xref: /linux/net/tipc/bcast.c (revision 8ea636848aca35b9f97c5b5dee30225cf2dd0fe6)
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 i, mtu, prim;
112 
113 	bb->primary_bearer = INVALID_BEARER_ID;
114 	bb->bcast_support = true;
115 
116 	if (!all_dests)
117 		return;
118 
119 	for (i = 0; i < MAX_BEARERS; i++) {
120 		if (!bb->dests[i])
121 			continue;
122 
123 		mtu = tipc_bearer_mtu(net, i);
124 		if (mtu < tipc_link_mtu(bb->link))
125 			tipc_link_set_mtu(bb->link, mtu);
126 		bb->bcast_support &= tipc_bearer_bcast_support(net, i);
127 		if (bb->dests[i] < all_dests)
128 			continue;
129 
130 		bb->primary_bearer = i;
131 
132 		/* Reduce risk that all nodes select same primary */
133 		if ((i ^ tipc_own_addr(net)) & 1)
134 			break;
135 	}
136 	prim = bb->primary_bearer;
137 	if (prim != INVALID_BEARER_ID)
138 		bb->bcast_support = tipc_bearer_bcast_support(net, prim);
139 }
140 
141 void tipc_bcast_inc_bearer_dst_cnt(struct net *net, int bearer_id)
142 {
143 	struct tipc_bc_base *bb = tipc_bc_base(net);
144 
145 	tipc_bcast_lock(net);
146 	bb->dests[bearer_id]++;
147 	tipc_bcbase_select_primary(net);
148 	tipc_bcast_unlock(net);
149 }
150 
151 void tipc_bcast_dec_bearer_dst_cnt(struct net *net, int bearer_id)
152 {
153 	struct tipc_bc_base *bb = tipc_bc_base(net);
154 
155 	tipc_bcast_lock(net);
156 	bb->dests[bearer_id]--;
157 	tipc_bcbase_select_primary(net);
158 	tipc_bcast_unlock(net);
159 }
160 
161 /* tipc_bcbase_xmit - broadcast a packet queue across one or more bearers
162  *
163  * Note that number of reachable destinations, as indicated in the dests[]
164  * array, may transitionally differ from the number of destinations indicated
165  * in each sent buffer. We can sustain this. Excess destination nodes will
166  * drop and never acknowledge the unexpected packets, and missing destinations
167  * will either require retransmission (if they are just about to be added to
168  * the bearer), or be removed from the buffer's 'ackers' counter (if they
169  * just went down)
170  */
171 static void tipc_bcbase_xmit(struct net *net, struct sk_buff_head *xmitq)
172 {
173 	int bearer_id;
174 	struct tipc_bc_base *bb = tipc_bc_base(net);
175 	struct sk_buff *skb, *_skb;
176 	struct sk_buff_head _xmitq;
177 
178 	if (skb_queue_empty(xmitq))
179 		return;
180 
181 	/* The typical case: at least one bearer has links to all nodes */
182 	bearer_id = bb->primary_bearer;
183 	if (bearer_id >= 0) {
184 		tipc_bearer_bc_xmit(net, bearer_id, xmitq);
185 		return;
186 	}
187 
188 	/* We have to transmit across all bearers */
189 	__skb_queue_head_init(&_xmitq);
190 	for (bearer_id = 0; bearer_id < MAX_BEARERS; bearer_id++) {
191 		if (!bb->dests[bearer_id])
192 			continue;
193 
194 		skb_queue_walk(xmitq, skb) {
195 			_skb = pskb_copy_for_clone(skb, GFP_ATOMIC);
196 			if (!_skb)
197 				break;
198 			__skb_queue_tail(&_xmitq, _skb);
199 		}
200 		tipc_bearer_bc_xmit(net, bearer_id, &_xmitq);
201 	}
202 	__skb_queue_purge(xmitq);
203 	__skb_queue_purge(&_xmitq);
204 }
205 
206 static void tipc_bcast_select_xmit_method(struct net *net, int dests,
207 					  struct tipc_mc_method *method)
208 {
209 	struct tipc_bc_base *bb = tipc_bc_base(net);
210 	unsigned long exp = method->expires;
211 
212 	/* Broadcast supported by used bearer/bearers? */
213 	if (!bb->bcast_support) {
214 		method->rcast = true;
215 		return;
216 	}
217 	/* Any destinations which don't support replicast ? */
218 	if (!bb->rcast_support) {
219 		method->rcast = false;
220 		return;
221 	}
222 	/* Can current method be changed ? */
223 	method->expires = jiffies + TIPC_METHOD_EXPIRE;
224 	if (method->mandatory)
225 		return;
226 
227 	if (!(tipc_net(net)->capabilities & TIPC_MCAST_RBCTL) &&
228 	    time_before(jiffies, exp))
229 		return;
230 
231 	/* Configuration as force 'broadcast' method */
232 	if (bb->force_bcast) {
233 		method->rcast = false;
234 		return;
235 	}
236 	/* Configuration as force 'replicast' method */
237 	if (bb->force_rcast) {
238 		method->rcast = true;
239 		return;
240 	}
241 	/* Configuration as 'autoselect' or default method */
242 	/* Determine method to use now */
243 	method->rcast = dests <= bb->bc_threshold;
244 }
245 
246 /* tipc_bcast_xmit - broadcast the buffer chain to all external nodes
247  * @net: the applicable net namespace
248  * @pkts: chain of buffers containing message
249  * @cong_link_cnt: set to 1 if broadcast link is congested, otherwise 0
250  * Consumes the buffer chain.
251  * Returns 0 if success, otherwise errno: -EHOSTUNREACH,-EMSGSIZE
252  */
253 int tipc_bcast_xmit(struct net *net, struct sk_buff_head *pkts,
254 		    u16 *cong_link_cnt)
255 {
256 	struct tipc_link *l = tipc_bc_sndlink(net);
257 	struct sk_buff_head xmitq;
258 	int rc = 0;
259 
260 	__skb_queue_head_init(&xmitq);
261 	tipc_bcast_lock(net);
262 	if (tipc_link_bc_peers(l))
263 		rc = tipc_link_xmit(l, pkts, &xmitq);
264 	tipc_bcast_unlock(net);
265 	tipc_bcbase_xmit(net, &xmitq);
266 	__skb_queue_purge(pkts);
267 	if (rc == -ELINKCONG) {
268 		*cong_link_cnt = 1;
269 		rc = 0;
270 	}
271 	return rc;
272 }
273 
274 /* tipc_rcast_xmit - replicate and send a message to given destination nodes
275  * @net: the applicable net namespace
276  * @pkts: chain of buffers containing message
277  * @dests: list of destination nodes
278  * @cong_link_cnt: returns number of congested links
279  * @cong_links: returns identities of congested links
280  * Returns 0 if success, otherwise errno
281  */
282 static int tipc_rcast_xmit(struct net *net, struct sk_buff_head *pkts,
283 			   struct tipc_nlist *dests, u16 *cong_link_cnt)
284 {
285 	struct tipc_dest *dst, *tmp;
286 	struct sk_buff_head _pkts;
287 	u32 dnode, selector;
288 
289 	selector = msg_link_selector(buf_msg(skb_peek(pkts)));
290 	__skb_queue_head_init(&_pkts);
291 
292 	list_for_each_entry_safe(dst, tmp, &dests->list, list) {
293 		dnode = dst->node;
294 		if (!tipc_msg_pskb_copy(dnode, pkts, &_pkts))
295 			return -ENOMEM;
296 
297 		/* Any other return value than -ELINKCONG is ignored */
298 		if (tipc_node_xmit(net, &_pkts, dnode, selector) == -ELINKCONG)
299 			(*cong_link_cnt)++;
300 	}
301 	return 0;
302 }
303 
304 /* tipc_mcast_send_sync - deliver a dummy message with SYN bit
305  * @net: the applicable net namespace
306  * @skb: socket buffer to copy
307  * @method: send method to be used
308  * @dests: destination nodes for message.
309  * Returns 0 if success, otherwise errno
310  */
311 static int tipc_mcast_send_sync(struct net *net, struct sk_buff *skb,
312 				struct tipc_mc_method *method,
313 				struct tipc_nlist *dests)
314 {
315 	struct tipc_msg *hdr, *_hdr;
316 	struct sk_buff_head tmpq;
317 	struct sk_buff *_skb;
318 	u16 cong_link_cnt;
319 	int rc = 0;
320 
321 	/* Is a cluster supporting with new capabilities ? */
322 	if (!(tipc_net(net)->capabilities & TIPC_MCAST_RBCTL))
323 		return 0;
324 
325 	hdr = buf_msg(skb);
326 	if (msg_user(hdr) == MSG_FRAGMENTER)
327 		hdr = msg_inner_hdr(hdr);
328 	if (msg_type(hdr) != TIPC_MCAST_MSG)
329 		return 0;
330 
331 	/* Allocate dummy message */
332 	_skb = tipc_buf_acquire(MCAST_H_SIZE, GFP_KERNEL);
333 	if (!_skb)
334 		return -ENOMEM;
335 
336 	/* Preparing for 'synching' header */
337 	msg_set_syn(hdr, 1);
338 
339 	/* Copy skb's header into a dummy header */
340 	skb_copy_to_linear_data(_skb, hdr, MCAST_H_SIZE);
341 	skb_orphan(_skb);
342 
343 	/* Reverse method for dummy message */
344 	_hdr = buf_msg(_skb);
345 	msg_set_size(_hdr, MCAST_H_SIZE);
346 	msg_set_is_rcast(_hdr, !msg_is_rcast(hdr));
347 	msg_set_errcode(_hdr, TIPC_ERR_NO_PORT);
348 
349 	__skb_queue_head_init(&tmpq);
350 	__skb_queue_tail(&tmpq, _skb);
351 	if (method->rcast)
352 		rc = tipc_bcast_xmit(net, &tmpq, &cong_link_cnt);
353 	else
354 		rc = tipc_rcast_xmit(net, &tmpq, dests, &cong_link_cnt);
355 
356 	/* This queue should normally be empty by now */
357 	__skb_queue_purge(&tmpq);
358 
359 	return rc;
360 }
361 
362 /* tipc_mcast_xmit - deliver message to indicated destination nodes
363  *                   and to identified node local sockets
364  * @net: the applicable net namespace
365  * @pkts: chain of buffers containing message
366  * @method: send method to be used
367  * @dests: destination nodes for message.
368  * @cong_link_cnt: returns number of encountered congested destination links
369  * Consumes buffer chain.
370  * Returns 0 if success, otherwise errno
371  */
372 int tipc_mcast_xmit(struct net *net, struct sk_buff_head *pkts,
373 		    struct tipc_mc_method *method, struct tipc_nlist *dests,
374 		    u16 *cong_link_cnt)
375 {
376 	struct sk_buff_head inputq, localq;
377 	bool rcast = method->rcast;
378 	struct tipc_msg *hdr;
379 	struct sk_buff *skb;
380 	int rc = 0;
381 
382 	skb_queue_head_init(&inputq);
383 	__skb_queue_head_init(&localq);
384 
385 	/* Clone packets before they are consumed by next call */
386 	if (dests->local && !tipc_msg_reassemble(pkts, &localq)) {
387 		rc = -ENOMEM;
388 		goto exit;
389 	}
390 	/* Send according to determined transmit method */
391 	if (dests->remote) {
392 		tipc_bcast_select_xmit_method(net, dests->remote, method);
393 
394 		skb = skb_peek(pkts);
395 		hdr = buf_msg(skb);
396 		if (msg_user(hdr) == MSG_FRAGMENTER)
397 			hdr = msg_inner_hdr(hdr);
398 		msg_set_is_rcast(hdr, method->rcast);
399 
400 		/* Switch method ? */
401 		if (rcast != method->rcast) {
402 			rc = tipc_mcast_send_sync(net, skb, method, dests);
403 			if (unlikely(rc)) {
404 				pr_err("Unable to send SYN: method %d, rc %d\n",
405 				       rcast, rc);
406 				goto exit;
407 			}
408 		}
409 
410 		if (method->rcast)
411 			rc = tipc_rcast_xmit(net, pkts, dests, cong_link_cnt);
412 		else
413 			rc = tipc_bcast_xmit(net, pkts, cong_link_cnt);
414 	}
415 
416 	if (dests->local) {
417 		tipc_loopback_trace(net, &localq);
418 		tipc_sk_mcast_rcv(net, &localq, &inputq);
419 	}
420 exit:
421 	/* This queue should normally be empty by now */
422 	__skb_queue_purge(pkts);
423 	return rc;
424 }
425 
426 /* tipc_bcast_rcv - receive a broadcast packet, and deliver to rcv link
427  *
428  * RCU is locked, no other locks set
429  */
430 int tipc_bcast_rcv(struct net *net, struct tipc_link *l, struct sk_buff *skb)
431 {
432 	struct tipc_msg *hdr = buf_msg(skb);
433 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
434 	struct sk_buff_head xmitq;
435 	int rc;
436 
437 	__skb_queue_head_init(&xmitq);
438 
439 	if (msg_mc_netid(hdr) != tipc_netid(net) || !tipc_link_is_up(l)) {
440 		kfree_skb(skb);
441 		return 0;
442 	}
443 
444 	tipc_bcast_lock(net);
445 	if (msg_user(hdr) == BCAST_PROTOCOL)
446 		rc = tipc_link_bc_nack_rcv(l, skb, &xmitq);
447 	else
448 		rc = tipc_link_rcv(l, skb, NULL);
449 	tipc_bcast_unlock(net);
450 
451 	tipc_bcbase_xmit(net, &xmitq);
452 
453 	/* Any socket wakeup messages ? */
454 	if (!skb_queue_empty(inputq))
455 		tipc_sk_rcv(net, inputq);
456 
457 	return rc;
458 }
459 
460 /* tipc_bcast_ack_rcv - receive and handle a broadcast acknowledge
461  *
462  * RCU is locked, no other locks set
463  */
464 void tipc_bcast_ack_rcv(struct net *net, struct tipc_link *l,
465 			struct tipc_msg *hdr)
466 {
467 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
468 	u16 acked = msg_bcast_ack(hdr);
469 	struct sk_buff_head xmitq;
470 
471 	/* Ignore bc acks sent by peer before bcast synch point was received */
472 	if (msg_bc_ack_invalid(hdr))
473 		return;
474 
475 	__skb_queue_head_init(&xmitq);
476 
477 	tipc_bcast_lock(net);
478 	tipc_link_bc_ack_rcv(l, acked, 0, NULL, &xmitq, NULL);
479 	tipc_bcast_unlock(net);
480 
481 	tipc_bcbase_xmit(net, &xmitq);
482 
483 	/* Any socket wakeup messages ? */
484 	if (!skb_queue_empty(inputq))
485 		tipc_sk_rcv(net, inputq);
486 }
487 
488 /* tipc_bcast_synch_rcv -  check and update rcv link with peer's send state
489  *
490  * RCU is locked, no other locks set
491  */
492 int tipc_bcast_sync_rcv(struct net *net, struct tipc_link *l,
493 			struct tipc_msg *hdr,
494 			struct sk_buff_head *retrq)
495 {
496 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
497 	struct tipc_gap_ack_blks *ga;
498 	struct sk_buff_head xmitq;
499 	int rc = 0;
500 
501 	__skb_queue_head_init(&xmitq);
502 
503 	tipc_bcast_lock(net);
504 	if (msg_type(hdr) != STATE_MSG) {
505 		tipc_link_bc_init_rcv(l, hdr);
506 	} else if (!msg_bc_ack_invalid(hdr)) {
507 		tipc_get_gap_ack_blks(&ga, l, hdr, false);
508 		if (!sysctl_tipc_bc_retruni)
509 			retrq = &xmitq;
510 		rc = tipc_link_bc_ack_rcv(l, msg_bcast_ack(hdr),
511 					  msg_bc_gap(hdr), ga, &xmitq,
512 					  retrq);
513 		rc |= tipc_link_bc_sync_rcv(l, hdr, &xmitq);
514 	}
515 	tipc_bcast_unlock(net);
516 
517 	tipc_bcbase_xmit(net, &xmitq);
518 
519 	/* Any socket wakeup messages ? */
520 	if (!skb_queue_empty(inputq))
521 		tipc_sk_rcv(net, inputq);
522 	return rc;
523 }
524 
525 /* tipc_bcast_add_peer - add a peer node to broadcast link and bearer
526  *
527  * RCU is locked, node lock is set
528  */
529 void tipc_bcast_add_peer(struct net *net, struct tipc_link *uc_l,
530 			 struct sk_buff_head *xmitq)
531 {
532 	struct tipc_link *snd_l = tipc_bc_sndlink(net);
533 
534 	tipc_bcast_lock(net);
535 	tipc_link_add_bc_peer(snd_l, uc_l, xmitq);
536 	tipc_bcbase_select_primary(net);
537 	tipc_bcbase_calc_bc_threshold(net);
538 	tipc_bcast_unlock(net);
539 }
540 
541 /* tipc_bcast_remove_peer - remove a peer node from broadcast link and bearer
542  *
543  * RCU is locked, node lock is set
544  */
545 void tipc_bcast_remove_peer(struct net *net, struct tipc_link *rcv_l)
546 {
547 	struct tipc_link *snd_l = tipc_bc_sndlink(net);
548 	struct sk_buff_head *inputq = &tipc_bc_base(net)->inputq;
549 	struct sk_buff_head xmitq;
550 
551 	__skb_queue_head_init(&xmitq);
552 
553 	tipc_bcast_lock(net);
554 	tipc_link_remove_bc_peer(snd_l, rcv_l, &xmitq);
555 	tipc_bcbase_select_primary(net);
556 	tipc_bcbase_calc_bc_threshold(net);
557 	tipc_bcast_unlock(net);
558 
559 	tipc_bcbase_xmit(net, &xmitq);
560 
561 	/* Any socket wakeup messages ? */
562 	if (!skb_queue_empty(inputq))
563 		tipc_sk_rcv(net, inputq);
564 }
565 
566 int tipc_bclink_reset_stats(struct net *net, struct tipc_link *l)
567 {
568 	if (!l)
569 		return -ENOPROTOOPT;
570 
571 	tipc_bcast_lock(net);
572 	tipc_link_reset_stats(l);
573 	tipc_bcast_unlock(net);
574 	return 0;
575 }
576 
577 static int tipc_bc_link_set_queue_limits(struct net *net, u32 max_win)
578 {
579 	struct tipc_link *l = tipc_bc_sndlink(net);
580 
581 	if (!l)
582 		return -ENOPROTOOPT;
583 	if (max_win < BCLINK_WIN_MIN)
584 		max_win = BCLINK_WIN_MIN;
585 	if (max_win > TIPC_MAX_LINK_WIN)
586 		return -EINVAL;
587 	tipc_bcast_lock(net);
588 	tipc_link_set_queue_limits(l, BCLINK_WIN_MIN, max_win);
589 	tipc_bcast_unlock(net);
590 	return 0;
591 }
592 
593 static int tipc_bc_link_set_broadcast_mode(struct net *net, u32 bc_mode)
594 {
595 	struct tipc_bc_base *bb = tipc_bc_base(net);
596 
597 	switch (bc_mode) {
598 	case BCLINK_MODE_BCAST:
599 		if (!bb->bcast_support)
600 			return -ENOPROTOOPT;
601 
602 		bb->force_bcast = true;
603 		bb->force_rcast = false;
604 		break;
605 	case BCLINK_MODE_RCAST:
606 		if (!bb->rcast_support)
607 			return -ENOPROTOOPT;
608 
609 		bb->force_bcast = false;
610 		bb->force_rcast = true;
611 		break;
612 	case BCLINK_MODE_SEL:
613 		if (!bb->bcast_support || !bb->rcast_support)
614 			return -ENOPROTOOPT;
615 
616 		bb->force_bcast = false;
617 		bb->force_rcast = false;
618 		break;
619 	default:
620 		return -EINVAL;
621 	}
622 
623 	return 0;
624 }
625 
626 static int tipc_bc_link_set_broadcast_ratio(struct net *net, u32 bc_ratio)
627 {
628 	struct tipc_bc_base *bb = tipc_bc_base(net);
629 
630 	if (!bb->bcast_support || !bb->rcast_support)
631 		return -ENOPROTOOPT;
632 
633 	if (bc_ratio > 100 || bc_ratio <= 0)
634 		return -EINVAL;
635 
636 	bb->rc_ratio = bc_ratio;
637 	tipc_bcast_lock(net);
638 	tipc_bcbase_calc_bc_threshold(net);
639 	tipc_bcast_unlock(net);
640 
641 	return 0;
642 }
643 
644 int tipc_nl_bc_link_set(struct net *net, struct nlattr *attrs[])
645 {
646 	int err;
647 	u32 win;
648 	u32 bc_mode;
649 	u32 bc_ratio;
650 	struct nlattr *props[TIPC_NLA_PROP_MAX + 1];
651 
652 	if (!attrs[TIPC_NLA_LINK_PROP])
653 		return -EINVAL;
654 
655 	err = tipc_nl_parse_link_prop(attrs[TIPC_NLA_LINK_PROP], props);
656 	if (err)
657 		return err;
658 
659 	if (!props[TIPC_NLA_PROP_WIN] &&
660 	    !props[TIPC_NLA_PROP_BROADCAST] &&
661 	    !props[TIPC_NLA_PROP_BROADCAST_RATIO]) {
662 		return -EOPNOTSUPP;
663 	}
664 
665 	if (props[TIPC_NLA_PROP_BROADCAST]) {
666 		bc_mode = nla_get_u32(props[TIPC_NLA_PROP_BROADCAST]);
667 		err = tipc_bc_link_set_broadcast_mode(net, bc_mode);
668 	}
669 
670 	if (!err && props[TIPC_NLA_PROP_BROADCAST_RATIO]) {
671 		bc_ratio = nla_get_u32(props[TIPC_NLA_PROP_BROADCAST_RATIO]);
672 		err = tipc_bc_link_set_broadcast_ratio(net, bc_ratio);
673 	}
674 
675 	if (!err && props[TIPC_NLA_PROP_WIN]) {
676 		win = nla_get_u32(props[TIPC_NLA_PROP_WIN]);
677 		err = tipc_bc_link_set_queue_limits(net, win);
678 	}
679 
680 	return err;
681 }
682 
683 int tipc_bcast_init(struct net *net)
684 {
685 	struct tipc_net *tn = tipc_net(net);
686 	struct tipc_bc_base *bb = NULL;
687 	struct tipc_link *l = NULL;
688 
689 	bb = kzalloc(sizeof(*bb), GFP_KERNEL);
690 	if (!bb)
691 		goto enomem;
692 	tn->bcbase = bb;
693 	spin_lock_init(&tipc_net(net)->bclock);
694 
695 	if (!tipc_link_bc_create(net, 0, 0, NULL,
696 				 FB_MTU,
697 				 BCLINK_WIN_DEFAULT,
698 				 BCLINK_WIN_DEFAULT,
699 				 0,
700 				 &bb->inputq,
701 				 NULL,
702 				 NULL,
703 				 &l))
704 		goto enomem;
705 	bb->link = l;
706 	tn->bcl = l;
707 	bb->rc_ratio = 10;
708 	bb->rcast_support = true;
709 	return 0;
710 enomem:
711 	kfree(bb);
712 	kfree(l);
713 	return -ENOMEM;
714 }
715 
716 void tipc_bcast_stop(struct net *net)
717 {
718 	struct tipc_net *tn = net_generic(net, tipc_net_id);
719 
720 	synchronize_net();
721 	kfree(tn->bcbase);
722 	kfree(tn->bcl);
723 }
724 
725 void tipc_nlist_init(struct tipc_nlist *nl, u32 self)
726 {
727 	memset(nl, 0, sizeof(*nl));
728 	INIT_LIST_HEAD(&nl->list);
729 	nl->self = self;
730 }
731 
732 void tipc_nlist_add(struct tipc_nlist *nl, u32 node)
733 {
734 	if (node == nl->self)
735 		nl->local = true;
736 	else if (tipc_dest_push(&nl->list, node, 0))
737 		nl->remote++;
738 }
739 
740 void tipc_nlist_del(struct tipc_nlist *nl, u32 node)
741 {
742 	if (node == nl->self)
743 		nl->local = false;
744 	else if (tipc_dest_del(&nl->list, node, 0))
745 		nl->remote--;
746 }
747 
748 void tipc_nlist_purge(struct tipc_nlist *nl)
749 {
750 	tipc_dest_list_purge(&nl->list);
751 	nl->remote = 0;
752 	nl->local = false;
753 }
754 
755 u32 tipc_bcast_get_mode(struct net *net)
756 {
757 	struct tipc_bc_base *bb = tipc_bc_base(net);
758 
759 	if (bb->force_bcast)
760 		return BCLINK_MODE_BCAST;
761 
762 	if (bb->force_rcast)
763 		return BCLINK_MODE_RCAST;
764 
765 	if (bb->bcast_support && bb->rcast_support)
766 		return BCLINK_MODE_SEL;
767 
768 	return 0;
769 }
770 
771 u32 tipc_bcast_get_broadcast_ratio(struct net *net)
772 {
773 	struct tipc_bc_base *bb = tipc_bc_base(net);
774 
775 	return bb->rc_ratio;
776 }
777 
778 void tipc_mcast_filter_msg(struct net *net, struct sk_buff_head *defq,
779 			   struct sk_buff_head *inputq)
780 {
781 	struct sk_buff *skb, *_skb, *tmp;
782 	struct tipc_msg *hdr, *_hdr;
783 	bool match = false;
784 	u32 node, port;
785 
786 	skb = skb_peek(inputq);
787 	if (!skb)
788 		return;
789 
790 	hdr = buf_msg(skb);
791 
792 	if (likely(!msg_is_syn(hdr) && skb_queue_empty(defq)))
793 		return;
794 
795 	node = msg_orignode(hdr);
796 	if (node == tipc_own_addr(net))
797 		return;
798 
799 	port = msg_origport(hdr);
800 
801 	/* Has the twin SYN message already arrived ? */
802 	skb_queue_walk(defq, _skb) {
803 		_hdr = buf_msg(_skb);
804 		if (msg_orignode(_hdr) != node)
805 			continue;
806 		if (msg_origport(_hdr) != port)
807 			continue;
808 		match = true;
809 		break;
810 	}
811 
812 	if (!match) {
813 		if (!msg_is_syn(hdr))
814 			return;
815 		__skb_dequeue(inputq);
816 		__skb_queue_tail(defq, skb);
817 		return;
818 	}
819 
820 	/* Deliver non-SYN message from other link, otherwise queue it */
821 	if (!msg_is_syn(hdr)) {
822 		if (msg_is_rcast(hdr) != msg_is_rcast(_hdr))
823 			return;
824 		__skb_dequeue(inputq);
825 		__skb_queue_tail(defq, skb);
826 		return;
827 	}
828 
829 	/* Queue non-SYN/SYN message from same link */
830 	if (msg_is_rcast(hdr) == msg_is_rcast(_hdr)) {
831 		__skb_dequeue(inputq);
832 		__skb_queue_tail(defq, skb);
833 		return;
834 	}
835 
836 	/* Matching SYN messages => return the one with data, if any */
837 	__skb_unlink(_skb, defq);
838 	if (msg_data_sz(hdr)) {
839 		kfree_skb(_skb);
840 	} else {
841 		__skb_dequeue(inputq);
842 		kfree_skb(skb);
843 		__skb_queue_tail(inputq, _skb);
844 	}
845 
846 	/* Deliver subsequent non-SYN messages from same peer */
847 	skb_queue_walk_safe(defq, _skb, tmp) {
848 		_hdr = buf_msg(_skb);
849 		if (msg_orignode(_hdr) != node)
850 			continue;
851 		if (msg_origport(_hdr) != port)
852 			continue;
853 		if (msg_is_syn(_hdr))
854 			break;
855 		__skb_unlink(_skb, defq);
856 		__skb_queue_tail(inputq, _skb);
857 	}
858 }
859