xref: /linux/net/tipc/socket.c (revision ec8a42e7343234802b9054874fe01810880289ce)
1 /*
2  * net/tipc/socket.c: TIPC socket API
3  *
4  * Copyright (c) 2001-2007, 2012-2019, Ericsson AB
5  * Copyright (c) 2004-2008, 2010-2013, Wind River Systems
6  * Copyright (c) 2020, Red Hat Inc
7  * All rights reserved.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions are met:
11  *
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. Neither the names of the copyright holders nor the names of its
18  *    contributors may be used to endorse or promote products derived from
19  *    this software without specific prior written permission.
20  *
21  * Alternatively, this software may be distributed under the terms of the
22  * GNU General Public License ("GPL") version 2 as published by the Free
23  * Software Foundation.
24  *
25  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
26  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
29  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
30  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
31  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
32  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
33  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
34  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35  * POSSIBILITY OF SUCH DAMAGE.
36  */
37 
38 #include <linux/rhashtable.h>
39 #include <linux/sched/signal.h>
40 
41 #include "core.h"
42 #include "name_table.h"
43 #include "node.h"
44 #include "link.h"
45 #include "name_distr.h"
46 #include "socket.h"
47 #include "bcast.h"
48 #include "netlink.h"
49 #include "group.h"
50 #include "trace.h"
51 
52 #define NAGLE_START_INIT	4
53 #define NAGLE_START_MAX		1024
54 #define CONN_TIMEOUT_DEFAULT    8000    /* default connect timeout = 8s */
55 #define CONN_PROBING_INTV	msecs_to_jiffies(3600000)  /* [ms] => 1 h */
56 #define TIPC_MAX_PORT		0xffffffff
57 #define TIPC_MIN_PORT		1
58 #define TIPC_ACK_RATE		4       /* ACK at 1/4 of rcv window size */
59 
60 enum {
61 	TIPC_LISTEN = TCP_LISTEN,
62 	TIPC_ESTABLISHED = TCP_ESTABLISHED,
63 	TIPC_OPEN = TCP_CLOSE,
64 	TIPC_DISCONNECTING = TCP_CLOSE_WAIT,
65 	TIPC_CONNECTING = TCP_SYN_SENT,
66 };
67 
68 struct sockaddr_pair {
69 	struct sockaddr_tipc sock;
70 	struct sockaddr_tipc member;
71 };
72 
73 /**
74  * struct tipc_sock - TIPC socket structure
75  * @sk: socket - interacts with 'port' and with user via the socket API
76  * @conn_type: TIPC type used when connection was established
77  * @conn_instance: TIPC instance used when connection was established
78  * @published: non-zero if port has one or more associated names
79  * @max_pkt: maximum packet size "hint" used when building messages sent by port
80  * @maxnagle: maximum size of msg which can be subject to nagle
81  * @portid: unique port identity in TIPC socket hash table
82  * @phdr: preformatted message header used when sending messages
83  * @cong_links: list of congested links
84  * @publications: list of publications for port
85  * @blocking_link: address of the congested link we are currently sleeping on
86  * @pub_count: total # of publications port has made during its lifetime
87  * @conn_timeout: the time we can wait for an unresponded setup request
88  * @probe_unacked: probe has not received ack yet
89  * @dupl_rcvcnt: number of bytes counted twice, in both backlog and rcv queue
90  * @cong_link_cnt: number of congested links
91  * @snt_unacked: # messages sent by socket, and not yet acked by peer
92  * @snd_win: send window size
93  * @peer_caps: peer capabilities mask
94  * @rcv_unacked: # messages read by user, but not yet acked back to peer
95  * @rcv_win: receive window size
96  * @peer: 'connected' peer for dgram/rdm
97  * @node: hash table node
98  * @mc_method: cookie for use between socket and broadcast layer
99  * @rcu: rcu struct for tipc_sock
100  * @group: TIPC communications group
101  * @oneway: message count in one direction (FIXME)
102  * @nagle_start: current nagle value
103  * @snd_backlog: send backlog count
104  * @msg_acc: messages accepted; used in managing backlog and nagle
105  * @pkt_cnt: TIPC socket packet count
106  * @expect_ack: whether this TIPC socket is expecting an ack
107  * @nodelay: setsockopt() TIPC_NODELAY setting
108  * @group_is_open: TIPC socket group is fully open (FIXME)
109  */
110 struct tipc_sock {
111 	struct sock sk;
112 	u32 conn_type;
113 	u32 conn_instance;
114 	int published;
115 	u32 max_pkt;
116 	u32 maxnagle;
117 	u32 portid;
118 	struct tipc_msg phdr;
119 	struct list_head cong_links;
120 	struct list_head publications;
121 	u32 pub_count;
122 	atomic_t dupl_rcvcnt;
123 	u16 conn_timeout;
124 	bool probe_unacked;
125 	u16 cong_link_cnt;
126 	u16 snt_unacked;
127 	u16 snd_win;
128 	u16 peer_caps;
129 	u16 rcv_unacked;
130 	u16 rcv_win;
131 	struct sockaddr_tipc peer;
132 	struct rhash_head node;
133 	struct tipc_mc_method mc_method;
134 	struct rcu_head rcu;
135 	struct tipc_group *group;
136 	u32 oneway;
137 	u32 nagle_start;
138 	u16 snd_backlog;
139 	u16 msg_acc;
140 	u16 pkt_cnt;
141 	bool expect_ack;
142 	bool nodelay;
143 	bool group_is_open;
144 };
145 
146 static int tipc_sk_backlog_rcv(struct sock *sk, struct sk_buff *skb);
147 static void tipc_data_ready(struct sock *sk);
148 static void tipc_write_space(struct sock *sk);
149 static void tipc_sock_destruct(struct sock *sk);
150 static int tipc_release(struct socket *sock);
151 static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags,
152 		       bool kern);
153 static void tipc_sk_timeout(struct timer_list *t);
154 static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
155 			   struct tipc_service_range const *seq);
156 static int tipc_sk_withdraw(struct tipc_sock *tsk, uint scope,
157 			    struct tipc_service_range const *seq);
158 static int tipc_sk_leave(struct tipc_sock *tsk);
159 static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid);
160 static int tipc_sk_insert(struct tipc_sock *tsk);
161 static void tipc_sk_remove(struct tipc_sock *tsk);
162 static int __tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dsz);
163 static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dsz);
164 static void tipc_sk_push_backlog(struct tipc_sock *tsk, bool nagle_ack);
165 
166 static const struct proto_ops packet_ops;
167 static const struct proto_ops stream_ops;
168 static const struct proto_ops msg_ops;
169 static struct proto tipc_proto;
170 static const struct rhashtable_params tsk_rht_params;
171 
172 static u32 tsk_own_node(struct tipc_sock *tsk)
173 {
174 	return msg_prevnode(&tsk->phdr);
175 }
176 
177 static u32 tsk_peer_node(struct tipc_sock *tsk)
178 {
179 	return msg_destnode(&tsk->phdr);
180 }
181 
182 static u32 tsk_peer_port(struct tipc_sock *tsk)
183 {
184 	return msg_destport(&tsk->phdr);
185 }
186 
187 static  bool tsk_unreliable(struct tipc_sock *tsk)
188 {
189 	return msg_src_droppable(&tsk->phdr) != 0;
190 }
191 
192 static void tsk_set_unreliable(struct tipc_sock *tsk, bool unreliable)
193 {
194 	msg_set_src_droppable(&tsk->phdr, unreliable ? 1 : 0);
195 }
196 
197 static bool tsk_unreturnable(struct tipc_sock *tsk)
198 {
199 	return msg_dest_droppable(&tsk->phdr) != 0;
200 }
201 
202 static void tsk_set_unreturnable(struct tipc_sock *tsk, bool unreturnable)
203 {
204 	msg_set_dest_droppable(&tsk->phdr, unreturnable ? 1 : 0);
205 }
206 
207 static int tsk_importance(struct tipc_sock *tsk)
208 {
209 	return msg_importance(&tsk->phdr);
210 }
211 
212 static struct tipc_sock *tipc_sk(const struct sock *sk)
213 {
214 	return container_of(sk, struct tipc_sock, sk);
215 }
216 
217 int tsk_set_importance(struct sock *sk, int imp)
218 {
219 	if (imp > TIPC_CRITICAL_IMPORTANCE)
220 		return -EINVAL;
221 	msg_set_importance(&tipc_sk(sk)->phdr, (u32)imp);
222 	return 0;
223 }
224 
225 static bool tsk_conn_cong(struct tipc_sock *tsk)
226 {
227 	return tsk->snt_unacked > tsk->snd_win;
228 }
229 
230 static u16 tsk_blocks(int len)
231 {
232 	return ((len / FLOWCTL_BLK_SZ) + 1);
233 }
234 
235 /* tsk_blocks(): translate a buffer size in bytes to number of
236  * advertisable blocks, taking into account the ratio truesize(len)/len
237  * We can trust that this ratio is always < 4 for len >= FLOWCTL_BLK_SZ
238  */
239 static u16 tsk_adv_blocks(int len)
240 {
241 	return len / FLOWCTL_BLK_SZ / 4;
242 }
243 
244 /* tsk_inc(): increment counter for sent or received data
245  * - If block based flow control is not supported by peer we
246  *   fall back to message based ditto, incrementing the counter
247  */
248 static u16 tsk_inc(struct tipc_sock *tsk, int msglen)
249 {
250 	if (likely(tsk->peer_caps & TIPC_BLOCK_FLOWCTL))
251 		return ((msglen / FLOWCTL_BLK_SZ) + 1);
252 	return 1;
253 }
254 
255 /* tsk_set_nagle - enable/disable nagle property by manipulating maxnagle
256  */
257 static void tsk_set_nagle(struct tipc_sock *tsk)
258 {
259 	struct sock *sk = &tsk->sk;
260 
261 	tsk->maxnagle = 0;
262 	if (sk->sk_type != SOCK_STREAM)
263 		return;
264 	if (tsk->nodelay)
265 		return;
266 	if (!(tsk->peer_caps & TIPC_NAGLE))
267 		return;
268 	/* Limit node local buffer size to avoid receive queue overflow */
269 	if (tsk->max_pkt == MAX_MSG_SIZE)
270 		tsk->maxnagle = 1500;
271 	else
272 		tsk->maxnagle = tsk->max_pkt;
273 }
274 
275 /**
276  * tsk_advance_rx_queue - discard first buffer in socket receive queue
277  * @sk: network socket
278  *
279  * Caller must hold socket lock
280  */
281 static void tsk_advance_rx_queue(struct sock *sk)
282 {
283 	trace_tipc_sk_advance_rx(sk, NULL, TIPC_DUMP_SK_RCVQ, " ");
284 	kfree_skb(__skb_dequeue(&sk->sk_receive_queue));
285 }
286 
287 /* tipc_sk_respond() : send response message back to sender
288  */
289 static void tipc_sk_respond(struct sock *sk, struct sk_buff *skb, int err)
290 {
291 	u32 selector;
292 	u32 dnode;
293 	u32 onode = tipc_own_addr(sock_net(sk));
294 
295 	if (!tipc_msg_reverse(onode, &skb, err))
296 		return;
297 
298 	trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_NONE, "@sk_respond!");
299 	dnode = msg_destnode(buf_msg(skb));
300 	selector = msg_origport(buf_msg(skb));
301 	tipc_node_xmit_skb(sock_net(sk), skb, dnode, selector);
302 }
303 
304 /**
305  * tsk_rej_rx_queue - reject all buffers in socket receive queue
306  * @sk: network socket
307  * @error: response error code
308  *
309  * Caller must hold socket lock
310  */
311 static void tsk_rej_rx_queue(struct sock *sk, int error)
312 {
313 	struct sk_buff *skb;
314 
315 	while ((skb = __skb_dequeue(&sk->sk_receive_queue)))
316 		tipc_sk_respond(sk, skb, error);
317 }
318 
319 static bool tipc_sk_connected(struct sock *sk)
320 {
321 	return sk->sk_state == TIPC_ESTABLISHED;
322 }
323 
324 /* tipc_sk_type_connectionless - check if the socket is datagram socket
325  * @sk: socket
326  *
327  * Returns true if connection less, false otherwise
328  */
329 static bool tipc_sk_type_connectionless(struct sock *sk)
330 {
331 	return sk->sk_type == SOCK_RDM || sk->sk_type == SOCK_DGRAM;
332 }
333 
334 /* tsk_peer_msg - verify if message was sent by connected port's peer
335  *
336  * Handles cases where the node's network address has changed from
337  * the default of <0.0.0> to its configured setting.
338  */
339 static bool tsk_peer_msg(struct tipc_sock *tsk, struct tipc_msg *msg)
340 {
341 	struct sock *sk = &tsk->sk;
342 	u32 self = tipc_own_addr(sock_net(sk));
343 	u32 peer_port = tsk_peer_port(tsk);
344 	u32 orig_node, peer_node;
345 
346 	if (unlikely(!tipc_sk_connected(sk)))
347 		return false;
348 
349 	if (unlikely(msg_origport(msg) != peer_port))
350 		return false;
351 
352 	orig_node = msg_orignode(msg);
353 	peer_node = tsk_peer_node(tsk);
354 
355 	if (likely(orig_node == peer_node))
356 		return true;
357 
358 	if (!orig_node && peer_node == self)
359 		return true;
360 
361 	if (!peer_node && orig_node == self)
362 		return true;
363 
364 	return false;
365 }
366 
367 /* tipc_set_sk_state - set the sk_state of the socket
368  * @sk: socket
369  *
370  * Caller must hold socket lock
371  *
372  * Returns 0 on success, errno otherwise
373  */
374 static int tipc_set_sk_state(struct sock *sk, int state)
375 {
376 	int oldsk_state = sk->sk_state;
377 	int res = -EINVAL;
378 
379 	switch (state) {
380 	case TIPC_OPEN:
381 		res = 0;
382 		break;
383 	case TIPC_LISTEN:
384 	case TIPC_CONNECTING:
385 		if (oldsk_state == TIPC_OPEN)
386 			res = 0;
387 		break;
388 	case TIPC_ESTABLISHED:
389 		if (oldsk_state == TIPC_CONNECTING ||
390 		    oldsk_state == TIPC_OPEN)
391 			res = 0;
392 		break;
393 	case TIPC_DISCONNECTING:
394 		if (oldsk_state == TIPC_CONNECTING ||
395 		    oldsk_state == TIPC_ESTABLISHED)
396 			res = 0;
397 		break;
398 	}
399 
400 	if (!res)
401 		sk->sk_state = state;
402 
403 	return res;
404 }
405 
406 static int tipc_sk_sock_err(struct socket *sock, long *timeout)
407 {
408 	struct sock *sk = sock->sk;
409 	int err = sock_error(sk);
410 	int typ = sock->type;
411 
412 	if (err)
413 		return err;
414 	if (typ == SOCK_STREAM || typ == SOCK_SEQPACKET) {
415 		if (sk->sk_state == TIPC_DISCONNECTING)
416 			return -EPIPE;
417 		else if (!tipc_sk_connected(sk))
418 			return -ENOTCONN;
419 	}
420 	if (!*timeout)
421 		return -EAGAIN;
422 	if (signal_pending(current))
423 		return sock_intr_errno(*timeout);
424 
425 	return 0;
426 }
427 
428 #define tipc_wait_for_cond(sock_, timeo_, condition_)			       \
429 ({                                                                             \
430 	DEFINE_WAIT_FUNC(wait_, woken_wake_function);                          \
431 	struct sock *sk_;						       \
432 	int rc_;							       \
433 									       \
434 	while ((rc_ = !(condition_))) {					       \
435 		/* coupled with smp_wmb() in tipc_sk_proto_rcv() */            \
436 		smp_rmb();                                                     \
437 		sk_ = (sock_)->sk;					       \
438 		rc_ = tipc_sk_sock_err((sock_), timeo_);		       \
439 		if (rc_)						       \
440 			break;						       \
441 		add_wait_queue(sk_sleep(sk_), &wait_);                         \
442 		release_sock(sk_);					       \
443 		*(timeo_) = wait_woken(&wait_, TASK_INTERRUPTIBLE, *(timeo_)); \
444 		sched_annotate_sleep();				               \
445 		lock_sock(sk_);						       \
446 		remove_wait_queue(sk_sleep(sk_), &wait_);		       \
447 	}								       \
448 	rc_;								       \
449 })
450 
451 /**
452  * tipc_sk_create - create a TIPC socket
453  * @net: network namespace (must be default network)
454  * @sock: pre-allocated socket structure
455  * @protocol: protocol indicator (must be 0)
456  * @kern: caused by kernel or by userspace?
457  *
458  * This routine creates additional data structures used by the TIPC socket,
459  * initializes them, and links them together.
460  *
461  * Return: 0 on success, errno otherwise
462  */
463 static int tipc_sk_create(struct net *net, struct socket *sock,
464 			  int protocol, int kern)
465 {
466 	const struct proto_ops *ops;
467 	struct sock *sk;
468 	struct tipc_sock *tsk;
469 	struct tipc_msg *msg;
470 
471 	/* Validate arguments */
472 	if (unlikely(protocol != 0))
473 		return -EPROTONOSUPPORT;
474 
475 	switch (sock->type) {
476 	case SOCK_STREAM:
477 		ops = &stream_ops;
478 		break;
479 	case SOCK_SEQPACKET:
480 		ops = &packet_ops;
481 		break;
482 	case SOCK_DGRAM:
483 	case SOCK_RDM:
484 		ops = &msg_ops;
485 		break;
486 	default:
487 		return -EPROTOTYPE;
488 	}
489 
490 	/* Allocate socket's protocol area */
491 	sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto, kern);
492 	if (sk == NULL)
493 		return -ENOMEM;
494 
495 	tsk = tipc_sk(sk);
496 	tsk->max_pkt = MAX_PKT_DEFAULT;
497 	tsk->maxnagle = 0;
498 	tsk->nagle_start = NAGLE_START_INIT;
499 	INIT_LIST_HEAD(&tsk->publications);
500 	INIT_LIST_HEAD(&tsk->cong_links);
501 	msg = &tsk->phdr;
502 
503 	/* Finish initializing socket data structures */
504 	sock->ops = ops;
505 	sock_init_data(sock, sk);
506 	tipc_set_sk_state(sk, TIPC_OPEN);
507 	if (tipc_sk_insert(tsk)) {
508 		pr_warn("Socket create failed; port number exhausted\n");
509 		return -EINVAL;
510 	}
511 
512 	/* Ensure tsk is visible before we read own_addr. */
513 	smp_mb();
514 
515 	tipc_msg_init(tipc_own_addr(net), msg, TIPC_LOW_IMPORTANCE,
516 		      TIPC_NAMED_MSG, NAMED_H_SIZE, 0);
517 
518 	msg_set_origport(msg, tsk->portid);
519 	timer_setup(&sk->sk_timer, tipc_sk_timeout, 0);
520 	sk->sk_shutdown = 0;
521 	sk->sk_backlog_rcv = tipc_sk_backlog_rcv;
522 	sk->sk_rcvbuf = sysctl_tipc_rmem[1];
523 	sk->sk_data_ready = tipc_data_ready;
524 	sk->sk_write_space = tipc_write_space;
525 	sk->sk_destruct = tipc_sock_destruct;
526 	tsk->conn_timeout = CONN_TIMEOUT_DEFAULT;
527 	tsk->group_is_open = true;
528 	atomic_set(&tsk->dupl_rcvcnt, 0);
529 
530 	/* Start out with safe limits until we receive an advertised window */
531 	tsk->snd_win = tsk_adv_blocks(RCVBUF_MIN);
532 	tsk->rcv_win = tsk->snd_win;
533 
534 	if (tipc_sk_type_connectionless(sk)) {
535 		tsk_set_unreturnable(tsk, true);
536 		if (sock->type == SOCK_DGRAM)
537 			tsk_set_unreliable(tsk, true);
538 	}
539 	__skb_queue_head_init(&tsk->mc_method.deferredq);
540 	trace_tipc_sk_create(sk, NULL, TIPC_DUMP_NONE, " ");
541 	return 0;
542 }
543 
544 static void tipc_sk_callback(struct rcu_head *head)
545 {
546 	struct tipc_sock *tsk = container_of(head, struct tipc_sock, rcu);
547 
548 	sock_put(&tsk->sk);
549 }
550 
551 /* Caller should hold socket lock for the socket. */
552 static void __tipc_shutdown(struct socket *sock, int error)
553 {
554 	struct sock *sk = sock->sk;
555 	struct tipc_sock *tsk = tipc_sk(sk);
556 	struct net *net = sock_net(sk);
557 	long timeout = msecs_to_jiffies(CONN_TIMEOUT_DEFAULT);
558 	u32 dnode = tsk_peer_node(tsk);
559 	struct sk_buff *skb;
560 
561 	/* Avoid that hi-prio shutdown msgs bypass msgs in link wakeup queue */
562 	tipc_wait_for_cond(sock, &timeout, (!tsk->cong_link_cnt &&
563 					    !tsk_conn_cong(tsk)));
564 
565 	/* Push out delayed messages if in Nagle mode */
566 	tipc_sk_push_backlog(tsk, false);
567 	/* Remove pending SYN */
568 	__skb_queue_purge(&sk->sk_write_queue);
569 
570 	/* Remove partially received buffer if any */
571 	skb = skb_peek(&sk->sk_receive_queue);
572 	if (skb && TIPC_SKB_CB(skb)->bytes_read) {
573 		__skb_unlink(skb, &sk->sk_receive_queue);
574 		kfree_skb(skb);
575 	}
576 
577 	/* Reject all unreceived messages if connectionless */
578 	if (tipc_sk_type_connectionless(sk)) {
579 		tsk_rej_rx_queue(sk, error);
580 		return;
581 	}
582 
583 	switch (sk->sk_state) {
584 	case TIPC_CONNECTING:
585 	case TIPC_ESTABLISHED:
586 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
587 		tipc_node_remove_conn(net, dnode, tsk->portid);
588 		/* Send a FIN+/- to its peer */
589 		skb = __skb_dequeue(&sk->sk_receive_queue);
590 		if (skb) {
591 			__skb_queue_purge(&sk->sk_receive_queue);
592 			tipc_sk_respond(sk, skb, error);
593 			break;
594 		}
595 		skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE,
596 				      TIPC_CONN_MSG, SHORT_H_SIZE, 0, dnode,
597 				      tsk_own_node(tsk), tsk_peer_port(tsk),
598 				      tsk->portid, error);
599 		if (skb)
600 			tipc_node_xmit_skb(net, skb, dnode, tsk->portid);
601 		break;
602 	case TIPC_LISTEN:
603 		/* Reject all SYN messages */
604 		tsk_rej_rx_queue(sk, error);
605 		break;
606 	default:
607 		__skb_queue_purge(&sk->sk_receive_queue);
608 		break;
609 	}
610 }
611 
612 /**
613  * tipc_release - destroy a TIPC socket
614  * @sock: socket to destroy
615  *
616  * This routine cleans up any messages that are still queued on the socket.
617  * For DGRAM and RDM socket types, all queued messages are rejected.
618  * For SEQPACKET and STREAM socket types, the first message is rejected
619  * and any others are discarded.  (If the first message on a STREAM socket
620  * is partially-read, it is discarded and the next one is rejected instead.)
621  *
622  * NOTE: Rejected messages are not necessarily returned to the sender!  They
623  * are returned or discarded according to the "destination droppable" setting
624  * specified for the message by the sender.
625  *
626  * Return: 0 on success, errno otherwise
627  */
628 static int tipc_release(struct socket *sock)
629 {
630 	struct sock *sk = sock->sk;
631 	struct tipc_sock *tsk;
632 
633 	/*
634 	 * Exit if socket isn't fully initialized (occurs when a failed accept()
635 	 * releases a pre-allocated child socket that was never used)
636 	 */
637 	if (sk == NULL)
638 		return 0;
639 
640 	tsk = tipc_sk(sk);
641 	lock_sock(sk);
642 
643 	trace_tipc_sk_release(sk, NULL, TIPC_DUMP_ALL, " ");
644 	__tipc_shutdown(sock, TIPC_ERR_NO_PORT);
645 	sk->sk_shutdown = SHUTDOWN_MASK;
646 	tipc_sk_leave(tsk);
647 	tipc_sk_withdraw(tsk, 0, NULL);
648 	__skb_queue_purge(&tsk->mc_method.deferredq);
649 	sk_stop_timer(sk, &sk->sk_timer);
650 	tipc_sk_remove(tsk);
651 
652 	sock_orphan(sk);
653 	/* Reject any messages that accumulated in backlog queue */
654 	release_sock(sk);
655 	tipc_dest_list_purge(&tsk->cong_links);
656 	tsk->cong_link_cnt = 0;
657 	call_rcu(&tsk->rcu, tipc_sk_callback);
658 	sock->sk = NULL;
659 
660 	return 0;
661 }
662 
663 /**
664  * __tipc_bind - associate or disassocate TIPC name(s) with a socket
665  * @sock: socket structure
666  * @skaddr: socket address describing name(s) and desired operation
667  * @alen: size of socket address data structure
668  *
669  * Name and name sequence binding is indicated using a positive scope value;
670  * a negative scope value unbinds the specified name.  Specifying no name
671  * (i.e. a socket address length of 0) unbinds all names from the socket.
672  *
673  * Return: 0 on success, errno otherwise
674  *
675  * NOTE: This routine doesn't need to take the socket lock since it doesn't
676  *       access any non-constant socket information.
677  */
678 static int __tipc_bind(struct socket *sock, struct sockaddr *skaddr, int alen)
679 {
680 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)skaddr;
681 	struct tipc_sock *tsk = tipc_sk(sock->sk);
682 
683 	if (unlikely(!alen))
684 		return tipc_sk_withdraw(tsk, 0, NULL);
685 
686 	if (addr->addrtype == TIPC_SERVICE_ADDR)
687 		addr->addr.nameseq.upper = addr->addr.nameseq.lower;
688 
689 	if (tsk->group)
690 		return -EACCES;
691 
692 	if (addr->scope >= 0)
693 		return tipc_sk_publish(tsk, addr->scope, &addr->addr.nameseq);
694 	else
695 		return tipc_sk_withdraw(tsk, -addr->scope, &addr->addr.nameseq);
696 }
697 
698 int tipc_sk_bind(struct socket *sock, struct sockaddr *skaddr, int alen)
699 {
700 	int res;
701 
702 	lock_sock(sock->sk);
703 	res = __tipc_bind(sock, skaddr, alen);
704 	release_sock(sock->sk);
705 	return res;
706 }
707 
708 static int tipc_bind(struct socket *sock, struct sockaddr *skaddr, int alen)
709 {
710 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)skaddr;
711 
712 	if (alen) {
713 		if (alen < sizeof(struct sockaddr_tipc))
714 			return -EINVAL;
715 		if (addr->family != AF_TIPC)
716 			return -EAFNOSUPPORT;
717 		if (addr->addrtype > TIPC_SERVICE_ADDR)
718 			return -EAFNOSUPPORT;
719 		if (addr->addr.nameseq.type < TIPC_RESERVED_TYPES) {
720 			pr_warn_once("Can't bind to reserved service type %u\n",
721 				     addr->addr.nameseq.type);
722 			return -EACCES;
723 		}
724 	}
725 	return tipc_sk_bind(sock, skaddr, alen);
726 }
727 
728 /**
729  * tipc_getname - get port ID of socket or peer socket
730  * @sock: socket structure
731  * @uaddr: area for returned socket address
732  * @peer: 0 = own ID, 1 = current peer ID, 2 = current/former peer ID
733  *
734  * Return: 0 on success, errno otherwise
735  *
736  * NOTE: This routine doesn't need to take the socket lock since it only
737  *       accesses socket information that is unchanging (or which changes in
738  *       a completely predictable manner).
739  */
740 static int tipc_getname(struct socket *sock, struct sockaddr *uaddr,
741 			int peer)
742 {
743 	struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr;
744 	struct sock *sk = sock->sk;
745 	struct tipc_sock *tsk = tipc_sk(sk);
746 
747 	memset(addr, 0, sizeof(*addr));
748 	if (peer) {
749 		if ((!tipc_sk_connected(sk)) &&
750 		    ((peer != 2) || (sk->sk_state != TIPC_DISCONNECTING)))
751 			return -ENOTCONN;
752 		addr->addr.id.ref = tsk_peer_port(tsk);
753 		addr->addr.id.node = tsk_peer_node(tsk);
754 	} else {
755 		addr->addr.id.ref = tsk->portid;
756 		addr->addr.id.node = tipc_own_addr(sock_net(sk));
757 	}
758 
759 	addr->addrtype = TIPC_SOCKET_ADDR;
760 	addr->family = AF_TIPC;
761 	addr->scope = 0;
762 	addr->addr.name.domain = 0;
763 
764 	return sizeof(*addr);
765 }
766 
767 /**
768  * tipc_poll - read and possibly block on pollmask
769  * @file: file structure associated with the socket
770  * @sock: socket for which to calculate the poll bits
771  * @wait: ???
772  *
773  * Return: pollmask value
774  *
775  * COMMENTARY:
776  * It appears that the usual socket locking mechanisms are not useful here
777  * since the pollmask info is potentially out-of-date the moment this routine
778  * exits.  TCP and other protocols seem to rely on higher level poll routines
779  * to handle any preventable race conditions, so TIPC will do the same ...
780  *
781  * IMPORTANT: The fact that a read or write operation is indicated does NOT
782  * imply that the operation will succeed, merely that it should be performed
783  * and will not block.
784  */
785 static __poll_t tipc_poll(struct file *file, struct socket *sock,
786 			      poll_table *wait)
787 {
788 	struct sock *sk = sock->sk;
789 	struct tipc_sock *tsk = tipc_sk(sk);
790 	__poll_t revents = 0;
791 
792 	sock_poll_wait(file, sock, wait);
793 	trace_tipc_sk_poll(sk, NULL, TIPC_DUMP_ALL, " ");
794 
795 	if (sk->sk_shutdown & RCV_SHUTDOWN)
796 		revents |= EPOLLRDHUP | EPOLLIN | EPOLLRDNORM;
797 	if (sk->sk_shutdown == SHUTDOWN_MASK)
798 		revents |= EPOLLHUP;
799 
800 	switch (sk->sk_state) {
801 	case TIPC_ESTABLISHED:
802 		if (!tsk->cong_link_cnt && !tsk_conn_cong(tsk))
803 			revents |= EPOLLOUT;
804 		fallthrough;
805 	case TIPC_LISTEN:
806 	case TIPC_CONNECTING:
807 		if (!skb_queue_empty_lockless(&sk->sk_receive_queue))
808 			revents |= EPOLLIN | EPOLLRDNORM;
809 		break;
810 	case TIPC_OPEN:
811 		if (tsk->group_is_open && !tsk->cong_link_cnt)
812 			revents |= EPOLLOUT;
813 		if (!tipc_sk_type_connectionless(sk))
814 			break;
815 		if (skb_queue_empty_lockless(&sk->sk_receive_queue))
816 			break;
817 		revents |= EPOLLIN | EPOLLRDNORM;
818 		break;
819 	case TIPC_DISCONNECTING:
820 		revents = EPOLLIN | EPOLLRDNORM | EPOLLHUP;
821 		break;
822 	}
823 	return revents;
824 }
825 
826 /**
827  * tipc_sendmcast - send multicast message
828  * @sock: socket structure
829  * @seq: destination address
830  * @msg: message to send
831  * @dlen: length of data to send
832  * @timeout: timeout to wait for wakeup
833  *
834  * Called from function tipc_sendmsg(), which has done all sanity checks
835  * Return: the number of bytes sent on success, or errno
836  */
837 static int tipc_sendmcast(struct  socket *sock, struct tipc_service_range *seq,
838 			  struct msghdr *msg, size_t dlen, long timeout)
839 {
840 	struct sock *sk = sock->sk;
841 	struct tipc_sock *tsk = tipc_sk(sk);
842 	struct tipc_msg *hdr = &tsk->phdr;
843 	struct net *net = sock_net(sk);
844 	int mtu = tipc_bcast_get_mtu(net);
845 	struct tipc_mc_method *method = &tsk->mc_method;
846 	struct sk_buff_head pkts;
847 	struct tipc_nlist dsts;
848 	int rc;
849 
850 	if (tsk->group)
851 		return -EACCES;
852 
853 	/* Block or return if any destination link is congested */
854 	rc = tipc_wait_for_cond(sock, &timeout, !tsk->cong_link_cnt);
855 	if (unlikely(rc))
856 		return rc;
857 
858 	/* Lookup destination nodes */
859 	tipc_nlist_init(&dsts, tipc_own_addr(net));
860 	tipc_nametbl_lookup_dst_nodes(net, seq->type, seq->lower,
861 				      seq->upper, &dsts);
862 	if (!dsts.local && !dsts.remote)
863 		return -EHOSTUNREACH;
864 
865 	/* Build message header */
866 	msg_set_type(hdr, TIPC_MCAST_MSG);
867 	msg_set_hdr_sz(hdr, MCAST_H_SIZE);
868 	msg_set_lookup_scope(hdr, TIPC_CLUSTER_SCOPE);
869 	msg_set_destport(hdr, 0);
870 	msg_set_destnode(hdr, 0);
871 	msg_set_nametype(hdr, seq->type);
872 	msg_set_namelower(hdr, seq->lower);
873 	msg_set_nameupper(hdr, seq->upper);
874 
875 	/* Build message as chain of buffers */
876 	__skb_queue_head_init(&pkts);
877 	rc = tipc_msg_build(hdr, msg, 0, dlen, mtu, &pkts);
878 
879 	/* Send message if build was successful */
880 	if (unlikely(rc == dlen)) {
881 		trace_tipc_sk_sendmcast(sk, skb_peek(&pkts),
882 					TIPC_DUMP_SK_SNDQ, " ");
883 		rc = tipc_mcast_xmit(net, &pkts, method, &dsts,
884 				     &tsk->cong_link_cnt);
885 	}
886 
887 	tipc_nlist_purge(&dsts);
888 
889 	return rc ? rc : dlen;
890 }
891 
892 /**
893  * tipc_send_group_msg - send a message to a member in the group
894  * @net: network namespace
895  * @tsk: tipc socket
896  * @m: message to send
897  * @mb: group member
898  * @dnode: destination node
899  * @dport: destination port
900  * @dlen: total length of message data
901  */
902 static int tipc_send_group_msg(struct net *net, struct tipc_sock *tsk,
903 			       struct msghdr *m, struct tipc_member *mb,
904 			       u32 dnode, u32 dport, int dlen)
905 {
906 	u16 bc_snd_nxt = tipc_group_bc_snd_nxt(tsk->group);
907 	struct tipc_mc_method *method = &tsk->mc_method;
908 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
909 	struct tipc_msg *hdr = &tsk->phdr;
910 	struct sk_buff_head pkts;
911 	int mtu, rc;
912 
913 	/* Complete message header */
914 	msg_set_type(hdr, TIPC_GRP_UCAST_MSG);
915 	msg_set_hdr_sz(hdr, GROUP_H_SIZE);
916 	msg_set_destport(hdr, dport);
917 	msg_set_destnode(hdr, dnode);
918 	msg_set_grp_bc_seqno(hdr, bc_snd_nxt);
919 
920 	/* Build message as chain of buffers */
921 	__skb_queue_head_init(&pkts);
922 	mtu = tipc_node_get_mtu(net, dnode, tsk->portid, false);
923 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
924 	if (unlikely(rc != dlen))
925 		return rc;
926 
927 	/* Send message */
928 	rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
929 	if (unlikely(rc == -ELINKCONG)) {
930 		tipc_dest_push(&tsk->cong_links, dnode, 0);
931 		tsk->cong_link_cnt++;
932 	}
933 
934 	/* Update send window */
935 	tipc_group_update_member(mb, blks);
936 
937 	/* A broadcast sent within next EXPIRE period must follow same path */
938 	method->rcast = true;
939 	method->mandatory = true;
940 	return dlen;
941 }
942 
943 /**
944  * tipc_send_group_unicast - send message to a member in the group
945  * @sock: socket structure
946  * @m: message to send
947  * @dlen: total length of message data
948  * @timeout: timeout to wait for wakeup
949  *
950  * Called from function tipc_sendmsg(), which has done all sanity checks
951  * Return: the number of bytes sent on success, or errno
952  */
953 static int tipc_send_group_unicast(struct socket *sock, struct msghdr *m,
954 				   int dlen, long timeout)
955 {
956 	struct sock *sk = sock->sk;
957 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
958 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
959 	struct tipc_sock *tsk = tipc_sk(sk);
960 	struct net *net = sock_net(sk);
961 	struct tipc_member *mb = NULL;
962 	u32 node, port;
963 	int rc;
964 
965 	node = dest->addr.id.node;
966 	port = dest->addr.id.ref;
967 	if (!port && !node)
968 		return -EHOSTUNREACH;
969 
970 	/* Block or return if destination link or member is congested */
971 	rc = tipc_wait_for_cond(sock, &timeout,
972 				!tipc_dest_find(&tsk->cong_links, node, 0) &&
973 				tsk->group &&
974 				!tipc_group_cong(tsk->group, node, port, blks,
975 						 &mb));
976 	if (unlikely(rc))
977 		return rc;
978 
979 	if (unlikely(!mb))
980 		return -EHOSTUNREACH;
981 
982 	rc = tipc_send_group_msg(net, tsk, m, mb, node, port, dlen);
983 
984 	return rc ? rc : dlen;
985 }
986 
987 /**
988  * tipc_send_group_anycast - send message to any member with given identity
989  * @sock: socket structure
990  * @m: message to send
991  * @dlen: total length of message data
992  * @timeout: timeout to wait for wakeup
993  *
994  * Called from function tipc_sendmsg(), which has done all sanity checks
995  * Return: the number of bytes sent on success, or errno
996  */
997 static int tipc_send_group_anycast(struct socket *sock, struct msghdr *m,
998 				   int dlen, long timeout)
999 {
1000 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1001 	struct sock *sk = sock->sk;
1002 	struct tipc_sock *tsk = tipc_sk(sk);
1003 	struct list_head *cong_links = &tsk->cong_links;
1004 	int blks = tsk_blocks(GROUP_H_SIZE + dlen);
1005 	struct tipc_msg *hdr = &tsk->phdr;
1006 	struct tipc_member *first = NULL;
1007 	struct tipc_member *mbr = NULL;
1008 	struct net *net = sock_net(sk);
1009 	u32 node, port, exclude;
1010 	struct list_head dsts;
1011 	u32 type, inst, scope;
1012 	int lookups = 0;
1013 	int dstcnt, rc;
1014 	bool cong;
1015 
1016 	INIT_LIST_HEAD(&dsts);
1017 
1018 	type = msg_nametype(hdr);
1019 	inst = dest->addr.name.name.instance;
1020 	scope = msg_lookup_scope(hdr);
1021 
1022 	while (++lookups < 4) {
1023 		exclude = tipc_group_exclude(tsk->group);
1024 
1025 		first = NULL;
1026 
1027 		/* Look for a non-congested destination member, if any */
1028 		while (1) {
1029 			if (!tipc_nametbl_lookup(net, type, inst, scope, &dsts,
1030 						 &dstcnt, exclude, false))
1031 				return -EHOSTUNREACH;
1032 			tipc_dest_pop(&dsts, &node, &port);
1033 			cong = tipc_group_cong(tsk->group, node, port, blks,
1034 					       &mbr);
1035 			if (!cong)
1036 				break;
1037 			if (mbr == first)
1038 				break;
1039 			if (!first)
1040 				first = mbr;
1041 		}
1042 
1043 		/* Start over if destination was not in member list */
1044 		if (unlikely(!mbr))
1045 			continue;
1046 
1047 		if (likely(!cong && !tipc_dest_find(cong_links, node, 0)))
1048 			break;
1049 
1050 		/* Block or return if destination link or member is congested */
1051 		rc = tipc_wait_for_cond(sock, &timeout,
1052 					!tipc_dest_find(cong_links, node, 0) &&
1053 					tsk->group &&
1054 					!tipc_group_cong(tsk->group, node, port,
1055 							 blks, &mbr));
1056 		if (unlikely(rc))
1057 			return rc;
1058 
1059 		/* Send, unless destination disappeared while waiting */
1060 		if (likely(mbr))
1061 			break;
1062 	}
1063 
1064 	if (unlikely(lookups >= 4))
1065 		return -EHOSTUNREACH;
1066 
1067 	rc = tipc_send_group_msg(net, tsk, m, mbr, node, port, dlen);
1068 
1069 	return rc ? rc : dlen;
1070 }
1071 
1072 /**
1073  * tipc_send_group_bcast - send message to all members in communication group
1074  * @sock: socket structure
1075  * @m: message to send
1076  * @dlen: total length of message data
1077  * @timeout: timeout to wait for wakeup
1078  *
1079  * Called from function tipc_sendmsg(), which has done all sanity checks
1080  * Return: the number of bytes sent on success, or errno
1081  */
1082 static int tipc_send_group_bcast(struct socket *sock, struct msghdr *m,
1083 				 int dlen, long timeout)
1084 {
1085 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1086 	struct sock *sk = sock->sk;
1087 	struct net *net = sock_net(sk);
1088 	struct tipc_sock *tsk = tipc_sk(sk);
1089 	struct tipc_nlist *dsts;
1090 	struct tipc_mc_method *method = &tsk->mc_method;
1091 	bool ack = method->mandatory && method->rcast;
1092 	int blks = tsk_blocks(MCAST_H_SIZE + dlen);
1093 	struct tipc_msg *hdr = &tsk->phdr;
1094 	int mtu = tipc_bcast_get_mtu(net);
1095 	struct sk_buff_head pkts;
1096 	int rc = -EHOSTUNREACH;
1097 
1098 	/* Block or return if any destination link or member is congested */
1099 	rc = tipc_wait_for_cond(sock, &timeout,
1100 				!tsk->cong_link_cnt && tsk->group &&
1101 				!tipc_group_bc_cong(tsk->group, blks));
1102 	if (unlikely(rc))
1103 		return rc;
1104 
1105 	dsts = tipc_group_dests(tsk->group);
1106 	if (!dsts->local && !dsts->remote)
1107 		return -EHOSTUNREACH;
1108 
1109 	/* Complete message header */
1110 	if (dest) {
1111 		msg_set_type(hdr, TIPC_GRP_MCAST_MSG);
1112 		msg_set_nameinst(hdr, dest->addr.name.name.instance);
1113 	} else {
1114 		msg_set_type(hdr, TIPC_GRP_BCAST_MSG);
1115 		msg_set_nameinst(hdr, 0);
1116 	}
1117 	msg_set_hdr_sz(hdr, GROUP_H_SIZE);
1118 	msg_set_destport(hdr, 0);
1119 	msg_set_destnode(hdr, 0);
1120 	msg_set_grp_bc_seqno(hdr, tipc_group_bc_snd_nxt(tsk->group));
1121 
1122 	/* Avoid getting stuck with repeated forced replicasts */
1123 	msg_set_grp_bc_ack_req(hdr, ack);
1124 
1125 	/* Build message as chain of buffers */
1126 	__skb_queue_head_init(&pkts);
1127 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
1128 	if (unlikely(rc != dlen))
1129 		return rc;
1130 
1131 	/* Send message */
1132 	rc = tipc_mcast_xmit(net, &pkts, method, dsts, &tsk->cong_link_cnt);
1133 	if (unlikely(rc))
1134 		return rc;
1135 
1136 	/* Update broadcast sequence number and send windows */
1137 	tipc_group_update_bc_members(tsk->group, blks, ack);
1138 
1139 	/* Broadcast link is now free to choose method for next broadcast */
1140 	method->mandatory = false;
1141 	method->expires = jiffies;
1142 
1143 	return dlen;
1144 }
1145 
1146 /**
1147  * tipc_send_group_mcast - send message to all members with given identity
1148  * @sock: socket structure
1149  * @m: message to send
1150  * @dlen: total length of message data
1151  * @timeout: timeout to wait for wakeup
1152  *
1153  * Called from function tipc_sendmsg(), which has done all sanity checks
1154  * Return: the number of bytes sent on success, or errno
1155  */
1156 static int tipc_send_group_mcast(struct socket *sock, struct msghdr *m,
1157 				 int dlen, long timeout)
1158 {
1159 	struct sock *sk = sock->sk;
1160 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1161 	struct tipc_sock *tsk = tipc_sk(sk);
1162 	struct tipc_group *grp = tsk->group;
1163 	struct tipc_msg *hdr = &tsk->phdr;
1164 	struct net *net = sock_net(sk);
1165 	u32 type, inst, scope, exclude;
1166 	struct list_head dsts;
1167 	u32 dstcnt;
1168 
1169 	INIT_LIST_HEAD(&dsts);
1170 
1171 	type = msg_nametype(hdr);
1172 	inst = dest->addr.name.name.instance;
1173 	scope = msg_lookup_scope(hdr);
1174 	exclude = tipc_group_exclude(grp);
1175 
1176 	if (!tipc_nametbl_lookup(net, type, inst, scope, &dsts,
1177 				 &dstcnt, exclude, true))
1178 		return -EHOSTUNREACH;
1179 
1180 	if (dstcnt == 1) {
1181 		tipc_dest_pop(&dsts, &dest->addr.id.node, &dest->addr.id.ref);
1182 		return tipc_send_group_unicast(sock, m, dlen, timeout);
1183 	}
1184 
1185 	tipc_dest_list_purge(&dsts);
1186 	return tipc_send_group_bcast(sock, m, dlen, timeout);
1187 }
1188 
1189 /**
1190  * tipc_sk_mcast_rcv - Deliver multicast messages to all destination sockets
1191  * @net: the associated network namespace
1192  * @arrvq: queue with arriving messages, to be cloned after destination lookup
1193  * @inputq: queue with cloned messages, delivered to socket after dest lookup
1194  *
1195  * Multi-threaded: parallel calls with reference to same queues may occur
1196  */
1197 void tipc_sk_mcast_rcv(struct net *net, struct sk_buff_head *arrvq,
1198 		       struct sk_buff_head *inputq)
1199 {
1200 	u32 self = tipc_own_addr(net);
1201 	u32 type, lower, upper, scope;
1202 	struct sk_buff *skb, *_skb;
1203 	u32 portid, onode;
1204 	struct sk_buff_head tmpq;
1205 	struct list_head dports;
1206 	struct tipc_msg *hdr;
1207 	int user, mtyp, hlen;
1208 	bool exact;
1209 
1210 	__skb_queue_head_init(&tmpq);
1211 	INIT_LIST_HEAD(&dports);
1212 
1213 	skb = tipc_skb_peek(arrvq, &inputq->lock);
1214 	for (; skb; skb = tipc_skb_peek(arrvq, &inputq->lock)) {
1215 		hdr = buf_msg(skb);
1216 		user = msg_user(hdr);
1217 		mtyp = msg_type(hdr);
1218 		hlen = skb_headroom(skb) + msg_hdr_sz(hdr);
1219 		onode = msg_orignode(hdr);
1220 		type = msg_nametype(hdr);
1221 
1222 		if (mtyp == TIPC_GRP_UCAST_MSG || user == GROUP_PROTOCOL) {
1223 			spin_lock_bh(&inputq->lock);
1224 			if (skb_peek(arrvq) == skb) {
1225 				__skb_dequeue(arrvq);
1226 				__skb_queue_tail(inputq, skb);
1227 			}
1228 			kfree_skb(skb);
1229 			spin_unlock_bh(&inputq->lock);
1230 			continue;
1231 		}
1232 
1233 		/* Group messages require exact scope match */
1234 		if (msg_in_group(hdr)) {
1235 			lower = 0;
1236 			upper = ~0;
1237 			scope = msg_lookup_scope(hdr);
1238 			exact = true;
1239 		} else {
1240 			/* TIPC_NODE_SCOPE means "any scope" in this context */
1241 			if (onode == self)
1242 				scope = TIPC_NODE_SCOPE;
1243 			else
1244 				scope = TIPC_CLUSTER_SCOPE;
1245 			exact = false;
1246 			lower = msg_namelower(hdr);
1247 			upper = msg_nameupper(hdr);
1248 		}
1249 
1250 		/* Create destination port list: */
1251 		tipc_nametbl_mc_lookup(net, type, lower, upper,
1252 				       scope, exact, &dports);
1253 
1254 		/* Clone message per destination */
1255 		while (tipc_dest_pop(&dports, NULL, &portid)) {
1256 			_skb = __pskb_copy(skb, hlen, GFP_ATOMIC);
1257 			if (_skb) {
1258 				msg_set_destport(buf_msg(_skb), portid);
1259 				__skb_queue_tail(&tmpq, _skb);
1260 				continue;
1261 			}
1262 			pr_warn("Failed to clone mcast rcv buffer\n");
1263 		}
1264 		/* Append to inputq if not already done by other thread */
1265 		spin_lock_bh(&inputq->lock);
1266 		if (skb_peek(arrvq) == skb) {
1267 			skb_queue_splice_tail_init(&tmpq, inputq);
1268 			kfree_skb(__skb_dequeue(arrvq));
1269 		}
1270 		spin_unlock_bh(&inputq->lock);
1271 		__skb_queue_purge(&tmpq);
1272 		kfree_skb(skb);
1273 	}
1274 	tipc_sk_rcv(net, inputq);
1275 }
1276 
1277 /* tipc_sk_push_backlog(): send accumulated buffers in socket write queue
1278  *                         when socket is in Nagle mode
1279  */
1280 static void tipc_sk_push_backlog(struct tipc_sock *tsk, bool nagle_ack)
1281 {
1282 	struct sk_buff_head *txq = &tsk->sk.sk_write_queue;
1283 	struct sk_buff *skb = skb_peek_tail(txq);
1284 	struct net *net = sock_net(&tsk->sk);
1285 	u32 dnode = tsk_peer_node(tsk);
1286 	int rc;
1287 
1288 	if (nagle_ack) {
1289 		tsk->pkt_cnt += skb_queue_len(txq);
1290 		if (!tsk->pkt_cnt || tsk->msg_acc / tsk->pkt_cnt < 2) {
1291 			tsk->oneway = 0;
1292 			if (tsk->nagle_start < NAGLE_START_MAX)
1293 				tsk->nagle_start *= 2;
1294 			tsk->expect_ack = false;
1295 			pr_debug("tsk %10u: bad nagle %u -> %u, next start %u!\n",
1296 				 tsk->portid, tsk->msg_acc, tsk->pkt_cnt,
1297 				 tsk->nagle_start);
1298 		} else {
1299 			tsk->nagle_start = NAGLE_START_INIT;
1300 			if (skb) {
1301 				msg_set_ack_required(buf_msg(skb));
1302 				tsk->expect_ack = true;
1303 			} else {
1304 				tsk->expect_ack = false;
1305 			}
1306 		}
1307 		tsk->msg_acc = 0;
1308 		tsk->pkt_cnt = 0;
1309 	}
1310 
1311 	if (!skb || tsk->cong_link_cnt)
1312 		return;
1313 
1314 	/* Do not send SYN again after congestion */
1315 	if (msg_is_syn(buf_msg(skb)))
1316 		return;
1317 
1318 	if (tsk->msg_acc)
1319 		tsk->pkt_cnt += skb_queue_len(txq);
1320 	tsk->snt_unacked += tsk->snd_backlog;
1321 	tsk->snd_backlog = 0;
1322 	rc = tipc_node_xmit(net, txq, dnode, tsk->portid);
1323 	if (rc == -ELINKCONG)
1324 		tsk->cong_link_cnt = 1;
1325 }
1326 
1327 /**
1328  * tipc_sk_conn_proto_rcv - receive a connection mng protocol message
1329  * @tsk: receiving socket
1330  * @skb: pointer to message buffer.
1331  * @inputq: buffer list containing the buffers
1332  * @xmitq: output message area
1333  */
1334 static void tipc_sk_conn_proto_rcv(struct tipc_sock *tsk, struct sk_buff *skb,
1335 				   struct sk_buff_head *inputq,
1336 				   struct sk_buff_head *xmitq)
1337 {
1338 	struct tipc_msg *hdr = buf_msg(skb);
1339 	u32 onode = tsk_own_node(tsk);
1340 	struct sock *sk = &tsk->sk;
1341 	int mtyp = msg_type(hdr);
1342 	bool was_cong;
1343 
1344 	/* Ignore if connection cannot be validated: */
1345 	if (!tsk_peer_msg(tsk, hdr)) {
1346 		trace_tipc_sk_drop_msg(sk, skb, TIPC_DUMP_NONE, "@proto_rcv!");
1347 		goto exit;
1348 	}
1349 
1350 	if (unlikely(msg_errcode(hdr))) {
1351 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
1352 		tipc_node_remove_conn(sock_net(sk), tsk_peer_node(tsk),
1353 				      tsk_peer_port(tsk));
1354 		sk->sk_state_change(sk);
1355 
1356 		/* State change is ignored if socket already awake,
1357 		 * - convert msg to abort msg and add to inqueue
1358 		 */
1359 		msg_set_user(hdr, TIPC_CRITICAL_IMPORTANCE);
1360 		msg_set_type(hdr, TIPC_CONN_MSG);
1361 		msg_set_size(hdr, BASIC_H_SIZE);
1362 		msg_set_hdr_sz(hdr, BASIC_H_SIZE);
1363 		__skb_queue_tail(inputq, skb);
1364 		return;
1365 	}
1366 
1367 	tsk->probe_unacked = false;
1368 
1369 	if (mtyp == CONN_PROBE) {
1370 		msg_set_type(hdr, CONN_PROBE_REPLY);
1371 		if (tipc_msg_reverse(onode, &skb, TIPC_OK))
1372 			__skb_queue_tail(xmitq, skb);
1373 		return;
1374 	} else if (mtyp == CONN_ACK) {
1375 		was_cong = tsk_conn_cong(tsk);
1376 		tipc_sk_push_backlog(tsk, msg_nagle_ack(hdr));
1377 		tsk->snt_unacked -= msg_conn_ack(hdr);
1378 		if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL)
1379 			tsk->snd_win = msg_adv_win(hdr);
1380 		if (was_cong && !tsk_conn_cong(tsk))
1381 			sk->sk_write_space(sk);
1382 	} else if (mtyp != CONN_PROBE_REPLY) {
1383 		pr_warn("Received unknown CONN_PROTO msg\n");
1384 	}
1385 exit:
1386 	kfree_skb(skb);
1387 }
1388 
1389 /**
1390  * tipc_sendmsg - send message in connectionless manner
1391  * @sock: socket structure
1392  * @m: message to send
1393  * @dsz: amount of user data to be sent
1394  *
1395  * Message must have an destination specified explicitly.
1396  * Used for SOCK_RDM and SOCK_DGRAM messages,
1397  * and for 'SYN' messages on SOCK_SEQPACKET and SOCK_STREAM connections.
1398  * (Note: 'SYN+' is prohibited on SOCK_STREAM.)
1399  *
1400  * Return: the number of bytes sent on success, or errno otherwise
1401  */
1402 static int tipc_sendmsg(struct socket *sock,
1403 			struct msghdr *m, size_t dsz)
1404 {
1405 	struct sock *sk = sock->sk;
1406 	int ret;
1407 
1408 	lock_sock(sk);
1409 	ret = __tipc_sendmsg(sock, m, dsz);
1410 	release_sock(sk);
1411 
1412 	return ret;
1413 }
1414 
1415 static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dlen)
1416 {
1417 	struct sock *sk = sock->sk;
1418 	struct net *net = sock_net(sk);
1419 	struct tipc_sock *tsk = tipc_sk(sk);
1420 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1421 	long timeout = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
1422 	struct list_head *clinks = &tsk->cong_links;
1423 	bool syn = !tipc_sk_type_connectionless(sk);
1424 	struct tipc_group *grp = tsk->group;
1425 	struct tipc_msg *hdr = &tsk->phdr;
1426 	struct tipc_service_range *seq;
1427 	struct sk_buff_head pkts;
1428 	u32 dport = 0, dnode = 0;
1429 	u32 type = 0, inst = 0;
1430 	int mtu, rc;
1431 
1432 	if (unlikely(dlen > TIPC_MAX_USER_MSG_SIZE))
1433 		return -EMSGSIZE;
1434 
1435 	if (likely(dest)) {
1436 		if (unlikely(m->msg_namelen < sizeof(*dest)))
1437 			return -EINVAL;
1438 		if (unlikely(dest->family != AF_TIPC))
1439 			return -EINVAL;
1440 	}
1441 
1442 	if (grp) {
1443 		if (!dest)
1444 			return tipc_send_group_bcast(sock, m, dlen, timeout);
1445 		if (dest->addrtype == TIPC_SERVICE_ADDR)
1446 			return tipc_send_group_anycast(sock, m, dlen, timeout);
1447 		if (dest->addrtype == TIPC_SOCKET_ADDR)
1448 			return tipc_send_group_unicast(sock, m, dlen, timeout);
1449 		if (dest->addrtype == TIPC_ADDR_MCAST)
1450 			return tipc_send_group_mcast(sock, m, dlen, timeout);
1451 		return -EINVAL;
1452 	}
1453 
1454 	if (unlikely(!dest)) {
1455 		dest = &tsk->peer;
1456 		if (!syn && dest->family != AF_TIPC)
1457 			return -EDESTADDRREQ;
1458 	}
1459 
1460 	if (unlikely(syn)) {
1461 		if (sk->sk_state == TIPC_LISTEN)
1462 			return -EPIPE;
1463 		if (sk->sk_state != TIPC_OPEN)
1464 			return -EISCONN;
1465 		if (tsk->published)
1466 			return -EOPNOTSUPP;
1467 		if (dest->addrtype == TIPC_SERVICE_ADDR) {
1468 			tsk->conn_type = dest->addr.name.name.type;
1469 			tsk->conn_instance = dest->addr.name.name.instance;
1470 		}
1471 		msg_set_syn(hdr, 1);
1472 	}
1473 
1474 	seq = &dest->addr.nameseq;
1475 	if (dest->addrtype == TIPC_ADDR_MCAST)
1476 		return tipc_sendmcast(sock, seq, m, dlen, timeout);
1477 
1478 	if (dest->addrtype == TIPC_SERVICE_ADDR) {
1479 		type = dest->addr.name.name.type;
1480 		inst = dest->addr.name.name.instance;
1481 		dnode = dest->addr.name.domain;
1482 		dport = tipc_nametbl_translate(net, type, inst, &dnode);
1483 		if (unlikely(!dport && !dnode))
1484 			return -EHOSTUNREACH;
1485 	} else if (dest->addrtype == TIPC_SOCKET_ADDR) {
1486 		dnode = dest->addr.id.node;
1487 	} else {
1488 		return -EINVAL;
1489 	}
1490 
1491 	/* Block or return if destination link is congested */
1492 	rc = tipc_wait_for_cond(sock, &timeout,
1493 				!tipc_dest_find(clinks, dnode, 0));
1494 	if (unlikely(rc))
1495 		return rc;
1496 
1497 	if (dest->addrtype == TIPC_SERVICE_ADDR) {
1498 		msg_set_type(hdr, TIPC_NAMED_MSG);
1499 		msg_set_hdr_sz(hdr, NAMED_H_SIZE);
1500 		msg_set_nametype(hdr, type);
1501 		msg_set_nameinst(hdr, inst);
1502 		msg_set_lookup_scope(hdr, tipc_node2scope(dnode));
1503 		msg_set_destnode(hdr, dnode);
1504 		msg_set_destport(hdr, dport);
1505 	} else { /* TIPC_SOCKET_ADDR */
1506 		msg_set_type(hdr, TIPC_DIRECT_MSG);
1507 		msg_set_lookup_scope(hdr, 0);
1508 		msg_set_destnode(hdr, dnode);
1509 		msg_set_destport(hdr, dest->addr.id.ref);
1510 		msg_set_hdr_sz(hdr, BASIC_H_SIZE);
1511 	}
1512 
1513 	__skb_queue_head_init(&pkts);
1514 	mtu = tipc_node_get_mtu(net, dnode, tsk->portid, true);
1515 	rc = tipc_msg_build(hdr, m, 0, dlen, mtu, &pkts);
1516 	if (unlikely(rc != dlen))
1517 		return rc;
1518 	if (unlikely(syn && !tipc_msg_skb_clone(&pkts, &sk->sk_write_queue))) {
1519 		__skb_queue_purge(&pkts);
1520 		return -ENOMEM;
1521 	}
1522 
1523 	trace_tipc_sk_sendmsg(sk, skb_peek(&pkts), TIPC_DUMP_SK_SNDQ, " ");
1524 	rc = tipc_node_xmit(net, &pkts, dnode, tsk->portid);
1525 	if (unlikely(rc == -ELINKCONG)) {
1526 		tipc_dest_push(clinks, dnode, 0);
1527 		tsk->cong_link_cnt++;
1528 		rc = 0;
1529 	}
1530 
1531 	if (unlikely(syn && !rc))
1532 		tipc_set_sk_state(sk, TIPC_CONNECTING);
1533 
1534 	return rc ? rc : dlen;
1535 }
1536 
1537 /**
1538  * tipc_sendstream - send stream-oriented data
1539  * @sock: socket structure
1540  * @m: data to send
1541  * @dsz: total length of data to be transmitted
1542  *
1543  * Used for SOCK_STREAM data.
1544  *
1545  * Return: the number of bytes sent on success (or partial success),
1546  * or errno if no data sent
1547  */
1548 static int tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dsz)
1549 {
1550 	struct sock *sk = sock->sk;
1551 	int ret;
1552 
1553 	lock_sock(sk);
1554 	ret = __tipc_sendstream(sock, m, dsz);
1555 	release_sock(sk);
1556 
1557 	return ret;
1558 }
1559 
1560 static int __tipc_sendstream(struct socket *sock, struct msghdr *m, size_t dlen)
1561 {
1562 	struct sock *sk = sock->sk;
1563 	DECLARE_SOCKADDR(struct sockaddr_tipc *, dest, m->msg_name);
1564 	long timeout = sock_sndtimeo(sk, m->msg_flags & MSG_DONTWAIT);
1565 	struct sk_buff_head *txq = &sk->sk_write_queue;
1566 	struct tipc_sock *tsk = tipc_sk(sk);
1567 	struct tipc_msg *hdr = &tsk->phdr;
1568 	struct net *net = sock_net(sk);
1569 	struct sk_buff *skb;
1570 	u32 dnode = tsk_peer_node(tsk);
1571 	int maxnagle = tsk->maxnagle;
1572 	int maxpkt = tsk->max_pkt;
1573 	int send, sent = 0;
1574 	int blocks, rc = 0;
1575 
1576 	if (unlikely(dlen > INT_MAX))
1577 		return -EMSGSIZE;
1578 
1579 	/* Handle implicit connection setup */
1580 	if (unlikely(dest)) {
1581 		rc = __tipc_sendmsg(sock, m, dlen);
1582 		if (dlen && dlen == rc) {
1583 			tsk->peer_caps = tipc_node_get_capabilities(net, dnode);
1584 			tsk->snt_unacked = tsk_inc(tsk, dlen + msg_hdr_sz(hdr));
1585 		}
1586 		return rc;
1587 	}
1588 
1589 	do {
1590 		rc = tipc_wait_for_cond(sock, &timeout,
1591 					(!tsk->cong_link_cnt &&
1592 					 !tsk_conn_cong(tsk) &&
1593 					 tipc_sk_connected(sk)));
1594 		if (unlikely(rc))
1595 			break;
1596 		send = min_t(size_t, dlen - sent, TIPC_MAX_USER_MSG_SIZE);
1597 		blocks = tsk->snd_backlog;
1598 		if (tsk->oneway++ >= tsk->nagle_start && maxnagle &&
1599 		    send <= maxnagle) {
1600 			rc = tipc_msg_append(hdr, m, send, maxnagle, txq);
1601 			if (unlikely(rc < 0))
1602 				break;
1603 			blocks += rc;
1604 			tsk->msg_acc++;
1605 			if (blocks <= 64 && tsk->expect_ack) {
1606 				tsk->snd_backlog = blocks;
1607 				sent += send;
1608 				break;
1609 			} else if (blocks > 64) {
1610 				tsk->pkt_cnt += skb_queue_len(txq);
1611 			} else {
1612 				skb = skb_peek_tail(txq);
1613 				if (skb) {
1614 					msg_set_ack_required(buf_msg(skb));
1615 					tsk->expect_ack = true;
1616 				} else {
1617 					tsk->expect_ack = false;
1618 				}
1619 				tsk->msg_acc = 0;
1620 				tsk->pkt_cnt = 0;
1621 			}
1622 		} else {
1623 			rc = tipc_msg_build(hdr, m, sent, send, maxpkt, txq);
1624 			if (unlikely(rc != send))
1625 				break;
1626 			blocks += tsk_inc(tsk, send + MIN_H_SIZE);
1627 		}
1628 		trace_tipc_sk_sendstream(sk, skb_peek(txq),
1629 					 TIPC_DUMP_SK_SNDQ, " ");
1630 		rc = tipc_node_xmit(net, txq, dnode, tsk->portid);
1631 		if (unlikely(rc == -ELINKCONG)) {
1632 			tsk->cong_link_cnt = 1;
1633 			rc = 0;
1634 		}
1635 		if (likely(!rc)) {
1636 			tsk->snt_unacked += blocks;
1637 			tsk->snd_backlog = 0;
1638 			sent += send;
1639 		}
1640 	} while (sent < dlen && !rc);
1641 
1642 	return sent ? sent : rc;
1643 }
1644 
1645 /**
1646  * tipc_send_packet - send a connection-oriented message
1647  * @sock: socket structure
1648  * @m: message to send
1649  * @dsz: length of data to be transmitted
1650  *
1651  * Used for SOCK_SEQPACKET messages.
1652  *
1653  * Return: the number of bytes sent on success, or errno otherwise
1654  */
1655 static int tipc_send_packet(struct socket *sock, struct msghdr *m, size_t dsz)
1656 {
1657 	if (dsz > TIPC_MAX_USER_MSG_SIZE)
1658 		return -EMSGSIZE;
1659 
1660 	return tipc_sendstream(sock, m, dsz);
1661 }
1662 
1663 /* tipc_sk_finish_conn - complete the setup of a connection
1664  */
1665 static void tipc_sk_finish_conn(struct tipc_sock *tsk, u32 peer_port,
1666 				u32 peer_node)
1667 {
1668 	struct sock *sk = &tsk->sk;
1669 	struct net *net = sock_net(sk);
1670 	struct tipc_msg *msg = &tsk->phdr;
1671 
1672 	msg_set_syn(msg, 0);
1673 	msg_set_destnode(msg, peer_node);
1674 	msg_set_destport(msg, peer_port);
1675 	msg_set_type(msg, TIPC_CONN_MSG);
1676 	msg_set_lookup_scope(msg, 0);
1677 	msg_set_hdr_sz(msg, SHORT_H_SIZE);
1678 
1679 	sk_reset_timer(sk, &sk->sk_timer, jiffies + CONN_PROBING_INTV);
1680 	tipc_set_sk_state(sk, TIPC_ESTABLISHED);
1681 	tipc_node_add_conn(net, peer_node, tsk->portid, peer_port);
1682 	tsk->max_pkt = tipc_node_get_mtu(net, peer_node, tsk->portid, true);
1683 	tsk->peer_caps = tipc_node_get_capabilities(net, peer_node);
1684 	tsk_set_nagle(tsk);
1685 	__skb_queue_purge(&sk->sk_write_queue);
1686 	if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL)
1687 		return;
1688 
1689 	/* Fall back to message based flow control */
1690 	tsk->rcv_win = FLOWCTL_MSG_WIN;
1691 	tsk->snd_win = FLOWCTL_MSG_WIN;
1692 }
1693 
1694 /**
1695  * tipc_sk_set_orig_addr - capture sender's address for received message
1696  * @m: descriptor for message info
1697  * @skb: received message
1698  *
1699  * Note: Address is not captured if not requested by receiver.
1700  */
1701 static void tipc_sk_set_orig_addr(struct msghdr *m, struct sk_buff *skb)
1702 {
1703 	DECLARE_SOCKADDR(struct sockaddr_pair *, srcaddr, m->msg_name);
1704 	struct tipc_msg *hdr = buf_msg(skb);
1705 
1706 	if (!srcaddr)
1707 		return;
1708 
1709 	srcaddr->sock.family = AF_TIPC;
1710 	srcaddr->sock.addrtype = TIPC_SOCKET_ADDR;
1711 	srcaddr->sock.scope = 0;
1712 	srcaddr->sock.addr.id.ref = msg_origport(hdr);
1713 	srcaddr->sock.addr.id.node = msg_orignode(hdr);
1714 	srcaddr->sock.addr.name.domain = 0;
1715 	m->msg_namelen = sizeof(struct sockaddr_tipc);
1716 
1717 	if (!msg_in_group(hdr))
1718 		return;
1719 
1720 	/* Group message users may also want to know sending member's id */
1721 	srcaddr->member.family = AF_TIPC;
1722 	srcaddr->member.addrtype = TIPC_SERVICE_ADDR;
1723 	srcaddr->member.scope = 0;
1724 	srcaddr->member.addr.name.name.type = msg_nametype(hdr);
1725 	srcaddr->member.addr.name.name.instance = TIPC_SKB_CB(skb)->orig_member;
1726 	srcaddr->member.addr.name.domain = 0;
1727 	m->msg_namelen = sizeof(*srcaddr);
1728 }
1729 
1730 /**
1731  * tipc_sk_anc_data_recv - optionally capture ancillary data for received message
1732  * @m: descriptor for message info
1733  * @skb: received message buffer
1734  * @tsk: TIPC port associated with message
1735  *
1736  * Note: Ancillary data is not captured if not requested by receiver.
1737  *
1738  * Return: 0 if successful, otherwise errno
1739  */
1740 static int tipc_sk_anc_data_recv(struct msghdr *m, struct sk_buff *skb,
1741 				 struct tipc_sock *tsk)
1742 {
1743 	struct tipc_msg *msg;
1744 	u32 anc_data[3];
1745 	u32 err;
1746 	u32 dest_type;
1747 	int has_name;
1748 	int res;
1749 
1750 	if (likely(m->msg_controllen == 0))
1751 		return 0;
1752 	msg = buf_msg(skb);
1753 
1754 	/* Optionally capture errored message object(s) */
1755 	err = msg ? msg_errcode(msg) : 0;
1756 	if (unlikely(err)) {
1757 		anc_data[0] = err;
1758 		anc_data[1] = msg_data_sz(msg);
1759 		res = put_cmsg(m, SOL_TIPC, TIPC_ERRINFO, 8, anc_data);
1760 		if (res)
1761 			return res;
1762 		if (anc_data[1]) {
1763 			if (skb_linearize(skb))
1764 				return -ENOMEM;
1765 			msg = buf_msg(skb);
1766 			res = put_cmsg(m, SOL_TIPC, TIPC_RETDATA, anc_data[1],
1767 				       msg_data(msg));
1768 			if (res)
1769 				return res;
1770 		}
1771 	}
1772 
1773 	/* Optionally capture message destination object */
1774 	dest_type = msg ? msg_type(msg) : TIPC_DIRECT_MSG;
1775 	switch (dest_type) {
1776 	case TIPC_NAMED_MSG:
1777 		has_name = 1;
1778 		anc_data[0] = msg_nametype(msg);
1779 		anc_data[1] = msg_namelower(msg);
1780 		anc_data[2] = msg_namelower(msg);
1781 		break;
1782 	case TIPC_MCAST_MSG:
1783 		has_name = 1;
1784 		anc_data[0] = msg_nametype(msg);
1785 		anc_data[1] = msg_namelower(msg);
1786 		anc_data[2] = msg_nameupper(msg);
1787 		break;
1788 	case TIPC_CONN_MSG:
1789 		has_name = (tsk->conn_type != 0);
1790 		anc_data[0] = tsk->conn_type;
1791 		anc_data[1] = tsk->conn_instance;
1792 		anc_data[2] = tsk->conn_instance;
1793 		break;
1794 	default:
1795 		has_name = 0;
1796 	}
1797 	if (has_name) {
1798 		res = put_cmsg(m, SOL_TIPC, TIPC_DESTNAME, 12, anc_data);
1799 		if (res)
1800 			return res;
1801 	}
1802 
1803 	return 0;
1804 }
1805 
1806 static struct sk_buff *tipc_sk_build_ack(struct tipc_sock *tsk)
1807 {
1808 	struct sock *sk = &tsk->sk;
1809 	struct sk_buff *skb = NULL;
1810 	struct tipc_msg *msg;
1811 	u32 peer_port = tsk_peer_port(tsk);
1812 	u32 dnode = tsk_peer_node(tsk);
1813 
1814 	if (!tipc_sk_connected(sk))
1815 		return NULL;
1816 	skb = tipc_msg_create(CONN_MANAGER, CONN_ACK, INT_H_SIZE, 0,
1817 			      dnode, tsk_own_node(tsk), peer_port,
1818 			      tsk->portid, TIPC_OK);
1819 	if (!skb)
1820 		return NULL;
1821 	msg = buf_msg(skb);
1822 	msg_set_conn_ack(msg, tsk->rcv_unacked);
1823 	tsk->rcv_unacked = 0;
1824 
1825 	/* Adjust to and advertize the correct window limit */
1826 	if (tsk->peer_caps & TIPC_BLOCK_FLOWCTL) {
1827 		tsk->rcv_win = tsk_adv_blocks(tsk->sk.sk_rcvbuf);
1828 		msg_set_adv_win(msg, tsk->rcv_win);
1829 	}
1830 	return skb;
1831 }
1832 
1833 static void tipc_sk_send_ack(struct tipc_sock *tsk)
1834 {
1835 	struct sk_buff *skb;
1836 
1837 	skb = tipc_sk_build_ack(tsk);
1838 	if (!skb)
1839 		return;
1840 
1841 	tipc_node_xmit_skb(sock_net(&tsk->sk), skb, tsk_peer_node(tsk),
1842 			   msg_link_selector(buf_msg(skb)));
1843 }
1844 
1845 static int tipc_wait_for_rcvmsg(struct socket *sock, long *timeop)
1846 {
1847 	struct sock *sk = sock->sk;
1848 	DEFINE_WAIT_FUNC(wait, woken_wake_function);
1849 	long timeo = *timeop;
1850 	int err = sock_error(sk);
1851 
1852 	if (err)
1853 		return err;
1854 
1855 	for (;;) {
1856 		if (timeo && skb_queue_empty(&sk->sk_receive_queue)) {
1857 			if (sk->sk_shutdown & RCV_SHUTDOWN) {
1858 				err = -ENOTCONN;
1859 				break;
1860 			}
1861 			add_wait_queue(sk_sleep(sk), &wait);
1862 			release_sock(sk);
1863 			timeo = wait_woken(&wait, TASK_INTERRUPTIBLE, timeo);
1864 			sched_annotate_sleep();
1865 			lock_sock(sk);
1866 			remove_wait_queue(sk_sleep(sk), &wait);
1867 		}
1868 		err = 0;
1869 		if (!skb_queue_empty(&sk->sk_receive_queue))
1870 			break;
1871 		err = -EAGAIN;
1872 		if (!timeo)
1873 			break;
1874 		err = sock_intr_errno(timeo);
1875 		if (signal_pending(current))
1876 			break;
1877 
1878 		err = sock_error(sk);
1879 		if (err)
1880 			break;
1881 	}
1882 	*timeop = timeo;
1883 	return err;
1884 }
1885 
1886 /**
1887  * tipc_recvmsg - receive packet-oriented message
1888  * @sock: network socket
1889  * @m: descriptor for message info
1890  * @buflen: length of user buffer area
1891  * @flags: receive flags
1892  *
1893  * Used for SOCK_DGRAM, SOCK_RDM, and SOCK_SEQPACKET messages.
1894  * If the complete message doesn't fit in user area, truncate it.
1895  *
1896  * Return: size of returned message data, errno otherwise
1897  */
1898 static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
1899 			size_t buflen,	int flags)
1900 {
1901 	struct sock *sk = sock->sk;
1902 	bool connected = !tipc_sk_type_connectionless(sk);
1903 	struct tipc_sock *tsk = tipc_sk(sk);
1904 	int rc, err, hlen, dlen, copy;
1905 	struct sk_buff_head xmitq;
1906 	struct tipc_msg *hdr;
1907 	struct sk_buff *skb;
1908 	bool grp_evt;
1909 	long timeout;
1910 
1911 	/* Catch invalid receive requests */
1912 	if (unlikely(!buflen))
1913 		return -EINVAL;
1914 
1915 	lock_sock(sk);
1916 	if (unlikely(connected && sk->sk_state == TIPC_OPEN)) {
1917 		rc = -ENOTCONN;
1918 		goto exit;
1919 	}
1920 	timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
1921 
1922 	/* Step rcv queue to first msg with data or error; wait if necessary */
1923 	do {
1924 		rc = tipc_wait_for_rcvmsg(sock, &timeout);
1925 		if (unlikely(rc))
1926 			goto exit;
1927 		skb = skb_peek(&sk->sk_receive_queue);
1928 		hdr = buf_msg(skb);
1929 		dlen = msg_data_sz(hdr);
1930 		hlen = msg_hdr_sz(hdr);
1931 		err = msg_errcode(hdr);
1932 		grp_evt = msg_is_grp_evt(hdr);
1933 		if (likely(dlen || err))
1934 			break;
1935 		tsk_advance_rx_queue(sk);
1936 	} while (1);
1937 
1938 	/* Collect msg meta data, including error code and rejected data */
1939 	tipc_sk_set_orig_addr(m, skb);
1940 	rc = tipc_sk_anc_data_recv(m, skb, tsk);
1941 	if (unlikely(rc))
1942 		goto exit;
1943 	hdr = buf_msg(skb);
1944 
1945 	/* Capture data if non-error msg, otherwise just set return value */
1946 	if (likely(!err)) {
1947 		copy = min_t(int, dlen, buflen);
1948 		if (unlikely(copy != dlen))
1949 			m->msg_flags |= MSG_TRUNC;
1950 		rc = skb_copy_datagram_msg(skb, hlen, m, copy);
1951 	} else {
1952 		copy = 0;
1953 		rc = 0;
1954 		if (err != TIPC_CONN_SHUTDOWN && connected && !m->msg_control)
1955 			rc = -ECONNRESET;
1956 	}
1957 	if (unlikely(rc))
1958 		goto exit;
1959 
1960 	/* Mark message as group event if applicable */
1961 	if (unlikely(grp_evt)) {
1962 		if (msg_grp_evt(hdr) == TIPC_WITHDRAWN)
1963 			m->msg_flags |= MSG_EOR;
1964 		m->msg_flags |= MSG_OOB;
1965 		copy = 0;
1966 	}
1967 
1968 	/* Caption of data or error code/rejected data was successful */
1969 	if (unlikely(flags & MSG_PEEK))
1970 		goto exit;
1971 
1972 	/* Send group flow control advertisement when applicable */
1973 	if (tsk->group && msg_in_group(hdr) && !grp_evt) {
1974 		__skb_queue_head_init(&xmitq);
1975 		tipc_group_update_rcv_win(tsk->group, tsk_blocks(hlen + dlen),
1976 					  msg_orignode(hdr), msg_origport(hdr),
1977 					  &xmitq);
1978 		tipc_node_distr_xmit(sock_net(sk), &xmitq);
1979 	}
1980 
1981 	tsk_advance_rx_queue(sk);
1982 
1983 	if (likely(!connected))
1984 		goto exit;
1985 
1986 	/* Send connection flow control advertisement when applicable */
1987 	tsk->rcv_unacked += tsk_inc(tsk, hlen + dlen);
1988 	if (tsk->rcv_unacked >= tsk->rcv_win / TIPC_ACK_RATE)
1989 		tipc_sk_send_ack(tsk);
1990 exit:
1991 	release_sock(sk);
1992 	return rc ? rc : copy;
1993 }
1994 
1995 /**
1996  * tipc_recvstream - receive stream-oriented data
1997  * @sock: network socket
1998  * @m: descriptor for message info
1999  * @buflen: total size of user buffer area
2000  * @flags: receive flags
2001  *
2002  * Used for SOCK_STREAM messages only.  If not enough data is available
2003  * will optionally wait for more; never truncates data.
2004  *
2005  * Return: size of returned message data, errno otherwise
2006  */
2007 static int tipc_recvstream(struct socket *sock, struct msghdr *m,
2008 			   size_t buflen, int flags)
2009 {
2010 	struct sock *sk = sock->sk;
2011 	struct tipc_sock *tsk = tipc_sk(sk);
2012 	struct sk_buff *skb;
2013 	struct tipc_msg *hdr;
2014 	struct tipc_skb_cb *skb_cb;
2015 	bool peek = flags & MSG_PEEK;
2016 	int offset, required, copy, copied = 0;
2017 	int hlen, dlen, err, rc;
2018 	long timeout;
2019 
2020 	/* Catch invalid receive attempts */
2021 	if (unlikely(!buflen))
2022 		return -EINVAL;
2023 
2024 	lock_sock(sk);
2025 
2026 	if (unlikely(sk->sk_state == TIPC_OPEN)) {
2027 		rc = -ENOTCONN;
2028 		goto exit;
2029 	}
2030 	required = sock_rcvlowat(sk, flags & MSG_WAITALL, buflen);
2031 	timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
2032 
2033 	do {
2034 		/* Look at first msg in receive queue; wait if necessary */
2035 		rc = tipc_wait_for_rcvmsg(sock, &timeout);
2036 		if (unlikely(rc))
2037 			break;
2038 		skb = skb_peek(&sk->sk_receive_queue);
2039 		skb_cb = TIPC_SKB_CB(skb);
2040 		hdr = buf_msg(skb);
2041 		dlen = msg_data_sz(hdr);
2042 		hlen = msg_hdr_sz(hdr);
2043 		err = msg_errcode(hdr);
2044 
2045 		/* Discard any empty non-errored (SYN-) message */
2046 		if (unlikely(!dlen && !err)) {
2047 			tsk_advance_rx_queue(sk);
2048 			continue;
2049 		}
2050 
2051 		/* Collect msg meta data, incl. error code and rejected data */
2052 		if (!copied) {
2053 			tipc_sk_set_orig_addr(m, skb);
2054 			rc = tipc_sk_anc_data_recv(m, skb, tsk);
2055 			if (rc)
2056 				break;
2057 			hdr = buf_msg(skb);
2058 		}
2059 
2060 		/* Copy data if msg ok, otherwise return error/partial data */
2061 		if (likely(!err)) {
2062 			offset = skb_cb->bytes_read;
2063 			copy = min_t(int, dlen - offset, buflen - copied);
2064 			rc = skb_copy_datagram_msg(skb, hlen + offset, m, copy);
2065 			if (unlikely(rc))
2066 				break;
2067 			copied += copy;
2068 			offset += copy;
2069 			if (unlikely(offset < dlen)) {
2070 				if (!peek)
2071 					skb_cb->bytes_read = offset;
2072 				break;
2073 			}
2074 		} else {
2075 			rc = 0;
2076 			if ((err != TIPC_CONN_SHUTDOWN) && !m->msg_control)
2077 				rc = -ECONNRESET;
2078 			if (copied || rc)
2079 				break;
2080 		}
2081 
2082 		if (unlikely(peek))
2083 			break;
2084 
2085 		tsk_advance_rx_queue(sk);
2086 
2087 		/* Send connection flow control advertisement when applicable */
2088 		tsk->rcv_unacked += tsk_inc(tsk, hlen + dlen);
2089 		if (tsk->rcv_unacked >= tsk->rcv_win / TIPC_ACK_RATE)
2090 			tipc_sk_send_ack(tsk);
2091 
2092 		/* Exit if all requested data or FIN/error received */
2093 		if (copied == buflen || err)
2094 			break;
2095 
2096 	} while (!skb_queue_empty(&sk->sk_receive_queue) || copied < required);
2097 exit:
2098 	release_sock(sk);
2099 	return copied ? copied : rc;
2100 }
2101 
2102 /**
2103  * tipc_write_space - wake up thread if port congestion is released
2104  * @sk: socket
2105  */
2106 static void tipc_write_space(struct sock *sk)
2107 {
2108 	struct socket_wq *wq;
2109 
2110 	rcu_read_lock();
2111 	wq = rcu_dereference(sk->sk_wq);
2112 	if (skwq_has_sleeper(wq))
2113 		wake_up_interruptible_sync_poll(&wq->wait, EPOLLOUT |
2114 						EPOLLWRNORM | EPOLLWRBAND);
2115 	rcu_read_unlock();
2116 }
2117 
2118 /**
2119  * tipc_data_ready - wake up threads to indicate messages have been received
2120  * @sk: socket
2121  */
2122 static void tipc_data_ready(struct sock *sk)
2123 {
2124 	struct socket_wq *wq;
2125 
2126 	rcu_read_lock();
2127 	wq = rcu_dereference(sk->sk_wq);
2128 	if (skwq_has_sleeper(wq))
2129 		wake_up_interruptible_sync_poll(&wq->wait, EPOLLIN |
2130 						EPOLLRDNORM | EPOLLRDBAND);
2131 	rcu_read_unlock();
2132 }
2133 
2134 static void tipc_sock_destruct(struct sock *sk)
2135 {
2136 	__skb_queue_purge(&sk->sk_receive_queue);
2137 }
2138 
2139 static void tipc_sk_proto_rcv(struct sock *sk,
2140 			      struct sk_buff_head *inputq,
2141 			      struct sk_buff_head *xmitq)
2142 {
2143 	struct sk_buff *skb = __skb_dequeue(inputq);
2144 	struct tipc_sock *tsk = tipc_sk(sk);
2145 	struct tipc_msg *hdr = buf_msg(skb);
2146 	struct tipc_group *grp = tsk->group;
2147 	bool wakeup = false;
2148 
2149 	switch (msg_user(hdr)) {
2150 	case CONN_MANAGER:
2151 		tipc_sk_conn_proto_rcv(tsk, skb, inputq, xmitq);
2152 		return;
2153 	case SOCK_WAKEUP:
2154 		tipc_dest_del(&tsk->cong_links, msg_orignode(hdr), 0);
2155 		/* coupled with smp_rmb() in tipc_wait_for_cond() */
2156 		smp_wmb();
2157 		tsk->cong_link_cnt--;
2158 		wakeup = true;
2159 		tipc_sk_push_backlog(tsk, false);
2160 		break;
2161 	case GROUP_PROTOCOL:
2162 		tipc_group_proto_rcv(grp, &wakeup, hdr, inputq, xmitq);
2163 		break;
2164 	case TOP_SRV:
2165 		tipc_group_member_evt(tsk->group, &wakeup, &sk->sk_rcvbuf,
2166 				      hdr, inputq, xmitq);
2167 		break;
2168 	default:
2169 		break;
2170 	}
2171 
2172 	if (wakeup)
2173 		sk->sk_write_space(sk);
2174 
2175 	kfree_skb(skb);
2176 }
2177 
2178 /**
2179  * tipc_sk_filter_connect - check incoming message for a connection-based socket
2180  * @tsk: TIPC socket
2181  * @skb: pointer to message buffer.
2182  * @xmitq: for Nagle ACK if any
2183  * Return: true if message should be added to receive queue, false otherwise
2184  */
2185 static bool tipc_sk_filter_connect(struct tipc_sock *tsk, struct sk_buff *skb,
2186 				   struct sk_buff_head *xmitq)
2187 {
2188 	struct sock *sk = &tsk->sk;
2189 	struct net *net = sock_net(sk);
2190 	struct tipc_msg *hdr = buf_msg(skb);
2191 	bool con_msg = msg_connected(hdr);
2192 	u32 pport = tsk_peer_port(tsk);
2193 	u32 pnode = tsk_peer_node(tsk);
2194 	u32 oport = msg_origport(hdr);
2195 	u32 onode = msg_orignode(hdr);
2196 	int err = msg_errcode(hdr);
2197 	unsigned long delay;
2198 
2199 	if (unlikely(msg_mcast(hdr)))
2200 		return false;
2201 	tsk->oneway = 0;
2202 
2203 	switch (sk->sk_state) {
2204 	case TIPC_CONNECTING:
2205 		/* Setup ACK */
2206 		if (likely(con_msg)) {
2207 			if (err)
2208 				break;
2209 			tipc_sk_finish_conn(tsk, oport, onode);
2210 			msg_set_importance(&tsk->phdr, msg_importance(hdr));
2211 			/* ACK+ message with data is added to receive queue */
2212 			if (msg_data_sz(hdr))
2213 				return true;
2214 			/* Empty ACK-, - wake up sleeping connect() and drop */
2215 			sk->sk_state_change(sk);
2216 			msg_set_dest_droppable(hdr, 1);
2217 			return false;
2218 		}
2219 		/* Ignore connectionless message if not from listening socket */
2220 		if (oport != pport || onode != pnode)
2221 			return false;
2222 
2223 		/* Rejected SYN */
2224 		if (err != TIPC_ERR_OVERLOAD)
2225 			break;
2226 
2227 		/* Prepare for new setup attempt if we have a SYN clone */
2228 		if (skb_queue_empty(&sk->sk_write_queue))
2229 			break;
2230 		get_random_bytes(&delay, 2);
2231 		delay %= (tsk->conn_timeout / 4);
2232 		delay = msecs_to_jiffies(delay + 100);
2233 		sk_reset_timer(sk, &sk->sk_timer, jiffies + delay);
2234 		return false;
2235 	case TIPC_OPEN:
2236 	case TIPC_DISCONNECTING:
2237 		return false;
2238 	case TIPC_LISTEN:
2239 		/* Accept only SYN message */
2240 		if (!msg_is_syn(hdr) &&
2241 		    tipc_node_get_capabilities(net, onode) & TIPC_SYN_BIT)
2242 			return false;
2243 		if (!con_msg && !err)
2244 			return true;
2245 		return false;
2246 	case TIPC_ESTABLISHED:
2247 		if (!skb_queue_empty(&sk->sk_write_queue))
2248 			tipc_sk_push_backlog(tsk, false);
2249 		/* Accept only connection-based messages sent by peer */
2250 		if (likely(con_msg && !err && pport == oport &&
2251 			   pnode == onode)) {
2252 			if (msg_ack_required(hdr)) {
2253 				struct sk_buff *skb;
2254 
2255 				skb = tipc_sk_build_ack(tsk);
2256 				if (skb) {
2257 					msg_set_nagle_ack(buf_msg(skb));
2258 					__skb_queue_tail(xmitq, skb);
2259 				}
2260 			}
2261 			return true;
2262 		}
2263 		if (!tsk_peer_msg(tsk, hdr))
2264 			return false;
2265 		if (!err)
2266 			return true;
2267 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2268 		tipc_node_remove_conn(net, pnode, tsk->portid);
2269 		sk->sk_state_change(sk);
2270 		return true;
2271 	default:
2272 		pr_err("Unknown sk_state %u\n", sk->sk_state);
2273 	}
2274 	/* Abort connection setup attempt */
2275 	tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2276 	sk->sk_err = ECONNREFUSED;
2277 	sk->sk_state_change(sk);
2278 	return true;
2279 }
2280 
2281 /**
2282  * rcvbuf_limit - get proper overload limit of socket receive queue
2283  * @sk: socket
2284  * @skb: message
2285  *
2286  * For connection oriented messages, irrespective of importance,
2287  * default queue limit is 2 MB.
2288  *
2289  * For connectionless messages, queue limits are based on message
2290  * importance as follows:
2291  *
2292  * TIPC_LOW_IMPORTANCE       (2 MB)
2293  * TIPC_MEDIUM_IMPORTANCE    (4 MB)
2294  * TIPC_HIGH_IMPORTANCE      (8 MB)
2295  * TIPC_CRITICAL_IMPORTANCE  (16 MB)
2296  *
2297  * Return: overload limit according to corresponding message importance
2298  */
2299 static unsigned int rcvbuf_limit(struct sock *sk, struct sk_buff *skb)
2300 {
2301 	struct tipc_sock *tsk = tipc_sk(sk);
2302 	struct tipc_msg *hdr = buf_msg(skb);
2303 
2304 	if (unlikely(msg_in_group(hdr)))
2305 		return READ_ONCE(sk->sk_rcvbuf);
2306 
2307 	if (unlikely(!msg_connected(hdr)))
2308 		return READ_ONCE(sk->sk_rcvbuf) << msg_importance(hdr);
2309 
2310 	if (likely(tsk->peer_caps & TIPC_BLOCK_FLOWCTL))
2311 		return READ_ONCE(sk->sk_rcvbuf);
2312 
2313 	return FLOWCTL_MSG_LIM;
2314 }
2315 
2316 /**
2317  * tipc_sk_filter_rcv - validate incoming message
2318  * @sk: socket
2319  * @skb: pointer to message.
2320  * @xmitq: output message area (FIXME)
2321  *
2322  * Enqueues message on receive queue if acceptable; optionally handles
2323  * disconnect indication for a connected socket.
2324  *
2325  * Called with socket lock already taken
2326  */
2327 static void tipc_sk_filter_rcv(struct sock *sk, struct sk_buff *skb,
2328 			       struct sk_buff_head *xmitq)
2329 {
2330 	bool sk_conn = !tipc_sk_type_connectionless(sk);
2331 	struct tipc_sock *tsk = tipc_sk(sk);
2332 	struct tipc_group *grp = tsk->group;
2333 	struct tipc_msg *hdr = buf_msg(skb);
2334 	struct net *net = sock_net(sk);
2335 	struct sk_buff_head inputq;
2336 	int mtyp = msg_type(hdr);
2337 	int limit, err = TIPC_OK;
2338 
2339 	trace_tipc_sk_filter_rcv(sk, skb, TIPC_DUMP_ALL, " ");
2340 	TIPC_SKB_CB(skb)->bytes_read = 0;
2341 	__skb_queue_head_init(&inputq);
2342 	__skb_queue_tail(&inputq, skb);
2343 
2344 	if (unlikely(!msg_isdata(hdr)))
2345 		tipc_sk_proto_rcv(sk, &inputq, xmitq);
2346 
2347 	if (unlikely(grp))
2348 		tipc_group_filter_msg(grp, &inputq, xmitq);
2349 
2350 	if (unlikely(!grp) && mtyp == TIPC_MCAST_MSG)
2351 		tipc_mcast_filter_msg(net, &tsk->mc_method.deferredq, &inputq);
2352 
2353 	/* Validate and add to receive buffer if there is space */
2354 	while ((skb = __skb_dequeue(&inputq))) {
2355 		hdr = buf_msg(skb);
2356 		limit = rcvbuf_limit(sk, skb);
2357 		if ((sk_conn && !tipc_sk_filter_connect(tsk, skb, xmitq)) ||
2358 		    (!sk_conn && msg_connected(hdr)) ||
2359 		    (!grp && msg_in_group(hdr)))
2360 			err = TIPC_ERR_NO_PORT;
2361 		else if (sk_rmem_alloc_get(sk) + skb->truesize >= limit) {
2362 			trace_tipc_sk_dump(sk, skb, TIPC_DUMP_ALL,
2363 					   "err_overload2!");
2364 			atomic_inc(&sk->sk_drops);
2365 			err = TIPC_ERR_OVERLOAD;
2366 		}
2367 
2368 		if (unlikely(err)) {
2369 			if (tipc_msg_reverse(tipc_own_addr(net), &skb, err)) {
2370 				trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_NONE,
2371 						      "@filter_rcv!");
2372 				__skb_queue_tail(xmitq, skb);
2373 			}
2374 			err = TIPC_OK;
2375 			continue;
2376 		}
2377 		__skb_queue_tail(&sk->sk_receive_queue, skb);
2378 		skb_set_owner_r(skb, sk);
2379 		trace_tipc_sk_overlimit2(sk, skb, TIPC_DUMP_ALL,
2380 					 "rcvq >90% allocated!");
2381 		sk->sk_data_ready(sk);
2382 	}
2383 }
2384 
2385 /**
2386  * tipc_sk_backlog_rcv - handle incoming message from backlog queue
2387  * @sk: socket
2388  * @skb: message
2389  *
2390  * Caller must hold socket lock
2391  */
2392 static int tipc_sk_backlog_rcv(struct sock *sk, struct sk_buff *skb)
2393 {
2394 	unsigned int before = sk_rmem_alloc_get(sk);
2395 	struct sk_buff_head xmitq;
2396 	unsigned int added;
2397 
2398 	__skb_queue_head_init(&xmitq);
2399 
2400 	tipc_sk_filter_rcv(sk, skb, &xmitq);
2401 	added = sk_rmem_alloc_get(sk) - before;
2402 	atomic_add(added, &tipc_sk(sk)->dupl_rcvcnt);
2403 
2404 	/* Send pending response/rejected messages, if any */
2405 	tipc_node_distr_xmit(sock_net(sk), &xmitq);
2406 	return 0;
2407 }
2408 
2409 /**
2410  * tipc_sk_enqueue - extract all buffers with destination 'dport' from
2411  *                   inputq and try adding them to socket or backlog queue
2412  * @inputq: list of incoming buffers with potentially different destinations
2413  * @sk: socket where the buffers should be enqueued
2414  * @dport: port number for the socket
2415  * @xmitq: output queue
2416  *
2417  * Caller must hold socket lock
2418  */
2419 static void tipc_sk_enqueue(struct sk_buff_head *inputq, struct sock *sk,
2420 			    u32 dport, struct sk_buff_head *xmitq)
2421 {
2422 	unsigned long time_limit = jiffies + 2;
2423 	struct sk_buff *skb;
2424 	unsigned int lim;
2425 	atomic_t *dcnt;
2426 	u32 onode;
2427 
2428 	while (skb_queue_len(inputq)) {
2429 		if (unlikely(time_after_eq(jiffies, time_limit)))
2430 			return;
2431 
2432 		skb = tipc_skb_dequeue(inputq, dport);
2433 		if (unlikely(!skb))
2434 			return;
2435 
2436 		/* Add message directly to receive queue if possible */
2437 		if (!sock_owned_by_user(sk)) {
2438 			tipc_sk_filter_rcv(sk, skb, xmitq);
2439 			continue;
2440 		}
2441 
2442 		/* Try backlog, compensating for double-counted bytes */
2443 		dcnt = &tipc_sk(sk)->dupl_rcvcnt;
2444 		if (!sk->sk_backlog.len)
2445 			atomic_set(dcnt, 0);
2446 		lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt);
2447 		if (likely(!sk_add_backlog(sk, skb, lim))) {
2448 			trace_tipc_sk_overlimit1(sk, skb, TIPC_DUMP_ALL,
2449 						 "bklg & rcvq >90% allocated!");
2450 			continue;
2451 		}
2452 
2453 		trace_tipc_sk_dump(sk, skb, TIPC_DUMP_ALL, "err_overload!");
2454 		/* Overload => reject message back to sender */
2455 		onode = tipc_own_addr(sock_net(sk));
2456 		atomic_inc(&sk->sk_drops);
2457 		if (tipc_msg_reverse(onode, &skb, TIPC_ERR_OVERLOAD)) {
2458 			trace_tipc_sk_rej_msg(sk, skb, TIPC_DUMP_ALL,
2459 					      "@sk_enqueue!");
2460 			__skb_queue_tail(xmitq, skb);
2461 		}
2462 		break;
2463 	}
2464 }
2465 
2466 /**
2467  * tipc_sk_rcv - handle a chain of incoming buffers
2468  * @net: the associated network namespace
2469  * @inputq: buffer list containing the buffers
2470  * Consumes all buffers in list until inputq is empty
2471  * Note: may be called in multiple threads referring to the same queue
2472  */
2473 void tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq)
2474 {
2475 	struct sk_buff_head xmitq;
2476 	u32 dnode, dport = 0;
2477 	int err;
2478 	struct tipc_sock *tsk;
2479 	struct sock *sk;
2480 	struct sk_buff *skb;
2481 
2482 	__skb_queue_head_init(&xmitq);
2483 	while (skb_queue_len(inputq)) {
2484 		dport = tipc_skb_peek_port(inputq, dport);
2485 		tsk = tipc_sk_lookup(net, dport);
2486 
2487 		if (likely(tsk)) {
2488 			sk = &tsk->sk;
2489 			if (likely(spin_trylock_bh(&sk->sk_lock.slock))) {
2490 				tipc_sk_enqueue(inputq, sk, dport, &xmitq);
2491 				spin_unlock_bh(&sk->sk_lock.slock);
2492 			}
2493 			/* Send pending response/rejected messages, if any */
2494 			tipc_node_distr_xmit(sock_net(sk), &xmitq);
2495 			sock_put(sk);
2496 			continue;
2497 		}
2498 		/* No destination socket => dequeue skb if still there */
2499 		skb = tipc_skb_dequeue(inputq, dport);
2500 		if (!skb)
2501 			return;
2502 
2503 		/* Try secondary lookup if unresolved named message */
2504 		err = TIPC_ERR_NO_PORT;
2505 		if (tipc_msg_lookup_dest(net, skb, &err))
2506 			goto xmit;
2507 
2508 		/* Prepare for message rejection */
2509 		if (!tipc_msg_reverse(tipc_own_addr(net), &skb, err))
2510 			continue;
2511 
2512 		trace_tipc_sk_rej_msg(NULL, skb, TIPC_DUMP_NONE, "@sk_rcv!");
2513 xmit:
2514 		dnode = msg_destnode(buf_msg(skb));
2515 		tipc_node_xmit_skb(net, skb, dnode, dport);
2516 	}
2517 }
2518 
2519 static int tipc_wait_for_connect(struct socket *sock, long *timeo_p)
2520 {
2521 	DEFINE_WAIT_FUNC(wait, woken_wake_function);
2522 	struct sock *sk = sock->sk;
2523 	int done;
2524 
2525 	do {
2526 		int err = sock_error(sk);
2527 		if (err)
2528 			return err;
2529 		if (!*timeo_p)
2530 			return -ETIMEDOUT;
2531 		if (signal_pending(current))
2532 			return sock_intr_errno(*timeo_p);
2533 		if (sk->sk_state == TIPC_DISCONNECTING)
2534 			break;
2535 
2536 		add_wait_queue(sk_sleep(sk), &wait);
2537 		done = sk_wait_event(sk, timeo_p, tipc_sk_connected(sk),
2538 				     &wait);
2539 		remove_wait_queue(sk_sleep(sk), &wait);
2540 	} while (!done);
2541 	return 0;
2542 }
2543 
2544 static bool tipc_sockaddr_is_sane(struct sockaddr_tipc *addr)
2545 {
2546 	if (addr->family != AF_TIPC)
2547 		return false;
2548 	if (addr->addrtype == TIPC_SERVICE_RANGE)
2549 		return (addr->addr.nameseq.lower <= addr->addr.nameseq.upper);
2550 	return (addr->addrtype == TIPC_SERVICE_ADDR ||
2551 		addr->addrtype == TIPC_SOCKET_ADDR);
2552 }
2553 
2554 /**
2555  * tipc_connect - establish a connection to another TIPC port
2556  * @sock: socket structure
2557  * @dest: socket address for destination port
2558  * @destlen: size of socket address data structure
2559  * @flags: file-related flags associated with socket
2560  *
2561  * Return: 0 on success, errno otherwise
2562  */
2563 static int tipc_connect(struct socket *sock, struct sockaddr *dest,
2564 			int destlen, int flags)
2565 {
2566 	struct sock *sk = sock->sk;
2567 	struct tipc_sock *tsk = tipc_sk(sk);
2568 	struct sockaddr_tipc *dst = (struct sockaddr_tipc *)dest;
2569 	struct msghdr m = {NULL,};
2570 	long timeout = (flags & O_NONBLOCK) ? 0 : tsk->conn_timeout;
2571 	int previous;
2572 	int res = 0;
2573 
2574 	if (destlen != sizeof(struct sockaddr_tipc))
2575 		return -EINVAL;
2576 
2577 	lock_sock(sk);
2578 
2579 	if (tsk->group) {
2580 		res = -EINVAL;
2581 		goto exit;
2582 	}
2583 
2584 	if (dst->family == AF_UNSPEC) {
2585 		memset(&tsk->peer, 0, sizeof(struct sockaddr_tipc));
2586 		if (!tipc_sk_type_connectionless(sk))
2587 			res = -EINVAL;
2588 		goto exit;
2589 	}
2590 	if (!tipc_sockaddr_is_sane(dst)) {
2591 		res = -EINVAL;
2592 		goto exit;
2593 	}
2594 	/* DGRAM/RDM connect(), just save the destaddr */
2595 	if (tipc_sk_type_connectionless(sk)) {
2596 		memcpy(&tsk->peer, dest, destlen);
2597 		goto exit;
2598 	} else if (dst->addrtype == TIPC_SERVICE_RANGE) {
2599 		res = -EINVAL;
2600 		goto exit;
2601 	}
2602 
2603 	previous = sk->sk_state;
2604 
2605 	switch (sk->sk_state) {
2606 	case TIPC_OPEN:
2607 		/* Send a 'SYN-' to destination */
2608 		m.msg_name = dest;
2609 		m.msg_namelen = destlen;
2610 
2611 		/* If connect is in non-blocking case, set MSG_DONTWAIT to
2612 		 * indicate send_msg() is never blocked.
2613 		 */
2614 		if (!timeout)
2615 			m.msg_flags = MSG_DONTWAIT;
2616 
2617 		res = __tipc_sendmsg(sock, &m, 0);
2618 		if ((res < 0) && (res != -EWOULDBLOCK))
2619 			goto exit;
2620 
2621 		/* Just entered TIPC_CONNECTING state; the only
2622 		 * difference is that return value in non-blocking
2623 		 * case is EINPROGRESS, rather than EALREADY.
2624 		 */
2625 		res = -EINPROGRESS;
2626 		fallthrough;
2627 	case TIPC_CONNECTING:
2628 		if (!timeout) {
2629 			if (previous == TIPC_CONNECTING)
2630 				res = -EALREADY;
2631 			goto exit;
2632 		}
2633 		timeout = msecs_to_jiffies(timeout);
2634 		/* Wait until an 'ACK' or 'RST' arrives, or a timeout occurs */
2635 		res = tipc_wait_for_connect(sock, &timeout);
2636 		break;
2637 	case TIPC_ESTABLISHED:
2638 		res = -EISCONN;
2639 		break;
2640 	default:
2641 		res = -EINVAL;
2642 	}
2643 
2644 exit:
2645 	release_sock(sk);
2646 	return res;
2647 }
2648 
2649 /**
2650  * tipc_listen - allow socket to listen for incoming connections
2651  * @sock: socket structure
2652  * @len: (unused)
2653  *
2654  * Return: 0 on success, errno otherwise
2655  */
2656 static int tipc_listen(struct socket *sock, int len)
2657 {
2658 	struct sock *sk = sock->sk;
2659 	int res;
2660 
2661 	lock_sock(sk);
2662 	res = tipc_set_sk_state(sk, TIPC_LISTEN);
2663 	release_sock(sk);
2664 
2665 	return res;
2666 }
2667 
2668 static int tipc_wait_for_accept(struct socket *sock, long timeo)
2669 {
2670 	struct sock *sk = sock->sk;
2671 	DEFINE_WAIT(wait);
2672 	int err;
2673 
2674 	/* True wake-one mechanism for incoming connections: only
2675 	 * one process gets woken up, not the 'whole herd'.
2676 	 * Since we do not 'race & poll' for established sockets
2677 	 * anymore, the common case will execute the loop only once.
2678 	*/
2679 	for (;;) {
2680 		prepare_to_wait_exclusive(sk_sleep(sk), &wait,
2681 					  TASK_INTERRUPTIBLE);
2682 		if (timeo && skb_queue_empty(&sk->sk_receive_queue)) {
2683 			release_sock(sk);
2684 			timeo = schedule_timeout(timeo);
2685 			lock_sock(sk);
2686 		}
2687 		err = 0;
2688 		if (!skb_queue_empty(&sk->sk_receive_queue))
2689 			break;
2690 		err = -EAGAIN;
2691 		if (!timeo)
2692 			break;
2693 		err = sock_intr_errno(timeo);
2694 		if (signal_pending(current))
2695 			break;
2696 	}
2697 	finish_wait(sk_sleep(sk), &wait);
2698 	return err;
2699 }
2700 
2701 /**
2702  * tipc_accept - wait for connection request
2703  * @sock: listening socket
2704  * @new_sock: new socket that is to be connected
2705  * @flags: file-related flags associated with socket
2706  * @kern: caused by kernel or by userspace?
2707  *
2708  * Return: 0 on success, errno otherwise
2709  */
2710 static int tipc_accept(struct socket *sock, struct socket *new_sock, int flags,
2711 		       bool kern)
2712 {
2713 	struct sock *new_sk, *sk = sock->sk;
2714 	struct sk_buff *buf;
2715 	struct tipc_sock *new_tsock;
2716 	struct tipc_msg *msg;
2717 	long timeo;
2718 	int res;
2719 
2720 	lock_sock(sk);
2721 
2722 	if (sk->sk_state != TIPC_LISTEN) {
2723 		res = -EINVAL;
2724 		goto exit;
2725 	}
2726 	timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
2727 	res = tipc_wait_for_accept(sock, timeo);
2728 	if (res)
2729 		goto exit;
2730 
2731 	buf = skb_peek(&sk->sk_receive_queue);
2732 
2733 	res = tipc_sk_create(sock_net(sock->sk), new_sock, 0, kern);
2734 	if (res)
2735 		goto exit;
2736 	security_sk_clone(sock->sk, new_sock->sk);
2737 
2738 	new_sk = new_sock->sk;
2739 	new_tsock = tipc_sk(new_sk);
2740 	msg = buf_msg(buf);
2741 
2742 	/* we lock on new_sk; but lockdep sees the lock on sk */
2743 	lock_sock_nested(new_sk, SINGLE_DEPTH_NESTING);
2744 
2745 	/*
2746 	 * Reject any stray messages received by new socket
2747 	 * before the socket lock was taken (very, very unlikely)
2748 	 */
2749 	tsk_rej_rx_queue(new_sk, TIPC_ERR_NO_PORT);
2750 
2751 	/* Connect new socket to it's peer */
2752 	tipc_sk_finish_conn(new_tsock, msg_origport(msg), msg_orignode(msg));
2753 
2754 	tsk_set_importance(new_sk, msg_importance(msg));
2755 	if (msg_named(msg)) {
2756 		new_tsock->conn_type = msg_nametype(msg);
2757 		new_tsock->conn_instance = msg_nameinst(msg);
2758 	}
2759 
2760 	/*
2761 	 * Respond to 'SYN-' by discarding it & returning 'ACK'-.
2762 	 * Respond to 'SYN+' by queuing it on new socket.
2763 	 */
2764 	if (!msg_data_sz(msg)) {
2765 		struct msghdr m = {NULL,};
2766 
2767 		tsk_advance_rx_queue(sk);
2768 		__tipc_sendstream(new_sock, &m, 0);
2769 	} else {
2770 		__skb_dequeue(&sk->sk_receive_queue);
2771 		__skb_queue_head(&new_sk->sk_receive_queue, buf);
2772 		skb_set_owner_r(buf, new_sk);
2773 	}
2774 	release_sock(new_sk);
2775 exit:
2776 	release_sock(sk);
2777 	return res;
2778 }
2779 
2780 /**
2781  * tipc_shutdown - shutdown socket connection
2782  * @sock: socket structure
2783  * @how: direction to close (must be SHUT_RDWR)
2784  *
2785  * Terminates connection (if necessary), then purges socket's receive queue.
2786  *
2787  * Return: 0 on success, errno otherwise
2788  */
2789 static int tipc_shutdown(struct socket *sock, int how)
2790 {
2791 	struct sock *sk = sock->sk;
2792 	int res;
2793 
2794 	if (how != SHUT_RDWR)
2795 		return -EINVAL;
2796 
2797 	lock_sock(sk);
2798 
2799 	trace_tipc_sk_shutdown(sk, NULL, TIPC_DUMP_ALL, " ");
2800 	__tipc_shutdown(sock, TIPC_CONN_SHUTDOWN);
2801 	sk->sk_shutdown = SHUTDOWN_MASK;
2802 
2803 	if (sk->sk_state == TIPC_DISCONNECTING) {
2804 		/* Discard any unreceived messages */
2805 		__skb_queue_purge(&sk->sk_receive_queue);
2806 
2807 		res = 0;
2808 	} else {
2809 		res = -ENOTCONN;
2810 	}
2811 	/* Wake up anyone sleeping in poll. */
2812 	sk->sk_state_change(sk);
2813 
2814 	release_sock(sk);
2815 	return res;
2816 }
2817 
2818 static void tipc_sk_check_probing_state(struct sock *sk,
2819 					struct sk_buff_head *list)
2820 {
2821 	struct tipc_sock *tsk = tipc_sk(sk);
2822 	u32 pnode = tsk_peer_node(tsk);
2823 	u32 pport = tsk_peer_port(tsk);
2824 	u32 self = tsk_own_node(tsk);
2825 	u32 oport = tsk->portid;
2826 	struct sk_buff *skb;
2827 
2828 	if (tsk->probe_unacked) {
2829 		tipc_set_sk_state(sk, TIPC_DISCONNECTING);
2830 		sk->sk_err = ECONNABORTED;
2831 		tipc_node_remove_conn(sock_net(sk), pnode, pport);
2832 		sk->sk_state_change(sk);
2833 		return;
2834 	}
2835 	/* Prepare new probe */
2836 	skb = tipc_msg_create(CONN_MANAGER, CONN_PROBE, INT_H_SIZE, 0,
2837 			      pnode, self, pport, oport, TIPC_OK);
2838 	if (skb)
2839 		__skb_queue_tail(list, skb);
2840 	tsk->probe_unacked = true;
2841 	sk_reset_timer(sk, &sk->sk_timer, jiffies + CONN_PROBING_INTV);
2842 }
2843 
2844 static void tipc_sk_retry_connect(struct sock *sk, struct sk_buff_head *list)
2845 {
2846 	struct tipc_sock *tsk = tipc_sk(sk);
2847 
2848 	/* Try again later if dest link is congested */
2849 	if (tsk->cong_link_cnt) {
2850 		sk_reset_timer(sk, &sk->sk_timer, msecs_to_jiffies(100));
2851 		return;
2852 	}
2853 	/* Prepare SYN for retransmit */
2854 	tipc_msg_skb_clone(&sk->sk_write_queue, list);
2855 }
2856 
2857 static void tipc_sk_timeout(struct timer_list *t)
2858 {
2859 	struct sock *sk = from_timer(sk, t, sk_timer);
2860 	struct tipc_sock *tsk = tipc_sk(sk);
2861 	u32 pnode = tsk_peer_node(tsk);
2862 	struct sk_buff_head list;
2863 	int rc = 0;
2864 
2865 	__skb_queue_head_init(&list);
2866 	bh_lock_sock(sk);
2867 
2868 	/* Try again later if socket is busy */
2869 	if (sock_owned_by_user(sk)) {
2870 		sk_reset_timer(sk, &sk->sk_timer, jiffies + HZ / 20);
2871 		bh_unlock_sock(sk);
2872 		sock_put(sk);
2873 		return;
2874 	}
2875 
2876 	if (sk->sk_state == TIPC_ESTABLISHED)
2877 		tipc_sk_check_probing_state(sk, &list);
2878 	else if (sk->sk_state == TIPC_CONNECTING)
2879 		tipc_sk_retry_connect(sk, &list);
2880 
2881 	bh_unlock_sock(sk);
2882 
2883 	if (!skb_queue_empty(&list))
2884 		rc = tipc_node_xmit(sock_net(sk), &list, pnode, tsk->portid);
2885 
2886 	/* SYN messages may cause link congestion */
2887 	if (rc == -ELINKCONG) {
2888 		tipc_dest_push(&tsk->cong_links, pnode, 0);
2889 		tsk->cong_link_cnt = 1;
2890 	}
2891 	sock_put(sk);
2892 }
2893 
2894 static int tipc_sk_publish(struct tipc_sock *tsk, uint scope,
2895 			   struct tipc_service_range const *seq)
2896 {
2897 	struct sock *sk = &tsk->sk;
2898 	struct net *net = sock_net(sk);
2899 	struct publication *publ;
2900 	u32 key;
2901 
2902 	if (scope != TIPC_NODE_SCOPE)
2903 		scope = TIPC_CLUSTER_SCOPE;
2904 
2905 	if (tipc_sk_connected(sk))
2906 		return -EINVAL;
2907 	key = tsk->portid + tsk->pub_count + 1;
2908 	if (key == tsk->portid)
2909 		return -EADDRINUSE;
2910 
2911 	publ = tipc_nametbl_publish(net, seq->type, seq->lower, seq->upper,
2912 				    scope, tsk->portid, key);
2913 	if (unlikely(!publ))
2914 		return -EINVAL;
2915 
2916 	list_add(&publ->binding_sock, &tsk->publications);
2917 	tsk->pub_count++;
2918 	tsk->published = 1;
2919 	return 0;
2920 }
2921 
2922 static int tipc_sk_withdraw(struct tipc_sock *tsk, uint scope,
2923 			    struct tipc_service_range const *seq)
2924 {
2925 	struct net *net = sock_net(&tsk->sk);
2926 	struct publication *publ;
2927 	struct publication *safe;
2928 	int rc = -EINVAL;
2929 
2930 	if (scope != TIPC_NODE_SCOPE)
2931 		scope = TIPC_CLUSTER_SCOPE;
2932 
2933 	list_for_each_entry_safe(publ, safe, &tsk->publications, binding_sock) {
2934 		if (seq) {
2935 			if (publ->scope != scope)
2936 				continue;
2937 			if (publ->type != seq->type)
2938 				continue;
2939 			if (publ->lower != seq->lower)
2940 				continue;
2941 			if (publ->upper != seq->upper)
2942 				break;
2943 			tipc_nametbl_withdraw(net, publ->type, publ->lower,
2944 					      publ->upper, publ->key);
2945 			rc = 0;
2946 			break;
2947 		}
2948 		tipc_nametbl_withdraw(net, publ->type, publ->lower,
2949 				      publ->upper, publ->key);
2950 		rc = 0;
2951 	}
2952 	if (list_empty(&tsk->publications))
2953 		tsk->published = 0;
2954 	return rc;
2955 }
2956 
2957 /* tipc_sk_reinit: set non-zero address in all existing sockets
2958  *                 when we go from standalone to network mode.
2959  */
2960 void tipc_sk_reinit(struct net *net)
2961 {
2962 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2963 	struct rhashtable_iter iter;
2964 	struct tipc_sock *tsk;
2965 	struct tipc_msg *msg;
2966 
2967 	rhashtable_walk_enter(&tn->sk_rht, &iter);
2968 
2969 	do {
2970 		rhashtable_walk_start(&iter);
2971 
2972 		while ((tsk = rhashtable_walk_next(&iter)) && !IS_ERR(tsk)) {
2973 			sock_hold(&tsk->sk);
2974 			rhashtable_walk_stop(&iter);
2975 			lock_sock(&tsk->sk);
2976 			msg = &tsk->phdr;
2977 			msg_set_prevnode(msg, tipc_own_addr(net));
2978 			msg_set_orignode(msg, tipc_own_addr(net));
2979 			release_sock(&tsk->sk);
2980 			rhashtable_walk_start(&iter);
2981 			sock_put(&tsk->sk);
2982 		}
2983 
2984 		rhashtable_walk_stop(&iter);
2985 	} while (tsk == ERR_PTR(-EAGAIN));
2986 
2987 	rhashtable_walk_exit(&iter);
2988 }
2989 
2990 static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid)
2991 {
2992 	struct tipc_net *tn = net_generic(net, tipc_net_id);
2993 	struct tipc_sock *tsk;
2994 
2995 	rcu_read_lock();
2996 	tsk = rhashtable_lookup(&tn->sk_rht, &portid, tsk_rht_params);
2997 	if (tsk)
2998 		sock_hold(&tsk->sk);
2999 	rcu_read_unlock();
3000 
3001 	return tsk;
3002 }
3003 
3004 static int tipc_sk_insert(struct tipc_sock *tsk)
3005 {
3006 	struct sock *sk = &tsk->sk;
3007 	struct net *net = sock_net(sk);
3008 	struct tipc_net *tn = net_generic(net, tipc_net_id);
3009 	u32 remaining = (TIPC_MAX_PORT - TIPC_MIN_PORT) + 1;
3010 	u32 portid = prandom_u32() % remaining + TIPC_MIN_PORT;
3011 
3012 	while (remaining--) {
3013 		portid++;
3014 		if ((portid < TIPC_MIN_PORT) || (portid > TIPC_MAX_PORT))
3015 			portid = TIPC_MIN_PORT;
3016 		tsk->portid = portid;
3017 		sock_hold(&tsk->sk);
3018 		if (!rhashtable_lookup_insert_fast(&tn->sk_rht, &tsk->node,
3019 						   tsk_rht_params))
3020 			return 0;
3021 		sock_put(&tsk->sk);
3022 	}
3023 
3024 	return -1;
3025 }
3026 
3027 static void tipc_sk_remove(struct tipc_sock *tsk)
3028 {
3029 	struct sock *sk = &tsk->sk;
3030 	struct tipc_net *tn = net_generic(sock_net(sk), tipc_net_id);
3031 
3032 	if (!rhashtable_remove_fast(&tn->sk_rht, &tsk->node, tsk_rht_params)) {
3033 		WARN_ON(refcount_read(&sk->sk_refcnt) == 1);
3034 		__sock_put(sk);
3035 	}
3036 }
3037 
3038 static const struct rhashtable_params tsk_rht_params = {
3039 	.nelem_hint = 192,
3040 	.head_offset = offsetof(struct tipc_sock, node),
3041 	.key_offset = offsetof(struct tipc_sock, portid),
3042 	.key_len = sizeof(u32), /* portid */
3043 	.max_size = 1048576,
3044 	.min_size = 256,
3045 	.automatic_shrinking = true,
3046 };
3047 
3048 int tipc_sk_rht_init(struct net *net)
3049 {
3050 	struct tipc_net *tn = net_generic(net, tipc_net_id);
3051 
3052 	return rhashtable_init(&tn->sk_rht, &tsk_rht_params);
3053 }
3054 
3055 void tipc_sk_rht_destroy(struct net *net)
3056 {
3057 	struct tipc_net *tn = net_generic(net, tipc_net_id);
3058 
3059 	/* Wait for socket readers to complete */
3060 	synchronize_net();
3061 
3062 	rhashtable_destroy(&tn->sk_rht);
3063 }
3064 
3065 static int tipc_sk_join(struct tipc_sock *tsk, struct tipc_group_req *mreq)
3066 {
3067 	struct net *net = sock_net(&tsk->sk);
3068 	struct tipc_group *grp = tsk->group;
3069 	struct tipc_msg *hdr = &tsk->phdr;
3070 	struct tipc_service_range seq;
3071 	int rc;
3072 
3073 	if (mreq->type < TIPC_RESERVED_TYPES)
3074 		return -EACCES;
3075 	if (mreq->scope > TIPC_NODE_SCOPE)
3076 		return -EINVAL;
3077 	if (grp)
3078 		return -EACCES;
3079 	grp = tipc_group_create(net, tsk->portid, mreq, &tsk->group_is_open);
3080 	if (!grp)
3081 		return -ENOMEM;
3082 	tsk->group = grp;
3083 	msg_set_lookup_scope(hdr, mreq->scope);
3084 	msg_set_nametype(hdr, mreq->type);
3085 	msg_set_dest_droppable(hdr, true);
3086 	seq.type = mreq->type;
3087 	seq.lower = mreq->instance;
3088 	seq.upper = seq.lower;
3089 	tipc_nametbl_build_group(net, grp, mreq->type, mreq->scope);
3090 	rc = tipc_sk_publish(tsk, mreq->scope, &seq);
3091 	if (rc) {
3092 		tipc_group_delete(net, grp);
3093 		tsk->group = NULL;
3094 		return rc;
3095 	}
3096 	/* Eliminate any risk that a broadcast overtakes sent JOINs */
3097 	tsk->mc_method.rcast = true;
3098 	tsk->mc_method.mandatory = true;
3099 	tipc_group_join(net, grp, &tsk->sk.sk_rcvbuf);
3100 	return rc;
3101 }
3102 
3103 static int tipc_sk_leave(struct tipc_sock *tsk)
3104 {
3105 	struct net *net = sock_net(&tsk->sk);
3106 	struct tipc_group *grp = tsk->group;
3107 	struct tipc_service_range seq;
3108 	int scope;
3109 
3110 	if (!grp)
3111 		return -EINVAL;
3112 	tipc_group_self(grp, &seq, &scope);
3113 	tipc_group_delete(net, grp);
3114 	tsk->group = NULL;
3115 	tipc_sk_withdraw(tsk, scope, &seq);
3116 	return 0;
3117 }
3118 
3119 /**
3120  * tipc_setsockopt - set socket option
3121  * @sock: socket structure
3122  * @lvl: option level
3123  * @opt: option identifier
3124  * @ov: pointer to new option value
3125  * @ol: length of option value
3126  *
3127  * For stream sockets only, accepts and ignores all IPPROTO_TCP options
3128  * (to ease compatibility).
3129  *
3130  * Return: 0 on success, errno otherwise
3131  */
3132 static int tipc_setsockopt(struct socket *sock, int lvl, int opt,
3133 			   sockptr_t ov, unsigned int ol)
3134 {
3135 	struct sock *sk = sock->sk;
3136 	struct tipc_sock *tsk = tipc_sk(sk);
3137 	struct tipc_group_req mreq;
3138 	u32 value = 0;
3139 	int res = 0;
3140 
3141 	if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM))
3142 		return 0;
3143 	if (lvl != SOL_TIPC)
3144 		return -ENOPROTOOPT;
3145 
3146 	switch (opt) {
3147 	case TIPC_IMPORTANCE:
3148 	case TIPC_SRC_DROPPABLE:
3149 	case TIPC_DEST_DROPPABLE:
3150 	case TIPC_CONN_TIMEOUT:
3151 	case TIPC_NODELAY:
3152 		if (ol < sizeof(value))
3153 			return -EINVAL;
3154 		if (copy_from_sockptr(&value, ov, sizeof(u32)))
3155 			return -EFAULT;
3156 		break;
3157 	case TIPC_GROUP_JOIN:
3158 		if (ol < sizeof(mreq))
3159 			return -EINVAL;
3160 		if (copy_from_sockptr(&mreq, ov, sizeof(mreq)))
3161 			return -EFAULT;
3162 		break;
3163 	default:
3164 		if (!sockptr_is_null(ov) || ol)
3165 			return -EINVAL;
3166 	}
3167 
3168 	lock_sock(sk);
3169 
3170 	switch (opt) {
3171 	case TIPC_IMPORTANCE:
3172 		res = tsk_set_importance(sk, value);
3173 		break;
3174 	case TIPC_SRC_DROPPABLE:
3175 		if (sock->type != SOCK_STREAM)
3176 			tsk_set_unreliable(tsk, value);
3177 		else
3178 			res = -ENOPROTOOPT;
3179 		break;
3180 	case TIPC_DEST_DROPPABLE:
3181 		tsk_set_unreturnable(tsk, value);
3182 		break;
3183 	case TIPC_CONN_TIMEOUT:
3184 		tipc_sk(sk)->conn_timeout = value;
3185 		break;
3186 	case TIPC_MCAST_BROADCAST:
3187 		tsk->mc_method.rcast = false;
3188 		tsk->mc_method.mandatory = true;
3189 		break;
3190 	case TIPC_MCAST_REPLICAST:
3191 		tsk->mc_method.rcast = true;
3192 		tsk->mc_method.mandatory = true;
3193 		break;
3194 	case TIPC_GROUP_JOIN:
3195 		res = tipc_sk_join(tsk, &mreq);
3196 		break;
3197 	case TIPC_GROUP_LEAVE:
3198 		res = tipc_sk_leave(tsk);
3199 		break;
3200 	case TIPC_NODELAY:
3201 		tsk->nodelay = !!value;
3202 		tsk_set_nagle(tsk);
3203 		break;
3204 	default:
3205 		res = -EINVAL;
3206 	}
3207 
3208 	release_sock(sk);
3209 
3210 	return res;
3211 }
3212 
3213 /**
3214  * tipc_getsockopt - get socket option
3215  * @sock: socket structure
3216  * @lvl: option level
3217  * @opt: option identifier
3218  * @ov: receptacle for option value
3219  * @ol: receptacle for length of option value
3220  *
3221  * For stream sockets only, returns 0 length result for all IPPROTO_TCP options
3222  * (to ease compatibility).
3223  *
3224  * Return: 0 on success, errno otherwise
3225  */
3226 static int tipc_getsockopt(struct socket *sock, int lvl, int opt,
3227 			   char __user *ov, int __user *ol)
3228 {
3229 	struct sock *sk = sock->sk;
3230 	struct tipc_sock *tsk = tipc_sk(sk);
3231 	struct tipc_service_range seq;
3232 	int len, scope;
3233 	u32 value;
3234 	int res;
3235 
3236 	if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM))
3237 		return put_user(0, ol);
3238 	if (lvl != SOL_TIPC)
3239 		return -ENOPROTOOPT;
3240 	res = get_user(len, ol);
3241 	if (res)
3242 		return res;
3243 
3244 	lock_sock(sk);
3245 
3246 	switch (opt) {
3247 	case TIPC_IMPORTANCE:
3248 		value = tsk_importance(tsk);
3249 		break;
3250 	case TIPC_SRC_DROPPABLE:
3251 		value = tsk_unreliable(tsk);
3252 		break;
3253 	case TIPC_DEST_DROPPABLE:
3254 		value = tsk_unreturnable(tsk);
3255 		break;
3256 	case TIPC_CONN_TIMEOUT:
3257 		value = tsk->conn_timeout;
3258 		/* no need to set "res", since already 0 at this point */
3259 		break;
3260 	case TIPC_NODE_RECVQ_DEPTH:
3261 		value = 0; /* was tipc_queue_size, now obsolete */
3262 		break;
3263 	case TIPC_SOCK_RECVQ_DEPTH:
3264 		value = skb_queue_len(&sk->sk_receive_queue);
3265 		break;
3266 	case TIPC_SOCK_RECVQ_USED:
3267 		value = sk_rmem_alloc_get(sk);
3268 		break;
3269 	case TIPC_GROUP_JOIN:
3270 		seq.type = 0;
3271 		if (tsk->group)
3272 			tipc_group_self(tsk->group, &seq, &scope);
3273 		value = seq.type;
3274 		break;
3275 	default:
3276 		res = -EINVAL;
3277 	}
3278 
3279 	release_sock(sk);
3280 
3281 	if (res)
3282 		return res;	/* "get" failed */
3283 
3284 	if (len < sizeof(value))
3285 		return -EINVAL;
3286 
3287 	if (copy_to_user(ov, &value, sizeof(value)))
3288 		return -EFAULT;
3289 
3290 	return put_user(sizeof(value), ol);
3291 }
3292 
3293 static int tipc_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
3294 {
3295 	struct net *net = sock_net(sock->sk);
3296 	struct tipc_sioc_nodeid_req nr = {0};
3297 	struct tipc_sioc_ln_req lnr;
3298 	void __user *argp = (void __user *)arg;
3299 
3300 	switch (cmd) {
3301 	case SIOCGETLINKNAME:
3302 		if (copy_from_user(&lnr, argp, sizeof(lnr)))
3303 			return -EFAULT;
3304 		if (!tipc_node_get_linkname(net,
3305 					    lnr.bearer_id & 0xffff, lnr.peer,
3306 					    lnr.linkname, TIPC_MAX_LINK_NAME)) {
3307 			if (copy_to_user(argp, &lnr, sizeof(lnr)))
3308 				return -EFAULT;
3309 			return 0;
3310 		}
3311 		return -EADDRNOTAVAIL;
3312 	case SIOCGETNODEID:
3313 		if (copy_from_user(&nr, argp, sizeof(nr)))
3314 			return -EFAULT;
3315 		if (!tipc_node_get_id(net, nr.peer, nr.node_id))
3316 			return -EADDRNOTAVAIL;
3317 		if (copy_to_user(argp, &nr, sizeof(nr)))
3318 			return -EFAULT;
3319 		return 0;
3320 	default:
3321 		return -ENOIOCTLCMD;
3322 	}
3323 }
3324 
3325 static int tipc_socketpair(struct socket *sock1, struct socket *sock2)
3326 {
3327 	struct tipc_sock *tsk2 = tipc_sk(sock2->sk);
3328 	struct tipc_sock *tsk1 = tipc_sk(sock1->sk);
3329 	u32 onode = tipc_own_addr(sock_net(sock1->sk));
3330 
3331 	tsk1->peer.family = AF_TIPC;
3332 	tsk1->peer.addrtype = TIPC_SOCKET_ADDR;
3333 	tsk1->peer.scope = TIPC_NODE_SCOPE;
3334 	tsk1->peer.addr.id.ref = tsk2->portid;
3335 	tsk1->peer.addr.id.node = onode;
3336 	tsk2->peer.family = AF_TIPC;
3337 	tsk2->peer.addrtype = TIPC_SOCKET_ADDR;
3338 	tsk2->peer.scope = TIPC_NODE_SCOPE;
3339 	tsk2->peer.addr.id.ref = tsk1->portid;
3340 	tsk2->peer.addr.id.node = onode;
3341 
3342 	tipc_sk_finish_conn(tsk1, tsk2->portid, onode);
3343 	tipc_sk_finish_conn(tsk2, tsk1->portid, onode);
3344 	return 0;
3345 }
3346 
3347 /* Protocol switches for the various types of TIPC sockets */
3348 
3349 static const struct proto_ops msg_ops = {
3350 	.owner		= THIS_MODULE,
3351 	.family		= AF_TIPC,
3352 	.release	= tipc_release,
3353 	.bind		= tipc_bind,
3354 	.connect	= tipc_connect,
3355 	.socketpair	= tipc_socketpair,
3356 	.accept		= sock_no_accept,
3357 	.getname	= tipc_getname,
3358 	.poll		= tipc_poll,
3359 	.ioctl		= tipc_ioctl,
3360 	.listen		= sock_no_listen,
3361 	.shutdown	= tipc_shutdown,
3362 	.setsockopt	= tipc_setsockopt,
3363 	.getsockopt	= tipc_getsockopt,
3364 	.sendmsg	= tipc_sendmsg,
3365 	.recvmsg	= tipc_recvmsg,
3366 	.mmap		= sock_no_mmap,
3367 	.sendpage	= sock_no_sendpage
3368 };
3369 
3370 static const struct proto_ops packet_ops = {
3371 	.owner		= THIS_MODULE,
3372 	.family		= AF_TIPC,
3373 	.release	= tipc_release,
3374 	.bind		= tipc_bind,
3375 	.connect	= tipc_connect,
3376 	.socketpair	= tipc_socketpair,
3377 	.accept		= tipc_accept,
3378 	.getname	= tipc_getname,
3379 	.poll		= tipc_poll,
3380 	.ioctl		= tipc_ioctl,
3381 	.listen		= tipc_listen,
3382 	.shutdown	= tipc_shutdown,
3383 	.setsockopt	= tipc_setsockopt,
3384 	.getsockopt	= tipc_getsockopt,
3385 	.sendmsg	= tipc_send_packet,
3386 	.recvmsg	= tipc_recvmsg,
3387 	.mmap		= sock_no_mmap,
3388 	.sendpage	= sock_no_sendpage
3389 };
3390 
3391 static const struct proto_ops stream_ops = {
3392 	.owner		= THIS_MODULE,
3393 	.family		= AF_TIPC,
3394 	.release	= tipc_release,
3395 	.bind		= tipc_bind,
3396 	.connect	= tipc_connect,
3397 	.socketpair	= tipc_socketpair,
3398 	.accept		= tipc_accept,
3399 	.getname	= tipc_getname,
3400 	.poll		= tipc_poll,
3401 	.ioctl		= tipc_ioctl,
3402 	.listen		= tipc_listen,
3403 	.shutdown	= tipc_shutdown,
3404 	.setsockopt	= tipc_setsockopt,
3405 	.getsockopt	= tipc_getsockopt,
3406 	.sendmsg	= tipc_sendstream,
3407 	.recvmsg	= tipc_recvstream,
3408 	.mmap		= sock_no_mmap,
3409 	.sendpage	= sock_no_sendpage
3410 };
3411 
3412 static const struct net_proto_family tipc_family_ops = {
3413 	.owner		= THIS_MODULE,
3414 	.family		= AF_TIPC,
3415 	.create		= tipc_sk_create
3416 };
3417 
3418 static struct proto tipc_proto = {
3419 	.name		= "TIPC",
3420 	.owner		= THIS_MODULE,
3421 	.obj_size	= sizeof(struct tipc_sock),
3422 	.sysctl_rmem	= sysctl_tipc_rmem
3423 };
3424 
3425 /**
3426  * tipc_socket_init - initialize TIPC socket interface
3427  *
3428  * Return: 0 on success, errno otherwise
3429  */
3430 int tipc_socket_init(void)
3431 {
3432 	int res;
3433 
3434 	res = proto_register(&tipc_proto, 1);
3435 	if (res) {
3436 		pr_err("Failed to register TIPC protocol type\n");
3437 		goto out;
3438 	}
3439 
3440 	res = sock_register(&tipc_family_ops);
3441 	if (res) {
3442 		pr_err("Failed to register TIPC socket type\n");
3443 		proto_unregister(&tipc_proto);
3444 		goto out;
3445 	}
3446  out:
3447 	return res;
3448 }
3449 
3450 /**
3451  * tipc_socket_stop - stop TIPC socket interface
3452  */
3453 void tipc_socket_stop(void)
3454 {
3455 	sock_unregister(tipc_family_ops.family);
3456 	proto_unregister(&tipc_proto);
3457 }
3458 
3459 /* Caller should hold socket lock for the passed tipc socket. */
3460 static int __tipc_nl_add_sk_con(struct sk_buff *skb, struct tipc_sock *tsk)
3461 {
3462 	u32 peer_node;
3463 	u32 peer_port;
3464 	struct nlattr *nest;
3465 
3466 	peer_node = tsk_peer_node(tsk);
3467 	peer_port = tsk_peer_port(tsk);
3468 
3469 	nest = nla_nest_start_noflag(skb, TIPC_NLA_SOCK_CON);
3470 	if (!nest)
3471 		return -EMSGSIZE;
3472 
3473 	if (nla_put_u32(skb, TIPC_NLA_CON_NODE, peer_node))
3474 		goto msg_full;
3475 	if (nla_put_u32(skb, TIPC_NLA_CON_SOCK, peer_port))
3476 		goto msg_full;
3477 
3478 	if (tsk->conn_type != 0) {
3479 		if (nla_put_flag(skb, TIPC_NLA_CON_FLAG))
3480 			goto msg_full;
3481 		if (nla_put_u32(skb, TIPC_NLA_CON_TYPE, tsk->conn_type))
3482 			goto msg_full;
3483 		if (nla_put_u32(skb, TIPC_NLA_CON_INST, tsk->conn_instance))
3484 			goto msg_full;
3485 	}
3486 	nla_nest_end(skb, nest);
3487 
3488 	return 0;
3489 
3490 msg_full:
3491 	nla_nest_cancel(skb, nest);
3492 
3493 	return -EMSGSIZE;
3494 }
3495 
3496 static int __tipc_nl_add_sk_info(struct sk_buff *skb, struct tipc_sock
3497 			  *tsk)
3498 {
3499 	struct net *net = sock_net(skb->sk);
3500 	struct sock *sk = &tsk->sk;
3501 
3502 	if (nla_put_u32(skb, TIPC_NLA_SOCK_REF, tsk->portid) ||
3503 	    nla_put_u32(skb, TIPC_NLA_SOCK_ADDR, tipc_own_addr(net)))
3504 		return -EMSGSIZE;
3505 
3506 	if (tipc_sk_connected(sk)) {
3507 		if (__tipc_nl_add_sk_con(skb, tsk))
3508 			return -EMSGSIZE;
3509 	} else if (!list_empty(&tsk->publications)) {
3510 		if (nla_put_flag(skb, TIPC_NLA_SOCK_HAS_PUBL))
3511 			return -EMSGSIZE;
3512 	}
3513 	return 0;
3514 }
3515 
3516 /* Caller should hold socket lock for the passed tipc socket. */
3517 static int __tipc_nl_add_sk(struct sk_buff *skb, struct netlink_callback *cb,
3518 			    struct tipc_sock *tsk)
3519 {
3520 	struct nlattr *attrs;
3521 	void *hdr;
3522 
3523 	hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
3524 			  &tipc_genl_family, NLM_F_MULTI, TIPC_NL_SOCK_GET);
3525 	if (!hdr)
3526 		goto msg_cancel;
3527 
3528 	attrs = nla_nest_start_noflag(skb, TIPC_NLA_SOCK);
3529 	if (!attrs)
3530 		goto genlmsg_cancel;
3531 
3532 	if (__tipc_nl_add_sk_info(skb, tsk))
3533 		goto attr_msg_cancel;
3534 
3535 	nla_nest_end(skb, attrs);
3536 	genlmsg_end(skb, hdr);
3537 
3538 	return 0;
3539 
3540 attr_msg_cancel:
3541 	nla_nest_cancel(skb, attrs);
3542 genlmsg_cancel:
3543 	genlmsg_cancel(skb, hdr);
3544 msg_cancel:
3545 	return -EMSGSIZE;
3546 }
3547 
3548 int tipc_nl_sk_walk(struct sk_buff *skb, struct netlink_callback *cb,
3549 		    int (*skb_handler)(struct sk_buff *skb,
3550 				       struct netlink_callback *cb,
3551 				       struct tipc_sock *tsk))
3552 {
3553 	struct rhashtable_iter *iter = (void *)cb->args[4];
3554 	struct tipc_sock *tsk;
3555 	int err;
3556 
3557 	rhashtable_walk_start(iter);
3558 	while ((tsk = rhashtable_walk_next(iter)) != NULL) {
3559 		if (IS_ERR(tsk)) {
3560 			err = PTR_ERR(tsk);
3561 			if (err == -EAGAIN) {
3562 				err = 0;
3563 				continue;
3564 			}
3565 			break;
3566 		}
3567 
3568 		sock_hold(&tsk->sk);
3569 		rhashtable_walk_stop(iter);
3570 		lock_sock(&tsk->sk);
3571 		err = skb_handler(skb, cb, tsk);
3572 		if (err) {
3573 			release_sock(&tsk->sk);
3574 			sock_put(&tsk->sk);
3575 			goto out;
3576 		}
3577 		release_sock(&tsk->sk);
3578 		rhashtable_walk_start(iter);
3579 		sock_put(&tsk->sk);
3580 	}
3581 	rhashtable_walk_stop(iter);
3582 out:
3583 	return skb->len;
3584 }
3585 EXPORT_SYMBOL(tipc_nl_sk_walk);
3586 
3587 int tipc_dump_start(struct netlink_callback *cb)
3588 {
3589 	return __tipc_dump_start(cb, sock_net(cb->skb->sk));
3590 }
3591 EXPORT_SYMBOL(tipc_dump_start);
3592 
3593 int __tipc_dump_start(struct netlink_callback *cb, struct net *net)
3594 {
3595 	/* tipc_nl_name_table_dump() uses cb->args[0...3]. */
3596 	struct rhashtable_iter *iter = (void *)cb->args[4];
3597 	struct tipc_net *tn = tipc_net(net);
3598 
3599 	if (!iter) {
3600 		iter = kmalloc(sizeof(*iter), GFP_KERNEL);
3601 		if (!iter)
3602 			return -ENOMEM;
3603 
3604 		cb->args[4] = (long)iter;
3605 	}
3606 
3607 	rhashtable_walk_enter(&tn->sk_rht, iter);
3608 	return 0;
3609 }
3610 
3611 int tipc_dump_done(struct netlink_callback *cb)
3612 {
3613 	struct rhashtable_iter *hti = (void *)cb->args[4];
3614 
3615 	rhashtable_walk_exit(hti);
3616 	kfree(hti);
3617 	return 0;
3618 }
3619 EXPORT_SYMBOL(tipc_dump_done);
3620 
3621 int tipc_sk_fill_sock_diag(struct sk_buff *skb, struct netlink_callback *cb,
3622 			   struct tipc_sock *tsk, u32 sk_filter_state,
3623 			   u64 (*tipc_diag_gen_cookie)(struct sock *sk))
3624 {
3625 	struct sock *sk = &tsk->sk;
3626 	struct nlattr *attrs;
3627 	struct nlattr *stat;
3628 
3629 	/*filter response w.r.t sk_state*/
3630 	if (!(sk_filter_state & (1 << sk->sk_state)))
3631 		return 0;
3632 
3633 	attrs = nla_nest_start_noflag(skb, TIPC_NLA_SOCK);
3634 	if (!attrs)
3635 		goto msg_cancel;
3636 
3637 	if (__tipc_nl_add_sk_info(skb, tsk))
3638 		goto attr_msg_cancel;
3639 
3640 	if (nla_put_u32(skb, TIPC_NLA_SOCK_TYPE, (u32)sk->sk_type) ||
3641 	    nla_put_u32(skb, TIPC_NLA_SOCK_TIPC_STATE, (u32)sk->sk_state) ||
3642 	    nla_put_u32(skb, TIPC_NLA_SOCK_INO, sock_i_ino(sk)) ||
3643 	    nla_put_u32(skb, TIPC_NLA_SOCK_UID,
3644 			from_kuid_munged(sk_user_ns(NETLINK_CB(cb->skb).sk),
3645 					 sock_i_uid(sk))) ||
3646 	    nla_put_u64_64bit(skb, TIPC_NLA_SOCK_COOKIE,
3647 			      tipc_diag_gen_cookie(sk),
3648 			      TIPC_NLA_SOCK_PAD))
3649 		goto attr_msg_cancel;
3650 
3651 	stat = nla_nest_start_noflag(skb, TIPC_NLA_SOCK_STAT);
3652 	if (!stat)
3653 		goto attr_msg_cancel;
3654 
3655 	if (nla_put_u32(skb, TIPC_NLA_SOCK_STAT_RCVQ,
3656 			skb_queue_len(&sk->sk_receive_queue)) ||
3657 	    nla_put_u32(skb, TIPC_NLA_SOCK_STAT_SENDQ,
3658 			skb_queue_len(&sk->sk_write_queue)) ||
3659 	    nla_put_u32(skb, TIPC_NLA_SOCK_STAT_DROP,
3660 			atomic_read(&sk->sk_drops)))
3661 		goto stat_msg_cancel;
3662 
3663 	if (tsk->cong_link_cnt &&
3664 	    nla_put_flag(skb, TIPC_NLA_SOCK_STAT_LINK_CONG))
3665 		goto stat_msg_cancel;
3666 
3667 	if (tsk_conn_cong(tsk) &&
3668 	    nla_put_flag(skb, TIPC_NLA_SOCK_STAT_CONN_CONG))
3669 		goto stat_msg_cancel;
3670 
3671 	nla_nest_end(skb, stat);
3672 
3673 	if (tsk->group)
3674 		if (tipc_group_fill_sock_diag(tsk->group, skb))
3675 			goto stat_msg_cancel;
3676 
3677 	nla_nest_end(skb, attrs);
3678 
3679 	return 0;
3680 
3681 stat_msg_cancel:
3682 	nla_nest_cancel(skb, stat);
3683 attr_msg_cancel:
3684 	nla_nest_cancel(skb, attrs);
3685 msg_cancel:
3686 	return -EMSGSIZE;
3687 }
3688 EXPORT_SYMBOL(tipc_sk_fill_sock_diag);
3689 
3690 int tipc_nl_sk_dump(struct sk_buff *skb, struct netlink_callback *cb)
3691 {
3692 	return tipc_nl_sk_walk(skb, cb, __tipc_nl_add_sk);
3693 }
3694 
3695 /* Caller should hold socket lock for the passed tipc socket. */
3696 static int __tipc_nl_add_sk_publ(struct sk_buff *skb,
3697 				 struct netlink_callback *cb,
3698 				 struct publication *publ)
3699 {
3700 	void *hdr;
3701 	struct nlattr *attrs;
3702 
3703 	hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq,
3704 			  &tipc_genl_family, NLM_F_MULTI, TIPC_NL_PUBL_GET);
3705 	if (!hdr)
3706 		goto msg_cancel;
3707 
3708 	attrs = nla_nest_start_noflag(skb, TIPC_NLA_PUBL);
3709 	if (!attrs)
3710 		goto genlmsg_cancel;
3711 
3712 	if (nla_put_u32(skb, TIPC_NLA_PUBL_KEY, publ->key))
3713 		goto attr_msg_cancel;
3714 	if (nla_put_u32(skb, TIPC_NLA_PUBL_TYPE, publ->type))
3715 		goto attr_msg_cancel;
3716 	if (nla_put_u32(skb, TIPC_NLA_PUBL_LOWER, publ->lower))
3717 		goto attr_msg_cancel;
3718 	if (nla_put_u32(skb, TIPC_NLA_PUBL_UPPER, publ->upper))
3719 		goto attr_msg_cancel;
3720 
3721 	nla_nest_end(skb, attrs);
3722 	genlmsg_end(skb, hdr);
3723 
3724 	return 0;
3725 
3726 attr_msg_cancel:
3727 	nla_nest_cancel(skb, attrs);
3728 genlmsg_cancel:
3729 	genlmsg_cancel(skb, hdr);
3730 msg_cancel:
3731 	return -EMSGSIZE;
3732 }
3733 
3734 /* Caller should hold socket lock for the passed tipc socket. */
3735 static int __tipc_nl_list_sk_publ(struct sk_buff *skb,
3736 				  struct netlink_callback *cb,
3737 				  struct tipc_sock *tsk, u32 *last_publ)
3738 {
3739 	int err;
3740 	struct publication *p;
3741 
3742 	if (*last_publ) {
3743 		list_for_each_entry(p, &tsk->publications, binding_sock) {
3744 			if (p->key == *last_publ)
3745 				break;
3746 		}
3747 		if (p->key != *last_publ) {
3748 			/* We never set seq or call nl_dump_check_consistent()
3749 			 * this means that setting prev_seq here will cause the
3750 			 * consistence check to fail in the netlink callback
3751 			 * handler. Resulting in the last NLMSG_DONE message
3752 			 * having the NLM_F_DUMP_INTR flag set.
3753 			 */
3754 			cb->prev_seq = 1;
3755 			*last_publ = 0;
3756 			return -EPIPE;
3757 		}
3758 	} else {
3759 		p = list_first_entry(&tsk->publications, struct publication,
3760 				     binding_sock);
3761 	}
3762 
3763 	list_for_each_entry_from(p, &tsk->publications, binding_sock) {
3764 		err = __tipc_nl_add_sk_publ(skb, cb, p);
3765 		if (err) {
3766 			*last_publ = p->key;
3767 			return err;
3768 		}
3769 	}
3770 	*last_publ = 0;
3771 
3772 	return 0;
3773 }
3774 
3775 int tipc_nl_publ_dump(struct sk_buff *skb, struct netlink_callback *cb)
3776 {
3777 	int err;
3778 	u32 tsk_portid = cb->args[0];
3779 	u32 last_publ = cb->args[1];
3780 	u32 done = cb->args[2];
3781 	struct net *net = sock_net(skb->sk);
3782 	struct tipc_sock *tsk;
3783 
3784 	if (!tsk_portid) {
3785 		struct nlattr **attrs = genl_dumpit_info(cb)->attrs;
3786 		struct nlattr *sock[TIPC_NLA_SOCK_MAX + 1];
3787 
3788 		if (!attrs[TIPC_NLA_SOCK])
3789 			return -EINVAL;
3790 
3791 		err = nla_parse_nested_deprecated(sock, TIPC_NLA_SOCK_MAX,
3792 						  attrs[TIPC_NLA_SOCK],
3793 						  tipc_nl_sock_policy, NULL);
3794 		if (err)
3795 			return err;
3796 
3797 		if (!sock[TIPC_NLA_SOCK_REF])
3798 			return -EINVAL;
3799 
3800 		tsk_portid = nla_get_u32(sock[TIPC_NLA_SOCK_REF]);
3801 	}
3802 
3803 	if (done)
3804 		return 0;
3805 
3806 	tsk = tipc_sk_lookup(net, tsk_portid);
3807 	if (!tsk)
3808 		return -EINVAL;
3809 
3810 	lock_sock(&tsk->sk);
3811 	err = __tipc_nl_list_sk_publ(skb, cb, tsk, &last_publ);
3812 	if (!err)
3813 		done = 1;
3814 	release_sock(&tsk->sk);
3815 	sock_put(&tsk->sk);
3816 
3817 	cb->args[0] = tsk_portid;
3818 	cb->args[1] = last_publ;
3819 	cb->args[2] = done;
3820 
3821 	return skb->len;
3822 }
3823 
3824 /**
3825  * tipc_sk_filtering - check if a socket should be traced
3826  * @sk: the socket to be examined
3827  *
3828  * @sysctl_tipc_sk_filter is used as the socket tuple for filtering:
3829  * (portid, sock type, name type, name lower, name upper)
3830  *
3831  * Return: true if the socket meets the socket tuple data
3832  * (value 0 = 'any') or when there is no tuple set (all = 0),
3833  * otherwise false
3834  */
3835 bool tipc_sk_filtering(struct sock *sk)
3836 {
3837 	struct tipc_sock *tsk;
3838 	struct publication *p;
3839 	u32 _port, _sktype, _type, _lower, _upper;
3840 	u32 type = 0, lower = 0, upper = 0;
3841 
3842 	if (!sk)
3843 		return true;
3844 
3845 	tsk = tipc_sk(sk);
3846 
3847 	_port = sysctl_tipc_sk_filter[0];
3848 	_sktype = sysctl_tipc_sk_filter[1];
3849 	_type = sysctl_tipc_sk_filter[2];
3850 	_lower = sysctl_tipc_sk_filter[3];
3851 	_upper = sysctl_tipc_sk_filter[4];
3852 
3853 	if (!_port && !_sktype && !_type && !_lower && !_upper)
3854 		return true;
3855 
3856 	if (_port)
3857 		return (_port == tsk->portid);
3858 
3859 	if (_sktype && _sktype != sk->sk_type)
3860 		return false;
3861 
3862 	if (tsk->published) {
3863 		p = list_first_entry_or_null(&tsk->publications,
3864 					     struct publication, binding_sock);
3865 		if (p) {
3866 			type = p->type;
3867 			lower = p->lower;
3868 			upper = p->upper;
3869 		}
3870 	}
3871 
3872 	if (!tipc_sk_type_connectionless(sk)) {
3873 		type = tsk->conn_type;
3874 		lower = tsk->conn_instance;
3875 		upper = tsk->conn_instance;
3876 	}
3877 
3878 	if ((_type && _type != type) || (_lower && _lower != lower) ||
3879 	    (_upper && _upper != upper))
3880 		return false;
3881 
3882 	return true;
3883 }
3884 
3885 u32 tipc_sock_get_portid(struct sock *sk)
3886 {
3887 	return (sk) ? (tipc_sk(sk))->portid : 0;
3888 }
3889 
3890 /**
3891  * tipc_sk_overlimit1 - check if socket rx queue is about to be overloaded,
3892  *			both the rcv and backlog queues are considered
3893  * @sk: tipc sk to be checked
3894  * @skb: tipc msg to be checked
3895  *
3896  * Return: true if the socket rx queue allocation is > 90%, otherwise false
3897  */
3898 
3899 bool tipc_sk_overlimit1(struct sock *sk, struct sk_buff *skb)
3900 {
3901 	atomic_t *dcnt = &tipc_sk(sk)->dupl_rcvcnt;
3902 	unsigned int lim = rcvbuf_limit(sk, skb) + atomic_read(dcnt);
3903 	unsigned int qsize = sk->sk_backlog.len + sk_rmem_alloc_get(sk);
3904 
3905 	return (qsize > lim * 90 / 100);
3906 }
3907 
3908 /**
3909  * tipc_sk_overlimit2 - check if socket rx queue is about to be overloaded,
3910  *			only the rcv queue is considered
3911  * @sk: tipc sk to be checked
3912  * @skb: tipc msg to be checked
3913  *
3914  * Return: true if the socket rx queue allocation is > 90%, otherwise false
3915  */
3916 
3917 bool tipc_sk_overlimit2(struct sock *sk, struct sk_buff *skb)
3918 {
3919 	unsigned int lim = rcvbuf_limit(sk, skb);
3920 	unsigned int qsize = sk_rmem_alloc_get(sk);
3921 
3922 	return (qsize > lim * 90 / 100);
3923 }
3924 
3925 /**
3926  * tipc_sk_dump - dump TIPC socket
3927  * @sk: tipc sk to be dumped
3928  * @dqueues: bitmask to decide if any socket queue to be dumped?
3929  *           - TIPC_DUMP_NONE: don't dump socket queues
3930  *           - TIPC_DUMP_SK_SNDQ: dump socket send queue
3931  *           - TIPC_DUMP_SK_RCVQ: dump socket rcv queue
3932  *           - TIPC_DUMP_SK_BKLGQ: dump socket backlog queue
3933  *           - TIPC_DUMP_ALL: dump all the socket queues above
3934  * @buf: returned buffer of dump data in format
3935  */
3936 int tipc_sk_dump(struct sock *sk, u16 dqueues, char *buf)
3937 {
3938 	int i = 0;
3939 	size_t sz = (dqueues) ? SK_LMAX : SK_LMIN;
3940 	struct tipc_sock *tsk;
3941 	struct publication *p;
3942 	bool tsk_connected;
3943 
3944 	if (!sk) {
3945 		i += scnprintf(buf, sz, "sk data: (null)\n");
3946 		return i;
3947 	}
3948 
3949 	tsk = tipc_sk(sk);
3950 	tsk_connected = !tipc_sk_type_connectionless(sk);
3951 
3952 	i += scnprintf(buf, sz, "sk data: %u", sk->sk_type);
3953 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_state);
3954 	i += scnprintf(buf + i, sz - i, " %x", tsk_own_node(tsk));
3955 	i += scnprintf(buf + i, sz - i, " %u", tsk->portid);
3956 	i += scnprintf(buf + i, sz - i, " | %u", tsk_connected);
3957 	if (tsk_connected) {
3958 		i += scnprintf(buf + i, sz - i, " %x", tsk_peer_node(tsk));
3959 		i += scnprintf(buf + i, sz - i, " %u", tsk_peer_port(tsk));
3960 		i += scnprintf(buf + i, sz - i, " %u", tsk->conn_type);
3961 		i += scnprintf(buf + i, sz - i, " %u", tsk->conn_instance);
3962 	}
3963 	i += scnprintf(buf + i, sz - i, " | %u", tsk->published);
3964 	if (tsk->published) {
3965 		p = list_first_entry_or_null(&tsk->publications,
3966 					     struct publication, binding_sock);
3967 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->type : 0);
3968 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->lower : 0);
3969 		i += scnprintf(buf + i, sz - i, " %u", (p) ? p->upper : 0);
3970 	}
3971 	i += scnprintf(buf + i, sz - i, " | %u", tsk->snd_win);
3972 	i += scnprintf(buf + i, sz - i, " %u", tsk->rcv_win);
3973 	i += scnprintf(buf + i, sz - i, " %u", tsk->max_pkt);
3974 	i += scnprintf(buf + i, sz - i, " %x", tsk->peer_caps);
3975 	i += scnprintf(buf + i, sz - i, " %u", tsk->cong_link_cnt);
3976 	i += scnprintf(buf + i, sz - i, " %u", tsk->snt_unacked);
3977 	i += scnprintf(buf + i, sz - i, " %u", tsk->rcv_unacked);
3978 	i += scnprintf(buf + i, sz - i, " %u", atomic_read(&tsk->dupl_rcvcnt));
3979 	i += scnprintf(buf + i, sz - i, " %u", sk->sk_shutdown);
3980 	i += scnprintf(buf + i, sz - i, " | %d", sk_wmem_alloc_get(sk));
3981 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_sndbuf);
3982 	i += scnprintf(buf + i, sz - i, " | %d", sk_rmem_alloc_get(sk));
3983 	i += scnprintf(buf + i, sz - i, " %d", sk->sk_rcvbuf);
3984 	i += scnprintf(buf + i, sz - i, " | %d\n", READ_ONCE(sk->sk_backlog.len));
3985 
3986 	if (dqueues & TIPC_DUMP_SK_SNDQ) {
3987 		i += scnprintf(buf + i, sz - i, "sk_write_queue: ");
3988 		i += tipc_list_dump(&sk->sk_write_queue, false, buf + i);
3989 	}
3990 
3991 	if (dqueues & TIPC_DUMP_SK_RCVQ) {
3992 		i += scnprintf(buf + i, sz - i, "sk_receive_queue: ");
3993 		i += tipc_list_dump(&sk->sk_receive_queue, false, buf + i);
3994 	}
3995 
3996 	if (dqueues & TIPC_DUMP_SK_BKLGQ) {
3997 		i += scnprintf(buf + i, sz - i, "sk_backlog:\n  head ");
3998 		i += tipc_skb_dump(sk->sk_backlog.head, false, buf + i);
3999 		if (sk->sk_backlog.tail != sk->sk_backlog.head) {
4000 			i += scnprintf(buf + i, sz - i, "  tail ");
4001 			i += tipc_skb_dump(sk->sk_backlog.tail, false,
4002 					   buf + i);
4003 		}
4004 	}
4005 
4006 	return i;
4007 }
4008