link.c 57.6 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 ";
53
static const char tipc_bclink_name[] = "broadcast-link";
P
Per Liden 已提交
54

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
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 }
};

71 72 73 74 75 76 77 78
/* 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 }
};

79 80 81 82 83 84 85 86
/* Send states for broadcast NACKs
 */
enum {
	BC_NACK_SND_CONDITIONAL,
	BC_NACK_SND_UNCONDITIONAL,
	BC_NACK_SND_SUPPRESS,
};

87 88 89 90
/*
 * Interval between NACKs when packets arrive out of order
 */
#define TIPC_NACK_INTV (TIPC_MIN_LINK_WIN * 2)
91 92 93
/*
 * Out-of-range value for link session numbers
 */
94
#define WILDCARD_SESSION 0x10000
95

96
/* Link FSM states:
P
Per Liden 已提交
97
 */
98
enum {
99 100 101 102 103 104 105
	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
106 107 108 109
};

/* Link FSM state checking routines
 */
110
static int link_is_up(struct tipc_link *l)
111
{
112
	return l->state & (LINK_ESTABLISHED | LINK_SYNCHING);
113
}
P
Per Liden 已提交
114

115 116
static int tipc_link_proto_rcv(struct tipc_link *l, struct sk_buff *skb,
			       struct sk_buff_head *xmitq);
117 118 119
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);
120 121
static void link_reset_statistics(struct tipc_link *l_ptr);
static void link_print(struct tipc_link *l_ptr, const char *str);
122 123 124 125 126
static void tipc_link_build_nack_msg(struct tipc_link *l,
				     struct sk_buff_head *xmitq);
static void tipc_link_build_bc_init_msg(struct tipc_link *l,
					struct sk_buff_head *xmitq);
static bool tipc_link_release_pkts(struct tipc_link *l, u16 to);
127

P
Per Liden 已提交
128
/*
S
Sam Ravnborg 已提交
129
 *  Simple non-static link routines (i.e. referenced outside this file)
P
Per Liden 已提交
130
 */
131
bool tipc_link_is_up(struct tipc_link *l)
P
Per Liden 已提交
132
{
133 134 135
	return link_is_up(l);
}

136 137 138 139 140
bool tipc_link_peer_is_down(struct tipc_link *l)
{
	return l->state == LINK_PEER_RESET;
}

141 142 143 144 145
bool tipc_link_is_reset(struct tipc_link *l)
{
	return l->state & (LINK_RESET | LINK_FAILINGOVER | LINK_ESTABLISHING);
}

146 147 148 149 150
bool tipc_link_is_establishing(struct tipc_link *l)
{
	return l->state == LINK_ESTABLISHING;
}

151 152 153 154 155 156 157 158 159 160 161 162 163
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
164 165
}

166 167 168 169 170 171 172 173 174 175
bool link_is_bc_sndlink(struct tipc_link *l)
{
	return !l->bc_sndlink;
}

bool link_is_bc_rcvlink(struct tipc_link *l)
{
	return ((l->bc_rcvlink == l) && !link_is_bc_sndlink(l));
}

176
int tipc_link_is_active(struct tipc_link *l)
P
Per Liden 已提交
177
{
178 179 180
	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
181 182
}

183 184 185
void tipc_link_add_bc_peer(struct tipc_link *snd_l,
			   struct tipc_link *uc_l,
			   struct sk_buff_head *xmitq)
186
{
187 188 189 190 191
	struct tipc_link *rcv_l = uc_l->bc_rcvlink;

	snd_l->ackers++;
	rcv_l->acked = snd_l->snd_nxt - 1;
	tipc_link_build_bc_init_msg(uc_l, xmitq);
192 193
}

194 195 196
void tipc_link_remove_bc_peer(struct tipc_link *snd_l,
			      struct tipc_link *rcv_l,
			      struct sk_buff_head *xmitq)
197
{
198 199 200 201 202 203 204 205 206 207
	u16 ack = snd_l->snd_nxt - 1;

	snd_l->ackers--;
	tipc_link_bc_ack_rcv(rcv_l, ack, xmitq);
	tipc_link_reset(rcv_l);
	rcv_l->state = LINK_RESET;
	if (!snd_l->ackers) {
		tipc_link_reset(snd_l);
		__skb_queue_purge(xmitq);
	}
208 209 210 211 212 213 214
}

int tipc_link_bc_peers(struct tipc_link *l)
{
	return l->ackers;
}

215 216 217 218 219 220 221 222 223 224
void tipc_link_set_mtu(struct tipc_link *l, int mtu)
{
	l->mtu = mtu;
}

int tipc_link_mtu(struct tipc_link *l)
{
	return l->mtu;
}

J
Jon Paul Maloy 已提交
225 226 227 228 229
static u32 link_own_addr(struct tipc_link *l)
{
	return msg_prevnode(l->pmsg);
}

P
Per Liden 已提交
230
/**
231
 * tipc_link_create - create a new link
J
Jon Paul Maloy 已提交
232
 * @n: pointer to associated node
233 234 235 236 237 238 239 240
 * @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
241
 * @ownnode: identity of own node
242
 * @peer: node id of peer node
243
 * @peer_caps: bitmap describing peer node capabilities
J
Jon Paul Maloy 已提交
244
 * @maddr: media address to be used
245 246
 * @bc_sndlink: the namespace global link used for broadcast sending
 * @bc_rcvlink: the peer specific link used for broadcast reception
J
Jon Paul Maloy 已提交
247 248 249
 * @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
250
 *
J
Jon Paul Maloy 已提交
251
 * Returns true if link was created, otherwise false
P
Per Liden 已提交
252
 */
253 254 255
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,
256 257 258 259 260 261
		      u16 peer_caps,
		      struct tipc_media_addr *maddr,
		      struct tipc_link *bc_sndlink,
		      struct tipc_link *bc_rcvlink,
		      struct sk_buff_head *inputq,
		      struct sk_buff_head *namedq,
J
Jon Paul Maloy 已提交
262
		      struct tipc_link **link)
P
Per Liden 已提交
263
{
J
Jon Paul Maloy 已提交
264 265
	struct tipc_link *l;
	struct tipc_msg *hdr;
266

J
Jon Paul Maloy 已提交
267 268 269 270
	l = kzalloc(sizeof(*l), GFP_ATOMIC);
	if (!l)
		return false;
	*link = l;
271 272 273 274 275 276
	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);
277

J
Jon Paul Maloy 已提交
278 279 280 281
	/* 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));
282
	strcpy((char *)msg_data(hdr), if_name);
P
Per Liden 已提交
283

J
Jon Paul Maloy 已提交
284
	l->addr = peer;
285
	l->peer_caps = peer_caps;
J
Jon Paul Maloy 已提交
286 287 288
	l->media_addr = maddr;
	l->owner = n;
	l->peer_session = WILDCARD_SESSION;
289 290 291 292 293 294 295
	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);
296
	l->ackers = 1;
297 298
	l->bc_sndlink = bc_sndlink;
	l->bc_rcvlink = bc_rcvlink;
J
Jon Paul Maloy 已提交
299 300 301 302 303 304 305 306 307
	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
308 309
}

310 311 312 313 314 315 316 317 318 319 320
/**
 * tipc_link_bc_create - create new link to be used for broadcast
 * @n: pointer to associated node
 * @mtu: mtu to be used
 * @window: send window 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
 *
 * Returns true if link was created, otherwise false
 */
321 322
bool tipc_link_bc_create(struct tipc_node *n, u32 ownnode, u32 peer,
			 int mtu, int window, u16 peer_caps,
323 324
			 struct sk_buff_head *inputq,
			 struct sk_buff_head *namedq,
325
			 struct tipc_link *bc_sndlink,
326 327 328 329 330
			 struct tipc_link **link)
{
	struct tipc_link *l;

	if (!tipc_link_create(n, "", MAX_BEARERS, 0, 'Z', mtu, 0, window,
331 332
			      0, ownnode, peer, peer_caps, NULL, bc_sndlink,
			      NULL, inputq, namedq, link))
333 334 335 336 337
		return false;

	l = *link;
	strcpy(l->name, tipc_bclink_name);
	tipc_link_reset(l);
338
	l->state = LINK_RESET;
339
	l->ackers = 0;
340
	l->bc_rcvlink = l;
341

342 343 344
	/* Broadcast send link is always up */
	if (link_is_bc_sndlink(l))
		l->state = LINK_ESTABLISHED;
345

346
	return true;
347 348
}

349 350 351 352 353
/**
 * tipc_link_fsm_evt - link finite state machine
 * @l: pointer to link
 * @evt: state machine event to be processed
 */
354
int tipc_link_fsm_evt(struct tipc_link *l, int evt)
355
{
356
	int rc = 0;
357 358

	switch (l->state) {
359
	case LINK_RESETTING:
360
		switch (evt) {
361 362
		case LINK_PEER_RESET_EVT:
			l->state = LINK_PEER_RESET;
363
			break;
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
		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;
381
			break;
382 383 384 385 386 387
		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:
388
			break;
389 390
		case LINK_SYNCH_BEGIN_EVT:
		case LINK_SYNCH_END_EVT:
391
		default:
392
			goto illegal_evt;
393 394
		}
		break;
395
	case LINK_PEER_RESET:
396
		switch (evt) {
397 398
		case LINK_RESET_EVT:
			l->state = LINK_ESTABLISHING;
399
			break;
400 401 402
		case LINK_PEER_RESET_EVT:
		case LINK_ESTABLISH_EVT:
		case LINK_FAILURE_EVT:
403
			break;
404 405 406 407 408 409 410 411 412 413 414 415
		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;
416
			break;
417 418 419 420 421 422 423 424
		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:
425
		default:
426
			goto illegal_evt;
427 428
		}
		break;
429
	case LINK_ESTABLISHING:
430
		switch (evt) {
431 432
		case LINK_ESTABLISH_EVT:
			l->state = LINK_ESTABLISHED;
433
			break;
434 435 436 437
		case LINK_FAILOVER_BEGIN_EVT:
			l->state = LINK_FAILINGOVER;
			break;
		case LINK_RESET_EVT:
438 439
			l->state = LINK_RESET;
			break;
440
		case LINK_FAILURE_EVT:
441
		case LINK_PEER_RESET_EVT:
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
		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;
459
			break;
460 461
		case LINK_RESET_EVT:
			l->state = LINK_RESET;
462
			break;
463
		case LINK_ESTABLISH_EVT:
464
		case LINK_SYNCH_END_EVT:
465
			break;
466 467 468 469 470
		case LINK_SYNCH_BEGIN_EVT:
			l->state = LINK_SYNCHING;
			break;
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_FAILOVER_END_EVT:
471
		default:
472
			goto illegal_evt;
473 474
		}
		break;
475
	case LINK_SYNCHING:
476
		switch (evt) {
477 478 479 480 481 482 483
		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;
484
			break;
485 486
		case LINK_RESET_EVT:
			l->state = LINK_RESET;
487
			break;
488 489
		case LINK_ESTABLISH_EVT:
		case LINK_SYNCH_BEGIN_EVT:
490
			break;
491 492 493 494 495
		case LINK_SYNCH_END_EVT:
			l->state = LINK_ESTABLISHED;
			break;
		case LINK_FAILOVER_BEGIN_EVT:
		case LINK_FAILOVER_END_EVT:
496
		default:
497
			goto illegal_evt;
498 499 500
		}
		break;
	default:
501
		pr_err("Unknown FSM state %x in %s\n", l->state, l->name);
502
	}
503 504 505 506
	return rc;
illegal_evt:
	pr_err("Illegal FSM event %x in state %x on link %s\n",
	       evt, l->state, l->name);
507 508 509
	return rc;
}

510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
/* 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]++;
}

551 552
/* tipc_link_timeout - perform periodic task as instructed from node timeout
 */
553 554 555 556 557
/* 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;
558 559 560
	int mtyp = STATE_MSG;
	bool xmit = false;
	bool prb = false;
561 562 563
	u16 bc_snt = l->bc_sndlink->snd_nxt - 1;
	u16 bc_acked = l->bc_rcvlink->acked;
	bool bc_up = link_is_up(l->bc_rcvlink);
564 565

	link_profile_stats(l);
566

567 568 569
	switch (l->state) {
	case LINK_ESTABLISHED:
	case LINK_SYNCHING:
570
		if (!l->silent_intv_cnt) {
571
			if (bc_up && (bc_acked != bc_snt))
572 573 574 575 576
				xmit = true;
		} else if (l->silent_intv_cnt <= l->abort_limit) {
			xmit = true;
			prb = true;
		} else {
577
			rc |= tipc_link_fsm_evt(l, LINK_FAILURE_EVT);
578 579
		}
		l->silent_intv_cnt++;
580 581
		break;
	case LINK_RESET:
582 583
		xmit = true;
		mtyp = RESET_MSG;
584 585
		break;
	case LINK_ESTABLISHING:
586 587
		xmit = true;
		mtyp = ACTIVATE_MSG;
588 589
		break;
	case LINK_PEER_RESET:
590
	case LINK_RESETTING:
591 592 593 594
	case LINK_FAILINGOVER:
		break;
	default:
		break;
595
	}
596

597 598 599
	if (xmit)
		tipc_link_build_proto_msg(l, mtyp, prb, 0, 0, 0, xmitq);

600 601 602
	return rc;
}

P
Per Liden 已提交
603
/**
604
 * link_schedule_user - schedule a message sender for wakeup after congestion
605
 * @link: congested link
606
 * @list: message that was attempted sent
607
 * Create pseudo msg to send back to user when congestion abates
608
 * Does not consume buffer list
P
Per Liden 已提交
609
 */
610
static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list)
P
Per Liden 已提交
611
{
612 613 614 615 616 617 618 619 620
	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);
621
		return -ENOBUFS;
622 623 624 625 626 627 628 629 630
	}
	/* 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)
631
		return -ENOBUFS;
632 633 634
	TIPC_SKB_CB(skb)->chain_sz = skb_queue_len(list);
	TIPC_SKB_CB(skb)->chain_imp = imp;
	skb_queue_tail(&link->wakeupq, skb);
635
	link->stats.link_congs++;
636
	return -ELINKCONG;
P
Per Liden 已提交
637 638
}

639 640 641 642 643 644
/**
 * 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
 */
645
void link_prepare_wakeup(struct tipc_link *l)
P
Per Liden 已提交
646
{
647 648
	int pnd[TIPC_SYSTEM_IMPORTANCE + 1] = {0,};
	int imp, lim;
649
	struct sk_buff *skb, *tmp;
650

651 652 653 654 655
	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
656
			break;
657
		skb_unlink(skb, &l->wakeupq);
658
		skb_queue_tail(l->inputq, skb);
P
Per Liden 已提交
659 660 661 662
	}
}

/**
663
 * tipc_link_reset_fragments - purge link's inbound message fragments queue
P
Per Liden 已提交
664 665
 * @l_ptr: pointer to link
 */
666
void tipc_link_reset_fragments(struct tipc_link *l_ptr)
P
Per Liden 已提交
667
{
668 669
	kfree_skb(l_ptr->reasm_buf);
	l_ptr->reasm_buf = NULL;
P
Per Liden 已提交
670 671
}

672
void tipc_link_purge_backlog(struct tipc_link *l)
673 674 675 676 677 678 679 680 681
{
	__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;
}

682
/**
683
 * tipc_link_purge_queues - purge all pkt queues associated with link
P
Per Liden 已提交
684 685
 * @l_ptr: pointer to link
 */
686
void tipc_link_purge_queues(struct tipc_link *l_ptr)
P
Per Liden 已提交
687
{
J
Jon Paul Maloy 已提交
688 689
	__skb_queue_purge(&l_ptr->deferdq);
	__skb_queue_purge(&l_ptr->transmq);
690
	tipc_link_purge_backlog(l_ptr);
691
	tipc_link_reset_fragments(l_ptr);
P
Per Liden 已提交
692 693
}

694
void tipc_link_reset(struct tipc_link *l)
P
Per Liden 已提交
695
{
696
	/* Link is down, accept any session */
697
	l->peer_session = WILDCARD_SESSION;
P
Per Liden 已提交
698

699 700
	/* 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
701

702 703
	/* Prepare for renewed mtu size negotiation */
	l->mtu = l->advertised_mtu;
P
Per Liden 已提交
704

705
	/* Clean up all queues: */
706 707
	__skb_queue_purge(&l->transmq);
	__skb_queue_purge(&l->deferdq);
708
	skb_queue_splice_init(&l->wakeupq, l->inputq);
709 710 711 712 713 714 715 716 717

	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;
718
	l->acked = 0;
719 720 721
	l->silent_intv_cnt = 0;
	l->stats.recv_info = 0;
	l->stale_count = 0;
722
	l->bc_peer_is_up = false;
723
	link_reset_statistics(l);
P
Per Liden 已提交
724 725
}

726
/**
727
 * __tipc_link_xmit(): same as tipc_link_xmit, but destlink is known & locked
728
 * @link: link to use
729 730
 * @list: chain of buffers containing message
 *
731
 * Consumes the buffer chain, except when returning an error code,
732 733
 * Returns 0 if success, or errno: -ELINKCONG, -EMSGSIZE or -ENOBUFS
 * Messages at TIPC_SYSTEM_IMPORTANCE are always accepted
734
 */
735 736
int __tipc_link_xmit(struct net *net, struct tipc_link *link,
		     struct sk_buff_head *list)
737
{
738
	struct tipc_msg *msg = buf_msg(skb_peek(list));
J
Jon Paul Maloy 已提交
739
	unsigned int maxwin = link->window;
740
	unsigned int i, imp = msg_importance(msg);
741
	uint mtu = link->mtu;
742 743
	u16 ack = mod(link->rcv_nxt - 1);
	u16 seqno = link->snd_nxt;
744
	u16 bc_ack = link->bc_rcvlink->rcv_nxt - 1;
J
Jon Paul Maloy 已提交
745
	struct tipc_media_addr *addr = link->media_addr;
J
Jon Paul Maloy 已提交
746 747
	struct sk_buff_head *transmq = &link->transmq;
	struct sk_buff_head *backlogq = &link->backlogq;
748
	struct sk_buff *skb, *bskb;
749

750 751 752 753 754
	/* 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);
	}
755
	if (unlikely(msg_size(msg) > mtu))
756
		return -EMSGSIZE;
757

J
Jon Paul Maloy 已提交
758
	/* Prepare each packet for sending, and add to relevant queue: */
759 760
	while (skb_queue_len(list)) {
		skb = skb_peek(list);
761
		msg = buf_msg(skb);
J
Jon Paul Maloy 已提交
762 763
		msg_set_seqno(msg, seqno);
		msg_set_ack(msg, ack);
764
		msg_set_bcast_ack(msg, bc_ack);
765

J
Jon Paul Maloy 已提交
766
		if (likely(skb_queue_len(transmq) < maxwin)) {
767
			__skb_dequeue(list);
J
Jon Paul Maloy 已提交
768 769 770 771 772 773
			__skb_queue_tail(transmq, skb);
			tipc_bearer_send(net, link->bearer_id, skb, addr);
			link->rcv_unacked = 0;
			seqno++;
			continue;
		}
774 775
		if (tipc_msg_bundle(skb_peek_tail(backlogq), msg, mtu)) {
			kfree_skb(__skb_dequeue(list));
776 777
			link->stats.sent_bundled++;
			continue;
J
Jon Paul Maloy 已提交
778
		}
779 780 781 782
		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++;
783 784
			link->stats.sent_bundled++;
			link->stats.sent_bundles++;
785
			continue;
786
		}
787 788
		link->backlog[imp].len += skb_queue_len(list);
		skb_queue_splice_tail_init(list, backlogq);
789
	}
790
	link->snd_nxt = seqno;
791 792 793
	return 0;
}

794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813
/**
 * 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;
814
	u16 bc_ack = l->bc_rcvlink->rcv_nxt - 1;
815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832
	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);
833
		msg_set_bcast_ack(hdr, bc_ack);
834 835 836 837 838 839 840 841

		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);
842
			TIPC_SKB_CB(skb)->ackers = l->ackers;
843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866
			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;
}

867
/*
868 869 870 871 872 873
 * 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
874
 */
J
Jon Paul Maloy 已提交
875
void tipc_link_push_packets(struct tipc_link *link)
P
Per Liden 已提交
876
{
J
Jon Paul Maloy 已提交
877
	struct sk_buff *skb;
878
	struct tipc_msg *msg;
879
	u16 seqno = link->snd_nxt;
880
	u16 ack = mod(link->rcv_nxt - 1);
P
Per Liden 已提交
881

J
Jon Paul Maloy 已提交
882 883 884
	while (skb_queue_len(&link->transmq) < link->window) {
		skb = __skb_dequeue(&link->backlogq);
		if (!skb)
885
			break;
886
		TIPC_SKB_CB(skb)->ackers = link->ackers;
J
Jon Paul Maloy 已提交
887
		msg = buf_msg(skb);
888
		link->backlog[msg_importance(msg)].len--;
J
Jon Paul Maloy 已提交
889
		msg_set_ack(msg, ack);
890 891
		msg_set_seqno(msg, seqno);
		seqno = mod(seqno + 1);
J
Jon Paul Maloy 已提交
892 893 894 895
		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
896
				 skb, link->media_addr);
P
Per Liden 已提交
897
	}
898
	link->snd_nxt = seqno;
P
Per Liden 已提交
899 900
}

901 902 903 904 905 906
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;
907
	u16 bc_ack = l->bc_rcvlink->rcv_nxt - 1;
908 909 910 911 912 913 914 915 916 917 918 919 920

	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);
921
		TIPC_SKB_CB(skb)->ackers = l->ackers;
922
		msg_set_seqno(hdr, seqno);
923 924
		msg_set_ack(hdr, ack);
		msg_set_bcast_ack(hdr, bc_ack);
925 926 927 928 929 930
		l->rcv_unacked = 0;
		seqno++;
	}
	l->snd_nxt = seqno;
}

931
static void link_retransmit_failure(struct tipc_link *l, struct sk_buff *skb)
932
{
933 934 935 936 937 938 939 940
	struct tipc_msg *hdr = buf_msg(skb);

	pr_warn("Retransmission failure on link <%s>\n", l->name);
	link_print(l, "Resetting link ");
	pr_info("Failed msg: usr %u, typ %u, len %u, err %u\n",
		msg_user(hdr), msg_type(hdr), msg_size(hdr), msg_errcode(hdr));
	pr_info("sqno %u, prev: %x, src: %x\n",
		msg_seqno(hdr), msg_prevnode(hdr), msg_orignode(hdr));
941 942
}

943
void tipc_link_retransmit(struct tipc_link *l_ptr, struct sk_buff *skb,
944
			  u32 retransmits)
P
Per Liden 已提交
945 946 947
{
	struct tipc_msg *msg;

948
	if (!skb)
949 950
		return;

951
	msg = buf_msg(skb);
952

953
	/* Detect repeated retransmit failures */
954
	if (l_ptr->last_retransm == msg_seqno(msg)) {
955
		if (++l_ptr->stale_count > 100) {
956
			link_retransmit_failure(l_ptr, skb);
957
			return;
958 959
		}
	} else {
960
		l_ptr->last_retransm = msg_seqno(msg);
961
		l_ptr->stale_count = 1;
P
Per Liden 已提交
962
	}
963

J
Jon Paul Maloy 已提交
964 965
	skb_queue_walk_from(&l_ptr->transmq, skb) {
		if (!retransmits)
966 967
			break;
		msg = buf_msg(skb);
968
		msg_set_ack(msg, mod(l_ptr->rcv_nxt - 1));
969
		msg_set_bcast_ack(msg, l_ptr->owner->bclink.last_in);
970
		tipc_bearer_send(l_ptr->owner->net, l_ptr->bearer_id, skb,
J
Jon Paul Maloy 已提交
971
				 l_ptr->media_addr);
972 973
		retransmits--;
		l_ptr->stats.retransmitted++;
P
Per Liden 已提交
974 975 976
	}
}

977 978
int tipc_link_retrans(struct tipc_link *l, u16 from, u16 to,
		      struct sk_buff_head *xmitq)
979 980 981
{
	struct sk_buff *_skb, *skb = skb_peek(&l->transmq);
	struct tipc_msg *hdr;
982
	u16 ack = l->rcv_nxt - 1;
983
	u16 bc_ack = l->bc_rcvlink->rcv_nxt - 1;
984 985 986 987 988 989 990 991 992 993

	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);
994
		return tipc_link_fsm_evt(l, LINK_FAILURE_EVT);
995
	}
996 997

	/* Move forward to where retransmission should start */
998
	skb_queue_walk(&l->transmq, skb) {
999 1000 1001 1002 1003 1004 1005
		if (!less(buf_seqno(skb), from))
			break;
	}

	skb_queue_walk_from(&l->transmq, skb) {
		if (more(buf_seqno(skb), to))
			break;
1006 1007 1008 1009 1010
		hdr = buf_msg(skb);
		_skb = __pskb_copy(skb, MIN_H_SIZE, GFP_ATOMIC);
		if (!_skb)
			return 0;
		hdr = buf_msg(_skb);
1011 1012
		msg_set_ack(hdr, ack);
		msg_set_bcast_ack(hdr, bc_ack);
1013 1014 1015 1016 1017 1018 1019
		_skb->priority = TC_PRIO_CONTROL;
		__skb_queue_tail(xmitq, _skb);
		l->stats.retransmitted++;
	}
	return 0;
}

1020
/* tipc_data_input - deliver data and name distr msgs to upper layer
1021
 *
1022
 * Consumes buffer if message is of right type
1023 1024
 * Node lock must be held
 */
1025
static bool tipc_data_input(struct tipc_link *l, struct sk_buff *skb,
1026
			    struct sk_buff_head *inputq)
1027
{
1028
	switch (msg_user(buf_msg(skb))) {
1029 1030 1031 1032 1033
	case TIPC_LOW_IMPORTANCE:
	case TIPC_MEDIUM_IMPORTANCE:
	case TIPC_HIGH_IMPORTANCE:
	case TIPC_CRITICAL_IMPORTANCE:
	case CONN_MANAGER:
1034
		skb_queue_tail(inputq, skb);
1035
		return true;
1036
	case NAME_DISTRIBUTOR:
1037 1038
		l->bc_rcvlink->state = LINK_ESTABLISHED;
		skb_queue_tail(l->namedq, skb);
1039 1040
		return true;
	case MSG_BUNDLER:
1041
	case TUNNEL_PROTOCOL:
1042
	case MSG_FRAGMENTER:
1043
	case BCAST_PROTOCOL:
1044
		return false;
1045
	default:
1046 1047 1048 1049
		pr_warn("Dropping received illegal msg type\n");
		kfree_skb(skb);
		return false;
	};
1050
}
1051 1052 1053 1054

/* tipc_link_input - process packet that has passed link protocol check
 *
 * Consumes buffer
1055
 */
1056 1057
static int tipc_link_input(struct tipc_link *l, struct sk_buff *skb,
			   struct sk_buff_head *inputq)
1058
{
1059 1060
	struct tipc_msg *hdr = buf_msg(skb);
	struct sk_buff **reasm_skb = &l->reasm_buf;
1061
	struct sk_buff *iskb;
1062
	struct sk_buff_head tmpq;
1063
	int usr = msg_user(hdr);
1064
	int rc = 0;
1065 1066
	int pos = 0;
	int ipos = 0;
1067

1068 1069 1070 1071
	if (unlikely(usr == TUNNEL_PROTOCOL)) {
		if (msg_type(hdr) == SYNCH_MSG) {
			__skb_queue_purge(&l->deferdq);
			goto drop;
1072
		}
1073 1074 1075 1076 1077 1078 1079
		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;
1080
		if (tipc_data_input(l, skb, inputq))
1081 1082 1083 1084
			return rc;
		usr = msg_user(hdr);
		reasm_skb = &l->failover_reasm_skb;
	}
1085

1086
	if (usr == MSG_BUNDLER) {
1087
		skb_queue_head_init(&tmpq);
1088 1089
		l->stats.recv_bundles++;
		l->stats.recv_bundled += msg_msgcnt(hdr);
1090
		while (tipc_msg_extract(skb, &iskb, &pos))
1091 1092
			tipc_data_input(l, iskb, &tmpq);
		tipc_skb_queue_splice_tail(&tmpq, inputq);
1093
		return 0;
1094 1095 1096 1097
	} else if (usr == MSG_FRAGMENTER) {
		l->stats.recv_fragments++;
		if (tipc_buf_append(reasm_skb, &skb)) {
			l->stats.recv_fragmented++;
1098
			tipc_data_input(l, skb, inputq);
1099 1100
		} else if (!*reasm_skb && !link_is_bc_rcvlink(l)) {
			pr_warn_ratelimited("Unable to build fragment list\n");
1101
			return tipc_link_fsm_evt(l, LINK_FAILURE_EVT);
1102
		}
1103
		return 0;
1104
	} else if (usr == BCAST_PROTOCOL) {
1105 1106 1107
		tipc_bcast_lock(l->owner->net);
		tipc_link_bc_init_rcv(l->bc_rcvlink, hdr);
		tipc_bcast_unlock(l->owner->net);
1108 1109 1110
	}
drop:
	kfree_skb(skb);
1111
	return 0;
1112 1113
}

1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
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;
}

1129
/* tipc_link_build_ack_msg: prepare link acknowledge message for transmission
1130 1131 1132
 *
 * Note that sending of broadcast ack is coordinated among nodes, to reduce
 * risk of ack storms towards the sender
1133
 */
1134
int tipc_link_build_ack_msg(struct tipc_link *l, struct sk_buff_head *xmitq)
1135
{
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147
	if (!l)
		return 0;

	/* Broadcast ACK must be sent via a unicast link => defer to caller */
	if (link_is_bc_rcvlink(l)) {
		if (((l->rcv_nxt ^ link_own_addr(l)) & 0xf) != 0xf)
			return 0;
		l->rcv_unacked = 0;
		return TIPC_LINK_SND_BC_ACK;
	}

	/* Unicast ACK */
1148 1149 1150
	l->rcv_unacked = 0;
	l->stats.sent_acks++;
	tipc_link_build_proto_msg(l, STATE_MSG, 0, 0, 0, 0, xmitq);
1151
	return 0;
1152 1153
}

1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165
/* 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);
}

1166 1167 1168 1169 1170 1171 1172
/* 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;

1173 1174 1175
	if (link_is_bc_rcvlink(l))
		return;

1176 1177 1178 1179
	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);
}

1180
/* tipc_link_rcv - process TIPC packets/messages arriving from off-node
1181
 * @l: the link that should handle the message
1182 1183 1184 1185 1186 1187
 * @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)
{
1188
	struct sk_buff_head *defq = &l->deferdq;
1189
	struct tipc_msg *hdr;
1190
	u16 seqno, rcv_nxt, win_lim;
1191 1192
	int rc = 0;

1193
	do {
1194
		hdr = buf_msg(skb);
1195 1196
		seqno = msg_seqno(hdr);
		rcv_nxt = l->rcv_nxt;
1197
		win_lim = rcv_nxt + TIPC_MAX_LINK_WIN;
1198 1199

		/* Verify and update link state */
1200 1201
		if (unlikely(msg_user(hdr) == LINK_PROTOCOL))
			return tipc_link_proto_rcv(l, skb, xmitq);
1202

1203
		if (unlikely(!link_is_up(l))) {
1204 1205 1206
			if (l->state == LINK_ESTABLISHING)
				rc = TIPC_LINK_UP_EVT;
			goto drop;
1207 1208
		}

1209
		/* Don't send probe at next timeout expiration */
1210 1211
		l->silent_intv_cnt = 0;

1212 1213 1214 1215 1216 1217
		/* Drop if outside receive window */
		if (unlikely(less(seqno, rcv_nxt) || more(seqno, win_lim))) {
			l->stats.duplicates++;
			goto drop;
		}

1218 1219 1220 1221 1222 1223 1224
		/* 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);
		}

1225 1226
		/* Defer delivery if sequence gap */
		if (unlikely(seqno != rcv_nxt)) {
1227
			__tipc_skb_queue_sorted(defq, seqno, skb);
1228 1229
			tipc_link_build_nack_msg(l, xmitq);
			break;
1230 1231
		}

1232
		/* Deliver packet */
1233 1234
		l->rcv_nxt++;
		l->stats.recv_info++;
1235
		if (!tipc_data_input(l, skb, l->inputq))
1236
			rc |= tipc_link_input(l, skb, l->inputq);
1237
		if (unlikely(++l->rcv_unacked >= TIPC_MIN_LINK_WIN))
1238 1239 1240
			rc |= tipc_link_build_ack_msg(l, xmitq);
		if (unlikely(rc & ~TIPC_LINK_SND_BC_ACK))
			break;
1241 1242 1243 1244 1245
	} while ((skb = __skb_dequeue(defq)));

	return rc;
drop:
	kfree_skb(skb);
1246 1247 1248
	return rc;
}

1249
/**
1250 1251 1252
 * 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
1253
 */
1254
u32 tipc_link_defer_pkt(struct sk_buff_head *list, struct sk_buff *skb)
P
Per Liden 已提交
1255
{
1256
	struct sk_buff *skb1;
1257
	u16 seq_no = buf_seqno(skb);
P
Per Liden 已提交
1258 1259

	/* Empty queue ? */
1260 1261
	if (skb_queue_empty(list)) {
		__skb_queue_tail(list, skb);
P
Per Liden 已提交
1262 1263 1264 1265
		return 1;
	}

	/* Last ? */
1266 1267
	if (less(buf_seqno(skb_peek_tail(list)), seq_no)) {
		__skb_queue_tail(list, skb);
P
Per Liden 已提交
1268 1269 1270
		return 1;
	}

1271
	/* Locate insertion point in queue, then insert; discard if duplicate */
1272
	skb_queue_walk(list, skb1) {
1273
		u16 curr_seqno = buf_seqno(skb1);
P
Per Liden 已提交
1274

1275
		if (seq_no == curr_seqno) {
1276
			kfree_skb(skb);
1277
			return 0;
P
Per Liden 已提交
1278
		}
1279 1280

		if (less(seq_no, curr_seqno))
P
Per Liden 已提交
1281
			break;
1282
	}
P
Per Liden 已提交
1283

1284
	__skb_queue_before(list, skb1, skb);
1285
	return 1;
P
Per Liden 已提交
1286 1287 1288 1289 1290
}

/*
 * Send protocol message to the other endpoint.
 */
1291
void tipc_link_proto_xmit(struct tipc_link *l, u32 msg_typ, int probe_msg,
1292
			  u32 gap, u32 tolerance, u32 priority)
P
Per Liden 已提交
1293
{
1294 1295
	struct sk_buff *skb = NULL;
	struct sk_buff_head xmitq;
P
Per Liden 已提交
1296

1297 1298 1299 1300 1301
	__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
1302
		return;
J
Jon Paul Maloy 已提交
1303
	tipc_bearer_send(l->owner->net, l->bearer_id, skb, l->media_addr);
1304 1305
	l->rcv_unacked = 0;
	kfree_skb(skb);
P
Per Liden 已提交
1306 1307
}

1308 1309 1310 1311 1312 1313
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;
1314
	bool node_up = link_is_up(l->bc_rcvlink);
1315 1316

	/* Don't send protocol message during reset or link failover */
1317
	if (tipc_link_is_blocked(l))
1318 1319 1320 1321
		return;

	msg_set_type(hdr, mtyp);
	msg_set_net_plane(hdr, l->net_plane);
1322 1323 1324 1325
	msg_set_next_sent(hdr, l->snd_nxt);
	msg_set_ack(hdr, l->rcv_nxt - 1);
	msg_set_bcast_ack(hdr, l->bc_rcvlink->rcv_nxt - 1);
	msg_set_last_bcast(hdr, l->bc_sndlink->snd_nxt - 1);
1326 1327 1328 1329 1330 1331
	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 */
1332
	msg_set_seqno(hdr, l->snd_nxt + U16_MAX / 2);
1333 1334 1335 1336 1337 1338 1339

	if (mtyp == STATE_MSG) {
		if (!tipc_link_is_up(l))
			return;

		/* Override rcvgap if there are packets in deferred queue */
		if (!skb_queue_empty(&l->deferdq))
1340
			rcvgap = buf_seqno(skb_peek(&l->deferdq)) - l->rcv_nxt;
1341 1342 1343 1344 1345 1346 1347 1348
		if (rcvgap) {
			msg_set_seq_gap(hdr, rcvgap);
			l->stats.sent_nacks++;
		}
		msg_set_probe(hdr, probe);
		if (probe)
			l->stats.sent_probes++;
		l->stats.sent_states++;
1349
		l->rcv_unacked = 0;
1350 1351 1352
	} else {
		/* RESET_MSG or ACTIVATE_MSG */
		msg_set_max_pkt(hdr, l->advertised_mtu);
1353
		msg_set_ack(hdr, l->rcv_nxt - 1);
1354 1355 1356 1357 1358 1359 1360
		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;
1361
	__skb_queue_tail(xmitq, skb);
P
Per Liden 已提交
1362 1363
}

1364
/* tipc_link_tnl_prepare(): prepare and return a list of tunnel packets
1365
 * with contents of the link's transmit and backlog queues.
P
Per Liden 已提交
1366
 */
1367 1368
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
1369
{
1370 1371 1372 1373 1374
	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
1375

1376
	if (!tnl)
P
Per Liden 已提交
1377 1378
		return;

1379 1380
	skb_queue_head_init(&tnlq);
	skb_queue_head_init(&tmpxq);
1381

1382 1383 1384 1385 1386 1387
	/* 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
1388 1389
		return;
	}
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
	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
1402
	skb_queue_walk(queue, skb) {
1403 1404 1405 1406 1407 1408 1409 1410
		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
1411 1412
			return;
		}
1413 1414 1415
		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
1416
	}
1417 1418 1419
	if (queue != &l->backlogq) {
		queue = &l->backlogq;
		goto tnl;
1420
	}
1421

1422
	tipc_link_xmit(tnl, &tnlq, xmitq);
P
Per Liden 已提交
1423

1424 1425 1426 1427
	if (mtyp == FAILOVER_MSG) {
		tnl->drop_point = l->rcv_nxt;
		tnl->failover_reasm_skb = l->reasm_buf;
		l->reasm_buf = NULL;
1428
	}
P
Per Liden 已提交
1429 1430
}

1431 1432 1433 1434 1435 1436 1437 1438 1439 1440
/* 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;
1441 1442
	u16 ack = msg_ack(hdr);
	u16 gap = msg_seq_gap(hdr);
1443 1444 1445
	u16 peers_snd_nxt =  msg_next_sent(hdr);
	u16 peers_tol = msg_link_tolerance(hdr);
	u16 peers_prio = msg_linkprio(hdr);
1446
	u16 rcv_nxt = l->rcv_nxt;
1447
	int mtyp = msg_type(hdr);
1448 1449 1450
	char *if_name;
	int rc = 0;

1451
	if (tipc_link_is_blocked(l) || !xmitq)
1452 1453 1454 1455 1456
		goto exit;

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

1457
	switch (mtyp) {
1458 1459 1460 1461 1462 1463 1464
	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' */
1465

1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483
	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;

1484 1485 1486 1487 1488 1489 1490 1491
		/* 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;

1492 1493 1494 1495 1496
		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;
1497

1498
	case STATE_MSG:
1499

1500 1501 1502 1503 1504 1505 1506 1507
		/* 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++;
1508 1509 1510 1511

		if (!link_is_up(l)) {
			if (l->state == LINK_ESTABLISHING)
				rc = TIPC_LINK_UP_EVT;
1512
			break;
1513
		}
1514

1515
		/* Send NACK if peer has sent pkts we haven't received yet */
1516
		if (more(peers_snd_nxt, rcv_nxt) && !tipc_link_is_synching(l))
1517 1518 1519
			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
1520
						  0, 0, xmitq);
1521
		tipc_link_release_pkts(l, ack);
1522 1523

		/* If NACK, retransmit will now start at right position */
1524 1525
		if (gap) {
			rc = tipc_link_retrans(l, ack + 1, ack + gap, xmitq);
1526 1527
			l->stats.recv_nacks++;
		}
1528

1529 1530 1531 1532 1533 1534 1535 1536 1537
		tipc_link_advance_backlog(l, xmitq);
		if (unlikely(!skb_queue_empty(&l->wakeupq)))
			link_prepare_wakeup(l);
	}
exit:
	kfree_skb(skb);
	return rc;
}

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 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 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 1715 1716 1717 1718 1719
/* tipc_link_build_bc_proto_msg() - create broadcast protocol message
 */
static bool tipc_link_build_bc_proto_msg(struct tipc_link *l, bool bcast,
					 u16 peers_snd_nxt,
					 struct sk_buff_head *xmitq)
{
	struct sk_buff *skb;
	struct tipc_msg *hdr;
	struct sk_buff *dfrd_skb = skb_peek(&l->deferdq);
	u16 ack = l->rcv_nxt - 1;
	u16 gap_to = peers_snd_nxt - 1;

	skb = tipc_msg_create(BCAST_PROTOCOL, STATE_MSG, INT_H_SIZE,
			      0, l->addr, link_own_addr(l), 0, 0, 0);
	if (!skb)
		return false;
	hdr = buf_msg(skb);
	msg_set_last_bcast(hdr, l->bc_sndlink->snd_nxt - 1);
	msg_set_bcast_ack(hdr, ack);
	msg_set_bcgap_after(hdr, ack);
	if (dfrd_skb)
		gap_to = buf_seqno(dfrd_skb) - 1;
	msg_set_bcgap_to(hdr, gap_to);
	msg_set_non_seq(hdr, bcast);
	__skb_queue_tail(xmitq, skb);
	return true;
}

/* tipc_link_build_bc_init_msg() - synchronize broadcast link endpoints.
 *
 * Give a newly added peer node the sequence number where it should
 * start receiving and acking broadcast packets.
 */
void tipc_link_build_bc_init_msg(struct tipc_link *l,
				 struct sk_buff_head *xmitq)
{
	struct sk_buff_head list;

	__skb_queue_head_init(&list);
	if (!tipc_link_build_bc_proto_msg(l->bc_rcvlink, false, 0, &list))
		return;
	tipc_link_xmit(l, &list, xmitq);
}

/* tipc_link_bc_init_rcv - receive initial broadcast synch data from peer
 */
void tipc_link_bc_init_rcv(struct tipc_link *l, struct tipc_msg *hdr)
{
	int mtyp = msg_type(hdr);
	u16 peers_snd_nxt = msg_bc_snd_nxt(hdr);

	if (link_is_up(l))
		return;

	if (msg_user(hdr) == BCAST_PROTOCOL) {
		l->rcv_nxt = peers_snd_nxt;
		l->state = LINK_ESTABLISHED;
		return;
	}

	if (l->peer_caps & TIPC_BCAST_SYNCH)
		return;

	if (msg_peer_node_is_up(hdr))
		return;

	/* Compatibility: accept older, less safe initial synch data */
	if ((mtyp == RESET_MSG) || (mtyp == ACTIVATE_MSG))
		l->rcv_nxt = peers_snd_nxt;
}

/* tipc_link_bc_sync_rcv - update rcv link according to peer's send state
 */
void tipc_link_bc_sync_rcv(struct tipc_link *l, struct tipc_msg *hdr,
			   struct sk_buff_head *xmitq)
{
	u16 peers_snd_nxt = msg_bc_snd_nxt(hdr);

	if (!link_is_up(l))
		return;

	if (!msg_peer_node_is_up(hdr))
		return;

	l->bc_peer_is_up = true;

	/* Ignore if peers_snd_nxt goes beyond receive window */
	if (more(peers_snd_nxt, l->rcv_nxt + l->window))
		return;

	if (!more(peers_snd_nxt, l->rcv_nxt)) {
		l->nack_state = BC_NACK_SND_CONDITIONAL;
		return;
	}

	/* Don't NACK if one was recently sent or peeked */
	if (l->nack_state == BC_NACK_SND_SUPPRESS) {
		l->nack_state = BC_NACK_SND_UNCONDITIONAL;
		return;
	}

	/* Conditionally delay NACK sending until next synch rcv */
	if (l->nack_state == BC_NACK_SND_CONDITIONAL) {
		l->nack_state = BC_NACK_SND_UNCONDITIONAL;
		if ((peers_snd_nxt - l->rcv_nxt) < TIPC_MIN_LINK_WIN)
			return;
	}

	/* Send NACK now but suppress next one */
	tipc_link_build_bc_proto_msg(l, true, peers_snd_nxt, xmitq);
	l->nack_state = BC_NACK_SND_SUPPRESS;
}

void tipc_link_bc_ack_rcv(struct tipc_link *l, u16 acked,
			  struct sk_buff_head *xmitq)
{
	struct sk_buff *skb, *tmp;
	struct tipc_link *snd_l = l->bc_sndlink;

	if (!link_is_up(l) || !l->bc_peer_is_up)
		return;

	if (!more(acked, l->acked))
		return;

	/* Skip over packets peer has already acked */
	skb_queue_walk(&snd_l->transmq, skb) {
		if (more(buf_seqno(skb), l->acked))
			break;
	}

	/* Update/release the packets peer is acking now */
	skb_queue_walk_from_safe(&snd_l->transmq, skb, tmp) {
		if (more(buf_seqno(skb), acked))
			break;
		if (!--TIPC_SKB_CB(skb)->ackers) {
			__skb_unlink(skb, &snd_l->transmq);
			kfree_skb(skb);
		}
	}
	l->acked = acked;
	tipc_link_advance_backlog(snd_l, xmitq);
	if (unlikely(!skb_queue_empty(&snd_l->wakeupq)))
		link_prepare_wakeup(snd_l);
}

/* tipc_link_bc_nack_rcv(): receive broadcast nack message
 */
int tipc_link_bc_nack_rcv(struct tipc_link *l, struct sk_buff *skb,
			  struct sk_buff_head *xmitq)
{
	struct tipc_msg *hdr = buf_msg(skb);
	u32 dnode = msg_destnode(hdr);
	int mtyp = msg_type(hdr);
	u16 acked = msg_bcast_ack(hdr);
	u16 from = acked + 1;
	u16 to = msg_bcgap_to(hdr);
	u16 peers_snd_nxt = to + 1;
	int rc = 0;

	kfree_skb(skb);

	if (!tipc_link_is_up(l) || !l->bc_peer_is_up)
		return 0;

	if (mtyp != STATE_MSG)
		return 0;

	if (dnode == link_own_addr(l)) {
		tipc_link_bc_ack_rcv(l, acked, xmitq);
		rc = tipc_link_retrans(l->bc_sndlink, from, to, xmitq);
		l->stats.recv_nacks++;
		return rc;
	}

	/* Msg for other node => suppress own NACK at next sync if applicable */
	if (more(peers_snd_nxt, l->rcv_nxt) && !less(l->rcv_nxt, from))
		l->nack_state = BC_NACK_SND_SUPPRESS;

	return 0;
}

1720
void tipc_link_set_queue_limits(struct tipc_link *l, u32 win)
P
Per Liden 已提交
1721
{
1722
	int max_bulk = TIPC_MAX_PUBLICATIONS / (l->mtu / ITEM_SIZE);
1723 1724

	l->window = win;
1725 1726 1727 1728 1729
	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
1730 1731
}

1732
/* tipc_link_find_owner - locate owner node of link by link's name
1733
 * @net: the applicable net namespace
1734 1735
 * @name: pointer to link name string
 * @bearer_id: pointer to index in 'node->links' array where the link was found.
1736
 *
1737
 * Returns pointer to node owning the link, or 0 if no matching link is found.
P
Per Liden 已提交
1738
 */
1739 1740
static struct tipc_node *tipc_link_find_owner(struct net *net,
					      const char *link_name,
1741
					      unsigned int *bearer_id)
P
Per Liden 已提交
1742
{
1743
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1744
	struct tipc_link *l_ptr;
1745
	struct tipc_node *n_ptr;
1746
	struct tipc_node *found_node = NULL;
1747
	int i;
P
Per Liden 已提交
1748

1749
	*bearer_id = 0;
1750
	rcu_read_lock();
1751
	list_for_each_entry_rcu(n_ptr, &tn->node_list, list) {
1752
		tipc_node_lock(n_ptr);
1753
		for (i = 0; i < MAX_BEARERS; i++) {
1754
			l_ptr = n_ptr->links[i].link;
1755 1756 1757 1758 1759
			if (l_ptr && !strcmp(l_ptr->name, link_name)) {
				*bearer_id = i;
				found_node = n_ptr;
				break;
			}
1760
		}
1761
		tipc_node_unlock(n_ptr);
1762 1763
		if (found_node)
			break;
1764
	}
1765 1766
	rcu_read_unlock();

1767
	return found_node;
P
Per Liden 已提交
1768 1769 1770 1771 1772 1773
}

/**
 * link_reset_statistics - reset link statistics
 * @l_ptr: pointer to link
 */
1774
static void link_reset_statistics(struct tipc_link *l_ptr)
P
Per Liden 已提交
1775 1776
{
	memset(&l_ptr->stats, 0, sizeof(l_ptr->stats));
1777 1778
	l_ptr->stats.sent_info = l_ptr->snd_nxt;
	l_ptr->stats.recv_info = l_ptr->rcv_nxt;
P
Per Liden 已提交
1779 1780
}

1781
static void link_print(struct tipc_link *l, const char *str)
P
Per Liden 已提交
1782
{
1783
	struct sk_buff *hskb = skb_peek(&l->transmq);
1784
	u16 head = hskb ? msg_seqno(buf_msg(hskb)) : l->snd_nxt - 1;
1785
	u16 tail = l->snd_nxt - 1;
1786

1787
	pr_info("%s Link <%s> state %x\n", str, l->name, l->state);
1788 1789 1790
	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
1791
}
1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829

/* 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;
}
1830

1831 1832 1833 1834 1835 1836 1837 1838 1839
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];
1840
	struct net *net = sock_net(skb->sk);
1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855

	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]);

1856 1857 1858
	if (strcmp(name, tipc_bclink_name) == 0)
		return tipc_nl_bc_link_set(net, attrs);

1859
	node = tipc_link_find_owner(net, name, &bearer_id);
1860 1861 1862 1863 1864
	if (!node)
		return -EINVAL;

	tipc_node_lock(node);

1865
	link = node->links[bearer_id].link;
1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884
	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]);
1885
			link->tolerance = tol;
1886
			tipc_link_proto_xmit(link, STATE_MSG, 0, 0, tol, 0);
1887 1888 1889 1890 1891 1892
		}
		if (props[TIPC_NLA_PROP_PRIO]) {
			u32 prio;

			prio = nla_get_u32(props[TIPC_NLA_PROP_PRIO]);
			link->priority = prio;
1893
			tipc_link_proto_xmit(link, STATE_MSG, 0, 0, 0, prio);
1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907
		}
		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;
}
1908 1909

static int __tipc_nl_add_stats(struct sk_buff *skb, struct tipc_stats *s)
1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974
{
	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 */
1975
static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg,
1976
			      struct tipc_link *link, int nlflags)
1977 1978 1979 1980 1981
{
	int err;
	void *hdr;
	struct nlattr *attrs;
	struct nlattr *prop;
1982
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1983

1984
	hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
1985
			  nlflags, TIPC_NL_LINK_GET);
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
	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,
1996
			tipc_cluster_mask(tn->own_addr)))
1997
		goto attr_msg_full;
1998
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_MTU, link->mtu))
1999
		goto attr_msg_full;
2000
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, link->rcv_nxt))
2001
		goto attr_msg_full;
2002
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, link->snd_nxt))
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
		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,
2020
			link->window))
2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045
		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  */
2046 2047
static int __tipc_nl_add_node_links(struct net *net, struct tipc_nl_msg *msg,
				    struct tipc_node *node, u32 *prev_link)
2048 2049 2050 2051 2052 2053 2054
{
	u32 i;
	int err;

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

2055
		if (!node->links[i].link)
2056 2057
			continue;

2058 2059
		err = __tipc_nl_add_link(net, msg,
					 node->links[i].link, NLM_F_MULTI);
2060 2061 2062 2063 2064 2065 2066 2067 2068 2069
		if (err)
			return err;
	}
	*prev_link = 0;

	return 0;
}

int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
2070 2071
	struct net *net = sock_net(skb->sk);
	struct tipc_net *tn = net_generic(net, tipc_net_id);
2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
	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) {
2088
		node = tipc_node_find(net, prev_node);
2089 2090 2091 2092 2093 2094 2095 2096 2097 2098
		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;
		}
2099
		tipc_node_put(node);
2100

2101 2102
		list_for_each_entry_continue_rcu(node, &tn->node_list,
						 list) {
2103
			tipc_node_lock(node);
2104 2105
			err = __tipc_nl_add_node_links(net, &msg, node,
						       &prev_link);
2106 2107 2108 2109 2110 2111 2112
			tipc_node_unlock(node);
			if (err)
				goto out;

			prev_node = node->addr;
		}
	} else {
2113
		err = tipc_nl_add_bc_link(net, &msg);
2114 2115 2116
		if (err)
			goto out;

2117
		list_for_each_entry_rcu(node, &tn->node_list, list) {
2118
			tipc_node_lock(node);
2119 2120
			err = __tipc_nl_add_node_links(net, &msg, node,
						       &prev_link);
2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140
			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)
{
2141
	struct net *net = genl_info_net(info);
2142 2143 2144 2145
	struct tipc_nl_msg msg;
	char *name;
	int err;

2146 2147 2148
	msg.portid = info->snd_portid;
	msg.seq = info->snd_seq;

2149 2150 2151 2152
	if (!info->attrs[TIPC_NLA_LINK_NAME])
		return -EINVAL;
	name = nla_data(info->attrs[TIPC_NLA_LINK_NAME]);

2153 2154
	msg.skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
	if (!msg.skb)
2155 2156
		return -ENOMEM;

2157 2158 2159 2160 2161 2162 2163 2164 2165 2166
	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;
2167

2168 2169 2170
		node = tipc_link_find_owner(net, name, &bearer_id);
		if (!node)
			return -EINVAL;
2171

2172
		tipc_node_lock(node);
2173
		link = node->links[bearer_id].link;
2174 2175 2176 2177 2178
		if (!link) {
			tipc_node_unlock(node);
			nlmsg_free(msg.skb);
			return -EINVAL;
		}
2179

2180 2181 2182 2183 2184 2185 2186
		err = __tipc_nl_add_link(net, &msg, link, 0);
		tipc_node_unlock(node);
		if (err) {
			nlmsg_free(msg.skb);
			return err;
		}
	}
2187

2188
	return genlmsg_reply(msg.skb, info);
2189
}
2190 2191 2192 2193 2194 2195 2196 2197 2198

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];
2199
	struct net *net = sock_net(skb->sk);
2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215

	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) {
2216
		err = tipc_bclink_reset_stats(net);
2217 2218 2219 2220 2221
		if (err)
			return err;
		return 0;
	}

2222
	node = tipc_link_find_owner(net, link_name, &bearer_id);
2223 2224 2225 2226 2227
	if (!node)
		return -EINVAL;

	tipc_node_lock(node);

2228
	link = node->links[bearer_id].link;
2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239
	if (!link) {
		tipc_node_unlock(node);
		return -EINVAL;
	}

	link_reset_statistics(link);

	tipc_node_unlock(node);

	return 0;
}