xref: /openbmc/linux/net/tipc/bcast.c (revision 034f90b3)
1 /*
2  * net/tipc/bcast.c: TIPC broadcast code
3  *
4  * Copyright (c) 2004-2006, 2014-2015, 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 "socket.h"
39 #include "msg.h"
40 #include "bcast.h"
41 #include "name_distr.h"
42 #include "core.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 
47 const char tipc_bclink_name[] = "broadcast-link";
48 
49 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
50 			   struct tipc_node_map *nm_b,
51 			   struct tipc_node_map *nm_diff);
52 static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node);
53 static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node);
54 
55 static void tipc_bclink_lock(struct net *net)
56 {
57 	struct tipc_net *tn = net_generic(net, tipc_net_id);
58 
59 	spin_lock_bh(&tn->bclink->lock);
60 }
61 
62 static void tipc_bclink_unlock(struct net *net)
63 {
64 	struct tipc_net *tn = net_generic(net, tipc_net_id);
65 	struct tipc_node *node = NULL;
66 
67 	if (likely(!tn->bclink->flags)) {
68 		spin_unlock_bh(&tn->bclink->lock);
69 		return;
70 	}
71 
72 	if (tn->bclink->flags & TIPC_BCLINK_RESET) {
73 		tn->bclink->flags &= ~TIPC_BCLINK_RESET;
74 		node = tipc_bclink_retransmit_to(net);
75 	}
76 	spin_unlock_bh(&tn->bclink->lock);
77 
78 	if (node)
79 		tipc_link_reset_all(node);
80 }
81 
82 void tipc_bclink_input(struct net *net)
83 {
84 	struct tipc_net *tn = net_generic(net, tipc_net_id);
85 
86 	tipc_sk_mcast_rcv(net, &tn->bclink->arrvq, &tn->bclink->inputq);
87 }
88 
89 uint  tipc_bclink_get_mtu(void)
90 {
91 	return MAX_PKT_DEFAULT_MCAST;
92 }
93 
94 void tipc_bclink_set_flags(struct net *net, unsigned int flags)
95 {
96 	struct tipc_net *tn = net_generic(net, tipc_net_id);
97 
98 	tn->bclink->flags |= flags;
99 }
100 
101 static u32 bcbuf_acks(struct sk_buff *buf)
102 {
103 	return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle;
104 }
105 
106 static void bcbuf_set_acks(struct sk_buff *buf, u32 acks)
107 {
108 	TIPC_SKB_CB(buf)->handle = (void *)(unsigned long)acks;
109 }
110 
111 static void bcbuf_decr_acks(struct sk_buff *buf)
112 {
113 	bcbuf_set_acks(buf, bcbuf_acks(buf) - 1);
114 }
115 
116 void tipc_bclink_add_node(struct net *net, u32 addr)
117 {
118 	struct tipc_net *tn = net_generic(net, tipc_net_id);
119 
120 	tipc_bclink_lock(net);
121 	tipc_nmap_add(&tn->bclink->bcast_nodes, addr);
122 	tipc_bclink_unlock(net);
123 }
124 
125 void tipc_bclink_remove_node(struct net *net, u32 addr)
126 {
127 	struct tipc_net *tn = net_generic(net, tipc_net_id);
128 
129 	tipc_bclink_lock(net);
130 	tipc_nmap_remove(&tn->bclink->bcast_nodes, addr);
131 	tipc_bclink_unlock(net);
132 }
133 
134 static void bclink_set_last_sent(struct net *net)
135 {
136 	struct tipc_net *tn = net_generic(net, tipc_net_id);
137 	struct tipc_link *bcl = tn->bcl;
138 
139 	if (bcl->next_out)
140 		bcl->fsm_msg_cnt = mod(buf_seqno(bcl->next_out) - 1);
141 	else
142 		bcl->fsm_msg_cnt = mod(bcl->next_out_no - 1);
143 }
144 
145 u32 tipc_bclink_get_last_sent(struct net *net)
146 {
147 	struct tipc_net *tn = net_generic(net, tipc_net_id);
148 
149 	return tn->bcl->fsm_msg_cnt;
150 }
151 
152 static void bclink_update_last_sent(struct tipc_node *node, u32 seqno)
153 {
154 	node->bclink.last_sent = less_eq(node->bclink.last_sent, seqno) ?
155 						seqno : node->bclink.last_sent;
156 }
157 
158 
159 /**
160  * tipc_bclink_retransmit_to - get most recent node to request retransmission
161  *
162  * Called with bclink_lock locked
163  */
164 struct tipc_node *tipc_bclink_retransmit_to(struct net *net)
165 {
166 	struct tipc_net *tn = net_generic(net, tipc_net_id);
167 
168 	return tn->bclink->retransmit_to;
169 }
170 
171 /**
172  * bclink_retransmit_pkt - retransmit broadcast packets
173  * @after: sequence number of last packet to *not* retransmit
174  * @to: sequence number of last packet to retransmit
175  *
176  * Called with bclink_lock locked
177  */
178 static void bclink_retransmit_pkt(struct tipc_net *tn, u32 after, u32 to)
179 {
180 	struct sk_buff *skb;
181 	struct tipc_link *bcl = tn->bcl;
182 
183 	skb_queue_walk(&bcl->outqueue, skb) {
184 		if (more(buf_seqno(skb), after)) {
185 			tipc_link_retransmit(bcl, skb, mod(to - after));
186 			break;
187 		}
188 	}
189 }
190 
191 /**
192  * tipc_bclink_wakeup_users - wake up pending users
193  *
194  * Called with no locks taken
195  */
196 void tipc_bclink_wakeup_users(struct net *net)
197 {
198 	struct tipc_net *tn = net_generic(net, tipc_net_id);
199 
200 	tipc_sk_rcv(net, &tn->bclink->link.wakeupq);
201 }
202 
203 /**
204  * tipc_bclink_acknowledge - handle acknowledgement of broadcast packets
205  * @n_ptr: node that sent acknowledgement info
206  * @acked: broadcast sequence # that has been acknowledged
207  *
208  * Node is locked, bclink_lock unlocked.
209  */
210 void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked)
211 {
212 	struct sk_buff *skb, *tmp;
213 	struct sk_buff *next;
214 	unsigned int released = 0;
215 	struct net *net = n_ptr->net;
216 	struct tipc_net *tn = net_generic(net, tipc_net_id);
217 
218 	tipc_bclink_lock(net);
219 	/* Bail out if tx queue is empty (no clean up is required) */
220 	skb = skb_peek(&tn->bcl->outqueue);
221 	if (!skb)
222 		goto exit;
223 
224 	/* Determine which messages need to be acknowledged */
225 	if (acked == INVALID_LINK_SEQ) {
226 		/*
227 		 * Contact with specified node has been lost, so need to
228 		 * acknowledge sent messages only (if other nodes still exist)
229 		 * or both sent and unsent messages (otherwise)
230 		 */
231 		if (tn->bclink->bcast_nodes.count)
232 			acked = tn->bcl->fsm_msg_cnt;
233 		else
234 			acked = tn->bcl->next_out_no;
235 	} else {
236 		/*
237 		 * Bail out if specified sequence number does not correspond
238 		 * to a message that has been sent and not yet acknowledged
239 		 */
240 		if (less(acked, buf_seqno(skb)) ||
241 		    less(tn->bcl->fsm_msg_cnt, acked) ||
242 		    less_eq(acked, n_ptr->bclink.acked))
243 			goto exit;
244 	}
245 
246 	/* Skip over packets that node has previously acknowledged */
247 	skb_queue_walk(&tn->bcl->outqueue, skb) {
248 		if (more(buf_seqno(skb), n_ptr->bclink.acked))
249 			break;
250 	}
251 
252 	/* Update packets that node is now acknowledging */
253 	skb_queue_walk_from_safe(&tn->bcl->outqueue, skb, tmp) {
254 		if (more(buf_seqno(skb), acked))
255 			break;
256 
257 		next = tipc_skb_queue_next(&tn->bcl->outqueue, skb);
258 		if (skb != tn->bcl->next_out) {
259 			bcbuf_decr_acks(skb);
260 		} else {
261 			bcbuf_set_acks(skb, 0);
262 			tn->bcl->next_out = next;
263 			bclink_set_last_sent(net);
264 		}
265 
266 		if (bcbuf_acks(skb) == 0) {
267 			__skb_unlink(skb, &tn->bcl->outqueue);
268 			kfree_skb(skb);
269 			released = 1;
270 		}
271 	}
272 	n_ptr->bclink.acked = acked;
273 
274 	/* Try resolving broadcast link congestion, if necessary */
275 	if (unlikely(tn->bcl->next_out)) {
276 		tipc_link_push_packets(tn->bcl);
277 		bclink_set_last_sent(net);
278 	}
279 	if (unlikely(released && !skb_queue_empty(&tn->bcl->wakeupq)))
280 		n_ptr->action_flags |= TIPC_WAKEUP_BCAST_USERS;
281 exit:
282 	tipc_bclink_unlock(net);
283 }
284 
285 /**
286  * tipc_bclink_update_link_state - update broadcast link state
287  *
288  * RCU and node lock set
289  */
290 void tipc_bclink_update_link_state(struct tipc_node *n_ptr,
291 				   u32 last_sent)
292 {
293 	struct sk_buff *buf;
294 	struct net *net = n_ptr->net;
295 	struct tipc_net *tn = net_generic(net, tipc_net_id);
296 
297 	/* Ignore "stale" link state info */
298 	if (less_eq(last_sent, n_ptr->bclink.last_in))
299 		return;
300 
301 	/* Update link synchronization state; quit if in sync */
302 	bclink_update_last_sent(n_ptr, last_sent);
303 
304 	if (n_ptr->bclink.last_sent == n_ptr->bclink.last_in)
305 		return;
306 
307 	/* Update out-of-sync state; quit if loss is still unconfirmed */
308 	if ((++n_ptr->bclink.oos_state) == 1) {
309 		if (n_ptr->bclink.deferred_size < (TIPC_MIN_LINK_WIN / 2))
310 			return;
311 		n_ptr->bclink.oos_state++;
312 	}
313 
314 	/* Don't NACK if one has been recently sent (or seen) */
315 	if (n_ptr->bclink.oos_state & 0x1)
316 		return;
317 
318 	/* Send NACK */
319 	buf = tipc_buf_acquire(INT_H_SIZE);
320 	if (buf) {
321 		struct tipc_msg *msg = buf_msg(buf);
322 		struct sk_buff *skb = skb_peek(&n_ptr->bclink.deferred_queue);
323 		u32 to = skb ? buf_seqno(skb) - 1 : n_ptr->bclink.last_sent;
324 
325 		tipc_msg_init(tn->own_addr, msg, BCAST_PROTOCOL, STATE_MSG,
326 			      INT_H_SIZE, n_ptr->addr);
327 		msg_set_non_seq(msg, 1);
328 		msg_set_mc_netid(msg, tn->net_id);
329 		msg_set_bcast_ack(msg, n_ptr->bclink.last_in);
330 		msg_set_bcgap_after(msg, n_ptr->bclink.last_in);
331 		msg_set_bcgap_to(msg, to);
332 
333 		tipc_bclink_lock(net);
334 		tipc_bearer_send(net, MAX_BEARERS, buf, NULL);
335 		tn->bcl->stats.sent_nacks++;
336 		tipc_bclink_unlock(net);
337 		kfree_skb(buf);
338 
339 		n_ptr->bclink.oos_state++;
340 	}
341 }
342 
343 /**
344  * bclink_peek_nack - monitor retransmission requests sent by other nodes
345  *
346  * Delay any upcoming NACK by this node if another node has already
347  * requested the first message this node is going to ask for.
348  */
349 static void bclink_peek_nack(struct net *net, struct tipc_msg *msg)
350 {
351 	struct tipc_node *n_ptr = tipc_node_find(net, msg_destnode(msg));
352 
353 	if (unlikely(!n_ptr))
354 		return;
355 
356 	tipc_node_lock(n_ptr);
357 
358 	if (n_ptr->bclink.recv_permitted &&
359 	    (n_ptr->bclink.last_in != n_ptr->bclink.last_sent) &&
360 	    (n_ptr->bclink.last_in == msg_bcgap_after(msg)))
361 		n_ptr->bclink.oos_state = 2;
362 
363 	tipc_node_unlock(n_ptr);
364 }
365 
366 /* tipc_bclink_xmit - deliver buffer chain to all nodes in cluster
367  *                    and to identified node local sockets
368  * @net: the applicable net namespace
369  * @list: chain of buffers containing message
370  * Consumes the buffer chain, except when returning -ELINKCONG
371  * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE
372  */
373 int tipc_bclink_xmit(struct net *net, struct sk_buff_head *list)
374 {
375 	struct tipc_net *tn = net_generic(net, tipc_net_id);
376 	struct tipc_link *bcl = tn->bcl;
377 	struct tipc_bclink *bclink = tn->bclink;
378 	int rc = 0;
379 	int bc = 0;
380 	struct sk_buff *skb;
381 	struct sk_buff_head arrvq;
382 	struct sk_buff_head inputq;
383 
384 	/* Prepare clone of message for local node */
385 	skb = tipc_msg_reassemble(list);
386 	if (unlikely(!skb)) {
387 		__skb_queue_purge(list);
388 		return -EHOSTUNREACH;
389 	}
390 
391 	/* Broadcast to all nodes */
392 	if (likely(bclink)) {
393 		tipc_bclink_lock(net);
394 		if (likely(bclink->bcast_nodes.count)) {
395 			rc = __tipc_link_xmit(net, bcl, list);
396 			if (likely(!rc)) {
397 				u32 len = skb_queue_len(&bcl->outqueue);
398 
399 				bclink_set_last_sent(net);
400 				bcl->stats.queue_sz_counts++;
401 				bcl->stats.accu_queue_sz += len;
402 			}
403 			bc = 1;
404 		}
405 		tipc_bclink_unlock(net);
406 	}
407 
408 	if (unlikely(!bc))
409 		__skb_queue_purge(list);
410 
411 	if (unlikely(rc)) {
412 		kfree_skb(skb);
413 		return rc;
414 	}
415 	/* Deliver message clone */
416 	__skb_queue_head_init(&arrvq);
417 	skb_queue_head_init(&inputq);
418 	__skb_queue_tail(&arrvq, skb);
419 	tipc_sk_mcast_rcv(net, &arrvq, &inputq);
420 	return rc;
421 }
422 
423 /**
424  * bclink_accept_pkt - accept an incoming, in-sequence broadcast packet
425  *
426  * Called with both sending node's lock and bclink_lock taken.
427  */
428 static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
429 {
430 	struct tipc_net *tn = net_generic(node->net, tipc_net_id);
431 
432 	bclink_update_last_sent(node, seqno);
433 	node->bclink.last_in = seqno;
434 	node->bclink.oos_state = 0;
435 	tn->bcl->stats.recv_info++;
436 
437 	/*
438 	 * Unicast an ACK periodically, ensuring that
439 	 * all nodes in the cluster don't ACK at the same time
440 	 */
441 	if (((seqno - tn->own_addr) % TIPC_MIN_LINK_WIN) == 0) {
442 		tipc_link_proto_xmit(node->active_links[node->addr & 1],
443 				     STATE_MSG, 0, 0, 0, 0, 0);
444 		tn->bcl->stats.sent_acks++;
445 	}
446 }
447 
448 /**
449  * tipc_bclink_rcv - receive a broadcast packet, and deliver upwards
450  *
451  * RCU is locked, no other locks set
452  */
453 void tipc_bclink_rcv(struct net *net, struct sk_buff *buf)
454 {
455 	struct tipc_net *tn = net_generic(net, tipc_net_id);
456 	struct tipc_link *bcl = tn->bcl;
457 	struct tipc_msg *msg = buf_msg(buf);
458 	struct tipc_node *node;
459 	u32 next_in;
460 	u32 seqno;
461 	int deferred = 0;
462 	int pos = 0;
463 	struct sk_buff *iskb;
464 	struct sk_buff_head *arrvq, *inputq;
465 
466 	/* Screen out unwanted broadcast messages */
467 	if (msg_mc_netid(msg) != tn->net_id)
468 		goto exit;
469 
470 	node = tipc_node_find(net, msg_prevnode(msg));
471 	if (unlikely(!node))
472 		goto exit;
473 
474 	tipc_node_lock(node);
475 	if (unlikely(!node->bclink.recv_permitted))
476 		goto unlock;
477 
478 	/* Handle broadcast protocol message */
479 	if (unlikely(msg_user(msg) == BCAST_PROTOCOL)) {
480 		if (msg_type(msg) != STATE_MSG)
481 			goto unlock;
482 		if (msg_destnode(msg) == tn->own_addr) {
483 			tipc_bclink_acknowledge(node, msg_bcast_ack(msg));
484 			tipc_node_unlock(node);
485 			tipc_bclink_lock(net);
486 			bcl->stats.recv_nacks++;
487 			tn->bclink->retransmit_to = node;
488 			bclink_retransmit_pkt(tn, msg_bcgap_after(msg),
489 					      msg_bcgap_to(msg));
490 			tipc_bclink_unlock(net);
491 		} else {
492 			tipc_node_unlock(node);
493 			bclink_peek_nack(net, msg);
494 		}
495 		goto exit;
496 	}
497 
498 	/* Handle in-sequence broadcast message */
499 	seqno = msg_seqno(msg);
500 	next_in = mod(node->bclink.last_in + 1);
501 	arrvq = &tn->bclink->arrvq;
502 	inputq = &tn->bclink->inputq;
503 
504 	if (likely(seqno == next_in)) {
505 receive:
506 		/* Deliver message to destination */
507 		if (likely(msg_isdata(msg))) {
508 			tipc_bclink_lock(net);
509 			bclink_accept_pkt(node, seqno);
510 			spin_lock_bh(&inputq->lock);
511 			__skb_queue_tail(arrvq, buf);
512 			spin_unlock_bh(&inputq->lock);
513 			node->action_flags |= TIPC_BCAST_MSG_EVT;
514 			tipc_bclink_unlock(net);
515 			tipc_node_unlock(node);
516 		} else if (msg_user(msg) == MSG_BUNDLER) {
517 			tipc_bclink_lock(net);
518 			bclink_accept_pkt(node, seqno);
519 			bcl->stats.recv_bundles++;
520 			bcl->stats.recv_bundled += msg_msgcnt(msg);
521 			pos = 0;
522 			while (tipc_msg_extract(buf, &iskb, &pos)) {
523 				spin_lock_bh(&inputq->lock);
524 				__skb_queue_tail(arrvq, iskb);
525 				spin_unlock_bh(&inputq->lock);
526 			}
527 			node->action_flags |= TIPC_BCAST_MSG_EVT;
528 			tipc_bclink_unlock(net);
529 			tipc_node_unlock(node);
530 		} else if (msg_user(msg) == MSG_FRAGMENTER) {
531 			tipc_buf_append(&node->bclink.reasm_buf, &buf);
532 			if (unlikely(!buf && !node->bclink.reasm_buf))
533 				goto unlock;
534 			tipc_bclink_lock(net);
535 			bclink_accept_pkt(node, seqno);
536 			bcl->stats.recv_fragments++;
537 			if (buf) {
538 				bcl->stats.recv_fragmented++;
539 				msg = buf_msg(buf);
540 				tipc_bclink_unlock(net);
541 				goto receive;
542 			}
543 			tipc_bclink_unlock(net);
544 			tipc_node_unlock(node);
545 		} else {
546 			tipc_bclink_lock(net);
547 			bclink_accept_pkt(node, seqno);
548 			tipc_bclink_unlock(net);
549 			tipc_node_unlock(node);
550 			kfree_skb(buf);
551 		}
552 		buf = NULL;
553 
554 		/* Determine new synchronization state */
555 		tipc_node_lock(node);
556 		if (unlikely(!tipc_node_is_up(node)))
557 			goto unlock;
558 
559 		if (node->bclink.last_in == node->bclink.last_sent)
560 			goto unlock;
561 
562 		if (skb_queue_empty(&node->bclink.deferred_queue)) {
563 			node->bclink.oos_state = 1;
564 			goto unlock;
565 		}
566 
567 		msg = buf_msg(skb_peek(&node->bclink.deferred_queue));
568 		seqno = msg_seqno(msg);
569 		next_in = mod(next_in + 1);
570 		if (seqno != next_in)
571 			goto unlock;
572 
573 		/* Take in-sequence message from deferred queue & deliver it */
574 		buf = __skb_dequeue(&node->bclink.deferred_queue);
575 		goto receive;
576 	}
577 
578 	/* Handle out-of-sequence broadcast message */
579 	if (less(next_in, seqno)) {
580 		deferred = tipc_link_defer_pkt(&node->bclink.deferred_queue,
581 					       buf);
582 		bclink_update_last_sent(node, seqno);
583 		buf = NULL;
584 	}
585 
586 	tipc_bclink_lock(net);
587 
588 	if (deferred)
589 		bcl->stats.deferred_recv++;
590 	else
591 		bcl->stats.duplicates++;
592 
593 	tipc_bclink_unlock(net);
594 
595 unlock:
596 	tipc_node_unlock(node);
597 exit:
598 	kfree_skb(buf);
599 }
600 
601 u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr)
602 {
603 	return (n_ptr->bclink.recv_permitted &&
604 		(tipc_bclink_get_last_sent(n_ptr->net) != n_ptr->bclink.acked));
605 }
606 
607 
608 /**
609  * tipc_bcbearer_send - send a packet through the broadcast pseudo-bearer
610  *
611  * Send packet over as many bearers as necessary to reach all nodes
612  * that have joined the broadcast link.
613  *
614  * Returns 0 (packet sent successfully) under all circumstances,
615  * since the broadcast link's pseudo-bearer never blocks
616  */
617 static int tipc_bcbearer_send(struct net *net, struct sk_buff *buf,
618 			      struct tipc_bearer *unused1,
619 			      struct tipc_media_addr *unused2)
620 {
621 	int bp_index;
622 	struct tipc_msg *msg = buf_msg(buf);
623 	struct tipc_net *tn = net_generic(net, tipc_net_id);
624 	struct tipc_bcbearer *bcbearer = tn->bcbearer;
625 	struct tipc_bclink *bclink = tn->bclink;
626 
627 	/* Prepare broadcast link message for reliable transmission,
628 	 * if first time trying to send it;
629 	 * preparation is skipped for broadcast link protocol messages
630 	 * since they are sent in an unreliable manner and don't need it
631 	 */
632 	if (likely(!msg_non_seq(buf_msg(buf)))) {
633 		bcbuf_set_acks(buf, bclink->bcast_nodes.count);
634 		msg_set_non_seq(msg, 1);
635 		msg_set_mc_netid(msg, tn->net_id);
636 		tn->bcl->stats.sent_info++;
637 
638 		if (WARN_ON(!bclink->bcast_nodes.count)) {
639 			dump_stack();
640 			return 0;
641 		}
642 	}
643 
644 	/* Send buffer over bearers until all targets reached */
645 	bcbearer->remains = bclink->bcast_nodes;
646 
647 	for (bp_index = 0; bp_index < MAX_BEARERS; bp_index++) {
648 		struct tipc_bearer *p = bcbearer->bpairs[bp_index].primary;
649 		struct tipc_bearer *s = bcbearer->bpairs[bp_index].secondary;
650 		struct tipc_bearer *bp[2] = {p, s};
651 		struct tipc_bearer *b = bp[msg_link_selector(msg)];
652 		struct sk_buff *tbuf;
653 
654 		if (!p)
655 			break; /* No more bearers to try */
656 		if (!b)
657 			b = p;
658 		tipc_nmap_diff(&bcbearer->remains, &b->nodes,
659 			       &bcbearer->remains_new);
660 		if (bcbearer->remains_new.count == bcbearer->remains.count)
661 			continue; /* Nothing added by bearer pair */
662 
663 		if (bp_index == 0) {
664 			/* Use original buffer for first bearer */
665 			tipc_bearer_send(net, b->identity, buf, &b->bcast_addr);
666 		} else {
667 			/* Avoid concurrent buffer access */
668 			tbuf = pskb_copy_for_clone(buf, GFP_ATOMIC);
669 			if (!tbuf)
670 				break;
671 			tipc_bearer_send(net, b->identity, tbuf,
672 					 &b->bcast_addr);
673 			kfree_skb(tbuf); /* Bearer keeps a clone */
674 		}
675 		if (bcbearer->remains_new.count == 0)
676 			break; /* All targets reached */
677 
678 		bcbearer->remains = bcbearer->remains_new;
679 	}
680 
681 	return 0;
682 }
683 
684 /**
685  * tipc_bcbearer_sort - create sets of bearer pairs used by broadcast bearer
686  */
687 void tipc_bcbearer_sort(struct net *net, struct tipc_node_map *nm_ptr,
688 			u32 node, bool action)
689 {
690 	struct tipc_net *tn = net_generic(net, tipc_net_id);
691 	struct tipc_bcbearer *bcbearer = tn->bcbearer;
692 	struct tipc_bcbearer_pair *bp_temp = bcbearer->bpairs_temp;
693 	struct tipc_bcbearer_pair *bp_curr;
694 	struct tipc_bearer *b;
695 	int b_index;
696 	int pri;
697 
698 	tipc_bclink_lock(net);
699 
700 	if (action)
701 		tipc_nmap_add(nm_ptr, node);
702 	else
703 		tipc_nmap_remove(nm_ptr, node);
704 
705 	/* Group bearers by priority (can assume max of two per priority) */
706 	memset(bp_temp, 0, sizeof(bcbearer->bpairs_temp));
707 
708 	rcu_read_lock();
709 	for (b_index = 0; b_index < MAX_BEARERS; b_index++) {
710 		b = rcu_dereference_rtnl(tn->bearer_list[b_index]);
711 		if (!b || !b->nodes.count)
712 			continue;
713 
714 		if (!bp_temp[b->priority].primary)
715 			bp_temp[b->priority].primary = b;
716 		else
717 			bp_temp[b->priority].secondary = b;
718 	}
719 	rcu_read_unlock();
720 
721 	/* Create array of bearer pairs for broadcasting */
722 	bp_curr = bcbearer->bpairs;
723 	memset(bcbearer->bpairs, 0, sizeof(bcbearer->bpairs));
724 
725 	for (pri = TIPC_MAX_LINK_PRI; pri >= 0; pri--) {
726 
727 		if (!bp_temp[pri].primary)
728 			continue;
729 
730 		bp_curr->primary = bp_temp[pri].primary;
731 
732 		if (bp_temp[pri].secondary) {
733 			if (tipc_nmap_equal(&bp_temp[pri].primary->nodes,
734 					    &bp_temp[pri].secondary->nodes)) {
735 				bp_curr->secondary = bp_temp[pri].secondary;
736 			} else {
737 				bp_curr++;
738 				bp_curr->primary = bp_temp[pri].secondary;
739 			}
740 		}
741 
742 		bp_curr++;
743 	}
744 
745 	tipc_bclink_unlock(net);
746 }
747 
748 static int __tipc_nl_add_bc_link_stat(struct sk_buff *skb,
749 				      struct tipc_stats *stats)
750 {
751 	int i;
752 	struct nlattr *nest;
753 
754 	struct nla_map {
755 		__u32 key;
756 		__u32 val;
757 	};
758 
759 	struct nla_map map[] = {
760 		{TIPC_NLA_STATS_RX_INFO, stats->recv_info},
761 		{TIPC_NLA_STATS_RX_FRAGMENTS, stats->recv_fragments},
762 		{TIPC_NLA_STATS_RX_FRAGMENTED, stats->recv_fragmented},
763 		{TIPC_NLA_STATS_RX_BUNDLES, stats->recv_bundles},
764 		{TIPC_NLA_STATS_RX_BUNDLED, stats->recv_bundled},
765 		{TIPC_NLA_STATS_TX_INFO, stats->sent_info},
766 		{TIPC_NLA_STATS_TX_FRAGMENTS, stats->sent_fragments},
767 		{TIPC_NLA_STATS_TX_FRAGMENTED, stats->sent_fragmented},
768 		{TIPC_NLA_STATS_TX_BUNDLES, stats->sent_bundles},
769 		{TIPC_NLA_STATS_TX_BUNDLED, stats->sent_bundled},
770 		{TIPC_NLA_STATS_RX_NACKS, stats->recv_nacks},
771 		{TIPC_NLA_STATS_RX_DEFERRED, stats->deferred_recv},
772 		{TIPC_NLA_STATS_TX_NACKS, stats->sent_nacks},
773 		{TIPC_NLA_STATS_TX_ACKS, stats->sent_acks},
774 		{TIPC_NLA_STATS_RETRANSMITTED, stats->retransmitted},
775 		{TIPC_NLA_STATS_DUPLICATES, stats->duplicates},
776 		{TIPC_NLA_STATS_LINK_CONGS, stats->link_congs},
777 		{TIPC_NLA_STATS_MAX_QUEUE, stats->max_queue_sz},
778 		{TIPC_NLA_STATS_AVG_QUEUE, stats->queue_sz_counts ?
779 			(stats->accu_queue_sz / stats->queue_sz_counts) : 0}
780 	};
781 
782 	nest = nla_nest_start(skb, TIPC_NLA_LINK_STATS);
783 	if (!nest)
784 		return -EMSGSIZE;
785 
786 	for (i = 0; i <  ARRAY_SIZE(map); i++)
787 		if (nla_put_u32(skb, map[i].key, map[i].val))
788 			goto msg_full;
789 
790 	nla_nest_end(skb, nest);
791 
792 	return 0;
793 msg_full:
794 	nla_nest_cancel(skb, nest);
795 
796 	return -EMSGSIZE;
797 }
798 
799 int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg)
800 {
801 	int err;
802 	void *hdr;
803 	struct nlattr *attrs;
804 	struct nlattr *prop;
805 	struct tipc_net *tn = net_generic(net, tipc_net_id);
806 	struct tipc_link *bcl = tn->bcl;
807 
808 	if (!bcl)
809 		return 0;
810 
811 	tipc_bclink_lock(net);
812 
813 	hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
814 			  NLM_F_MULTI, TIPC_NL_LINK_GET);
815 	if (!hdr)
816 		return -EMSGSIZE;
817 
818 	attrs = nla_nest_start(msg->skb, TIPC_NLA_LINK);
819 	if (!attrs)
820 		goto msg_full;
821 
822 	/* The broadcast link is always up */
823 	if (nla_put_flag(msg->skb, TIPC_NLA_LINK_UP))
824 		goto attr_msg_full;
825 
826 	if (nla_put_flag(msg->skb, TIPC_NLA_LINK_BROADCAST))
827 		goto attr_msg_full;
828 	if (nla_put_string(msg->skb, TIPC_NLA_LINK_NAME, bcl->name))
829 		goto attr_msg_full;
830 	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, bcl->next_in_no))
831 		goto attr_msg_full;
832 	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, bcl->next_out_no))
833 		goto attr_msg_full;
834 
835 	prop = nla_nest_start(msg->skb, TIPC_NLA_LINK_PROP);
836 	if (!prop)
837 		goto attr_msg_full;
838 	if (nla_put_u32(msg->skb, TIPC_NLA_PROP_WIN, bcl->queue_limit[0]))
839 		goto prop_msg_full;
840 	nla_nest_end(msg->skb, prop);
841 
842 	err = __tipc_nl_add_bc_link_stat(msg->skb, &bcl->stats);
843 	if (err)
844 		goto attr_msg_full;
845 
846 	tipc_bclink_unlock(net);
847 	nla_nest_end(msg->skb, attrs);
848 	genlmsg_end(msg->skb, hdr);
849 
850 	return 0;
851 
852 prop_msg_full:
853 	nla_nest_cancel(msg->skb, prop);
854 attr_msg_full:
855 	nla_nest_cancel(msg->skb, attrs);
856 msg_full:
857 	tipc_bclink_unlock(net);
858 	genlmsg_cancel(msg->skb, hdr);
859 
860 	return -EMSGSIZE;
861 }
862 
863 int tipc_bclink_reset_stats(struct net *net)
864 {
865 	struct tipc_net *tn = net_generic(net, tipc_net_id);
866 	struct tipc_link *bcl = tn->bcl;
867 
868 	if (!bcl)
869 		return -ENOPROTOOPT;
870 
871 	tipc_bclink_lock(net);
872 	memset(&bcl->stats, 0, sizeof(bcl->stats));
873 	tipc_bclink_unlock(net);
874 	return 0;
875 }
876 
877 int tipc_bclink_set_queue_limits(struct net *net, u32 limit)
878 {
879 	struct tipc_net *tn = net_generic(net, tipc_net_id);
880 	struct tipc_link *bcl = tn->bcl;
881 
882 	if (!bcl)
883 		return -ENOPROTOOPT;
884 	if ((limit < TIPC_MIN_LINK_WIN) || (limit > TIPC_MAX_LINK_WIN))
885 		return -EINVAL;
886 
887 	tipc_bclink_lock(net);
888 	tipc_link_set_queue_limits(bcl, limit);
889 	tipc_bclink_unlock(net);
890 	return 0;
891 }
892 
893 int tipc_bclink_init(struct net *net)
894 {
895 	struct tipc_net *tn = net_generic(net, tipc_net_id);
896 	struct tipc_bcbearer *bcbearer;
897 	struct tipc_bclink *bclink;
898 	struct tipc_link *bcl;
899 
900 	bcbearer = kzalloc(sizeof(*bcbearer), GFP_ATOMIC);
901 	if (!bcbearer)
902 		return -ENOMEM;
903 
904 	bclink = kzalloc(sizeof(*bclink), GFP_ATOMIC);
905 	if (!bclink) {
906 		kfree(bcbearer);
907 		return -ENOMEM;
908 	}
909 
910 	bcl = &bclink->link;
911 	bcbearer->bearer.media = &bcbearer->media;
912 	bcbearer->media.send_msg = tipc_bcbearer_send;
913 	sprintf(bcbearer->media.name, "tipc-broadcast");
914 
915 	spin_lock_init(&bclink->lock);
916 	__skb_queue_head_init(&bcl->outqueue);
917 	__skb_queue_head_init(&bcl->deferred_queue);
918 	skb_queue_head_init(&bcl->wakeupq);
919 	bcl->next_out_no = 1;
920 	spin_lock_init(&bclink->node.lock);
921 	__skb_queue_head_init(&bclink->arrvq);
922 	skb_queue_head_init(&bclink->inputq);
923 	bcl->owner = &bclink->node;
924 	bcl->owner->net = net;
925 	bcl->max_pkt = MAX_PKT_DEFAULT_MCAST;
926 	tipc_link_set_queue_limits(bcl, BCLINK_WIN_DEFAULT);
927 	bcl->bearer_id = MAX_BEARERS;
928 	rcu_assign_pointer(tn->bearer_list[MAX_BEARERS], &bcbearer->bearer);
929 	bcl->state = WORKING_WORKING;
930 	bcl->pmsg = (struct tipc_msg *)&bcl->proto_msg;
931 	msg_set_prevnode(bcl->pmsg, tn->own_addr);
932 	strlcpy(bcl->name, tipc_bclink_name, TIPC_MAX_LINK_NAME);
933 	tn->bcbearer = bcbearer;
934 	tn->bclink = bclink;
935 	tn->bcl = bcl;
936 	return 0;
937 }
938 
939 void tipc_bclink_stop(struct net *net)
940 {
941 	struct tipc_net *tn = net_generic(net, tipc_net_id);
942 
943 	tipc_bclink_lock(net);
944 	tipc_link_purge_queues(tn->bcl);
945 	tipc_bclink_unlock(net);
946 
947 	RCU_INIT_POINTER(tn->bearer_list[BCBEARER], NULL);
948 	synchronize_net();
949 	kfree(tn->bcbearer);
950 	kfree(tn->bclink);
951 }
952 
953 /**
954  * tipc_nmap_add - add a node to a node map
955  */
956 static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node)
957 {
958 	int n = tipc_node(node);
959 	int w = n / WSIZE;
960 	u32 mask = (1 << (n % WSIZE));
961 
962 	if ((nm_ptr->map[w] & mask) == 0) {
963 		nm_ptr->count++;
964 		nm_ptr->map[w] |= mask;
965 	}
966 }
967 
968 /**
969  * tipc_nmap_remove - remove a node from a node map
970  */
971 static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node)
972 {
973 	int n = tipc_node(node);
974 	int w = n / WSIZE;
975 	u32 mask = (1 << (n % WSIZE));
976 
977 	if ((nm_ptr->map[w] & mask) != 0) {
978 		nm_ptr->map[w] &= ~mask;
979 		nm_ptr->count--;
980 	}
981 }
982 
983 /**
984  * tipc_nmap_diff - find differences between node maps
985  * @nm_a: input node map A
986  * @nm_b: input node map B
987  * @nm_diff: output node map A-B (i.e. nodes of A that are not in B)
988  */
989 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
990 			   struct tipc_node_map *nm_b,
991 			   struct tipc_node_map *nm_diff)
992 {
993 	int stop = ARRAY_SIZE(nm_a->map);
994 	int w;
995 	int b;
996 	u32 map;
997 
998 	memset(nm_diff, 0, sizeof(*nm_diff));
999 	for (w = 0; w < stop; w++) {
1000 		map = nm_a->map[w] ^ (nm_a->map[w] & nm_b->map[w]);
1001 		nm_diff->map[w] = map;
1002 		if (map != 0) {
1003 			for (b = 0 ; b < WSIZE; b++) {
1004 				if (map & (1 << b))
1005 					nm_diff->count++;
1006 			}
1007 		}
1008 	}
1009 }
1010