xref: /openbmc/linux/net/tipc/bcast.c (revision d2168146)
1 /*
2  * net/tipc/bcast.c: TIPC broadcast code
3  *
4  * Copyright (c) 2004-2006, 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 "core.h"
39 #include "link.h"
40 #include "port.h"
41 #include "bcast.h"
42 #include "name_distr.h"
43 
44 #define	MAX_PKT_DEFAULT_MCAST	1500	/* bcast link max packet size (fixed) */
45 #define	BCLINK_WIN_DEFAULT	20	/* bcast link window size (default) */
46 #define	BCBEARER		MAX_BEARERS
47 
48 /**
49  * struct tipc_bcbearer_pair - a pair of bearers used by broadcast link
50  * @primary: pointer to primary bearer
51  * @secondary: pointer to secondary bearer
52  *
53  * Bearers must have same priority and same set of reachable destinations
54  * to be paired.
55  */
56 
57 struct tipc_bcbearer_pair {
58 	struct tipc_bearer *primary;
59 	struct tipc_bearer *secondary;
60 };
61 
62 /**
63  * struct tipc_bcbearer - bearer used by broadcast link
64  * @bearer: (non-standard) broadcast bearer structure
65  * @media: (non-standard) broadcast media structure
66  * @bpairs: array of bearer pairs
67  * @bpairs_temp: temporary array of bearer pairs used by tipc_bcbearer_sort()
68  * @remains: temporary node map used by tipc_bcbearer_send()
69  * @remains_new: temporary node map used tipc_bcbearer_send()
70  *
71  * Note: The fields labelled "temporary" are incorporated into the bearer
72  * to avoid consuming potentially limited stack space through the use of
73  * large local variables within multicast routines.  Concurrent access is
74  * prevented through use of the spinlock "bclink_lock".
75  */
76 struct tipc_bcbearer {
77 	struct tipc_bearer bearer;
78 	struct tipc_media media;
79 	struct tipc_bcbearer_pair bpairs[MAX_BEARERS];
80 	struct tipc_bcbearer_pair bpairs_temp[TIPC_MAX_LINK_PRI + 1];
81 	struct tipc_node_map remains;
82 	struct tipc_node_map remains_new;
83 };
84 
85 /**
86  * struct tipc_bclink - link used for broadcast messages
87  * @lock: spinlock governing access to structure
88  * @link: (non-standard) broadcast link structure
89  * @node: (non-standard) node structure representing b'cast link's peer node
90  * @flags: represent bclink states
91  * @bcast_nodes: map of broadcast-capable nodes
92  * @retransmit_to: node that most recently requested a retransmit
93  *
94  * Handles sequence numbering, fragmentation, bundling, etc.
95  */
96 struct tipc_bclink {
97 	spinlock_t lock;
98 	struct tipc_link link;
99 	struct tipc_node node;
100 	unsigned int flags;
101 	struct tipc_node_map bcast_nodes;
102 	struct tipc_node *retransmit_to;
103 };
104 
105 static struct tipc_bcbearer *bcbearer;
106 static struct tipc_bclink *bclink;
107 static struct tipc_link *bcl;
108 
109 const char tipc_bclink_name[] = "broadcast-link";
110 
111 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
112 			   struct tipc_node_map *nm_b,
113 			   struct tipc_node_map *nm_diff);
114 static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node);
115 static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node);
116 
117 static void tipc_bclink_lock(void)
118 {
119 	spin_lock_bh(&bclink->lock);
120 }
121 
122 static void tipc_bclink_unlock(void)
123 {
124 	struct tipc_node *node = NULL;
125 
126 	if (likely(!bclink->flags)) {
127 		spin_unlock_bh(&bclink->lock);
128 		return;
129 	}
130 
131 	if (bclink->flags & TIPC_BCLINK_RESET) {
132 		bclink->flags &= ~TIPC_BCLINK_RESET;
133 		node = tipc_bclink_retransmit_to();
134 	}
135 	spin_unlock_bh(&bclink->lock);
136 
137 	if (node)
138 		tipc_link_reset_all(node);
139 }
140 
141 void tipc_bclink_set_flags(unsigned int flags)
142 {
143 	bclink->flags |= flags;
144 }
145 
146 static u32 bcbuf_acks(struct sk_buff *buf)
147 {
148 	return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle;
149 }
150 
151 static void bcbuf_set_acks(struct sk_buff *buf, u32 acks)
152 {
153 	TIPC_SKB_CB(buf)->handle = (void *)(unsigned long)acks;
154 }
155 
156 static void bcbuf_decr_acks(struct sk_buff *buf)
157 {
158 	bcbuf_set_acks(buf, bcbuf_acks(buf) - 1);
159 }
160 
161 void tipc_bclink_add_node(u32 addr)
162 {
163 	tipc_bclink_lock();
164 	tipc_nmap_add(&bclink->bcast_nodes, addr);
165 	tipc_bclink_unlock();
166 }
167 
168 void tipc_bclink_remove_node(u32 addr)
169 {
170 	tipc_bclink_lock();
171 	tipc_nmap_remove(&bclink->bcast_nodes, addr);
172 	tipc_bclink_unlock();
173 }
174 
175 static void bclink_set_last_sent(void)
176 {
177 	if (bcl->next_out)
178 		bcl->fsm_msg_cnt = mod(buf_seqno(bcl->next_out) - 1);
179 	else
180 		bcl->fsm_msg_cnt = mod(bcl->next_out_no - 1);
181 }
182 
183 u32 tipc_bclink_get_last_sent(void)
184 {
185 	return bcl->fsm_msg_cnt;
186 }
187 
188 static void bclink_update_last_sent(struct tipc_node *node, u32 seqno)
189 {
190 	node->bclink.last_sent = less_eq(node->bclink.last_sent, seqno) ?
191 						seqno : node->bclink.last_sent;
192 }
193 
194 
195 /**
196  * tipc_bclink_retransmit_to - get most recent node to request retransmission
197  *
198  * Called with bclink_lock locked
199  */
200 struct tipc_node *tipc_bclink_retransmit_to(void)
201 {
202 	return bclink->retransmit_to;
203 }
204 
205 /**
206  * bclink_retransmit_pkt - retransmit broadcast packets
207  * @after: sequence number of last packet to *not* retransmit
208  * @to: sequence number of last packet to retransmit
209  *
210  * Called with bclink_lock locked
211  */
212 static void bclink_retransmit_pkt(u32 after, u32 to)
213 {
214 	struct sk_buff *buf;
215 
216 	buf = bcl->first_out;
217 	while (buf && less_eq(buf_seqno(buf), after))
218 		buf = buf->next;
219 	tipc_link_retransmit(bcl, buf, mod(to - after));
220 }
221 
222 /**
223  * tipc_bclink_acknowledge - handle acknowledgement of broadcast packets
224  * @n_ptr: node that sent acknowledgement info
225  * @acked: broadcast sequence # that has been acknowledged
226  *
227  * Node is locked, bclink_lock unlocked.
228  */
229 void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked)
230 {
231 	struct sk_buff *crs;
232 	struct sk_buff *next;
233 	unsigned int released = 0;
234 
235 	tipc_bclink_lock();
236 	/* Bail out if tx queue is empty (no clean up is required) */
237 	crs = bcl->first_out;
238 	if (!crs)
239 		goto exit;
240 
241 	/* Determine which messages need to be acknowledged */
242 	if (acked == INVALID_LINK_SEQ) {
243 		/*
244 		 * Contact with specified node has been lost, so need to
245 		 * acknowledge sent messages only (if other nodes still exist)
246 		 * or both sent and unsent messages (otherwise)
247 		 */
248 		if (bclink->bcast_nodes.count)
249 			acked = bcl->fsm_msg_cnt;
250 		else
251 			acked = bcl->next_out_no;
252 	} else {
253 		/*
254 		 * Bail out if specified sequence number does not correspond
255 		 * to a message that has been sent and not yet acknowledged
256 		 */
257 		if (less(acked, buf_seqno(crs)) ||
258 		    less(bcl->fsm_msg_cnt, acked) ||
259 		    less_eq(acked, n_ptr->bclink.acked))
260 			goto exit;
261 	}
262 
263 	/* Skip over packets that node has previously acknowledged */
264 	while (crs && less_eq(buf_seqno(crs), n_ptr->bclink.acked))
265 		crs = crs->next;
266 
267 	/* Update packets that node is now acknowledging */
268 
269 	while (crs && less_eq(buf_seqno(crs), acked)) {
270 		next = crs->next;
271 
272 		if (crs != bcl->next_out)
273 			bcbuf_decr_acks(crs);
274 		else {
275 			bcbuf_set_acks(crs, 0);
276 			bcl->next_out = next;
277 			bclink_set_last_sent();
278 		}
279 
280 		if (bcbuf_acks(crs) == 0) {
281 			bcl->first_out = next;
282 			bcl->out_queue_size--;
283 			kfree_skb(crs);
284 			released = 1;
285 		}
286 		crs = next;
287 	}
288 	n_ptr->bclink.acked = acked;
289 
290 	/* Try resolving broadcast link congestion, if necessary */
291 
292 	if (unlikely(bcl->next_out)) {
293 		tipc_link_push_queue(bcl);
294 		bclink_set_last_sent();
295 	}
296 	if (unlikely(released && !list_empty(&bcl->waiting_ports)))
297 		tipc_link_wakeup_ports(bcl, 0);
298 exit:
299 	tipc_bclink_unlock();
300 }
301 
302 /**
303  * tipc_bclink_update_link_state - update broadcast link state
304  *
305  * RCU and node lock set
306  */
307 void tipc_bclink_update_link_state(struct tipc_node *n_ptr, u32 last_sent)
308 {
309 	struct sk_buff *buf;
310 
311 	/* Ignore "stale" link state info */
312 
313 	if (less_eq(last_sent, n_ptr->bclink.last_in))
314 		return;
315 
316 	/* Update link synchronization state; quit if in sync */
317 
318 	bclink_update_last_sent(n_ptr, last_sent);
319 
320 	if (n_ptr->bclink.last_sent == n_ptr->bclink.last_in)
321 		return;
322 
323 	/* Update out-of-sync state; quit if loss is still unconfirmed */
324 
325 	if ((++n_ptr->bclink.oos_state) == 1) {
326 		if (n_ptr->bclink.deferred_size < (TIPC_MIN_LINK_WIN / 2))
327 			return;
328 		n_ptr->bclink.oos_state++;
329 	}
330 
331 	/* Don't NACK if one has been recently sent (or seen) */
332 
333 	if (n_ptr->bclink.oos_state & 0x1)
334 		return;
335 
336 	/* Send NACK */
337 
338 	buf = tipc_buf_acquire(INT_H_SIZE);
339 	if (buf) {
340 		struct tipc_msg *msg = buf_msg(buf);
341 
342 		tipc_msg_init(msg, BCAST_PROTOCOL, STATE_MSG,
343 			      INT_H_SIZE, n_ptr->addr);
344 		msg_set_non_seq(msg, 1);
345 		msg_set_mc_netid(msg, tipc_net_id);
346 		msg_set_bcast_ack(msg, n_ptr->bclink.last_in);
347 		msg_set_bcgap_after(msg, n_ptr->bclink.last_in);
348 		msg_set_bcgap_to(msg, n_ptr->bclink.deferred_head
349 				 ? buf_seqno(n_ptr->bclink.deferred_head) - 1
350 				 : n_ptr->bclink.last_sent);
351 
352 		tipc_bclink_lock();
353 		tipc_bearer_send(MAX_BEARERS, buf, NULL);
354 		bcl->stats.sent_nacks++;
355 		tipc_bclink_unlock();
356 		kfree_skb(buf);
357 
358 		n_ptr->bclink.oos_state++;
359 	}
360 }
361 
362 /**
363  * bclink_peek_nack - monitor retransmission requests sent by other nodes
364  *
365  * Delay any upcoming NACK by this node if another node has already
366  * requested the first message this node is going to ask for.
367  */
368 static void bclink_peek_nack(struct tipc_msg *msg)
369 {
370 	struct tipc_node *n_ptr = tipc_node_find(msg_destnode(msg));
371 
372 	if (unlikely(!n_ptr))
373 		return;
374 
375 	tipc_node_lock(n_ptr);
376 
377 	if (n_ptr->bclink.recv_permitted &&
378 	    (n_ptr->bclink.last_in != n_ptr->bclink.last_sent) &&
379 	    (n_ptr->bclink.last_in == msg_bcgap_after(msg)))
380 		n_ptr->bclink.oos_state = 2;
381 
382 	tipc_node_unlock(n_ptr);
383 }
384 
385 /*
386  * tipc_bclink_xmit - broadcast a packet to all nodes in cluster
387  */
388 int tipc_bclink_xmit(struct sk_buff *buf)
389 {
390 	int res;
391 
392 	tipc_bclink_lock();
393 
394 	if (!bclink->bcast_nodes.count) {
395 		res = msg_data_sz(buf_msg(buf));
396 		kfree_skb(buf);
397 		goto exit;
398 	}
399 
400 	res = __tipc_link_xmit(bcl, buf);
401 	if (likely(res >= 0)) {
402 		bclink_set_last_sent();
403 		bcl->stats.queue_sz_counts++;
404 		bcl->stats.accu_queue_sz += bcl->out_queue_size;
405 	}
406 exit:
407 	tipc_bclink_unlock();
408 	return res;
409 }
410 
411 /**
412  * bclink_accept_pkt - accept an incoming, in-sequence broadcast packet
413  *
414  * Called with both sending node's lock and bclink_lock taken.
415  */
416 static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
417 {
418 	bclink_update_last_sent(node, seqno);
419 	node->bclink.last_in = seqno;
420 	node->bclink.oos_state = 0;
421 	bcl->stats.recv_info++;
422 
423 	/*
424 	 * Unicast an ACK periodically, ensuring that
425 	 * all nodes in the cluster don't ACK at the same time
426 	 */
427 
428 	if (((seqno - tipc_own_addr) % TIPC_MIN_LINK_WIN) == 0) {
429 		tipc_link_proto_xmit(node->active_links[node->addr & 1],
430 				     STATE_MSG, 0, 0, 0, 0, 0);
431 		bcl->stats.sent_acks++;
432 	}
433 }
434 
435 /**
436  * tipc_bclink_rcv - receive a broadcast packet, and deliver upwards
437  *
438  * RCU is locked, no other locks set
439  */
440 void tipc_bclink_rcv(struct sk_buff *buf)
441 {
442 	struct tipc_msg *msg = buf_msg(buf);
443 	struct tipc_node *node;
444 	u32 next_in;
445 	u32 seqno;
446 	int deferred;
447 
448 	/* Screen out unwanted broadcast messages */
449 
450 	if (msg_mc_netid(msg) != tipc_net_id)
451 		goto exit;
452 
453 	node = tipc_node_find(msg_prevnode(msg));
454 	if (unlikely(!node))
455 		goto exit;
456 
457 	tipc_node_lock(node);
458 	if (unlikely(!node->bclink.recv_permitted))
459 		goto unlock;
460 
461 	/* Handle broadcast protocol message */
462 
463 	if (unlikely(msg_user(msg) == BCAST_PROTOCOL)) {
464 		if (msg_type(msg) != STATE_MSG)
465 			goto unlock;
466 		if (msg_destnode(msg) == tipc_own_addr) {
467 			tipc_bclink_acknowledge(node, msg_bcast_ack(msg));
468 			tipc_node_unlock(node);
469 			tipc_bclink_lock();
470 			bcl->stats.recv_nacks++;
471 			bclink->retransmit_to = node;
472 			bclink_retransmit_pkt(msg_bcgap_after(msg),
473 					      msg_bcgap_to(msg));
474 			tipc_bclink_unlock();
475 		} else {
476 			tipc_node_unlock(node);
477 			bclink_peek_nack(msg);
478 		}
479 		goto exit;
480 	}
481 
482 	/* Handle in-sequence broadcast message */
483 
484 	seqno = msg_seqno(msg);
485 	next_in = mod(node->bclink.last_in + 1);
486 
487 	if (likely(seqno == next_in)) {
488 receive:
489 		/* Deliver message to destination */
490 
491 		if (likely(msg_isdata(msg))) {
492 			tipc_bclink_lock();
493 			bclink_accept_pkt(node, seqno);
494 			tipc_bclink_unlock();
495 			tipc_node_unlock(node);
496 			if (likely(msg_mcast(msg)))
497 				tipc_port_mcast_rcv(buf, NULL);
498 			else
499 				kfree_skb(buf);
500 		} else if (msg_user(msg) == MSG_BUNDLER) {
501 			tipc_bclink_lock();
502 			bclink_accept_pkt(node, seqno);
503 			bcl->stats.recv_bundles++;
504 			bcl->stats.recv_bundled += msg_msgcnt(msg);
505 			tipc_bclink_unlock();
506 			tipc_node_unlock(node);
507 			tipc_link_bundle_rcv(buf);
508 		} else if (msg_user(msg) == MSG_FRAGMENTER) {
509 			tipc_buf_append(&node->bclink.reasm_buf, &buf);
510 			if (unlikely(!buf && !node->bclink.reasm_buf))
511 				goto unlock;
512 			tipc_bclink_lock();
513 			bclink_accept_pkt(node, seqno);
514 			bcl->stats.recv_fragments++;
515 			if (buf) {
516 				bcl->stats.recv_fragmented++;
517 				msg = buf_msg(buf);
518 				tipc_bclink_unlock();
519 				goto receive;
520 			}
521 			tipc_bclink_unlock();
522 			tipc_node_unlock(node);
523 		} else if (msg_user(msg) == NAME_DISTRIBUTOR) {
524 			tipc_bclink_lock();
525 			bclink_accept_pkt(node, seqno);
526 			tipc_bclink_unlock();
527 			tipc_node_unlock(node);
528 			tipc_named_rcv(buf);
529 		} else {
530 			tipc_bclink_lock();
531 			bclink_accept_pkt(node, seqno);
532 			tipc_bclink_unlock();
533 			tipc_node_unlock(node);
534 			kfree_skb(buf);
535 		}
536 		buf = NULL;
537 
538 		/* Determine new synchronization state */
539 
540 		tipc_node_lock(node);
541 		if (unlikely(!tipc_node_is_up(node)))
542 			goto unlock;
543 
544 		if (node->bclink.last_in == node->bclink.last_sent)
545 			goto unlock;
546 
547 		if (!node->bclink.deferred_head) {
548 			node->bclink.oos_state = 1;
549 			goto unlock;
550 		}
551 
552 		msg = buf_msg(node->bclink.deferred_head);
553 		seqno = msg_seqno(msg);
554 		next_in = mod(next_in + 1);
555 		if (seqno != next_in)
556 			goto unlock;
557 
558 		/* Take in-sequence message from deferred queue & deliver it */
559 
560 		buf = node->bclink.deferred_head;
561 		node->bclink.deferred_head = buf->next;
562 		buf->next = NULL;
563 		node->bclink.deferred_size--;
564 		goto receive;
565 	}
566 
567 	/* Handle out-of-sequence broadcast message */
568 
569 	if (less(next_in, seqno)) {
570 		deferred = tipc_link_defer_pkt(&node->bclink.deferred_head,
571 					       &node->bclink.deferred_tail,
572 					       buf);
573 		node->bclink.deferred_size += deferred;
574 		bclink_update_last_sent(node, seqno);
575 		buf = NULL;
576 	} else
577 		deferred = 0;
578 
579 	tipc_bclink_lock();
580 
581 	if (deferred)
582 		bcl->stats.deferred_recv++;
583 	else
584 		bcl->stats.duplicates++;
585 
586 	tipc_bclink_unlock();
587 
588 unlock:
589 	tipc_node_unlock(node);
590 exit:
591 	kfree_skb(buf);
592 }
593 
594 u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr)
595 {
596 	return (n_ptr->bclink.recv_permitted &&
597 		(tipc_bclink_get_last_sent() != n_ptr->bclink.acked));
598 }
599 
600 
601 /**
602  * tipc_bcbearer_send - send a packet through the broadcast pseudo-bearer
603  *
604  * Send packet over as many bearers as necessary to reach all nodes
605  * that have joined the broadcast link.
606  *
607  * Returns 0 (packet sent successfully) under all circumstances,
608  * since the broadcast link's pseudo-bearer never blocks
609  */
610 static int tipc_bcbearer_send(struct sk_buff *buf, struct tipc_bearer *unused1,
611 			      struct tipc_media_addr *unused2)
612 {
613 	int bp_index;
614 
615 	/* Prepare broadcast link message for reliable transmission,
616 	 * if first time trying to send it;
617 	 * preparation is skipped for broadcast link protocol messages
618 	 * since they are sent in an unreliable manner and don't need it
619 	 */
620 	if (likely(!msg_non_seq(buf_msg(buf)))) {
621 		struct tipc_msg *msg;
622 
623 		bcbuf_set_acks(buf, bclink->bcast_nodes.count);
624 		msg = buf_msg(buf);
625 		msg_set_non_seq(msg, 1);
626 		msg_set_mc_netid(msg, tipc_net_id);
627 		bcl->stats.sent_info++;
628 
629 		if (WARN_ON(!bclink->bcast_nodes.count)) {
630 			dump_stack();
631 			return 0;
632 		}
633 	}
634 
635 	/* Send buffer over bearers until all targets reached */
636 	bcbearer->remains = bclink->bcast_nodes;
637 
638 	for (bp_index = 0; bp_index < MAX_BEARERS; bp_index++) {
639 		struct tipc_bearer *p = bcbearer->bpairs[bp_index].primary;
640 		struct tipc_bearer *s = bcbearer->bpairs[bp_index].secondary;
641 		struct tipc_bearer *b = p;
642 		struct sk_buff *tbuf;
643 
644 		if (!p)
645 			break; /* No more bearers to try */
646 
647 		tipc_nmap_diff(&bcbearer->remains, &b->nodes,
648 			       &bcbearer->remains_new);
649 		if (bcbearer->remains_new.count == bcbearer->remains.count)
650 			continue; /* Nothing added by bearer pair */
651 
652 		if (bp_index == 0) {
653 			/* Use original buffer for first bearer */
654 			tipc_bearer_send(b->identity, buf, &b->bcast_addr);
655 		} else {
656 			/* Avoid concurrent buffer access */
657 			tbuf = pskb_copy_for_clone(buf, GFP_ATOMIC);
658 			if (!tbuf)
659 				break;
660 			tipc_bearer_send(b->identity, tbuf, &b->bcast_addr);
661 			kfree_skb(tbuf); /* Bearer keeps a clone */
662 		}
663 
664 		/* Swap bearers for next packet */
665 		if (s) {
666 			bcbearer->bpairs[bp_index].primary = s;
667 			bcbearer->bpairs[bp_index].secondary = p;
668 		}
669 
670 		if (bcbearer->remains_new.count == 0)
671 			break; /* All targets reached */
672 
673 		bcbearer->remains = bcbearer->remains_new;
674 	}
675 
676 	return 0;
677 }
678 
679 /**
680  * tipc_bcbearer_sort - create sets of bearer pairs used by broadcast bearer
681  */
682 void tipc_bcbearer_sort(struct tipc_node_map *nm_ptr, u32 node, bool action)
683 {
684 	struct tipc_bcbearer_pair *bp_temp = bcbearer->bpairs_temp;
685 	struct tipc_bcbearer_pair *bp_curr;
686 	struct tipc_bearer *b;
687 	int b_index;
688 	int pri;
689 
690 	tipc_bclink_lock();
691 
692 	if (action)
693 		tipc_nmap_add(nm_ptr, node);
694 	else
695 		tipc_nmap_remove(nm_ptr, node);
696 
697 	/* Group bearers by priority (can assume max of two per priority) */
698 	memset(bp_temp, 0, sizeof(bcbearer->bpairs_temp));
699 
700 	rcu_read_lock();
701 	for (b_index = 0; b_index < MAX_BEARERS; b_index++) {
702 		b = rcu_dereference_rtnl(bearer_list[b_index]);
703 		if (!b || !b->nodes.count)
704 			continue;
705 
706 		if (!bp_temp[b->priority].primary)
707 			bp_temp[b->priority].primary = b;
708 		else
709 			bp_temp[b->priority].secondary = b;
710 	}
711 	rcu_read_unlock();
712 
713 	/* Create array of bearer pairs for broadcasting */
714 	bp_curr = bcbearer->bpairs;
715 	memset(bcbearer->bpairs, 0, sizeof(bcbearer->bpairs));
716 
717 	for (pri = TIPC_MAX_LINK_PRI; pri >= 0; pri--) {
718 
719 		if (!bp_temp[pri].primary)
720 			continue;
721 
722 		bp_curr->primary = bp_temp[pri].primary;
723 
724 		if (bp_temp[pri].secondary) {
725 			if (tipc_nmap_equal(&bp_temp[pri].primary->nodes,
726 					    &bp_temp[pri].secondary->nodes)) {
727 				bp_curr->secondary = bp_temp[pri].secondary;
728 			} else {
729 				bp_curr++;
730 				bp_curr->primary = bp_temp[pri].secondary;
731 			}
732 		}
733 
734 		bp_curr++;
735 	}
736 
737 	tipc_bclink_unlock();
738 }
739 
740 
741 int tipc_bclink_stats(char *buf, const u32 buf_size)
742 {
743 	int ret;
744 	struct tipc_stats *s;
745 
746 	if (!bcl)
747 		return 0;
748 
749 	tipc_bclink_lock();
750 
751 	s = &bcl->stats;
752 
753 	ret = tipc_snprintf(buf, buf_size, "Link <%s>\n"
754 			    "  Window:%u packets\n",
755 			    bcl->name, bcl->queue_limit[0]);
756 	ret += tipc_snprintf(buf + ret, buf_size - ret,
757 			     "  RX packets:%u fragments:%u/%u bundles:%u/%u\n",
758 			     s->recv_info, s->recv_fragments,
759 			     s->recv_fragmented, s->recv_bundles,
760 			     s->recv_bundled);
761 	ret += tipc_snprintf(buf + ret, buf_size - ret,
762 			     "  TX packets:%u fragments:%u/%u bundles:%u/%u\n",
763 			     s->sent_info, s->sent_fragments,
764 			     s->sent_fragmented, s->sent_bundles,
765 			     s->sent_bundled);
766 	ret += tipc_snprintf(buf + ret, buf_size - ret,
767 			     "  RX naks:%u defs:%u dups:%u\n",
768 			     s->recv_nacks, s->deferred_recv, s->duplicates);
769 	ret += tipc_snprintf(buf + ret, buf_size - ret,
770 			     "  TX naks:%u acks:%u dups:%u\n",
771 			     s->sent_nacks, s->sent_acks, s->retransmitted);
772 	ret += tipc_snprintf(buf + ret, buf_size - ret,
773 			     "  Congestion link:%u  Send queue max:%u avg:%u\n",
774 			     s->link_congs, s->max_queue_sz,
775 			     s->queue_sz_counts ?
776 			     (s->accu_queue_sz / s->queue_sz_counts) : 0);
777 
778 	tipc_bclink_unlock();
779 	return ret;
780 }
781 
782 int tipc_bclink_reset_stats(void)
783 {
784 	if (!bcl)
785 		return -ENOPROTOOPT;
786 
787 	tipc_bclink_lock();
788 	memset(&bcl->stats, 0, sizeof(bcl->stats));
789 	tipc_bclink_unlock();
790 	return 0;
791 }
792 
793 int tipc_bclink_set_queue_limits(u32 limit)
794 {
795 	if (!bcl)
796 		return -ENOPROTOOPT;
797 	if ((limit < TIPC_MIN_LINK_WIN) || (limit > TIPC_MAX_LINK_WIN))
798 		return -EINVAL;
799 
800 	tipc_bclink_lock();
801 	tipc_link_set_queue_limits(bcl, limit);
802 	tipc_bclink_unlock();
803 	return 0;
804 }
805 
806 int tipc_bclink_init(void)
807 {
808 	bcbearer = kzalloc(sizeof(*bcbearer), GFP_ATOMIC);
809 	if (!bcbearer)
810 		return -ENOMEM;
811 
812 	bclink = kzalloc(sizeof(*bclink), GFP_ATOMIC);
813 	if (!bclink) {
814 		kfree(bcbearer);
815 		return -ENOMEM;
816 	}
817 
818 	bcl = &bclink->link;
819 	bcbearer->bearer.media = &bcbearer->media;
820 	bcbearer->media.send_msg = tipc_bcbearer_send;
821 	sprintf(bcbearer->media.name, "tipc-broadcast");
822 
823 	spin_lock_init(&bclink->lock);
824 	INIT_LIST_HEAD(&bcl->waiting_ports);
825 	bcl->next_out_no = 1;
826 	spin_lock_init(&bclink->node.lock);
827 	bcl->owner = &bclink->node;
828 	bcl->max_pkt = MAX_PKT_DEFAULT_MCAST;
829 	tipc_link_set_queue_limits(bcl, BCLINK_WIN_DEFAULT);
830 	bcl->bearer_id = MAX_BEARERS;
831 	rcu_assign_pointer(bearer_list[MAX_BEARERS], &bcbearer->bearer);
832 	bcl->state = WORKING_WORKING;
833 	strlcpy(bcl->name, tipc_bclink_name, TIPC_MAX_LINK_NAME);
834 	return 0;
835 }
836 
837 void tipc_bclink_stop(void)
838 {
839 	tipc_bclink_lock();
840 	tipc_link_purge_queues(bcl);
841 	tipc_bclink_unlock();
842 
843 	RCU_INIT_POINTER(bearer_list[BCBEARER], NULL);
844 	synchronize_net();
845 	kfree(bcbearer);
846 	kfree(bclink);
847 }
848 
849 /**
850  * tipc_nmap_add - add a node to a node map
851  */
852 static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node)
853 {
854 	int n = tipc_node(node);
855 	int w = n / WSIZE;
856 	u32 mask = (1 << (n % WSIZE));
857 
858 	if ((nm_ptr->map[w] & mask) == 0) {
859 		nm_ptr->count++;
860 		nm_ptr->map[w] |= mask;
861 	}
862 }
863 
864 /**
865  * tipc_nmap_remove - remove a node from a node map
866  */
867 static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node)
868 {
869 	int n = tipc_node(node);
870 	int w = n / WSIZE;
871 	u32 mask = (1 << (n % WSIZE));
872 
873 	if ((nm_ptr->map[w] & mask) != 0) {
874 		nm_ptr->map[w] &= ~mask;
875 		nm_ptr->count--;
876 	}
877 }
878 
879 /**
880  * tipc_nmap_diff - find differences between node maps
881  * @nm_a: input node map A
882  * @nm_b: input node map B
883  * @nm_diff: output node map A-B (i.e. nodes of A that are not in B)
884  */
885 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
886 			   struct tipc_node_map *nm_b,
887 			   struct tipc_node_map *nm_diff)
888 {
889 	int stop = ARRAY_SIZE(nm_a->map);
890 	int w;
891 	int b;
892 	u32 map;
893 
894 	memset(nm_diff, 0, sizeof(*nm_diff));
895 	for (w = 0; w < stop; w++) {
896 		map = nm_a->map[w] ^ (nm_a->map[w] & nm_b->map[w]);
897 		nm_diff->map[w] = map;
898 		if (map != 0) {
899 			for (b = 0 ; b < WSIZE; b++) {
900 				if (map & (1 << b))
901 					nm_diff->count++;
902 			}
903 		}
904 	}
905 }
906 
907 /**
908  * tipc_port_list_add - add a port to a port list, ensuring no duplicates
909  */
910 void tipc_port_list_add(struct tipc_port_list *pl_ptr, u32 port)
911 {
912 	struct tipc_port_list *item = pl_ptr;
913 	int i;
914 	int item_sz = PLSIZE;
915 	int cnt = pl_ptr->count;
916 
917 	for (; ; cnt -= item_sz, item = item->next) {
918 		if (cnt < PLSIZE)
919 			item_sz = cnt;
920 		for (i = 0; i < item_sz; i++)
921 			if (item->ports[i] == port)
922 				return;
923 		if (i < PLSIZE) {
924 			item->ports[i] = port;
925 			pl_ptr->count++;
926 			return;
927 		}
928 		if (!item->next) {
929 			item->next = kmalloc(sizeof(*item), GFP_ATOMIC);
930 			if (!item->next) {
931 				pr_warn("Incomplete multicast delivery, no memory\n");
932 				return;
933 			}
934 			item->next->next = NULL;
935 		}
936 	}
937 }
938 
939 /**
940  * tipc_port_list_free - free dynamically created entries in port_list chain
941  *
942  */
943 void tipc_port_list_free(struct tipc_port_list *pl_ptr)
944 {
945 	struct tipc_port_list *item;
946 	struct tipc_port_list *next;
947 
948 	for (item = pl_ptr->next; item; item = next) {
949 		next = item->next;
950 		kfree(item);
951 	}
952 }
953