• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * net/tipc/group.c: TIPC group messaging code
3  *
4  * Copyright (c) 2017, Ericsson AB
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions are met:
9  *
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the names of the copyright holders nor the names of its
16  *    contributors may be used to endorse or promote products derived from
17  *    this software without specific prior written permission.
18  *
19  * Alternatively, this software may be distributed under the terms of the
20  * GNU General Public License ("GPL") version 2 as published by the Free
21  * Software Foundation.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
24  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
27  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
28  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
29  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
30  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
31  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
32  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
33  * POSSIBILITY OF SUCH DAMAGE.
34  */
35 
36 #include "core.h"
37 #include "addr.h"
38 #include "group.h"
39 #include "bcast.h"
40 #include "topsrv.h"
41 #include "msg.h"
42 #include "socket.h"
43 #include "node.h"
44 #include "name_table.h"
45 #include "subscr.h"
46 
47 #define ADV_UNIT (((MAX_MSG_SIZE + MAX_H_SIZE) / FLOWCTL_BLK_SZ) + 1)
48 #define ADV_IDLE ADV_UNIT
49 #define ADV_ACTIVE (ADV_UNIT * 12)
50 
51 enum mbr_state {
52 	MBR_JOINING,
53 	MBR_PUBLISHED,
54 	MBR_JOINED,
55 	MBR_PENDING,
56 	MBR_ACTIVE,
57 	MBR_RECLAIMING,
58 	MBR_REMITTED,
59 	MBR_LEAVING
60 };
61 
62 struct tipc_member {
63 	struct rb_node tree_node;
64 	struct list_head list;
65 	struct list_head small_win;
66 	struct sk_buff_head deferredq;
67 	struct tipc_group *group;
68 	u32 node;
69 	u32 port;
70 	u32 instance;
71 	enum mbr_state state;
72 	u16 advertised;
73 	u16 window;
74 	u16 bc_rcv_nxt;
75 	u16 bc_syncpt;
76 	u16 bc_acked;
77 };
78 
79 struct tipc_group {
80 	struct rb_root members;
81 	struct list_head small_win;
82 	struct list_head pending;
83 	struct list_head active;
84 	struct tipc_nlist dests;
85 	struct net *net;
86 	int subid;
87 	u32 type;
88 	u32 instance;
89 	u32 scope;
90 	u32 portid;
91 	u16 member_cnt;
92 	u16 active_cnt;
93 	u16 max_active;
94 	u16 bc_snd_nxt;
95 	u16 bc_ackers;
96 	bool *open;
97 	bool loopback;
98 	bool events;
99 };
100 
101 static void tipc_group_proto_xmit(struct tipc_group *grp, struct tipc_member *m,
102 				  int mtyp, struct sk_buff_head *xmitq);
103 
tipc_group_open(struct tipc_member * m,bool * wakeup)104 static void tipc_group_open(struct tipc_member *m, bool *wakeup)
105 {
106 	*wakeup = false;
107 	if (list_empty(&m->small_win))
108 		return;
109 	list_del_init(&m->small_win);
110 	*m->group->open = true;
111 	*wakeup = true;
112 }
113 
tipc_group_decr_active(struct tipc_group * grp,struct tipc_member * m)114 static void tipc_group_decr_active(struct tipc_group *grp,
115 				   struct tipc_member *m)
116 {
117 	if (m->state == MBR_ACTIVE || m->state == MBR_RECLAIMING ||
118 	    m->state == MBR_REMITTED)
119 		grp->active_cnt--;
120 }
121 
tipc_group_rcvbuf_limit(struct tipc_group * grp)122 static int tipc_group_rcvbuf_limit(struct tipc_group *grp)
123 {
124 	int max_active, active_pool, idle_pool;
125 	int mcnt = grp->member_cnt + 1;
126 
127 	/* Limit simultaneous reception from other members */
128 	max_active = min(mcnt / 8, 64);
129 	max_active = max(max_active, 16);
130 	grp->max_active = max_active;
131 
132 	/* Reserve blocks for active and idle members */
133 	active_pool = max_active * ADV_ACTIVE;
134 	idle_pool = (mcnt - max_active) * ADV_IDLE;
135 
136 	/* Scale to bytes, considering worst-case truesize/msgsize ratio */
137 	return (active_pool + idle_pool) * FLOWCTL_BLK_SZ * 4;
138 }
139 
tipc_group_bc_snd_nxt(struct tipc_group * grp)140 u16 tipc_group_bc_snd_nxt(struct tipc_group *grp)
141 {
142 	return grp->bc_snd_nxt;
143 }
144 
tipc_group_is_receiver(struct tipc_member * m)145 static bool tipc_group_is_receiver(struct tipc_member *m)
146 {
147 	return m && m->state != MBR_JOINING && m->state != MBR_LEAVING;
148 }
149 
tipc_group_is_sender(struct tipc_member * m)150 static bool tipc_group_is_sender(struct tipc_member *m)
151 {
152 	return m && m->state != MBR_JOINING && m->state != MBR_PUBLISHED;
153 }
154 
tipc_group_exclude(struct tipc_group * grp)155 u32 tipc_group_exclude(struct tipc_group *grp)
156 {
157 	if (!grp->loopback)
158 		return grp->portid;
159 	return 0;
160 }
161 
tipc_group_create(struct net * net,u32 portid,struct tipc_group_req * mreq,bool * group_is_open)162 struct tipc_group *tipc_group_create(struct net *net, u32 portid,
163 				     struct tipc_group_req *mreq,
164 				     bool *group_is_open)
165 {
166 	u32 filter = TIPC_SUB_PORTS | TIPC_SUB_NO_STATUS;
167 	bool global = mreq->scope != TIPC_NODE_SCOPE;
168 	struct tipc_group *grp;
169 	u32 type = mreq->type;
170 
171 	grp = kzalloc(sizeof(*grp), GFP_ATOMIC);
172 	if (!grp)
173 		return NULL;
174 	tipc_nlist_init(&grp->dests, tipc_own_addr(net));
175 	INIT_LIST_HEAD(&grp->small_win);
176 	INIT_LIST_HEAD(&grp->active);
177 	INIT_LIST_HEAD(&grp->pending);
178 	grp->members = RB_ROOT;
179 	grp->net = net;
180 	grp->portid = portid;
181 	grp->type = type;
182 	grp->instance = mreq->instance;
183 	grp->scope = mreq->scope;
184 	grp->loopback = mreq->flags & TIPC_GROUP_LOOPBACK;
185 	grp->events = mreq->flags & TIPC_GROUP_MEMBER_EVTS;
186 	grp->open = group_is_open;
187 	*grp->open = false;
188 	filter |= global ? TIPC_SUB_CLUSTER_SCOPE : TIPC_SUB_NODE_SCOPE;
189 	if (tipc_topsrv_kern_subscr(net, portid, type, 0, ~0,
190 				    filter, &grp->subid))
191 		return grp;
192 	kfree(grp);
193 	return NULL;
194 }
195 
tipc_group_join(struct net * net,struct tipc_group * grp,int * sk_rcvbuf)196 void tipc_group_join(struct net *net, struct tipc_group *grp, int *sk_rcvbuf)
197 {
198 	struct rb_root *tree = &grp->members;
199 	struct tipc_member *m, *tmp;
200 	struct sk_buff_head xmitq;
201 
202 	__skb_queue_head_init(&xmitq);
203 	rbtree_postorder_for_each_entry_safe(m, tmp, tree, tree_node) {
204 		tipc_group_proto_xmit(grp, m, GRP_JOIN_MSG, &xmitq);
205 		tipc_group_update_member(m, 0);
206 	}
207 	tipc_node_distr_xmit(net, &xmitq);
208 	*sk_rcvbuf = tipc_group_rcvbuf_limit(grp);
209 }
210 
tipc_group_delete(struct net * net,struct tipc_group * grp)211 void tipc_group_delete(struct net *net, struct tipc_group *grp)
212 {
213 	struct rb_root *tree = &grp->members;
214 	struct tipc_member *m, *tmp;
215 	struct sk_buff_head xmitq;
216 
217 	__skb_queue_head_init(&xmitq);
218 
219 	rbtree_postorder_for_each_entry_safe(m, tmp, tree, tree_node) {
220 		tipc_group_proto_xmit(grp, m, GRP_LEAVE_MSG, &xmitq);
221 		__skb_queue_purge(&m->deferredq);
222 		list_del(&m->list);
223 		kfree(m);
224 	}
225 	tipc_node_distr_xmit(net, &xmitq);
226 	tipc_nlist_purge(&grp->dests);
227 	tipc_topsrv_kern_unsubscr(net, grp->subid);
228 	kfree(grp);
229 }
230 
tipc_group_find_member(struct tipc_group * grp,u32 node,u32 port)231 static struct tipc_member *tipc_group_find_member(struct tipc_group *grp,
232 						  u32 node, u32 port)
233 {
234 	struct rb_node *n = grp->members.rb_node;
235 	u64 nkey, key = (u64)node << 32 | port;
236 	struct tipc_member *m;
237 
238 	while (n) {
239 		m = container_of(n, struct tipc_member, tree_node);
240 		nkey = (u64)m->node << 32 | m->port;
241 		if (key < nkey)
242 			n = n->rb_left;
243 		else if (key > nkey)
244 			n = n->rb_right;
245 		else
246 			return m;
247 	}
248 	return NULL;
249 }
250 
tipc_group_find_dest(struct tipc_group * grp,u32 node,u32 port)251 static struct tipc_member *tipc_group_find_dest(struct tipc_group *grp,
252 						u32 node, u32 port)
253 {
254 	struct tipc_member *m;
255 
256 	m = tipc_group_find_member(grp, node, port);
257 	if (m && tipc_group_is_receiver(m))
258 		return m;
259 	return NULL;
260 }
261 
tipc_group_find_node(struct tipc_group * grp,u32 node)262 static struct tipc_member *tipc_group_find_node(struct tipc_group *grp,
263 						u32 node)
264 {
265 	struct tipc_member *m;
266 	struct rb_node *n;
267 
268 	for (n = rb_first(&grp->members); n; n = rb_next(n)) {
269 		m = container_of(n, struct tipc_member, tree_node);
270 		if (m->node == node)
271 			return m;
272 	}
273 	return NULL;
274 }
275 
tipc_group_add_to_tree(struct tipc_group * grp,struct tipc_member * m)276 static int tipc_group_add_to_tree(struct tipc_group *grp,
277 				  struct tipc_member *m)
278 {
279 	u64 nkey, key = (u64)m->node << 32 | m->port;
280 	struct rb_node **n, *parent = NULL;
281 	struct tipc_member *tmp;
282 
283 	n = &grp->members.rb_node;
284 	while (*n) {
285 		tmp = container_of(*n, struct tipc_member, tree_node);
286 		parent = *n;
287 		tmp = container_of(parent, struct tipc_member, tree_node);
288 		nkey = (u64)tmp->node << 32 | tmp->port;
289 		if (key < nkey)
290 			n = &(*n)->rb_left;
291 		else if (key > nkey)
292 			n = &(*n)->rb_right;
293 		else
294 			return -EEXIST;
295 	}
296 	rb_link_node(&m->tree_node, parent, n);
297 	rb_insert_color(&m->tree_node, &grp->members);
298 	return 0;
299 }
300 
tipc_group_create_member(struct tipc_group * grp,u32 node,u32 port,u32 instance,int state)301 static struct tipc_member *tipc_group_create_member(struct tipc_group *grp,
302 						    u32 node, u32 port,
303 						    u32 instance, int state)
304 {
305 	struct tipc_member *m;
306 	int ret;
307 
308 	m = kzalloc(sizeof(*m), GFP_ATOMIC);
309 	if (!m)
310 		return NULL;
311 	INIT_LIST_HEAD(&m->list);
312 	INIT_LIST_HEAD(&m->small_win);
313 	__skb_queue_head_init(&m->deferredq);
314 	m->group = grp;
315 	m->node = node;
316 	m->port = port;
317 	m->instance = instance;
318 	m->bc_acked = grp->bc_snd_nxt - 1;
319 	ret = tipc_group_add_to_tree(grp, m);
320 	if (ret < 0) {
321 		kfree(m);
322 		return NULL;
323 	}
324 	grp->member_cnt++;
325 	tipc_nlist_add(&grp->dests, m->node);
326 	m->state = state;
327 	return m;
328 }
329 
tipc_group_add_member(struct tipc_group * grp,u32 node,u32 port,u32 instance)330 void tipc_group_add_member(struct tipc_group *grp, u32 node,
331 			   u32 port, u32 instance)
332 {
333 	tipc_group_create_member(grp, node, port, instance, MBR_PUBLISHED);
334 }
335 
tipc_group_delete_member(struct tipc_group * grp,struct tipc_member * m)336 static void tipc_group_delete_member(struct tipc_group *grp,
337 				     struct tipc_member *m)
338 {
339 	rb_erase(&m->tree_node, &grp->members);
340 	grp->member_cnt--;
341 
342 	/* Check if we were waiting for replicast ack from this member */
343 	if (grp->bc_ackers && less(m->bc_acked, grp->bc_snd_nxt - 1))
344 		grp->bc_ackers--;
345 
346 	list_del_init(&m->list);
347 	list_del_init(&m->small_win);
348 	tipc_group_decr_active(grp, m);
349 
350 	/* If last member on a node, remove node from dest list */
351 	if (!tipc_group_find_node(grp, m->node))
352 		tipc_nlist_del(&grp->dests, m->node);
353 
354 	kfree(m);
355 }
356 
tipc_group_dests(struct tipc_group * grp)357 struct tipc_nlist *tipc_group_dests(struct tipc_group *grp)
358 {
359 	return &grp->dests;
360 }
361 
tipc_group_self(struct tipc_group * grp,struct tipc_name_seq * seq,int * scope)362 void tipc_group_self(struct tipc_group *grp, struct tipc_name_seq *seq,
363 		     int *scope)
364 {
365 	seq->type = grp->type;
366 	seq->lower = grp->instance;
367 	seq->upper = grp->instance;
368 	*scope = grp->scope;
369 }
370 
tipc_group_update_member(struct tipc_member * m,int len)371 void tipc_group_update_member(struct tipc_member *m, int len)
372 {
373 	struct tipc_group *grp = m->group;
374 	struct tipc_member *_m, *tmp;
375 
376 	if (!tipc_group_is_receiver(m))
377 		return;
378 
379 	m->window -= len;
380 
381 	if (m->window >= ADV_IDLE)
382 		return;
383 
384 	list_del_init(&m->small_win);
385 
386 	/* Sort member into small_window members' list */
387 	list_for_each_entry_safe(_m, tmp, &grp->small_win, small_win) {
388 		if (_m->window > m->window)
389 			break;
390 	}
391 	list_add_tail(&m->small_win, &_m->small_win);
392 }
393 
tipc_group_update_bc_members(struct tipc_group * grp,int len,bool ack)394 void tipc_group_update_bc_members(struct tipc_group *grp, int len, bool ack)
395 {
396 	u16 prev = grp->bc_snd_nxt - 1;
397 	struct tipc_member *m;
398 	struct rb_node *n;
399 	u16 ackers = 0;
400 
401 	for (n = rb_first(&grp->members); n; n = rb_next(n)) {
402 		m = container_of(n, struct tipc_member, tree_node);
403 		if (tipc_group_is_receiver(m)) {
404 			tipc_group_update_member(m, len);
405 			m->bc_acked = prev;
406 			ackers++;
407 		}
408 	}
409 
410 	/* Mark number of acknowledges to expect, if any */
411 	if (ack)
412 		grp->bc_ackers = ackers;
413 	grp->bc_snd_nxt++;
414 }
415 
tipc_group_cong(struct tipc_group * grp,u32 dnode,u32 dport,int len,struct tipc_member ** mbr)416 bool tipc_group_cong(struct tipc_group *grp, u32 dnode, u32 dport,
417 		     int len, struct tipc_member **mbr)
418 {
419 	struct sk_buff_head xmitq;
420 	struct tipc_member *m;
421 	int adv, state;
422 
423 	m = tipc_group_find_dest(grp, dnode, dport);
424 	if (!tipc_group_is_receiver(m)) {
425 		*mbr = NULL;
426 		return false;
427 	}
428 	*mbr = m;
429 
430 	if (m->window >= len)
431 		return false;
432 
433 	*grp->open = false;
434 
435 	/* If not fully advertised, do it now to prevent mutual blocking */
436 	adv = m->advertised;
437 	state = m->state;
438 	if (state == MBR_JOINED && adv == ADV_IDLE)
439 		return true;
440 	if (state == MBR_ACTIVE && adv == ADV_ACTIVE)
441 		return true;
442 	if (state == MBR_PENDING && adv == ADV_IDLE)
443 		return true;
444 	__skb_queue_head_init(&xmitq);
445 	tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, &xmitq);
446 	tipc_node_distr_xmit(grp->net, &xmitq);
447 	return true;
448 }
449 
tipc_group_bc_cong(struct tipc_group * grp,int len)450 bool tipc_group_bc_cong(struct tipc_group *grp, int len)
451 {
452 	struct tipc_member *m = NULL;
453 
454 	/* If prev bcast was replicast, reject until all receivers have acked */
455 	if (grp->bc_ackers) {
456 		*grp->open = false;
457 		return true;
458 	}
459 	if (list_empty(&grp->small_win))
460 		return false;
461 
462 	m = list_first_entry(&grp->small_win, struct tipc_member, small_win);
463 	if (m->window >= len)
464 		return false;
465 
466 	return tipc_group_cong(grp, m->node, m->port, len, &m);
467 }
468 
469 /* tipc_group_sort_msg() - sort msg into queue by bcast sequence number
470  */
tipc_group_sort_msg(struct sk_buff * skb,struct sk_buff_head * defq)471 static void tipc_group_sort_msg(struct sk_buff *skb, struct sk_buff_head *defq)
472 {
473 	struct tipc_msg *_hdr, *hdr = buf_msg(skb);
474 	u16 bc_seqno = msg_grp_bc_seqno(hdr);
475 	struct sk_buff *_skb, *tmp;
476 	int mtyp = msg_type(hdr);
477 
478 	/* Bcast/mcast may be bypassed by ucast or other bcast, - sort it in */
479 	if (mtyp == TIPC_GRP_BCAST_MSG || mtyp == TIPC_GRP_MCAST_MSG) {
480 		skb_queue_walk_safe(defq, _skb, tmp) {
481 			_hdr = buf_msg(_skb);
482 			if (!less(bc_seqno, msg_grp_bc_seqno(_hdr)))
483 				continue;
484 			__skb_queue_before(defq, _skb, skb);
485 			return;
486 		}
487 		/* Bcast was not bypassed, - add to tail */
488 	}
489 	/* Unicasts are never bypassed, - always add to tail */
490 	__skb_queue_tail(defq, skb);
491 }
492 
493 /* tipc_group_filter_msg() - determine if we should accept arriving message
494  */
tipc_group_filter_msg(struct tipc_group * grp,struct sk_buff_head * inputq,struct sk_buff_head * xmitq)495 void tipc_group_filter_msg(struct tipc_group *grp, struct sk_buff_head *inputq,
496 			   struct sk_buff_head *xmitq)
497 {
498 	struct sk_buff *skb = __skb_dequeue(inputq);
499 	bool ack, deliver, update, leave = false;
500 	struct sk_buff_head *defq;
501 	struct tipc_member *m;
502 	struct tipc_msg *hdr;
503 	u32 node, port;
504 	int mtyp, blks;
505 
506 	if (!skb)
507 		return;
508 
509 	hdr = buf_msg(skb);
510 	node =  msg_orignode(hdr);
511 	port = msg_origport(hdr);
512 
513 	if (!msg_in_group(hdr))
514 		goto drop;
515 
516 	m = tipc_group_find_member(grp, node, port);
517 	if (!tipc_group_is_sender(m))
518 		goto drop;
519 
520 	if (less(msg_grp_bc_seqno(hdr), m->bc_rcv_nxt))
521 		goto drop;
522 
523 	TIPC_SKB_CB(skb)->orig_member = m->instance;
524 	defq = &m->deferredq;
525 	tipc_group_sort_msg(skb, defq);
526 
527 	while ((skb = skb_peek(defq))) {
528 		hdr = buf_msg(skb);
529 		mtyp = msg_type(hdr);
530 		blks = msg_blocks(hdr);
531 		deliver = true;
532 		ack = false;
533 		update = false;
534 
535 		if (more(msg_grp_bc_seqno(hdr), m->bc_rcv_nxt))
536 			break;
537 
538 		/* Decide what to do with message */
539 		switch (mtyp) {
540 		case TIPC_GRP_MCAST_MSG:
541 			if (msg_nameinst(hdr) != grp->instance) {
542 				update = true;
543 				deliver = false;
544 			}
545 			fallthrough;
546 		case TIPC_GRP_BCAST_MSG:
547 			m->bc_rcv_nxt++;
548 			ack = msg_grp_bc_ack_req(hdr);
549 			break;
550 		case TIPC_GRP_UCAST_MSG:
551 			break;
552 		case TIPC_GRP_MEMBER_EVT:
553 			if (m->state == MBR_LEAVING)
554 				leave = true;
555 			if (!grp->events)
556 				deliver = false;
557 			break;
558 		default:
559 			break;
560 		}
561 
562 		/* Execute decisions */
563 		__skb_dequeue(defq);
564 		if (deliver)
565 			__skb_queue_tail(inputq, skb);
566 		else
567 			kfree_skb(skb);
568 
569 		if (ack)
570 			tipc_group_proto_xmit(grp, m, GRP_ACK_MSG, xmitq);
571 
572 		if (leave) {
573 			__skb_queue_purge(defq);
574 			tipc_group_delete_member(grp, m);
575 			break;
576 		}
577 		if (!update)
578 			continue;
579 
580 		tipc_group_update_rcv_win(grp, blks, node, port, xmitq);
581 	}
582 	return;
583 drop:
584 	kfree_skb(skb);
585 }
586 
tipc_group_update_rcv_win(struct tipc_group * grp,int blks,u32 node,u32 port,struct sk_buff_head * xmitq)587 void tipc_group_update_rcv_win(struct tipc_group *grp, int blks, u32 node,
588 			       u32 port, struct sk_buff_head *xmitq)
589 {
590 	struct list_head *active = &grp->active;
591 	int max_active = grp->max_active;
592 	int reclaim_limit = max_active * 3 / 4;
593 	int active_cnt = grp->active_cnt;
594 	struct tipc_member *m, *rm, *pm;
595 
596 	m = tipc_group_find_member(grp, node, port);
597 	if (!m)
598 		return;
599 
600 	m->advertised -= blks;
601 
602 	switch (m->state) {
603 	case MBR_JOINED:
604 		/* First, decide if member can go active */
605 		if (active_cnt <= max_active) {
606 			m->state = MBR_ACTIVE;
607 			list_add_tail(&m->list, active);
608 			grp->active_cnt++;
609 			tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq);
610 		} else {
611 			m->state = MBR_PENDING;
612 			list_add_tail(&m->list, &grp->pending);
613 		}
614 
615 		if (active_cnt < reclaim_limit)
616 			break;
617 
618 		/* Reclaim from oldest active member, if possible */
619 		if (!list_empty(active)) {
620 			rm = list_first_entry(active, struct tipc_member, list);
621 			rm->state = MBR_RECLAIMING;
622 			list_del_init(&rm->list);
623 			tipc_group_proto_xmit(grp, rm, GRP_RECLAIM_MSG, xmitq);
624 			break;
625 		}
626 		/* Nobody to reclaim from; - revert oldest pending to JOINED */
627 		pm = list_first_entry(&grp->pending, struct tipc_member, list);
628 		list_del_init(&pm->list);
629 		pm->state = MBR_JOINED;
630 		tipc_group_proto_xmit(grp, pm, GRP_ADV_MSG, xmitq);
631 		break;
632 	case MBR_ACTIVE:
633 		if (!list_is_last(&m->list, &grp->active))
634 			list_move_tail(&m->list, &grp->active);
635 		if (m->advertised > (ADV_ACTIVE * 3 / 4))
636 			break;
637 		tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq);
638 		break;
639 	case MBR_REMITTED:
640 		if (m->advertised > ADV_IDLE)
641 			break;
642 		m->state = MBR_JOINED;
643 		grp->active_cnt--;
644 		if (m->advertised < ADV_IDLE) {
645 			pr_warn_ratelimited("Rcv unexpected msg after REMIT\n");
646 			tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq);
647 		}
648 
649 		if (list_empty(&grp->pending))
650 			return;
651 
652 		/* Set oldest pending member to active and advertise */
653 		pm = list_first_entry(&grp->pending, struct tipc_member, list);
654 		pm->state = MBR_ACTIVE;
655 		list_move_tail(&pm->list, &grp->active);
656 		grp->active_cnt++;
657 		tipc_group_proto_xmit(grp, pm, GRP_ADV_MSG, xmitq);
658 		break;
659 	case MBR_RECLAIMING:
660 	case MBR_JOINING:
661 	case MBR_LEAVING:
662 	default:
663 		break;
664 	}
665 }
666 
tipc_group_create_event(struct tipc_group * grp,struct tipc_member * m,u32 event,u16 seqno,struct sk_buff_head * inputq)667 static void tipc_group_create_event(struct tipc_group *grp,
668 				    struct tipc_member *m,
669 				    u32 event, u16 seqno,
670 				    struct sk_buff_head *inputq)
671 {	u32 dnode = tipc_own_addr(grp->net);
672 	struct tipc_event evt;
673 	struct sk_buff *skb;
674 	struct tipc_msg *hdr;
675 
676 	memset(&evt, 0, sizeof(evt));
677 	evt.event = event;
678 	evt.found_lower = m->instance;
679 	evt.found_upper = m->instance;
680 	evt.port.ref = m->port;
681 	evt.port.node = m->node;
682 	evt.s.seq.type = grp->type;
683 	evt.s.seq.lower = m->instance;
684 	evt.s.seq.upper = m->instance;
685 
686 	skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, TIPC_GRP_MEMBER_EVT,
687 			      GROUP_H_SIZE, sizeof(evt), dnode, m->node,
688 			      grp->portid, m->port, 0);
689 	if (!skb)
690 		return;
691 
692 	hdr = buf_msg(skb);
693 	msg_set_nametype(hdr, grp->type);
694 	msg_set_grp_evt(hdr, event);
695 	msg_set_dest_droppable(hdr, true);
696 	msg_set_grp_bc_seqno(hdr, seqno);
697 	memcpy(msg_data(hdr), &evt, sizeof(evt));
698 	TIPC_SKB_CB(skb)->orig_member = m->instance;
699 	__skb_queue_tail(inputq, skb);
700 }
701 
tipc_group_proto_xmit(struct tipc_group * grp,struct tipc_member * m,int mtyp,struct sk_buff_head * xmitq)702 static void tipc_group_proto_xmit(struct tipc_group *grp, struct tipc_member *m,
703 				  int mtyp, struct sk_buff_head *xmitq)
704 {
705 	struct tipc_msg *hdr;
706 	struct sk_buff *skb;
707 	int adv = 0;
708 
709 	skb = tipc_msg_create(GROUP_PROTOCOL, mtyp, INT_H_SIZE, 0,
710 			      m->node, tipc_own_addr(grp->net),
711 			      m->port, grp->portid, 0);
712 	if (!skb)
713 		return;
714 
715 	if (m->state == MBR_ACTIVE)
716 		adv = ADV_ACTIVE - m->advertised;
717 	else if (m->state == MBR_JOINED || m->state == MBR_PENDING)
718 		adv = ADV_IDLE - m->advertised;
719 
720 	hdr = buf_msg(skb);
721 
722 	if (mtyp == GRP_JOIN_MSG) {
723 		msg_set_grp_bc_syncpt(hdr, grp->bc_snd_nxt);
724 		msg_set_adv_win(hdr, adv);
725 		m->advertised += adv;
726 	} else if (mtyp == GRP_LEAVE_MSG) {
727 		msg_set_grp_bc_syncpt(hdr, grp->bc_snd_nxt);
728 	} else if (mtyp == GRP_ADV_MSG) {
729 		msg_set_adv_win(hdr, adv);
730 		m->advertised += adv;
731 	} else if (mtyp == GRP_ACK_MSG) {
732 		msg_set_grp_bc_acked(hdr, m->bc_rcv_nxt);
733 	} else if (mtyp == GRP_REMIT_MSG) {
734 		msg_set_grp_remitted(hdr, m->window);
735 	}
736 	msg_set_dest_droppable(hdr, true);
737 	__skb_queue_tail(xmitq, skb);
738 }
739 
tipc_group_proto_rcv(struct tipc_group * grp,bool * usr_wakeup,struct tipc_msg * hdr,struct sk_buff_head * inputq,struct sk_buff_head * xmitq)740 void tipc_group_proto_rcv(struct tipc_group *grp, bool *usr_wakeup,
741 			  struct tipc_msg *hdr, struct sk_buff_head *inputq,
742 			  struct sk_buff_head *xmitq)
743 {
744 	u32 node = msg_orignode(hdr);
745 	u32 port = msg_origport(hdr);
746 	struct tipc_member *m, *pm;
747 	u16 remitted, in_flight;
748 
749 	if (!grp)
750 		return;
751 
752 	if (grp->scope == TIPC_NODE_SCOPE && node != tipc_own_addr(grp->net))
753 		return;
754 
755 	m = tipc_group_find_member(grp, node, port);
756 
757 	switch (msg_type(hdr)) {
758 	case GRP_JOIN_MSG:
759 		if (!m)
760 			m = tipc_group_create_member(grp, node, port,
761 						     0, MBR_JOINING);
762 		if (!m)
763 			return;
764 		m->bc_syncpt = msg_grp_bc_syncpt(hdr);
765 		m->bc_rcv_nxt = m->bc_syncpt;
766 		m->window += msg_adv_win(hdr);
767 
768 		/* Wait until PUBLISH event is received if necessary */
769 		if (m->state != MBR_PUBLISHED)
770 			return;
771 
772 		/* Member can be taken into service */
773 		m->state = MBR_JOINED;
774 		tipc_group_open(m, usr_wakeup);
775 		tipc_group_update_member(m, 0);
776 		tipc_group_proto_xmit(grp, m, GRP_ADV_MSG, xmitq);
777 		tipc_group_create_event(grp, m, TIPC_PUBLISHED,
778 					m->bc_syncpt, inputq);
779 		return;
780 	case GRP_LEAVE_MSG:
781 		if (!m)
782 			return;
783 		m->bc_syncpt = msg_grp_bc_syncpt(hdr);
784 		list_del_init(&m->list);
785 		tipc_group_open(m, usr_wakeup);
786 		tipc_group_decr_active(grp, m);
787 		m->state = MBR_LEAVING;
788 		tipc_group_create_event(grp, m, TIPC_WITHDRAWN,
789 					m->bc_syncpt, inputq);
790 		return;
791 	case GRP_ADV_MSG:
792 		if (!m)
793 			return;
794 		m->window += msg_adv_win(hdr);
795 		tipc_group_open(m, usr_wakeup);
796 		return;
797 	case GRP_ACK_MSG:
798 		if (!m)
799 			return;
800 		m->bc_acked = msg_grp_bc_acked(hdr);
801 		if (--grp->bc_ackers)
802 			return;
803 		list_del_init(&m->small_win);
804 		*m->group->open = true;
805 		*usr_wakeup = true;
806 		tipc_group_update_member(m, 0);
807 		return;
808 	case GRP_RECLAIM_MSG:
809 		if (!m)
810 			return;
811 		tipc_group_proto_xmit(grp, m, GRP_REMIT_MSG, xmitq);
812 		m->window = ADV_IDLE;
813 		tipc_group_open(m, usr_wakeup);
814 		return;
815 	case GRP_REMIT_MSG:
816 		if (!m || m->state != MBR_RECLAIMING)
817 			return;
818 
819 		remitted = msg_grp_remitted(hdr);
820 
821 		/* Messages preceding the REMIT still in receive queue */
822 		if (m->advertised > remitted) {
823 			m->state = MBR_REMITTED;
824 			in_flight = m->advertised - remitted;
825 			m->advertised = ADV_IDLE + in_flight;
826 			return;
827 		}
828 		/* This should never happen */
829 		if (m->advertised < remitted)
830 			pr_warn_ratelimited("Unexpected REMIT msg\n");
831 
832 		/* All messages preceding the REMIT have been read */
833 		m->state = MBR_JOINED;
834 		grp->active_cnt--;
835 		m->advertised = ADV_IDLE;
836 
837 		/* Set oldest pending member to active and advertise */
838 		if (list_empty(&grp->pending))
839 			return;
840 		pm = list_first_entry(&grp->pending, struct tipc_member, list);
841 		pm->state = MBR_ACTIVE;
842 		list_move_tail(&pm->list, &grp->active);
843 		grp->active_cnt++;
844 		if (pm->advertised <= (ADV_ACTIVE * 3 / 4))
845 			tipc_group_proto_xmit(grp, pm, GRP_ADV_MSG, xmitq);
846 		return;
847 	default:
848 		pr_warn("Received unknown GROUP_PROTO message\n");
849 	}
850 }
851 
852 /* tipc_group_member_evt() - receive and handle a member up/down event
853  */
tipc_group_member_evt(struct tipc_group * grp,bool * usr_wakeup,int * sk_rcvbuf,struct tipc_msg * hdr,struct sk_buff_head * inputq,struct sk_buff_head * xmitq)854 void tipc_group_member_evt(struct tipc_group *grp,
855 			   bool *usr_wakeup,
856 			   int *sk_rcvbuf,
857 			   struct tipc_msg *hdr,
858 			   struct sk_buff_head *inputq,
859 			   struct sk_buff_head *xmitq)
860 {
861 	struct tipc_event *evt = (void *)msg_data(hdr);
862 	u32 instance = evt->found_lower;
863 	u32 node = evt->port.node;
864 	u32 port = evt->port.ref;
865 	int event = evt->event;
866 	struct tipc_member *m;
867 	struct net *net;
868 	u32 self;
869 
870 	if (!grp)
871 		return;
872 
873 	net = grp->net;
874 	self = tipc_own_addr(net);
875 	if (!grp->loopback && node == self && port == grp->portid)
876 		return;
877 
878 	m = tipc_group_find_member(grp, node, port);
879 
880 	switch (event) {
881 	case TIPC_PUBLISHED:
882 		/* Send and wait for arrival of JOIN message if necessary */
883 		if (!m) {
884 			m = tipc_group_create_member(grp, node, port, instance,
885 						     MBR_PUBLISHED);
886 			if (!m)
887 				break;
888 			tipc_group_update_member(m, 0);
889 			tipc_group_proto_xmit(grp, m, GRP_JOIN_MSG, xmitq);
890 			break;
891 		}
892 
893 		if (m->state != MBR_JOINING)
894 			break;
895 
896 		/* Member can be taken into service */
897 		m->instance = instance;
898 		m->state = MBR_JOINED;
899 		tipc_group_open(m, usr_wakeup);
900 		tipc_group_update_member(m, 0);
901 		tipc_group_proto_xmit(grp, m, GRP_JOIN_MSG, xmitq);
902 		tipc_group_create_event(grp, m, TIPC_PUBLISHED,
903 					m->bc_syncpt, inputq);
904 		break;
905 	case TIPC_WITHDRAWN:
906 		if (!m)
907 			break;
908 
909 		tipc_group_decr_active(grp, m);
910 		m->state = MBR_LEAVING;
911 		list_del_init(&m->list);
912 		tipc_group_open(m, usr_wakeup);
913 
914 		/* Only send event if no LEAVE message can be expected */
915 		if (!tipc_node_is_up(net, node))
916 			tipc_group_create_event(grp, m, TIPC_WITHDRAWN,
917 						m->bc_rcv_nxt, inputq);
918 		break;
919 	default:
920 		break;
921 	}
922 	*sk_rcvbuf = tipc_group_rcvbuf_limit(grp);
923 }
924 
tipc_group_fill_sock_diag(struct tipc_group * grp,struct sk_buff * skb)925 int tipc_group_fill_sock_diag(struct tipc_group *grp, struct sk_buff *skb)
926 {
927 	struct nlattr *group = nla_nest_start_noflag(skb, TIPC_NLA_SOCK_GROUP);
928 
929 	if (!group)
930 		return -EMSGSIZE;
931 
932 	if (nla_put_u32(skb, TIPC_NLA_SOCK_GROUP_ID,
933 			grp->type) ||
934 	    nla_put_u32(skb, TIPC_NLA_SOCK_GROUP_INSTANCE,
935 			grp->instance) ||
936 	    nla_put_u32(skb, TIPC_NLA_SOCK_GROUP_BC_SEND_NEXT,
937 			grp->bc_snd_nxt))
938 		goto group_msg_cancel;
939 
940 	if (grp->scope == TIPC_NODE_SCOPE)
941 		if (nla_put_flag(skb, TIPC_NLA_SOCK_GROUP_NODE_SCOPE))
942 			goto group_msg_cancel;
943 
944 	if (grp->scope == TIPC_CLUSTER_SCOPE)
945 		if (nla_put_flag(skb, TIPC_NLA_SOCK_GROUP_CLUSTER_SCOPE))
946 			goto group_msg_cancel;
947 
948 	if (*grp->open)
949 		if (nla_put_flag(skb, TIPC_NLA_SOCK_GROUP_OPEN))
950 			goto group_msg_cancel;
951 
952 	nla_nest_end(skb, group);
953 	return 0;
954 
955 group_msg_cancel:
956 	nla_nest_cancel(skb, group);
957 	return -1;
958 }
959