link.c 56.8 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 51 52 53
/*
 * Error message prefixes
 */
static const char *link_co_err = "Link changeover error, ";
static const char *link_rst_msg = "Resetting link ";
static const char *link_unk_evt = "Unknown link event ";
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
/*
 * Interval between NACKs when packets arrive out of order
 */
#define TIPC_NACK_INTV (TIPC_MIN_LINK_WIN * 2)
83 84 85
/*
 * Out-of-range value for link session numbers
 */
86
#define WILDCARD_SESSION 0x10000
87

88
/* State value stored in 'failover_pkts'
P
Per Liden 已提交
89
 */
90
#define FIRST_FAILOVER 0xffffu
P
Per Liden 已提交
91

92
/* Link FSM states and events:
P
Per Liden 已提交
93
 */
94
enum {
95 96 97 98
	TIPC_LINK_WORKING,
	TIPC_LINK_PROBING,
	TIPC_LINK_RESETTING,
	TIPC_LINK_ESTABLISHING
99 100 101 102 103 104 105 106 107 108 109
};

enum {
	PEER_RESET_EVT    = RESET_MSG,
	ACTIVATE_EVT      = ACTIVATE_MSG,
	TRAFFIC_EVT,      /* Any other valid msg from peer */
	SILENCE_EVT       /* Peer was silent during last timer interval*/
};

/* Link FSM state checking routines
 */
110
static int link_working(struct tipc_link *l)
111
{
112
	return l->state == TIPC_LINK_WORKING;
113 114
}

115
static int link_probing(struct tipc_link *l)
116
{
117
	return l->state == TIPC_LINK_PROBING;
118 119
}

120
static int link_resetting(struct tipc_link *l)
121
{
122
	return l->state == TIPC_LINK_RESETTING;
123 124
}

125
static int link_establishing(struct tipc_link *l)
126
{
127
	return l->state == TIPC_LINK_ESTABLISHING;
128
}
P
Per Liden 已提交
129

130 131
static int tipc_link_proto_rcv(struct tipc_link *l, struct sk_buff *skb,
			       struct sk_buff_head *xmitq);
132 133 134
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);
135 136
static void link_reset_statistics(struct tipc_link *l_ptr);
static void link_print(struct tipc_link *l_ptr, const char *str);
137 138
static void tipc_link_build_bcast_sync_msg(struct tipc_link *l,
					   struct sk_buff_head *xmitq);
139
static void tipc_link_sync_rcv(struct tipc_node *n, struct sk_buff *buf);
140
static int tipc_link_input(struct tipc_link *l, struct sk_buff *skb);
141
static bool tipc_data_input(struct tipc_link *l, struct sk_buff *skb);
142
static int tipc_link_failover_rcv(struct tipc_link *l, struct sk_buff **skb);
143

P
Per Liden 已提交
144
/*
S
Sam Ravnborg 已提交
145
 *  Simple link routines
P
Per Liden 已提交
146
 */
S
Sam Ravnborg 已提交
147
static unsigned int align(unsigned int i)
P
Per Liden 已提交
148 149 150 151
{
	return (i + 3) & ~3u;
}

152 153
static struct tipc_link *tipc_parallel_link(struct tipc_link *l)
{
154 155 156 157 158
	struct tipc_node *n = l->owner;

	if (node_active_link(n, 0) != l)
		return node_active_link(n, 0);
	return node_active_link(n, 1);
159 160
}

P
Per Liden 已提交
161
/*
S
Sam Ravnborg 已提交
162
 *  Simple non-static link routines (i.e. referenced outside this file)
P
Per Liden 已提交
163
 */
164
int tipc_link_is_up(struct tipc_link *l_ptr)
P
Per Liden 已提交
165 166 167
{
	if (!l_ptr)
		return 0;
168
	return link_working(l_ptr) || link_probing(l_ptr);
P
Per Liden 已提交
169 170
}

171
int tipc_link_is_active(struct tipc_link *l)
P
Per Liden 已提交
172
{
173 174 175
	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
176 177 178
}

/**
179
 * tipc_link_create - create a new link
180
 * @n_ptr: pointer to associated node
P
Per Liden 已提交
181 182
 * @b_ptr: pointer to associated bearer
 * @media_addr: media address to use when sending messages over link
183
 *
P
Per Liden 已提交
184 185
 * Returns pointer to link.
 */
186
struct tipc_link *tipc_link_create(struct tipc_node *n_ptr,
187
				   struct tipc_bearer *b_ptr,
188 189 190
				   const struct tipc_media_addr *media_addr,
				   struct sk_buff_head *inputq,
				   struct sk_buff_head *namedq)
P
Per Liden 已提交
191
{
192
	struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id);
193
	struct tipc_link *l_ptr;
P
Per Liden 已提交
194 195
	struct tipc_msg *msg;
	char *if_name;
196 197 198
	char addr_string[16];
	u32 peer = n_ptr->addr;

199
	if (n_ptr->link_cnt >= MAX_BEARERS) {
200
		tipc_addr_string_fill(addr_string, n_ptr->addr);
201 202
		pr_err("Cannot establish %uth link to %s. Max %u allowed.\n",
		       n_ptr->link_cnt, addr_string, MAX_BEARERS);
203 204 205
		return NULL;
	}

206
	if (n_ptr->links[b_ptr->identity].link) {
207
		tipc_addr_string_fill(addr_string, n_ptr->addr);
208 209
		pr_err("Attempt to establish second link on <%s> to %s\n",
		       b_ptr->name, addr_string);
210 211
		return NULL;
	}
P
Per Liden 已提交
212

213
	l_ptr = kzalloc(sizeof(*l_ptr), GFP_ATOMIC);
P
Per Liden 已提交
214
	if (!l_ptr) {
215
		pr_warn("Link creation failed, no memory\n");
P
Per Liden 已提交
216 217 218
		return NULL;
	}
	l_ptr->addr = peer;
219
	if_name = strchr(b_ptr->name, ':') + 1;
220
	sprintf(l_ptr->name, "%u.%u.%u:%s-%u.%u.%u:unknown",
221 222
		tipc_zone(tn->own_addr), tipc_cluster(tn->own_addr),
		tipc_node(tn->own_addr),
P
Per Liden 已提交
223 224
		if_name,
		tipc_zone(peer), tipc_cluster(peer), tipc_node(peer));
225
		/* note: peer i/f name is updated by reset/activate message */
P
Per Liden 已提交
226
	memcpy(&l_ptr->media_addr, media_addr, sizeof(*media_addr));
227
	l_ptr->owner = n_ptr;
228
	l_ptr->peer_session = WILDCARD_SESSION;
229
	l_ptr->bearer_id = b_ptr->identity;
230
	l_ptr->tolerance = b_ptr->tolerance;
231 232
	l_ptr->snd_nxt = 1;
	l_ptr->rcv_nxt = 1;
233
	l_ptr->state = TIPC_LINK_RESETTING;
P
Per Liden 已提交
234 235 236

	l_ptr->pmsg = (struct tipc_msg *)&l_ptr->proto_msg;
	msg = l_ptr->pmsg;
237
	tipc_msg_init(tn->own_addr, msg, LINK_PROTOCOL, RESET_MSG, INT_H_SIZE,
238
		      l_ptr->addr);
P
Per Liden 已提交
239
	msg_set_size(msg, sizeof(l_ptr->proto_msg));
240
	msg_set_session(msg, (tn->random & 0xffff));
P
Per Liden 已提交
241 242
	msg_set_bearer_id(msg, b_ptr->identity);
	strcpy((char *)msg_data(msg), if_name);
243
	l_ptr->net_plane = b_ptr->net_plane;
244 245
	l_ptr->advertised_mtu = b_ptr->mtu;
	l_ptr->mtu = l_ptr->advertised_mtu;
246 247
	l_ptr->priority = b_ptr->priority;
	tipc_link_set_queue_limits(l_ptr, b_ptr->window);
248
	l_ptr->snd_nxt = 1;
J
Jon Paul Maloy 已提交
249 250 251
	__skb_queue_head_init(&l_ptr->transmq);
	__skb_queue_head_init(&l_ptr->backlogq);
	__skb_queue_head_init(&l_ptr->deferdq);
252
	skb_queue_head_init(&l_ptr->wakeupq);
253 254 255
	l_ptr->inputq = inputq;
	l_ptr->namedq = namedq;
	skb_queue_head_init(l_ptr->inputq);
P
Per Liden 已提交
256
	link_reset_statistics(l_ptr);
257
	tipc_node_attach_link(n_ptr, l_ptr);
P
Per Liden 已提交
258 259 260
	return l_ptr;
}

261 262 263 264 265 266 267 268 269 270
/* 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.
 */
static void tipc_link_build_bcast_sync_msg(struct tipc_link *l,
					   struct sk_buff_head *xmitq)
{
	struct sk_buff *skb;
	struct sk_buff_head list;
271
	u16 last_sent;
272 273 274 275 276

	skb = tipc_msg_create(BCAST_PROTOCOL, STATE_MSG, INT_H_SIZE,
			      0, l->addr, link_own_addr(l), 0, 0, 0);
	if (!skb)
		return;
277 278
	last_sent = tipc_bclink_get_last_sent(l->owner->net);
	msg_set_last_bcast(buf_msg(skb), last_sent);
279 280 281 282 283
	__skb_queue_head_init(&list);
	__skb_queue_tail(&list, skb);
	tipc_link_xmit(l, &list, xmitq);
}

284 285 286 287 288 289 290 291 292 293 294 295
/**
 * tipc_link_fsm_evt - link finite state machine
 * @l: pointer to link
 * @evt: state machine event to be processed
 * @xmitq: queue to prepend created protocol message, if any
 */
static int tipc_link_fsm_evt(struct tipc_link *l, int evt,
			     struct sk_buff_head *xmitq)
{
	int mtyp = 0, rc = 0;
	struct tipc_link *pl;
	enum {
296 297 298 299 300 301 302
		LINK_RESET     = 1,
		LINK_ACTIVATE  = (1 << 1),
		SND_PROBE      = (1 << 2),
		SND_STATE      = (1 << 3),
		SND_RESET      = (1 << 4),
		SND_ACTIVATE   = (1 << 5),
		SND_BCAST_SYNC = (1 << 6)
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
	} actions = 0;

	if (l->exec_mode == TIPC_LINK_BLOCKED)
		return rc;

	switch (l->state) {
	case TIPC_LINK_WORKING:
		switch (evt) {
		case TRAFFIC_EVT:
		case ACTIVATE_EVT:
			break;
		case SILENCE_EVT:
			l->state = TIPC_LINK_PROBING;
			actions |= SND_PROBE;
			break;
		case PEER_RESET_EVT:
			actions |= LINK_RESET | SND_ACTIVATE;
			break;
		default:
			pr_debug("%s%u WORKING\n", link_unk_evt, evt);
		}
		break;
	case TIPC_LINK_PROBING:
		switch (evt) {
		case TRAFFIC_EVT:
		case ACTIVATE_EVT:
			l->state = TIPC_LINK_WORKING;
			break;
		case PEER_RESET_EVT:
			actions |= LINK_RESET | SND_ACTIVATE;
			break;
		case SILENCE_EVT:
			if (l->silent_intv_cnt <= l->abort_limit) {
				actions |= SND_PROBE;
				break;
			}
			actions |= LINK_RESET | SND_RESET;
			break;
		default:
			pr_err("%s%u PROBING\n", link_unk_evt, evt);
		}
		break;
	case TIPC_LINK_RESETTING:
		switch (evt) {
		case TRAFFIC_EVT:
			break;
		case ACTIVATE_EVT:
			pl = node_active_link(l->owner, 0);
			if (pl && link_probing(pl))
				break;
353
			l->state = TIPC_LINK_WORKING;
354
			actions |= LINK_ACTIVATE;
355 356
			if (!l->owner->working_links)
				actions |= SND_BCAST_SYNC;
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
			break;
		case PEER_RESET_EVT:
			l->state = TIPC_LINK_ESTABLISHING;
			actions |= SND_ACTIVATE;
			break;
		case SILENCE_EVT:
			actions |= SND_RESET;
			break;
		default:
			pr_err("%s%u in RESETTING\n", link_unk_evt, evt);
		}
		break;
	case TIPC_LINK_ESTABLISHING:
		switch (evt) {
		case TRAFFIC_EVT:
		case ACTIVATE_EVT:
			pl = node_active_link(l->owner, 0);
			if (pl && link_probing(pl))
				break;
376
			l->state = TIPC_LINK_WORKING;
377
			actions |= LINK_ACTIVATE;
378 379
			if (!l->owner->working_links)
				actions |= SND_BCAST_SYNC;
380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
			break;
		case PEER_RESET_EVT:
			break;
		case SILENCE_EVT:
			actions |= SND_ACTIVATE;
			break;
		default:
			pr_err("%s%u ESTABLISHING\n", link_unk_evt, evt);
		}
		break;
	default:
		pr_err("Unknown link state %u/%u\n", l->state, evt);
	}

	/* Perform actions as decided by FSM */
	if (actions & LINK_RESET) {
		l->exec_mode = TIPC_LINK_BLOCKED;
		rc |= TIPC_LINK_DOWN_EVT;
	}
	if (actions & LINK_ACTIVATE) {
		l->exec_mode = TIPC_LINK_OPEN;
		rc |= TIPC_LINK_UP_EVT;
	}
	if (actions & (SND_STATE | SND_PROBE))
		mtyp = STATE_MSG;
	if (actions & SND_RESET)
		mtyp = RESET_MSG;
	if (actions & SND_ACTIVATE)
		mtyp = ACTIVATE_MSG;
	if (actions & (SND_PROBE | SND_STATE | SND_RESET | SND_ACTIVATE))
		tipc_link_build_proto_msg(l, mtyp, actions & SND_PROBE,
					  0, 0, 0, xmitq);
412 413
	if (actions & SND_BCAST_SYNC)
		tipc_link_build_bcast_sync_msg(l, xmitq);
414 415 416
	return rc;
}

417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472
/* 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;

	link_profile_stats(l);
	if (l->silent_intv_cnt)
		rc = tipc_link_fsm_evt(l, SILENCE_EVT, xmitq);
	else if (link_working(l) && tipc_bclink_acks_missing(l->owner))
		tipc_link_build_proto_msg(l, STATE_MSG, 0, 0, 0, 0, xmitq);
	l->silent_intv_cnt++;
	return rc;
}

P
Per Liden 已提交
473
/**
474
 * link_schedule_user - schedule a message sender for wakeup after congestion
475
 * @link: congested link
476
 * @list: message that was attempted sent
477
 * Create pseudo msg to send back to user when congestion abates
478
 * Does not consume buffer list
P
Per Liden 已提交
479
 */
480
static int link_schedule_user(struct tipc_link *link, struct sk_buff_head *list)
P
Per Liden 已提交
481
{
482 483 484 485 486 487 488 489 490
	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);
491
		return -ENOBUFS;
492 493 494 495 496 497 498 499 500
	}
	/* 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)
501
		return -ENOBUFS;
502 503 504
	TIPC_SKB_CB(skb)->chain_sz = skb_queue_len(list);
	TIPC_SKB_CB(skb)->chain_imp = imp;
	skb_queue_tail(&link->wakeupq, skb);
505
	link->stats.link_congs++;
506
	return -ELINKCONG;
P
Per Liden 已提交
507 508
}

509 510 511 512 513 514
/**
 * 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
 */
515
void link_prepare_wakeup(struct tipc_link *l)
P
Per Liden 已提交
516
{
517 518
	int pnd[TIPC_SYSTEM_IMPORTANCE + 1] = {0,};
	int imp, lim;
519
	struct sk_buff *skb, *tmp;
520

521 522 523 524 525
	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
526
			break;
527
		skb_unlink(skb, &l->wakeupq);
528 529
		skb_queue_tail(l->inputq, skb);
		l->owner->inputq = l->inputq;
530
		l->owner->action_flags |= TIPC_MSG_EVT;
P
Per Liden 已提交
531 532 533 534
	}
}

/**
535
 * tipc_link_reset_fragments - purge link's inbound message fragments queue
P
Per Liden 已提交
536 537
 * @l_ptr: pointer to link
 */
538
void tipc_link_reset_fragments(struct tipc_link *l_ptr)
P
Per Liden 已提交
539
{
540 541
	kfree_skb(l_ptr->reasm_buf);
	l_ptr->reasm_buf = NULL;
P
Per Liden 已提交
542 543
}

544
void tipc_link_purge_backlog(struct tipc_link *l)
545 546 547 548 549 550 551 552 553
{
	__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;
}

554
/**
555
 * tipc_link_purge_queues - purge all pkt queues associated with link
P
Per Liden 已提交
556 557
 * @l_ptr: pointer to link
 */
558
void tipc_link_purge_queues(struct tipc_link *l_ptr)
P
Per Liden 已提交
559
{
J
Jon Paul Maloy 已提交
560 561
	__skb_queue_purge(&l_ptr->deferdq);
	__skb_queue_purge(&l_ptr->transmq);
562
	tipc_link_purge_backlog(l_ptr);
563
	tipc_link_reset_fragments(l_ptr);
P
Per Liden 已提交
564 565
}

566
void tipc_link_reset(struct tipc_link *l_ptr)
P
Per Liden 已提交
567 568
{
	u32 prev_state = l_ptr->state;
569
	int was_active_link = tipc_link_is_active(l_ptr);
570
	struct tipc_node *owner = l_ptr->owner;
571
	struct tipc_link *pl = tipc_parallel_link(l_ptr);
572

573
	msg_set_session(l_ptr->pmsg, ((msg_session(l_ptr->pmsg) + 1) & 0xffff));
P
Per Liden 已提交
574

575
	/* Link is down, accept any session */
576
	l_ptr->peer_session = WILDCARD_SESSION;
P
Per Liden 已提交
577

578 579
	/* Prepare for renewed mtu size negotiation */
	l_ptr->mtu = l_ptr->advertised_mtu;
580

581
	l_ptr->state = TIPC_LINK_RESETTING;
P
Per Liden 已提交
582

583 584
	if ((prev_state == TIPC_LINK_RESETTING) ||
	    (prev_state == TIPC_LINK_ESTABLISHING))
P
Per Liden 已提交
585 586
		return;

587
	tipc_node_link_down(l_ptr->owner, l_ptr->bearer_id);
588
	tipc_bearer_remove_dest(owner->net, l_ptr->bearer_id, l_ptr->addr);
589

590
	if (was_active_link && tipc_node_is_up(l_ptr->owner) && (pl != l_ptr)) {
591
		l_ptr->exec_mode = TIPC_LINK_BLOCKED;
592
		l_ptr->failover_checkpt = l_ptr->rcv_nxt;
593
		pl->failover_pkts = FIRST_FAILOVER;
594
		pl->failover_checkpt = l_ptr->rcv_nxt;
595 596 597
		pl->failover_skb = l_ptr->reasm_buf;
	} else {
		kfree_skb(l_ptr->reasm_buf);
P
Per Liden 已提交
598
	}
599
	/* Clean up all queues, except inputq: */
J
Jon Paul Maloy 已提交
600 601
	__skb_queue_purge(&l_ptr->transmq);
	__skb_queue_purge(&l_ptr->deferdq);
602
	if (!owner->inputq)
603
		owner->inputq = l_ptr->inputq;
604 605
	skb_queue_splice_init(&l_ptr->wakeupq, owner->inputq);
	if (!skb_queue_empty(owner->inputq))
606
		owner->action_flags |= TIPC_MSG_EVT;
607
	tipc_link_purge_backlog(l_ptr);
608
	l_ptr->reasm_buf = NULL;
J
Jon Paul Maloy 已提交
609
	l_ptr->rcv_unacked = 0;
610
	l_ptr->snd_nxt = 1;
611
	l_ptr->rcv_nxt = 1;
612
	l_ptr->silent_intv_cnt = 0;
613
	l_ptr->stats.recv_info = 0;
P
Per Liden 已提交
614 615 616 617
	l_ptr->stale_count = 0;
	link_reset_statistics(l_ptr);
}

618
/**
619
 * __tipc_link_xmit(): same as tipc_link_xmit, but destlink is known & locked
620
 * @link: link to use
621 622
 * @list: chain of buffers containing message
 *
623
 * Consumes the buffer chain, except when returning an error code,
624 625
 * Returns 0 if success, or errno: -ELINKCONG, -EMSGSIZE or -ENOBUFS
 * Messages at TIPC_SYSTEM_IMPORTANCE are always accepted
626
 */
627 628
int __tipc_link_xmit(struct net *net, struct tipc_link *link,
		     struct sk_buff_head *list)
629
{
630
	struct tipc_msg *msg = buf_msg(skb_peek(list));
J
Jon Paul Maloy 已提交
631
	unsigned int maxwin = link->window;
632
	unsigned int i, imp = msg_importance(msg);
633
	uint mtu = link->mtu;
634 635
	u16 ack = mod(link->rcv_nxt - 1);
	u16 seqno = link->snd_nxt;
636
	u16 bc_last_in = link->owner->bclink.last_in;
637
	struct tipc_media_addr *addr = &link->media_addr;
J
Jon Paul Maloy 已提交
638 639
	struct sk_buff_head *transmq = &link->transmq;
	struct sk_buff_head *backlogq = &link->backlogq;
640
	struct sk_buff *skb, *bskb;
641

642 643 644 645 646
	/* 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);
	}
647
	if (unlikely(msg_size(msg) > mtu))
648
		return -EMSGSIZE;
649

J
Jon Paul Maloy 已提交
650
	/* Prepare each packet for sending, and add to relevant queue: */
651 652
	while (skb_queue_len(list)) {
		skb = skb_peek(list);
653
		msg = buf_msg(skb);
J
Jon Paul Maloy 已提交
654 655
		msg_set_seqno(msg, seqno);
		msg_set_ack(msg, ack);
656 657
		msg_set_bcast_ack(msg, bc_last_in);

J
Jon Paul Maloy 已提交
658
		if (likely(skb_queue_len(transmq) < maxwin)) {
659
			__skb_dequeue(list);
J
Jon Paul Maloy 已提交
660 661 662 663 664 665
			__skb_queue_tail(transmq, skb);
			tipc_bearer_send(net, link->bearer_id, skb, addr);
			link->rcv_unacked = 0;
			seqno++;
			continue;
		}
666 667
		if (tipc_msg_bundle(skb_peek_tail(backlogq), msg, mtu)) {
			kfree_skb(__skb_dequeue(list));
668 669
			link->stats.sent_bundled++;
			continue;
J
Jon Paul Maloy 已提交
670
		}
671 672 673 674
		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++;
675 676
			link->stats.sent_bundled++;
			link->stats.sent_bundles++;
677
			continue;
678
		}
679 680
		link->backlog[imp].len += skb_queue_len(list);
		skb_queue_splice_tail_init(list, backlogq);
681
	}
682
	link->snd_nxt = seqno;
683 684 685
	return 0;
}

686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757
/**
 * 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;
}

758 759
static void skb2list(struct sk_buff *skb, struct sk_buff_head *list)
{
760
	skb_queue_head_init(list);
761 762 763 764 765 766 767 768
	__skb_queue_tail(list, skb);
}

static int __tipc_link_xmit_skb(struct tipc_link *link, struct sk_buff *skb)
{
	struct sk_buff_head head;

	skb2list(skb, &head);
769
	return __tipc_link_xmit(link->owner->net, link, &head);
770 771
}

772
/*
773
 * tipc_link_sync_rcv - synchronize broadcast link endpoints.
774 775 776 777 778 779
 * 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
 */
780
static void tipc_link_sync_rcv(struct tipc_node *n, struct sk_buff *buf)
781 782 783 784 785 786 787 788
{
	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);
}

789
/*
790 791 792 793 794 795
 * 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
796
 */
J
Jon Paul Maloy 已提交
797
void tipc_link_push_packets(struct tipc_link *link)
P
Per Liden 已提交
798
{
J
Jon Paul Maloy 已提交
799
	struct sk_buff *skb;
800
	struct tipc_msg *msg;
801
	u16 seqno = link->snd_nxt;
802
	u16 ack = mod(link->rcv_nxt - 1);
P
Per Liden 已提交
803

J
Jon Paul Maloy 已提交
804 805 806
	while (skb_queue_len(&link->transmq) < link->window) {
		skb = __skb_dequeue(&link->backlogq);
		if (!skb)
807
			break;
J
Jon Paul Maloy 已提交
808
		msg = buf_msg(skb);
809
		link->backlog[msg_importance(msg)].len--;
J
Jon Paul Maloy 已提交
810
		msg_set_ack(msg, ack);
811 812
		msg_set_seqno(msg, seqno);
		seqno = mod(seqno + 1);
J
Jon Paul Maloy 已提交
813 814 815 816 817
		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,
				 skb, &link->media_addr);
P
Per Liden 已提交
818
	}
819
	link->snd_nxt = seqno;
P
Per Liden 已提交
820 821
}

822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849
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;
}

850
static void link_retransmit_failure(struct tipc_link *l_ptr,
851
				    struct sk_buff *buf)
852 853
{
	struct tipc_msg *msg = buf_msg(buf);
854
	struct net *net = l_ptr->owner->net;
855

856
	pr_warn("Retransmission failure on link <%s>\n", l_ptr->name);
857 858 859

	if (l_ptr->addr) {
		/* Handle failure on standard link */
860 861 862 863 864 865
		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));
866 867
	} else {
		/* Handle failure on broadcast link */
868
		struct tipc_node *n_ptr;
869 870
		char addr_string[16];

871 872 873
		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
874

875
		n_ptr = tipc_bclink_retransmit_to(net);
876

877
		tipc_addr_string_fill(addr_string, n_ptr->addr);
878
		pr_info("Broadcast link info for %s\n", addr_string);
879 880
		pr_info("Reception permitted: %d,  Acked: %u\n",
			n_ptr->bclink.recv_permitted,
881 882 883 884 885
			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);
886

887
		n_ptr->action_flags |= TIPC_BCAST_RESET;
888 889 890 891
		l_ptr->stale_count = 0;
	}
}

892
void tipc_link_retransmit(struct tipc_link *l_ptr, struct sk_buff *skb,
893
			  u32 retransmits)
P
Per Liden 已提交
894 895 896
{
	struct tipc_msg *msg;

897
	if (!skb)
898 899
		return;

900
	msg = buf_msg(skb);
901

902
	/* Detect repeated retransmit failures */
903
	if (l_ptr->last_retransm == msg_seqno(msg)) {
904
		if (++l_ptr->stale_count > 100) {
905
			link_retransmit_failure(l_ptr, skb);
906
			return;
907 908
		}
	} else {
909
		l_ptr->last_retransm = msg_seqno(msg);
910
		l_ptr->stale_count = 1;
P
Per Liden 已提交
911
	}
912

J
Jon Paul Maloy 已提交
913 914
	skb_queue_walk_from(&l_ptr->transmq, skb) {
		if (!retransmits)
915 916
			break;
		msg = buf_msg(skb);
917
		msg_set_ack(msg, mod(l_ptr->rcv_nxt - 1));
918
		msg_set_bcast_ack(msg, l_ptr->owner->bclink.last_in);
919 920
		tipc_bearer_send(l_ptr->owner->net, l_ptr->bearer_id, skb,
				 &l_ptr->media_addr);
921 922
		retransmits--;
		l_ptr->stats.retransmitted++;
P
Per Liden 已提交
923 924 925
	}
}

926 927 928 929 930 931 932 933 934 935 936 937 938 939 940
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);
941
		l->exec_mode = TIPC_LINK_BLOCKED;
942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961
		return TIPC_LINK_DOWN_EVT;
	}
	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;
}

962 963 964 965 966 967 968 969 970 971 972 973 974 975
/* link_synch(): check if all packets arrived before the synch
 *               point have been consumed
 * Returns true if the parallel links are synched, otherwise false
 */
static bool link_synch(struct tipc_link *l)
{
	unsigned int post_synch;
	struct tipc_link *pl;

	pl  = tipc_parallel_link(l);
	if (pl == l)
		goto synched;

	/* Was last pre-synch packet added to input queue ? */
976
	if (less_eq(pl->rcv_nxt, l->synch_point))
977 978 979
		return false;

	/* Is it still in the input queue ? */
980
	post_synch = mod(pl->rcv_nxt - l->synch_point) - 1;
981
	if (skb_queue_len(pl->inputq) > post_synch)
982 983
		return false;
synched:
984
	l->exec_mode = TIPC_LINK_OPEN;
985 986 987
	return true;
}

988
/* tipc_data_input - deliver data and name distr msgs to upper layer
989
 *
990
 * Consumes buffer if message is of right type
991 992
 * Node lock must be held
 */
993
static bool tipc_data_input(struct tipc_link *link, struct sk_buff *skb)
994
{
995 996 997
	struct tipc_node *node = link->owner;
	struct tipc_msg *msg = buf_msg(skb);
	u32 dport = msg_destport(msg);
998 999

	switch (msg_user(msg)) {
1000 1001 1002 1003 1004
	case TIPC_LOW_IMPORTANCE:
	case TIPC_MEDIUM_IMPORTANCE:
	case TIPC_HIGH_IMPORTANCE:
	case TIPC_CRITICAL_IMPORTANCE:
	case CONN_MANAGER:
1005 1006
		if (tipc_skb_queue_tail(link->inputq, skb, dport)) {
			node->inputq = link->inputq;
1007
			node->action_flags |= TIPC_MSG_EVT;
1008
		}
1009
		return true;
1010
	case NAME_DISTRIBUTOR:
1011
		node->bclink.recv_permitted = true;
1012 1013 1014
		node->namedq = link->namedq;
		skb_queue_tail(link->namedq, skb);
		if (skb_queue_len(link->namedq) == 1)
1015 1016 1017
			node->action_flags |= TIPC_NAMED_MSG_EVT;
		return true;
	case MSG_BUNDLER:
1018
	case TUNNEL_PROTOCOL:
1019
	case MSG_FRAGMENTER:
1020
	case BCAST_PROTOCOL:
1021
		return false;
1022
	default:
1023 1024 1025 1026
		pr_warn("Dropping received illegal msg type\n");
		kfree_skb(skb);
		return false;
	};
1027
}
1028 1029 1030 1031 1032

/* tipc_link_input - process packet that has passed link protocol check
 *
 * Consumes buffer
 * Node lock must be held
1033
 */
1034
static int tipc_link_input(struct tipc_link *link, struct sk_buff *skb)
1035
{
1036 1037 1038 1039
	struct tipc_node *node = link->owner;
	struct tipc_msg *msg = buf_msg(skb);
	struct sk_buff *iskb;
	int pos = 0;
1040
	int rc = 0;
1041

1042
	switch (msg_user(msg)) {
1043
	case TUNNEL_PROTOCOL:
1044
		if (msg_dup(msg)) {
1045
			link->exec_mode = TIPC_LINK_TUNNEL;
1046
			link->synch_point = msg_seqno(msg_get_wrapped(msg));
1047 1048
			kfree_skb(skb);
			break;
1049
		}
1050 1051
		rc |= tipc_link_failover_rcv(link, &skb);
		if (!skb)
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
			break;
		if (msg_user(buf_msg(skb)) != MSG_BUNDLER) {
			tipc_data_input(link, skb);
			break;
		}
	case MSG_BUNDLER:
		link->stats.recv_bundles++;
		link->stats.recv_bundled += msg_msgcnt(msg);

		while (tipc_msg_extract(skb, &iskb, &pos))
			tipc_data_input(link, iskb);
1063
		break;
1064 1065 1066 1067 1068 1069
	case MSG_FRAGMENTER:
		link->stats.recv_fragments++;
		if (tipc_buf_append(&link->reasm_buf, &skb)) {
			link->stats.recv_fragmented++;
			tipc_data_input(link, skb);
		} else if (!link->reasm_buf) {
1070 1071
			link->exec_mode = TIPC_LINK_BLOCKED;
			rc |= TIPC_LINK_DOWN_EVT;
1072
		}
1073
		break;
1074 1075
	case BCAST_PROTOCOL:
		tipc_link_sync_rcv(node, skb);
1076 1077
		break;
	default:
1078 1079
		break;
	};
1080
	return rc;
1081 1082
}

1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173
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;
}

/* tipc_link_rcv - process TIPC packets/messages arriving from off-node
 * @link: the link that should handle the message
 * @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)
{
	struct sk_buff_head *arrvq = &l->deferdq;
	struct sk_buff *tmp;
	struct tipc_msg *hdr;
	u16 seqno, rcv_nxt;
	int rc = 0;

	if (unlikely(!__tipc_skb_queue_sorted(arrvq, skb))) {
		if (!(skb_queue_len(arrvq) % TIPC_NACK_INTV))
			tipc_link_build_proto_msg(l, STATE_MSG, 0,
						  0, 0, 0, xmitq);
		return rc;
	}

	skb_queue_walk_safe(arrvq, skb, tmp) {
		hdr = buf_msg(skb);

		/* Verify and update link state */
		if (unlikely(msg_user(hdr) == LINK_PROTOCOL)) {
			__skb_dequeue(arrvq);
			rc |= tipc_link_proto_rcv(l, skb, xmitq);
			continue;
		}

		if (unlikely(!link_working(l))) {
			rc |= tipc_link_fsm_evt(l, TRAFFIC_EVT, xmitq);
			if (!link_working(l)) {
				kfree_skb(__skb_dequeue(arrvq));
				return rc;
			}
		}

		l->silent_intv_cnt = 0;

		/* 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);
		}

		/* Defer reception if there is a gap in the sequence */
		seqno = msg_seqno(hdr);
		rcv_nxt = l->rcv_nxt;
		if (unlikely(less(rcv_nxt, seqno))) {
			l->stats.deferred_recv++;
			return rc;
		}

		__skb_dequeue(arrvq);

		/* Drop if packet already received */
		if (unlikely(more(rcv_nxt, seqno))) {
			l->stats.duplicates++;
			kfree_skb(skb);
			return rc;
		}

		/* Synchronize with parallel link if applicable */
		if (unlikely(l->exec_mode == TIPC_LINK_TUNNEL))
			if (!msg_dup(hdr) && !link_synch(l)) {
				kfree_skb(skb);
				return rc;
			}

		/* Packet can be delivered */
		l->rcv_nxt++;
		l->stats.recv_info++;
		if (unlikely(!tipc_data_input(l, skb)))
1174
			rc |= tipc_link_input(l, skb);
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186

		/* Ack at regular intervals */
		if (unlikely(++l->rcv_unacked >= TIPC_MIN_LINK_WIN)) {
			l->rcv_unacked = 0;
			l->stats.sent_acks++;
			tipc_link_build_proto_msg(l, STATE_MSG,
						  0, 0, 0, 0, xmitq);
		}
	}
	return rc;
}

1187
/**
1188 1189 1190
 * 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
1191
 */
1192
u32 tipc_link_defer_pkt(struct sk_buff_head *list, struct sk_buff *skb)
P
Per Liden 已提交
1193
{
1194
	struct sk_buff *skb1;
1195
	u16 seq_no = buf_seqno(skb);
P
Per Liden 已提交
1196 1197

	/* Empty queue ? */
1198 1199
	if (skb_queue_empty(list)) {
		__skb_queue_tail(list, skb);
P
Per Liden 已提交
1200 1201 1202 1203
		return 1;
	}

	/* Last ? */
1204 1205
	if (less(buf_seqno(skb_peek_tail(list)), seq_no)) {
		__skb_queue_tail(list, skb);
P
Per Liden 已提交
1206 1207 1208
		return 1;
	}

1209
	/* Locate insertion point in queue, then insert; discard if duplicate */
1210
	skb_queue_walk(list, skb1) {
1211
		u16 curr_seqno = buf_seqno(skb1);
P
Per Liden 已提交
1212

1213
		if (seq_no == curr_seqno) {
1214
			kfree_skb(skb);
1215
			return 0;
P
Per Liden 已提交
1216
		}
1217 1218

		if (less(seq_no, curr_seqno))
P
Per Liden 已提交
1219
			break;
1220
	}
P
Per Liden 已提交
1221

1222
	__skb_queue_before(list, skb1, skb);
1223
	return 1;
P
Per Liden 已提交
1224 1225 1226 1227 1228
}

/*
 * Send protocol message to the other endpoint.
 */
1229
void tipc_link_proto_xmit(struct tipc_link *l, u32 msg_typ, int probe_msg,
1230
			  u32 gap, u32 tolerance, u32 priority)
P
Per Liden 已提交
1231
{
1232 1233
	struct sk_buff *skb = NULL;
	struct sk_buff_head xmitq;
P
Per Liden 已提交
1234

1235 1236 1237 1238 1239
	__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
1240
		return;
1241 1242 1243
	tipc_bearer_send(l->owner->net, l->bearer_id, skb, &l->media_addr);
	l->rcv_unacked = 0;
	kfree_skb(skb);
P
Per Liden 已提交
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 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304
/* 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 */
	if (l->exec_mode == TIPC_LINK_BLOCKED)
		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);
		msg_set_ack(hdr, l->failover_checkpt - 1);
		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;
	__skb_queue_head(xmitq, skb);
}
P
Per Liden 已提交
1305

1306 1307
/* tipc_link_tunnel_xmit(): Tunnel one packet via a link belonging to
 * a different bearer. Owner node is locked.
P
Per Liden 已提交
1308
 */
1309 1310 1311 1312
static void tipc_link_tunnel_xmit(struct tipc_link *l_ptr,
				  struct tipc_msg *tunnel_hdr,
				  struct tipc_msg *msg,
				  u32 selector)
P
Per Liden 已提交
1313
{
1314
	struct tipc_link *tunnel;
1315
	struct sk_buff *skb;
P
Per Liden 已提交
1316 1317
	u32 length = msg_size(msg);

1318
	tunnel = node_active_link(l_ptr->owner, selector & 1);
1319
	if (!tipc_link_is_up(tunnel)) {
1320
		pr_warn("%stunnel link no longer available\n", link_co_err);
P
Per Liden 已提交
1321
		return;
1322
	}
P
Per Liden 已提交
1323
	msg_set_size(tunnel_hdr, length + INT_H_SIZE);
1324 1325
	skb = tipc_buf_acquire(length + INT_H_SIZE);
	if (!skb) {
1326
		pr_warn("%sunable to send tunnel msg\n", link_co_err);
P
Per Liden 已提交
1327
		return;
1328
	}
1329 1330 1331
	skb_copy_to_linear_data(skb, tunnel_hdr, INT_H_SIZE);
	skb_copy_to_linear_data_offset(skb, INT_H_SIZE, msg, length);
	__tipc_link_xmit_skb(tunnel, skb);
P
Per Liden 已提交
1332 1333 1334
}


1335 1336 1337 1338 1339
/* tipc_link_failover_send_queue(): A link has gone down, but a second
 * link is still active. We can do failover. Tunnel the failing link's
 * whole send queue via the remaining link. This way, we don't lose
 * any packets, and sequence order is preserved for subsequent traffic
 * sent over the remaining link. Owner node is locked.
P
Per Liden 已提交
1340
 */
1341
void tipc_link_failover_send_queue(struct tipc_link *l_ptr)
P
Per Liden 已提交
1342
{
J
Jon Paul Maloy 已提交
1343
	int msgcount;
1344
	struct tipc_link *tunnel = node_active_link(l_ptr->owner, 0);
P
Per Liden 已提交
1345
	struct tipc_msg tunnel_hdr;
1346
	struct sk_buff *skb;
1347
	int split_bundles;
P
Per Liden 已提交
1348 1349 1350 1351

	if (!tunnel)
		return;

1352 1353
	tipc_msg_init(link_own_addr(l_ptr), &tunnel_hdr, TUNNEL_PROTOCOL,
		      FAILOVER_MSG, INT_H_SIZE, l_ptr->addr);
1354 1355 1356 1357 1358

	skb_queue_walk(&l_ptr->backlogq, skb) {
		msg_set_seqno(buf_msg(skb), l_ptr->snd_nxt);
		l_ptr->snd_nxt = mod(l_ptr->snd_nxt + 1);
	}
J
Jon Paul Maloy 已提交
1359
	skb_queue_splice_tail_init(&l_ptr->backlogq, &l_ptr->transmq);
1360
	tipc_link_purge_backlog(l_ptr);
J
Jon Paul Maloy 已提交
1361
	msgcount = skb_queue_len(&l_ptr->transmq);
P
Per Liden 已提交
1362 1363
	msg_set_bearer_id(&tunnel_hdr, l_ptr->peer_bearer_id);
	msg_set_msgcnt(&tunnel_hdr, msgcount);
1364

J
Jon Paul Maloy 已提交
1365
	if (skb_queue_empty(&l_ptr->transmq)) {
1366 1367 1368
		skb = tipc_buf_acquire(INT_H_SIZE);
		if (skb) {
			skb_copy_to_linear_data(skb, &tunnel_hdr, INT_H_SIZE);
P
Per Liden 已提交
1369
			msg_set_size(&tunnel_hdr, INT_H_SIZE);
1370
			__tipc_link_xmit_skb(tunnel, skb);
P
Per Liden 已提交
1371
		} else {
1372 1373
			pr_warn("%sunable to send changeover msg\n",
				link_co_err);
P
Per Liden 已提交
1374 1375 1376
		}
		return;
	}
1377

1378 1379
	split_bundles = (node_active_link(l_ptr->owner, 0) !=
			 node_active_link(l_ptr->owner, 0));
1380

J
Jon Paul Maloy 已提交
1381
	skb_queue_walk(&l_ptr->transmq, skb) {
1382
		struct tipc_msg *msg = buf_msg(skb);
P
Per Liden 已提交
1383 1384 1385

		if ((msg_user(msg) == MSG_BUNDLER) && split_bundles) {
			struct tipc_msg *m = msg_get_wrapped(msg);
1386
			unchar *pos = (unchar *)m;
P
Per Liden 已提交
1387

1388
			msgcount = msg_msgcnt(msg);
P
Per Liden 已提交
1389
			while (msgcount--) {
1390
				msg_set_seqno(m, msg_seqno(msg));
1391 1392
				tipc_link_tunnel_xmit(l_ptr, &tunnel_hdr, m,
						      msg_link_selector(m));
P
Per Liden 已提交
1393 1394 1395 1396
				pos += align(msg_size(m));
				m = (struct tipc_msg *)pos;
			}
		} else {
1397 1398
			tipc_link_tunnel_xmit(l_ptr, &tunnel_hdr, msg,
					      msg_link_selector(msg));
P
Per Liden 已提交
1399 1400 1401 1402
		}
	}
}

1403
/* tipc_link_dup_queue_xmit(): A second link has become active. Tunnel a
1404 1405 1406 1407 1408 1409 1410 1411
 * duplicate of the first link's send queue via the new link. This way, we
 * are guaranteed that currently queued packets from a socket are delivered
 * before future traffic from the same socket, even if this is using the
 * new link. The last arriving copy of each duplicate packet is dropped at
 * the receiving end by the regular protocol check, so packet cardinality
 * and sequence order is preserved per sender/receiver socket pair.
 * Owner node is locked.
 */
J
Jon Paul Maloy 已提交
1412 1413
void tipc_link_dup_queue_xmit(struct tipc_link *link,
			      struct tipc_link *tnl)
P
Per Liden 已提交
1414
{
1415
	struct sk_buff *skb;
J
Jon Paul Maloy 已提交
1416 1417 1418
	struct tipc_msg tnl_hdr;
	struct sk_buff_head *queue = &link->transmq;
	int mcnt;
1419
	u16 seqno;
J
Jon Paul Maloy 已提交
1420

1421 1422
	tipc_msg_init(link_own_addr(link), &tnl_hdr, TUNNEL_PROTOCOL,
		      SYNCH_MSG, INT_H_SIZE, link->addr);
J
Jon Paul Maloy 已提交
1423 1424 1425 1426 1427 1428
	mcnt = skb_queue_len(&link->transmq) + skb_queue_len(&link->backlogq);
	msg_set_msgcnt(&tnl_hdr, mcnt);
	msg_set_bearer_id(&tnl_hdr, link->peer_bearer_id);

tunnel_queue:
	skb_queue_walk(queue, skb) {
1429 1430
		struct sk_buff *outskb;
		struct tipc_msg *msg = buf_msg(skb);
J
Jon Paul Maloy 已提交
1431
		u32 len = msg_size(msg);
P
Per Liden 已提交
1432

1433
		msg_set_ack(msg, mod(link->rcv_nxt - 1));
J
Jon Paul Maloy 已提交
1434 1435 1436
		msg_set_bcast_ack(msg, link->owner->bclink.last_in);
		msg_set_size(&tnl_hdr, len + INT_H_SIZE);
		outskb = tipc_buf_acquire(len + INT_H_SIZE);
1437
		if (outskb == NULL) {
1438 1439
			pr_warn("%sunable to send duplicate msg\n",
				link_co_err);
P
Per Liden 已提交
1440 1441
			return;
		}
J
Jon Paul Maloy 已提交
1442 1443 1444 1445 1446
		skb_copy_to_linear_data(outskb, &tnl_hdr, INT_H_SIZE);
		skb_copy_to_linear_data_offset(outskb, INT_H_SIZE,
					       skb->data, len);
		__tipc_link_xmit_skb(tnl, outskb);
		if (!tipc_link_is_up(link))
P
Per Liden 已提交
1447 1448
			return;
	}
J
Jon Paul Maloy 已提交
1449 1450
	if (queue == &link->backlogq)
		return;
1451 1452 1453 1454 1455
	seqno = link->snd_nxt;
	skb_queue_walk(&link->backlogq, skb) {
		msg_set_seqno(buf_msg(skb), seqno);
		seqno = mod(seqno + 1);
	}
J
Jon Paul Maloy 已提交
1456 1457
	queue = &link->backlogq;
	goto tunnel_queue;
P
Per Liden 已提交
1458 1459
}

1460
/*  tipc_link_failover_rcv(): Receive a tunnelled FAILOVER_MSG packet
1461 1462
 *  Owner node is locked.
 */
1463 1464
static int tipc_link_failover_rcv(struct tipc_link *link,
				  struct sk_buff **skb)
1465
{
1466 1467
	struct tipc_msg *msg = buf_msg(*skb);
	struct sk_buff *iskb = NULL;
1468
	struct tipc_link *pl = NULL;
1469
	int bearer_id = msg_bearer_id(msg);
1470
	int pos = 0;
1471
	int rc = 0;
1472

1473
	if (msg_type(msg) != FAILOVER_MSG) {
1474 1475
		pr_warn("%sunknown tunnel pkt received\n", link_co_err);
		goto exit;
1476
	}
1477 1478
	if (bearer_id >= MAX_BEARERS)
		goto exit;
1479 1480

	if (bearer_id == link->bearer_id)
1481
		goto exit;
P
Per Liden 已提交
1482

1483
	pl = link->owner->links[bearer_id].link;
1484 1485 1486

	if (link->failover_pkts == FIRST_FAILOVER)
		link->failover_pkts = msg_msgcnt(msg);
1487

1488
	/* Should we expect an inner packet? */
1489
	if (!link->failover_pkts)
1490
		goto exit;
1491

1492 1493 1494
	if (!tipc_msg_extract(*skb, &iskb, &pos)) {
		pr_warn("%sno inner failover pkt\n", link_co_err);
		*skb = NULL;
P
Per Liden 已提交
1495
		goto exit;
1496
	}
1497
	link->failover_pkts--;
1498
	*skb = NULL;
P
Per Liden 已提交
1499

1500 1501
	/* Was this packet already delivered? */
	if (less(buf_seqno(iskb), link->failover_checkpt)) {
1502 1503 1504 1505 1506 1507
		kfree_skb(iskb);
		iskb = NULL;
		goto exit;
	}
	if (msg_user(buf_msg(iskb)) == MSG_FRAGMENTER) {
		link->stats.recv_fragments++;
1508 1509 1510 1511 1512
		if (!tipc_buf_append(&link->failover_skb, &iskb) &&
		    !link->failover_skb) {
			link->exec_mode = TIPC_LINK_BLOCKED;
			rc |= TIPC_LINK_DOWN_EVT;
		}
1513
	}
P
Per Liden 已提交
1514
exit:
1515
	if (!link->failover_pkts && pl)
1516
		pl->exec_mode = TIPC_LINK_OPEN;
1517 1518
	kfree_skb(*skb);
	*skb = iskb;
1519
	return rc;
P
Per Liden 已提交
1520 1521
}

1522 1523 1524 1525 1526 1527 1528 1529 1530 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 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
/* 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);
	char *if_name;
	int rc = 0;

	if (l->exec_mode == TIPC_LINK_BLOCKED)
		goto exit;

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

	switch (msg_type(hdr)) {
	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' */
	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;

		l->peer_session = msg_session(hdr);
		l->peer_bearer_id = msg_bearer_id(hdr);
		rc = tipc_link_fsm_evt(l, msg_type(hdr), xmitq);
		if (l->mtu > msg_max_pkt(hdr))
			l->mtu = msg_max_pkt(hdr);
		break;
	case STATE_MSG:
		/* 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++;
		rc = tipc_link_fsm_evt(l, TRAFFIC_EVT, xmitq);
		if (!tipc_link_is_up(l))
			break;

		/* Has peer sent packets we haven't received yet ? */
		if (more(peers_snd_nxt, l->rcv_nxt))
			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
1595
						  0, 0, xmitq);
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
		tipc_link_release_pkts(l, msg_ack(hdr));

		/* If NACK, retransmit will now start at right position */
		if (nacked_gap) {
			rc |= tipc_link_retransm(l, nacked_gap, xmitq);
			l->stats.recv_nacks++;
		}
		tipc_link_advance_backlog(l, xmitq);
		if (unlikely(!skb_queue_empty(&l->wakeupq)))
			link_prepare_wakeup(l);
	}
exit:
	kfree_skb(skb);
	return rc;
}

1612
void tipc_link_set_queue_limits(struct tipc_link *l, u32 win)
P
Per Liden 已提交
1613
{
1614
	int max_bulk = TIPC_MAX_PUBLICATIONS / (l->mtu / ITEM_SIZE);
1615 1616

	l->window = win;
1617 1618 1619 1620 1621
	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
1622 1623
}

1624
/* tipc_link_find_owner - locate owner node of link by link's name
1625
 * @net: the applicable net namespace
1626 1627
 * @name: pointer to link name string
 * @bearer_id: pointer to index in 'node->links' array where the link was found.
1628
 *
1629
 * Returns pointer to node owning the link, or 0 if no matching link is found.
P
Per Liden 已提交
1630
 */
1631 1632
static struct tipc_node *tipc_link_find_owner(struct net *net,
					      const char *link_name,
1633
					      unsigned int *bearer_id)
P
Per Liden 已提交
1634
{
1635
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1636
	struct tipc_link *l_ptr;
1637
	struct tipc_node *n_ptr;
1638
	struct tipc_node *found_node = NULL;
1639
	int i;
P
Per Liden 已提交
1640

1641
	*bearer_id = 0;
1642
	rcu_read_lock();
1643
	list_for_each_entry_rcu(n_ptr, &tn->node_list, list) {
1644
		tipc_node_lock(n_ptr);
1645
		for (i = 0; i < MAX_BEARERS; i++) {
1646
			l_ptr = n_ptr->links[i].link;
1647 1648 1649 1650 1651
			if (l_ptr && !strcmp(l_ptr->name, link_name)) {
				*bearer_id = i;
				found_node = n_ptr;
				break;
			}
1652
		}
1653
		tipc_node_unlock(n_ptr);
1654 1655
		if (found_node)
			break;
1656
	}
1657 1658
	rcu_read_unlock();

1659
	return found_node;
P
Per Liden 已提交
1660 1661 1662 1663 1664 1665
}

/**
 * link_reset_statistics - reset link statistics
 * @l_ptr: pointer to link
 */
1666
static void link_reset_statistics(struct tipc_link *l_ptr)
P
Per Liden 已提交
1667 1668
{
	memset(&l_ptr->stats, 0, sizeof(l_ptr->stats));
1669 1670
	l_ptr->stats.sent_info = l_ptr->snd_nxt;
	l_ptr->stats.recv_info = l_ptr->rcv_nxt;
P
Per Liden 已提交
1671 1672
}

1673
static void link_print(struct tipc_link *l, const char *str)
P
Per Liden 已提交
1674
{
1675 1676 1677
	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;
1678

1679
	pr_info("%s Link <%s>:", str, l->name);
1680

1681
	if (link_probing(l))
1682
		pr_cont(":P\n");
1683
	else if (link_establishing(l))
1684
		pr_cont(":E\n");
1685
	else if (link_resetting(l))
1686
		pr_cont(":R\n");
1687
	else if (link_working(l))
1688
		pr_cont(":W\n");
1689 1690
	else
		pr_cont("\n");
1691 1692 1693 1694

	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
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 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733

/* 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;
}
1734

1735 1736 1737 1738 1739 1740 1741 1742 1743
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];
1744
	struct net *net = sock_net(skb->sk);
1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759

	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]);

1760 1761 1762
	if (strcmp(name, tipc_bclink_name) == 0)
		return tipc_nl_bc_link_set(net, attrs);

1763
	node = tipc_link_find_owner(net, name, &bearer_id);
1764 1765 1766 1767 1768
	if (!node)
		return -EINVAL;

	tipc_node_lock(node);

1769
	link = node->links[bearer_id].link;
1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788
	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]);
1789
			link->tolerance = tol;
1790
			tipc_link_proto_xmit(link, STATE_MSG, 0, 0, tol, 0);
1791 1792 1793 1794 1795 1796
		}
		if (props[TIPC_NLA_PROP_PRIO]) {
			u32 prio;

			prio = nla_get_u32(props[TIPC_NLA_PROP_PRIO]);
			link->priority = prio;
1797
			tipc_link_proto_xmit(link, STATE_MSG, 0, 0, 0, prio);
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811
		}
		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;
}
1812 1813

static int __tipc_nl_add_stats(struct sk_buff *skb, struct tipc_stats *s)
1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878
{
	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 */
1879
static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg,
1880
			      struct tipc_link *link, int nlflags)
1881 1882 1883 1884 1885
{
	int err;
	void *hdr;
	struct nlattr *attrs;
	struct nlattr *prop;
1886
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1887

1888
	hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
1889
			  nlflags, TIPC_NL_LINK_GET);
1890 1891 1892 1893 1894 1895 1896 1897 1898 1899
	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,
1900
			tipc_cluster_mask(tn->own_addr)))
1901
		goto attr_msg_full;
1902
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_MTU, link->mtu))
1903
		goto attr_msg_full;
1904
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, link->rcv_nxt))
1905
		goto attr_msg_full;
1906
	if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, link->snd_nxt))
1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923
		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,
1924
			link->window))
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
		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  */
1950 1951
static int __tipc_nl_add_node_links(struct net *net, struct tipc_nl_msg *msg,
				    struct tipc_node *node, u32 *prev_link)
1952 1953 1954 1955 1956 1957 1958
{
	u32 i;
	int err;

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

1959
		if (!node->links[i].link)
1960 1961
			continue;

1962 1963
		err = __tipc_nl_add_link(net, msg,
					 node->links[i].link, NLM_F_MULTI);
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
		if (err)
			return err;
	}
	*prev_link = 0;

	return 0;
}

int tipc_nl_link_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
1974 1975
	struct net *net = sock_net(skb->sk);
	struct tipc_net *tn = net_generic(net, tipc_net_id);
1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
	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) {
1992
		node = tipc_node_find(net, prev_node);
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
		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;
		}
2003
		tipc_node_put(node);
2004

2005 2006
		list_for_each_entry_continue_rcu(node, &tn->node_list,
						 list) {
2007
			tipc_node_lock(node);
2008 2009
			err = __tipc_nl_add_node_links(net, &msg, node,
						       &prev_link);
2010 2011 2012 2013 2014 2015 2016
			tipc_node_unlock(node);
			if (err)
				goto out;

			prev_node = node->addr;
		}
	} else {
2017
		err = tipc_nl_add_bc_link(net, &msg);
2018 2019 2020
		if (err)
			goto out;

2021
		list_for_each_entry_rcu(node, &tn->node_list, list) {
2022
			tipc_node_lock(node);
2023 2024
			err = __tipc_nl_add_node_links(net, &msg, node,
						       &prev_link);
2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044
			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)
{
2045
	struct net *net = genl_info_net(info);
2046 2047 2048 2049
	struct tipc_nl_msg msg;
	char *name;
	int err;

2050 2051 2052
	msg.portid = info->snd_portid;
	msg.seq = info->snd_seq;

2053 2054 2055 2056
	if (!info->attrs[TIPC_NLA_LINK_NAME])
		return -EINVAL;
	name = nla_data(info->attrs[TIPC_NLA_LINK_NAME]);

2057 2058
	msg.skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
	if (!msg.skb)
2059 2060
		return -ENOMEM;

2061 2062 2063 2064 2065 2066 2067 2068 2069 2070
	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;
2071

2072 2073 2074
		node = tipc_link_find_owner(net, name, &bearer_id);
		if (!node)
			return -EINVAL;
2075

2076
		tipc_node_lock(node);
2077
		link = node->links[bearer_id].link;
2078 2079 2080 2081 2082
		if (!link) {
			tipc_node_unlock(node);
			nlmsg_free(msg.skb);
			return -EINVAL;
		}
2083

2084 2085 2086 2087 2088 2089 2090
		err = __tipc_nl_add_link(net, &msg, link, 0);
		tipc_node_unlock(node);
		if (err) {
			nlmsg_free(msg.skb);
			return err;
		}
	}
2091

2092
	return genlmsg_reply(msg.skb, info);
2093
}
2094 2095 2096 2097 2098 2099 2100 2101 2102

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];
2103
	struct net *net = sock_net(skb->sk);
2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119

	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) {
2120
		err = tipc_bclink_reset_stats(net);
2121 2122 2123 2124 2125
		if (err)
			return err;
		return 0;
	}

2126
	node = tipc_link_find_owner(net, link_name, &bearer_id);
2127 2128 2129 2130 2131
	if (!node)
		return -EINVAL;

	tipc_node_lock(node);

2132
	link = node->links[bearer_id].link;
2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143
	if (!link) {
		tipc_node_unlock(node);
		return -EINVAL;
	}

	link_reset_statistics(link);

	tipc_node_unlock(node);

	return 0;
}