xref: /openbmc/linux/net/tipc/bcast.c (revision c4c3c32d)
1 /*
2  * net/tipc/bcast.c: TIPC broadcast code
3  *
4  * Copyright (c) 2004-2006, 2014-2017, Ericsson AB
5  * Copyright (c) 2004, Intel Corporation.
6  * Copyright (c) 2005, 2010-2011, Wind River Systems
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/tipc_config.h>
39 #include "socket.h"
40 #include "msg.h"
41 #include "bcast.h"
42 #include "link.h"
43 #include "name_table.h"
44 
45 #define BCLINK_WIN_DEFAULT  50	/* bcast link window size (default) */
46 #define BCLINK_WIN_MIN      32	/* bcast minimum link window size */
47 
48 const char tipc_bclink_name[] = "broadcast-link";
49 unsigned long sysctl_tipc_bc_retruni __read_mostly;
50 
51 /**
52  * struct tipc_bc_base - base structure for keeping broadcast send state
53  * @link: broadcast send link structure
54  * @inputq: data input queue; will only carry SOCK_WAKEUP messages
55  * @dests: array keeping number of reachable destinations per bearer
56  * @primary_bearer: a bearer having links to all broadcast destinations, if any
57  * @bcast_support: indicates if primary bearer, if any, supports broadcast
58  * @force_bcast: forces broadcast for multicast traffic
59  * @rcast_support: indicates if all peer nodes support replicast
60  * @force_rcast: forces replicast for multicast traffic
61  * @rc_ratio: dest count as percentage of cluster size where send method changes
62  * @bc_threshold: calculated from rc_ratio; if dests > threshold use broadcast
63  */
64 struct tipc_bc_base {
65 	struct tipc_link *link;
66 	struct sk_buff_head inputq;
67 	int dests[MAX_BEARERS];
68 	int primary_bearer;
69 	bool bcast_support;
70 	bool force_bcast;
71 	bool rcast_support;
72 	bool force_rcast;
73 	int rc_ratio;
74 	int bc_threshold;
75 };
76 
77 static struct tipc_bc_base *tipc_bc_base(struct net *net)
78 {
79 	return tipc_net(net)->bcbase;
80 }
81 
82 /* tipc_bcast_get_mtu(): -get the MTU currently used by broadcast link
83  * Note: the MTU is decremented to give room for a tunnel header, in
84  * case the message needs to be sent as replicast
85  */
86 int tipc_bcast_get_mtu(struct net *net)
87 {
88 	return tipc_link_mss(tipc_bc_sndlink(net));
89 }
90 
91 void tipc_bcast_toggle_rcast(struct net *net, bool supp)
92 {
93 	tipc_bc_base(net)->rcast_support = supp;
94 }
95 
96 static void tipc_bcbase_calc_bc_threshold(struct net *net)
97 {
98 	struct tipc_bc_base *bb = tipc_bc_base(net);
99 	int cluster_size = tipc_link_bc_peers(tipc_bc_sndlink(net));
100 
101 	bb->bc_threshold = 1 + (cluster_size * bb->rc_ratio / 100);
102 }
103 
104 /* tipc_bcbase_select_primary(): find a bearer with links to all destinations,
105  *                               if any, and make it primary bearer
106  */
107 static void tipc_bcbase_select_primary(struct net *net)
108 {
109 	struct tipc_bc_base *bb = tipc_bc_base(net);
110 	int all_dests =  tipc_link_bc_peers(bb->link);
111 	int max_win = tipc_link_max_win(bb->link);
112 	int min_win = tipc_link_min_win(bb->link);
113 	int i, mtu, prim;
114 
115 	bb->primary_bearer = INVALID_BEARER_ID;
116 	bb->bcast_support = true;
117 
118 	if (!all_dests)
119 		return;
120 
121 	for (i = 0; i < MAX_BEARERS; i++) {
122 		if (!bb->dests[i])
123 			continue;
124 
125 		mtu = tipc_bearer_mtu(net, i);
126 		if (mtu < tipc_link_mtu(bb->link)) {
127 			tipc_link_set_mtu(bb->link, mtu);
128 			tipc_link_set_queue_limits(bb->link,
129 						   min_win,
130 						   max_win);
131 		}
132 		bb->bcast_support &= tipc_bearer_bcast_support(net, i);
133 		if (bb->dests[i] < all_dests)
134 			continue;
135 
136 		bb->primary_bearer = i;
137 
138 		/* Reduce risk that all nodes select same primary */
139 		if ((i ^ tipc_own_addr(net)) & 1)
140 			break;
141 	}
142 	prim = bb->primary_bearer;
143 	if (prim != INVALID_BEARER_ID)
144 		bb->bcast_support = tipc_bearer_bcast_support(net, prim);
145 }
146 
147 void tipc_bcast_inc_bearer_dst_cnt(struct net *net, int bearer_id)
148 {
149 	struct tipc_bc_base *bb = tipc_bc_base(net);
150 
151 	tipc_bcast_lock(net);
152 	bb->dests[bearer_id]++;
153 	tipc_bcbase_select_primary(net);
154 	tipc_bcast_unlock(net);
155 }
156 
157 void tipc_bcast_dec_bearer_dst_cnt(struct net *net, int bearer_id)
158 {
159 	struct tipc_bc_base *bb = tipc_bc_base(net);
160 
161 	tipc_bcast_lock(net);
162 	bb->dests[bearer_id]--;
163 	tipc_bcbase_select_primary(net);
164 	tipc_bcast_unlock(net);
165 }
166 
167 /* tipc_bcbase_xmit - broadcast a packet queue across one or more bearers
168  *
169  * Note that number of reachable destinations, as indicated in the dests[]
170  * array, may transitionally differ from the number of destinations indicated
171  * in each sent buffer. We can sustain this. Excess destination nodes will
172  * drop and never acknowledge the unexpected packets, and missing destinations
173  * will either require retransmission (if they are just about to be added to
174  * the bearer), or be removed from the buffer's 'ackers' counter (if they
175  * just went down)
176  */
177 static void tipc_bcbase_xmit(struct net *net, struct sk_buff_head *xmitq)
178 {
179 	int bearer_id;
180 	struct tipc_bc_base *bb = tipc_bc_base(net);
181 	struct sk_buff *skb, *_skb;
182 	struct sk_buff_head _xmitq;
183 
184 	if (skb_queue_empty(xmitq))
185 		return;
186 
187 	/* The typical case: at least one bearer has links to all nodes */
188 	bearer_id = bb->primary_bearer;
189 	if (bearer_id >= 0) {
190 		tipc_bearer_bc_xmit(net, bearer_id, xmitq);
191 		return;
192 	}
193 
194 	/* We have to transmit across all bearers */
195 	__skb_queue_head_init(&_xmitq);
196 	for (bearer_id = 0; bearer_id < MAX_BEARERS; bearer_id++) {
197 		if (!bb->dests[bearer_id])
198 			continue;
199 
200 		skb_queue_walk(xmitq, skb) {
201 			_skb = pskb_copy_for_clone(skb, GFP_ATOMIC);
202 			if (!_skb)
203 				break;
204 			__skb_queue_tail(&_xmitq, _skb);
205 		}
206 		tipc_bearer_bc_xmit(net, bearer_id, &_xmitq);
207 	}
208 	__skb_queue_purge(xmitq);
209 	__skb_queue_purge(&_xmitq);
210 }
211 
212 static void tipc_bcast_select_xmit_method(struct net *net, int dests,
213 					  struct tipc_mc_method *method)
214 {
215 	struct tipc_bc_base *bb = tipc_bc_base(net);
216 	unsigned long exp = method->expires;
217 
218 	/* Broadcast supported by used bearer/bearers? */
219 	if (!bb->bcast_support) {
220 		method->rcast = true;
221 		return;
222 	}
223 	/* Any destinations which don't support replicast ? */
224 	if (!bb->rcast_support) {
225 		method->rcast = false;
226 		return;
227 	}
228 	/* Can current method be changed ? */
229 	method->expires = jiffies + TIPC_METHOD_EXPIRE;
230 	if (method->mandatory)
231 		return;
232 
233 	if (!(tipc_net(net)->capabilities & TIPC_MCAST_RBCTL) &&
234 	    time_before(jiffies, exp))
235 		return;
236 
237 	/* Configuration as force 'broadcast' method */
238 	if (bb->force_bcast) {
239 		method->rcast = false;
240 		return;
241 	}
242 	/* Configuration as force 'replicast' method */
243 	if (bb->force_rcast) {
244 		method->rcast = true;
245 		return;
246 	}
247 	/* Configuration as 'autoselect' or default method */
248 	/* Determine method to use now */
249 	method->rcast = dests <= bb->bc_threshold;
250 }
251 
252 /* tipc_bcast_xmit - broadcast the buffer chain to all external nodes
253  * @net: the applicable net namespace
254  * @pkts: chain of buffers containing message
255  * @cong_link_cnt: set to 1 if broadcast link is congested, otherwise 0
256  * Consumes the buffer chain.
257  * Returns 0 if success, otherwise errno: -EHOSTUNREACH,-EMSGSIZE
258  */
259 int tipc_bcast_xmit(struct net *net, struct sk_buff_head *pkts,
260 		    u16 *cong_link_cnt)
261 {
262 	struct tipc_link *l = tipc_bc_sndlink(net);
263 	struct sk_buff_head xmitq;
264 	int rc = 0;
265 
266 	__skb_queue_head_init(&xmitq);
267 	tipc_bcast_lock(net);
268 	if (tipc_link_bc_peers(l))
269 		rc = tipc_link_xmit(l, pkts, &xmitq);
270 	tipc_bcast_unlock(net);
271 	tipc_bcbase_xmit(net, &xmitq);
272 	__skb_queue_purge(pkts);
273 	if (rc == -ELINKCONG) {
274 		*cong_link_cnt = 1;
275 		rc = 0;
276 	}
277 	return rc;
278 }
279 
280 /* tipc_rcast_xmit - replicate and send a message to given destination nodes
281  * @net: the applicable net namespace
282  * @pkts: chain of buffers containing message
283  * @dests: list of destination nodes
284  * @cong_link_cnt: returns number of congested links
285  * @cong_links: returns identities of congested links
286  * Returns 0 if success, otherwise errno
287  */
288 static int tipc_rcast_xmit(struct net *net, struct sk_buff_head *pkts,
289 			   struct tipc_nlist *dests, u16 *cong_link_cnt)
290 {
291 	struct tipc_dest *dst, *tmp;
292 	struct sk_buff_head _pkts;
293 	u32 dnode, selector;
294 
295 	selector = msg_link_selector(buf_msg(skb_peek(pkts)));
296 	__skb_queue_head_init(&_pkts);
297 
298 	list_for_each_entry_safe(dst, tmp, &dests->list, list) {
299 		dnode = dst->node;
300 		if (!tipc_msg_pskb_copy(dnode, pkts, &_pkts))
301 			return -ENOMEM;
302 
303 		/* Any other return value than -ELINKCONG is ignored */
304 		if (tipc_node_xmit(net, &_pkts, dnode, selector) == -ELINKCONG)
305 			(*cong_link_cnt)++;
306 	}
307 	return 0;
308 }
309 
310 /* tipc_mcast_send_sync - deliver a dummy message with SYN bit
311  * @net: the applicable net namespace
312  * @skb: socket buffer to copy
313  * @method: send method to be used
314  * @dests: destination nodes for message.
315  * Returns 0 if success, otherwise errno
316  */
317 static int tipc_mcast_send_sync(struct net *net, struct sk_buff *skb,
318 				struct tipc_mc_method *method,
319 				struct tipc_nlist *dests)
320 {
321 	struct tipc_msg *hdr, *_hdr;
322 	struct sk_buff_head tmpq;
323 	struct sk_buff *_skb;
324 	u16 cong_link_cnt;
325 	int rc = 0;
326 
327 	/* Is a cluster supporting with new capabilities ? */
328 	if (!(tipc_net(net)->capabilities & TIPC_MCAST_RBCTL))
329 		return 0;
330 
331 	hdr = buf_msg(skb);
332 	if (msg_user(hdr) == MSG_FRAGMENTER)
333 		hdr = msg_inner_hdr(hdr);
334 	if (msg_type(hdr) != TIPC_MCAST_MSG)
335 		return 0;
336 
337 	/* Allocate dummy message */
338 	_skb = tipc_buf_acquire(MCAST_H_SIZE, GFP_KERNEL);
339 	if (!_skb)
340 		return -ENOMEM;
341 
342 	/* Preparing for 'synching' header */
343 	msg_set_syn(hdr, 1);
344 
345 	/* Copy skb's header into a dummy header */
346 	skb_copy_to_linear_data(_skb, hdr, MCAST_H_SIZE);
347 	skb_orphan(_skb);
348 
349 	/* Reverse method for dummy message */
350 	_hdr = buf_msg(_skb);
351 	msg_set_size(_hdr, MCAST_H_SIZE);
352 	msg_set_is_rcast(_hdr, !msg_is_rcast(hdr));
353 	msg_set_errcode(_hdr, TIPC_ERR_NO_PORT);
354 
355 	__skb_queue_head_init(&tmpq);
356 	__skb_queue_tail(&tmpq, _skb);
357 	if (method->rcast)
358 		rc = tipc_bcast_xmit(net, &tmpq, &cong_link_cnt);
359 	else
360 		rc = tipc_rcast_xmit(net, &tmpq, dests, &cong_link_cnt);
361 
362 	/* This queue should normally be empty by now */
363 	__skb_queue_purge(&tmpq);
364 
365 	return rc;
366 }
367 
368 /* tipc_mcast_xmit - deliver message to indicated destination nodes
369  *                   and to identified node local sockets
370  * @net: the applicable net namespace
371  * @pkts: chain of buffers containing message
372  * @method: send method to be used
373  * @dests: destination nodes for message.
374  * @cong_link_cnt: returns number of encountered congested destination links
375  * Consumes buffer chain.
376  * Returns 0 if success, otherwise errno
377  */
378 int tipc_mcast_xmit(struct net *net, struct sk_buff_head *pkts,
379 		    struct tipc_mc_method *method, struct tipc_nlist *dests,
380 		    u16 *cong_link_cnt)
381 {
382 	struct sk_buff_head inputq, localq;
383 	bool rcast = method->rcast;
384 	struct tipc_msg *hdr;
385 	struct sk_buff *skb;
386 	int rc = 0;
387 
388 	skb_queue_head_init(&inputq);
389 	__skb_queue_head_init(&localq);
390 
391 	/* Clone packets before they are consumed by next call */
392 	if (dests->local && !tipc_msg_reassemble(pkts, &localq)) {
393 		rc = -ENOMEM;
394 		goto exit;
395 	}
396 	/* Send according to determined transmit method */
397 	if (dests->remote) {
398 		tipc_bcast_select_xmit_method(net, dests->remote, method);
399 
400 		skb = skb_peek(pkts);
401 		hdr = buf_msg(skb);
402 		if (msg_user(hdr) == MSG_FRAGMENTER)
403 			hdr = msg_inner_hdr(hdr);
404 		msg_set_is_rcast(hdr, method->rcast);
405 
406 		/* Switch method ? */
407 		if (rcast != method->rcast) {
408 			rc = tipc_mcast_send_sync(net, skb, method, dests);
409 			if (unlikely(rc)) {
410 				pr_err("Unable to send SYN: method %d, rc %d\n",
411 				       rcast, rc);
412 				goto exit;
413 			}
414 		}
415 
416 		if (method->rcast)
417 			rc = tipc_rcast_xmit(net, pkts, dests, cong_link_cnt);
418 		else
419 			rc = tipc_bcast_xmit(net, pkts, cong_link_cnt);
420 	}
421 
422 	if (dests->local) {
423 		tipc_loopback_trace(net, &localq);
424 		tipc_sk_mcast_rcv(net, &localq, &inputq);
425 	}
426 exit:
427 	/* This queue should normally be empty by now */
428 	__skb_queue_purge(pkts);
429 	return rc;
430 }
431 
432 /* tipc_bcast_rcv - receive a broadcast packet, and deliver to rcv link
433  *
434  * RCU is locked, no other locks set
435  */
436 int tipc_bcast_rcv(struct net *net, struct tipc_link *l, struct sk_buff *skb)
437 {
438 	struct tipc_msg *hdr = buf_msg(skb);
439 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
440 	struct sk_buff_head xmitq;
441 	int rc;
442 
443 	__skb_queue_head_init(&xmitq);
444 
445 	if (msg_mc_netid(hdr) != tipc_netid(net) || !tipc_link_is_up(l)) {
446 		kfree_skb(skb);
447 		return 0;
448 	}
449 
450 	tipc_bcast_lock(net);
451 	if (msg_user(hdr) == BCAST_PROTOCOL)
452 		rc = tipc_link_bc_nack_rcv(l, skb, &xmitq);
453 	else
454 		rc = tipc_link_rcv(l, skb, NULL);
455 	tipc_bcast_unlock(net);
456 
457 	tipc_bcbase_xmit(net, &xmitq);
458 
459 	/* Any socket wakeup messages ? */
460 	if (!skb_queue_empty(inputq))
461 		tipc_sk_rcv(net, inputq);
462 
463 	return rc;
464 }
465 
466 /* tipc_bcast_ack_rcv - receive and handle a broadcast acknowledge
467  *
468  * RCU is locked, no other locks set
469  */
470 void tipc_bcast_ack_rcv(struct net *net, struct tipc_link *l,
471 			struct tipc_msg *hdr)
472 {
473 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
474 	u16 acked = msg_bcast_ack(hdr);
475 	struct sk_buff_head xmitq;
476 
477 	/* Ignore bc acks sent by peer before bcast synch point was received */
478 	if (msg_bc_ack_invalid(hdr))
479 		return;
480 
481 	__skb_queue_head_init(&xmitq);
482 
483 	tipc_bcast_lock(net);
484 	tipc_link_bc_ack_rcv(l, acked, 0, NULL, &xmitq, NULL);
485 	tipc_bcast_unlock(net);
486 
487 	tipc_bcbase_xmit(net, &xmitq);
488 
489 	/* Any socket wakeup messages ? */
490 	if (!skb_queue_empty(inputq))
491 		tipc_sk_rcv(net, inputq);
492 }
493 
494 /* tipc_bcast_synch_rcv -  check and update rcv link with peer's send state
495  *
496  * RCU is locked, no other locks set
497  */
498 int tipc_bcast_sync_rcv(struct net *net, struct tipc_link *l,
499 			struct tipc_msg *hdr,
500 			struct sk_buff_head *retrq)
501 {
502 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
503 	struct tipc_gap_ack_blks *ga;
504 	struct sk_buff_head xmitq;
505 	int rc = 0;
506 
507 	__skb_queue_head_init(&xmitq);
508 
509 	tipc_bcast_lock(net);
510 	if (msg_type(hdr) != STATE_MSG) {
511 		tipc_link_bc_init_rcv(l, hdr);
512 	} else if (!msg_bc_ack_invalid(hdr)) {
513 		tipc_get_gap_ack_blks(&ga, l, hdr, false);
514 		if (!sysctl_tipc_bc_retruni)
515 			retrq = &xmitq;
516 		rc = tipc_link_bc_ack_rcv(l, msg_bcast_ack(hdr),
517 					  msg_bc_gap(hdr), ga, &xmitq,
518 					  retrq);
519 		rc |= tipc_link_bc_sync_rcv(l, hdr, &xmitq);
520 	}
521 	tipc_bcast_unlock(net);
522 
523 	tipc_bcbase_xmit(net, &xmitq);
524 
525 	/* Any socket wakeup messages ? */
526 	if (!skb_queue_empty(inputq))
527 		tipc_sk_rcv(net, inputq);
528 	return rc;
529 }
530 
531 /* tipc_bcast_add_peer - add a peer node to broadcast link and bearer
532  *
533  * RCU is locked, node lock is set
534  */
535 void tipc_bcast_add_peer(struct net *net, struct tipc_link *uc_l,
536 			 struct sk_buff_head *xmitq)
537 {
538 	struct tipc_link *snd_l = tipc_bc_sndlink(net);
539 
540 	tipc_bcast_lock(net);
541 	tipc_link_add_bc_peer(snd_l, uc_l, xmitq);
542 	tipc_bcbase_select_primary(net);
543 	tipc_bcbase_calc_bc_threshold(net);
544 	tipc_bcast_unlock(net);
545 }
546 
547 /* tipc_bcast_remove_peer - remove a peer node from broadcast link and bearer
548  *
549  * RCU is locked, node lock is set
550  */
551 void tipc_bcast_remove_peer(struct net *net, struct tipc_link *rcv_l)
552 {
553 	struct tipc_link *snd_l = tipc_bc_sndlink(net);
554 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
555 	struct sk_buff_head xmitq;
556 
557 	__skb_queue_head_init(&xmitq);
558 
559 	tipc_bcast_lock(net);
560 	tipc_link_remove_bc_peer(snd_l, rcv_l, &xmitq);
561 	tipc_bcbase_select_primary(net);
562 	tipc_bcbase_calc_bc_threshold(net);
563 	tipc_bcast_unlock(net);
564 
565 	tipc_bcbase_xmit(net, &xmitq);
566 
567 	/* Any socket wakeup messages ? */
568 	if (!skb_queue_empty(inputq))
569 		tipc_sk_rcv(net, inputq);
570 }
571 
572 int tipc_bclink_reset_stats(struct net *net, struct tipc_link *l)
573 {
574 	if (!l)
575 		return -ENOPROTOOPT;
576 
577 	tipc_bcast_lock(net);
578 	tipc_link_reset_stats(l);
579 	tipc_bcast_unlock(net);
580 	return 0;
581 }
582 
583 static int tipc_bc_link_set_queue_limits(struct net *net, u32 max_win)
584 {
585 	struct tipc_link *l = tipc_bc_sndlink(net);
586 
587 	if (!l)
588 		return -ENOPROTOOPT;
589 	if (max_win < BCLINK_WIN_MIN)
590 		max_win = BCLINK_WIN_MIN;
591 	if (max_win > TIPC_MAX_LINK_WIN)
592 		return -EINVAL;
593 	tipc_bcast_lock(net);
594 	tipc_link_set_queue_limits(l, tipc_link_min_win(l), max_win);
595 	tipc_bcast_unlock(net);
596 	return 0;
597 }
598 
599 static int tipc_bc_link_set_broadcast_mode(struct net *net, u32 bc_mode)
600 {
601 	struct tipc_bc_base *bb = tipc_bc_base(net);
602 
603 	switch (bc_mode) {
604 	case BCLINK_MODE_BCAST:
605 		if (!bb->bcast_support)
606 			return -ENOPROTOOPT;
607 
608 		bb->force_bcast = true;
609 		bb->force_rcast = false;
610 		break;
611 	case BCLINK_MODE_RCAST:
612 		if (!bb->rcast_support)
613 			return -ENOPROTOOPT;
614 
615 		bb->force_bcast = false;
616 		bb->force_rcast = true;
617 		break;
618 	case BCLINK_MODE_SEL:
619 		if (!bb->bcast_support || !bb->rcast_support)
620 			return -ENOPROTOOPT;
621 
622 		bb->force_bcast = false;
623 		bb->force_rcast = false;
624 		break;
625 	default:
626 		return -EINVAL;
627 	}
628 
629 	return 0;
630 }
631 
632 static int tipc_bc_link_set_broadcast_ratio(struct net *net, u32 bc_ratio)
633 {
634 	struct tipc_bc_base *bb = tipc_bc_base(net);
635 
636 	if (!bb->bcast_support || !bb->rcast_support)
637 		return -ENOPROTOOPT;
638 
639 	if (bc_ratio > 100 || bc_ratio <= 0)
640 		return -EINVAL;
641 
642 	bb->rc_ratio = bc_ratio;
643 	tipc_bcast_lock(net);
644 	tipc_bcbase_calc_bc_threshold(net);
645 	tipc_bcast_unlock(net);
646 
647 	return 0;
648 }
649 
650 int tipc_nl_bc_link_set(struct net *net, struct nlattr *attrs[])
651 {
652 	int err;
653 	u32 win;
654 	u32 bc_mode;
655 	u32 bc_ratio;
656 	struct nlattr *props[TIPC_NLA_PROP_MAX + 1];
657 
658 	if (!attrs[TIPC_NLA_LINK_PROP])
659 		return -EINVAL;
660 
661 	err = tipc_nl_parse_link_prop(attrs[TIPC_NLA_LINK_PROP], props);
662 	if (err)
663 		return err;
664 
665 	if (!props[TIPC_NLA_PROP_WIN] &&
666 	    !props[TIPC_NLA_PROP_BROADCAST] &&
667 	    !props[TIPC_NLA_PROP_BROADCAST_RATIO]) {
668 		return -EOPNOTSUPP;
669 	}
670 
671 	if (props[TIPC_NLA_PROP_BROADCAST]) {
672 		bc_mode = nla_get_u32(props[TIPC_NLA_PROP_BROADCAST]);
673 		err = tipc_bc_link_set_broadcast_mode(net, bc_mode);
674 	}
675 
676 	if (!err && props[TIPC_NLA_PROP_BROADCAST_RATIO]) {
677 		bc_ratio = nla_get_u32(props[TIPC_NLA_PROP_BROADCAST_RATIO]);
678 		err = tipc_bc_link_set_broadcast_ratio(net, bc_ratio);
679 	}
680 
681 	if (!err && props[TIPC_NLA_PROP_WIN]) {
682 		win = nla_get_u32(props[TIPC_NLA_PROP_WIN]);
683 		err = tipc_bc_link_set_queue_limits(net, win);
684 	}
685 
686 	return err;
687 }
688 
689 int tipc_bcast_init(struct net *net)
690 {
691 	struct tipc_net *tn = tipc_net(net);
692 	struct tipc_bc_base *bb = NULL;
693 	struct tipc_link *l = NULL;
694 
695 	bb = kzalloc(sizeof(*bb), GFP_KERNEL);
696 	if (!bb)
697 		goto enomem;
698 	tn->bcbase = bb;
699 	spin_lock_init(&tipc_net(net)->bclock);
700 
701 	if (!tipc_link_bc_create(net, 0, 0, NULL,
702 				 one_page_mtu,
703 				 BCLINK_WIN_DEFAULT,
704 				 BCLINK_WIN_DEFAULT,
705 				 0,
706 				 &bb->inputq,
707 				 NULL,
708 				 NULL,
709 				 &l))
710 		goto enomem;
711 	bb->link = l;
712 	tn->bcl = l;
713 	bb->rc_ratio = 10;
714 	bb->rcast_support = true;
715 	return 0;
716 enomem:
717 	kfree(bb);
718 	kfree(l);
719 	return -ENOMEM;
720 }
721 
722 void tipc_bcast_stop(struct net *net)
723 {
724 	struct tipc_net *tn = net_generic(net, tipc_net_id);
725 
726 	synchronize_net();
727 	kfree(tn->bcbase);
728 	kfree(tn->bcl);
729 }
730 
731 void tipc_nlist_init(struct tipc_nlist *nl, u32 self)
732 {
733 	memset(nl, 0, sizeof(*nl));
734 	INIT_LIST_HEAD(&nl->list);
735 	nl->self = self;
736 }
737 
738 void tipc_nlist_add(struct tipc_nlist *nl, u32 node)
739 {
740 	if (node == nl->self)
741 		nl->local = true;
742 	else if (tipc_dest_push(&nl->list, node, 0))
743 		nl->remote++;
744 }
745 
746 void tipc_nlist_del(struct tipc_nlist *nl, u32 node)
747 {
748 	if (node == nl->self)
749 		nl->local = false;
750 	else if (tipc_dest_del(&nl->list, node, 0))
751 		nl->remote--;
752 }
753 
754 void tipc_nlist_purge(struct tipc_nlist *nl)
755 {
756 	tipc_dest_list_purge(&nl->list);
757 	nl->remote = 0;
758 	nl->local = false;
759 }
760 
761 u32 tipc_bcast_get_mode(struct net *net)
762 {
763 	struct tipc_bc_base *bb = tipc_bc_base(net);
764 
765 	if (bb->force_bcast)
766 		return BCLINK_MODE_BCAST;
767 
768 	if (bb->force_rcast)
769 		return BCLINK_MODE_RCAST;
770 
771 	if (bb->bcast_support && bb->rcast_support)
772 		return BCLINK_MODE_SEL;
773 
774 	return 0;
775 }
776 
777 u32 tipc_bcast_get_broadcast_ratio(struct net *net)
778 {
779 	struct tipc_bc_base *bb = tipc_bc_base(net);
780 
781 	return bb->rc_ratio;
782 }
783 
784 void tipc_mcast_filter_msg(struct net *net, struct sk_buff_head *defq,
785 			   struct sk_buff_head *inputq)
786 {
787 	struct sk_buff *skb, *_skb, *tmp;
788 	struct tipc_msg *hdr, *_hdr;
789 	bool match = false;
790 	u32 node, port;
791 
792 	skb = skb_peek(inputq);
793 	if (!skb)
794 		return;
795 
796 	hdr = buf_msg(skb);
797 
798 	if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
799 		return;
800 
801 	node = msg_orignode(hdr);
802 	if (node == tipc_own_addr(net))
803 		return;
804 
805 	port = msg_origport(hdr);
806 
807 	/* Has the twin SYN message already arrived ? */
808 	skb_queue_walk(defq, _skb) {
809 		_hdr = buf_msg(_skb);
810 		if (msg_orignode(_hdr) != node)
811 			continue;
812 		if (msg_origport(_hdr) != port)
813 			continue;
814 		match = true;
815 		break;
816 	}
817 
818 	if (!match) {
819 		if (!msg_is_syn(hdr))
820 			return;
821 		__skb_dequeue(inputq);
822 		__skb_queue_tail(defq, skb);
823 		return;
824 	}
825 
826 	/* Deliver non-SYN message from other link, otherwise queue it */
827 	if (!msg_is_syn(hdr)) {
828 		if (msg_is_rcast(hdr) != msg_is_rcast(_hdr))
829 			return;
830 		__skb_dequeue(inputq);
831 		__skb_queue_tail(defq, skb);
832 		return;
833 	}
834 
835 	/* Queue non-SYN/SYN message from same link */
836 	if (msg_is_rcast(hdr) == msg_is_rcast(_hdr)) {
837 		__skb_dequeue(inputq);
838 		__skb_queue_tail(defq, skb);
839 		return;
840 	}
841 
842 	/* Matching SYN messages => return the one with data, if any */
843 	__skb_unlink(_skb, defq);
844 	if (msg_data_sz(hdr)) {
845 		kfree_skb(_skb);
846 	} else {
847 		__skb_dequeue(inputq);
848 		kfree_skb(skb);
849 		__skb_queue_tail(inputq, _skb);
850 	}
851 
852 	/* Deliver subsequent non-SYN messages from same peer */
853 	skb_queue_walk_safe(defq, _skb, tmp) {
854 		_hdr = buf_msg(_skb);
855 		if (msg_orignode(_hdr) != node)
856 			continue;
857 		if (msg_origport(_hdr) != port)
858 			continue;
859 		if (msg_is_syn(_hdr))
860 			break;
861 		__skb_unlink(_skb, defq);
862 		__skb_queue_tail(inputq, _skb);
863 	}
864 }
865