xref: /linux/net/tipc/socket.c (revision 693d345818e106318710ac150ae252b73765d0fa)
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 	DEFINE_WAIT_FUNC(wait_, woken_wake_function);                          \
383 	struct sock *sk_;						       \
384 	int rc_;							       \
385 									       \
386 	while ((rc_ = !(condition_))) {					       \
387 		/* coupled with smp_wmb() in tipc_sk_proto_rcv() */            \
388 		smp_rmb();                                                     \
389 		sk_ = (sock_)->sk;					       \
390 		rc_ = tipc_sk_sock_err((sock_), timeo_);		       \
391 		if (rc_)						       \
392 			break;						       \
393 		add_wait_queue(sk_sleep(sk_), &wait_);                         \
394 		release_sock(sk_);					       \
395 		*(timeo_) = wait_woken(&wait_, TASK_INTERRUPTIBLE, *(timeo_)); \
396 		sched_annotate_sleep();				               \
397 		lock_sock(sk_);						       \
398 		remove_wait_queue(sk_sleep(sk_), &wait_);		       \
399 	}								       \
400 	rc_;								       \
401 })
402 
403 /**
404  * tipc_sk_create - create a TIPC socket
405  * @net: network namespace (must be default network)
406  * @sock: pre-allocated socket structure
407  * @protocol: protocol indicator (must be 0)
408  * @kern: caused by kernel or by userspace?
409  *
410  * This routine creates additional data structures used by the TIPC socket,
411  * initializes them, and links them together.
412  *
413  * Returns 0 on success, errno otherwise
414  */
415 static int tipc_sk_create(struct net *net, struct socket *sock,
416 			  int protocol, int kern)
417 {
418 	const struct proto_ops *ops;
419 	struct sock *sk;
420 	struct tipc_sock *tsk;
421 	struct tipc_msg *msg;
422 
423 	/* Validate arguments */
424 	if (unlikely(protocol != 0))
425 		return -EPROTONOSUPPORT;
426 
427 	switch (sock->type) {
428 	case SOCK_STREAM:
429 		ops = &stream_ops;
430 		break;
431 	case SOCK_SEQPACKET:
432 		ops = &packet_ops;
433 		break;
434 	case SOCK_DGRAM:
435 	case SOCK_RDM:
436 		ops = &msg_ops;
437 		break;
438 	default:
439 		return -EPROTOTYPE;
440 	}
441 
442 	/* Allocate socket's protocol area */
443 	sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto, kern);
444 	if (sk == NULL)
445 		return -ENOMEM;
446 
447 	tsk = tipc_sk(sk);
448 	tsk->max_pkt = MAX_PKT_DEFAULT;
449 	INIT_LIST_HEAD(&tsk->publications);
450 	INIT_LIST_HEAD(&tsk->cong_links);
451 	msg = &tsk->phdr;
452 
453 	/* Finish initializing socket data structures */
454 	sock->ops = ops;
455 	sock_init_data(sock, sk);
456 	tipc_set_sk_state(sk, TIPC_OPEN);
457 	if (tipc_sk_insert(tsk)) {
458 		pr_warn("Socket create failed; port number exhausted\n");
459 		return -EINVAL;
460 	}
461 
462 	/* Ensure tsk is visible before we read own_addr. */
463 	smp_mb();
464 
465 	tipc_msg_init(tipc_own_addr(net), msg, TIPC_LOW_IMPORTANCE,
466 		      TIPC_NAMED_MSG, NAMED_H_SIZE, 0);
467 
468 	msg_set_origport(msg, tsk->portid);
469 	timer_setup(&sk->sk_timer, tipc_sk_timeout, 0);
470 	sk->sk_shutdown = 0;
471 	sk->sk_backlog_rcv = tipc_sk_backlog_rcv;
472 	sk->sk_rcvbuf = sysctl_tipc_rmem[1];
473 	sk->sk_data_ready = tipc_data_ready;
474 	sk->sk_write_space = tipc_write_space;
475 	sk->sk_destruct = tipc_sock_destruct;
476 	tsk->conn_timeout = CONN_TIMEOUT_DEFAULT;
477 	tsk->group_is_open = true;
478 	atomic_set(&tsk->dupl_rcvcnt, 0);
479 
480 	/* Start out with safe limits until we receive an advertised window */
481 	tsk->snd_win = tsk_adv_blocks(RCVBUF_MIN);
482 	tsk->rcv_win = tsk->snd_win;
483 
484 	if (tipc_sk_type_connectionless(sk)) {
485 		tsk_set_unreturnable(tsk, true);
486 		if (sock->type == SOCK_DGRAM)
487 			tsk_set_unreliable(tsk, true);
488 	}
489 	__skb_queue_head_init(&tsk->mc_method.deferredq);
490 	trace_tipc_sk_create(sk, NULL, TIPC_DUMP_NONE, " ");
491 	return 0;
492 }
493 
494 static void tipc_sk_callback(struct rcu_head *head)
495 {
496 	struct tipc_sock *tsk = container_of(head, struct tipc_sock, rcu);
497 
498 	sock_put(&tsk->sk);
499 }
500 
501 /* Caller should hold socket lock for the socket. */
502 static void __tipc_shutdown(struct socket *sock, int error)
503 {
504 	struct sock *sk = sock->sk;
505 	struct tipc_sock *tsk = tipc_sk(sk);
506 	struct net *net = sock_net(sk);
507 	long timeout = CONN_TIMEOUT_DEFAULT;
508 	u32 dnode = tsk_peer_node(tsk);
509 	struct sk_buff *skb;
510 
511 	/* Avoid that hi-prio shutdown msgs bypass msgs in link wakeup queue */
512 	tipc_wait_for_cond(sock, &timeout, (!tsk->cong_link_cnt &&
513 					    !tsk_conn_cong(tsk)));
514 
515 	/* Remove any pending SYN message */
516 	__skb_queue_purge(&sk->sk_write_queue);
517 
518 	/* Reject all unreceived messages, except on an active connection
519 	 * (which disconnects locally & sends a 'FIN+' to peer).
520 	 */
521 	while ((skb = __skb_dequeue(&sk->sk_receive_queue)) != NULL) {
522 		if (TIPC_SKB_CB(skb)->bytes_read) {
523 			kfree_skb(skb);
524 			continue;
525 		}
526 		if (!tipc_sk_type_connectionless(sk) &&
527 		    sk->sk_state != TIPC_DISCONNECTING) {
528 			tipc_set_sk_state(sk, TIPC_DISCONNECTING);
529 			tipc_node_remove_conn(net, dnode, tsk->portid);
530 		}
531 		tipc_sk_respond(sk, skb, error);
532 	}
533 
534 	if (tipc_sk_type_connectionless(sk))
535 		return;
536 
537 	if (sk->sk_state != TIPC_DISCONNECTING) {
538 		skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE,
539 				      TIPC_CONN_MSG, SHORT_H_SIZE, 0, dnode,
540 				      tsk_own_node(tsk), tsk_peer_port(tsk),
541 				      tsk->portid, error);
542 		if (skb)
543 			tipc_node_xmit_skb(net, skb, dnode, tsk->portid);
544 		tipc_node_remove_conn(net, dnode, tsk->portid);
545 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
546 	}
547 }
548 
549 /**
550  * tipc_release - destroy a TIPC socket
551  * @sock: socket to destroy
552  *
553  * This routine cleans up any messages that are still queued on the socket.
554  * For DGRAM and RDM socket types, all queued messages are rejected.
555  * For SEQPACKET and STREAM socket types, the first message is rejected
556  * and any others are discarded.  (If the first message on a STREAM socket
557  * is partially-read, it is discarded and the next one is rejected instead.)
558  *
559  * NOTE: Rejected messages are not necessarily returned to the sender!  They
560  * are returned or discarded according to the "destination droppable" setting
561  * specified for the message by the sender.
562  *
563  * Returns 0 on success, errno otherwise
564  */
565 static int tipc_release(struct socket *sock)
566 {
567 	struct sock *sk = sock->sk;
568 	struct tipc_sock *tsk;
569 
570 	/*
571 	 * Exit if socket isn't fully initialized (occurs when a failed accept()
572 	 * releases a pre-allocated child socket that was never used)
573 	 */
574 	if (sk == NULL)
575 		return 0;
576 
577 	tsk = tipc_sk(sk);
578 	lock_sock(sk);
579 
580 	trace_tipc_sk_release(sk, NULL, TIPC_DUMP_ALL, " ");
581 	__tipc_shutdown(sock, TIPC_ERR_NO_PORT);
582 	sk->sk_shutdown = SHUTDOWN_MASK;
583 	tipc_sk_leave(tsk);
584 	tipc_sk_withdraw(tsk, 0, NULL);
585 	__skb_queue_purge(&tsk->mc_method.deferredq);
586 	sk_stop_timer(sk, &sk->sk_timer);
587 	tipc_sk_remove(tsk);
588 
589 	sock_orphan(sk);
590 	/* Reject any messages that accumulated in backlog queue */
591 	release_sock(sk);
592 	tipc_dest_list_purge(&tsk->cong_links);
593 	tsk->cong_link_cnt = 0;
594 	call_rcu(&tsk->rcu, tipc_sk_callback);
595 	sock->sk = NULL;
596 
597 	return 0;
598 }
599 
600 /**
601  * tipc_bind - associate or disassocate TIPC name(s) with a socket
602  * @sock: socket structure
603  * @uaddr: socket address describing name(s) and desired operation
604  * @uaddr_len: size of socket address data structure
605  *
606  * Name and name sequence binding is indicated using a positive scope value;
607  * a negative scope value unbinds the specified name.  Specifying no name
608  * (i.e. a socket address length of 0) unbinds all names from the socket.
609  *
610  * Returns 0 on success, errno otherwise
611  *
612  * NOTE: This routine doesn't need to take the socket lock since it doesn't
613  *       access any non-constant socket information.
614  */
615 static int tipc_bind(struct socket *sock, struct sockaddr *uaddr,
616 		     int uaddr_len)
617 {
618 	struct sock *sk = sock->sk;
619 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr;
620 	struct tipc_sock *tsk = tipc_sk(sk);
621 	int res = -EINVAL;
622 
623 	lock_sock(sk);
624 	if (unlikely(!uaddr_len)) {
625 		res = tipc_sk_withdraw(tsk, 0, NULL);
626 		goto exit;
627 	}
628 	if (tsk->group) {
629 		res = -EACCES;
630 		goto exit;
631 	}
632 	if (uaddr_len < sizeof(struct sockaddr_tipc)) {
633 		res = -EINVAL;
634 		goto exit;
635 	}
636 	if (addr->family != AF_TIPC) {
637 		res = -EAFNOSUPPORT;
638 		goto exit;
639 	}
640 
641 	if (addr->addrtype == TIPC_ADDR_NAME)
642 		addr->addr.nameseq.upper = addr->addr.nameseq.lower;
643 	else if (addr->addrtype != TIPC_ADDR_NAMESEQ) {
644 		res = -EAFNOSUPPORT;
645 		goto exit;
646 	}
647 
648 	if ((addr->addr.nameseq.type < TIPC_RESERVED_TYPES) &&
649 	    (addr->addr.nameseq.type != TIPC_TOP_SRV) &&
650 	    (addr->addr.nameseq.type != TIPC_CFG_SRV)) {
651 		res = -EACCES;
652 		goto exit;
653 	}
654 
655 	res = (addr->scope >= 0) ?
656 		tipc_sk_publish(tsk, addr->scope, &addr->addr.nameseq) :
657 		tipc_sk_withdraw(tsk, -addr->scope, &addr->addr.nameseq);
658 exit:
659 	release_sock(sk);
660 	return res;
661 }
662 
663 /**
664  * tipc_getname - get port ID of socket or peer socket
665  * @sock: socket structure
666  * @uaddr: area for returned socket address
667  * @uaddr_len: area for returned length of socket address
668  * @peer: 0 = own ID, 1 = current peer ID, 2 = current/former peer ID
669  *
670  * Returns 0 on success, errno otherwise
671  *
672  * NOTE: This routine doesn't need to take the socket lock since it only
673  *       accesses socket information that is unchanging (or which changes in
674  *       a completely predictable manner).
675  */
676 static int tipc_getname(struct socket *sock, struct sockaddr *uaddr,
677 			int peer)
678 {
679 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr;
680 	struct sock *sk = sock->sk;
681 	struct tipc_sock *tsk = tipc_sk(sk);
682 
683 	memset(addr, 0, sizeof(*addr));
684 	if (peer) {
685 		if ((!tipc_sk_connected(sk)) &&
686 		    ((peer != 2) || (sk->sk_state != TIPC_DISCONNECTING)))
687 			return -ENOTCONN;
688 		addr->addr.id.ref = tsk_peer_port(tsk);
689 		addr->addr.id.node = tsk_peer_node(tsk);
690 	} else {
691 		addr->addr.id.ref = tsk->portid;
692 		addr->addr.id.node = tipc_own_addr(sock_net(sk));
693 	}
694 
695 	addr->addrtype = TIPC_ADDR_ID;
696 	addr->family = AF_TIPC;
697 	addr->scope = 0;
698 	addr->addr.name.domain = 0;
699 
700 	return sizeof(*addr);
701 }
702 
703 /**
704  * tipc_poll - read and possibly block on pollmask
705  * @file: file structure associated with the socket
706  * @sock: socket for which to calculate the poll bits
707  * @wait: ???
708  *
709  * Returns pollmask value
710  *
711  * COMMENTARY:
712  * It appears that the usual socket locking mechanisms are not useful here
713  * since the pollmask info is potentially out-of-date the moment this routine
714  * exits.  TCP and other protocols seem to rely on higher level poll routines
715  * to handle any preventable race conditions, so TIPC will do the same ...
716  *
717  * IMPORTANT: The fact that a read or write operation is indicated does NOT
718  * imply that the operation will succeed, merely that it should be performed
719  * and will not block.
720  */
721 static __poll_t tipc_poll(struct file *file, struct socket *sock,
722 			      poll_table *wait)
723 {
724 	struct sock *sk = sock->sk;
725 	struct tipc_sock *tsk = tipc_sk(sk);
726 	__poll_t revents = 0;
727 
728 	sock_poll_wait(file, sock, wait);
729 	trace_tipc_sk_poll(sk, NULL, TIPC_DUMP_ALL, " ");
730 
731 	if (sk->sk_shutdown & RCV_SHUTDOWN)
732 		revents |= EPOLLRDHUP | EPOLLIN | EPOLLRDNORM;
733 	if (sk->sk_shutdown == SHUTDOWN_MASK)
734 		revents |= EPOLLHUP;
735 
736 	switch (sk->sk_state) {
737 	case TIPC_ESTABLISHED:
738 		if (!tsk->cong_link_cnt && !tsk_conn_cong(tsk))
739 			revents |= EPOLLOUT;
740 		/* fall through */
741 	case TIPC_LISTEN:
742 	case TIPC_CONNECTING:
743 		if (!skb_queue_empty(&sk->sk_receive_queue))
744 			revents |= EPOLLIN | EPOLLRDNORM;
745 		break;
746 	case TIPC_OPEN:
747 		if (tsk->group_is_open && !tsk->cong_link_cnt)
748 			revents |= EPOLLOUT;
749 		if (!tipc_sk_type_connectionless(sk))
750 			break;
751 		if (skb_queue_empty(&sk->sk_receive_queue))
752 			break;
753 		revents |= EPOLLIN | EPOLLRDNORM;
754 		break;
755 	case TIPC_DISCONNECTING:
756 		revents = EPOLLIN | EPOLLRDNORM | EPOLLHUP;
757 		break;
758 	}
759 	return revents;
760 }
761 
762 /**
763  * tipc_sendmcast - send multicast message
764  * @sock: socket structure
765  * @seq: destination address
766  * @msg: message to send
767  * @dlen: length of data to send
768  * @timeout: timeout to wait for wakeup
769  *
770  * Called from function tipc_sendmsg(), which has done all sanity checks
771  * Returns the number of bytes sent on success, or errno
772  */
773 static int tipc_sendmcast(struct  socket *sock, struct tipc_name_seq *seq,
774 			  struct msghdr *msg, size_t dlen, long timeout)
775 {
776 	struct sock *sk = sock->sk;
777 	struct tipc_sock *tsk = tipc_sk(sk);
778 	struct tipc_msg *hdr = &tsk->phdr;
779 	struct net *net = sock_net(sk);
780 	int mtu = tipc_bcast_get_mtu(net);
781 	struct tipc_mc_method *method = &tsk->mc_method;
782 	struct sk_buff_head pkts;
783 	struct tipc_nlist dsts;
784 	int rc;
785 
786 	if (tsk->group)
787 		return -EACCES;
788 
789 	/* Block or return if any destination link is congested */
790 	rc = tipc_wait_for_cond(sock, &timeout, !tsk->cong_link_cnt);
791 	if (unlikely(rc))
792 		return rc;
793 
794 	/* Lookup destination nodes */
795 	tipc_nlist_init(&dsts, tipc_own_addr(net));
796 	tipc_nametbl_lookup_dst_nodes(net, seq->type, seq->lower,
797 				      seq->upper, &dsts);
798 	if (!dsts.local && !dsts.remote)
799 		return -EHOSTUNREACH;
800 
801 	/* Build message header */
802 	msg_set_type(hdr, TIPC_MCAST_MSG);
803 	msg_set_hdr_sz(hdr, MCAST_H_SIZE);
804 	msg_set_lookup_scope(hdr, TIPC_CLUSTER_SCOPE);
805 	msg_set_destport(hdr, 0);
806 	msg_set_destnode(hdr, 0);
807 	msg_set_nametype(hdr, seq->type);
808 	msg_set_namelower(hdr, seq->lower);
809 	msg_set_nameupper(hdr, seq->upper);
810 
811 	/* Build message as chain of buffers */
812 	__skb_queue_head_init(&pkts);
813 	rc = tipc_msg_build(hdr, msg, 0, dlen, mtu, &pkts);
814 
815 	/* Send message if build was successful */
816 	if (unlikely(rc == dlen)) {
817 		trace_tipc_sk_sendmcast(sk, skb_peek(&pkts),
818 					TIPC_DUMP_SK_SNDQ, " ");
819 		rc = tipc_mcast_xmit(net, &pkts, method, &dsts,
820 				     &tsk->cong_link_cnt);
821 	}
822 
823 	tipc_nlist_purge(&dsts);
824 
825 	return rc ? rc : dlen;
826 }
827 
828 /**
829  * tipc_send_group_msg - send a message to a member in the group
830  * @net: network namespace
831  * @m: message to send
832  * @mb: group member
833  * @dnode: destination node
834  * @dport: destination port
835  * @dlen: total length of message data
836  */
837 static int tipc_send_group_msg(struct net *net, struct tipc_sock *tsk,
838 			       struct msghdr *m, struct tipc_member *mb,
839 			       u32 dnode, u32 dport, int dlen)
840 {
841 	u16 bc_snd_nxt = tipc_group_bc_snd_nxt(tsk->group);
842 	struct tipc_mc_method *method = &tsk->mc_method;
843 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
844 	struct tipc_msg *hdr = &tsk->phdr;
845 	struct sk_buff_head pkts;
846 	int mtu, rc;
847 
848 	/* Complete message header */
849 	msg_set_type(hdr, TIPC_GRP_UCAST_MSG);
850 	msg_set_hdr_sz(hdr, GROUP_H_SIZE);
851 	msg_set_destport(hdr, dport);
852 	msg_set_destnode(hdr, dnode);
853 	msg_set_grp_bc_seqno(hdr, bc_snd_nxt);
854 
855 	/* Build message as chain of buffers */
856 	__skb_queue_head_init(&pkts);
857 	mtu = tipc_node_get_mtu(net, dnode, tsk->portid);
858 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
859 	if (unlikely(rc != dlen))
860 		return rc;
861 
862 	/* Send message */
863 	rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
864 	if (unlikely(rc == -ELINKCONG)) {
865 		tipc_dest_push(&tsk->cong_links, dnode, 0);
866 		tsk->cong_link_cnt++;
867 	}
868 
869 	/* Update send window */
870 	tipc_group_update_member(mb, blks);
871 
872 	/* A broadcast sent within next EXPIRE period must follow same path */
873 	method->rcast = true;
874 	method->mandatory = true;
875 	return dlen;
876 }
877 
878 /**
879  * tipc_send_group_unicast - send message to a member in the group
880  * @sock: socket structure
881  * @m: message to send
882  * @dlen: total length of message data
883  * @timeout: timeout to wait for wakeup
884  *
885  * Called from function tipc_sendmsg(), which has done all sanity checks
886  * Returns the number of bytes sent on success, or errno
887  */
888 static int tipc_send_group_unicast(struct socket *sock, struct msghdr *m,
889 				   int dlen, long timeout)
890 {
891 	struct sock *sk = sock->sk;
892 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
893 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
894 	struct tipc_sock *tsk = tipc_sk(sk);
895 	struct net *net = sock_net(sk);
896 	struct tipc_member *mb = NULL;
897 	u32 node, port;
898 	int rc;
899 
900 	node = dest->addr.id.node;
901 	port = dest->addr.id.ref;
902 	if (!port && !node)
903 		return -EHOSTUNREACH;
904 
905 	/* Block or return if destination link or member is congested */
906 	rc = tipc_wait_for_cond(sock, &timeout,
907 				!tipc_dest_find(&tsk->cong_links, node, 0) &&
908 				tsk->group &&
909 				!tipc_group_cong(tsk->group, node, port, blks,
910 						 &mb));
911 	if (unlikely(rc))
912 		return rc;
913 
914 	if (unlikely(!mb))
915 		return -EHOSTUNREACH;
916 
917 	rc = tipc_send_group_msg(net, tsk, m, mb, node, port, dlen);
918 
919 	return rc ? rc : dlen;
920 }
921 
922 /**
923  * tipc_send_group_anycast - send message to any member with given identity
924  * @sock: socket structure
925  * @m: message to send
926  * @dlen: total length of message data
927  * @timeout: timeout to wait for wakeup
928  *
929  * Called from function tipc_sendmsg(), which has done all sanity checks
930  * Returns the number of bytes sent on success, or errno
931  */
932 static int tipc_send_group_anycast(struct socket *sock, struct msghdr *m,
933 				   int dlen, long timeout)
934 {
935 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
936 	struct sock *sk = sock->sk;
937 	struct tipc_sock *tsk = tipc_sk(sk);
938 	struct list_head *cong_links = &tsk->cong_links;
939 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
940 	struct tipc_msg *hdr = &tsk->phdr;
941 	struct tipc_member *first = NULL;
942 	struct tipc_member *mbr = NULL;
943 	struct net *net = sock_net(sk);
944 	u32 node, port, exclude;
945 	struct list_head dsts;
946 	u32 type, inst, scope;
947 	int lookups = 0;
948 	int dstcnt, rc;
949 	bool cong;
950 
951 	INIT_LIST_HEAD(&dsts);
952 
953 	type = msg_nametype(hdr);
954 	inst = dest->addr.name.name.instance;
955 	scope = msg_lookup_scope(hdr);
956 
957 	while (++lookups < 4) {
958 		exclude = tipc_group_exclude(tsk->group);
959 
960 		first = NULL;
961 
962 		/* Look for a non-congested destination member, if any */
963 		while (1) {
964 			if (!tipc_nametbl_lookup(net, type, inst, scope, &dsts,
965 						 &dstcnt, exclude, false))
966 				return -EHOSTUNREACH;
967 			tipc_dest_pop(&dsts, &node, &port);
968 			cong = tipc_group_cong(tsk->group, node, port, blks,
969 					       &mbr);
970 			if (!cong)
971 				break;
972 			if (mbr == first)
973 				break;
974 			if (!first)
975 				first = mbr;
976 		}
977 
978 		/* Start over if destination was not in member list */
979 		if (unlikely(!mbr))
980 			continue;
981 
982 		if (likely(!cong && !tipc_dest_find(cong_links, node, 0)))
983 			break;
984 
985 		/* Block or return if destination link or member is congested */
986 		rc = tipc_wait_for_cond(sock, &timeout,
987 					!tipc_dest_find(cong_links, node, 0) &&
988 					tsk->group &&
989 					!tipc_group_cong(tsk->group, node, port,
990 							 blks, &mbr));
991 		if (unlikely(rc))
992 			return rc;
993 
994 		/* Send, unless destination disappeared while waiting */
995 		if (likely(mbr))
996 			break;
997 	}
998 
999 	if (unlikely(lookups >= 4))
1000 		return -EHOSTUNREACH;
1001 
1002 	rc = tipc_send_group_msg(net, tsk, m, mbr, node, port, dlen);
1003 
1004 	return rc ? rc : dlen;
1005 }
1006 
1007 /**
1008  * tipc_send_group_bcast - send message to all members in communication group
1009  * @sk: socket structure
1010  * @m: message to send
1011  * @dlen: total length of message data
1012  * @timeout: timeout to wait for wakeup
1013  *
1014  * Called from function tipc_sendmsg(), which has done all sanity checks
1015  * Returns the number of bytes sent on success, or errno
1016  */
1017 static int tipc_send_group_bcast(struct socket *sock, struct msghdr *m,
1018 				 int dlen, long timeout)
1019 {
1020 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1021 	struct sock *sk = sock->sk;
1022 	struct net *net = sock_net(sk);
1023 	struct tipc_sock *tsk = tipc_sk(sk);
1024 	struct tipc_nlist *dsts;
1025 	struct tipc_mc_method *method = &tsk->mc_method;
1026 	bool ack = method->mandatory && method->rcast;
1027 	int blks = tsk_blocks(MCAST_H_SIZE + dlen);
1028 	struct tipc_msg *hdr = &tsk->phdr;
1029 	int mtu = tipc_bcast_get_mtu(net);
1030 	struct sk_buff_head pkts;
1031 	int rc = -EHOSTUNREACH;
1032 
1033 	/* Block or return if any destination link or member is congested */
1034 	rc = tipc_wait_for_cond(sock, &timeout,
1035 				!tsk->cong_link_cnt && tsk->group &&
1036 				!tipc_group_bc_cong(tsk->group, blks));
1037 	if (unlikely(rc))
1038 		return rc;
1039 
1040 	dsts = tipc_group_dests(tsk->group);
1041 	if (!dsts->local && !dsts->remote)
1042 		return -EHOSTUNREACH;
1043 
1044 	/* Complete message header */
1045 	if (dest) {
1046 		msg_set_type(hdr, TIPC_GRP_MCAST_MSG);
1047 		msg_set_nameinst(hdr, dest->addr.name.name.instance);
1048 	} else {
1049 		msg_set_type(hdr, TIPC_GRP_BCAST_MSG);
1050 		msg_set_nameinst(hdr, 0);
1051 	}
1052 	msg_set_hdr_sz(hdr, GROUP_H_SIZE);
1053 	msg_set_destport(hdr, 0);
1054 	msg_set_destnode(hdr, 0);
1055 	msg_set_grp_bc_seqno(hdr, tipc_group_bc_snd_nxt(tsk->group));
1056 
1057 	/* Avoid getting stuck with repeated forced replicasts */
1058 	msg_set_grp_bc_ack_req(hdr, ack);
1059 
1060 	/* Build message as chain of buffers */
1061 	__skb_queue_head_init(&pkts);
1062 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
1063 	if (unlikely(rc != dlen))
1064 		return rc;
1065 
1066 	/* Send message */
1067 	rc = tipc_mcast_xmit(net, &pkts, method, dsts, &tsk->cong_link_cnt);
1068 	if (unlikely(rc))
1069 		return rc;
1070 
1071 	/* Update broadcast sequence number and send windows */
1072 	tipc_group_update_bc_members(tsk->group, blks, ack);
1073 
1074 	/* Broadcast link is now free to choose method for next broadcast */
1075 	method->mandatory = false;
1076 	method->expires = jiffies;
1077 
1078 	return dlen;
1079 }
1080 
1081 /**
1082  * tipc_send_group_mcast - send message to all members with given identity
1083  * @sock: socket structure
1084  * @m: message to send
1085  * @dlen: total length of message data
1086  * @timeout: timeout to wait for wakeup
1087  *
1088  * Called from function tipc_sendmsg(), which has done all sanity checks
1089  * Returns the number of bytes sent on success, or errno
1090  */
1091 static int tipc_send_group_mcast(struct socket *sock, struct msghdr *m,
1092 				 int dlen, long timeout)
1093 {
1094 	struct sock *sk = sock->sk;
1095 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1096 	struct tipc_sock *tsk = tipc_sk(sk);
1097 	struct tipc_group *grp = tsk->group;
1098 	struct tipc_msg *hdr = &tsk->phdr;
1099 	struct net *net = sock_net(sk);
1100 	u32 type, inst, scope, exclude;
1101 	struct list_head dsts;
1102 	u32 dstcnt;
1103 
1104 	INIT_LIST_HEAD(&dsts);
1105 
1106 	type = msg_nametype(hdr);
1107 	inst = dest->addr.name.name.instance;
1108 	scope = msg_lookup_scope(hdr);
1109 	exclude = tipc_group_exclude(grp);
1110 
1111 	if (!tipc_nametbl_lookup(net, type, inst, scope, &dsts,
1112 				 &dstcnt, exclude, true))
1113 		return -EHOSTUNREACH;
1114 
1115 	if (dstcnt == 1) {
1116 		tipc_dest_pop(&dsts, &dest->addr.id.node, &dest->addr.id.ref);
1117 		return tipc_send_group_unicast(sock, m, dlen, timeout);
1118 	}
1119 
1120 	tipc_dest_list_purge(&dsts);
1121 	return tipc_send_group_bcast(sock, m, dlen, timeout);
1122 }
1123 
1124 /**
1125  * tipc_sk_mcast_rcv - Deliver multicast messages to all destination sockets
1126  * @arrvq: queue with arriving messages, to be cloned after destination lookup
1127  * @inputq: queue with cloned messages, delivered to socket after dest lookup
1128  *
1129  * Multi-threaded: parallel calls with reference to same queues may occur
1130  */
1131 void tipc_sk_mcast_rcv(struct net *net, struct sk_buff_head *arrvq,
1132 		       struct sk_buff_head *inputq)
1133 {
1134 	u32 self = tipc_own_addr(net);
1135 	u32 type, lower, upper, scope;
1136 	struct sk_buff *skb, *_skb;
1137 	u32 portid, onode;
1138 	struct sk_buff_head tmpq;
1139 	struct list_head dports;
1140 	struct tipc_msg *hdr;
1141 	int user, mtyp, hlen;
1142 	bool exact;
1143 
1144 	__skb_queue_head_init(&tmpq);
1145 	INIT_LIST_HEAD(&dports);
1146 
1147 	skb = tipc_skb_peek(arrvq, &inputq->lock);
1148 	for (; skb; skb = tipc_skb_peek(arrvq, &inputq->lock)) {
1149 		hdr = buf_msg(skb);
1150 		user = msg_user(hdr);
1151 		mtyp = msg_type(hdr);
1152 		hlen = skb_headroom(skb) + msg_hdr_sz(hdr);
1153 		onode = msg_orignode(hdr);
1154 		type = msg_nametype(hdr);
1155 
1156 		if (mtyp == TIPC_GRP_UCAST_MSG || user == GROUP_PROTOCOL) {
1157 			spin_lock_bh(&inputq->lock);
1158 			if (skb_peek(arrvq) == skb) {
1159 				__skb_dequeue(arrvq);
1160 				__skb_queue_tail(inputq, skb);
1161 			}
1162 			kfree_skb(skb);
1163 			spin_unlock_bh(&inputq->lock);
1164 			continue;
1165 		}
1166 
1167 		/* Group messages require exact scope match */
1168 		if (msg_in_group(hdr)) {
1169 			lower = 0;
1170 			upper = ~0;
1171 			scope = msg_lookup_scope(hdr);
1172 			exact = true;
1173 		} else {
1174 			/* TIPC_NODE_SCOPE means "any scope" in this context */
1175 			if (onode == self)
1176 				scope = TIPC_NODE_SCOPE;
1177 			else
1178 				scope = TIPC_CLUSTER_SCOPE;
1179 			exact = false;
1180 			lower = msg_namelower(hdr);
1181 			upper = msg_nameupper(hdr);
1182 		}
1183 
1184 		/* Create destination port list: */
1185 		tipc_nametbl_mc_lookup(net, type, lower, upper,
1186 				       scope, exact, &dports);
1187 
1188 		/* Clone message per destination */
1189 		while (tipc_dest_pop(&dports, NULL, &portid)) {
1190 			_skb = __pskb_copy(skb, hlen, GFP_ATOMIC);
1191 			if (_skb) {
1192 				msg_set_destport(buf_msg(_skb), portid);
1193 				__skb_queue_tail(&tmpq, _skb);
1194 				continue;
1195 			}
1196 			pr_warn("Failed to clone mcast rcv buffer\n");
1197 		}
1198 		/* Append to inputq if not already done by other thread */
1199 		spin_lock_bh(&inputq->lock);
1200 		if (skb_peek(arrvq) == skb) {
1201 			skb_queue_splice_tail_init(&tmpq, inputq);
1202 			kfree_skb(__skb_dequeue(arrvq));
1203 		}
1204 		spin_unlock_bh(&inputq->lock);
1205 		__skb_queue_purge(&tmpq);
1206 		kfree_skb(skb);
1207 	}
1208 	tipc_sk_rcv(net, inputq);
1209 }
1210 
1211 /**
1212  * tipc_sk_conn_proto_rcv - receive a connection mng protocol message
1213  * @tsk: receiving socket
1214  * @skb: pointer to message buffer.
1215  */
1216 static void tipc_sk_conn_proto_rcv(struct tipc_sock *tsk, struct sk_buff *skb,
1217 				   struct sk_buff_head *inputq,
1218 				   struct sk_buff_head *xmitq)
1219 {
1220 	struct tipc_msg *hdr = buf_msg(skb);
1221 	u32 onode = tsk_own_node(tsk);
1222 	struct sock *sk = &tsk->sk;
1223 	int mtyp = msg_type(hdr);
1224 	bool conn_cong;
1225 
1226 	/* Ignore if connection cannot be validated: */
1227 	if (!tsk_peer_msg(tsk, hdr)) {
1228 		trace_tipc_sk_drop_msg(sk, skb, TIPC_DUMP_NONE, "@proto_rcv!");
1229 		goto exit;
1230 	}
1231 
1232 	if (unlikely(msg_errcode(hdr))) {
1233 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
1234 		tipc_node_remove_conn(sock_net(sk), tsk_peer_node(tsk),
1235 				      tsk_peer_port(tsk));
1236 		sk->sk_state_change(sk);
1237 
1238 		/* State change is ignored if socket already awake,
1239 		 * - convert msg to abort msg and add to inqueue
1240 		 */
1241 		msg_set_user(hdr, TIPC_CRITICAL_IMPORTANCE);
1242 		msg_set_type(hdr, TIPC_CONN_MSG);
1243 		msg_set_size(hdr, BASIC_H_SIZE);
1244 		msg_set_hdr_sz(hdr, BASIC_H_SIZE);
1245 		__skb_queue_tail(inputq, skb);
1246 		return;
1247 	}
1248 
1249 	tsk->probe_unacked = false;
1250 
1251 	if (mtyp == CONN_PROBE) {
1252 		msg_set_type(hdr, CONN_PROBE_REPLY);
1253 		if (tipc_msg_reverse(onode, &skb, TIPC_OK))
1254 			__skb_queue_tail(xmitq, skb);
1255 		return;
1256 	} else if (mtyp == CONN_ACK) {
1257 		conn_cong = tsk_conn_cong(tsk);
1258 		tsk->snt_unacked -= msg_conn_ack(hdr);
1259 		if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL)
1260 			tsk->snd_win = msg_adv_win(hdr);
1261 		if (conn_cong)
1262 			sk->sk_write_space(sk);
1263 	} else if (mtyp != CONN_PROBE_REPLY) {
1264 		pr_warn("Received unknown CONN_PROTO msg\n");
1265 	}
1266 exit:
1267 	kfree_skb(skb);
1268 }
1269 
1270 /**
1271  * tipc_sendmsg - send message in connectionless manner
1272  * @sock: socket structure
1273  * @m: message to send
1274  * @dsz: amount of user data to be sent
1275  *
1276  * Message must have an destination specified explicitly.
1277  * Used for SOCK_RDM and SOCK_DGRAM messages,
1278  * and for 'SYN' messages on SOCK_SEQPACKET and SOCK_STREAM connections.
1279  * (Note: 'SYN+' is prohibited on SOCK_STREAM.)
1280  *
1281  * Returns the number of bytes sent on success, or errno otherwise
1282  */
1283 static int tipc_sendmsg(struct socket *sock,
1284 			struct msghdr *m, size_t dsz)
1285 {
1286 	struct sock *sk = sock->sk;
1287 	int ret;
1288 
1289 	lock_sock(sk);
1290 	ret = __tipc_sendmsg(sock, m, dsz);
1291 	release_sock(sk);
1292 
1293 	return ret;
1294 }
1295 
1296 static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dlen)
1297 {
1298 	struct sock *sk = sock->sk;
1299 	struct net *net = sock_net(sk);
1300 	struct tipc_sock *tsk = tipc_sk(sk);
1301 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1302 	long timeout = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
1303 	struct list_head *clinks = &tsk->cong_links;
1304 	bool syn = !tipc_sk_type_connectionless(sk);
1305 	struct tipc_group *grp = tsk->group;
1306 	struct tipc_msg *hdr = &tsk->phdr;
1307 	struct tipc_name_seq *seq;
1308 	struct sk_buff_head pkts;
1309 	u32 dport, dnode = 0;
1310 	u32 type, inst;
1311 	int mtu, rc;
1312 
1313 	if (unlikely(dlen > TIPC_MAX_USER_MSG_SIZE))
1314 		return -EMSGSIZE;
1315 
1316 	if (likely(dest)) {
1317 		if (unlikely(m->msg_namelen < sizeof(*dest)))
1318 			return -EINVAL;
1319 		if (unlikely(dest->family != AF_TIPC))
1320 			return -EINVAL;
1321 	}
1322 
1323 	if (grp) {
1324 		if (!dest)
1325 			return tipc_send_group_bcast(sock, m, dlen, timeout);
1326 		if (dest->addrtype == TIPC_ADDR_NAME)
1327 			return tipc_send_group_anycast(sock, m, dlen, timeout);
1328 		if (dest->addrtype == TIPC_ADDR_ID)
1329 			return tipc_send_group_unicast(sock, m, dlen, timeout);
1330 		if (dest->addrtype == TIPC_ADDR_MCAST)
1331 			return tipc_send_group_mcast(sock, m, dlen, timeout);
1332 		return -EINVAL;
1333 	}
1334 
1335 	if (unlikely(!dest)) {
1336 		dest = &tsk->peer;
1337 		if (!syn && dest->family != AF_TIPC)
1338 			return -EDESTADDRREQ;
1339 	}
1340 
1341 	if (unlikely(syn)) {
1342 		if (sk->sk_state == TIPC_LISTEN)
1343 			return -EPIPE;
1344 		if (sk->sk_state != TIPC_OPEN)
1345 			return -EISCONN;
1346 		if (tsk->published)
1347 			return -EOPNOTSUPP;
1348 		if (dest->addrtype == TIPC_ADDR_NAME) {
1349 			tsk->conn_type = dest->addr.name.name.type;
1350 			tsk->conn_instance = dest->addr.name.name.instance;
1351 		}
1352 		msg_set_syn(hdr, 1);
1353 	}
1354 
1355 	seq = &dest->addr.nameseq;
1356 	if (dest->addrtype == TIPC_ADDR_MCAST)
1357 		return tipc_sendmcast(sock, seq, m, dlen, timeout);
1358 
1359 	if (dest->addrtype == TIPC_ADDR_NAME) {
1360 		type = dest->addr.name.name.type;
1361 		inst = dest->addr.name.name.instance;
1362 		dnode = dest->addr.name.domain;
1363 		msg_set_type(hdr, TIPC_NAMED_MSG);
1364 		msg_set_hdr_sz(hdr, NAMED_H_SIZE);
1365 		msg_set_nametype(hdr, type);
1366 		msg_set_nameinst(hdr, inst);
1367 		msg_set_lookup_scope(hdr, tipc_node2scope(dnode));
1368 		dport = tipc_nametbl_translate(net, type, inst, &dnode);
1369 		msg_set_destnode(hdr, dnode);
1370 		msg_set_destport(hdr, dport);
1371 		if (unlikely(!dport && !dnode))
1372 			return -EHOSTUNREACH;
1373 	} else if (dest->addrtype == TIPC_ADDR_ID) {
1374 		dnode = dest->addr.id.node;
1375 		msg_set_type(hdr, TIPC_DIRECT_MSG);
1376 		msg_set_lookup_scope(hdr, 0);
1377 		msg_set_destnode(hdr, dnode);
1378 		msg_set_destport(hdr, dest->addr.id.ref);
1379 		msg_set_hdr_sz(hdr, BASIC_H_SIZE);
1380 	} else {
1381 		return -EINVAL;
1382 	}
1383 
1384 	/* Block or return if destination link is congested */
1385 	rc = tipc_wait_for_cond(sock, &timeout,
1386 				!tipc_dest_find(clinks, dnode, 0));
1387 	if (unlikely(rc))
1388 		return rc;
1389 
1390 	__skb_queue_head_init(&pkts);
1391 	mtu = tipc_node_get_mtu(net, dnode, tsk->portid);
1392 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
1393 	if (unlikely(rc != dlen))
1394 		return rc;
1395 	if (unlikely(syn && !tipc_msg_skb_clone(&pkts, &sk->sk_write_queue)))
1396 		return -ENOMEM;
1397 
1398 	trace_tipc_sk_sendmsg(sk, skb_peek(&pkts), TIPC_DUMP_SK_SNDQ, " ");
1399 	rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
1400 	if (unlikely(rc == -ELINKCONG)) {
1401 		tipc_dest_push(clinks, dnode, 0);
1402 		tsk->cong_link_cnt++;
1403 		rc = 0;
1404 	}
1405 
1406 	if (unlikely(syn && !rc))
1407 		tipc_set_sk_state(sk, TIPC_CONNECTING);
1408 
1409 	return rc ? rc : dlen;
1410 }
1411 
1412 /**
1413  * tipc_sendstream - send stream-oriented data
1414  * @sock: socket structure
1415  * @m: data to send
1416  * @dsz: total length of data to be transmitted
1417  *
1418  * Used for SOCK_STREAM data.
1419  *
1420  * Returns the number of bytes sent on success (or partial success),
1421  * or errno if no data sent
1422  */
1423 static int tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dsz)
1424 {
1425 	struct sock *sk = sock->sk;
1426 	int ret;
1427 
1428 	lock_sock(sk);
1429 	ret = __tipc_sendstream(sock, m, dsz);
1430 	release_sock(sk);
1431 
1432 	return ret;
1433 }
1434 
1435 static int __tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dlen)
1436 {
1437 	struct sock *sk = sock->sk;
1438 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1439 	long timeout = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
1440 	struct tipc_sock *tsk = tipc_sk(sk);
1441 	struct tipc_msg *hdr = &tsk->phdr;
1442 	struct net *net = sock_net(sk);
1443 	struct sk_buff_head pkts;
1444 	u32 dnode = tsk_peer_node(tsk);
1445 	int send, sent = 0;
1446 	int rc = 0;
1447 
1448 	__skb_queue_head_init(&pkts);
1449 
1450 	if (unlikely(dlen > INT_MAX))
1451 		return -EMSGSIZE;
1452 
1453 	/* Handle implicit connection setup */
1454 	if (unlikely(dest)) {
1455 		rc = __tipc_sendmsg(sock, m, dlen);
1456 		if (dlen && dlen == rc) {
1457 			tsk->peer_caps = tipc_node_get_capabilities(net, dnode);
1458 			tsk->snt_unacked = tsk_inc(tsk, dlen + msg_hdr_sz(hdr));
1459 		}
1460 		return rc;
1461 	}
1462 
1463 	do {
1464 		rc = tipc_wait_for_cond(sock, &timeout,
1465 					(!tsk->cong_link_cnt &&
1466 					 !tsk_conn_cong(tsk) &&
1467 					 tipc_sk_connected(sk)));
1468 		if (unlikely(rc))
1469 			break;
1470 
1471 		send = min_t(size_t, dlen - sent, TIPC_MAX_USER_MSG_SIZE);
1472 		rc = tipc_msg_build(hdr, m, sent, send, tsk->max_pkt, &pkts);
1473 		if (unlikely(rc != send))
1474 			break;
1475 
1476 		trace_tipc_sk_sendstream(sk, skb_peek(&pkts),
1477 					 TIPC_DUMP_SK_SNDQ, " ");
1478 		rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
1479 		if (unlikely(rc == -ELINKCONG)) {
1480 			tsk->cong_link_cnt = 1;
1481 			rc = 0;
1482 		}
1483 		if (likely(!rc)) {
1484 			tsk->snt_unacked += tsk_inc(tsk, send + MIN_H_SIZE);
1485 			sent += send;
1486 		}
1487 	} while (sent < dlen && !rc);
1488 
1489 	return sent ? sent : rc;
1490 }
1491 
1492 /**
1493  * tipc_send_packet - send a connection-oriented message
1494  * @sock: socket structure
1495  * @m: message to send
1496  * @dsz: length of data to be transmitted
1497  *
1498  * Used for SOCK_SEQPACKET messages.
1499  *
1500  * Returns the number of bytes sent on success, or errno otherwise
1501  */
1502 static int tipc_send_packet(struct socket *sock, struct msghdr *m, size_t dsz)
1503 {
1504 	if (dsz > TIPC_MAX_USER_MSG_SIZE)
1505 		return -EMSGSIZE;
1506 
1507 	return tipc_sendstream(sock, m, dsz);
1508 }
1509 
1510 /* tipc_sk_finish_conn - complete the setup of a connection
1511  */
1512 static void tipc_sk_finish_conn(struct tipc_sock *tsk, u32 peer_port,
1513 				u32 peer_node)
1514 {
1515 	struct sock *sk = &tsk->sk;
1516 	struct net *net = sock_net(sk);
1517 	struct tipc_msg *msg = &tsk->phdr;
1518 
1519 	msg_set_syn(msg, 0);
1520 	msg_set_destnode(msg, peer_node);
1521 	msg_set_destport(msg, peer_port);
1522 	msg_set_type(msg, TIPC_CONN_MSG);
1523 	msg_set_lookup_scope(msg, 0);
1524 	msg_set_hdr_sz(msg, SHORT_H_SIZE);
1525 
1526 	sk_reset_timer(sk, &sk->sk_timer, jiffies + CONN_PROBING_INTV);
1527 	tipc_set_sk_state(sk, TIPC_ESTABLISHED);
1528 	tipc_node_add_conn(net, peer_node, tsk->portid, peer_port);
1529 	tsk->max_pkt = tipc_node_get_mtu(net, peer_node, tsk->portid);
1530 	tsk->peer_caps = tipc_node_get_capabilities(net, peer_node);
1531 	__skb_queue_purge(&sk->sk_write_queue);
1532 	if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL)
1533 		return;
1534 
1535 	/* Fall back to message based flow control */
1536 	tsk->rcv_win = FLOWCTL_MSG_WIN;
1537 	tsk->snd_win = FLOWCTL_MSG_WIN;
1538 }
1539 
1540 /**
1541  * tipc_sk_set_orig_addr - capture sender's address for received message
1542  * @m: descriptor for message info
1543  * @hdr: received message header
1544  *
1545  * Note: Address is not captured if not requested by receiver.
1546  */
1547 static void tipc_sk_set_orig_addr(struct msghdr *m, struct sk_buff *skb)
1548 {
1549 	DECLARE_SOCKADDR(struct sockaddr_pair *, srcaddr, m->msg_name);
1550 	struct tipc_msg *hdr = buf_msg(skb);
1551 
1552 	if (!srcaddr)
1553 		return;
1554 
1555 	srcaddr->sock.family = AF_TIPC;
1556 	srcaddr->sock.addrtype = TIPC_ADDR_ID;
1557 	srcaddr->sock.scope = 0;
1558 	srcaddr->sock.addr.id.ref = msg_origport(hdr);
1559 	srcaddr->sock.addr.id.node = msg_orignode(hdr);
1560 	srcaddr->sock.addr.name.domain = 0;
1561 	m->msg_namelen = sizeof(struct sockaddr_tipc);
1562 
1563 	if (!msg_in_group(hdr))
1564 		return;
1565 
1566 	/* Group message users may also want to know sending member's id */
1567 	srcaddr->member.family = AF_TIPC;
1568 	srcaddr->member.addrtype = TIPC_ADDR_NAME;
1569 	srcaddr->member.scope = 0;
1570 	srcaddr->member.addr.name.name.type = msg_nametype(hdr);
1571 	srcaddr->member.addr.name.name.instance = TIPC_SKB_CB(skb)->orig_member;
1572 	srcaddr->member.addr.name.domain = 0;
1573 	m->msg_namelen = sizeof(*srcaddr);
1574 }
1575 
1576 /**
1577  * tipc_sk_anc_data_recv - optionally capture ancillary data for received message
1578  * @m: descriptor for message info
1579  * @skb: received message buffer
1580  * @tsk: TIPC port associated with message
1581  *
1582  * Note: Ancillary data is not captured if not requested by receiver.
1583  *
1584  * Returns 0 if successful, otherwise errno
1585  */
1586 static int tipc_sk_anc_data_recv(struct msghdr *m, struct sk_buff *skb,
1587 				 struct tipc_sock *tsk)
1588 {
1589 	struct tipc_msg *msg;
1590 	u32 anc_data[3];
1591 	u32 err;
1592 	u32 dest_type;
1593 	int has_name;
1594 	int res;
1595 
1596 	if (likely(m->msg_controllen == 0))
1597 		return 0;
1598 	msg = buf_msg(skb);
1599 
1600 	/* Optionally capture errored message object(s) */
1601 	err = msg ? msg_errcode(msg) : 0;
1602 	if (unlikely(err)) {
1603 		anc_data[0] = err;
1604 		anc_data[1] = msg_data_sz(msg);
1605 		res = put_cmsg(m, SOL_TIPC, TIPC_ERRINFO, 8, anc_data);
1606 		if (res)
1607 			return res;
1608 		if (anc_data[1]) {
1609 			if (skb_linearize(skb))
1610 				return -ENOMEM;
1611 			msg = buf_msg(skb);
1612 			res = put_cmsg(m, SOL_TIPC, TIPC_RETDATA, anc_data[1],
1613 				       msg_data(msg));
1614 			if (res)
1615 				return res;
1616 		}
1617 	}
1618 
1619 	/* Optionally capture message destination object */
1620 	dest_type = msg ? msg_type(msg) : TIPC_DIRECT_MSG;
1621 	switch (dest_type) {
1622 	case TIPC_NAMED_MSG:
1623 		has_name = 1;
1624 		anc_data[0] = msg_nametype(msg);
1625 		anc_data[1] = msg_namelower(msg);
1626 		anc_data[2] = msg_namelower(msg);
1627 		break;
1628 	case TIPC_MCAST_MSG:
1629 		has_name = 1;
1630 		anc_data[0] = msg_nametype(msg);
1631 		anc_data[1] = msg_namelower(msg);
1632 		anc_data[2] = msg_nameupper(msg);
1633 		break;
1634 	case TIPC_CONN_MSG:
1635 		has_name = (tsk->conn_type != 0);
1636 		anc_data[0] = tsk->conn_type;
1637 		anc_data[1] = tsk->conn_instance;
1638 		anc_data[2] = tsk->conn_instance;
1639 		break;
1640 	default:
1641 		has_name = 0;
1642 	}
1643 	if (has_name) {
1644 		res = put_cmsg(m, SOL_TIPC, TIPC_DESTNAME, 12, anc_data);
1645 		if (res)
1646 			return res;
1647 	}
1648 
1649 	return 0;
1650 }
1651 
1652 static void tipc_sk_send_ack(struct tipc_sock *tsk)
1653 {
1654 	struct sock *sk = &tsk->sk;
1655 	struct net *net = sock_net(sk);
1656 	struct sk_buff *skb = NULL;
1657 	struct tipc_msg *msg;
1658 	u32 peer_port = tsk_peer_port(tsk);
1659 	u32 dnode = tsk_peer_node(tsk);
1660 
1661 	if (!tipc_sk_connected(sk))
1662 		return;
1663 	skb = tipc_msg_create(CONN_MANAGER, CONN_ACK, INT_H_SIZE, 0,
1664 			      dnode, tsk_own_node(tsk), peer_port,
1665 			      tsk->portid, TIPC_OK);
1666 	if (!skb)
1667 		return;
1668 	msg = buf_msg(skb);
1669 	msg_set_conn_ack(msg, tsk->rcv_unacked);
1670 	tsk->rcv_unacked = 0;
1671 
1672 	/* Adjust to and advertize the correct window limit */
1673 	if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL) {
1674 		tsk->rcv_win = tsk_adv_blocks(tsk->sk.sk_rcvbuf);
1675 		msg_set_adv_win(msg, tsk->rcv_win);
1676 	}
1677 	tipc_node_xmit_skb(net, skb, dnode, msg_link_selector(msg));
1678 }
1679 
1680 static int tipc_wait_for_rcvmsg(struct socket *sock, long *timeop)
1681 {
1682 	struct sock *sk = sock->sk;
1683 	DEFINE_WAIT_FUNC(wait, woken_wake_function);
1684 	long timeo = *timeop;
1685 	int err = sock_error(sk);
1686 
1687 	if (err)
1688 		return err;
1689 
1690 	for (;;) {
1691 		if (timeo && skb_queue_empty(&sk->sk_receive_queue)) {
1692 			if (sk->sk_shutdown & RCV_SHUTDOWN) {
1693 				err = -ENOTCONN;
1694 				break;
1695 			}
1696 			add_wait_queue(sk_sleep(sk), &wait);
1697 			release_sock(sk);
1698 			timeo = wait_woken(&wait, TASK_INTERRUPTIBLE, timeo);
1699 			sched_annotate_sleep();
1700 			lock_sock(sk);
1701 			remove_wait_queue(sk_sleep(sk), &wait);
1702 		}
1703 		err = 0;
1704 		if (!skb_queue_empty(&sk->sk_receive_queue))
1705 			break;
1706 		err = -EAGAIN;
1707 		if (!timeo)
1708 			break;
1709 		err = sock_intr_errno(timeo);
1710 		if (signal_pending(current))
1711 			break;
1712 
1713 		err = sock_error(sk);
1714 		if (err)
1715 			break;
1716 	}
1717 	*timeop = timeo;
1718 	return err;
1719 }
1720 
1721 /**
1722  * tipc_recvmsg - receive packet-oriented message
1723  * @m: descriptor for message info
1724  * @buflen: length of user buffer area
1725  * @flags: receive flags
1726  *
1727  * Used for SOCK_DGRAM, SOCK_RDM, and SOCK_SEQPACKET messages.
1728  * If the complete message doesn't fit in user area, truncate it.
1729  *
1730  * Returns size of returned message data, errno otherwise
1731  */
1732 static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
1733 			size_t buflen,	int flags)
1734 {
1735 	struct sock *sk = sock->sk;
1736 	bool connected = !tipc_sk_type_connectionless(sk);
1737 	struct tipc_sock *tsk = tipc_sk(sk);
1738 	int rc, err, hlen, dlen, copy;
1739 	struct sk_buff_head xmitq;
1740 	struct tipc_msg *hdr;
1741 	struct sk_buff *skb;
1742 	bool grp_evt;
1743 	long timeout;
1744 
1745 	/* Catch invalid receive requests */
1746 	if (unlikely(!buflen))
1747 		return -EINVAL;
1748 
1749 	lock_sock(sk);
1750 	if (unlikely(connected && sk->sk_state == TIPC_OPEN)) {
1751 		rc = -ENOTCONN;
1752 		goto exit;
1753 	}
1754 	timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
1755 
1756 	/* Step rcv queue to first msg with data or error; wait if necessary */
1757 	do {
1758 		rc = tipc_wait_for_rcvmsg(sock, &timeout);
1759 		if (unlikely(rc))
1760 			goto exit;
1761 		skb = skb_peek(&sk->sk_receive_queue);
1762 		hdr = buf_msg(skb);
1763 		dlen = msg_data_sz(hdr);
1764 		hlen = msg_hdr_sz(hdr);
1765 		err = msg_errcode(hdr);
1766 		grp_evt = msg_is_grp_evt(hdr);
1767 		if (likely(dlen || err))
1768 			break;
1769 		tsk_advance_rx_queue(sk);
1770 	} while (1);
1771 
1772 	/* Collect msg meta data, including error code and rejected data */
1773 	tipc_sk_set_orig_addr(m, skb);
1774 	rc = tipc_sk_anc_data_recv(m, skb, tsk);
1775 	if (unlikely(rc))
1776 		goto exit;
1777 	hdr = buf_msg(skb);
1778 
1779 	/* Capture data if non-error msg, otherwise just set return value */
1780 	if (likely(!err)) {
1781 		copy = min_t(int, dlen, buflen);
1782 		if (unlikely(copy != dlen))
1783 			m->msg_flags |= MSG_TRUNC;
1784 		rc = skb_copy_datagram_msg(skb, hlen, m, copy);
1785 	} else {
1786 		copy = 0;
1787 		rc = 0;
1788 		if (err != TIPC_CONN_SHUTDOWN && connected && !m->msg_control)
1789 			rc = -ECONNRESET;
1790 	}
1791 	if (unlikely(rc))
1792 		goto exit;
1793 
1794 	/* Mark message as group event if applicable */
1795 	if (unlikely(grp_evt)) {
1796 		if (msg_grp_evt(hdr) == TIPC_WITHDRAWN)
1797 			m->msg_flags |= MSG_EOR;
1798 		m->msg_flags |= MSG_OOB;
1799 		copy = 0;
1800 	}
1801 
1802 	/* Caption of data or error code/rejected data was successful */
1803 	if (unlikely(flags & MSG_PEEK))
1804 		goto exit;
1805 
1806 	/* Send group flow control advertisement when applicable */
1807 	if (tsk->group && msg_in_group(hdr) && !grp_evt) {
1808 		__skb_queue_head_init(&xmitq);
1809 		tipc_group_update_rcv_win(tsk->group, tsk_blocks(hlen + dlen),
1810 					  msg_orignode(hdr), msg_origport(hdr),
1811 					  &xmitq);
1812 		tipc_node_distr_xmit(sock_net(sk), &xmitq);
1813 	}
1814 
1815 	tsk_advance_rx_queue(sk);
1816 
1817 	if (likely(!connected))
1818 		goto exit;
1819 
1820 	/* Send connection flow control advertisement when applicable */
1821 	tsk->rcv_unacked += tsk_inc(tsk, hlen + dlen);
1822 	if (tsk->rcv_unacked >= tsk->rcv_win / TIPC_ACK_RATE)
1823 		tipc_sk_send_ack(tsk);
1824 exit:
1825 	release_sock(sk);
1826 	return rc ? rc : copy;
1827 }
1828 
1829 /**
1830  * tipc_recvstream - receive stream-oriented data
1831  * @m: descriptor for message info
1832  * @buflen: total size of user buffer area
1833  * @flags: receive flags
1834  *
1835  * Used for SOCK_STREAM messages only.  If not enough data is available
1836  * will optionally wait for more; never truncates data.
1837  *
1838  * Returns size of returned message data, errno otherwise
1839  */
1840 static int tipc_recvstream(struct socket *sock, struct msghdr *m,
1841 			   size_t buflen, int flags)
1842 {
1843 	struct sock *sk = sock->sk;
1844 	struct tipc_sock *tsk = tipc_sk(sk);
1845 	struct sk_buff *skb;
1846 	struct tipc_msg *hdr;
1847 	struct tipc_skb_cb *skb_cb;
1848 	bool peek = flags & MSG_PEEK;
1849 	int offset, required, copy, copied = 0;
1850 	int hlen, dlen, err, rc;
1851 	long timeout;
1852 
1853 	/* Catch invalid receive attempts */
1854 	if (unlikely(!buflen))
1855 		return -EINVAL;
1856 
1857 	lock_sock(sk);
1858 
1859 	if (unlikely(sk->sk_state == TIPC_OPEN)) {
1860 		rc = -ENOTCONN;
1861 		goto exit;
1862 	}
1863 	required = sock_rcvlowat(sk, flags & MSG_WAITALL, buflen);
1864 	timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
1865 
1866 	do {
1867 		/* Look at first msg in receive queue; wait if necessary */
1868 		rc = tipc_wait_for_rcvmsg(sock, &timeout);
1869 		if (unlikely(rc))
1870 			break;
1871 		skb = skb_peek(&sk->sk_receive_queue);
1872 		skb_cb = TIPC_SKB_CB(skb);
1873 		hdr = buf_msg(skb);
1874 		dlen = msg_data_sz(hdr);
1875 		hlen = msg_hdr_sz(hdr);
1876 		err = msg_errcode(hdr);
1877 
1878 		/* Discard any empty non-errored (SYN-) message */
1879 		if (unlikely(!dlen && !err)) {
1880 			tsk_advance_rx_queue(sk);
1881 			continue;
1882 		}
1883 
1884 		/* Collect msg meta data, incl. error code and rejected data */
1885 		if (!copied) {
1886 			tipc_sk_set_orig_addr(m, skb);
1887 			rc = tipc_sk_anc_data_recv(m, skb, tsk);
1888 			if (rc)
1889 				break;
1890 			hdr = buf_msg(skb);
1891 		}
1892 
1893 		/* Copy data if msg ok, otherwise return error/partial data */
1894 		if (likely(!err)) {
1895 			offset = skb_cb->bytes_read;
1896 			copy = min_t(int, dlen - offset, buflen - copied);
1897 			rc = skb_copy_datagram_msg(skb, hlen + offset, m, copy);
1898 			if (unlikely(rc))
1899 				break;
1900 			copied += copy;
1901 			offset += copy;
1902 			if (unlikely(offset < dlen)) {
1903 				if (!peek)
1904 					skb_cb->bytes_read = offset;
1905 				break;
1906 			}
1907 		} else {
1908 			rc = 0;
1909 			if ((err != TIPC_CONN_SHUTDOWN) && !m->msg_control)
1910 				rc = -ECONNRESET;
1911 			if (copied || rc)
1912 				break;
1913 		}
1914 
1915 		if (unlikely(peek))
1916 			break;
1917 
1918 		tsk_advance_rx_queue(sk);
1919 
1920 		/* Send connection flow control advertisement when applicable */
1921 		tsk->rcv_unacked += tsk_inc(tsk, hlen + dlen);
1922 		if (unlikely(tsk->rcv_unacked >= tsk->rcv_win / TIPC_ACK_RATE))
1923 			tipc_sk_send_ack(tsk);
1924 
1925 		/* Exit if all requested data or FIN/error received */
1926 		if (copied == buflen || err)
1927 			break;
1928 
1929 	} while (!skb_queue_empty(&sk->sk_receive_queue) || copied < required);
1930 exit:
1931 	release_sock(sk);
1932 	return copied ? copied : rc;
1933 }
1934 
1935 /**
1936  * tipc_write_space - wake up thread if port congestion is released
1937  * @sk: socket
1938  */
1939 static void tipc_write_space(struct sock *sk)
1940 {
1941 	struct socket_wq *wq;
1942 
1943 	rcu_read_lock();
1944 	wq = rcu_dereference(sk->sk_wq);
1945 	if (skwq_has_sleeper(wq))
1946 		wake_up_interruptible_sync_poll(&wq->wait, EPOLLOUT |
1947 						EPOLLWRNORM | EPOLLWRBAND);
1948 	rcu_read_unlock();
1949 }
1950 
1951 /**
1952  * tipc_data_ready - wake up threads to indicate messages have been received
1953  * @sk: socket
1954  * @len: the length of messages
1955  */
1956 static void tipc_data_ready(struct sock *sk)
1957 {
1958 	struct socket_wq *wq;
1959 
1960 	rcu_read_lock();
1961 	wq = rcu_dereference(sk->sk_wq);
1962 	if (skwq_has_sleeper(wq))
1963 		wake_up_interruptible_sync_poll(&wq->wait, EPOLLIN |
1964 						EPOLLRDNORM | EPOLLRDBAND);
1965 	rcu_read_unlock();
1966 }
1967 
1968 static void tipc_sock_destruct(struct sock *sk)
1969 {
1970 	__skb_queue_purge(&sk->sk_receive_queue);
1971 }
1972 
1973 static void tipc_sk_proto_rcv(struct sock *sk,
1974 			      struct sk_buff_head *inputq,
1975 			      struct sk_buff_head *xmitq)
1976 {
1977 	struct sk_buff *skb = __skb_dequeue(inputq);
1978 	struct tipc_sock *tsk = tipc_sk(sk);
1979 	struct tipc_msg *hdr = buf_msg(skb);
1980 	struct tipc_group *grp = tsk->group;
1981 	bool wakeup = false;
1982 
1983 	switch (msg_user(hdr)) {
1984 	case CONN_MANAGER:
1985 		tipc_sk_conn_proto_rcv(tsk, skb, inputq, xmitq);
1986 		return;
1987 	case SOCK_WAKEUP:
1988 		tipc_dest_del(&tsk->cong_links, msg_orignode(hdr), 0);
1989 		/* coupled with smp_rmb() in tipc_wait_for_cond() */
1990 		smp_wmb();
1991 		tsk->cong_link_cnt--;
1992 		wakeup = true;
1993 		break;
1994 	case GROUP_PROTOCOL:
1995 		tipc_group_proto_rcv(grp, &wakeup, hdr, inputq, xmitq);
1996 		break;
1997 	case TOP_SRV:
1998 		tipc_group_member_evt(tsk->group, &wakeup, &sk->sk_rcvbuf,
1999 				      hdr, inputq, xmitq);
2000 		break;
2001 	default:
2002 		break;
2003 	}
2004 
2005 	if (wakeup)
2006 		sk->sk_write_space(sk);
2007 
2008 	kfree_skb(skb);
2009 }
2010 
2011 /**
2012  * tipc_sk_filter_connect - check incoming message for a connection-based socket
2013  * @tsk: TIPC socket
2014  * @skb: pointer to message buffer.
2015  * Returns true if message should be added to receive queue, false otherwise
2016  */
2017 static bool tipc_sk_filter_connect(struct tipc_sock *tsk, struct sk_buff *skb)
2018 {
2019 	struct sock *sk = &tsk->sk;
2020 	struct net *net = sock_net(sk);
2021 	struct tipc_msg *hdr = buf_msg(skb);
2022 	bool con_msg = msg_connected(hdr);
2023 	u32 pport = tsk_peer_port(tsk);
2024 	u32 pnode = tsk_peer_node(tsk);
2025 	u32 oport = msg_origport(hdr);
2026 	u32 onode = msg_orignode(hdr);
2027 	int err = msg_errcode(hdr);
2028 	unsigned long delay;
2029 
2030 	if (unlikely(msg_mcast(hdr)))
2031 		return false;
2032 
2033 	switch (sk->sk_state) {
2034 	case TIPC_CONNECTING:
2035 		/* Setup ACK */
2036 		if (likely(con_msg)) {
2037 			if (err)
2038 				break;
2039 			tipc_sk_finish_conn(tsk, oport, onode);
2040 			msg_set_importance(&tsk->phdr, msg_importance(hdr));
2041 			/* ACK+ message with data is added to receive queue */
2042 			if (msg_data_sz(hdr))
2043 				return true;
2044 			/* Empty ACK-, - wake up sleeping connect() and drop */
2045 			sk->sk_state_change(sk);
2046 			msg_set_dest_droppable(hdr, 1);
2047 			return false;
2048 		}
2049 		/* Ignore connectionless message if not from listening socket */
2050 		if (oport != pport || onode != pnode)
2051 			return false;
2052 
2053 		/* Rejected SYN */
2054 		if (err != TIPC_ERR_OVERLOAD)
2055 			break;
2056 
2057 		/* Prepare for new setup attempt if we have a SYN clone */
2058 		if (skb_queue_empty(&sk->sk_write_queue))
2059 			break;
2060 		get_random_bytes(&delay, 2);
2061 		delay %= (tsk->conn_timeout / 4);
2062 		delay = msecs_to_jiffies(delay + 100);
2063 		sk_reset_timer(sk, &sk->sk_timer, jiffies + delay);
2064 		return false;
2065 	case TIPC_OPEN:
2066 	case TIPC_DISCONNECTING:
2067 		return false;
2068 	case TIPC_LISTEN:
2069 		/* Accept only SYN message */
2070 		if (!msg_is_syn(hdr) &&
2071 		    tipc_node_get_capabilities(net, onode) & TIPC_SYN_BIT)
2072 			return false;
2073 		if (!con_msg && !err)
2074 			return true;
2075 		return false;
2076 	case TIPC_ESTABLISHED:
2077 		/* Accept only connection-based messages sent by peer */
2078 		if (likely(con_msg && !err && pport == oport && pnode == onode))
2079 			return true;
2080 		if (!tsk_peer_msg(tsk, hdr))
2081 			return false;
2082 		if (!err)
2083 			return true;
2084 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2085 		tipc_node_remove_conn(net, pnode, tsk->portid);
2086 		sk->sk_state_change(sk);
2087 		return true;
2088 	default:
2089 		pr_err("Unknown sk_state %u\n", sk->sk_state);
2090 	}
2091 	/* Abort connection setup attempt */
2092 	tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2093 	sk->sk_err = ECONNREFUSED;
2094 	sk->sk_state_change(sk);
2095 	return true;
2096 }
2097 
2098 /**
2099  * rcvbuf_limit - get proper overload limit of socket receive queue
2100  * @sk: socket
2101  * @skb: message
2102  *
2103  * For connection oriented messages, irrespective of importance,
2104  * default queue limit is 2 MB.
2105  *
2106  * For connectionless messages, queue limits are based on message
2107  * importance as follows:
2108  *
2109  * TIPC_LOW_IMPORTANCE       (2 MB)
2110  * TIPC_MEDIUM_IMPORTANCE    (4 MB)
2111  * TIPC_HIGH_IMPORTANCE      (8 MB)
2112  * TIPC_CRITICAL_IMPORTANCE  (16 MB)
2113  *
2114  * Returns overload limit according to corresponding message importance
2115  */
2116 static unsigned int rcvbuf_limit(struct sock *sk, struct sk_buff *skb)
2117 {
2118 	struct tipc_sock *tsk = tipc_sk(sk);
2119 	struct tipc_msg *hdr = buf_msg(skb);
2120 
2121 	if (unlikely(msg_in_group(hdr)))
2122 		return sk->sk_rcvbuf;
2123 
2124 	if (unlikely(!msg_connected(hdr)))
2125 		return sk->sk_rcvbuf << msg_importance(hdr);
2126 
2127 	if (likely(tsk->peer_caps & TIPC_BLOCK_FLOWCTL))
2128 		return sk->sk_rcvbuf;
2129 
2130 	return FLOWCTL_MSG_LIM;
2131 }
2132 
2133 /**
2134  * tipc_sk_filter_rcv - validate incoming message
2135  * @sk: socket
2136  * @skb: pointer to message.
2137  *
2138  * Enqueues message on receive queue if acceptable; optionally handles
2139  * disconnect indication for a connected socket.
2140  *
2141  * Called with socket lock already taken
2142  *
2143  */
2144 static void tipc_sk_filter_rcv(struct sock *sk, struct sk_buff *skb,
2145 			       struct sk_buff_head *xmitq)
2146 {
2147 	bool sk_conn = !tipc_sk_type_connectionless(sk);
2148 	struct tipc_sock *tsk = tipc_sk(sk);
2149 	struct tipc_group *grp = tsk->group;
2150 	struct tipc_msg *hdr = buf_msg(skb);
2151 	struct net *net = sock_net(sk);
2152 	struct sk_buff_head inputq;
2153 	int mtyp = msg_type(hdr);
2154 	int limit, err = TIPC_OK;
2155 
2156 	trace_tipc_sk_filter_rcv(sk, skb, TIPC_DUMP_ALL, " ");
2157 	TIPC_SKB_CB(skb)->bytes_read = 0;
2158 	__skb_queue_head_init(&inputq);
2159 	__skb_queue_tail(&inputq, skb);
2160 
2161 	if (unlikely(!msg_isdata(hdr)))
2162 		tipc_sk_proto_rcv(sk, &inputq, xmitq);
2163 
2164 	if (unlikely(grp))
2165 		tipc_group_filter_msg(grp, &inputq, xmitq);
2166 
2167 	if (unlikely(!grp) && mtyp == TIPC_MCAST_MSG)
2168 		tipc_mcast_filter_msg(net, &tsk->mc_method.deferredq, &inputq);
2169 
2170 	/* Validate and add to receive buffer if there is space */
2171 	while ((skb = __skb_dequeue(&inputq))) {
2172 		hdr = buf_msg(skb);
2173 		limit = rcvbuf_limit(sk, skb);
2174 		if ((sk_conn && !tipc_sk_filter_connect(tsk, skb)) ||
2175 		    (!sk_conn && msg_connected(hdr)) ||
2176 		    (!grp && msg_in_group(hdr)))
2177 			err = TIPC_ERR_NO_PORT;
2178 		else if (sk_rmem_alloc_get(sk) + skb->truesize >= limit) {
2179 			trace_tipc_sk_dump(sk, skb, TIPC_DUMP_ALL,
2180 					   "err_overload2!");
2181 			atomic_inc(&sk->sk_drops);
2182 			err = TIPC_ERR_OVERLOAD;
2183 		}
2184 
2185 		if (unlikely(err)) {
2186 			if (tipc_msg_reverse(tipc_own_addr(net), &skb, err)) {
2187 				trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_NONE,
2188 						      "@filter_rcv!");
2189 				__skb_queue_tail(xmitq, skb);
2190 			}
2191 			err = TIPC_OK;
2192 			continue;
2193 		}
2194 		__skb_queue_tail(&sk->sk_receive_queue, skb);
2195 		skb_set_owner_r(skb, sk);
2196 		trace_tipc_sk_overlimit2(sk, skb, TIPC_DUMP_ALL,
2197 					 "rcvq >90% allocated!");
2198 		sk->sk_data_ready(sk);
2199 	}
2200 }
2201 
2202 /**
2203  * tipc_sk_backlog_rcv - handle incoming message from backlog queue
2204  * @sk: socket
2205  * @skb: message
2206  *
2207  * Caller must hold socket lock
2208  */
2209 static int tipc_sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
2210 {
2211 	unsigned int before = sk_rmem_alloc_get(sk);
2212 	struct sk_buff_head xmitq;
2213 	unsigned int added;
2214 
2215 	__skb_queue_head_init(&xmitq);
2216 
2217 	tipc_sk_filter_rcv(sk, skb, &xmitq);
2218 	added = sk_rmem_alloc_get(sk) - before;
2219 	atomic_add(added, &tipc_sk(sk)->dupl_rcvcnt);
2220 
2221 	/* Send pending response/rejected messages, if any */
2222 	tipc_node_distr_xmit(sock_net(sk), &xmitq);
2223 	return 0;
2224 }
2225 
2226 /**
2227  * tipc_sk_enqueue - extract all buffers with destination 'dport' from
2228  *                   inputq and try adding them to socket or backlog queue
2229  * @inputq: list of incoming buffers with potentially different destinations
2230  * @sk: socket where the buffers should be enqueued
2231  * @dport: port number for the socket
2232  *
2233  * Caller must hold socket lock
2234  */
2235 static void tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk,
2236 			    u32 dport, struct sk_buff_head *xmitq)
2237 {
2238 	unsigned long time_limit = jiffies + 2;
2239 	struct sk_buff *skb;
2240 	unsigned int lim;
2241 	atomic_t *dcnt;
2242 	u32 onode;
2243 
2244 	while (skb_queue_len(inputq)) {
2245 		if (unlikely(time_after_eq(jiffies, time_limit)))
2246 			return;
2247 
2248 		skb = tipc_skb_dequeue(inputq, dport);
2249 		if (unlikely(!skb))
2250 			return;
2251 
2252 		/* Add message directly to receive queue if possible */
2253 		if (!sock_owned_by_user(sk)) {
2254 			tipc_sk_filter_rcv(sk, skb, xmitq);
2255 			continue;
2256 		}
2257 
2258 		/* Try backlog, compensating for double-counted bytes */
2259 		dcnt = &tipc_sk(sk)->dupl_rcvcnt;
2260 		if (!sk->sk_backlog.len)
2261 			atomic_set(dcnt, 0);
2262 		lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt);
2263 		if (likely(!sk_add_backlog(sk, skb, lim))) {
2264 			trace_tipc_sk_overlimit1(sk, skb, TIPC_DUMP_ALL,
2265 						 "bklg & rcvq >90% allocated!");
2266 			continue;
2267 		}
2268 
2269 		trace_tipc_sk_dump(sk, skb, TIPC_DUMP_ALL, "err_overload!");
2270 		/* Overload => reject message back to sender */
2271 		onode = tipc_own_addr(sock_net(sk));
2272 		atomic_inc(&sk->sk_drops);
2273 		if (tipc_msg_reverse(onode, &skb, TIPC_ERR_OVERLOAD)) {
2274 			trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_ALL,
2275 					      "@sk_enqueue!");
2276 			__skb_queue_tail(xmitq, skb);
2277 		}
2278 		break;
2279 	}
2280 }
2281 
2282 /**
2283  * tipc_sk_rcv - handle a chain of incoming buffers
2284  * @inputq: buffer list containing the buffers
2285  * Consumes all buffers in list until inputq is empty
2286  * Note: may be called in multiple threads referring to the same queue
2287  */
2288 void tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq)
2289 {
2290 	struct sk_buff_head xmitq;
2291 	u32 dnode, dport = 0;
2292 	int err;
2293 	struct tipc_sock *tsk;
2294 	struct sock *sk;
2295 	struct sk_buff *skb;
2296 
2297 	__skb_queue_head_init(&xmitq);
2298 	while (skb_queue_len(inputq)) {
2299 		dport = tipc_skb_peek_port(inputq, dport);
2300 		tsk = tipc_sk_lookup(net, dport);
2301 
2302 		if (likely(tsk)) {
2303 			sk = &tsk->sk;
2304 			if (likely(spin_trylock_bh(&sk->sk_lock.slock))) {
2305 				tipc_sk_enqueue(inputq, sk, dport, &xmitq);
2306 				spin_unlock_bh(&sk->sk_lock.slock);
2307 			}
2308 			/* Send pending response/rejected messages, if any */
2309 			tipc_node_distr_xmit(sock_net(sk), &xmitq);
2310 			sock_put(sk);
2311 			continue;
2312 		}
2313 		/* No destination socket => dequeue skb if still there */
2314 		skb = tipc_skb_dequeue(inputq, dport);
2315 		if (!skb)
2316 			return;
2317 
2318 		/* Try secondary lookup if unresolved named message */
2319 		err = TIPC_ERR_NO_PORT;
2320 		if (tipc_msg_lookup_dest(net, skb, &err))
2321 			goto xmit;
2322 
2323 		/* Prepare for message rejection */
2324 		if (!tipc_msg_reverse(tipc_own_addr(net), &skb, err))
2325 			continue;
2326 
2327 		trace_tipc_sk_rej_msg(NULL, skb, TIPC_DUMP_NONE, "@sk_rcv!");
2328 xmit:
2329 		dnode = msg_destnode(buf_msg(skb));
2330 		tipc_node_xmit_skb(net, skb, dnode, dport);
2331 	}
2332 }
2333 
2334 static int tipc_wait_for_connect(struct socket *sock, long *timeo_p)
2335 {
2336 	DEFINE_WAIT_FUNC(wait, woken_wake_function);
2337 	struct sock *sk = sock->sk;
2338 	int done;
2339 
2340 	do {
2341 		int err = sock_error(sk);
2342 		if (err)
2343 			return err;
2344 		if (!*timeo_p)
2345 			return -ETIMEDOUT;
2346 		if (signal_pending(current))
2347 			return sock_intr_errno(*timeo_p);
2348 
2349 		add_wait_queue(sk_sleep(sk), &wait);
2350 		done = sk_wait_event(sk, timeo_p,
2351 				     sk->sk_state != TIPC_CONNECTING, &wait);
2352 		remove_wait_queue(sk_sleep(sk), &wait);
2353 	} while (!done);
2354 	return 0;
2355 }
2356 
2357 static bool tipc_sockaddr_is_sane(struct sockaddr_tipc *addr)
2358 {
2359 	if (addr->family != AF_TIPC)
2360 		return false;
2361 	if (addr->addrtype == TIPC_SERVICE_RANGE)
2362 		return (addr->addr.nameseq.lower <= addr->addr.nameseq.upper);
2363 	return (addr->addrtype == TIPC_SERVICE_ADDR ||
2364 		addr->addrtype == TIPC_SOCKET_ADDR);
2365 }
2366 
2367 /**
2368  * tipc_connect - establish a connection to another TIPC port
2369  * @sock: socket structure
2370  * @dest: socket address for destination port
2371  * @destlen: size of socket address data structure
2372  * @flags: file-related flags associated with socket
2373  *
2374  * Returns 0 on success, errno otherwise
2375  */
2376 static int tipc_connect(struct socket *sock, struct sockaddr *dest,
2377 			int destlen, int flags)
2378 {
2379 	struct sock *sk = sock->sk;
2380 	struct tipc_sock *tsk = tipc_sk(sk);
2381 	struct sockaddr_tipc *dst = (struct sockaddr_tipc *)dest;
2382 	struct msghdr m = {NULL,};
2383 	long timeout = (flags & O_NONBLOCK) ? 0 : tsk->conn_timeout;
2384 	int previous;
2385 	int res = 0;
2386 
2387 	if (destlen != sizeof(struct sockaddr_tipc))
2388 		return -EINVAL;
2389 
2390 	lock_sock(sk);
2391 
2392 	if (tsk->group) {
2393 		res = -EINVAL;
2394 		goto exit;
2395 	}
2396 
2397 	if (dst->family == AF_UNSPEC) {
2398 		memset(&tsk->peer, 0, sizeof(struct sockaddr_tipc));
2399 		if (!tipc_sk_type_connectionless(sk))
2400 			res = -EINVAL;
2401 		goto exit;
2402 	}
2403 	if (!tipc_sockaddr_is_sane(dst)) {
2404 		res = -EINVAL;
2405 		goto exit;
2406 	}
2407 	/* DGRAM/RDM connect(), just save the destaddr */
2408 	if (tipc_sk_type_connectionless(sk)) {
2409 		memcpy(&tsk->peer, dest, destlen);
2410 		goto exit;
2411 	} else if (dst->addrtype == TIPC_SERVICE_RANGE) {
2412 		res = -EINVAL;
2413 		goto exit;
2414 	}
2415 
2416 	previous = sk->sk_state;
2417 
2418 	switch (sk->sk_state) {
2419 	case TIPC_OPEN:
2420 		/* Send a 'SYN-' to destination */
2421 		m.msg_name = dest;
2422 		m.msg_namelen = destlen;
2423 
2424 		/* If connect is in non-blocking case, set MSG_DONTWAIT to
2425 		 * indicate send_msg() is never blocked.
2426 		 */
2427 		if (!timeout)
2428 			m.msg_flags = MSG_DONTWAIT;
2429 
2430 		res = __tipc_sendmsg(sock, &m, 0);
2431 		if ((res < 0) && (res != -EWOULDBLOCK))
2432 			goto exit;
2433 
2434 		/* Just entered TIPC_CONNECTING state; the only
2435 		 * difference is that return value in non-blocking
2436 		 * case is EINPROGRESS, rather than EALREADY.
2437 		 */
2438 		res = -EINPROGRESS;
2439 		/* fall through */
2440 	case TIPC_CONNECTING:
2441 		if (!timeout) {
2442 			if (previous == TIPC_CONNECTING)
2443 				res = -EALREADY;
2444 			goto exit;
2445 		}
2446 		timeout = msecs_to_jiffies(timeout);
2447 		/* Wait until an 'ACK' or 'RST' arrives, or a timeout occurs */
2448 		res = tipc_wait_for_connect(sock, &timeout);
2449 		break;
2450 	case TIPC_ESTABLISHED:
2451 		res = -EISCONN;
2452 		break;
2453 	default:
2454 		res = -EINVAL;
2455 	}
2456 
2457 exit:
2458 	release_sock(sk);
2459 	return res;
2460 }
2461 
2462 /**
2463  * tipc_listen - allow socket to listen for incoming connections
2464  * @sock: socket structure
2465  * @len: (unused)
2466  *
2467  * Returns 0 on success, errno otherwise
2468  */
2469 static int tipc_listen(struct socket *sock, int len)
2470 {
2471 	struct sock *sk = sock->sk;
2472 	int res;
2473 
2474 	lock_sock(sk);
2475 	res = tipc_set_sk_state(sk, TIPC_LISTEN);
2476 	release_sock(sk);
2477 
2478 	return res;
2479 }
2480 
2481 static int tipc_wait_for_accept(struct socket *sock, long timeo)
2482 {
2483 	struct sock *sk = sock->sk;
2484 	DEFINE_WAIT(wait);
2485 	int err;
2486 
2487 	/* True wake-one mechanism for incoming connections: only
2488 	 * one process gets woken up, not the 'whole herd'.
2489 	 * Since we do not 'race & poll' for established sockets
2490 	 * anymore, the common case will execute the loop only once.
2491 	*/
2492 	for (;;) {
2493 		prepare_to_wait_exclusive(sk_sleep(sk), &wait,
2494 					  TASK_INTERRUPTIBLE);
2495 		if (timeo && skb_queue_empty(&sk->sk_receive_queue)) {
2496 			release_sock(sk);
2497 			timeo = schedule_timeout(timeo);
2498 			lock_sock(sk);
2499 		}
2500 		err = 0;
2501 		if (!skb_queue_empty(&sk->sk_receive_queue))
2502 			break;
2503 		err = -EAGAIN;
2504 		if (!timeo)
2505 			break;
2506 		err = sock_intr_errno(timeo);
2507 		if (signal_pending(current))
2508 			break;
2509 	}
2510 	finish_wait(sk_sleep(sk), &wait);
2511 	return err;
2512 }
2513 
2514 /**
2515  * tipc_accept - wait for connection request
2516  * @sock: listening socket
2517  * @newsock: new socket that is to be connected
2518  * @flags: file-related flags associated with socket
2519  *
2520  * Returns 0 on success, errno otherwise
2521  */
2522 static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags,
2523 		       bool kern)
2524 {
2525 	struct sock *new_sk, *sk = sock->sk;
2526 	struct sk_buff *buf;
2527 	struct tipc_sock *new_tsock;
2528 	struct tipc_msg *msg;
2529 	long timeo;
2530 	int res;
2531 
2532 	lock_sock(sk);
2533 
2534 	if (sk->sk_state != TIPC_LISTEN) {
2535 		res = -EINVAL;
2536 		goto exit;
2537 	}
2538 	timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
2539 	res = tipc_wait_for_accept(sock, timeo);
2540 	if (res)
2541 		goto exit;
2542 
2543 	buf = skb_peek(&sk->sk_receive_queue);
2544 
2545 	res = tipc_sk_create(sock_net(sock->sk), new_sock, 0, kern);
2546 	if (res)
2547 		goto exit;
2548 	security_sk_clone(sock->sk, new_sock->sk);
2549 
2550 	new_sk = new_sock->sk;
2551 	new_tsock = tipc_sk(new_sk);
2552 	msg = buf_msg(buf);
2553 
2554 	/* we lock on new_sk; but lockdep sees the lock on sk */
2555 	lock_sock_nested(new_sk, SINGLE_DEPTH_NESTING);
2556 
2557 	/*
2558 	 * Reject any stray messages received by new socket
2559 	 * before the socket lock was taken (very, very unlikely)
2560 	 */
2561 	tsk_rej_rx_queue(new_sk);
2562 
2563 	/* Connect new socket to it's peer */
2564 	tipc_sk_finish_conn(new_tsock, msg_origport(msg), msg_orignode(msg));
2565 
2566 	tsk_set_importance(new_tsock, msg_importance(msg));
2567 	if (msg_named(msg)) {
2568 		new_tsock->conn_type = msg_nametype(msg);
2569 		new_tsock->conn_instance = msg_nameinst(msg);
2570 	}
2571 
2572 	/*
2573 	 * Respond to 'SYN-' by discarding it & returning 'ACK'-.
2574 	 * Respond to 'SYN+' by queuing it on new socket.
2575 	 */
2576 	if (!msg_data_sz(msg)) {
2577 		struct msghdr m = {NULL,};
2578 
2579 		tsk_advance_rx_queue(sk);
2580 		__tipc_sendstream(new_sock, &m, 0);
2581 	} else {
2582 		__skb_dequeue(&sk->sk_receive_queue);
2583 		__skb_queue_head(&new_sk->sk_receive_queue, buf);
2584 		skb_set_owner_r(buf, new_sk);
2585 	}
2586 	release_sock(new_sk);
2587 exit:
2588 	release_sock(sk);
2589 	return res;
2590 }
2591 
2592 /**
2593  * tipc_shutdown - shutdown socket connection
2594  * @sock: socket structure
2595  * @how: direction to close (must be SHUT_RDWR)
2596  *
2597  * Terminates connection (if necessary), then purges socket's receive queue.
2598  *
2599  * Returns 0 on success, errno otherwise
2600  */
2601 static int tipc_shutdown(struct socket *sock, int how)
2602 {
2603 	struct sock *sk = sock->sk;
2604 	int res;
2605 
2606 	if (how != SHUT_RDWR)
2607 		return -EINVAL;
2608 
2609 	lock_sock(sk);
2610 
2611 	trace_tipc_sk_shutdown(sk, NULL, TIPC_DUMP_ALL, " ");
2612 	__tipc_shutdown(sock, TIPC_CONN_SHUTDOWN);
2613 	sk->sk_shutdown = SEND_SHUTDOWN;
2614 
2615 	if (sk->sk_state == TIPC_DISCONNECTING) {
2616 		/* Discard any unreceived messages */
2617 		__skb_queue_purge(&sk->sk_receive_queue);
2618 
2619 		/* Wake up anyone sleeping in poll */
2620 		sk->sk_state_change(sk);
2621 		res = 0;
2622 	} else {
2623 		res = -ENOTCONN;
2624 	}
2625 
2626 	release_sock(sk);
2627 	return res;
2628 }
2629 
2630 static void tipc_sk_check_probing_state(struct sock *sk,
2631 					struct sk_buff_head *list)
2632 {
2633 	struct tipc_sock *tsk = tipc_sk(sk);
2634 	u32 pnode = tsk_peer_node(tsk);
2635 	u32 pport = tsk_peer_port(tsk);
2636 	u32 self = tsk_own_node(tsk);
2637 	u32 oport = tsk->portid;
2638 	struct sk_buff *skb;
2639 
2640 	if (tsk->probe_unacked) {
2641 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2642 		sk->sk_err = ECONNABORTED;
2643 		tipc_node_remove_conn(sock_net(sk), pnode, pport);
2644 		sk->sk_state_change(sk);
2645 		return;
2646 	}
2647 	/* Prepare new probe */
2648 	skb = tipc_msg_create(CONN_MANAGER, CONN_PROBE, INT_H_SIZE, 0,
2649 			      pnode, self, pport, oport, TIPC_OK);
2650 	if (skb)
2651 		__skb_queue_tail(list, skb);
2652 	tsk->probe_unacked = true;
2653 	sk_reset_timer(sk, &sk->sk_timer, jiffies + CONN_PROBING_INTV);
2654 }
2655 
2656 static void tipc_sk_retry_connect(struct sock *sk, struct sk_buff_head *list)
2657 {
2658 	struct tipc_sock *tsk = tipc_sk(sk);
2659 
2660 	/* Try again later if dest link is congested */
2661 	if (tsk->cong_link_cnt) {
2662 		sk_reset_timer(sk, &sk->sk_timer, msecs_to_jiffies(100));
2663 		return;
2664 	}
2665 	/* Prepare SYN for retransmit */
2666 	tipc_msg_skb_clone(&sk->sk_write_queue, list);
2667 }
2668 
2669 static void tipc_sk_timeout(struct timer_list *t)
2670 {
2671 	struct sock *sk = from_timer(sk, t, sk_timer);
2672 	struct tipc_sock *tsk = tipc_sk(sk);
2673 	u32 pnode = tsk_peer_node(tsk);
2674 	struct sk_buff_head list;
2675 	int rc = 0;
2676 
2677 	__skb_queue_head_init(&list);
2678 	bh_lock_sock(sk);
2679 
2680 	/* Try again later if socket is busy */
2681 	if (sock_owned_by_user(sk)) {
2682 		sk_reset_timer(sk, &sk->sk_timer, jiffies + HZ / 20);
2683 		bh_unlock_sock(sk);
2684 		return;
2685 	}
2686 
2687 	if (sk->sk_state == TIPC_ESTABLISHED)
2688 		tipc_sk_check_probing_state(sk, &list);
2689 	else if (sk->sk_state == TIPC_CONNECTING)
2690 		tipc_sk_retry_connect(sk, &list);
2691 
2692 	bh_unlock_sock(sk);
2693 
2694 	if (!skb_queue_empty(&list))
2695 		rc = tipc_node_xmit(sock_net(sk), &list, pnode, tsk->portid);
2696 
2697 	/* SYN messages may cause link congestion */
2698 	if (rc == -ELINKCONG) {
2699 		tipc_dest_push(&tsk->cong_links, pnode, 0);
2700 		tsk->cong_link_cnt = 1;
2701 	}
2702 	sock_put(sk);
2703 }
2704 
2705 static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
2706 			   struct tipc_name_seq const *seq)
2707 {
2708 	struct sock *sk = &tsk->sk;
2709 	struct net *net = sock_net(sk);
2710 	struct publication *publ;
2711 	u32 key;
2712 
2713 	if (scope != TIPC_NODE_SCOPE)
2714 		scope = TIPC_CLUSTER_SCOPE;
2715 
2716 	if (tipc_sk_connected(sk))
2717 		return -EINVAL;
2718 	key = tsk->portid + tsk->pub_count + 1;
2719 	if (key == tsk->portid)
2720 		return -EADDRINUSE;
2721 
2722 	publ = tipc_nametbl_publish(net, seq->type, seq->lower, seq->upper,
2723 				    scope, tsk->portid, key);
2724 	if (unlikely(!publ))
2725 		return -EINVAL;
2726 
2727 	list_add(&publ->binding_sock, &tsk->publications);
2728 	tsk->pub_count++;
2729 	tsk->published = 1;
2730 	return 0;
2731 }
2732 
2733 static int tipc_sk_withdraw(struct tipc_sock *tsk, uint scope,
2734 			    struct tipc_name_seq const *seq)
2735 {
2736 	struct net *net = sock_net(&tsk->sk);
2737 	struct publication *publ;
2738 	struct publication *safe;
2739 	int rc = -EINVAL;
2740 
2741 	if (scope != TIPC_NODE_SCOPE)
2742 		scope = TIPC_CLUSTER_SCOPE;
2743 
2744 	list_for_each_entry_safe(publ, safe, &tsk->publications, binding_sock) {
2745 		if (seq) {
2746 			if (publ->scope != scope)
2747 				continue;
2748 			if (publ->type != seq->type)
2749 				continue;
2750 			if (publ->lower != seq->lower)
2751 				continue;
2752 			if (publ->upper != seq->upper)
2753 				break;
2754 			tipc_nametbl_withdraw(net, publ->type, publ->lower,
2755 					      publ->upper, publ->key);
2756 			rc = 0;
2757 			break;
2758 		}
2759 		tipc_nametbl_withdraw(net, publ->type, publ->lower,
2760 				      publ->upper, publ->key);
2761 		rc = 0;
2762 	}
2763 	if (list_empty(&tsk->publications))
2764 		tsk->published = 0;
2765 	return rc;
2766 }
2767 
2768 /* tipc_sk_reinit: set non-zero address in all existing sockets
2769  *                 when we go from standalone to network mode.
2770  */
2771 void tipc_sk_reinit(struct net *net)
2772 {
2773 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2774 	struct rhashtable_iter iter;
2775 	struct tipc_sock *tsk;
2776 	struct tipc_msg *msg;
2777 
2778 	rhashtable_walk_enter(&tn->sk_rht, &iter);
2779 
2780 	do {
2781 		rhashtable_walk_start(&iter);
2782 
2783 		while ((tsk = rhashtable_walk_next(&iter)) && !IS_ERR(tsk)) {
2784 			sock_hold(&tsk->sk);
2785 			rhashtable_walk_stop(&iter);
2786 			lock_sock(&tsk->sk);
2787 			msg = &tsk->phdr;
2788 			msg_set_prevnode(msg, tipc_own_addr(net));
2789 			msg_set_orignode(msg, tipc_own_addr(net));
2790 			release_sock(&tsk->sk);
2791 			rhashtable_walk_start(&iter);
2792 			sock_put(&tsk->sk);
2793 		}
2794 
2795 		rhashtable_walk_stop(&iter);
2796 	} while (tsk == ERR_PTR(-EAGAIN));
2797 
2798 	rhashtable_walk_exit(&iter);
2799 }
2800 
2801 static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid)
2802 {
2803 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2804 	struct tipc_sock *tsk;
2805 
2806 	rcu_read_lock();
2807 	tsk = rhashtable_lookup_fast(&tn->sk_rht, &portid, tsk_rht_params);
2808 	if (tsk)
2809 		sock_hold(&tsk->sk);
2810 	rcu_read_unlock();
2811 
2812 	return tsk;
2813 }
2814 
2815 static int tipc_sk_insert(struct tipc_sock *tsk)
2816 {
2817 	struct sock *sk = &tsk->sk;
2818 	struct net *net = sock_net(sk);
2819 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2820 	u32 remaining = (TIPC_MAX_PORT - TIPC_MIN_PORT) + 1;
2821 	u32 portid = prandom_u32() % remaining + TIPC_MIN_PORT;
2822 
2823 	while (remaining--) {
2824 		portid++;
2825 		if ((portid < TIPC_MIN_PORT) || (portid > TIPC_MAX_PORT))
2826 			portid = TIPC_MIN_PORT;
2827 		tsk->portid = portid;
2828 		sock_hold(&tsk->sk);
2829 		if (!rhashtable_lookup_insert_fast(&tn->sk_rht, &tsk->node,
2830 						   tsk_rht_params))
2831 			return 0;
2832 		sock_put(&tsk->sk);
2833 	}
2834 
2835 	return -1;
2836 }
2837 
2838 static void tipc_sk_remove(struct tipc_sock *tsk)
2839 {
2840 	struct sock *sk = &tsk->sk;
2841 	struct tipc_net *tn = net_generic(sock_net(sk), tipc_net_id);
2842 
2843 	if (!rhashtable_remove_fast(&tn->sk_rht, &tsk->node, tsk_rht_params)) {
2844 		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
2845 		__sock_put(sk);
2846 	}
2847 }
2848 
2849 static const struct rhashtable_params tsk_rht_params = {
2850 	.nelem_hint = 192,
2851 	.head_offset = offsetof(struct tipc_sock, node),
2852 	.key_offset = offsetof(struct tipc_sock, portid),
2853 	.key_len = sizeof(u32), /* portid */
2854 	.max_size = 1048576,
2855 	.min_size = 256,
2856 	.automatic_shrinking = true,
2857 };
2858 
2859 int tipc_sk_rht_init(struct net *net)
2860 {
2861 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2862 
2863 	return rhashtable_init(&tn->sk_rht, &tsk_rht_params);
2864 }
2865 
2866 void tipc_sk_rht_destroy(struct net *net)
2867 {
2868 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2869 
2870 	/* Wait for socket readers to complete */
2871 	synchronize_net();
2872 
2873 	rhashtable_destroy(&tn->sk_rht);
2874 }
2875 
2876 static int tipc_sk_join(struct tipc_sock *tsk, struct tipc_group_req *mreq)
2877 {
2878 	struct net *net = sock_net(&tsk->sk);
2879 	struct tipc_group *grp = tsk->group;
2880 	struct tipc_msg *hdr = &tsk->phdr;
2881 	struct tipc_name_seq seq;
2882 	int rc;
2883 
2884 	if (mreq->type < TIPC_RESERVED_TYPES)
2885 		return -EACCES;
2886 	if (mreq->scope > TIPC_NODE_SCOPE)
2887 		return -EINVAL;
2888 	if (grp)
2889 		return -EACCES;
2890 	grp = tipc_group_create(net, tsk->portid, mreq, &tsk->group_is_open);
2891 	if (!grp)
2892 		return -ENOMEM;
2893 	tsk->group = grp;
2894 	msg_set_lookup_scope(hdr, mreq->scope);
2895 	msg_set_nametype(hdr, mreq->type);
2896 	msg_set_dest_droppable(hdr, true);
2897 	seq.type = mreq->type;
2898 	seq.lower = mreq->instance;
2899 	seq.upper = seq.lower;
2900 	tipc_nametbl_build_group(net, grp, mreq->type, mreq->scope);
2901 	rc = tipc_sk_publish(tsk, mreq->scope, &seq);
2902 	if (rc) {
2903 		tipc_group_delete(net, grp);
2904 		tsk->group = NULL;
2905 		return rc;
2906 	}
2907 	/* Eliminate any risk that a broadcast overtakes sent JOINs */
2908 	tsk->mc_method.rcast = true;
2909 	tsk->mc_method.mandatory = true;
2910 	tipc_group_join(net, grp, &tsk->sk.sk_rcvbuf);
2911 	return rc;
2912 }
2913 
2914 static int tipc_sk_leave(struct tipc_sock *tsk)
2915 {
2916 	struct net *net = sock_net(&tsk->sk);
2917 	struct tipc_group *grp = tsk->group;
2918 	struct tipc_name_seq seq;
2919 	int scope;
2920 
2921 	if (!grp)
2922 		return -EINVAL;
2923 	tipc_group_self(grp, &seq, &scope);
2924 	tipc_group_delete(net, grp);
2925 	tsk->group = NULL;
2926 	tipc_sk_withdraw(tsk, scope, &seq);
2927 	return 0;
2928 }
2929 
2930 /**
2931  * tipc_setsockopt - set socket option
2932  * @sock: socket structure
2933  * @lvl: option level
2934  * @opt: option identifier
2935  * @ov: pointer to new option value
2936  * @ol: length of option value
2937  *
2938  * For stream sockets only, accepts and ignores all IPPROTO_TCP options
2939  * (to ease compatibility).
2940  *
2941  * Returns 0 on success, errno otherwise
2942  */
2943 static int tipc_setsockopt(struct socket *sock, int lvl, int opt,
2944 			   char __user *ov, unsigned int ol)
2945 {
2946 	struct sock *sk = sock->sk;
2947 	struct tipc_sock *tsk = tipc_sk(sk);
2948 	struct tipc_group_req mreq;
2949 	u32 value = 0;
2950 	int res = 0;
2951 
2952 	if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM))
2953 		return 0;
2954 	if (lvl != SOL_TIPC)
2955 		return -ENOPROTOOPT;
2956 
2957 	switch (opt) {
2958 	case TIPC_IMPORTANCE:
2959 	case TIPC_SRC_DROPPABLE:
2960 	case TIPC_DEST_DROPPABLE:
2961 	case TIPC_CONN_TIMEOUT:
2962 		if (ol < sizeof(value))
2963 			return -EINVAL;
2964 		if (get_user(value, (u32 __user *)ov))
2965 			return -EFAULT;
2966 		break;
2967 	case TIPC_GROUP_JOIN:
2968 		if (ol < sizeof(mreq))
2969 			return -EINVAL;
2970 		if (copy_from_user(&mreq, ov, sizeof(mreq)))
2971 			return -EFAULT;
2972 		break;
2973 	default:
2974 		if (ov || ol)
2975 			return -EINVAL;
2976 	}
2977 
2978 	lock_sock(sk);
2979 
2980 	switch (opt) {
2981 	case TIPC_IMPORTANCE:
2982 		res = tsk_set_importance(tsk, value);
2983 		break;
2984 	case TIPC_SRC_DROPPABLE:
2985 		if (sock->type != SOCK_STREAM)
2986 			tsk_set_unreliable(tsk, value);
2987 		else
2988 			res = -ENOPROTOOPT;
2989 		break;
2990 	case TIPC_DEST_DROPPABLE:
2991 		tsk_set_unreturnable(tsk, value);
2992 		break;
2993 	case TIPC_CONN_TIMEOUT:
2994 		tipc_sk(sk)->conn_timeout = value;
2995 		break;
2996 	case TIPC_MCAST_BROADCAST:
2997 		tsk->mc_method.rcast = false;
2998 		tsk->mc_method.mandatory = true;
2999 		break;
3000 	case TIPC_MCAST_REPLICAST:
3001 		tsk->mc_method.rcast = true;
3002 		tsk->mc_method.mandatory = true;
3003 		break;
3004 	case TIPC_GROUP_JOIN:
3005 		res = tipc_sk_join(tsk, &mreq);
3006 		break;
3007 	case TIPC_GROUP_LEAVE:
3008 		res = tipc_sk_leave(tsk);
3009 		break;
3010 	default:
3011 		res = -EINVAL;
3012 	}
3013 
3014 	release_sock(sk);
3015 
3016 	return res;
3017 }
3018 
3019 /**
3020  * tipc_getsockopt - get socket option
3021  * @sock: socket structure
3022  * @lvl: option level
3023  * @opt: option identifier
3024  * @ov: receptacle for option value
3025  * @ol: receptacle for length of option value
3026  *
3027  * For stream sockets only, returns 0 length result for all IPPROTO_TCP options
3028  * (to ease compatibility).
3029  *
3030  * Returns 0 on success, errno otherwise
3031  */
3032 static int tipc_getsockopt(struct socket *sock, int lvl, int opt,
3033 			   char __user *ov, int __user *ol)
3034 {
3035 	struct sock *sk = sock->sk;
3036 	struct tipc_sock *tsk = tipc_sk(sk);
3037 	struct tipc_name_seq seq;
3038 	int len, scope;
3039 	u32 value;
3040 	int res;
3041 
3042 	if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM))
3043 		return put_user(0, ol);
3044 	if (lvl != SOL_TIPC)
3045 		return -ENOPROTOOPT;
3046 	res = get_user(len, ol);
3047 	if (res)
3048 		return res;
3049 
3050 	lock_sock(sk);
3051 
3052 	switch (opt) {
3053 	case TIPC_IMPORTANCE:
3054 		value = tsk_importance(tsk);
3055 		break;
3056 	case TIPC_SRC_DROPPABLE:
3057 		value = tsk_unreliable(tsk);
3058 		break;
3059 	case TIPC_DEST_DROPPABLE:
3060 		value = tsk_unreturnable(tsk);
3061 		break;
3062 	case TIPC_CONN_TIMEOUT:
3063 		value = tsk->conn_timeout;
3064 		/* no need to set "res", since already 0 at this point */
3065 		break;
3066 	case TIPC_NODE_RECVQ_DEPTH:
3067 		value = 0; /* was tipc_queue_size, now obsolete */
3068 		break;
3069 	case TIPC_SOCK_RECVQ_DEPTH:
3070 		value = skb_queue_len(&sk->sk_receive_queue);
3071 		break;
3072 	case TIPC_SOCK_RECVQ_USED:
3073 		value = sk_rmem_alloc_get(sk);
3074 		break;
3075 	case TIPC_GROUP_JOIN:
3076 		seq.type = 0;
3077 		if (tsk->group)
3078 			tipc_group_self(tsk->group, &seq, &scope);
3079 		value = seq.type;
3080 		break;
3081 	default:
3082 		res = -EINVAL;
3083 	}
3084 
3085 	release_sock(sk);
3086 
3087 	if (res)
3088 		return res;	/* "get" failed */
3089 
3090 	if (len < sizeof(value))
3091 		return -EINVAL;
3092 
3093 	if (copy_to_user(ov, &value, sizeof(value)))
3094 		return -EFAULT;
3095 
3096 	return put_user(sizeof(value), ol);
3097 }
3098 
3099 static int tipc_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
3100 {
3101 	struct net *net = sock_net(sock->sk);
3102 	struct tipc_sioc_nodeid_req nr = {0};
3103 	struct tipc_sioc_ln_req lnr;
3104 	void __user *argp = (void __user *)arg;
3105 
3106 	switch (cmd) {
3107 	case SIOCGETLINKNAME:
3108 		if (copy_from_user(&lnr, argp, sizeof(lnr)))
3109 			return -EFAULT;
3110 		if (!tipc_node_get_linkname(net,
3111 					    lnr.bearer_id & 0xffff, lnr.peer,
3112 					    lnr.linkname, TIPC_MAX_LINK_NAME)) {
3113 			if (copy_to_user(argp, &lnr, sizeof(lnr)))
3114 				return -EFAULT;
3115 			return 0;
3116 		}
3117 		return -EADDRNOTAVAIL;
3118 	case SIOCGETNODEID:
3119 		if (copy_from_user(&nr, argp, sizeof(nr)))
3120 			return -EFAULT;
3121 		if (!tipc_node_get_id(net, nr.peer, nr.node_id))
3122 			return -EADDRNOTAVAIL;
3123 		if (copy_to_user(argp, &nr, sizeof(nr)))
3124 			return -EFAULT;
3125 		return 0;
3126 	default:
3127 		return -ENOIOCTLCMD;
3128 	}
3129 }
3130 
3131 static int tipc_socketpair(struct socket *sock1, struct socket *sock2)
3132 {
3133 	struct tipc_sock *tsk2 = tipc_sk(sock2->sk);
3134 	struct tipc_sock *tsk1 = tipc_sk(sock1->sk);
3135 	u32 onode = tipc_own_addr(sock_net(sock1->sk));
3136 
3137 	tsk1->peer.family = AF_TIPC;
3138 	tsk1->peer.addrtype = TIPC_ADDR_ID;
3139 	tsk1->peer.scope = TIPC_NODE_SCOPE;
3140 	tsk1->peer.addr.id.ref = tsk2->portid;
3141 	tsk1->peer.addr.id.node = onode;
3142 	tsk2->peer.family = AF_TIPC;
3143 	tsk2->peer.addrtype = TIPC_ADDR_ID;
3144 	tsk2->peer.scope = TIPC_NODE_SCOPE;
3145 	tsk2->peer.addr.id.ref = tsk1->portid;
3146 	tsk2->peer.addr.id.node = onode;
3147 
3148 	tipc_sk_finish_conn(tsk1, tsk2->portid, onode);
3149 	tipc_sk_finish_conn(tsk2, tsk1->portid, onode);
3150 	return 0;
3151 }
3152 
3153 /* Protocol switches for the various types of TIPC sockets */
3154 
3155 static const struct proto_ops msg_ops = {
3156 	.owner		= THIS_MODULE,
3157 	.family		= AF_TIPC,
3158 	.release	= tipc_release,
3159 	.bind		= tipc_bind,
3160 	.connect	= tipc_connect,
3161 	.socketpair	= tipc_socketpair,
3162 	.accept		= sock_no_accept,
3163 	.getname	= tipc_getname,
3164 	.poll		= tipc_poll,
3165 	.ioctl		= tipc_ioctl,
3166 	.listen		= sock_no_listen,
3167 	.shutdown	= tipc_shutdown,
3168 	.setsockopt	= tipc_setsockopt,
3169 	.getsockopt	= tipc_getsockopt,
3170 	.sendmsg	= tipc_sendmsg,
3171 	.recvmsg	= tipc_recvmsg,
3172 	.mmap		= sock_no_mmap,
3173 	.sendpage	= sock_no_sendpage
3174 };
3175 
3176 static const struct proto_ops packet_ops = {
3177 	.owner		= THIS_MODULE,
3178 	.family		= AF_TIPC,
3179 	.release	= tipc_release,
3180 	.bind		= tipc_bind,
3181 	.connect	= tipc_connect,
3182 	.socketpair	= tipc_socketpair,
3183 	.accept		= tipc_accept,
3184 	.getname	= tipc_getname,
3185 	.poll		= tipc_poll,
3186 	.ioctl		= tipc_ioctl,
3187 	.listen		= tipc_listen,
3188 	.shutdown	= tipc_shutdown,
3189 	.setsockopt	= tipc_setsockopt,
3190 	.getsockopt	= tipc_getsockopt,
3191 	.sendmsg	= tipc_send_packet,
3192 	.recvmsg	= tipc_recvmsg,
3193 	.mmap		= sock_no_mmap,
3194 	.sendpage	= sock_no_sendpage
3195 };
3196 
3197 static const struct proto_ops stream_ops = {
3198 	.owner		= THIS_MODULE,
3199 	.family		= AF_TIPC,
3200 	.release	= tipc_release,
3201 	.bind		= tipc_bind,
3202 	.connect	= tipc_connect,
3203 	.socketpair	= tipc_socketpair,
3204 	.accept		= tipc_accept,
3205 	.getname	= tipc_getname,
3206 	.poll		= tipc_poll,
3207 	.ioctl		= tipc_ioctl,
3208 	.listen		= tipc_listen,
3209 	.shutdown	= tipc_shutdown,
3210 	.setsockopt	= tipc_setsockopt,
3211 	.getsockopt	= tipc_getsockopt,
3212 	.sendmsg	= tipc_sendstream,
3213 	.recvmsg	= tipc_recvstream,
3214 	.mmap		= sock_no_mmap,
3215 	.sendpage	= sock_no_sendpage
3216 };
3217 
3218 static const struct net_proto_family tipc_family_ops = {
3219 	.owner		= THIS_MODULE,
3220 	.family		= AF_TIPC,
3221 	.create		= tipc_sk_create
3222 };
3223 
3224 static struct proto tipc_proto = {
3225 	.name		= "TIPC",
3226 	.owner		= THIS_MODULE,
3227 	.obj_size	= sizeof(struct tipc_sock),
3228 	.sysctl_rmem	= sysctl_tipc_rmem
3229 };
3230 
3231 /**
3232  * tipc_socket_init - initialize TIPC socket interface
3233  *
3234  * Returns 0 on success, errno otherwise
3235  */
3236 int tipc_socket_init(void)
3237 {
3238 	int res;
3239 
3240 	res = proto_register(&tipc_proto, 1);
3241 	if (res) {
3242 		pr_err("Failed to register TIPC protocol type\n");
3243 		goto out;
3244 	}
3245 
3246 	res = sock_register(&tipc_family_ops);
3247 	if (res) {
3248 		pr_err("Failed to register TIPC socket type\n");
3249 		proto_unregister(&tipc_proto);
3250 		goto out;
3251 	}
3252  out:
3253 	return res;
3254 }
3255 
3256 /**
3257  * tipc_socket_stop - stop TIPC socket interface
3258  */
3259 void tipc_socket_stop(void)
3260 {
3261 	sock_unregister(tipc_family_ops.family);
3262 	proto_unregister(&tipc_proto);
3263 }
3264 
3265 /* Caller should hold socket lock for the passed tipc socket. */
3266 static int __tipc_nl_add_sk_con(struct sk_buff *skb, struct tipc_sock *tsk)
3267 {
3268 	u32 peer_node;
3269 	u32 peer_port;
3270 	struct nlattr *nest;
3271 
3272 	peer_node = tsk_peer_node(tsk);
3273 	peer_port = tsk_peer_port(tsk);
3274 
3275 	nest = nla_nest_start_noflag(skb, TIPC_NLA_SOCK_CON);
3276 	if (!nest)
3277 		return -EMSGSIZE;
3278 
3279 	if (nla_put_u32(skb, TIPC_NLA_CON_NODE, peer_node))
3280 		goto msg_full;
3281 	if (nla_put_u32(skb, TIPC_NLA_CON_SOCK, peer_port))
3282 		goto msg_full;
3283 
3284 	if (tsk->conn_type != 0) {
3285 		if (nla_put_flag(skb, TIPC_NLA_CON_FLAG))
3286 			goto msg_full;
3287 		if (nla_put_u32(skb, TIPC_NLA_CON_TYPE, tsk->conn_type))
3288 			goto msg_full;
3289 		if (nla_put_u32(skb, TIPC_NLA_CON_INST, tsk->conn_instance))
3290 			goto msg_full;
3291 	}
3292 	nla_nest_end(skb, nest);
3293 
3294 	return 0;
3295 
3296 msg_full:
3297 	nla_nest_cancel(skb, nest);
3298 
3299 	return -EMSGSIZE;
3300 }
3301 
3302 static int __tipc_nl_add_sk_info(struct sk_buff *skb, struct tipc_sock
3303 			  *tsk)
3304 {
3305 	struct net *net = sock_net(skb->sk);
3306 	struct sock *sk = &tsk->sk;
3307 
3308 	if (nla_put_u32(skb, TIPC_NLA_SOCK_REF, tsk->portid) ||
3309 	    nla_put_u32(skb, TIPC_NLA_SOCK_ADDR, tipc_own_addr(net)))
3310 		return -EMSGSIZE;
3311 
3312 	if (tipc_sk_connected(sk)) {
3313 		if (__tipc_nl_add_sk_con(skb, tsk))
3314 			return -EMSGSIZE;
3315 	} else if (!list_empty(&tsk->publications)) {
3316 		if (nla_put_flag(skb, TIPC_NLA_SOCK_HAS_PUBL))
3317 			return -EMSGSIZE;
3318 	}
3319 	return 0;
3320 }
3321 
3322 /* Caller should hold socket lock for the passed tipc socket. */
3323 static int __tipc_nl_add_sk(struct sk_buff *skb, struct netlink_callback *cb,
3324 			    struct tipc_sock *tsk)
3325 {
3326 	struct nlattr *attrs;
3327 	void *hdr;
3328 
3329 	hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
3330 			  &tipc_genl_family, NLM_F_MULTI, TIPC_NL_SOCK_GET);
3331 	if (!hdr)
3332 		goto msg_cancel;
3333 
3334 	attrs = nla_nest_start_noflag(skb, TIPC_NLA_SOCK);
3335 	if (!attrs)
3336 		goto genlmsg_cancel;
3337 
3338 	if (__tipc_nl_add_sk_info(skb, tsk))
3339 		goto attr_msg_cancel;
3340 
3341 	nla_nest_end(skb, attrs);
3342 	genlmsg_end(skb, hdr);
3343 
3344 	return 0;
3345 
3346 attr_msg_cancel:
3347 	nla_nest_cancel(skb, attrs);
3348 genlmsg_cancel:
3349 	genlmsg_cancel(skb, hdr);
3350 msg_cancel:
3351 	return -EMSGSIZE;
3352 }
3353 
3354 int tipc_nl_sk_walk(struct sk_buff *skb, struct netlink_callback *cb,
3355 		    int (*skb_handler)(struct sk_buff *skb,
3356 				       struct netlink_callback *cb,
3357 				       struct tipc_sock *tsk))
3358 {
3359 	struct rhashtable_iter *iter = (void *)cb->args[4];
3360 	struct tipc_sock *tsk;
3361 	int err;
3362 
3363 	rhashtable_walk_start(iter);
3364 	while ((tsk = rhashtable_walk_next(iter)) != NULL) {
3365 		if (IS_ERR(tsk)) {
3366 			err = PTR_ERR(tsk);
3367 			if (err == -EAGAIN) {
3368 				err = 0;
3369 				continue;
3370 			}
3371 			break;
3372 		}
3373 
3374 		sock_hold(&tsk->sk);
3375 		rhashtable_walk_stop(iter);
3376 		lock_sock(&tsk->sk);
3377 		err = skb_handler(skb, cb, tsk);
3378 		if (err) {
3379 			release_sock(&tsk->sk);
3380 			sock_put(&tsk->sk);
3381 			goto out;
3382 		}
3383 		release_sock(&tsk->sk);
3384 		rhashtable_walk_start(iter);
3385 		sock_put(&tsk->sk);
3386 	}
3387 	rhashtable_walk_stop(iter);
3388 out:
3389 	return skb->len;
3390 }
3391 EXPORT_SYMBOL(tipc_nl_sk_walk);
3392 
3393 int tipc_dump_start(struct netlink_callback *cb)
3394 {
3395 	return __tipc_dump_start(cb, sock_net(cb->skb->sk));
3396 }
3397 EXPORT_SYMBOL(tipc_dump_start);
3398 
3399 int __tipc_dump_start(struct netlink_callback *cb, struct net *net)
3400 {
3401 	/* tipc_nl_name_table_dump() uses cb->args[0...3]. */
3402 	struct rhashtable_iter *iter = (void *)cb->args[4];
3403 	struct tipc_net *tn = tipc_net(net);
3404 
3405 	if (!iter) {
3406 		iter = kmalloc(sizeof(*iter), GFP_KERNEL);
3407 		if (!iter)
3408 			return -ENOMEM;
3409 
3410 		cb->args[4] = (long)iter;
3411 	}
3412 
3413 	rhashtable_walk_enter(&tn->sk_rht, iter);
3414 	return 0;
3415 }
3416 
3417 int tipc_dump_done(struct netlink_callback *cb)
3418 {
3419 	struct rhashtable_iter *hti = (void *)cb->args[4];
3420 
3421 	rhashtable_walk_exit(hti);
3422 	kfree(hti);
3423 	return 0;
3424 }
3425 EXPORT_SYMBOL(tipc_dump_done);
3426 
3427 int tipc_sk_fill_sock_diag(struct sk_buff *skb, struct netlink_callback *cb,
3428 			   struct tipc_sock *tsk, u32 sk_filter_state,
3429 			   u64 (*tipc_diag_gen_cookie)(struct sock *sk))
3430 {
3431 	struct sock *sk = &tsk->sk;
3432 	struct nlattr *attrs;
3433 	struct nlattr *stat;
3434 
3435 	/*filter response w.r.t sk_state*/
3436 	if (!(sk_filter_state & (1 << sk->sk_state)))
3437 		return 0;
3438 
3439 	attrs = nla_nest_start_noflag(skb, TIPC_NLA_SOCK);
3440 	if (!attrs)
3441 		goto msg_cancel;
3442 
3443 	if (__tipc_nl_add_sk_info(skb, tsk))
3444 		goto attr_msg_cancel;
3445 
3446 	if (nla_put_u32(skb, TIPC_NLA_SOCK_TYPE, (u32)sk->sk_type) ||
3447 	    nla_put_u32(skb, TIPC_NLA_SOCK_TIPC_STATE, (u32)sk->sk_state) ||
3448 	    nla_put_u32(skb, TIPC_NLA_SOCK_INO, sock_i_ino(sk)) ||
3449 	    nla_put_u32(skb, TIPC_NLA_SOCK_UID,
3450 			from_kuid_munged(sk_user_ns(NETLINK_CB(cb->skb).sk),
3451 					 sock_i_uid(sk))) ||
3452 	    nla_put_u64_64bit(skb, TIPC_NLA_SOCK_COOKIE,
3453 			      tipc_diag_gen_cookie(sk),
3454 			      TIPC_NLA_SOCK_PAD))
3455 		goto attr_msg_cancel;
3456 
3457 	stat = nla_nest_start_noflag(skb, TIPC_NLA_SOCK_STAT);
3458 	if (!stat)
3459 		goto attr_msg_cancel;
3460 
3461 	if (nla_put_u32(skb, TIPC_NLA_SOCK_STAT_RCVQ,
3462 			skb_queue_len(&sk->sk_receive_queue)) ||
3463 	    nla_put_u32(skb, TIPC_NLA_SOCK_STAT_SENDQ,
3464 			skb_queue_len(&sk->sk_write_queue)) ||
3465 	    nla_put_u32(skb, TIPC_NLA_SOCK_STAT_DROP,
3466 			atomic_read(&sk->sk_drops)))
3467 		goto stat_msg_cancel;
3468 
3469 	if (tsk->cong_link_cnt &&
3470 	    nla_put_flag(skb, TIPC_NLA_SOCK_STAT_LINK_CONG))
3471 		goto stat_msg_cancel;
3472 
3473 	if (tsk_conn_cong(tsk) &&
3474 	    nla_put_flag(skb, TIPC_NLA_SOCK_STAT_CONN_CONG))
3475 		goto stat_msg_cancel;
3476 
3477 	nla_nest_end(skb, stat);
3478 
3479 	if (tsk->group)
3480 		if (tipc_group_fill_sock_diag(tsk->group, skb))
3481 			goto stat_msg_cancel;
3482 
3483 	nla_nest_end(skb, attrs);
3484 
3485 	return 0;
3486 
3487 stat_msg_cancel:
3488 	nla_nest_cancel(skb, stat);
3489 attr_msg_cancel:
3490 	nla_nest_cancel(skb, attrs);
3491 msg_cancel:
3492 	return -EMSGSIZE;
3493 }
3494 EXPORT_SYMBOL(tipc_sk_fill_sock_diag);
3495 
3496 int tipc_nl_sk_dump(struct sk_buff *skb, struct netlink_callback *cb)
3497 {
3498 	return tipc_nl_sk_walk(skb, cb, __tipc_nl_add_sk);
3499 }
3500 
3501 /* Caller should hold socket lock for the passed tipc socket. */
3502 static int __tipc_nl_add_sk_publ(struct sk_buff *skb,
3503 				 struct netlink_callback *cb,
3504 				 struct publication *publ)
3505 {
3506 	void *hdr;
3507 	struct nlattr *attrs;
3508 
3509 	hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
3510 			  &tipc_genl_family, NLM_F_MULTI, TIPC_NL_PUBL_GET);
3511 	if (!hdr)
3512 		goto msg_cancel;
3513 
3514 	attrs = nla_nest_start_noflag(skb, TIPC_NLA_PUBL);
3515 	if (!attrs)
3516 		goto genlmsg_cancel;
3517 
3518 	if (nla_put_u32(skb, TIPC_NLA_PUBL_KEY, publ->key))
3519 		goto attr_msg_cancel;
3520 	if (nla_put_u32(skb, TIPC_NLA_PUBL_TYPE, publ->type))
3521 		goto attr_msg_cancel;
3522 	if (nla_put_u32(skb, TIPC_NLA_PUBL_LOWER, publ->lower))
3523 		goto attr_msg_cancel;
3524 	if (nla_put_u32(skb, TIPC_NLA_PUBL_UPPER, publ->upper))
3525 		goto attr_msg_cancel;
3526 
3527 	nla_nest_end(skb, attrs);
3528 	genlmsg_end(skb, hdr);
3529 
3530 	return 0;
3531 
3532 attr_msg_cancel:
3533 	nla_nest_cancel(skb, attrs);
3534 genlmsg_cancel:
3535 	genlmsg_cancel(skb, hdr);
3536 msg_cancel:
3537 	return -EMSGSIZE;
3538 }
3539 
3540 /* Caller should hold socket lock for the passed tipc socket. */
3541 static int __tipc_nl_list_sk_publ(struct sk_buff *skb,
3542 				  struct netlink_callback *cb,
3543 				  struct tipc_sock *tsk, u32 *last_publ)
3544 {
3545 	int err;
3546 	struct publication *p;
3547 
3548 	if (*last_publ) {
3549 		list_for_each_entry(p, &tsk->publications, binding_sock) {
3550 			if (p->key == *last_publ)
3551 				break;
3552 		}
3553 		if (p->key != *last_publ) {
3554 			/* We never set seq or call nl_dump_check_consistent()
3555 			 * this means that setting prev_seq here will cause the
3556 			 * consistence check to fail in the netlink callback
3557 			 * handler. Resulting in the last NLMSG_DONE message
3558 			 * having the NLM_F_DUMP_INTR flag set.
3559 			 */
3560 			cb->prev_seq = 1;
3561 			*last_publ = 0;
3562 			return -EPIPE;
3563 		}
3564 	} else {
3565 		p = list_first_entry(&tsk->publications, struct publication,
3566 				     binding_sock);
3567 	}
3568 
3569 	list_for_each_entry_from(p, &tsk->publications, binding_sock) {
3570 		err = __tipc_nl_add_sk_publ(skb, cb, p);
3571 		if (err) {
3572 			*last_publ = p->key;
3573 			return err;
3574 		}
3575 	}
3576 	*last_publ = 0;
3577 
3578 	return 0;
3579 }
3580 
3581 int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb)
3582 {
3583 	int err;
3584 	u32 tsk_portid = cb->args[0];
3585 	u32 last_publ = cb->args[1];
3586 	u32 done = cb->args[2];
3587 	struct net *net = sock_net(skb->sk);
3588 	struct tipc_sock *tsk;
3589 
3590 	if (!tsk_portid) {
3591 		struct nlattr **attrs;
3592 		struct nlattr *sock[TIPC_NLA_SOCK_MAX + 1];
3593 
3594 		err = tipc_nlmsg_parse(cb->nlh, &attrs);
3595 		if (err)
3596 			return err;
3597 
3598 		if (!attrs[TIPC_NLA_SOCK])
3599 			return -EINVAL;
3600 
3601 		err = nla_parse_nested_deprecated(sock, TIPC_NLA_SOCK_MAX,
3602 						  attrs[TIPC_NLA_SOCK],
3603 						  tipc_nl_sock_policy, NULL);
3604 		if (err)
3605 			return err;
3606 
3607 		if (!sock[TIPC_NLA_SOCK_REF])
3608 			return -EINVAL;
3609 
3610 		tsk_portid = nla_get_u32(sock[TIPC_NLA_SOCK_REF]);
3611 	}
3612 
3613 	if (done)
3614 		return 0;
3615 
3616 	tsk = tipc_sk_lookup(net, tsk_portid);
3617 	if (!tsk)
3618 		return -EINVAL;
3619 
3620 	lock_sock(&tsk->sk);
3621 	err = __tipc_nl_list_sk_publ(skb, cb, tsk, &last_publ);
3622 	if (!err)
3623 		done = 1;
3624 	release_sock(&tsk->sk);
3625 	sock_put(&tsk->sk);
3626 
3627 	cb->args[0] = tsk_portid;
3628 	cb->args[1] = last_publ;
3629 	cb->args[2] = done;
3630 
3631 	return skb->len;
3632 }
3633 
3634 /**
3635  * tipc_sk_filtering - check if a socket should be traced
3636  * @sk: the socket to be examined
3637  * @sysctl_tipc_sk_filter[]: the socket tuple for filtering,
3638  *  (portid, sock type, name type, name lower, name upper)
3639  *
3640  * Returns true if the socket meets the socket tuple data
3641  * (value 0 = 'any') or when there is no tuple set (all = 0),
3642  * otherwise false
3643  */
3644 bool tipc_sk_filtering(struct sock *sk)
3645 {
3646 	struct tipc_sock *tsk;
3647 	struct publication *p;
3648 	u32 _port, _sktype, _type, _lower, _upper;
3649 	u32 type = 0, lower = 0, upper = 0;
3650 
3651 	if (!sk)
3652 		return true;
3653 
3654 	tsk = tipc_sk(sk);
3655 
3656 	_port = sysctl_tipc_sk_filter[0];
3657 	_sktype = sysctl_tipc_sk_filter[1];
3658 	_type = sysctl_tipc_sk_filter[2];
3659 	_lower = sysctl_tipc_sk_filter[3];
3660 	_upper = sysctl_tipc_sk_filter[4];
3661 
3662 	if (!_port && !_sktype && !_type && !_lower && !_upper)
3663 		return true;
3664 
3665 	if (_port)
3666 		return (_port == tsk->portid);
3667 
3668 	if (_sktype && _sktype != sk->sk_type)
3669 		return false;
3670 
3671 	if (tsk->published) {
3672 		p = list_first_entry_or_null(&tsk->publications,
3673 					     struct publication, binding_sock);
3674 		if (p) {
3675 			type = p->type;
3676 			lower = p->lower;
3677 			upper = p->upper;
3678 		}
3679 	}
3680 
3681 	if (!tipc_sk_type_connectionless(sk)) {
3682 		type = tsk->conn_type;
3683 		lower = tsk->conn_instance;
3684 		upper = tsk->conn_instance;
3685 	}
3686 
3687 	if ((_type && _type != type) || (_lower && _lower != lower) ||
3688 	    (_upper && _upper != upper))
3689 		return false;
3690 
3691 	return true;
3692 }
3693 
3694 u32 tipc_sock_get_portid(struct sock *sk)
3695 {
3696 	return (sk) ? (tipc_sk(sk))->portid : 0;
3697 }
3698 
3699 /**
3700  * tipc_sk_overlimit1 - check if socket rx queue is about to be overloaded,
3701  *			both the rcv and backlog queues are considered
3702  * @sk: tipc sk to be checked
3703  * @skb: tipc msg to be checked
3704  *
3705  * Returns true if the socket rx queue allocation is > 90%, otherwise false
3706  */
3707 
3708 bool tipc_sk_overlimit1(struct sock *sk, struct sk_buff *skb)
3709 {
3710 	atomic_t *dcnt = &tipc_sk(sk)->dupl_rcvcnt;
3711 	unsigned int lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt);
3712 	unsigned int qsize = sk->sk_backlog.len + sk_rmem_alloc_get(sk);
3713 
3714 	return (qsize > lim * 90 / 100);
3715 }
3716 
3717 /**
3718  * tipc_sk_overlimit2 - check if socket rx queue is about to be overloaded,
3719  *			only the rcv queue is considered
3720  * @sk: tipc sk to be checked
3721  * @skb: tipc msg to be checked
3722  *
3723  * Returns true if the socket rx queue allocation is > 90%, otherwise false
3724  */
3725 
3726 bool tipc_sk_overlimit2(struct sock *sk, struct sk_buff *skb)
3727 {
3728 	unsigned int lim = rcvbuf_limit(sk, skb);
3729 	unsigned int qsize = sk_rmem_alloc_get(sk);
3730 
3731 	return (qsize > lim * 90 / 100);
3732 }
3733 
3734 /**
3735  * tipc_sk_dump - dump TIPC socket
3736  * @sk: tipc sk to be dumped
3737  * @dqueues: bitmask to decide if any socket queue to be dumped?
3738  *           - TIPC_DUMP_NONE: don't dump socket queues
3739  *           - TIPC_DUMP_SK_SNDQ: dump socket send queue
3740  *           - TIPC_DUMP_SK_RCVQ: dump socket rcv queue
3741  *           - TIPC_DUMP_SK_BKLGQ: dump socket backlog queue
3742  *           - TIPC_DUMP_ALL: dump all the socket queues above
3743  * @buf: returned buffer of dump data in format
3744  */
3745 int tipc_sk_dump(struct sock *sk, u16 dqueues, char *buf)
3746 {
3747 	int i = 0;
3748 	size_t sz = (dqueues) ? SK_LMAX : SK_LMIN;
3749 	struct tipc_sock *tsk;
3750 	struct publication *p;
3751 	bool tsk_connected;
3752 
3753 	if (!sk) {
3754 		i += scnprintf(buf, sz, "sk data: (null)\n");
3755 		return i;
3756 	}
3757 
3758 	tsk = tipc_sk(sk);
3759 	tsk_connected = !tipc_sk_type_connectionless(sk);
3760 
3761 	i += scnprintf(buf, sz, "sk data: %u", sk->sk_type);
3762 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_state);
3763 	i += scnprintf(buf + i, sz - i, " %x", tsk_own_node(tsk));
3764 	i += scnprintf(buf + i, sz - i, " %u", tsk->portid);
3765 	i += scnprintf(buf + i, sz - i, " | %u", tsk_connected);
3766 	if (tsk_connected) {
3767 		i += scnprintf(buf + i, sz - i, " %x", tsk_peer_node(tsk));
3768 		i += scnprintf(buf + i, sz - i, " %u", tsk_peer_port(tsk));
3769 		i += scnprintf(buf + i, sz - i, " %u", tsk->conn_type);
3770 		i += scnprintf(buf + i, sz - i, " %u", tsk->conn_instance);
3771 	}
3772 	i += scnprintf(buf + i, sz - i, " | %u", tsk->published);
3773 	if (tsk->published) {
3774 		p = list_first_entry_or_null(&tsk->publications,
3775 					     struct publication, binding_sock);
3776 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->type : 0);
3777 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->lower : 0);
3778 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->upper : 0);
3779 	}
3780 	i += scnprintf(buf + i, sz - i, " | %u", tsk->snd_win);
3781 	i += scnprintf(buf + i, sz - i, " %u", tsk->rcv_win);
3782 	i += scnprintf(buf + i, sz - i, " %u", tsk->max_pkt);
3783 	i += scnprintf(buf + i, sz - i, " %x", tsk->peer_caps);
3784 	i += scnprintf(buf + i, sz - i, " %u", tsk->cong_link_cnt);
3785 	i += scnprintf(buf + i, sz - i, " %u", tsk->snt_unacked);
3786 	i += scnprintf(buf + i, sz - i, " %u", tsk->rcv_unacked);
3787 	i += scnprintf(buf + i, sz - i, " %u", atomic_read(&tsk->dupl_rcvcnt));
3788 	i += scnprintf(buf + i, sz - i, " %u", sk->sk_shutdown);
3789 	i += scnprintf(buf + i, sz - i, " | %d", sk_wmem_alloc_get(sk));
3790 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_sndbuf);
3791 	i += scnprintf(buf + i, sz - i, " | %d", sk_rmem_alloc_get(sk));
3792 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_rcvbuf);
3793 	i += scnprintf(buf + i, sz - i, " | %d\n", sk->sk_backlog.len);
3794 
3795 	if (dqueues & TIPC_DUMP_SK_SNDQ) {
3796 		i += scnprintf(buf + i, sz - i, "sk_write_queue: ");
3797 		i += tipc_list_dump(&sk->sk_write_queue, false, buf + i);
3798 	}
3799 
3800 	if (dqueues & TIPC_DUMP_SK_RCVQ) {
3801 		i += scnprintf(buf + i, sz - i, "sk_receive_queue: ");
3802 		i += tipc_list_dump(&sk->sk_receive_queue, false, buf + i);
3803 	}
3804 
3805 	if (dqueues & TIPC_DUMP_SK_BKLGQ) {
3806 		i += scnprintf(buf + i, sz - i, "sk_backlog:\n  head ");
3807 		i += tipc_skb_dump(sk->sk_backlog.head, false, buf + i);
3808 		if (sk->sk_backlog.tail != sk->sk_backlog.head) {
3809 			i += scnprintf(buf + i, sz - i, "  tail ");
3810 			i += tipc_skb_dump(sk->sk_backlog.tail, false,
3811 					   buf + i);
3812 		}
3813 	}
3814 
3815 	return i;
3816 }
3817