link.c 51.4 KB
Newer Older
P
Per Liden 已提交
1 2
/*
 * net/tipc/link.c: TIPC link code
3
 *
4
 * Copyright (c) 1996-2007, 2012-2015, Ericsson AB
5
 * Copyright (c) 2004-2007, 2010-2013, Wind River Systems
P
Per Liden 已提交
6 7
 * All rights reserved.
 *
P
Per Liden 已提交
8
 * Redistribution and use in source and binary forms, with or without
P
Per Liden 已提交
9 10
 * modification, are permitted provided that the following conditions are met:
 *
P
Per Liden 已提交
11 12 13 14 15 16 17 18
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. Neither the names of the copyright holders nor the names of its
 *    contributors may be used to endorse or promote products derived from
 *    this software without specific prior written permission.
P
Per Liden 已提交
19
 *
P
Per Liden 已提交
20 21 22 23 24 25 26 27 28 29 30 31 32 33
 * Alternatively, this software may be distributed under the terms of the
 * GNU General Public License ("GPL") version 2 as published by the Free
 * Software Foundation.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
P
Per Liden 已提交
34 35 36 37
 * POSSIBILITY OF SUCH DAMAGE.
 */

#include "core.h"
38
#include "subscr.h"
P
Per Liden 已提交
39
#include "link.h"
40
#include "bcast.h"
41
#include "socket.h"
P
Per Liden 已提交
42 43
#include "name_distr.h"
#include "discover.h"
44
#include "netlink.h"
P
Per Liden 已提交
45

46 47
#include <linux/pkt_sched.h>

48 49 50
/*
 * Error message prefixes
 */
51
static const char *link_co_err = "Link tunneling error, ";
52
static const char *link_rst_msg = "Resetting link ";
P
Per Liden 已提交
53

54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69
static const struct nla_policy tipc_nl_link_policy[TIPC_NLA_LINK_MAX + 1] = {
	[TIPC_NLA_LINK_UNSPEC]		= { .type = NLA_UNSPEC },
	[TIPC_NLA_LINK_NAME] = {
		.type = NLA_STRING,
		.len = TIPC_MAX_LINK_NAME
	},
	[TIPC_NLA_LINK_MTU]		= { .type = NLA_U32 },
	[TIPC_NLA_LINK_BROADCAST]	= { .type = NLA_FLAG },
	[TIPC_NLA_LINK_UP]		= { .type = NLA_FLAG },
	[TIPC_NLA_LINK_ACTIVE]		= { .type = NLA_FLAG },
	[TIPC_NLA_LINK_PROP]		= { .type = NLA_NESTED },
	[TIPC_NLA_LINK_STATS]		= { .type = NLA_NESTED },
	[TIPC_NLA_LINK_RX]		= { .type = NLA_U32 },
	[TIPC_NLA_LINK_TX]		= { .type = NLA_U32 }
};

70 71 72 73 74 75 76 77
/* Properties valid for media, bearar and link */
static const struct nla_policy tipc_nl_prop_policy[TIPC_NLA_PROP_MAX + 1] = {
	[TIPC_NLA_PROP_UNSPEC]		= { .type = NLA_UNSPEC },
	[TIPC_NLA_PROP_PRIO]		= { .type = NLA_U32 },
	[TIPC_NLA_PROP_TOL]		= { .type = NLA_U32 },
	[TIPC_NLA_PROP_WIN]		= { .type = NLA_U32 }
};

78 79 80 81
/*
 * Interval between NACKs when packets arrive out of order
 */
#define TIPC_NACK_INTV (TIPC_MIN_LINK_WIN * 2)
82 83 84
/*
 * Out-of-range value for link session numbers
 */
85
#define WILDCARD_SESSION 0x10000
86

87
/* Link FSM states:
P
Per Liden 已提交
88
 */
89
enum {
90 91 92 93 94 95 96
	LINK_ESTABLISHED     = 0xe,
	LINK_ESTABLISHING    = 0xe  << 4,
	LINK_RESET           = 0x1  << 8,
	LINK_RESETTING       = 0x2  << 12,
	LINK_PEER_RESET      = 0xd  << 16,
	LINK_FAILINGOVER     = 0xf  << 20,
	LINK_SYNCHING        = 0xc  << 24
97 98 99 100
};

/* Link FSM state checking routines
 */
101
static int link_is_up(struct tipc_link *l)
102
{
103
	return l->state & (LINK_ESTABLISHED | LINK_SYNCHING);
104
}
P
Per Liden 已提交
105

106 107
static int tipc_link_proto_rcv(struct tipc_link *l, struct sk_buff *skb,
			       struct sk_buff_head *xmitq);
108 109 110
static void tipc_link_build_proto_msg(struct tipc_link *l, int mtyp, bool probe,
				      u16 rcvgap, int tolerance, int priority,
				      struct sk_buff_head *xmitq);
111 112
static void link_reset_statistics(struct tipc_link *l_ptr);
static void link_print(struct tipc_link *l_ptr, const char *str);
113
static void tipc_link_sync_rcv(struct tipc_node *n, struct sk_buff *buf);
114

P
Per Liden 已提交
115
/*
S
Sam Ravnborg 已提交
116
 *  Simple non-static link routines (i.e. referenced outside this file)
P
Per Liden 已提交
117
 */
118
bool tipc_link_is_up(struct tipc_link *l)
P
Per Liden 已提交
119
{
120 121 122
	return link_is_up(l);
}

123 124 125 126 127
bool tipc_link_peer_is_down(struct tipc_link *l)
{
	return l->state == LINK_PEER_RESET;
}

128 129 130 131 132
bool tipc_link_is_reset(struct tipc_link *l)
{
	return l->state & (LINK_RESET | LINK_FAILINGOVER | LINK_ESTABLISHING);
}

133 134 135 136 137
bool tipc_link_is_establishing(struct tipc_link *l)
{
	return l->state == LINK_ESTABLISHING;
}

138 139 140 141 142 143 144 145 146 147 148 149 150
bool tipc_link_is_synching(struct tipc_link *l)
{
	return l->state == LINK_SYNCHING;
}

bool tipc_link_is_failingover(struct tipc_link *l)
{
	return l->state == LINK_FAILINGOVER;
}

bool tipc_link_is_blocked(struct tipc_link *l)
{
	return l->state & (LINK_RESETTING | LINK_PEER_RESET | LINK_FAILINGOVER);
P
Per Liden 已提交
151 152
}

153
int tipc_link_is_active(struct tipc_link *l)
P
Per Liden 已提交
154
{
155 156 157
	struct tipc_node *n = l->owner;

	return (node_active_link(n, 0) == l) || (node_active_link(n, 1) == l);
P
Per Liden 已提交
158 159
}

J
Jon Paul Maloy 已提交
160 161 162 163 164
static u32 link_own_addr(struct tipc_link *l)
{
	return msg_prevnode(l->pmsg);
}

P
Per Liden 已提交
165
/**
166
 * tipc_link_create - create a new link
J
Jon Paul Maloy 已提交
167
 * @n: pointer to associated node
168 169 170 171 172 173 174 175
 * @if_name: associated interface name
 * @bearer_id: id (index) of associated bearer
 * @tolerance: link tolerance to be used by link
 * @net_plane: network plane (A,B,c..) this link belongs to
 * @mtu: mtu to be advertised by link
 * @priority: priority to be used by link
 * @window: send window to be used by link
 * @session: session to be used by link
J
Jon Paul Maloy 已提交
176
 * @ownnode: identity of own node
177
 * @peer: node id of peer node
J
Jon Paul Maloy 已提交
178 179 180 181
 * @maddr: media address to be used
 * @inputq: queue to put messages ready for delivery
 * @namedq: queue to put binding table update messages ready for delivery
 * @link: return value, pointer to put the created link
182
 *
J
Jon Paul Maloy 已提交
183
 * Returns true if link was created, otherwise false
P
Per Liden 已提交
184
 */
185 186 187 188
bool tipc_link_create(struct tipc_node *n, char *if_name, int bearer_id,
		      int tolerance, char net_plane, u32 mtu, int priority,
		      int window, u32 session, u32 ownnode, u32 peer,
		      struct tipc_media_addr *maddr,
J
Jon Paul Maloy 已提交
189 190
		      struct sk_buff_head *inputq, struct sk_buff_head *namedq,
		      struct tipc_link **link)
P
Per Liden 已提交
191
{
J
Jon Paul Maloy 已提交
192 193
	struct tipc_link *l;
	struct tipc_msg *hdr;
194

J
Jon Paul Maloy 已提交
195 196 197 198
	l = kzalloc(sizeof(*l), GFP_ATOMIC);
	if (!l)
		return false;
	*link = l;
199 200 201 202 203 204
	l->pmsg = (struct tipc_msg *)&l->proto_msg;
	hdr = l->pmsg;
	tipc_msg_init(ownnode, hdr, LINK_PROTOCOL, RESET_MSG, INT_H_SIZE, peer);
	msg_set_size(hdr, sizeof(l->proto_msg));
	msg_set_session(hdr, session);
	msg_set_bearer_id(hdr, l->bearer_id);
205

J
Jon Paul Maloy 已提交
206 207 208 209
	/* Note: peer i/f name is completed by reset/activate message */
	sprintf(l->name, "%u.%u.%u:%s-%u.%u.%u:unknown",
		tipc_zone(ownnode), tipc_cluster(ownnode), tipc_node(ownnode),
		if_name, tipc_zone(peer), tipc_cluster(peer), tipc_node(peer));
210
	strcpy((char *)msg_data(hdr), if_name);
P
Per Liden 已提交
211

J
Jon Paul Maloy 已提交
212 213 214 215
	l->addr = peer;
	l->media_addr = maddr;
	l->owner = n;
	l->peer_session = WILDCARD_SESSION;
216 217 218 219 220 221 222
	l->bearer_id = bearer_id;
	l->tolerance = tolerance;
	l->net_plane = net_plane;
	l->advertised_mtu = mtu;
	l->mtu = mtu;
	l->priority = priority;
	tipc_link_set_queue_limits(l, window);
J
Jon Paul Maloy 已提交
223 224 225 226 227 228 229 230 231
	l->inputq = inputq;
	l->namedq = namedq;
	l->state = LINK_RESETTING;
	__skb_queue_head_init(&l->transmq);
	__skb_queue_head_init(&l->backlogq);
	__skb_queue_head_init(&l->deferdq);
	skb_queue_head_init(&l->wakeupq);
	skb_queue_head_init(l->inputq);
	return true;
P
Per Liden 已提交
232 233
}

234 235 236 237 238
/* tipc_link_build_bcast_sync_msg() - synchronize broadcast link endpoints.
 *
 * Give a newly added peer node the sequence number where it should
 * start receiving and acking broadcast packets.
 */
239 240
void tipc_link_build_bcast_sync_msg(struct tipc_link *l,
				    struct sk_buff_head *xmitq)
241 242 243
{
	struct sk_buff *skb;
	struct sk_buff_head list;
244
	u16 last_sent;
245 246 247 248 249

	skb = tipc_msg_create(BCAST_PROTOCOL, STATE_MSG, INT_H_SIZE,
			      0, l->addr, link_own_addr(l), 0, 0, 0);
	if (!skb)
		return;
250 251
	last_sent = tipc_bclink_get_last_sent(l->owner->net);
	msg_set_last_bcast(buf_msg(skb), last_sent);
252 253 254 255 256
	__skb_queue_head_init(&list);
	__skb_queue_tail(&list, skb);
	tipc_link_xmit(l, &list, xmitq);
}

257 258 259 260 261
/**
 * tipc_link_fsm_evt - link finite state machine
 * @l: pointer to link
 * @evt: state machine event to be processed
 */
262
int tipc_link_fsm_evt(struct tipc_link *l, int evt)
263
{
264
	int rc = 0;
265 266

	switch (l->state) {
267
	case LINK_RESETTING:
268
		switch (evt) {
269 270
		case LINK_PEER_RESET_EVT:
			l->state = LINK_PEER_RESET;
271
			break;
272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288
		case LINK_RESET_EVT:
			l->state = LINK_RESET;
			break;
		case LINK_FAILURE_EVT:
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_ESTABLISH_EVT:
		case LINK_FAILOVER_END_EVT:
		case LINK_SYNCH_BEGIN_EVT:
		case LINK_SYNCH_END_EVT:
		default:
			goto illegal_evt;
		}
		break;
	case LINK_RESET:
		switch (evt) {
		case LINK_PEER_RESET_EVT:
			l->state = LINK_ESTABLISHING;
289
			break;
290 291 292 293 294 295
		case LINK_FAILOVER_BEGIN_EVT:
			l->state = LINK_FAILINGOVER;
		case LINK_FAILURE_EVT:
		case LINK_RESET_EVT:
		case LINK_ESTABLISH_EVT:
		case LINK_FAILOVER_END_EVT:
296
			break;
297 298
		case LINK_SYNCH_BEGIN_EVT:
		case LINK_SYNCH_END_EVT:
299
		default:
300
			goto illegal_evt;
301 302
		}
		break;
303
	case LINK_PEER_RESET:
304
		switch (evt) {
305 306
		case LINK_RESET_EVT:
			l->state = LINK_ESTABLISHING;
307
			break;
308 309 310
		case LINK_PEER_RESET_EVT:
		case LINK_ESTABLISH_EVT:
		case LINK_FAILURE_EVT:
311
			break;
312 313 314 315 316 317 318 319 320 321 322 323
		case LINK_SYNCH_BEGIN_EVT:
		case LINK_SYNCH_END_EVT:
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_FAILOVER_END_EVT:
		default:
			goto illegal_evt;
		}
		break;
	case LINK_FAILINGOVER:
		switch (evt) {
		case LINK_FAILOVER_END_EVT:
			l->state = LINK_RESET;
324
			break;
325 326 327 328 329 330 331 332
		case LINK_PEER_RESET_EVT:
		case LINK_RESET_EVT:
		case LINK_ESTABLISH_EVT:
		case LINK_FAILURE_EVT:
			break;
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_SYNCH_BEGIN_EVT:
		case LINK_SYNCH_END_EVT:
333
		default:
334
			goto illegal_evt;
335 336
		}
		break;
337
	case LINK_ESTABLISHING:
338
		switch (evt) {
339 340
		case LINK_ESTABLISH_EVT:
			l->state = LINK_ESTABLISHED;
341
			break;
342 343 344 345
		case LINK_FAILOVER_BEGIN_EVT:
			l->state = LINK_FAILINGOVER;
			break;
		case LINK_RESET_EVT:
346 347
			l->state = LINK_RESET;
			break;
348
		case LINK_FAILURE_EVT:
349
		case LINK_PEER_RESET_EVT:
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366
		case LINK_SYNCH_BEGIN_EVT:
		case LINK_FAILOVER_END_EVT:
			break;
		case LINK_SYNCH_END_EVT:
		default:
			goto illegal_evt;
		}
		break;
	case LINK_ESTABLISHED:
		switch (evt) {
		case LINK_PEER_RESET_EVT:
			l->state = LINK_PEER_RESET;
			rc |= TIPC_LINK_DOWN_EVT;
			break;
		case LINK_FAILURE_EVT:
			l->state = LINK_RESETTING;
			rc |= TIPC_LINK_DOWN_EVT;
367
			break;
368 369
		case LINK_RESET_EVT:
			l->state = LINK_RESET;
370
			break;
371
		case LINK_ESTABLISH_EVT:
372
		case LINK_SYNCH_END_EVT:
373
			break;
374 375 376 377 378
		case LINK_SYNCH_BEGIN_EVT:
			l->state = LINK_SYNCHING;
			break;
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_FAILOVER_END_EVT:
379
		default:
380
			goto illegal_evt;
381 382
		}
		break;
383
	case LINK_SYNCHING:
384
		switch (evt) {
385 386 387 388 389 390 391
		case LINK_PEER_RESET_EVT:
			l->state = LINK_PEER_RESET;
			rc |= TIPC_LINK_DOWN_EVT;
			break;
		case LINK_FAILURE_EVT:
			l->state = LINK_RESETTING;
			rc |= TIPC_LINK_DOWN_EVT;
392
			break;
393 394
		case LINK_RESET_EVT:
			l->state = LINK_RESET;
395
			break;
396 397
		case LINK_ESTABLISH_EVT:
		case LINK_SYNCH_BEGIN_EVT:
398
			break;
399 400 401 402 403
		case LINK_SYNCH_END_EVT:
			l->state = LINK_ESTABLISHED;
			break;
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_FAILOVER_END_EVT:
404
		default:
405
			goto illegal_evt;
406 407 408
		}
		break;
	default:
409
		pr_err("Unknown FSM state %x in %s\n", l->state, l->name);
410
	}
411 412 413 414
	return rc;
illegal_evt:
	pr_err("Illegal FSM event %x in state %x on link %s\n",
	       evt, l->state, l->name);
415 416 417
	return rc;
}

418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
/* link_profile_stats - update statistical profiling of traffic
 */
static void link_profile_stats(struct tipc_link *l)
{
	struct sk_buff *skb;
	struct tipc_msg *msg;
	int length;

	/* Update counters used in statistical profiling of send traffic */
	l->stats.accu_queue_sz += skb_queue_len(&l->transmq);
	l->stats.queue_sz_counts++;

	skb = skb_peek(&l->transmq);
	if (!skb)
		return;
	msg = buf_msg(skb);
	length = msg_size(msg);

	if (msg_user(msg) == MSG_FRAGMENTER) {
		if (msg_type(msg) != FIRST_FRAGMENT)
			return;
		length = msg_size(msg_get_wrapped(msg));
	}
	l->stats.msg_lengths_total += length;
	l->stats.msg_length_counts++;
	if (length <= 64)
		l->stats.msg_length_profile[0]++;
	else if (length <= 256)
		l->stats.msg_length_profile[1]++;
	else if (length <= 1024)
		l->stats.msg_length_profile[2]++;
	else if (length <= 4096)
		l->stats.msg_length_profile[3]++;
	else if (length <= 16384)
		l->stats.msg_length_profile[4]++;
	else if (length <= 32768)
		l->stats.msg_length_profile[5]++;
	else
		l->stats.msg_length_profile[6]++;
}

/* tipc_link_timeout - perform periodic task as instructed from node timeout
 */
int tipc_link_timeout(struct tipc_link *l, struct sk_buff_head *xmitq)
{
	int rc = 0;
464 465 466
	int mtyp = STATE_MSG;
	bool xmit = false;
	bool prb = false;
467 468

	link_profile_stats(l);
469

470 471 472
	switch (l->state) {
	case LINK_ESTABLISHED:
	case LINK_SYNCHING:
473 474 475 476 477 478 479
		if (!l->silent_intv_cnt) {
			if (tipc_bclink_acks_missing(l->owner))
				xmit = true;
		} else if (l->silent_intv_cnt <= l->abort_limit) {
			xmit = true;
			prb = true;
		} else {
480
			rc |= tipc_link_fsm_evt(l, LINK_FAILURE_EVT);
481 482
		}
		l->silent_intv_cnt++;
483 484
		break;
	case LINK_RESET:
485 486
		xmit = true;
		mtyp = RESET_MSG;
487 488
		break;
	case LINK_ESTABLISHING:
489 490
		xmit = true;
		mtyp = ACTIVATE_MSG;
491 492
		break;
	case LINK_PEER_RESET:
493
	case LINK_RESETTING:
494 495 496 497
	case LINK_FAILINGOVER:
		break;
	default:
		break;
498
	}
499

500 501 502
	if (xmit)
		tipc_link_build_proto_msg(l, mtyp, prb, 0, 0, 0, xmitq);

503 504 505
	return rc;
}

P
Per Liden 已提交
506
/**
507
 * link_schedule_user - schedule a message sender for wakeup after congestion
508
 * @link: congested link
509
 * @list: message that was attempted sent
510
 * Create pseudo msg to send back to user when congestion abates
511
 * Does not consume buffer list
P
Per Liden 已提交
512
 */
513
static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list)
P
Per Liden 已提交
514
{
515 516 517 518 519 520 521 522 523
	struct tipc_msg *msg = buf_msg(skb_peek(list));
	int imp = msg_importance(msg);
	u32 oport = msg_origport(msg);
	u32 addr = link_own_addr(link);
	struct sk_buff *skb;

	/* This really cannot happen...  */
	if (unlikely(imp > TIPC_CRITICAL_IMPORTANCE)) {
		pr_warn("%s<%s>, send queue full", link_rst_msg, link->name);
524
		return -ENOBUFS;
525 526 527 528 529 530 531 532 533
	}
	/* Non-blocking sender: */
	if (TIPC_SKB_CB(skb_peek(list))->wakeup_pending)
		return -ELINKCONG;

	/* Create and schedule wakeup pseudo message */
	skb = tipc_msg_create(SOCK_WAKEUP, 0, INT_H_SIZE, 0,
			      addr, addr, oport, 0, 0);
	if (!skb)
534
		return -ENOBUFS;
535 536 537
	TIPC_SKB_CB(skb)->chain_sz = skb_queue_len(list);
	TIPC_SKB_CB(skb)->chain_imp = imp;
	skb_queue_tail(&link->wakeupq, skb);
538
	link->stats.link_congs++;
539
	return -ELINKCONG;
P
Per Liden 已提交
540 541
}

542 543 544 545 546 547
/**
 * link_prepare_wakeup - prepare users for wakeup after congestion
 * @link: congested link
 * Move a number of waiting users, as permitted by available space in
 * the send queue, from link wait queue to node wait queue for wakeup
 */
548
void link_prepare_wakeup(struct tipc_link *l)
P
Per Liden 已提交
549
{
550 551
	int pnd[TIPC_SYSTEM_IMPORTANCE + 1] = {0,};
	int imp, lim;
552
	struct sk_buff *skb, *tmp;
553

554 555 556 557 558
	skb_queue_walk_safe(&l->wakeupq, skb, tmp) {
		imp = TIPC_SKB_CB(skb)->chain_imp;
		lim = l->window + l->backlog[imp].limit;
		pnd[imp] += TIPC_SKB_CB(skb)->chain_sz;
		if ((pnd[imp] + l->backlog[imp].len) >= lim)
P
Per Liden 已提交
559
			break;
560
		skb_unlink(skb, &l->wakeupq);
561
		skb_queue_tail(l->inputq, skb);
P
Per Liden 已提交
562 563 564 565
	}
}

/**
566
 * tipc_link_reset_fragments - purge link's inbound message fragments queue
P
Per Liden 已提交
567 568
 * @l_ptr: pointer to link
 */
569
void tipc_link_reset_fragments(struct tipc_link *l_ptr)
P
Per Liden 已提交
570
{
571 572
	kfree_skb(l_ptr->reasm_buf);
	l_ptr->reasm_buf = NULL;
P
Per Liden 已提交
573 574
}

575
void tipc_link_purge_backlog(struct tipc_link *l)
576 577 578 579 580 581 582 583 584
{
	__skb_queue_purge(&l->backlogq);
	l->backlog[TIPC_LOW_IMPORTANCE].len = 0;
	l->backlog[TIPC_MEDIUM_IMPORTANCE].len = 0;
	l->backlog[TIPC_HIGH_IMPORTANCE].len = 0;
	l->backlog[TIPC_CRITICAL_IMPORTANCE].len = 0;
	l->backlog[TIPC_SYSTEM_IMPORTANCE].len = 0;
}

585
/**
586
 * tipc_link_purge_queues - purge all pkt queues associated with link
P
Per Liden 已提交
587 588
 * @l_ptr: pointer to link
 */
589
void tipc_link_purge_queues(struct tipc_link *l_ptr)
P
Per Liden 已提交
590
{
J
Jon Paul Maloy 已提交
591 592
	__skb_queue_purge(&l_ptr->deferdq);
	__skb_queue_purge(&l_ptr->transmq);
593
	tipc_link_purge_backlog(l_ptr);
594
	tipc_link_reset_fragments(l_ptr);
P
Per Liden 已提交
595 596
}

597
void tipc_link_reset(struct tipc_link *l)
P
Per Liden 已提交
598
{
599
	/* Link is down, accept any session */
600
	l->peer_session = WILDCARD_SESSION;
P
Per Liden 已提交
601

602 603
	/* If peer is up, it only accepts an incremented session number */
	msg_set_session(l->pmsg, msg_session(l->pmsg) + 1);
P
Per Liden 已提交
604

605 606
	/* Prepare for renewed mtu size negotiation */
	l->mtu = l->advertised_mtu;
P
Per Liden 已提交
607

608
	/* Clean up all queues: */
609 610
	__skb_queue_purge(&l->transmq);
	__skb_queue_purge(&l->deferdq);
611
	skb_queue_splice_init(&l->wakeupq, l->inputq);
612 613 614 615 616 617 618 619 620 621 622 623 624

	tipc_link_purge_backlog(l);
	kfree_skb(l->reasm_buf);
	kfree_skb(l->failover_reasm_skb);
	l->reasm_buf = NULL;
	l->failover_reasm_skb = NULL;
	l->rcv_unacked = 0;
	l->snd_nxt = 1;
	l->rcv_nxt = 1;
	l->silent_intv_cnt = 0;
	l->stats.recv_info = 0;
	l->stale_count = 0;
	link_reset_statistics(l);
P
Per Liden 已提交
625 626
}

627
/**
628
 * __tipc_link_xmit(): same as tipc_link_xmit, but destlink is known & locked
629
 * @link: link to use
630 631
 * @list: chain of buffers containing message
 *
632
 * Consumes the buffer chain, except when returning an error code,
633 634
 * Returns 0 if success, or errno: -ELINKCONG, -EMSGSIZE or -ENOBUFS
 * Messages at TIPC_SYSTEM_IMPORTANCE are always accepted
635
 */
636 637
int __tipc_link_xmit(struct net *net, struct tipc_link *link,
		     struct sk_buff_head *list)
638
{
639
	struct tipc_msg *msg = buf_msg(skb_peek(list));
J
Jon Paul Maloy 已提交
640
	unsigned int maxwin = link->window;
641
	unsigned int i, imp = msg_importance(msg);
642
	uint mtu = link->mtu;
643 644
	u16 ack = mod(link->rcv_nxt - 1);
	u16 seqno = link->snd_nxt;
645
	u16 bc_last_in = link->owner->bclink.last_in;
J
Jon Paul Maloy 已提交
646
	struct tipc_media_addr *addr = link->media_addr;
J
Jon Paul Maloy 已提交
647 648
	struct sk_buff_head *transmq = &link->transmq;
	struct sk_buff_head *backlogq = &link->backlogq;
649
	struct sk_buff *skb, *bskb;
650

651 652 653 654 655
	/* Match msg importance against this and all higher backlog limits: */
	for (i = imp; i <= TIPC_SYSTEM_IMPORTANCE; i++) {
		if (unlikely(link->backlog[i].len >= link->backlog[i].limit))
			return link_schedule_user(link, list);
	}
656
	if (unlikely(msg_size(msg) > mtu))
657
		return -EMSGSIZE;
658

J
Jon Paul Maloy 已提交
659
	/* Prepare each packet for sending, and add to relevant queue: */
660 661
	while (skb_queue_len(list)) {
		skb = skb_peek(list);
662
		msg = buf_msg(skb);
J
Jon Paul Maloy 已提交
663 664
		msg_set_seqno(msg, seqno);
		msg_set_ack(msg, ack);
665 666
		msg_set_bcast_ack(msg, bc_last_in);

J
Jon Paul Maloy 已提交
667
		if (likely(skb_queue_len(transmq) < maxwin)) {
668
			__skb_dequeue(list);
J
Jon Paul Maloy 已提交
669 670 671 672 673 674
			__skb_queue_tail(transmq, skb);
			tipc_bearer_send(net, link->bearer_id, skb, addr);
			link->rcv_unacked = 0;
			seqno++;
			continue;
		}
675 676
		if (tipc_msg_bundle(skb_peek_tail(backlogq), msg, mtu)) {
			kfree_skb(__skb_dequeue(list));
677 678
			link->stats.sent_bundled++;
			continue;
J
Jon Paul Maloy 已提交
679
		}
680 681 682 683
		if (tipc_msg_make_bundle(&bskb, msg, mtu, link->addr)) {
			kfree_skb(__skb_dequeue(list));
			__skb_queue_tail(backlogq, bskb);
			link->backlog[msg_importance(buf_msg(bskb))].len++;
684 685
			link->stats.sent_bundled++;
			link->stats.sent_bundles++;
686
			continue;
687
		}
688 689
		link->backlog[imp].len += skb_queue_len(list);
		skb_queue_splice_tail_init(list, backlogq);
690
	}
691
	link->snd_nxt = seqno;
692 693 694
	return 0;
}

695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766
/**
 * tipc_link_xmit(): enqueue buffer list according to queue situation
 * @link: link to use
 * @list: chain of buffers containing message
 * @xmitq: returned list of packets to be sent by caller
 *
 * Consumes the buffer chain, except when returning -ELINKCONG,
 * since the caller then may want to make more send attempts.
 * Returns 0 if success, or errno: -ELINKCONG, -EMSGSIZE or -ENOBUFS
 * Messages at TIPC_SYSTEM_IMPORTANCE are always accepted
 */
int tipc_link_xmit(struct tipc_link *l, struct sk_buff_head *list,
		   struct sk_buff_head *xmitq)
{
	struct tipc_msg *hdr = buf_msg(skb_peek(list));
	unsigned int maxwin = l->window;
	unsigned int i, imp = msg_importance(hdr);
	unsigned int mtu = l->mtu;
	u16 ack = l->rcv_nxt - 1;
	u16 seqno = l->snd_nxt;
	u16 bc_last_in = l->owner->bclink.last_in;
	struct sk_buff_head *transmq = &l->transmq;
	struct sk_buff_head *backlogq = &l->backlogq;
	struct sk_buff *skb, *_skb, *bskb;

	/* Match msg importance against this and all higher backlog limits: */
	for (i = imp; i <= TIPC_SYSTEM_IMPORTANCE; i++) {
		if (unlikely(l->backlog[i].len >= l->backlog[i].limit))
			return link_schedule_user(l, list);
	}
	if (unlikely(msg_size(hdr) > mtu))
		return -EMSGSIZE;

	/* Prepare each packet for sending, and add to relevant queue: */
	while (skb_queue_len(list)) {
		skb = skb_peek(list);
		hdr = buf_msg(skb);
		msg_set_seqno(hdr, seqno);
		msg_set_ack(hdr, ack);
		msg_set_bcast_ack(hdr, bc_last_in);

		if (likely(skb_queue_len(transmq) < maxwin)) {
			_skb = skb_clone(skb, GFP_ATOMIC);
			if (!_skb)
				return -ENOBUFS;
			__skb_dequeue(list);
			__skb_queue_tail(transmq, skb);
			__skb_queue_tail(xmitq, _skb);
			l->rcv_unacked = 0;
			seqno++;
			continue;
		}
		if (tipc_msg_bundle(skb_peek_tail(backlogq), hdr, mtu)) {
			kfree_skb(__skb_dequeue(list));
			l->stats.sent_bundled++;
			continue;
		}
		if (tipc_msg_make_bundle(&bskb, hdr, mtu, l->addr)) {
			kfree_skb(__skb_dequeue(list));
			__skb_queue_tail(backlogq, bskb);
			l->backlog[msg_importance(buf_msg(bskb))].len++;
			l->stats.sent_bundled++;
			l->stats.sent_bundles++;
			continue;
		}
		l->backlog[imp].len += skb_queue_len(list);
		skb_queue_splice_tail_init(list, backlogq);
	}
	l->snd_nxt = seqno;
	return 0;
}

767
/*
768
 * tipc_link_sync_rcv - synchronize broadcast link endpoints.
769 770 771 772 773 774
 * Receive the sequence number where we should start receiving and
 * acking broadcast packets from a newly added peer node, and open
 * up for reception of such packets.
 *
 * Called with node locked
 */
775
static void tipc_link_sync_rcv(struct tipc_node *n, struct sk_buff *buf)
776 777 778 779 780 781 782 783
{
	struct tipc_msg *msg = buf_msg(buf);

	n->bclink.last_sent = n->bclink.last_in = msg_last_bcast(msg);
	n->bclink.recv_permitted = true;
	kfree_skb(buf);
}

784
/*
785 786 787 788 789 790
 * tipc_link_push_packets - push unsent packets to bearer
 *
 * Push out the unsent messages of a link where congestion
 * has abated. Node is locked.
 *
 * Called with node locked
P
Per Liden 已提交
791
 */
J
Jon Paul Maloy 已提交
792
void tipc_link_push_packets(struct tipc_link *link)
P
Per Liden 已提交
793
{
J
Jon Paul Maloy 已提交
794
	struct sk_buff *skb;
795
	struct tipc_msg *msg;
796
	u16 seqno = link->snd_nxt;
797
	u16 ack = mod(link->rcv_nxt - 1);
P
Per Liden 已提交
798

J
Jon Paul Maloy 已提交
799 800 801
	while (skb_queue_len(&link->transmq) < link->window) {
		skb = __skb_dequeue(&link->backlogq);
		if (!skb)
802
			break;
J
Jon Paul Maloy 已提交
803
		msg = buf_msg(skb);
804
		link->backlog[msg_importance(msg)].len--;
J
Jon Paul Maloy 已提交
805
		msg_set_ack(msg, ack);
806 807
		msg_set_seqno(msg, seqno);
		seqno = mod(seqno + 1);
J
Jon Paul Maloy 已提交
808 809 810 811
		msg_set_bcast_ack(msg, link->owner->bclink.last_in);
		link->rcv_unacked = 0;
		__skb_queue_tail(&link->transmq, skb);
		tipc_bearer_send(link->owner->net, link->bearer_id,
J
Jon Paul Maloy 已提交
812
				 skb, link->media_addr);
P
Per Liden 已提交
813
	}
814
	link->snd_nxt = seqno;
P
Per Liden 已提交
815 816
}

817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844
void tipc_link_advance_backlog(struct tipc_link *l, struct sk_buff_head *xmitq)
{
	struct sk_buff *skb, *_skb;
	struct tipc_msg *hdr;
	u16 seqno = l->snd_nxt;
	u16 ack = l->rcv_nxt - 1;

	while (skb_queue_len(&l->transmq) < l->window) {
		skb = skb_peek(&l->backlogq);
		if (!skb)
			break;
		_skb = skb_clone(skb, GFP_ATOMIC);
		if (!_skb)
			break;
		__skb_dequeue(&l->backlogq);
		hdr = buf_msg(skb);
		l->backlog[msg_importance(hdr)].len--;
		__skb_queue_tail(&l->transmq, skb);
		__skb_queue_tail(xmitq, _skb);
		msg_set_ack(hdr, ack);
		msg_set_seqno(hdr, seqno);
		msg_set_bcast_ack(hdr, l->owner->bclink.last_in);
		l->rcv_unacked = 0;
		seqno++;
	}
	l->snd_nxt = seqno;
}

845
static void link_retransmit_failure(struct tipc_link *l_ptr,
846
				    struct sk_buff *buf)
847 848
{
	struct tipc_msg *msg = buf_msg(buf);
849
	struct net *net = l_ptr->owner->net;
850

851
	pr_warn("Retransmission failure on link <%s>\n", l_ptr->name);
852 853 854

	if (l_ptr->addr) {
		/* Handle failure on standard link */
855 856 857 858 859 860
		link_print(l_ptr, "Resetting link ");
		pr_info("Failed msg: usr %u, typ %u, len %u, err %u\n",
			msg_user(msg), msg_type(msg), msg_size(msg),
			msg_errcode(msg));
		pr_info("sqno %u, prev: %x, src: %x\n",
			msg_seqno(msg), msg_prevnode(msg), msg_orignode(msg));
861 862
	} else {
		/* Handle failure on broadcast link */
863
		struct tipc_node *n_ptr;
864 865
		char addr_string[16];

866 867 868
		pr_info("Msg seq number: %u,  ", msg_seqno(msg));
		pr_cont("Outstanding acks: %lu\n",
			(unsigned long) TIPC_SKB_CB(buf)->handle);
J
Jeff Garzik 已提交
869

870
		n_ptr = tipc_bclink_retransmit_to(net);
871

872
		tipc_addr_string_fill(addr_string, n_ptr->addr);
873
		pr_info("Broadcast link info for %s\n", addr_string);
874 875
		pr_info("Reception permitted: %d,  Acked: %u\n",
			n_ptr->bclink.recv_permitted,
876 877 878 879 880
			n_ptr->bclink.acked);
		pr_info("Last in: %u,  Oos state: %u,  Last sent: %u\n",
			n_ptr->bclink.last_in,
			n_ptr->bclink.oos_state,
			n_ptr->bclink.last_sent);
881

882
		n_ptr->action_flags |= TIPC_BCAST_RESET;
883 884 885 886
		l_ptr->stale_count = 0;
	}
}

887
void tipc_link_retransmit(struct tipc_link *l_ptr, struct sk_buff *skb,
888
			  u32 retransmits)
P
Per Liden 已提交
889 890 891
{
	struct tipc_msg *msg;

892
	if (!skb)
893 894
		return;

895
	msg = buf_msg(skb);
896

897
	/* Detect repeated retransmit failures */
898
	if (l_ptr->last_retransm == msg_seqno(msg)) {
899
		if (++l_ptr->stale_count > 100) {
900
			link_retransmit_failure(l_ptr, skb);
901
			return;
902 903
		}
	} else {
904
		l_ptr->last_retransm = msg_seqno(msg);
905
		l_ptr->stale_count = 1;
P
Per Liden 已提交
906
	}
907

J
Jon Paul Maloy 已提交
908 909
	skb_queue_walk_from(&l_ptr->transmq, skb) {
		if (!retransmits)
910 911
			break;
		msg = buf_msg(skb);
912
		msg_set_ack(msg, mod(l_ptr->rcv_nxt - 1));
913
		msg_set_bcast_ack(msg, l_ptr->owner->bclink.last_in);
914
		tipc_bearer_send(l_ptr->owner->net, l_ptr->bearer_id, skb,
J
Jon Paul Maloy 已提交
915
				 l_ptr->media_addr);
916 917
		retransmits--;
		l_ptr->stats.retransmitted++;
P
Per Liden 已提交
918 919 920
	}
}

921 922 923 924 925 926 927 928 929 930 931 932 933 934 935
static int tipc_link_retransm(struct tipc_link *l, int retransm,
			      struct sk_buff_head *xmitq)
{
	struct sk_buff *_skb, *skb = skb_peek(&l->transmq);
	struct tipc_msg *hdr;

	if (!skb)
		return 0;

	/* Detect repeated retransmit failures on same packet */
	if (likely(l->last_retransm != buf_seqno(skb))) {
		l->last_retransm = buf_seqno(skb);
		l->stale_count = 1;
	} else if (++l->stale_count > 100) {
		link_retransmit_failure(l, skb);
936
		return tipc_link_fsm_evt(l, LINK_FAILURE_EVT);
937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955
	}
	skb_queue_walk(&l->transmq, skb) {
		if (!retransm)
			return 0;
		hdr = buf_msg(skb);
		_skb = __pskb_copy(skb, MIN_H_SIZE, GFP_ATOMIC);
		if (!_skb)
			return 0;
		hdr = buf_msg(_skb);
		msg_set_ack(hdr, l->rcv_nxt - 1);
		msg_set_bcast_ack(hdr, l->owner->bclink.last_in);
		_skb->priority = TC_PRIO_CONTROL;
		__skb_queue_tail(xmitq, _skb);
		retransm--;
		l->stats.retransmitted++;
	}
	return 0;
}

956
/* tipc_data_input - deliver data and name distr msgs to upper layer
957
 *
958
 * Consumes buffer if message is of right type
959 960
 * Node lock must be held
 */
961 962
static bool tipc_data_input(struct tipc_link *link, struct sk_buff *skb,
			    struct sk_buff_head *inputq)
963
{
964
	struct tipc_node *node = link->owner;
965

966
	switch (msg_user(buf_msg(skb))) {
967 968 969 970 971
	case TIPC_LOW_IMPORTANCE:
	case TIPC_MEDIUM_IMPORTANCE:
	case TIPC_HIGH_IMPORTANCE:
	case TIPC_CRITICAL_IMPORTANCE:
	case CONN_MANAGER:
972
		skb_queue_tail(inputq, skb);
973
		return true;
974
	case NAME_DISTRIBUTOR:
975
		node->bclink.recv_permitted = true;
976
		skb_queue_tail(link->namedq, skb);
977 978
		return true;
	case MSG_BUNDLER:
979
	case TUNNEL_PROTOCOL:
980
	case MSG_FRAGMENTER:
981
	case BCAST_PROTOCOL:
982
		return false;
983
	default:
984 985 986 987
		pr_warn("Dropping received illegal msg type\n");
		kfree_skb(skb);
		return false;
	};
988
}
989 990 991 992

/* tipc_link_input - process packet that has passed link protocol check
 *
 * Consumes buffer
993
 */
994 995
static int tipc_link_input(struct tipc_link *l, struct sk_buff *skb,
			   struct sk_buff_head *inputq)
996
{
997 998 999
	struct tipc_node *node = l->owner;
	struct tipc_msg *hdr = buf_msg(skb);
	struct sk_buff **reasm_skb = &l->reasm_buf;
1000
	struct sk_buff *iskb;
1001
	struct sk_buff_head tmpq;
1002
	int usr = msg_user(hdr);
1003
	int rc = 0;
1004 1005
	int pos = 0;
	int ipos = 0;
1006

1007 1008 1009 1010
	if (unlikely(usr == TUNNEL_PROTOCOL)) {
		if (msg_type(hdr) == SYNCH_MSG) {
			__skb_queue_purge(&l->deferdq);
			goto drop;
1011
		}
1012 1013 1014 1015 1016 1017 1018
		if (!tipc_msg_extract(skb, &iskb, &ipos))
			return rc;
		kfree_skb(skb);
		skb = iskb;
		hdr = buf_msg(skb);
		if (less(msg_seqno(hdr), l->drop_point))
			goto drop;
1019
		if (tipc_data_input(l, skb, inputq))
1020 1021 1022 1023
			return rc;
		usr = msg_user(hdr);
		reasm_skb = &l->failover_reasm_skb;
	}
1024

1025
	if (usr == MSG_BUNDLER) {
1026
		skb_queue_head_init(&tmpq);
1027 1028
		l->stats.recv_bundles++;
		l->stats.recv_bundled += msg_msgcnt(hdr);
1029
		while (tipc_msg_extract(skb, &iskb, &pos))
1030 1031
			tipc_data_input(l, iskb, &tmpq);
		tipc_skb_queue_splice_tail(&tmpq, inputq);
1032
		return 0;
1033 1034 1035 1036
	} else if (usr == MSG_FRAGMENTER) {
		l->stats.recv_fragments++;
		if (tipc_buf_append(reasm_skb, &skb)) {
			l->stats.recv_fragmented++;
1037
			tipc_data_input(l, skb, inputq);
1038
		} else if (!*reasm_skb) {
1039
			return tipc_link_fsm_evt(l, LINK_FAILURE_EVT);
1040
		}
1041
		return 0;
1042
	} else if (usr == BCAST_PROTOCOL) {
1043
		tipc_link_sync_rcv(node, skb);
1044
		return 0;
1045 1046 1047
	}
drop:
	kfree_skb(skb);
1048
	return 0;
1049 1050
}

1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
static bool tipc_link_release_pkts(struct tipc_link *l, u16 acked)
{
	bool released = false;
	struct sk_buff *skb, *tmp;

	skb_queue_walk_safe(&l->transmq, skb, tmp) {
		if (more(buf_seqno(skb), acked))
			break;
		__skb_unlink(skb, &l->transmq);
		kfree_skb(skb);
		released = true;
	}
	return released;
}

1066 1067 1068 1069 1070 1071 1072 1073 1074
/* tipc_link_build_ack_msg: prepare link acknowledge message for transmission
 */
void tipc_link_build_ack_msg(struct tipc_link *l, struct sk_buff_head *xmitq)
{
	l->rcv_unacked = 0;
	l->stats.sent_acks++;
	tipc_link_build_proto_msg(l, STATE_MSG, 0, 0, 0, 0, xmitq);
}

1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086
/* tipc_link_build_reset_msg: prepare link RESET or ACTIVATE message
 */
void tipc_link_build_reset_msg(struct tipc_link *l, struct sk_buff_head *xmitq)
{
	int mtyp = RESET_MSG;

	if (l->state == LINK_ESTABLISHING)
		mtyp = ACTIVATE_MSG;

	tipc_link_build_proto_msg(l, mtyp, 0, 0, 0, 0, xmitq);
}

1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
/* tipc_link_build_nack_msg: prepare link nack message for transmission
 */
static void tipc_link_build_nack_msg(struct tipc_link *l,
				     struct sk_buff_head *xmitq)
{
	u32 def_cnt = ++l->stats.deferred_recv;

	if ((skb_queue_len(&l->deferdq) == 1) || !(def_cnt % TIPC_NACK_INTV))
		tipc_link_build_proto_msg(l, STATE_MSG, 0, 0, 0, 0, xmitq);
}

1098
/* tipc_link_rcv - process TIPC packets/messages arriving from off-node
1099
 * @l: the link that should handle the message
1100 1101 1102 1103 1104 1105
 * @skb: TIPC packet
 * @xmitq: queue to place packets to be sent after this call
 */
int tipc_link_rcv(struct tipc_link *l, struct sk_buff *skb,
		  struct sk_buff_head *xmitq)
{
1106
	struct sk_buff_head *defq = &l->deferdq;
1107
	struct tipc_msg *hdr;
1108
	u16 seqno, rcv_nxt, win_lim;
1109 1110
	int rc = 0;

1111
	do {
1112
		hdr = buf_msg(skb);
1113 1114
		seqno = msg_seqno(hdr);
		rcv_nxt = l->rcv_nxt;
1115
		win_lim = rcv_nxt + TIPC_MAX_LINK_WIN;
1116 1117

		/* Verify and update link state */
1118 1119
		if (unlikely(msg_user(hdr) == LINK_PROTOCOL))
			return tipc_link_proto_rcv(l, skb, xmitq);
1120

1121
		if (unlikely(!link_is_up(l))) {
1122 1123 1124
			if (l->state == LINK_ESTABLISHING)
				rc = TIPC_LINK_UP_EVT;
			goto drop;
1125 1126
		}

1127
		/* Don't send probe at next timeout expiration */
1128 1129
		l->silent_intv_cnt = 0;

1130 1131 1132 1133 1134 1135
		/* Drop if outside receive window */
		if (unlikely(less(seqno, rcv_nxt) || more(seqno, win_lim))) {
			l->stats.duplicates++;
			goto drop;
		}

1136 1137 1138 1139 1140 1141 1142
		/* Forward queues and wake up waiting users */
		if (likely(tipc_link_release_pkts(l, msg_ack(hdr)))) {
			tipc_link_advance_backlog(l, xmitq);
			if (unlikely(!skb_queue_empty(&l->wakeupq)))
				link_prepare_wakeup(l);
		}

1143 1144
		/* Defer delivery if sequence gap */
		if (unlikely(seqno != rcv_nxt)) {
1145
			__tipc_skb_queue_sorted(defq, seqno, skb);
1146 1147
			tipc_link_build_nack_msg(l, xmitq);
			break;
1148 1149
		}

1150
		/* Deliver packet */
1151 1152
		l->rcv_nxt++;
		l->stats.recv_info++;
1153
		if (!tipc_data_input(l, skb, l->inputq))
1154
			rc = tipc_link_input(l, skb, l->inputq);
1155
		if (unlikely(rc))
1156 1157 1158 1159 1160 1161 1162 1163 1164
			break;
		if (unlikely(++l->rcv_unacked >= TIPC_MIN_LINK_WIN))
			tipc_link_build_ack_msg(l, xmitq);

	} while ((skb = __skb_dequeue(defq)));

	return rc;
drop:
	kfree_skb(skb);
1165 1166 1167
	return rc;
}

1168
/**
1169 1170 1171
 * tipc_link_defer_pkt - Add out-of-sequence message to deferred reception queue
 *
 * Returns increase in queue length (i.e. 0 or 1)
P
Per Liden 已提交
1172
 */
1173
u32 tipc_link_defer_pkt(struct sk_buff_head *list, struct sk_buff *skb)
P
Per Liden 已提交
1174
{
1175
	struct sk_buff *skb1;
1176
	u16 seq_no = buf_seqno(skb);
P
Per Liden 已提交
1177 1178

	/* Empty queue ? */
1179 1180
	if (skb_queue_empty(list)) {
		__skb_queue_tail(list, skb);
P
Per Liden 已提交
1181 1182 1183 1184
		return 1;
	}

	/* Last ? */
1185 1186
	if (less(buf_seqno(skb_peek_tail(list)), seq_no)) {
		__skb_queue_tail(list, skb);
P
Per Liden 已提交
1187 1188 1189
		return 1;
	}

1190
	/* Locate insertion point in queue, then insert; discard if duplicate */
1191
	skb_queue_walk(list, skb1) {
1192
		u16 curr_seqno = buf_seqno(skb1);
P
Per Liden 已提交
1193

1194
		if (seq_no == curr_seqno) {
1195
			kfree_skb(skb);
1196
			return 0;
P
Per Liden 已提交
1197
		}
1198 1199

		if (less(seq_no, curr_seqno))
P
Per Liden 已提交
1200
			break;
1201
	}
P
Per Liden 已提交
1202

1203
	__skb_queue_before(list, skb1, skb);
1204
	return 1;
P
Per Liden 已提交
1205 1206 1207 1208 1209
}

/*
 * Send protocol message to the other endpoint.
 */
1210
void tipc_link_proto_xmit(struct tipc_link *l, u32 msg_typ, int probe_msg,
1211
			  u32 gap, u32 tolerance, u32 priority)
P
Per Liden 已提交
1212
{
1213 1214
	struct sk_buff *skb = NULL;
	struct sk_buff_head xmitq;
P
Per Liden 已提交
1215

1216 1217 1218 1219 1220
	__skb_queue_head_init(&xmitq);
	tipc_link_build_proto_msg(l, msg_typ, probe_msg, gap,
				  tolerance, priority, &xmitq);
	skb = __skb_dequeue(&xmitq);
	if (!skb)
P
Per Liden 已提交
1221
		return;
J
Jon Paul Maloy 已提交
1222
	tipc_bearer_send(l->owner->net, l->bearer_id, skb, l->media_addr);
1223 1224
	l->rcv_unacked = 0;
	kfree_skb(skb);
P
Per Liden 已提交
1225 1226
}

1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
/* tipc_link_build_proto_msg: prepare link protocol message for transmission
 */
static void tipc_link_build_proto_msg(struct tipc_link *l, int mtyp, bool probe,
				      u16 rcvgap, int tolerance, int priority,
				      struct sk_buff_head *xmitq)
{
	struct sk_buff *skb = NULL;
	struct tipc_msg *hdr = l->pmsg;
	u16 snd_nxt = l->snd_nxt;
	u16 rcv_nxt = l->rcv_nxt;
	u16 rcv_last = rcv_nxt - 1;
	int node_up = l->owner->bclink.recv_permitted;

	/* Don't send protocol message during reset or link failover */
1241
	if (tipc_link_is_blocked(l))
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275
		return;

	msg_set_type(hdr, mtyp);
	msg_set_net_plane(hdr, l->net_plane);
	msg_set_bcast_ack(hdr, l->owner->bclink.last_in);
	msg_set_last_bcast(hdr, tipc_bclink_get_last_sent(l->owner->net));
	msg_set_link_tolerance(hdr, tolerance);
	msg_set_linkprio(hdr, priority);
	msg_set_redundant_link(hdr, node_up);
	msg_set_seq_gap(hdr, 0);

	/* Compatibility: created msg must not be in sequence with pkt flow */
	msg_set_seqno(hdr, snd_nxt + U16_MAX / 2);

	if (mtyp == STATE_MSG) {
		if (!tipc_link_is_up(l))
			return;
		msg_set_next_sent(hdr, snd_nxt);

		/* Override rcvgap if there are packets in deferred queue */
		if (!skb_queue_empty(&l->deferdq))
			rcvgap = buf_seqno(skb_peek(&l->deferdq)) - rcv_nxt;
		if (rcvgap) {
			msg_set_seq_gap(hdr, rcvgap);
			l->stats.sent_nacks++;
		}
		msg_set_ack(hdr, rcv_last);
		msg_set_probe(hdr, probe);
		if (probe)
			l->stats.sent_probes++;
		l->stats.sent_states++;
	} else {
		/* RESET_MSG or ACTIVATE_MSG */
		msg_set_max_pkt(hdr, l->advertised_mtu);
1276
		msg_set_ack(hdr, l->rcv_nxt - 1);
1277 1278 1279 1280 1281 1282 1283
		msg_set_next_sent(hdr, 1);
	}
	skb = tipc_buf_acquire(msg_size(hdr));
	if (!skb)
		return;
	skb_copy_to_linear_data(skb, hdr, msg_size(hdr));
	skb->priority = TC_PRIO_CONTROL;
1284
	__skb_queue_tail(xmitq, skb);
P
Per Liden 已提交
1285 1286
}

1287
/* tipc_link_tnl_prepare(): prepare and return a list of tunnel packets
1288
 * with contents of the link's transmit and backlog queues.
P
Per Liden 已提交
1289
 */
1290 1291
void tipc_link_tnl_prepare(struct tipc_link *l, struct tipc_link *tnl,
			   int mtyp, struct sk_buff_head *xmitq)
P
Per Liden 已提交
1292
{
1293 1294 1295 1296 1297
	struct sk_buff *skb, *tnlskb;
	struct tipc_msg *hdr, tnlhdr;
	struct sk_buff_head *queue = &l->transmq;
	struct sk_buff_head tmpxq, tnlq;
	u16 pktlen, pktcnt, seqno = l->snd_nxt;
P
Per Liden 已提交
1298

1299
	if (!tnl)
P
Per Liden 已提交
1300 1301
		return;

1302 1303
	skb_queue_head_init(&tnlq);
	skb_queue_head_init(&tmpxq);
1304

1305 1306 1307 1308 1309 1310
	/* At least one packet required for safe algorithm => add dummy */
	skb = tipc_msg_create(TIPC_LOW_IMPORTANCE, TIPC_DIRECT_MSG,
			      BASIC_H_SIZE, 0, l->addr, link_own_addr(l),
			      0, 0, TIPC_ERR_NO_PORT);
	if (!skb) {
		pr_warn("%sunable to create tunnel packet\n", link_co_err);
P
Per Liden 已提交
1311 1312
		return;
	}
1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324
	skb_queue_tail(&tnlq, skb);
	tipc_link_xmit(l, &tnlq, &tmpxq);
	__skb_queue_purge(&tmpxq);

	/* Initialize reusable tunnel packet header */
	tipc_msg_init(link_own_addr(l), &tnlhdr, TUNNEL_PROTOCOL,
		      mtyp, INT_H_SIZE, l->addr);
	pktcnt = skb_queue_len(&l->transmq) + skb_queue_len(&l->backlogq);
	msg_set_msgcnt(&tnlhdr, pktcnt);
	msg_set_bearer_id(&tnlhdr, l->peer_bearer_id);
tnl:
	/* Wrap each packet into a tunnel packet */
J
Jon Paul Maloy 已提交
1325
	skb_queue_walk(queue, skb) {
1326 1327 1328 1329 1330 1331 1332 1333
		hdr = buf_msg(skb);
		if (queue == &l->backlogq)
			msg_set_seqno(hdr, seqno++);
		pktlen = msg_size(hdr);
		msg_set_size(&tnlhdr, pktlen + INT_H_SIZE);
		tnlskb = tipc_buf_acquire(pktlen + INT_H_SIZE);
		if (!tnlskb) {
			pr_warn("%sunable to send packet\n", link_co_err);
P
Per Liden 已提交
1334 1335
			return;
		}
1336 1337 1338
		skb_copy_to_linear_data(tnlskb, &tnlhdr, INT_H_SIZE);
		skb_copy_to_linear_data_offset(tnlskb, INT_H_SIZE, hdr, pktlen);
		__skb_queue_tail(&tnlq, tnlskb);
P
Per Liden 已提交
1339
	}
1340 1341 1342
	if (queue != &l->backlogq) {
		queue = &l->backlogq;
		goto tnl;
1343
	}
1344

1345
	tipc_link_xmit(tnl, &tnlq, xmitq);
P
Per Liden 已提交
1346

1347 1348 1349 1350
	if (mtyp == FAILOVER_MSG) {
		tnl->drop_point = l->rcv_nxt;
		tnl->failover_reasm_skb = l->reasm_buf;
		l->reasm_buf = NULL;
1351
	}
P
Per Liden 已提交
1352 1353
}

1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
/* tipc_link_proto_rcv(): receive link level protocol message :
 * Note that network plane id propagates through the network, and may
 * change at any time. The node with lowest numerical id determines
 * network plane
 */
static int tipc_link_proto_rcv(struct tipc_link *l, struct sk_buff *skb,
			       struct sk_buff_head *xmitq)
{
	struct tipc_msg *hdr = buf_msg(skb);
	u16 rcvgap = 0;
	u16 nacked_gap = msg_seq_gap(hdr);
	u16 peers_snd_nxt =  msg_next_sent(hdr);
	u16 peers_tol = msg_link_tolerance(hdr);
	u16 peers_prio = msg_linkprio(hdr);
1368
	u16 rcv_nxt = l->rcv_nxt;
1369
	int mtyp = msg_type(hdr);
1370 1371 1372
	char *if_name;
	int rc = 0;

1373
	if (tipc_link_is_blocked(l))
1374 1375 1376 1377 1378
		goto exit;

	if (link_own_addr(l) > msg_prevnode(hdr))
		l->net_plane = msg_net_plane(hdr);

1379
	switch (mtyp) {
1380 1381 1382 1383 1384 1385 1386
	case RESET_MSG:

		/* Ignore duplicate RESET with old session number */
		if ((less_eq(msg_session(hdr), l->peer_session)) &&
		    (l->peer_session != WILDCARD_SESSION))
			break;
		/* fall thru' */
1387

1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405
	case ACTIVATE_MSG:

		/* Complete own link name with peer's interface name */
		if_name =  strrchr(l->name, ':') + 1;
		if (sizeof(l->name) - (if_name - l->name) <= TIPC_MAX_IF_NAME)
			break;
		if (msg_data_sz(hdr) < TIPC_MAX_IF_NAME)
			break;
		strncpy(if_name, msg_data(hdr),	TIPC_MAX_IF_NAME);

		/* Update own tolerance if peer indicates a non-zero value */
		if (in_range(peers_tol, TIPC_MIN_LINK_TOL, TIPC_MAX_LINK_TOL))
			l->tolerance = peers_tol;

		/* Update own priority if peer's priority is higher */
		if (in_range(peers_prio, l->priority + 1, TIPC_MAX_LINK_PRI))
			l->priority = peers_prio;

1406 1407 1408 1409 1410 1411 1412 1413
		/* ACTIVATE_MSG serves as PEER_RESET if link is already down */
		if ((mtyp == RESET_MSG) || !link_is_up(l))
			rc = tipc_link_fsm_evt(l, LINK_PEER_RESET_EVT);

		/* ACTIVATE_MSG takes up link if it was already locally reset */
		if ((mtyp == ACTIVATE_MSG) && (l->state == LINK_ESTABLISHING))
			rc = TIPC_LINK_UP_EVT;

1414 1415 1416 1417 1418
		l->peer_session = msg_session(hdr);
		l->peer_bearer_id = msg_bearer_id(hdr);
		if (l->mtu > msg_max_pkt(hdr))
			l->mtu = msg_max_pkt(hdr);
		break;
1419

1420
	case STATE_MSG:
1421

1422 1423 1424 1425 1426 1427 1428 1429
		/* Update own tolerance if peer indicates a non-zero value */
		if (in_range(peers_tol, TIPC_MIN_LINK_TOL, TIPC_MAX_LINK_TOL))
			l->tolerance = peers_tol;

		l->silent_intv_cnt = 0;
		l->stats.recv_states++;
		if (msg_probe(hdr))
			l->stats.recv_probes++;
1430 1431 1432 1433

		if (!link_is_up(l)) {
			if (l->state == LINK_ESTABLISHING)
				rc = TIPC_LINK_UP_EVT;
1434
			break;
1435
		}
1436

1437
		/* Send NACK if peer has sent pkts we haven't received yet */
1438
		if (more(peers_snd_nxt, rcv_nxt) && !tipc_link_is_synching(l))
1439 1440 1441
			rcvgap = peers_snd_nxt - l->rcv_nxt;
		if (rcvgap || (msg_probe(hdr)))
			tipc_link_build_proto_msg(l, STATE_MSG, 0, rcvgap,
J
Jon Paul Maloy 已提交
1442
						  0, 0, xmitq);
1443 1444 1445 1446
		tipc_link_release_pkts(l, msg_ack(hdr));

		/* If NACK, retransmit will now start at right position */
		if (nacked_gap) {
1447
			rc = tipc_link_retransm(l, nacked_gap, xmitq);
1448 1449
			l->stats.recv_nacks++;
		}
1450

1451 1452 1453 1454 1455 1456 1457 1458 1459
		tipc_link_advance_backlog(l, xmitq);
		if (unlikely(!skb_queue_empty(&l->wakeupq)))
			link_prepare_wakeup(l);
	}
exit:
	kfree_skb(skb);
	return rc;
}

1460
void tipc_link_set_queue_limits(struct tipc_link *l, u32 win)
P
Per Liden 已提交
1461
{
1462
	int max_bulk = TIPC_MAX_PUBLICATIONS / (l->mtu / ITEM_SIZE);
1463 1464

	l->window = win;
1465 1466 1467 1468 1469
	l->backlog[TIPC_LOW_IMPORTANCE].limit      = win / 2;
	l->backlog[TIPC_MEDIUM_IMPORTANCE].limit   = win;
	l->backlog[TIPC_HIGH_IMPORTANCE].limit     = win / 2 * 3;
	l->backlog[TIPC_CRITICAL_IMPORTANCE].limit = win * 2;
	l->backlog[TIPC_SYSTEM_IMPORTANCE].limit   = max_bulk;
P
Per Liden 已提交
1470 1471
}

1472
/* tipc_link_find_owner - locate owner node of link by link's name
1473
 * @net: the applicable net namespace
1474 1475
 * @name: pointer to link name string
 * @bearer_id: pointer to index in 'node->links' array where the link was found.
1476
 *
1477
 * Returns pointer to node owning the link, or 0 if no matching link is found.
P
Per Liden 已提交
1478
 */
1479 1480
static struct tipc_node *tipc_link_find_owner(struct net *net,
					      const char *link_name,
1481
					      unsigned int *bearer_id)
P
Per Liden 已提交
1482
{
1483
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1484
	struct tipc_link *l_ptr;
1485
	struct tipc_node *n_ptr;
1486
	struct tipc_node *found_node = NULL;
1487
	int i;
P
Per Liden 已提交
1488

1489
	*bearer_id = 0;
1490
	rcu_read_lock();
1491
	list_for_each_entry_rcu(n_ptr, &tn->node_list, list) {
1492
		tipc_node_lock(n_ptr);
1493
		for (i = 0; i < MAX_BEARERS; i++) {
1494
			l_ptr = n_ptr->links[i].link;
1495 1496 1497 1498 1499
			if (l_ptr && !strcmp(l_ptr->name, link_name)) {
				*bearer_id = i;
				found_node = n_ptr;
				break;
			}
1500
		}
1501
		tipc_node_unlock(n_ptr);
1502 1503
		if (found_node)
			break;
1504
	}
1505 1506
	rcu_read_unlock();

1507
	return found_node;
P
Per Liden 已提交
1508 1509 1510 1511 1512 1513
}

/**
 * link_reset_statistics - reset link statistics
 * @l_ptr: pointer to link
 */
1514
static void link_reset_statistics(struct tipc_link *l_ptr)
P
Per Liden 已提交
1515 1516
{
	memset(&l_ptr->stats, 0, sizeof(l_ptr->stats));
1517 1518
	l_ptr->stats.sent_info = l_ptr->snd_nxt;
	l_ptr->stats.recv_info = l_ptr->rcv_nxt;
P
Per Liden 已提交
1519 1520
}

1521
static void link_print(struct tipc_link *l, const char *str)
P
Per Liden 已提交
1522
{
1523 1524 1525
	struct sk_buff *hskb = skb_peek(&l->transmq);
	u16 head = hskb ? msg_seqno(buf_msg(hskb)) : l->snd_nxt;
	u16 tail = l->snd_nxt - 1;
1526

1527
	pr_info("%s Link <%s> state %x\n", str, l->name, l->state);
1528 1529 1530
	pr_info("XMTQ: %u [%u-%u], BKLGQ: %u, SNDNX: %u, RCVNX: %u\n",
		skb_queue_len(&l->transmq), head, tail,
		skb_queue_len(&l->backlogq), l->snd_nxt, l->rcv_nxt);
P
Per Liden 已提交
1531
}
1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569

/* Parse and validate nested (link) properties valid for media, bearer and link
 */
int tipc_nl_parse_link_prop(struct nlattr *prop, struct nlattr *props[])
{
	int err;

	err = nla_parse_nested(props, TIPC_NLA_PROP_MAX, prop,
			       tipc_nl_prop_policy);
	if (err)
		return err;

	if (props[TIPC_NLA_PROP_PRIO]) {
		u32 prio;

		prio = nla_get_u32(props[TIPC_NLA_PROP_PRIO]);
		if (prio > TIPC_MAX_LINK_PRI)
			return -EINVAL;
	}

	if (props[TIPC_NLA_PROP_TOL]) {
		u32 tol;

		tol = nla_get_u32(props[TIPC_NLA_PROP_TOL]);
		if ((tol < TIPC_MIN_LINK_TOL) || (tol > TIPC_MAX_LINK_TOL))
			return -EINVAL;
	}

	if (props[TIPC_NLA_PROP_WIN]) {
		u32 win;

		win = nla_get_u32(props[TIPC_NLA_PROP_WIN]);
		if ((win < TIPC_MIN_LINK_WIN) || (win > TIPC_MAX_LINK_WIN))
			return -EINVAL;
	}

	return 0;
}
1570

1571 1572 1573 1574 1575 1576 1577 1578 1579
int tipc_nl_link_set(struct sk_buff *skb, struct genl_info *info)
{
	int err;
	int res = 0;
	int bearer_id;
	char *name;
	struct tipc_link *link;
	struct tipc_node *node;
	struct nlattr *attrs[TIPC_NLA_LINK_MAX + 1];
1580
	struct net *net = sock_net(skb->sk);
1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595

	if (!info->attrs[TIPC_NLA_LINK])
		return -EINVAL;

	err = nla_parse_nested(attrs, TIPC_NLA_LINK_MAX,
			       info->attrs[TIPC_NLA_LINK],
			       tipc_nl_link_policy);
	if (err)
		return err;

	if (!attrs[TIPC_NLA_LINK_NAME])
		return -EINVAL;

	name = nla_data(attrs[TIPC_NLA_LINK_NAME]);

1596 1597 1598
	if (strcmp(name, tipc_bclink_name) == 0)
		return tipc_nl_bc_link_set(net, attrs);

1599
	node = tipc_link_find_owner(net, name, &bearer_id);
1600 1601 1602 1603 1604
	if (!node)
		return -EINVAL;

	tipc_node_lock(node);

1605
	link = node->links[bearer_id].link;
1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
	if (!link) {
		res = -EINVAL;
		goto out;
	}

	if (attrs[TIPC_NLA_LINK_PROP]) {
		struct nlattr *props[TIPC_NLA_PROP_MAX + 1];

		err = tipc_nl_parse_link_prop(attrs[TIPC_NLA_LINK_PROP],
					      props);
		if (err) {
			res = err;
			goto out;
		}

		if (props[TIPC_NLA_PROP_TOL]) {
			u32 tol;

			tol = nla_get_u32(props[TIPC_NLA_PROP_TOL]);
1625
			link->tolerance = tol;
1626
			tipc_link_proto_xmit(link, STATE_MSG, 0, 0, tol, 0);
1627 1628 1629 1630 1631 1632
		}
		if (props[TIPC_NLA_PROP_PRIO]) {
			u32 prio;

			prio = nla_get_u32(props[TIPC_NLA_PROP_PRIO]);
			link->priority = prio;
1633
			tipc_link_proto_xmit(link, STATE_MSG, 0, 0, 0, prio);
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647
		}
		if (props[TIPC_NLA_PROP_WIN]) {
			u32 win;

			win = nla_get_u32(props[TIPC_NLA_PROP_WIN]);
			tipc_link_set_queue_limits(link, win);
		}
	}

out:
	tipc_node_unlock(node);

	return res;
}
1648 1649

static int __tipc_nl_add_stats(struct sk_buff *skb, struct tipc_stats *s)
1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
{
	int i;
	struct nlattr *stats;

	struct nla_map {
		u32 key;
		u32 val;
	};

	struct nla_map map[] = {
		{TIPC_NLA_STATS_RX_INFO, s->recv_info},
		{TIPC_NLA_STATS_RX_FRAGMENTS, s->recv_fragments},
		{TIPC_NLA_STATS_RX_FRAGMENTED, s->recv_fragmented},
		{TIPC_NLA_STATS_RX_BUNDLES, s->recv_bundles},
		{TIPC_NLA_STATS_RX_BUNDLED, s->recv_bundled},
		{TIPC_NLA_STATS_TX_INFO, s->sent_info},
		{TIPC_NLA_STATS_TX_FRAGMENTS, s->sent_fragments},
		{TIPC_NLA_STATS_TX_FRAGMENTED, s->sent_fragmented},
		{TIPC_NLA_STATS_TX_BUNDLES, s->sent_bundles},
		{TIPC_NLA_STATS_TX_BUNDLED, s->sent_bundled},
		{TIPC_NLA_STATS_MSG_PROF_TOT, (s->msg_length_counts) ?
			s->msg_length_counts : 1},
		{TIPC_NLA_STATS_MSG_LEN_CNT, s->msg_length_counts},
		{TIPC_NLA_STATS_MSG_LEN_TOT, s->msg_lengths_total},
		{TIPC_NLA_STATS_MSG_LEN_P0, s->msg_length_profile[0]},
		{TIPC_NLA_STATS_MSG_LEN_P1, s->msg_length_profile[1]},
		{TIPC_NLA_STATS_MSG_LEN_P2, s->msg_length_profile[2]},
		{TIPC_NLA_STATS_MSG_LEN_P3, s->msg_length_profile[3]},
		{TIPC_NLA_STATS_MSG_LEN_P4, s->msg_length_profile[4]},
		{TIPC_NLA_STATS_MSG_LEN_P5, s->msg_length_profile[5]},
		{TIPC_NLA_STATS_MSG_LEN_P6, s->msg_length_profile[6]},
		{TIPC_NLA_STATS_RX_STATES, s->recv_states},
		{TIPC_NLA_STATS_RX_PROBES, s->recv_probes},
		{TIPC_NLA_STATS_RX_NACKS, s->recv_nacks},
		{TIPC_NLA_STATS_RX_DEFERRED, s->deferred_recv},
		{TIPC_NLA_STATS_TX_STATES, s->sent_states},
		{TIPC_NLA_STATS_TX_PROBES, s->sent_probes},
		{TIPC_NLA_STATS_TX_NACKS, s->sent_nacks},
		{TIPC_NLA_STATS_TX_ACKS, s->sent_acks},
		{TIPC_NLA_STATS_RETRANSMITTED, s->retransmitted},
		{TIPC_NLA_STATS_DUPLICATES, s->duplicates},
		{TIPC_NLA_STATS_LINK_CONGS, s->link_congs},
		{TIPC_NLA_STATS_MAX_QUEUE, s->max_queue_sz},
		{TIPC_NLA_STATS_AVG_QUEUE, s->queue_sz_counts ?
			(s->accu_queue_sz / s->queue_sz_counts) : 0}
	};

	stats = nla_nest_start(skb, TIPC_NLA_LINK_STATS);
	if (!stats)
		return -EMSGSIZE;

	for (i = 0; i <  ARRAY_SIZE(map); i++)
		if (nla_put_u32(skb, map[i].key, map[i].val))
			goto msg_full;

	nla_nest_end(skb, stats);

	return 0;
msg_full:
	nla_nest_cancel(skb, stats);

	return -EMSGSIZE;
}

/* Caller should hold appropriate locks to protect the link */
1715
static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg,
1716
			      struct tipc_link *link, int nlflags)
1717 1718 1719 1720 1721
{
	int err;
	void *hdr;
	struct nlattr *attrs;
	struct nlattr *prop;
1722
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1723

1724
	hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
1725
			  nlflags, TIPC_NL_LINK_GET);
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735
	if (!hdr)
		return -EMSGSIZE;

	attrs = nla_nest_start(msg->skb, TIPC_NLA_LINK);
	if (!attrs)
		goto msg_full;

	if (nla_put_string(msg->skb, TIPC_NLA_LINK_NAME, link->name))
		goto attr_msg_full;
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_DEST,
1736
			tipc_cluster_mask(tn->own_addr)))
1737
		goto attr_msg_full;
1738
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_MTU, link->mtu))
1739
		goto attr_msg_full;
1740
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, link->rcv_nxt))
1741
		goto attr_msg_full;
1742
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, link->snd_nxt))
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759
		goto attr_msg_full;

	if (tipc_link_is_up(link))
		if (nla_put_flag(msg->skb, TIPC_NLA_LINK_UP))
			goto attr_msg_full;
	if (tipc_link_is_active(link))
		if (nla_put_flag(msg->skb, TIPC_NLA_LINK_ACTIVE))
			goto attr_msg_full;

	prop = nla_nest_start(msg->skb, TIPC_NLA_LINK_PROP);
	if (!prop)
		goto attr_msg_full;
	if (nla_put_u32(msg->skb, TIPC_NLA_PROP_PRIO, link->priority))
		goto prop_msg_full;
	if (nla_put_u32(msg->skb, TIPC_NLA_PROP_TOL, link->tolerance))
		goto prop_msg_full;
	if (nla_put_u32(msg->skb, TIPC_NLA_PROP_WIN,
1760
			link->window))
1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785
		goto prop_msg_full;
	if (nla_put_u32(msg->skb, TIPC_NLA_PROP_PRIO, link->priority))
		goto prop_msg_full;
	nla_nest_end(msg->skb, prop);

	err = __tipc_nl_add_stats(msg->skb, &link->stats);
	if (err)
		goto attr_msg_full;

	nla_nest_end(msg->skb, attrs);
	genlmsg_end(msg->skb, hdr);

	return 0;

prop_msg_full:
	nla_nest_cancel(msg->skb, prop);
attr_msg_full:
	nla_nest_cancel(msg->skb, attrs);
msg_full:
	genlmsg_cancel(msg->skb, hdr);

	return -EMSGSIZE;
}

/* Caller should hold node lock  */
1786 1787
static int __tipc_nl_add_node_links(struct net *net, struct tipc_nl_msg *msg,
				    struct tipc_node *node, u32 *prev_link)
1788 1789 1790 1791 1792 1793 1794
{
	u32 i;
	int err;

	for (i = *prev_link; i < MAX_BEARERS; i++) {
		*prev_link = i;

1795
		if (!node->links[i].link)
1796 1797
			continue;

1798 1799
		err = __tipc_nl_add_link(net, msg,
					 node->links[i].link, NLM_F_MULTI);
1800 1801 1802 1803 1804 1805 1806 1807 1808 1809
		if (err)
			return err;
	}
	*prev_link = 0;

	return 0;
}

int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
1810 1811
	struct net *net = sock_net(skb->sk);
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827
	struct tipc_node *node;
	struct tipc_nl_msg msg;
	u32 prev_node = cb->args[0];
	u32 prev_link = cb->args[1];
	int done = cb->args[2];
	int err;

	if (done)
		return 0;

	msg.skb = skb;
	msg.portid = NETLINK_CB(cb->skb).portid;
	msg.seq = cb->nlh->nlmsg_seq;

	rcu_read_lock();
	if (prev_node) {
1828
		node = tipc_node_find(net, prev_node);
1829 1830 1831 1832 1833 1834 1835 1836 1837 1838
		if (!node) {
			/* We never set seq or call nl_dump_check_consistent()
			 * this means that setting prev_seq here will cause the
			 * consistence check to fail in the netlink callback
			 * handler. Resulting in the last NLMSG_DONE message
			 * having the NLM_F_DUMP_INTR flag set.
			 */
			cb->prev_seq = 1;
			goto out;
		}
1839
		tipc_node_put(node);
1840

1841 1842
		list_for_each_entry_continue_rcu(node, &tn->node_list,
						 list) {
1843
			tipc_node_lock(node);
1844 1845
			err = __tipc_nl_add_node_links(net, &msg, node,
						       &prev_link);
1846 1847 1848 1849 1850 1851 1852
			tipc_node_unlock(node);
			if (err)
				goto out;

			prev_node = node->addr;
		}
	} else {
1853
		err = tipc_nl_add_bc_link(net, &msg);
1854 1855 1856
		if (err)
			goto out;

1857
		list_for_each_entry_rcu(node, &tn->node_list, list) {
1858
			tipc_node_lock(node);
1859 1860
			err = __tipc_nl_add_node_links(net, &msg, node,
						       &prev_link);
1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880
			tipc_node_unlock(node);
			if (err)
				goto out;

			prev_node = node->addr;
		}
	}
	done = 1;
out:
	rcu_read_unlock();

	cb->args[0] = prev_node;
	cb->args[1] = prev_link;
	cb->args[2] = done;

	return skb->len;
}

int tipc_nl_link_get(struct sk_buff *skb, struct genl_info *info)
{
1881
	struct net *net = genl_info_net(info);
1882 1883 1884 1885
	struct tipc_nl_msg msg;
	char *name;
	int err;

1886 1887 1888
	msg.portid = info->snd_portid;
	msg.seq = info->snd_seq;

1889 1890 1891 1892
	if (!info->attrs[TIPC_NLA_LINK_NAME])
		return -EINVAL;
	name = nla_data(info->attrs[TIPC_NLA_LINK_NAME]);

1893 1894
	msg.skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
	if (!msg.skb)
1895 1896
		return -ENOMEM;

1897 1898 1899 1900 1901 1902 1903 1904 1905 1906
	if (strcmp(name, tipc_bclink_name) == 0) {
		err = tipc_nl_add_bc_link(net, &msg);
		if (err) {
			nlmsg_free(msg.skb);
			return err;
		}
	} else {
		int bearer_id;
		struct tipc_node *node;
		struct tipc_link *link;
1907

1908 1909 1910
		node = tipc_link_find_owner(net, name, &bearer_id);
		if (!node)
			return -EINVAL;
1911

1912
		tipc_node_lock(node);
1913
		link = node->links[bearer_id].link;
1914 1915 1916 1917 1918
		if (!link) {
			tipc_node_unlock(node);
			nlmsg_free(msg.skb);
			return -EINVAL;
		}
1919

1920 1921 1922 1923 1924 1925 1926
		err = __tipc_nl_add_link(net, &msg, link, 0);
		tipc_node_unlock(node);
		if (err) {
			nlmsg_free(msg.skb);
			return err;
		}
	}
1927

1928
	return genlmsg_reply(msg.skb, info);
1929
}
1930 1931 1932 1933 1934 1935 1936 1937 1938

int tipc_nl_link_reset_stats(struct sk_buff *skb, struct genl_info *info)
{
	int err;
	char *link_name;
	unsigned int bearer_id;
	struct tipc_link *link;
	struct tipc_node *node;
	struct nlattr *attrs[TIPC_NLA_LINK_MAX + 1];
1939
	struct net *net = sock_net(skb->sk);
1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955

	if (!info->attrs[TIPC_NLA_LINK])
		return -EINVAL;

	err = nla_parse_nested(attrs, TIPC_NLA_LINK_MAX,
			       info->attrs[TIPC_NLA_LINK],
			       tipc_nl_link_policy);
	if (err)
		return err;

	if (!attrs[TIPC_NLA_LINK_NAME])
		return -EINVAL;

	link_name = nla_data(attrs[TIPC_NLA_LINK_NAME]);

	if (strcmp(link_name, tipc_bclink_name) == 0) {
1956
		err = tipc_bclink_reset_stats(net);
1957 1958 1959 1960 1961
		if (err)
			return err;
		return 0;
	}

1962
	node = tipc_link_find_owner(net, link_name, &bearer_id);
1963 1964 1965 1966 1967
	if (!node)
		return -EINVAL;

	tipc_node_lock(node);

1968
	link = node->links[bearer_id].link;
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
	if (!link) {
		tipc_node_unlock(node);
		return -EINVAL;
	}

	link_reset_statistics(link);

	tipc_node_unlock(node);

	return 0;
}