)
1257 			tsk->snd_win = msg_adv_win(hdr);
1258 		if (conn_cong)
1259 			sk->sk_write_space(sk);
1260 	} else if (mtyp != CONN_PROBE_REPLY) {
1261 		pr_warn("Received unknown CONN_PROTO msg\n");
1262 	}
1263 exit:
1264 	kfree_skb(skb);
1265 }
1266 
1267 /**
1268  * tipc_sendmsg - send message in connectionless manner
1269  * @sock: socket structure
1270  * @m: message to send
1271  * @dsz: amount of user data to be sent
1272  *
1273  * Message must have an destination specified explicitly.
1274  * Used for SOCK_RDM and SOCK_DGRAM messages,
1275  * and for 'SYN' messages on SOCK_SEQPACKET and SOCK_STREAM connections.
1276  * (Note: 'SYN+' is prohibited on SOCK_STREAM.)
1277  *
1278  * Returns the number of bytes sent on success, or errno otherwise
1279  */
1280 static int tipc_sendmsg(struct socket *sock,
1281 			struct msghdr *m, size_t dsz)
1282 {
1283 	struct sock *sk = sock->sk;
1284 	int ret;
1285 
1286 	lock_sock(sk);
1287 	ret = __tipc_sendmsg(sock, m, dsz);
1288 	release_sock(sk);
1289 
1290 	return ret;
1291 }
1292 
1293 static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dlen)
1294 {
1295 	struct sock *sk = sock->sk;
1296 	struct net *net = sock_net(sk);
1297 	struct tipc_sock *tsk = tipc_sk(sk);
1298 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1299 	long timeout = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
1300 	struct list_head *clinks = &tsk->cong_links;
1301 	bool syn = !tipc_sk_type_connectionless(sk);
1302 	struct tipc_group *grp = tsk->group;
1303 	struct tipc_msg *hdr = &tsk->phdr;
1304 	struct tipc_name_seq *seq;
1305 	struct sk_buff_head pkts;
1306 	u32 dport, dnode = 0;
1307 	u32 type, inst;
1308 	int mtu, rc;
1309 
1310 	if (unlikely(dlen > TIPC_MAX_USER_MSG_SIZE))
1311 		return -EMSGSIZE;
1312 
1313 	if (likely(dest)) {
1314 		if (unlikely(m->msg_namelen < sizeof(*dest)))
1315 			return -EINVAL;
1316 		if (unlikely(dest->family != AF_TIPC))
1317 			return -EINVAL;
1318 	}
1319 
1320 	if (grp) {
1321 		if (!dest)
1322 			return tipc_send_group_bcast(sock, m, dlen, timeout);
1323 		if (dest->addrtype == TIPC_ADDR_NAME)
1324 			return tipc_send_group_anycast(sock, m, dlen, timeout);
1325 		if (dest->addrtype == TIPC_ADDR_ID)
1326 			return tipc_send_group_unicast(sock, m, dlen, timeout);
1327 		if (dest->addrtype == TIPC_ADDR_MCAST)
1328 			return tipc_send_group_mcast(sock, m, dlen, timeout);
1329 		return -EINVAL;
1330 	}
1331 
1332 	if (unlikely(!dest)) {
1333 		dest = &tsk->peer;
1334 		if (!syn || dest->family != AF_TIPC)
1335 			return -EDESTADDRREQ;
1336 	}
1337 
1338 	if (unlikely(syn)) {
1339 		if (sk->sk_state == TIPC_LISTEN)
1340 			return -EPIPE;
1341 		if (sk->sk_state != TIPC_OPEN)
1342 			return -EISCONN;
1343 		if (tsk->published)
1344 			return -EOPNOTSUPP;
1345 		if (dest->addrtype == TIPC_ADDR_NAME) {
1346 			tsk->conn_type = dest->addr.name.name.type;
1347 			tsk->conn_instance = dest->addr.name.name.instance;
1348 		}
1349 		msg_set_syn(hdr, 1);
1350 	}
1351 
1352 	seq = &dest->addr.nameseq;
1353 	if (dest->addrtype == TIPC_ADDR_MCAST)
1354 		return tipc_sendmcast(sock, seq, m, dlen, timeout);
1355 
1356 	if (dest->addrtype == TIPC_ADDR_NAME) {
1357 		type = dest->addr.name.name.type;
1358 		inst = dest->addr.name.name.instance;
1359 		dnode = dest->addr.name.domain;
1360 		msg_set_type(hdr, TIPC_NAMED_MSG);
1361 		msg_set_hdr_sz(hdr, NAMED_H_SIZE);
1362 		msg_set_nametype(hdr, type);
1363 		msg_set_nameinst(hdr, inst);
1364 		msg_set_lookup_scope(hdr, tipc_node2scope(dnode));
1365 		dport = tipc_nametbl_translate(net, type, inst, &dnode);
1366 		msg_set_destnode(hdr, dnode);
1367 		msg_set_destport(hdr, dport);
1368 		if (unlikely(!dport && !dnode))
1369 			return -EHOSTUNREACH;
1370 	} else if (dest->addrtype == TIPC_ADDR_ID) {
1371 		dnode = dest->addr.id.node;
1372 		msg_set_type(hdr, TIPC_DIRECT_MSG);
1373 		msg_set_lookup_scope(hdr, 0);
1374 		msg_set_destnode(hdr, dnode);
1375 		msg_set_destport(hdr, dest->addr.id.ref);
1376 		msg_set_hdr_sz(hdr, BASIC_H_SIZE);
1377 	} else {
1378 		return -EINVAL;
1379 	}
1380 
1381 	/* Block or return if destination link is congested */
1382 	rc = tipc_wait_for_cond(sock, &timeout,
1383 				!tipc_dest_find(clinks, dnode, 0));
1384 	if (unlikely(rc))
1385 		return rc;
1386 
1387 	skb_queue_head_init(&pkts);
1388 	mtu = tipc_node_get_mtu(net, dnode, tsk->portid);
1389 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
1390 	if (unlikely(rc != dlen))
1391 		return rc;
1392 	if (unlikely(syn && !tipc_msg_skb_clone(&pkts, &sk->sk_write_queue)))
1393 		return -ENOMEM;
1394 
1395 	trace_tipc_sk_sendmsg(sk, skb_peek(&pkts), TIPC_DUMP_SK_SNDQ, " ");
1396 	rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
1397 	if (unlikely(rc == -ELINKCONG)) {
1398 		tipc_dest_push(clinks, dnode, 0);
1399 		tsk->cong_link_cnt++;
1400 		rc = 0;
1401 	}
1402 
1403 	if (unlikely(syn && !rc))
1404 		tipc_set_sk_state(sk, TIPC_CONNECTING);
1405 
1406 	return rc ? rc : dlen;
1407 }
1408 
1409 /**
1410  * tipc_sendstream - send stream-oriented data
1411  * @sock: socket structure
1412  * @m: data to send
1413  * @dsz: total length of data to be transmitted
1414  *
1415  * Used for SOCK_STREAM data.
1416  *
1417  * Returns the number of bytes sent on success (or partial success),
1418  * or errno if no data sent
1419  */
1420 static int tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dsz)
1421 {
1422 	struct sock *sk = sock->sk;
1423 	int ret;
1424 
1425 	lock_sock(sk);
1426 	ret = __tipc_sendstream(sock, m, dsz);
1427 	release_sock(sk);
1428 
1429 	return ret;
1430 }
1431 
1432 static int __tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dlen)
1433 {
1434 	struct sock *sk = sock->sk;
1435 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1436 	long timeout = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
1437 	struct tipc_sock *tsk = tipc_sk(sk);
1438 	struct tipc_msg *hdr = &tsk->phdr;
1439 	struct net *net = sock_net(sk);
1440 	struct sk_buff_head pkts;
1441 	u32 dnode = tsk_peer_node(tsk);
1442 	int send, sent = 0;
1443 	int rc = 0;
1444 
1445 	skb_queue_head_init(&pkts);
1446 
1447 	if (unlikely(dlen > INT_MAX))
1448 		return -EMSGSIZE;
1449 
1450 	/* Handle implicit connection setup */
1451 	if (unlikely(dest)) {
1452 		rc = __tipc_sendmsg(sock, m, dlen);
1453 		if (dlen && dlen == rc) {
1454 			tsk->peer_caps = tipc_node_get_capabilities(net, dnode);
1455 			tsk->snt_unacked = tsk_inc(tsk, dlen + msg_hdr_sz(hdr));
1456 		}
1457 		return rc;
1458 	}
1459 
1460 	do {
1461 		rc = tipc_wait_for_cond(sock, &timeout,
1462 					(!tsk->cong_link_cnt &&
1463 					 !tsk_conn_cong(tsk) &&
1464 					 tipc_sk_connected(sk)));
1465 		if (unlikely(rc))
1466 			break;
1467 
1468 		send = min_t(size_t, dlen - sent, TIPC_MAX_USER_MSG_SIZE);
1469 		rc = tipc_msg_build(hdr, m, sent, send, tsk->max_pkt, &pkts);
1470 		if (unlikely(rc != send))
1471 			break;
1472 
1473 		trace_tipc_sk_sendstream(sk, skb_peek(&pkts),
1474 					 TIPC_DUMP_SK_SNDQ, " ");
1475 		rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
1476 		if (unlikely(rc == -ELINKCONG)) {
1477 			tsk->cong_link_cnt = 1;
1478 			rc = 0;
1479 		}
1480 		if (likely(!rc)) {
1481 			tsk->snt_unacked += tsk_inc(tsk, send + MIN_H_SIZE);
1482 			sent += send;
1483 		}
1484 	} while (sent < dlen && !rc);
1485 
1486 	return sent ? sent : rc;
1487 }
1488 
1489 /**
1490  * tipc_send_packet - send a connection-oriented message
1491  * @sock: socket structure
1492  * @m: message to send
1493  * @dsz: length of data to be transmitted
1494  *
1495  * Used for SOCK_SEQPACKET messages.
1496  *
1497  * Returns the number of bytes sent on success, or errno otherwise
1498  */
1499 static int tipc_send_packet(struct socket *sock, struct msghdr *m, size_t dsz)
1500 {
1501 	if (dsz > TIPC_MAX_USER_MSG_SIZE)
1502 		return -EMSGSIZE;
1503 
1504 	return tipc_sendstream(sock, m, dsz);
1505 }
1506 
1507 /* tipc_sk_finish_conn - complete the setup of a connection
1508  */
1509 static void tipc_sk_finish_conn(struct tipc_sock *tsk, u32 peer_port,
1510 				u32 peer_node)
1511 {
1512 	struct sock *sk = &tsk->sk;
1513 	struct net *net = sock_net(sk);
1514 	struct tipc_msg *msg = &tsk->phdr;
1515 
1516 	msg_set_syn(msg, 0);
1517 	msg_set_destnode(msg, peer_node);
1518 	msg_set_destport(msg, peer_port);
1519 	msg_set_type(msg, TIPC_CONN_MSG);
1520 	msg_set_lookup_scope(msg, 0);
1521 	msg_set_hdr_sz(msg, SHORT_H_SIZE);
1522 
1523 	sk_reset_timer(sk, &sk->sk_timer, jiffies + CONN_PROBING_INTV);
1524 	tipc_set_sk_state(sk, TIPC_ESTABLISHED);
1525 	tipc_node_add_conn(net, peer_node, tsk->portid, peer_port);
1526 	tsk->max_pkt = tipc_node_get_mtu(net, peer_node, tsk->portid);
1527 	tsk->peer_caps = tipc_node_get_capabilities(net, peer_node);
1528 	__skb_queue_purge(&sk->sk_write_queue);
1529 	if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL)
1530 		return;
1531 
1532 	/* Fall back to message based flow control */
1533 	tsk->rcv_win = FLOWCTL_MSG_WIN;
1534 	tsk->snd_win = FLOWCTL_MSG_WIN;
1535 }
1536 
1537 /**
1538  * tipc_sk_set_orig_addr - capture sender's address for received message
1539  * @m: descriptor for message info
1540  * @hdr: received message header
1541  *
1542  * Note: Address is not captured if not requested by receiver.
1543  */
1544 static void tipc_sk_set_orig_addr(struct msghdr *m, struct sk_buff *skb)
1545 {
1546 	DECLARE_SOCKADDR(struct sockaddr_pair *, srcaddr, m->msg_name);
1547 	struct tipc_msg *hdr = buf_msg(skb);
1548 
1549 	if (!srcaddr)
1550 		return;
1551 
1552 	srcaddr->sock.family = AF_TIPC;
1553 	srcaddr->sock.addrtype = TIPC_ADDR_ID;
1554 	srcaddr->sock.scope = 0;
1555 	srcaddr->sock.addr.id.ref = msg_origport(hdr);
1556 	srcaddr->sock.addr.id.node = msg_orignode(hdr);
1557 	srcaddr->sock.addr.name.domain = 0;
1558 	m->msg_namelen = sizeof(struct sockaddr_tipc);
1559 
1560 	if (!msg_in_group(hdr))
1561 		return;
1562 
1563 	/* Group message users may also want to know sending member's id */
1564 	srcaddr->member.family = AF_TIPC;
1565 	srcaddr->member.addrtype = TIPC_ADDR_NAME;
1566 	srcaddr->member.scope = 0;
1567 	srcaddr->member.addr.name.name.type = msg_nametype(hdr);
1568 	srcaddr->member.addr.name.name.instance = TIPC_SKB_CB(skb)->orig_member;
1569 	srcaddr->member.addr.name.domain = 0;
1570 	m->msg_namelen = sizeof(*srcaddr);
1571 }
1572 
1573 /**
1574  * tipc_sk_anc_data_recv - optionally capture ancillary data for received message
1575  * @m: descriptor for message info
1576  * @skb: received message buffer
1577  * @tsk: TIPC port associated with message
1578  *
1579  * Note: Ancillary data is not captured if not requested by receiver.
1580  *
1581  * Returns 0 if successful, otherwise errno
1582  */
1583 static int tipc_sk_anc_data_recv(struct msghdr *m, struct sk_buff *skb,
1584 				 struct tipc_sock *tsk)
1585 {
1586 	struct tipc_msg *msg;
1587 	u32 anc_data[3];
1588 	u32 err;
1589 	u32 dest_type;
1590 	int has_name;
1591 	int res;
1592 
1593 	if (likely(m->msg_controllen == 0))
1594 		return 0;
1595 	msg = buf_msg(skb);
1596 
1597 	/* Optionally capture errored message object(s) */
1598 	err = msg ? msg_errcode(msg) : 0;
1599 	if (unlikely(err)) {
1600 		anc_data[0] = err;
1601 		anc_data[1] = msg_data_sz(msg);
1602 		res = put_cmsg(m, SOL_TIPC, TIPC_ERRINFO, 8, anc_data);
1603 		if (res)
1604 			return res;
1605 		if (anc_data[1]) {
1606 			if (skb_linearize(skb))
1607 				return -ENOMEM;
1608 			msg = buf_msg(skb);
1609 			res = put_cmsg(m, SOL_TIPC, TIPC_RETDATA, anc_data[1],
1610 				       msg_data(msg));
1611 			if (res)
1612 				return res;
1613 		}
1614 	}
1615 
1616 	/* Optionally capture message destination object */
1617 	dest_type = msg ? msg_type(msg) : TIPC_DIRECT_MSG;
1618 	switch (dest_type) {
1619 	case TIPC_NAMED_MSG:
1620 		has_name = 1;
1621 		anc_data[0] = msg_nametype(msg);
1622 		anc_data[1] = msg_namelower(msg);
1623 		anc_data[2] = msg_namelower(msg);
1624 		break;
1625 	case TIPC_MCAST_MSG:
1626 		has_name = 1;
1627 		anc_data[0] = msg_nametype(msg);
1628 		anc_data[1] = msg_namelower(msg);
1629 		anc_data[2] = msg_nameupper(msg);
1630 		break;
1631 	case TIPC_CONN_MSG:
1632 		has_name = (tsk->conn_type != 0);
1633 		anc_data[0] = tsk->conn_type;
1634 		anc_data[1] = tsk->conn_instance;
1635 		anc_data[2] = tsk->conn_instance;
1636 		break;
1637 	default:
1638 		has_name = 0;
1639 	}
1640 	if (has_name) {
1641 		res = put_cmsg(m, SOL_TIPC, TIPC_DESTNAME, 12, anc_data);
1642 		if (res)
1643 			return res;
1644 	}
1645 
1646 	return 0;
1647 }
1648 
1649 static void tipc_sk_send_ack(struct tipc_sock *tsk)
1650 {
1651 	struct sock *sk = &tsk->sk;
1652 	struct net *net = sock_net(sk);
1653 	struct sk_buff *skb = NULL;
1654 	struct tipc_msg *msg;
1655 	u32 peer_port = tsk_peer_port(tsk);
1656 	u32 dnode = tsk_peer_node(tsk);
1657 
1658 	if (!tipc_sk_connected(sk))
1659 		return;
1660 	skb = tipc_msg_create(CONN_MANAGER, CONN_ACK, INT_H_SIZE, 0,
1661 			      dnode, tsk_own_node(tsk), peer_port,
1662 			      tsk->portid, TIPC_OK);
1663 	if (!skb)
1664 		return;
1665 	msg = buf_msg(skb);
1666 	msg_set_conn_ack(msg, tsk->rcv_unacked);
1667 	tsk->rcv_unacked = 0;
1668 
1669 	/* Adjust to and advertize the correct window limit */
1670 	if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL) {
1671 		tsk->rcv_win = tsk_adv_blocks(tsk->sk.sk_rcvbuf);
1672 		msg_set_adv_win(msg, tsk->rcv_win);
1673 	}
1674 	tipc_node_xmit_skb(net, skb, dnode, msg_link_selector(msg));
1675 }
1676 
1677 static int tipc_wait_for_rcvmsg(struct socket *sock, long *timeop)
1678 {
1679 	struct sock *sk = sock->sk;
1680 	DEFINE_WAIT(wait);
1681 	long timeo = *timeop;
1682 	int err = sock_error(sk);
1683 
1684 	if (err)
1685 		return err;
1686 
1687 	for (;;) {
1688 		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
1689 		if (timeo && skb_queue_empty(&sk->sk_receive_queue)) {
1690 			if (sk->sk_shutdown & RCV_SHUTDOWN) {
1691 				err = -ENOTCONN;
1692 				break;
1693 			}
1694 			release_sock(sk);
1695 			timeo = schedule_timeout(timeo);
1696 			lock_sock(sk);
1697 		}
1698 		err = 0;
1699 		if (!skb_queue_empty(&sk->sk_receive_queue))
1700 			break;
1701 		err = -EAGAIN;
1702 		if (!timeo)
1703 			break;
1704 		err = sock_intr_errno(timeo);
1705 		if (signal_pending(current))
1706 			break;
1707 
1708 		err = sock_error(sk);
1709 		if (err)
1710 			break;
1711 	}
1712 	finish_wait(sk_sleep(sk), &wait);
1713 	*timeop = timeo;
1714 	return err;
1715 }
1716 
1717 /**
1718  * tipc_recvmsg - receive packet-oriented message
1719  * @m: descriptor for message info
1720  * @buflen: length of user buffer area
1721  * @flags: receive flags
1722  *
1723  * Used for SOCK_DGRAM, SOCK_RDM, and SOCK_SEQPACKET messages.
1724  * If the complete message doesn't fit in user area, truncate it.
1725  *
1726  * Returns size of returned message data, errno otherwise
1727  */
1728 static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
1729 			size_t buflen,	int flags)
1730 {
1731 	struct sock *sk = sock->sk;
1732 	bool connected = !tipc_sk_type_connectionless(sk);
1733 	struct tipc_sock *tsk = tipc_sk(sk);
1734 	int rc, err, hlen, dlen, copy;
1735 	struct sk_buff_head xmitq;
1736 	struct tipc_msg *hdr;
1737 	struct sk_buff *skb;
1738 	bool grp_evt;
1739 	long timeout;
1740 
1741 	/* Catch invalid receive requests */
1742 	if (unlikely(!buflen))
1743 		return -EINVAL;
1744 
1745 	lock_sock(sk);
1746 	if (unlikely(connected && sk->sk_state == TIPC_OPEN)) {
1747 		rc = -ENOTCONN;
1748 		goto exit;
1749 	}
1750 	timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
1751 
1752 	/* Step rcv queue to first msg with data or error; wait if necessary */
1753 	do {
1754 		rc = tipc_wait_for_rcvmsg(sock, &timeout);
1755 		if (unlikely(rc))
1756 			goto exit;
1757 		skb = skb_peek(&sk->sk_receive_queue);
1758 		hdr = buf_msg(skb);
1759 		dlen = msg_data_sz(hdr);
1760 		hlen = msg_hdr_sz(hdr);
1761 		err = msg_errcode(hdr);
1762 		grp_evt = msg_is_grp_evt(hdr);
1763 		if (likely(dlen || err))
1764 			break;
1765 		tsk_advance_rx_queue(sk);
1766 	} while (1);
1767 
1768 	/* Collect msg meta data, including error code and rejected data */
1769 	tipc_sk_set_orig_addr(m, skb);
1770 	rc = tipc_sk_anc_data_recv(m, skb, tsk);
1771 	if (unlikely(rc))
1772 		goto exit;
1773 	hdr = buf_msg(skb);
1774 
1775 	/* Capture data if non-error msg, otherwise just set return value */
1776 	if (likely(!err)) {
1777 		copy = min_t(int, dlen, buflen);
1778 		if (unlikely(copy != dlen))
1779 			m->msg_flags |= MSG_TRUNC;
1780 		rc = skb_copy_datagram_msg(skb, hlen, m, copy);
1781 	} else {
1782 		copy = 0;
1783 		rc = 0;
1784 		if (err != TIPC_CONN_SHUTDOWN && connected && !m->msg_control)
1785 			rc = -ECONNRESET;
1786 	}
1787 	if (unlikely(rc))
1788 		goto exit;
1789 
1790 	/* Mark message as group event if applicable */
1791 	if (unlikely(grp_evt)) {
1792 		if (msg_grp_evt(hdr) == TIPC_WITHDRAWN)
1793 			m->msg_flags |= MSG_EOR;
1794 		m->msg_flags |= MSG_OOB;
1795 		copy = 0;
1796 	}
1797 
1798 	/* Caption of data or error code/rejected data was successful */
1799 	if (unlikely(flags & MSG_PEEK))
1800 		goto exit;
1801 
1802 	/* Send group flow control advertisement when applicable */
1803 	if (tsk->group && msg_in_group(hdr) && !grp_evt) {
1804 		skb_queue_head_init(&xmitq);
1805 		tipc_group_update_rcv_win(tsk->group, tsk_blocks(hlen + dlen),
1806 					  msg_orignode(hdr), msg_origport(hdr),
1807 					  &xmitq);
1808 		tipc_node_distr_xmit(sock_net(sk), &xmitq);
1809 	}
1810 
1811 	tsk_advance_rx_queue(sk);
1812 
1813 	if (likely(!connected))
1814 		goto exit;
1815 
1816 	/* Send connection flow control advertisement when applicable */
1817 	tsk->rcv_unacked += tsk_inc(tsk, hlen + dlen);
1818 	if (tsk->rcv_unacked >= tsk->rcv_win / TIPC_ACK_RATE)
1819 		tipc_sk_send_ack(tsk);
1820 exit:
1821 	release_sock(sk);
1822 	return rc ? rc : copy;
1823 }
1824 
1825 /**
1826  * tipc_recvstream - receive stream-oriented data
1827  * @m: descriptor for message info
1828  * @buflen: total size of user buffer area
1829  * @flags: receive flags
1830  *
1831  * Used for SOCK_STREAM messages only.  If not enough data is available
1832  * will optionally wait for more; never truncates data.
1833  *
1834  * Returns size of returned message data, errno otherwise
1835  */
1836 static int tipc_recvstream(struct socket *sock, struct msghdr *m,
1837 			   size_t buflen, int flags)
1838 {
1839 	struct sock *sk = sock->sk;
1840 	struct tipc_sock *tsk = tipc_sk(sk);
1841 	struct sk_buff *skb;
1842 	struct tipc_msg *hdr;
1843 	struct tipc_skb_cb *skb_cb;
1844 	bool peek = flags & MSG_PEEK;
1845 	int offset, required, copy, copied = 0;
1846 	int hlen, dlen, err, rc;
1847 	long timeout;
1848 
1849 	/* Catch invalid receive attempts */
1850 	if (unlikely(!buflen))
1851 		return -EINVAL;
1852 
1853 	lock_sock(sk);
1854 
1855 	if (unlikely(sk->sk_state == TIPC_OPEN)) {
1856 		rc = -ENOTCONN;
1857 		goto exit;
1858 	}
1859 	required = sock_rcvlowat(sk, flags & MSG_WAITALL, buflen);
1860 	timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
1861 
1862 	do {
1863 		/* Look at first msg in receive queue; wait if necessary */
1864 		rc = tipc_wait_for_rcvmsg(sock, &timeout);
1865 		if (unlikely(rc))
1866 			break;
1867 		skb = skb_peek(&sk->sk_receive_queue);
1868 		skb_cb = TIPC_SKB_CB(skb);
1869 		hdr = buf_msg(skb);
1870 		dlen = msg_data_sz(hdr);
1871 		hlen = msg_hdr_sz(hdr);
1872 		err = msg_errcode(hdr);
1873 
1874 		/* Discard any empty non-errored (SYN-) message */
1875 		if (unlikely(!dlen && !err)) {
1876 			tsk_advance_rx_queue(sk);
1877 			continue;
1878 		}
1879 
1880 		/* Collect msg meta data, incl. error code and rejected data */
1881 		if (!copied) {
1882 			tipc_sk_set_orig_addr(m, skb);
1883 			rc = tipc_sk_anc_data_recv(m, skb, tsk);
1884 			if (rc)
1885 				break;
1886 			hdr = buf_msg(skb);
1887 		}
1888 
1889 		/* Copy data if msg ok, otherwise return error/partial data */
1890 		if (likely(!err)) {
1891 			offset = skb_cb->bytes_read;
1892 			copy = min_t(int, dlen - offset, buflen - copied);
1893 			rc = skb_copy_datagram_msg(skb, hlen + offset, m, copy);
1894 			if (unlikely(rc))
1895 				break;
1896 			copied += copy;
1897 			offset += copy;
1898 			if (unlikely(offset < dlen)) {
1899 				if (!peek)
1900 					skb_cb->bytes_read = offset;
1901 				break;
1902 			}
1903 		} else {
1904 			rc = 0;
1905 			if ((err != TIPC_CONN_SHUTDOWN) && !m->msg_control)
1906 				rc = -ECONNRESET;
1907 			if (copied || rc)
1908 				break;
1909 		}
1910 
1911 		if (unlikely(peek))
1912 			break;
1913 
1914 		tsk_advance_rx_queue(sk);
1915 
1916 		/* Send connection flow control advertisement when applicable */
1917 		tsk->rcv_unacked += tsk_inc(tsk, hlen + dlen);
1918 		if (unlikely(tsk->rcv_unacked >= tsk->rcv_win / TIPC_ACK_RATE))
1919 			tipc_sk_send_ack(tsk);
1920 
1921 		/* Exit if all requested data or FIN/error received */
1922 		if (copied == buflen || err)
1923 			break;
1924 
1925 	} while (!skb_queue_empty(&sk->sk_receive_queue) || copied < required);
1926 exit:
1927 	release_sock(sk);
1928 	return copied ? copied : rc;
1929 }
1930 
1931 /**
1932  * tipc_write_space - wake up thread if port congestion is released
1933  * @sk: socket
1934  */
1935 static void tipc_write_space(struct sock *sk)
1936 {
1937 	struct socket_wq *wq;
1938 
1939 	rcu_read_lock();
1940 	wq = rcu_dereference(sk->sk_wq);
1941 	if (skwq_has_sleeper(wq))
1942 		wake_up_interruptible_sync_poll(&wq->wait, EPOLLOUT |
1943 						EPOLLWRNORM | EPOLLWRBAND);
1944 	rcu_read_unlock();
1945 }
1946 
1947 /**
1948  * tipc_data_ready - wake up threads to indicate messages have been received
1949  * @sk: socket
1950  * @len: the length of messages
1951  */
1952 static void tipc_data_ready(struct sock *sk)
1953 {
1954 	struct socket_wq *wq;
1955 
1956 	rcu_read_lock();
1957 	wq = rcu_dereference(sk->sk_wq);
1958 	if (skwq_has_sleeper(wq))
1959 		wake_up_interruptible_sync_poll(&wq->wait, EPOLLIN |
1960 						EPOLLRDNORM | EPOLLRDBAND);
1961 	rcu_read_unlock();
1962 }
1963 
1964 static void tipc_sock_destruct(struct sock *sk)
1965 {
1966 	__skb_queue_purge(&sk->sk_receive_queue);
1967 }
1968 
1969 static void tipc_sk_proto_rcv(struct sock *sk,
1970 			      struct sk_buff_head *inputq,
1971 			      struct sk_buff_head *xmitq)
1972 {
1973 	struct sk_buff *skb = __skb_dequeue(inputq);
1974 	struct tipc_sock *tsk = tipc_sk(sk);
1975 	struct tipc_msg *hdr = buf_msg(skb);
1976 	struct tipc_group *grp = tsk->group;
1977 	bool wakeup = false;
1978 
1979 	switch (msg_user(hdr)) {
1980 	case CONN_MANAGER:
1981 		tipc_sk_conn_proto_rcv(tsk, skb, inputq, xmitq);
1982 		return;
1983 	case SOCK_WAKEUP:
1984 		tipc_dest_del(&tsk->cong_links, msg_orignode(hdr), 0);
1985 		tsk->cong_link_cnt--;
1986 		wakeup = true;
1987 		break;
1988 	case GROUP_PROTOCOL:
1989 		tipc_group_proto_rcv(grp, &wakeup, hdr, inputq, xmitq);
1990 		break;
1991 	case TOP_SRV:
1992 		tipc_group_member_evt(tsk->group, &wakeup, &sk->sk_rcvbuf,
1993 				      hdr, inputq, xmitq);
1994 		break;
1995 	default:
1996 		break;
1997 	}
1998 
1999 	if (wakeup)
2000 		sk->sk_write_space(sk);
2001 
2002 	kfree_skb(skb);
2003 }
2004 
2005 /**
2006  * tipc_sk_filter_connect - check incoming message for a connection-based socket
2007  * @tsk: TIPC socket
2008  * @skb: pointer to message buffer.
2009  * Returns true if message should be added to receive queue, false otherwise
2010  */
2011 static bool tipc_sk_filter_connect(struct tipc_sock *tsk, struct sk_buff *skb)
2012 {
2013 	struct sock *sk = &tsk->sk;
2014 	struct net *net = sock_net(sk);
2015 	struct tipc_msg *hdr = buf_msg(skb);
2016 	bool con_msg = msg_connected(hdr);
2017 	u32 pport = tsk_peer_port(tsk);
2018 	u32 pnode = tsk_peer_node(tsk);
2019 	u32 oport = msg_origport(hdr);
2020 	u32 onode = msg_orignode(hdr);
2021 	int err = msg_errcode(hdr);
2022 	unsigned long delay;
2023 
2024 	if (unlikely(msg_mcast(hdr)))
2025 		return false;
2026 
2027 	switch (sk->sk_state) {
2028 	case TIPC_CONNECTING:
2029 		/* Setup ACK */
2030 		if (likely(con_msg)) {
2031 			if (err)
2032 				break;
2033 			tipc_sk_finish_conn(tsk, oport, onode);
2034 			msg_set_importance(&tsk->phdr, msg_importance(hdr));
2035 			/* ACK+ message with data is added to receive queue */
2036 			if (msg_data_sz(hdr))
2037 				return true;
2038 			/* Empty ACK-, - wake up sleeping connect() and drop */
2039 			sk->sk_data_ready(sk);
2040 			msg_set_dest_droppable(hdr, 1);
2041 			return false;
2042 		}
2043 		/* Ignore connectionless message if not from listening socket */
2044 		if (oport != pport || onode != pnode)
2045 			return false;
2046 
2047 		/* Rejected SYN */
2048 		if (err != TIPC_ERR_OVERLOAD)
2049 			break;
2050 
2051 		/* Prepare for new setup attempt if we have a SYN clone */
2052 		if (skb_queue_empty(&sk->sk_write_queue))
2053 			break;
2054 		get_random_bytes(&delay, 2);
2055 		delay %= (tsk->conn_timeout / 4);
2056 		delay = msecs_to_jiffies(delay + 100);
2057 		sk_reset_timer(sk, &sk->sk_timer, jiffies + delay);
2058 		return false;
2059 	case TIPC_OPEN:
2060 	case TIPC_DISCONNECTING:
2061 		return false;
2062 	case TIPC_LISTEN:
2063 		/* Accept only SYN message */
2064 		if (!msg_is_syn(hdr) &&
2065 		    tipc_node_get_capabilities(net, onode) & TIPC_SYN_BIT)
2066 			return false;
2067 		if (!con_msg && !err)
2068 			return true;
2069 		return false;
2070 	case TIPC_ESTABLISHED:
2071 		/* Accept only connection-based messages sent by peer */
2072 		if (likely(con_msg && !err && pport == oport && pnode == onode))
2073 			return true;
2074 		if (!tsk_peer_msg(tsk, hdr))
2075 			return false;
2076 		if (!err)
2077 			return true;
2078 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2079 		tipc_node_remove_conn(net, pnode, tsk->portid);
2080 		sk->sk_state_change(sk);
2081 		return true;
2082 	default:
2083 		pr_err("Unknown sk_state %u\n", sk->sk_state);
2084 	}
2085 	/* Abort connection setup attempt */
2086 	tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2087 	sk->sk_err = ECONNREFUSED;
2088 	sk->sk_state_change(sk);
2089 	return true;
2090 }
2091 
2092 /**
2093  * rcvbuf_limit - get proper overload limit of socket receive queue
2094  * @sk: socket
2095  * @skb: message
2096  *
2097  * For connection oriented messages, irrespective of importance,
2098  * default queue limit is 2 MB.
2099  *
2100  * For connectionless messages, queue limits are based on message
2101  * importance as follows:
2102  *
2103  * TIPC_LOW_IMPORTANCE       (2 MB)
2104  * TIPC_MEDIUM_IMPORTANCE    (4 MB)
2105  * TIPC_HIGH_IMPORTANCE      (8 MB)
2106  * TIPC_CRITICAL_IMPORTANCE  (16 MB)
2107  *
2108  * Returns overload limit according to corresponding message importance
2109  */
2110 static unsigned int rcvbuf_limit(struct sock *sk, struct sk_buff *skb)
2111 {
2112 	struct tipc_sock *tsk = tipc_sk(sk);
2113 	struct tipc_msg *hdr = buf_msg(skb);
2114 
2115 	if (unlikely(msg_in_group(hdr)))
2116 		return sk->sk_rcvbuf;
2117 
2118 	if (unlikely(!msg_connected(hdr)))
2119 		return sk->sk_rcvbuf << msg_importance(hdr);
2120 
2121 	if (likely(tsk->peer_caps & TIPC_BLOCK_FLOWCTL))
2122 		return sk->sk_rcvbuf;
2123 
2124 	return FLOWCTL_MSG_LIM;
2125 }
2126 
2127 /**
2128  * tipc_sk_filter_rcv - validate incoming message
2129  * @sk: socket
2130  * @skb: pointer to message.
2131  *
2132  * Enqueues message on receive queue if acceptable; optionally handles
2133  * disconnect indication for a connected socket.
2134  *
2135  * Called with socket lock already taken
2136  *
2137  */
2138 static void tipc_sk_filter_rcv(struct sock *sk, struct sk_buff *skb,
2139 			       struct sk_buff_head *xmitq)
2140 {
2141 	bool sk_conn = !tipc_sk_type_connectionless(sk);
2142 	struct tipc_sock *tsk = tipc_sk(sk);
2143 	struct tipc_group *grp = tsk->group;
2144 	struct tipc_msg *hdr = buf_msg(skb);
2145 	struct net *net = sock_net(sk);
2146 	struct sk_buff_head inputq;
2147 	int limit, err = TIPC_OK;
2148 
2149 	trace_tipc_sk_filter_rcv(sk, skb, TIPC_DUMP_ALL, " ");
2150 	TIPC_SKB_CB(skb)->bytes_read = 0;
2151 	__skb_queue_head_init(&inputq);
2152 	__skb_queue_tail(&inputq, skb);
2153 
2154 	if (unlikely(!msg_isdata(hdr)))
2155 		tipc_sk_proto_rcv(sk, &inputq, xmitq);
2156 
2157 	if (unlikely(grp))
2158 		tipc_group_filter_msg(grp, &inputq, xmitq);
2159 
2160 	/* Validate and add to receive buffer if there is space */
2161 	while ((skb = __skb_dequeue(&inputq))) {
2162 		hdr = buf_msg(skb);
2163 		limit = rcvbuf_limit(sk, skb);
2164 		if ((sk_conn && !tipc_sk_filter_connect(tsk, skb)) ||
2165 		    (!sk_conn && msg_connected(hdr)) ||
2166 		    (!grp && msg_in_group(hdr)))
2167 			err = TIPC_ERR_NO_PORT;
2168 		else if (sk_rmem_alloc_get(sk) + skb->truesize >= limit) {
2169 			trace_tipc_sk_dump(sk, skb, TIPC_DUMP_ALL,
2170 					   "err_overload2!");
2171 			atomic_inc(&sk->sk_drops);
2172 			err = TIPC_ERR_OVERLOAD;
2173 		}
2174 
2175 		if (unlikely(err)) {
2176 			if (tipc_msg_reverse(tipc_own_addr(net), &skb, err)) {
2177 				trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_NONE,
2178 						      "@filter_rcv!");
2179 				__skb_queue_tail(xmitq, skb);
2180 			}
2181 			err = TIPC_OK;
2182 			continue;
2183 		}
2184 		__skb_queue_tail(&sk->sk_receive_queue, skb);
2185 		skb_set_owner_r(skb, sk);
2186 		trace_tipc_sk_overlimit2(sk, skb, TIPC_DUMP_ALL,
2187 					 "rcvq >90% allocated!");
2188 		sk->sk_data_ready(sk);
2189 	}
2190 }
2191 
2192 /**
2193  * tipc_sk_backlog_rcv - handle incoming message from backlog queue
2194  * @sk: socket
2195  * @skb: message
2196  *
2197  * Caller must hold socket lock
2198  */
2199 static int tipc_sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
2200 {
2201 	unsigned int before = sk_rmem_alloc_get(sk);
2202 	struct sk_buff_head xmitq;
2203 	unsigned int added;
2204 
2205 	__skb_queue_head_init(&xmitq);
2206 
2207 	tipc_sk_filter_rcv(sk, skb, &xmitq);
2208 	added = sk_rmem_alloc_get(sk) - before;
2209 	atomic_add(added, &tipc_sk(sk)->dupl_rcvcnt);
2210 
2211 	/* Send pending response/rejected messages, if any */
2212 	tipc_node_distr_xmit(sock_net(sk), &xmitq);
2213 	return 0;
2214 }
2215 
2216 /**
2217  * tipc_sk_enqueue - extract all buffers with destination 'dport' from
2218  *                   inputq and try adding them to socket or backlog queue
2219  * @inputq: list of incoming buffers with potentially different destinations
2220  * @sk: socket where the buffers should be enqueued
2221  * @dport: port number for the socket
2222  *
2223  * Caller must hold socket lock
2224  */
2225 static void tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk,
2226 			    u32 dport, struct sk_buff_head *xmitq)
2227 {
2228 	unsigned long time_limit = jiffies + 2;
2229 	struct sk_buff *skb;
2230 	unsigned int lim;
2231 	atomic_t *dcnt;
2232 	u32 onode;
2233 
2234 	while (skb_queue_len(inputq)) {
2235 		if (unlikely(time_after_eq(jiffies, time_limit)))
2236 			return;
2237 
2238 		skb = tipc_skb_dequeue(inputq, dport);
2239 		if (unlikely(!skb))
2240 			return;
2241 
2242 		/* Add message directly to receive queue if possible */
2243 		if (!sock_owned_by_user(sk)) {
2244 			tipc_sk_filter_rcv(sk, skb, xmitq);
2245 			continue;
2246 		}
2247 
2248 		/* Try backlog, compensating for double-counted bytes */
2249 		dcnt = &tipc_sk(sk)->dupl_rcvcnt;
2250 		if (!sk->sk_backlog.len)
2251 			atomic_set(dcnt, 0);
2252 		lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt);
2253 		if (likely(!sk_add_backlog(sk, skb, lim))) {
2254 			trace_tipc_sk_overlimit1(sk, skb, TIPC_DUMP_ALL,
2255 						 "bklg & rcvq >90% allocated!");
2256 			continue;
2257 		}
2258 
2259 		trace_tipc_sk_dump(sk, skb, TIPC_DUMP_ALL, "err_overload!");
2260 		/* Overload => reject message back to sender */
2261 		onode = tipc_own_addr(sock_net(sk));
2262 		atomic_inc(&sk->sk_drops);
2263 		if (tipc_msg_reverse(onode, &skb, TIPC_ERR_OVERLOAD)) {
2264 			trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_ALL,
2265 					      "@sk_enqueue!");
2266 			__skb_queue_tail(xmitq, skb);
2267 		}
2268 		break;
2269 	}
2270 }
2271 
2272 /**
2273  * tipc_sk_rcv - handle a chain of incoming buffers
2274  * @inputq: buffer list containing the buffers
2275  * Consumes all buffers in list until inputq is empty
2276  * Note: may be called in multiple threads referring to the same queue
2277  */
2278 void tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq)
2279 {
2280 	struct sk_buff_head xmitq;
2281 	u32 dnode, dport = 0;
2282 	int err;
2283 	struct tipc_sock *tsk;
2284 	struct sock *sk;
2285 	struct sk_buff *skb;
2286 
2287 	__skb_queue_head_init(&xmitq);
2288 	while (skb_queue_len(inputq)) {
2289 		dport = tipc_skb_peek_port(inputq, dport);
2290 		tsk = tipc_sk_lookup(net, dport);
2291 
2292 		if (likely(tsk)) {
2293 			sk = &tsk->sk;
2294 			if (likely(spin_trylock_bh(&sk->sk_lock.slock))) {
2295 				tipc_sk_enqueue(inputq, sk, dport, &xmitq);
2296 				spin_unlock_bh(&sk->sk_lock.slock);
2297 			}
2298 			/* Send pending response/rejected messages, if any */
2299 			tipc_node_distr_xmit(sock_net(sk), &xmitq);
2300 			sock_put(sk);
2301 			continue;
2302 		}
2303 		/* No destination socket => dequeue skb if still there */
2304 		skb = tipc_skb_dequeue(inputq, dport);
2305 		if (!skb)
2306 			return;
2307 
2308 		/* Try secondary lookup if unresolved named message */
2309 		err = TIPC_ERR_NO_PORT;
2310 		if (tipc_msg_lookup_dest(net, skb, &err))
2311 			goto xmit;
2312 
2313 		/* Prepare for message rejection */
2314 		if (!tipc_msg_reverse(tipc_own_addr(net), &skb, err))
2315 			continue;
2316 
2317 		trace_tipc_sk_rej_msg(NULL, skb, TIPC_DUMP_NONE, "@sk_rcv!");
2318 xmit:
2319 		dnode = msg_destnode(buf_msg(skb));
2320 		tipc_node_xmit_skb(net, skb, dnode, dport);
2321 	}
2322 }
2323 
2324 static int tipc_wait_for_connect(struct socket *sock, long *timeo_p)
2325 {
2326 	DEFINE_WAIT_FUNC(wait, woken_wake_function);
2327 	struct sock *sk = sock->sk;
2328 	int done;
2329 
2330 	do {
2331 		int err = sock_error(sk);
2332 		if (err)
2333 			return err;
2334 		if (!*timeo_p)
2335 			return -ETIMEDOUT;
2336 		if (signal_pending(current))
2337 			return sock_intr_errno(*timeo_p);
2338 
2339 		add_wait_queue(sk_sleep(sk), &wait);
2340 		done = sk_wait_event(sk, timeo_p,
2341 				     sk->sk_state != TIPC_CONNECTING, &wait);
2342 		remove_wait_queue(sk_sleep(sk), &wait);
2343 	} while (!done);
2344 	return 0;
2345 }
2346 
2347 /**
2348  * tipc_connect - establish a connection to another TIPC port
2349  * @sock: socket structure
2350  * @dest: socket address for destination port
2351  * @destlen: size of socket address data structure
2352  * @flags: file-related flags associated with socket
2353  *
2354  * Returns 0 on success, errno otherwise
2355  */
2356 static int tipc_connect(struct socket *sock, struct sockaddr *dest,
2357 			int destlen, int flags)
2358 {
2359 	struct sock *sk = sock->sk;
2360 	struct tipc_sock *tsk = tipc_sk(sk);
2361 	struct sockaddr_tipc *dst = (struct sockaddr_tipc *)dest;
2362 	struct msghdr m = {NULL,};
2363 	long timeout = (flags & O_NONBLOCK) ? 0 : tsk->conn_timeout;
2364 	int previous;
2365 	int res = 0;
2366 
2367 	if (destlen != sizeof(struct sockaddr_tipc))
2368 		return -EINVAL;
2369 
2370 	lock_sock(sk);
2371 
2372 	if (tsk->group) {
2373 		res = -EINVAL;
2374 		goto exit;
2375 	}
2376 
2377 	if (dst->family == AF_UNSPEC) {
2378 		memset(&tsk->peer, 0, sizeof(struct sockaddr_tipc));
2379 		if (!tipc_sk_type_connectionless(sk))
2380 			res = -EINVAL;
2381 		goto exit;
2382 	} else if (dst->family != AF_TIPC) {
2383 		res = -EINVAL;
2384 	}
2385 	if (dst->addrtype != TIPC_ADDR_ID && dst->addrtype != TIPC_ADDR_NAME)
2386 		res = -EINVAL;
2387 	if (res)
2388 		goto exit;
2389 
2390 	/* DGRAM/RDM connect(), just save the destaddr */
2391 	if (tipc_sk_type_connectionless(sk)) {
2392 		memcpy(&tsk->peer, dest, destlen);
2393 		goto exit;
2394 	}
2395 
2396 	previous = sk->sk_state;
2397 
2398 	switch (sk->sk_state) {
2399 	case TIPC_OPEN:
2400 		/* Send a 'SYN-' to destination */
2401 		m.msg_name = dest;
2402 		m.msg_namelen = destlen;
2403 
2404 		/* If connect is in non-blocking case, set MSG_DONTWAIT to
2405 		 * indicate send_msg() is never blocked.
2406 		 */
2407 		if (!timeout)
2408 			m.msg_flags = MSG_DONTWAIT;
2409 
2410 		res = __tipc_sendmsg(sock, &m, 0);
2411 		if ((res < 0) && (res != -EWOULDBLOCK))
2412 			goto exit;
2413 
2414 		/* Just entered TIPC_CONNECTING state; the only
2415 		 * difference is that return value in non-blocking
2416 		 * case is EINPROGRESS, rather than EALREADY.
2417 		 */
2418 		res = -EINPROGRESS;
2419 		/* fall thru' */
2420 	case TIPC_CONNECTING:
2421 		if (!timeout) {
2422 			if (previous == TIPC_CONNECTING)
2423 				res = -EALREADY;
2424 			goto exit;
2425 		}
2426 		timeout = msecs_to_jiffies(timeout);
2427 		/* Wait until an 'ACK' or 'RST' arrives, or a timeout occurs */
2428 		res = tipc_wait_for_connect(sock, &timeout);
2429 		break;
2430 	case TIPC_ESTABLISHED:
2431 		res = -EISCONN;
2432 		break;
2433 	default:
2434 		res = -EINVAL;
2435 	}
2436 
2437 exit:
2438 	release_sock(sk);
2439 	return res;
2440 }
2441 
2442 /**
2443  * tipc_listen - allow socket to listen for incoming connections
2444  * @sock: socket structure
2445  * @len: (unused)
2446  *
2447  * Returns 0 on success, errno otherwise
2448  */
2449 static int tipc_listen(struct socket *sock, int len)
2450 {
2451 	struct sock *sk = sock->sk;
2452 	int res;
2453 
2454 	lock_sock(sk);
2455 	res = tipc_set_sk_state(sk, TIPC_LISTEN);
2456 	release_sock(sk);
2457 
2458 	return res;
2459 }
2460 
2461 static int tipc_wait_for_accept(struct socket *sock, long timeo)
2462 {
2463 	struct sock *sk = sock->sk;
2464 	DEFINE_WAIT(wait);
2465 	int err;
2466 
2467 	/* True wake-one mechanism for incoming connections: only
2468 	 * one process gets woken up, not the 'whole herd'.
2469 	 * Since we do not 'race & poll' for established sockets
2470 	 * anymore, the common case will execute the loop only once.
2471 	*/
2472 	for (;;) {
2473 		prepare_to_wait_exclusive(sk_sleep(sk), &wait,
2474 					  TASK_INTERRUPTIBLE);
2475 		if (timeo && skb_queue_empty(&sk->sk_receive_queue)) {
2476 			release_sock(sk);
2477 			timeo = schedule_timeout(timeo);
2478 			lock_sock(sk);
2479 		}
2480 		err = 0;
2481 		if (!skb_queue_empty(&sk->sk_receive_queue))
2482 			break;
2483 		err = -EAGAIN;
2484 		if (!timeo)
2485 			break;
2486 		err = sock_intr_errno(timeo);
2487 		if (signal_pending(current))
2488 			break;
2489 	}
2490 	finish_wait(sk_sleep(sk), &wait);
2491 	return err;
2492 }
2493 
2494 /**
2495  * tipc_accept - wait for connection request
2496  * @sock: listening socket
2497  * @newsock: new socket that is to be connected
2498  * @flags: file-related flags associated with socket
2499  *
2500  * Returns 0 on success, errno otherwise
2501  */
2502 static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags,
2503 		       bool kern)
2504 {
2505 	struct sock *new_sk, *sk = sock->sk;
2506 	struct sk_buff *buf;
2507 	struct tipc_sock *new_tsock;
2508 	struct tipc_msg *msg;
2509 	long timeo;
2510 	int res;
2511 
2512 	lock_sock(sk);
2513 
2514 	if (sk->sk_state != TIPC_LISTEN) {
2515 		res = -EINVAL;
2516 		goto exit;
2517 	}
2518 	timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
2519 	res = tipc_wait_for_accept(sock, timeo);
2520 	if (res)
2521 		goto exit;
2522 
2523 	buf = skb_peek(&sk->sk_receive_queue);
2524 
2525 	res = tipc_sk_create(sock_net(sock->sk), new_sock, 0, kern);
2526 	if (res)
2527 		goto exit;
2528 	security_sk_clone(sock->sk, new_sock->sk);
2529 
2530 	new_sk = new_sock->sk;
2531 	new_tsock = tipc_sk(new_sk);
2532 	msg = buf_msg(buf);
2533 
2534 	/* we lock on new_sk; but lockdep sees the lock on sk */
2535 	lock_sock_nested(new_sk, SINGLE_DEPTH_NESTING);
2536 
2537 	/*
2538 	 * Reject any stray messages received by new socket
2539 	 * before the socket lock was taken (very, very unlikely)
2540 	 */
2541 	tsk_rej_rx_queue(new_sk);
2542 
2543 	/* Connect new socket to it's peer */
2544 	tipc_sk_finish_conn(new_tsock, msg_origport(msg), msg_orignode(msg));
2545 
2546 	tsk_set_importance(new_tsock, msg_importance(msg));
2547 	if (msg_named(msg)) {
2548 		new_tsock->conn_type = msg_nametype(msg);
2549 		new_tsock->conn_instance = msg_nameinst(msg);
2550 	}
2551 
2552 	/*
2553 	 * Respond to 'SYN-' by discarding it & returning 'ACK'-.
2554 	 * Respond to 'SYN+' by queuing it on new socket.
2555 	 */
2556 	if (!msg_data_sz(msg)) {
2557 		struct msghdr m = {NULL,};
2558 
2559 		tsk_advance_rx_queue(sk);
2560 		__tipc_sendstream(new_sock, &m, 0);
2561 	} else {
2562 		__skb_dequeue(&sk->sk_receive_queue);
2563 		__skb_queue_head(&new_sk->sk_receive_queue, buf);
2564 		skb_set_owner_r(buf, new_sk);
2565 	}
2566 	release_sock(new_sk);
2567 exit:
2568 	release_sock(sk);
2569 	return res;
2570 }
2571 
2572 /**
2573  * tipc_shutdown - shutdown socket connection
2574  * @sock: socket structure
2575  * @how: direction to close (must be SHUT_RDWR)
2576  *
2577  * Terminates connection (if necessary), then purges socket's receive queue.
2578  *
2579  * Returns 0 on success, errno otherwise
2580  */
2581 static int tipc_shutdown(struct socket *sock, int how)
2582 {
2583 	struct sock *sk = sock->sk;
2584 	int res;
2585 
2586 	if (how != SHUT_RDWR)
2587 		return -EINVAL;
2588 
2589 	lock_sock(sk);
2590 
2591 	trace_tipc_sk_shutdown(sk, NULL, TIPC_DUMP_ALL, " ");
2592 	__tipc_shutdown(sock, TIPC_CONN_SHUTDOWN);
2593 	sk->sk_shutdown = SEND_SHUTDOWN;
2594 
2595 	if (sk->sk_state == TIPC_DISCONNECTING) {
2596 		/* Discard any unreceived messages */
2597 		__skb_queue_purge(&sk->sk_receive_queue);
2598 
2599 		/* Wake up anyone sleeping in poll */
2600 		sk->sk_state_change(sk);
2601 		res = 0;
2602 	} else {
2603 		res = -ENOTCONN;
2604 	}
2605 
2606 	release_sock(sk);
2607 	return res;
2608 }
2609 
2610 static void tipc_sk_check_probing_state(struct sock *sk,
2611 					struct sk_buff_head *list)
2612 {
2613 	struct tipc_sock *tsk = tipc_sk(sk);
2614 	u32 pnode = tsk_peer_node(tsk);
2615 	u32 pport = tsk_peer_port(tsk);
2616 	u32 self = tsk_own_node(tsk);
2617 	u32 oport = tsk->portid;
2618 	struct sk_buff *skb;
2619 
2620 	if (tsk->probe_unacked) {
2621 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2622 		sk->sk_err = ECONNABORTED;
2623 		tipc_node_remove_conn(sock_net(sk), pnode, pport);
2624 		sk->sk_state_change(sk);
2625 		return;
2626 	}
2627 	/* Prepare new probe */
2628 	skb = tipc_msg_create(CONN_MANAGER, CONN_PROBE, INT_H_SIZE, 0,
2629 			      pnode, self, pport, oport, TIPC_OK);
2630 	if (skb)
2631 		__skb_queue_tail(list, skb);
2632 	tsk->probe_unacked = true;
2633 	sk_reset_timer(sk, &sk->sk_timer, jiffies + CONN_PROBING_INTV);
2634 }
2635 
2636 static void tipc_sk_retry_connect(struct sock *sk, struct sk_buff_head *list)
2637 {
2638 	struct tipc_sock *tsk = tipc_sk(sk);
2639 
2640 	/* Try again later if dest link is congested */
2641 	if (tsk->cong_link_cnt) {
2642 		sk_reset_timer(sk, &sk->sk_timer, msecs_to_jiffies(100));
2643 		return;
2644 	}
2645 	/* Prepare SYN for retransmit */
2646 	tipc_msg_skb_clone(&sk->sk_write_queue, list);
2647 }
2648 
2649 static void tipc_sk_timeout(struct timer_list *t)
2650 {
2651 	struct sock *sk = from_timer(sk, t, sk_timer);
2652 	struct tipc_sock *tsk = tipc_sk(sk);
2653 	u32 pnode = tsk_peer_node(tsk);
2654 	struct sk_buff_head list;
2655 	int rc = 0;
2656 
2657 	skb_queue_head_init(&list);
2658 	bh_lock_sock(sk);
2659 
2660 	/* Try again later if socket is busy */
2661 	if (sock_owned_by_user(sk)) {
2662 		sk_reset_timer(sk, &sk->sk_timer, jiffies + HZ / 20);
2663 		bh_unlock_sock(sk);
2664 		return;
2665 	}
2666 
2667 	if (sk->sk_state == TIPC_ESTABLISHED)
2668 		tipc_sk_check_probing_state(sk, &list);
2669 	else if (sk->sk_state == TIPC_CONNECTING)
2670 		tipc_sk_retry_connect(sk, &list);
2671 
2672 	bh_unlock_sock(sk);
2673 
2674 	if (!skb_queue_empty(&list))
2675 		rc = tipc_node_xmit(sock_net(sk), &list, pnode, tsk->portid);
2676 
2677 	/* SYN messages may cause link congestion */
2678 	if (rc == -ELINKCONG) {
2679 		tipc_dest_push(&tsk->cong_links, pnode, 0);
2680 		tsk->cong_link_cnt = 1;
2681 	}
2682 	sock_put(sk);
2683 }
2684 
2685 static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
2686 			   struct tipc_name_seq const *seq)
2687 {
2688 	struct sock *sk = &tsk->sk;
2689 	struct net *net = sock_net(sk);
2690 	struct publication *publ;
2691 	u32 key;
2692 
2693 	if (scope != TIPC_NODE_SCOPE)
2694 		scope = TIPC_CLUSTER_SCOPE;
2695 
2696 	if (tipc_sk_connected(sk))
2697 		return -EINVAL;
2698 	key = tsk->portid + tsk->pub_count + 1;
2699 	if (key == tsk->portid)
2700 		return -EADDRINUSE;
2701 
2702 	publ = tipc_nametbl_publish(net, seq->type, seq->lower, seq->upper,
2703 				    scope, tsk->portid, key);
2704 	if (unlikely(!publ))
2705 		return -EINVAL;
2706 
2707 	list_add(&publ->binding_sock, &tsk->publications);
2708 	tsk->pub_count++;
2709 	tsk->published = 1;
2710 	return 0;
2711 }
2712 
2713 static int tipc_sk_withdraw(struct tipc_sock *tsk, uint scope,
2714 			    struct tipc_name_seq const *seq)
2715 {
2716 	struct net *net = sock_net(&tsk->sk);
2717 	struct publication *publ;
2718 	struct publication *safe;
2719 	int rc = -EINVAL;
2720 
2721 	if (scope != TIPC_NODE_SCOPE)
2722 		scope = TIPC_CLUSTER_SCOPE;
2723 
2724 	list_for_each_entry_safe(publ, safe, &tsk->publications, binding_sock) {
2725 		if (seq) {
2726 			if (publ->scope != scope)
2727 				continue;
2728 			if (publ->type != seq->type)
2729 				continue;
2730 			if (publ->lower != seq->lower)
2731 				continue;
2732 			if (publ->upper != seq->upper)
2733 				break;
2734 			tipc_nametbl_withdraw(net, publ->type, publ->lower,
2735 					      publ->upper, publ->key);
2736 			rc = 0;
2737 			break;
2738 		}
2739 		tipc_nametbl_withdraw(net, publ->type, publ->lower,
2740 				      publ->upper, publ->key);
2741 		rc = 0;
2742 	}
2743 	if (list_empty(&tsk->publications))
2744 		tsk->published = 0;
2745 	return rc;
2746 }
2747 
2748 /* tipc_sk_reinit: set non-zero address in all existing sockets
2749  *                 when we go from standalone to network mode.
2750  */
2751 void tipc_sk_reinit(struct net *net)
2752 {
2753 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2754 	struct rhashtable_iter iter;
2755 	struct tipc_sock *tsk;
2756 	struct tipc_msg *msg;
2757 
2758 	rhashtable_walk_enter(&tn->sk_rht, &iter);
2759 
2760 	do {
2761 		rhashtable_walk_start(&iter);
2762 
2763 		while ((tsk = rhashtable_walk_next(&iter)) && !IS_ERR(tsk)) {
2764 			sock_hold(&tsk->sk);
2765 			rhashtable_walk_stop(&iter);
2766 			lock_sock(&tsk->sk);
2767 			msg = &tsk->phdr;
2768 			msg_set_prevnode(msg, tipc_own_addr(net));
2769 			msg_set_orignode(msg, tipc_own_addr(net));
2770 			release_sock(&tsk->sk);
2771 			rhashtable_walk_start(&iter);
2772 			sock_put(&tsk->sk);
2773 		}
2774 
2775 		rhashtable_walk_stop(&iter);
2776 	} while (tsk == ERR_PTR(-EAGAIN));
2777 
2778 	rhashtable_walk_exit(&iter);
2779 }
2780 
2781 static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid)
2782 {
2783 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2784 	struct tipc_sock *tsk;
2785 
2786 	rcu_read_lock();
2787 	tsk = rhashtable_lookup_fast(&tn->sk_rht, &portid, tsk_rht_params);
2788 	if (tsk)
2789 		sock_hold(&tsk->sk);
2790 	rcu_read_unlock();
2791 
2792 	return tsk;
2793 }
2794 
2795 static int tipc_sk_insert(struct tipc_sock *tsk)
2796 {
2797 	struct sock *sk = &tsk->sk;
2798 	struct net *net = sock_net(sk);
2799 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2800 	u32 remaining = (TIPC_MAX_PORT - TIPC_MIN_PORT) + 1;
2801 	u32 portid = prandom_u32() % remaining + TIPC_MIN_PORT;
2802 
2803 	while (remaining--) {
2804 		portid++;
2805 		if ((portid < TIPC_MIN_PORT) || (portid > TIPC_MAX_PORT))
2806 			portid = TIPC_MIN_PORT;
2807 		tsk->portid = portid;
2808 		sock_hold(&tsk->sk);
2809 		if (!rhashtable_lookup_insert_fast(&tn->sk_rht, &tsk->node,
2810 						   tsk_rht_params))
2811 			return 0;
2812 		sock_put(&tsk->sk);
2813 	}
2814 
2815 	return -1;
2816 }
2817 
2818 static void tipc_sk_remove(struct tipc_sock *tsk)
2819 {
2820 	struct sock *sk = &tsk->sk;
2821 	struct tipc_net *tn = net_generic(sock_net(sk), tipc_net_id);
2822 
2823 	if (!rhashtable_remove_fast(&tn->sk_rht, &tsk->node, tsk_rht_params)) {
2824 		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
2825 		__sock_put(sk);
2826 	}
2827 }
2828 
2829 static const struct rhashtable_params tsk_rht_params = {
2830 	.nelem_hint = 192,
2831 	.head_offset = offsetof(struct tipc_sock, node),
2832 	.key_offset = offsetof(struct tipc_sock, portid),
2833 	.key_len = sizeof(u32), /* portid */
2834 	.max_size = 1048576,
2835 	.min_size = 256,
2836 	.automatic_shrinking = true,
2837 };
2838 
2839 int tipc_sk_rht_init(struct net *net)
2840 {
2841 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2842 
2843 	return rhashtable_init(&tn->sk_rht, &tsk_rht_params);
2844 }
2845 
2846 void tipc_sk_rht_destroy(struct net *net)
2847 {
2848 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2849 
2850 	/* Wait for socket readers to complete */
2851 	synchronize_net();
2852 
2853 	rhashtable_destroy(&tn->sk_rht);
2854 }
2855 
2856 static int tipc_sk_join(struct tipc_sock *tsk, struct tipc_group_req *mreq)
2857 {
2858 	struct net *net = sock_net(&tsk->sk);
2859 	struct tipc_group *grp = tsk->group;
2860 	struct tipc_msg *hdr = &tsk->phdr;
2861 	struct tipc_name_seq seq;
2862 	int rc;
2863 
2864 	if (mreq->type < TIPC_RESERVED_TYPES)
2865 		return -EACCES;
2866 	if (mreq->scope > TIPC_NODE_SCOPE)
2867 		return -EINVAL;
2868 	if (grp)
2869 		return -EACCES;
2870 	grp = tipc_group_create(net, tsk->portid, mreq, &tsk->group_is_open);
2871 	if (!grp)
2872 		return -ENOMEM;
2873 	tsk->group = grp;
2874 	msg_set_lookup_scope(hdr, mreq->scope);
2875 	msg_set_nametype(hdr, mreq->type);
2876 	msg_set_dest_droppable(hdr, true);
2877 	seq.type = mreq->type;
2878 	seq.lower = mreq->instance;
2879 	seq.upper = seq.lower;
2880 	tipc_nametbl_build_group(net, grp, mreq->type, mreq->scope);
2881 	rc = tipc_sk_publish(tsk, mreq->scope, &seq);
2882 	if (rc) {
2883 		tipc_group_delete(net, grp);
2884 		tsk->group = NULL;
2885 		return rc;
2886 	}
2887 	/* Eliminate any risk that a broadcast overtakes sent JOINs */
2888 	tsk->mc_method.rcast = true;
2889 	tsk->mc_method.mandatory = true;
2890 	tipc_group_join(net, grp, &tsk->sk.sk_rcvbuf);
2891 	return rc;
2892 }
2893 
2894 static int tipc_sk_leave(struct tipc_sock *tsk)
2895 {
2896 	struct net *net = sock_net(&tsk->sk);
2897 	struct tipc_group *grp = tsk->group;
2898 	struct tipc_name_seq seq;
2899 	int scope;
2900 
2901 	if (!grp)
2902 		return -EINVAL;
2903 	tipc_group_self(grp, &seq, &scope);
2904 	tipc_group_delete(net, grp);
2905 	tsk->group = NULL;
2906 	tipc_sk_withdraw(tsk, scope, &seq);
2907 	return 0;
2908 }
2909 
2910 /**
2911  * tipc_setsockopt - set socket option
2912  * @sock: socket structure
2913  * @lvl: option level
2914  * @opt: option identifier
2915  * @ov: pointer to new option value
2916  * @ol: length of option value
2917  *
2918  * For stream sockets only, accepts and ignores all IPPROTO_TCP options
2919  * (to ease compatibility).
2920  *
2921  * Returns 0 on success, errno otherwise
2922  */
2923 static int tipc_setsockopt(struct socket *sock, int lvl, int opt,
2924 			   char __user *ov, unsigned int ol)
2925 {
2926 	struct sock *sk = sock->sk;
2927 	struct tipc_sock *tsk = tipc_sk(sk);
2928 	struct tipc_group_req mreq;
2929 	u32 value = 0;
2930 	int res = 0;
2931 
2932 	if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM))
2933 		return 0;
2934 	if (lvl != SOL_TIPC)
2935 		return -ENOPROTOOPT;
2936 
2937 	switch (opt) {
2938 	case TIPC_IMPORTANCE:
2939 	case TIPC_SRC_DROPPABLE:
2940 	case TIPC_DEST_DROPPABLE:
2941 	case TIPC_CONN_TIMEOUT:
2942 		if (ol < sizeof(value))
2943 			return -EINVAL;
2944 		if (get_user(value, (u32 __user *)ov))
2945 			return -EFAULT;
2946 		break;
2947 	case TIPC_GROUP_JOIN:
2948 		if (ol < sizeof(mreq))
2949 			return -EINVAL;
2950 		if (copy_from_user(&mreq, ov, sizeof(mreq)))
2951 			return -EFAULT;
2952 		break;
2953 	default:
2954 		if (ov || ol)
2955 			return -EINVAL;
2956 	}
2957 
2958 	lock_sock(sk);
2959 
2960 	switch (opt) {
2961 	case TIPC_IMPORTANCE:
2962 		res = tsk_set_importance(tsk, value);
2963 		break;
2964 	case TIPC_SRC_DROPPABLE:
2965 		if (sock->type != SOCK_STREAM)
2966 			tsk_set_unreliable(tsk, value);
2967 		else
2968 			res = -ENOPROTOOPT;
2969 		break;
2970 	case TIPC_DEST_DROPPABLE:
2971 		tsk_set_unreturnable(tsk, value);
2972 		break;
2973 	case TIPC_CONN_TIMEOUT:
2974 		tipc_sk(sk)->conn_timeout = value;
2975 		break;
2976 	case TIPC_MCAST_BROADCAST:
2977 		tsk->mc_method.rcast = false;
2978 		tsk->mc_method.mandatory = true;
2979 		break;
2980 	case TIPC_MCAST_REPLICAST:
2981 		tsk->mc_method.rcast = true;
2982 		tsk->mc_method.mandatory = true;
2983 		break;
2984 	case TIPC_GROUP_JOIN:
2985 		res = tipc_sk_join(tsk, &mreq);
2986 		break;
2987 	case TIPC_GROUP_LEAVE:
2988 		res = tipc_sk_leave(tsk);
2989 		break;
2990 	default:
2991 		res = -EINVAL;
2992 	}
2993 
2994 	release_sock(sk);
2995 
2996 	return res;
2997 }
2998 
2999 /**
3000  * tipc_getsockopt - get socket option
3001  * @sock: socket structure
3002  * @lvl: option level
3003  * @opt: option identifier
3004  * @ov: receptacle for option value
3005  * @ol: receptacle for length of option value
3006  *
3007  * For stream sockets only, returns 0 length result for all IPPROTO_TCP options
3008  * (to ease compatibility).
3009  *
3010  * Returns 0 on success, errno otherwise
3011  */
3012 static int tipc_getsockopt(struct socket *sock, int lvl, int opt,
3013 			   char __user *ov, int __user *ol)
3014 {
3015 	struct sock *sk = sock->sk;
3016 	struct tipc_sock *tsk = tipc_sk(sk);
3017 	struct tipc_name_seq seq;
3018 	int len, scope;
3019 	u32 value;
3020 	int res;
3021 
3022 	if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM))
3023 		return put_user(0, ol);
3024 	if (lvl != SOL_TIPC)
3025 		return -ENOPROTOOPT;
3026 	res = get_user(len, ol);
3027 	if (res)
3028 		return res;
3029 
3030 	lock_sock(sk);
3031 
3032 	switch (opt) {
3033 	case TIPC_IMPORTANCE:
3034 		value = tsk_importance(tsk);
3035 		break;
3036 	case TIPC_SRC_DROPPABLE:
3037 		value = tsk_unreliable(tsk);
3038 		break;
3039 	case TIPC_DEST_DROPPABLE:
3040 		value = tsk_unreturnable(tsk);
3041 		break;
3042 	case TIPC_CONN_TIMEOUT:
3043 		value = tsk->conn_timeout;
3044 		/* no need to set "res", since already 0 at this point */
3045 		break;
3046 	case TIPC_NODE_RECVQ_DEPTH:
3047 		value = 0; /* was tipc_queue_size, now obsolete */
3048 		break;
3049 	case TIPC_SOCK_RECVQ_DEPTH:
3050 		value = skb_queue_len(&sk->sk_receive_queue);
3051 		break;
3052 	case TIPC_GROUP_JOIN:
3053 		seq.type = 0;
3054 		if (tsk->group)
3055 			tipc_group_self(tsk->group, &seq, &scope);
3056 		value = seq.type;
3057 		break;
3058 	default:
3059 		res = -EINVAL;
3060 	}
3061 
3062 	release_sock(sk);
3063 
3064 	if (res)
3065 		return res;	/* "get" failed */
3066 
3067 	if (len < sizeof(value))
3068 		return -EINVAL;
3069 
3070 	if (copy_to_user(ov, &value, sizeof(value)))
3071 		return -EFAULT;
3072 
3073 	return put_user(sizeof(value), ol);
3074 }
3075 
3076 static int tipc_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
3077 {
3078 	struct net *net = sock_net(sock->sk);
3079 	struct tipc_sioc_nodeid_req nr = {0};
3080 	struct tipc_sioc_ln_req lnr;
3081 	void __user *argp = (void __user *)arg;
3082 
3083 	switch (cmd) {
3084 	case SIOCGETLINKNAME:
3085 		if (copy_from_user(&lnr, argp, sizeof(lnr)))
3086 			return -EFAULT;
3087 		if (!tipc_node_get_linkname(net,
3088 					    lnr.bearer_id & 0xffff, lnr.peer,
3089 					    lnr.linkname, TIPC_MAX_LINK_NAME)) {
3090 			if (copy_to_user(argp, &lnr, sizeof(lnr)))
3091 				return -EFAULT;
3092 			return 0;
3093 		}
3094 		return -EADDRNOTAVAIL;
3095 	case SIOCGETNODEID:
3096 		if (copy_from_user(&nr, argp, sizeof(nr)))
3097 			return -EFAULT;
3098 		if (!tipc_node_get_id(net, nr.peer, nr.node_id))
3099 			return -EADDRNOTAVAIL;
3100 		if (copy_to_user(argp, &nr, sizeof(nr)))
3101 			return -EFAULT;
3102 		return 0;
3103 	default:
3104 		return -ENOIOCTLCMD;
3105 	}
3106 }
3107 
3108 static int tipc_socketpair(struct socket *sock1, struct socket *sock2)
3109 {
3110 	struct tipc_sock *tsk2 = tipc_sk(sock2->sk);
3111 	struct tipc_sock *tsk1 = tipc_sk(sock1->sk);
3112 	u32 onode = tipc_own_addr(sock_net(sock1->sk));
3113 
3114 	tsk1->peer.family = AF_TIPC;
3115 	tsk1->peer.addrtype = TIPC_ADDR_ID;
3116 	tsk1->peer.scope = TIPC_NODE_SCOPE;
3117 	tsk1->peer.addr.id.ref = tsk2->portid;
3118 	tsk1->peer.addr.id.node = onode;
3119 	tsk2->peer.family = AF_TIPC;
3120 	tsk2->peer.addrtype = TIPC_ADDR_ID;
3121 	tsk2->peer.scope = TIPC_NODE_SCOPE;
3122 	tsk2->peer.addr.id.ref = tsk1->portid;
3123 	tsk2->peer.addr.id.node = onode;
3124 
3125 	tipc_sk_finish_conn(tsk1, tsk2->portid, onode);
3126 	tipc_sk_finish_conn(tsk2, tsk1->portid, onode);
3127 	return 0;
3128 }
3129 
3130 /* Protocol switches for the various types of TIPC sockets */
3131 
3132 static const struct proto_ops msg_ops = {
3133 	.owner		= THIS_MODULE,
3134 	.family		= AF_TIPC,
3135 	.release	= tipc_release,
3136 	.bind		= tipc_bind,
3137 	.connect	= tipc_connect,
3138 	.socketpair	= tipc_socketpair,
3139 	.accept		= sock_no_accept,
3140 	.getname	= tipc_getname,
3141 	.poll		= tipc_poll,
3142 	.ioctl		= tipc_ioctl,
3143 	.listen		= sock_no_listen,
3144 	.shutdown	= tipc_shutdown,
3145 	.setsockopt	= tipc_setsockopt,
3146 	.getsockopt	= tipc_getsockopt,
3147 	.sendmsg	= tipc_sendmsg,
3148 	.recvmsg	= tipc_recvmsg,
3149 	.mmap		= sock_no_mmap,
3150 	.sendpage	= sock_no_sendpage
3151 };
3152 
3153 static const struct proto_ops packet_ops = {
3154 	.owner		= THIS_MODULE,
3155 	.family		= AF_TIPC,
3156 	.release	= tipc_release,
3157 	.bind		= tipc_bind,
3158 	.connect	= tipc_connect,
3159 	.socketpair	= tipc_socketpair,
3160 	.accept		= tipc_accept,
3161 	.getname	= tipc_getname,
3162 	.poll		= tipc_poll,
3163 	.ioctl		= tipc_ioctl,
3164 	.listen		= tipc_listen,
3165 	.shutdown	= tipc_shutdown,
3166 	.setsockopt	= tipc_setsockopt,
3167 	.getsockopt	= tipc_getsockopt,
3168 	.sendmsg	= tipc_send_packet,
3169 	.recvmsg	= tipc_recvmsg,
3170 	.mmap		= sock_no_mmap,
3171 	.sendpage	= sock_no_sendpage
3172 };
3173 
3174 static const struct proto_ops stream_ops = {
3175 	.owner		= THIS_MODULE,
3176 	.family		= AF_TIPC,
3177 	.release	= tipc_release,
3178 	.bind		= tipc_bind,
3179 	.connect	= tipc_connect,
3180 	.socketpair	= tipc_socketpair,
3181 	.accept		= tipc_accept,
3182 	.getname	= tipc_getname,
3183 	.poll		= tipc_poll,
3184 	.ioctl		= tipc_ioctl,
3185 	.listen		= tipc_listen,
3186 	.shutdown	= tipc_shutdown,
3187 	.setsockopt	= tipc_setsockopt,
3188 	.getsockopt	= tipc_getsockopt,
3189 	.sendmsg	= tipc_sendstream,
3190 	.recvmsg	= tipc_recvstream,
3191 	.mmap		= sock_no_mmap,
3192 	.sendpage	= sock_no_sendpage
3193 };
3194 
3195 static const struct net_proto_family tipc_family_ops = {
3196 	.owner		= THIS_MODULE,
3197 	.family		= AF_TIPC,
3198 	.create		= tipc_sk_create
3199 };
3200 
3201 static struct proto tipc_proto = {
3202 	.name		= "TIPC",
3203 	.owner		= THIS_MODULE,
3204 	.obj_size	= sizeof(struct tipc_sock),
3205 	.sysctl_rmem	= sysctl_tipc_rmem
3206 };
3207 
3208 /**
3209  * tipc_socket_init - initialize TIPC socket interface
3210  *
3211  * Returns 0 on success, errno otherwise
3212  */
3213 int tipc_socket_init(void)
3214 {
3215 	int res;
3216 
3217 	res = proto_register(&tipc_proto, 1);
3218 	if (res) {
3219 		pr_err("Failed to register TIPC protocol type\n");
3220 		goto out;
3221 	}
3222 
3223 	res = sock_register(&tipc_family_ops);
3224 	if (res) {
3225 		pr_err("Failed to register TIPC socket type\n");
3226 		proto_unregister(&tipc_proto);
3227 		goto out;
3228 	}
3229  out:
3230 	return res;
3231 }
3232 
3233 /**
3234  * tipc_socket_stop - stop TIPC socket interface
3235  */
3236 void tipc_socket_stop(void)
3237 {
3238 	sock_unregister(tipc_family_ops.family);
3239 	proto_unregister(&tipc_proto);
3240 }
3241 
3242 /* Caller should hold socket lock for the passed tipc socket. */
3243 static int __tipc_nl_add_sk_con(struct sk_buff *skb, struct tipc_sock *tsk)
3244 {
3245 	u32 peer_node;
3246 	u32 peer_port;
3247 	struct nlattr *nest;
3248 
3249 	peer_node = tsk_peer_node(tsk);
3250 	peer_port = tsk_peer_port(tsk);
3251 
3252 	nest = nla_nest_start(skb, TIPC_NLA_SOCK_CON);
3253 
3254 	if (nla_put_u32(skb, TIPC_NLA_CON_NODE, peer_node))
3255 		goto msg_full;
3256 	if (nla_put_u32(skb, TIPC_NLA_CON_SOCK, peer_port))
3257 		goto msg_full;
3258 
3259 	if (tsk->conn_type != 0) {
3260 		if (nla_put_flag(skb, TIPC_NLA_CON_FLAG))
3261 			goto msg_full;
3262 		if (nla_put_u32(skb, TIPC_NLA_CON_TYPE, tsk->conn_type))
3263 			goto msg_full;
3264 		if (nla_put_u32(skb, TIPC_NLA_CON_INST, tsk->conn_instance))
3265 			goto msg_full;
3266 	}
3267 	nla_nest_end(skb, nest);
3268 
3269 	return 0;
3270 
3271 msg_full:
3272 	nla_nest_cancel(skb, nest);
3273 
3274 	return -EMSGSIZE;
3275 }
3276 
3277 static int __tipc_nl_add_sk_info(struct sk_buff *skb, struct tipc_sock
3278 			  *tsk)
3279 {
3280 	struct net *net = sock_net(skb->sk);
3281 	struct sock *sk = &tsk->sk;
3282 
3283 	if (nla_put_u32(skb, TIPC_NLA_SOCK_REF, tsk->portid) ||
3284 	    nla_put_u32(skb, TIPC_NLA_SOCK_ADDR, tipc_own_addr(net)))
3285 		return -EMSGSIZE;
3286 
3287 	if (tipc_sk_connected(sk)) {
3288 		if (__tipc_nl_add_sk_con(skb, tsk))
3289 			return -EMSGSIZE;
3290 	} else if (!list_empty(&tsk->publications)) {
3291 		if (nla_put_flag(skb, TIPC_NLA_SOCK_HAS_PUBL))
3292 			return -EMSGSIZE;
3293 	}
3294 	return 0;
3295 }
3296 
3297 /* Caller should hold socket lock for the passed tipc socket. */
3298 static int __tipc_nl_add_sk(struct sk_buff *skb, struct netlink_callback *cb,
3299 			    struct tipc_sock *tsk)
3300 {
3301 	struct nlattr *attrs;
3302 	void *hdr;
3303 
3304 	hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
3305 			  &tipc_genl_family, NLM_F_MULTI, TIPC_NL_SOCK_GET);
3306 	if (!hdr)
3307 		goto msg_cancel;
3308 
3309 	attrs = nla_nest_start(skb, TIPC_NLA_SOCK);
3310 	if (!attrs)
3311 		goto genlmsg_cancel;
3312 
3313 	if (__tipc_nl_add_sk_info(skb, tsk))
3314 		goto attr_msg_cancel;
3315 
3316 	nla_nest_end(skb, attrs);
3317 	genlmsg_end(skb, hdr);
3318 
3319 	return 0;
3320 
3321 attr_msg_cancel:
3322 	nla_nest_cancel(skb, attrs);
3323 genlmsg_cancel:
3324 	genlmsg_cancel(skb, hdr);
3325 msg_cancel:
3326 	return -EMSGSIZE;
3327 }
3328 
3329 int tipc_nl_sk_walk(struct sk_buff *skb, struct netlink_callback *cb,
3330 		    int (*skb_handler)(struct sk_buff *skb,
3331 				       struct netlink_callback *cb,
3332 				       struct tipc_sock *tsk))
3333 {
3334 	struct rhashtable_iter *iter = (void *)cb->args[4];
3335 	struct tipc_sock *tsk;
3336 	int err;
3337 
3338 	rhashtable_walk_start(iter);
3339 	while ((tsk = rhashtable_walk_next(iter)) != NULL) {
3340 		if (IS_ERR(tsk)) {
3341 			err = PTR_ERR(tsk);
3342 			if (err == -EAGAIN) {
3343 				err = 0;
3344 				continue;
3345 			}
3346 			break;
3347 		}
3348 
3349 		sock_hold(&tsk->sk);
3350 		rhashtable_walk_stop(iter);
3351 		lock_sock(&tsk->sk);
3352 		err = skb_handler(skb, cb, tsk);
3353 		if (err) {
3354 			release_sock(&tsk->sk);
3355 			sock_put(&tsk->sk);
3356 			goto out;
3357 		}
3358 		release_sock(&tsk->sk);
3359 		rhashtable_walk_start(iter);
3360 		sock_put(&tsk->sk);
3361 	}
3362 	rhashtable_walk_stop(iter);
3363 out:
3364 	return skb->len;
3365 }
3366 EXPORT_SYMBOL(tipc_nl_sk_walk);
3367 
3368 int tipc_dump_start(struct netlink_callback *cb)
3369 {
3370 	return __tipc_dump_start(cb, sock_net(cb->skb->sk));
3371 }
3372 EXPORT_SYMBOL(tipc_dump_start);
3373 
3374 int __tipc_dump_start(struct netlink_callback *cb, struct net *net)
3375 {
3376 	/* tipc_nl_name_table_dump() uses cb->args[0...3]. */
3377 	struct rhashtable_iter *iter = (void *)cb->args[4];
3378 	struct tipc_net *tn = tipc_net(net);
3379 
3380 	if (!iter) {
3381 		iter = kmalloc(sizeof(*iter), GFP_KERNEL);
3382 		if (!iter)
3383 			return -ENOMEM;
3384 
3385 		cb->args[4] = (long)iter;
3386 	}
3387 
3388 	rhashtable_walk_enter(&tn->sk_rht, iter);
3389 	return 0;
3390 }
3391 
3392 int tipc_dump_done(struct netlink_callback *cb)
3393 {
3394 	struct rhashtable_iter *hti = (void *)cb->args[4];
3395 
3396 	rhashtable_walk_exit(hti);
3397 	kfree(hti);
3398 	return 0;
3399 }
3400 EXPORT_SYMBOL(tipc_dump_done);
3401 
3402 int tipc_sk_fill_sock_diag(struct sk_buff *skb, struct netlink_callback *cb,
3403 			   struct tipc_sock *tsk, u32 sk_filter_state,
3404 			   u64 (*tipc_diag_gen_cookie)(struct sock *sk))
3405 {
3406 	struct sock *sk = &tsk->sk;
3407 	struct nlattr *attrs;
3408 	struct nlattr *stat;
3409 
3410 	/*filter response w.r.t sk_state*/
3411 	if (!(sk_filter_state & (1 << sk->sk_state)))
3412 		return 0;
3413 
3414 	attrs = nla_nest_start(skb, TIPC_NLA_SOCK);
3415 	if (!attrs)
3416 		goto msg_cancel;
3417 
3418 	if (__tipc_nl_add_sk_info(skb, tsk))
3419 		goto attr_msg_cancel;
3420 
3421 	if (nla_put_u32(skb, TIPC_NLA_SOCK_TYPE, (u32)sk->sk_type) ||
3422 	    nla_put_u32(skb, TIPC_NLA_SOCK_TIPC_STATE, (u32)sk->sk_state) ||
3423 	    nla_put_u32(skb, TIPC_NLA_SOCK_INO, sock_i_ino(sk)) ||
3424 	    nla_put_u32(skb, TIPC_NLA_SOCK_UID,
3425 			from_kuid_munged(sk_user_ns(NETLINK_CB(cb->skb).sk),
3426 					 sock_i_uid(sk))) ||
3427 	    nla_put_u64_64bit(skb, TIPC_NLA_SOCK_COOKIE,
3428 			      tipc_diag_gen_cookie(sk),
3429 			      TIPC_NLA_SOCK_PAD))
3430 		goto attr_msg_cancel;
3431 
3432 	stat = nla_nest_start(skb, TIPC_NLA_SOCK_STAT);
3433 	if (!stat)
3434 		goto attr_msg_cancel;
3435 
3436 	if (nla_put_u32(skb, TIPC_NLA_SOCK_STAT_RCVQ,
3437 			skb_queue_len(&sk->sk_receive_queue)) ||
3438 	    nla_put_u32(skb, TIPC_NLA_SOCK_STAT_SENDQ,
3439 			skb_queue_len(&sk->sk_write_queue)) ||
3440 	    nla_put_u32(skb, TIPC_NLA_SOCK_STAT_DROP,
3441 			atomic_read(&sk->sk_drops)))
3442 		goto stat_msg_cancel;
3443 
3444 	if (tsk->cong_link_cnt &&
3445 	    nla_put_flag(skb, TIPC_NLA_SOCK_STAT_LINK_CONG))
3446 		goto stat_msg_cancel;
3447 
3448 	if (tsk_conn_cong(tsk) &&
3449 	    nla_put_flag(skb, TIPC_NLA_SOCK_STAT_CONN_CONG))
3450 		goto stat_msg_cancel;
3451 
3452 	nla_nest_end(skb, stat);
3453 
3454 	if (tsk->group)
3455 		if (tipc_group_fill_sock_diag(tsk->group, skb))
3456 			goto stat_msg_cancel;
3457 
3458 	nla_nest_end(skb, attrs);
3459 
3460 	return 0;
3461 
3462 stat_msg_cancel:
3463 	nla_nest_cancel(skb, stat);
3464 attr_msg_cancel:
3465 	nla_nest_cancel(skb, attrs);
3466 msg_cancel:
3467 	return -EMSGSIZE;
3468 }
3469 EXPORT_SYMBOL(tipc_sk_fill_sock_diag);
3470 
3471 int tipc_nl_sk_dump(struct sk_buff *skb, struct netlink_callback *cb)
3472 {
3473 	return tipc_nl_sk_walk(skb, cb, __tipc_nl_add_sk);
3474 }
3475 
3476 /* Caller should hold socket lock for the passed tipc socket. */
3477 static int __tipc_nl_add_sk_publ(struct sk_buff *skb,
3478 				 struct netlink_callback *cb,
3479 				 struct publication *publ)
3480 {
3481 	void *hdr;
3482 	struct nlattr *attrs;
3483 
3484 	hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
3485 			  &tipc_genl_family, NLM_F_MULTI, TIPC_NL_PUBL_GET);
3486 	if (!hdr)
3487 		goto msg_cancel;
3488 
3489 	attrs = nla_nest_start(skb, TIPC_NLA_PUBL);
3490 	if (!attrs)
3491 		goto genlmsg_cancel;
3492 
3493 	if (nla_put_u32(skb, TIPC_NLA_PUBL_KEY, publ->key))
3494 		goto attr_msg_cancel;
3495 	if (nla_put_u32(skb, TIPC_NLA_PUBL_TYPE, publ->type))
3496 		goto attr_msg_cancel;
3497 	if (nla_put_u32(skb, TIPC_NLA_PUBL_LOWER, publ->lower))
3498 		goto attr_msg_cancel;
3499 	if (nla_put_u32(skb, TIPC_NLA_PUBL_UPPER, publ->upper))
3500 		goto attr_msg_cancel;
3501 
3502 	nla_nest_end(skb, attrs);
3503 	genlmsg_end(skb, hdr);
3504 
3505 	return 0;
3506 
3507 attr_msg_cancel:
3508 	nla_nest_cancel(skb, attrs);
3509 genlmsg_cancel:
3510 	genlmsg_cancel(skb, hdr);
3511 msg_cancel:
3512 	return -EMSGSIZE;
3513 }
3514 
3515 /* Caller should hold socket lock for the passed tipc socket. */
3516 static int __tipc_nl_list_sk_publ(struct sk_buff *skb,
3517 				  struct netlink_callback *cb,
3518 				  struct tipc_sock *tsk, u32 *last_publ)
3519 {
3520 	int err;
3521 	struct publication *p;
3522 
3523 	if (*last_publ) {
3524 		list_for_each_entry(p, &tsk->publications, binding_sock) {
3525 			if (p->key == *last_publ)
3526 				break;
3527 		}
3528 		if (p->key != *last_publ) {
3529 			/* We never set seq or call nl_dump_check_consistent()
3530 			 * this means that setting prev_seq here will cause the
3531 			 * consistence check to fail in the netlink callback
3532 			 * handler. Resulting in the last NLMSG_DONE message
3533 			 * having the NLM_F_DUMP_INTR flag set.
3534 			 */
3535 			cb->prev_seq = 1;
3536 			*last_publ = 0;
3537 			return -EPIPE;
3538 		}
3539 	} else {
3540 		p = list_first_entry(&tsk->publications, struct publication,
3541 				     binding_sock);
3542 	}
3543 
3544 	list_for_each_entry_from(p, &tsk->publications, binding_sock) {
3545 		err = __tipc_nl_add_sk_publ(skb, cb, p);
3546 		if (err) {
3547 			*last_publ = p->key;
3548 			return err;
3549 		}
3550 	}
3551 	*last_publ = 0;
3552 
3553 	return 0;
3554 }
3555 
3556 int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb)
3557 {
3558 	int err;
3559 	u32 tsk_portid = cb->args[0];
3560 	u32 last_publ = cb->args[1];
3561 	u32 done = cb->args[2];
3562 	struct net *net = sock_net(skb->sk);
3563 	struct tipc_sock *tsk;
3564 
3565 	if (!tsk_portid) {
3566 		struct nlattr **attrs;
3567 		struct nlattr *sock[TIPC_NLA_SOCK_MAX + 1];
3568 
3569 		err = tipc_nlmsg_parse(cb->nlh, &attrs);
3570 		if (err)
3571 			return err;
3572 
3573 		if (!attrs[TIPC_NLA_SOCK])
3574 			return -EINVAL;
3575 
3576 		err = nla_parse_nested(sock, TIPC_NLA_SOCK_MAX,
3577 				       attrs[TIPC_NLA_SOCK],
3578 				       tipc_nl_sock_policy, NULL);
3579 		if (err)
3580 			return err;
3581 
3582 		if (!sock[TIPC_NLA_SOCK_REF])
3583 			return -EINVAL;
3584 
3585 		tsk_portid = nla_get_u32(sock[TIPC_NLA_SOCK_REF]);
3586 	}
3587 
3588 	if (done)
3589 		return 0;
3590 
3591 	tsk = tipc_sk_lookup(net, tsk_portid);
3592 	if (!tsk)
3593 		return -EINVAL;
3594 
3595 	lock_sock(&tsk->sk);
3596 	err = __tipc_nl_list_sk_publ(skb, cb, tsk, &last_publ);
3597 	if (!err)
3598 		done = 1;
3599 	release_sock(&tsk->sk);
3600 	sock_put(&tsk->sk);
3601 
3602 	cb->args[0] = tsk_portid;
3603 	cb->args[1] = last_publ;
3604 	cb->args[2] = done;
3605 
3606 	return skb->len;
3607 }
3608 
3609 /**
3610  * tipc_sk_filtering - check if a socket should be traced
3611  * @sk: the socket to be examined
3612  * @sysctl_tipc_sk_filter[]: the socket tuple for filtering,
3613  *  (portid, sock type, name type, name lower, name upper)
3614  *
3615  * Returns true if the socket meets the socket tuple data
3616  * (value 0 = 'any') or when there is no tuple set (all = 0),
3617  * otherwise false
3618  */
3619 bool tipc_sk_filtering(struct sock *sk)
3620 {
3621 	struct tipc_sock *tsk;
3622 	struct publication *p;
3623 	u32 _port, _sktype, _type, _lower, _upper;
3624 	u32 type = 0, lower = 0, upper = 0;
3625 
3626 	if (!sk)
3627 		return true;
3628 
3629 	tsk = tipc_sk(sk);
3630 
3631 	_port = sysctl_tipc_sk_filter[0];
3632 	_sktype = sysctl_tipc_sk_filter[1];
3633 	_type = sysctl_tipc_sk_filter[2];
3634 	_lower = sysctl_tipc_sk_filter[3];
3635 	_upper = sysctl_tipc_sk_filter[4];
3636 
3637 	if (!_port && !_sktype && !_type && !_lower && !_upper)
3638 		return true;
3639 
3640 	if (_port)
3641 		return (_port == tsk->portid);
3642 
3643 	if (_sktype && _sktype != sk->sk_type)
3644 		return false;
3645 
3646 	if (tsk->published) {
3647 		p = list_first_entry_or_null(&tsk->publications,
3648 					     struct publication, binding_sock);
3649 		if (p) {
3650 			type = p->type;
3651 			lower = p->lower;
3652 			upper = p->upper;
3653 		}
3654 	}
3655 
3656 	if (!tipc_sk_type_connectionless(sk)) {
3657 		type = tsk->conn_type;
3658 		lower = tsk->conn_instance;
3659 		upper = tsk->conn_instance;
3660 	}
3661 
3662 	if ((_type && _type != type) || (_lower && _lower != lower) ||
3663 	    (_upper && _upper != upper))
3664 		return false;
3665 
3666 	return true;
3667 }
3668 
3669 u32 tipc_sock_get_portid(struct sock *sk)
3670 {
3671 	return (sk) ? (tipc_sk(sk))->portid : 0;
3672 }
3673 
3674 /**
3675  * tipc_sk_overlimit1 - check if socket rx queue is about to be overloaded,
3676  *			both the rcv and backlog queues are considered
3677  * @sk: tipc sk to be checked
3678  * @skb: tipc msg to be checked
3679  *
3680  * Returns true if the socket rx queue allocation is > 90%, otherwise false
3681  */
3682 
3683 bool tipc_sk_overlimit1(struct sock *sk, struct sk_buff *skb)
3684 {
3685 	atomic_t *dcnt = &tipc_sk(sk)->dupl_rcvcnt;
3686 	unsigned int lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt);
3687 	unsigned int qsize = sk->sk_backlog.len + sk_rmem_alloc_get(sk);
3688 
3689 	return (qsize > lim * 90 / 100);
3690 }
3691 
3692 /**
3693  * tipc_sk_overlimit2 - check if socket rx queue is about to be overloaded,
3694  *			only the rcv queue is considered
3695  * @sk: tipc sk to be checked
3696  * @skb: tipc msg to be checked
3697  *
3698  * Returns true if the socket rx queue allocation is > 90%, otherwise false
3699  */
3700 
3701 bool tipc_sk_overlimit2(struct sock *sk, struct sk_buff *skb)
3702 {
3703 	unsigned int lim = rcvbuf_limit(sk, skb);
3704 	unsigned int qsize = sk_rmem_alloc_get(sk);
3705 
3706 	return (qsize > lim * 90 / 100);
3707 }
3708 
3709 /**
3710  * tipc_sk_dump - dump TIPC socket
3711  * @sk: tipc sk to be dumped
3712  * @dqueues: bitmask to decide if any socket queue to be dumped?
3713  *           - TIPC_DUMP_NONE: don't dump socket queues
3714  *           - TIPC_DUMP_SK_SNDQ: dump socket send queue
3715  *           - TIPC_DUMP_SK_RCVQ: dump socket rcv queue
3716  *           - TIPC_DUMP_SK_BKLGQ: dump socket backlog queue
3717  *           - TIPC_DUMP_ALL: dump all the socket queues above
3718  * @buf: returned buffer of dump data in format
3719  */
3720 int tipc_sk_dump(struct sock *sk, u16 dqueues, char *buf)
3721 {
3722 	int i = 0;
3723 	size_t sz = (dqueues) ? SK_LMAX : SK_LMIN;
3724 	struct tipc_sock *tsk;
3725 	struct publication *p;
3726 	bool tsk_connected;
3727 
3728 	if (!sk) {
3729 		i += scnprintf(buf, sz, "sk data: (null)\n");
3730 		return i;
3731 	}
3732 
3733 	tsk = tipc_sk(sk);
3734 	tsk_connected = !tipc_sk_type_connectionless(sk);
3735 
3736 	i += scnprintf(buf, sz, "sk data: %u", sk->sk_type);
3737 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_state);
3738 	i += scnprintf(buf + i, sz - i, " %x", tsk_own_node(tsk));
3739 	i += scnprintf(buf + i, sz - i, " %u", tsk->portid);
3740 	i += scnprintf(buf + i, sz - i, " | %u", tsk_connected);
3741 	if (tsk_connected) {
3742 		i += scnprintf(buf + i, sz - i, " %x", tsk_peer_node(tsk));
3743 		i += scnprintf(buf + i, sz - i, " %u", tsk_peer_port(tsk));
3744 		i += scnprintf(buf + i, sz - i, " %u", tsk->conn_type);
3745 		i += scnprintf(buf + i, sz - i, " %u", tsk->conn_instance);
3746 	}
3747 	i += scnprintf(buf + i, sz - i, " | %u", tsk->published);
3748 	if (tsk->published) {
3749 		p = list_first_entry_or_null(&tsk->publications,
3750 					     struct publication, binding_sock);
3751 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->type : 0);
3752 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->lower : 0);
3753 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->upper : 0);
3754 	}
3755 	i += scnprintf(buf + i, sz - i, " | %u", tsk->snd_win);
3756 	i += scnprintf(buf + i, sz - i, " %u", tsk->rcv_win);
3757 	i += scnprintf(buf + i, sz - i, " %u", tsk->max_pkt);
3758 	i += scnprintf(buf + i, sz - i, " %x", tsk->peer_caps);
3759 	i += scnprintf(buf + i, sz - i, " %u", tsk->cong_link_cnt);
3760 	i += scnprintf(buf + i, sz - i, " %u", tsk->snt_unacked);
3761 	i += scnprintf(buf + i, sz - i, " %u", tsk->rcv_unacked);
3762 	i += scnprintf(buf + i, sz - i, " %u", atomic_read(&tsk->dupl_rcvcnt));
3763 	i += scnprintf(buf + i, sz - i, " %u", sk->sk_shutdown);
3764 	i += scnprintf(buf + i, sz - i, " | %d", sk_wmem_alloc_get(sk));
3765 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_sndbuf);
3766 	i += scnprintf(buf + i, sz - i, " | %d", sk_rmem_alloc_get(sk));
3767 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_rcvbuf);
3768 	i += scnprintf(buf + i, sz - i, " | %d\n", sk->sk_backlog.len);
3769 
3770 	if (dqueues & TIPC_DUMP_SK_SNDQ) {
3771 		i += scnprintf(buf + i, sz - i, "sk_write_queue: ");
3772 		i += tipc_list_dump(&sk->sk_write_queue, false, buf + i);
3773 	}
3774 
3775 	if (dqueues & TIPC_DUMP_SK_RCVQ) {
3776 		i += scnprintf(buf + i, sz - i, "sk_receive_queue: ");
3777 		i += tipc_list_dump(&sk->sk_receive_queue, false, buf + i);
3778 	}
3779 
3780 	if (dqueues & TIPC_DUMP_SK_BKLGQ) {
3781 		i += scnprintf(buf + i, sz - i, "sk_backlog:\n  head ");
3782 		i += tipc_skb_dump(sk->sk_backlog.head, false, buf + i);
3783 		if (sk->sk_backlog.tail != sk->sk_backlog.head) {
3784 			i += scnprintf(buf + i, sz - i, "  tail ");
3785 			i += tipc_skb_dump(sk->sk_backlog.tail, false,
3786 					   buf + i);
3787 		}
3788 	}
3789 
3790 	return i;
3791 }
3792