net/tipc/bcast.c

   1 /*
   2  * net/tipc/bcast.c: TIPC broadcast code
   3  *
   4  * Copyright (c) 2004-2006, Ericsson AB
   5  * Copyright (c) 2004, Intel Corporation.
   6  * Copyright (c) 2005, 2010-2011, Wind River Systems
   7  * All rights reserved.
   8  *
   9  * Redistribution and use in source and binary forms, with or without
  10  * modification, are permitted provided that the following conditions are met:
  11  *
  12  * 1. Redistributions of source code must retain the above copyright
  13  *    notice, this list of conditions and the following disclaimer.
  14  * 2. Redistributions in binary form must reproduce the above copyright
  15  *    notice, this list of conditions and the following disclaimer in the
  16  *    documentation and/or other materials provided with the distribution.
  17  * 3. Neither the names of the copyright holders nor the names of its
  18  *    contributors may be used to endorse or promote products derived from
  19  *    this software without specific prior written permission.
  20  *
  21  * Alternatively, this software may be distributed under the terms of the
  22  * GNU General Public License ("GPL") version 2 as published by the Free
  23  * Software Foundation.
  24  *
  25  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
  26  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  28  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
  29  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
  30  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
  31  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
  32  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
  33  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  34  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  35  * POSSIBILITY OF SUCH DAMAGE.
  36  */
  37
  38 #include "core.h"
  39 #include "link.h"
  40 #include "port.h"
  41 #include "bcast.h"
  42 #include "name_distr.h"
  43
  44 #define MAX_PKT_DEFAULT_MCAST 1500      /* bcast link max packet size (fixed) */
  45
  46 #define BCLINK_WIN_DEFAULT 20           /* bcast link window size (default) */
  47
  48 /**
  49  * struct tipc_bcbearer_pair - a pair of bearers used by broadcast link
  50  * @primary: pointer to primary bearer
  51  * @secondary: pointer to secondary bearer
  52  *
  53  * Bearers must have same priority and same set of reachable destinations
  54  * to be paired.
  55  */
  56
  57 struct tipc_bcbearer_pair {
  58         struct tipc_bearer *primary;
  59         struct tipc_bearer *secondary;
  60 };
  61
  62 /**
  63  * struct tipc_bcbearer - bearer used by broadcast link
  64  * @bearer: (non-standard) broadcast bearer structure
  65  * @media: (non-standard) broadcast media structure
  66  * @bpairs: array of bearer pairs
  67  * @bpairs_temp: temporary array of bearer pairs used by tipc_bcbearer_sort()
  68  * @remains: temporary node map used by tipc_bcbearer_send()
  69  * @remains_new: temporary node map used tipc_bcbearer_send()
  70  *
  71  * Note: The fields labelled "temporary" are incorporated into the bearer
  72  * to avoid consuming potentially limited stack space through the use of
  73  * large local variables within multicast routines.  Concurrent access is
  74  * prevented through use of the spinlock "bc_lock".
  75  */
  76 struct tipc_bcbearer {
  77         struct tipc_bearer bearer;
  78         struct tipc_media media;
  79         struct tipc_bcbearer_pair bpairs[MAX_BEARERS];
  80         struct tipc_bcbearer_pair bpairs_temp[TIPC_MAX_LINK_PRI + 1];
  81         struct tipc_node_map remains;
  82         struct tipc_node_map remains_new;
  83 };
  84
  85 /**
  86  * struct tipc_bclink - link used for broadcast messages
  87  * @link: (non-standard) broadcast link structure
  88  * @node: (non-standard) node structure representing b'cast link's peer node
  89  * @bcast_nodes: map of broadcast-capable nodes
  90  * @retransmit_to: node that most recently requested a retransmit
  91  *
  92  * Handles sequence numbering, fragmentation, bundling, etc.
  93  */
  94 struct tipc_bclink {
  95         struct tipc_link link;
  96         struct tipc_node node;
  97         struct tipc_node_map bcast_nodes;
  98         struct tipc_node *retransmit_to;
  99 };
 100
 101 static struct tipc_bcbearer bcast_bearer;
 102 static struct tipc_bclink bcast_link;
 103
 104 static struct tipc_bcbearer *bcbearer = &bcast_bearer;
 105 static struct tipc_bclink *bclink = &bcast_link;
 106 static struct tipc_link *bcl = &bcast_link.link;
 107
 108 static DEFINE_SPINLOCK(bc_lock);
 109
 110 const char tipc_bclink_name[] = "broadcast-link";
 111
 112 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
 113                            struct tipc_node_map *nm_b,
 114                            struct tipc_node_map *nm_diff);
 115
 116 static u32 bcbuf_acks(struct sk_buff *buf)
 117 {
 118         return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle;
 119 }
 120
 121 static void bcbuf_set_acks(struct sk_buff *buf, u32 acks)
 122 {
 123         TIPC_SKB_CB(buf)->handle = (void *)(unsigned long)acks;
 124 }
 125
 126 static void bcbuf_decr_acks(struct sk_buff *buf)
 127 {
 128         bcbuf_set_acks(buf, bcbuf_acks(buf) - 1);
 129 }
 130
 131 void tipc_bclink_add_node(u32 addr)
 132 {
 133         spin_lock_bh(&bc_lock);
 134         tipc_nmap_add(&bclink->bcast_nodes, addr);
 135         spin_unlock_bh(&bc_lock);
 136 }
 137
 138 void tipc_bclink_remove_node(u32 addr)
 139 {
 140         spin_lock_bh(&bc_lock);
 141         tipc_nmap_remove(&bclink->bcast_nodes, addr);
 142         spin_unlock_bh(&bc_lock);
 143 }
 144
 145 static void bclink_set_last_sent(void)
 146 {
 147         if (bcl->next_out)
 148                 bcl->fsm_msg_cnt = mod(buf_seqno(bcl->next_out) - 1);
 149         else
 150                 bcl->fsm_msg_cnt = mod(bcl->next_out_no - 1);
 151 }
 152
 153 u32 tipc_bclink_get_last_sent(void)
 154 {
 155         return bcl->fsm_msg_cnt;
 156 }
 157
 158 static void bclink_update_last_sent(struct tipc_node *node, u32 seqno)
 159 {
 160         node->bclink.last_sent = less_eq(node->bclink.last_sent, seqno) ?
 161                                                 seqno : node->bclink.last_sent;
 162 }
 163
 164
 165 /**
 166  * tipc_bclink_retransmit_to - get most recent node to request retransmission
 167  *
 168  * Called with bc_lock locked
 169  */
 170 struct tipc_node *tipc_bclink_retransmit_to(void)
 171 {
 172         return bclink->retransmit_to;
 173 }
 174
 175 /**
 176  * bclink_retransmit_pkt - retransmit broadcast packets
 177  * @after: sequence number of last packet to *not* retransmit
 178  * @to: sequence number of last packet to retransmit
 179  *
 180  * Called with bc_lock locked
 181  */
 182 static void bclink_retransmit_pkt(u32 after, u32 to)
 183 {
 184         struct sk_buff *buf;
 185
 186         buf = bcl->first_out;
 187         while (buf && less_eq(buf_seqno(buf), after))
 188                 buf = buf->next;
 189         tipc_link_retransmit(bcl, buf, mod(to - after));
 190 }
 191
 192 /**
 193  * tipc_bclink_acknowledge - handle acknowledgement of broadcast packets
 194  * @n_ptr: node that sent acknowledgement info
 195  * @acked: broadcast sequence # that has been acknowledged
 196  *
 197  * Node is locked, bc_lock unlocked.
 198  */
 199 void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked)
 200 {
 201         struct sk_buff *crs;
 202         struct sk_buff *next;
 203         unsigned int released = 0;
 204
 205         spin_lock_bh(&bc_lock);
 206
 207         /* Bail out if tx queue is empty (no clean up is required) */
 208         crs = bcl->first_out;
 209         if (!crs)
 210                 goto exit;
 211
 212         /* Determine which messages need to be acknowledged */
 213         if (acked == INVALID_LINK_SEQ) {
 214                 /*
 215                  * Contact with specified node has been lost, so need to
 216                  * acknowledge sent messages only (if other nodes still exist)
 217                  * or both sent and unsent messages (otherwise)
 218                  */
 219                 if (bclink->bcast_nodes.count)
 220                         acked = bcl->fsm_msg_cnt;
 221                 else
 222                         acked = bcl->next_out_no;
 223         } else {
 224                 /*
 225                  * Bail out if specified sequence number does not correspond
 226                  * to a message that has been sent and not yet acknowledged
 227                  */
 228                 if (less(acked, buf_seqno(crs)) ||
 229                     less(bcl->fsm_msg_cnt, acked) ||
 230                     less_eq(acked, n_ptr->bclink.acked))
 231                         goto exit;
 232         }
 233
 234         /* Skip over packets that node has previously acknowledged */
 235         while (crs && less_eq(buf_seqno(crs), n_ptr->bclink.acked))
 236                 crs = crs->next;
 237
 238         /* Update packets that node is now acknowledging */
 239
 240         while (crs && less_eq(buf_seqno(crs), acked)) {
 241                 next = crs->next;
 242
 243                 if (crs != bcl->next_out)
 244                         bcbuf_decr_acks(crs);
 245                 else {
 246                         bcbuf_set_acks(crs, 0);
 247                         bcl->next_out = next;
 248                         bclink_set_last_sent();
 249                 }
 250
 251                 if (bcbuf_acks(crs) == 0) {
 252                         bcl->first_out = next;
 253                         bcl->out_queue_size--;
 254                         kfree_skb(crs);
 255                         released = 1;
 256                 }
 257                 crs = next;
 258         }
 259         n_ptr->bclink.acked = acked;
 260
 261         /* Try resolving broadcast link congestion, if necessary */
 262
 263         if (unlikely(bcl->next_out)) {
 264                 tipc_link_push_queue(bcl);
 265                 bclink_set_last_sent();
 266         }
 267         if (unlikely(released && !list_empty(&bcl->waiting_ports)))
 268                 tipc_link_wakeup_ports(bcl, 0);
 269 exit:
 270         spin_unlock_bh(&bc_lock);
 271 }
 272
 273 /**
 274  * tipc_bclink_update_link_state - update broadcast link state
 275  *
 276  * tipc_net_lock and node lock set
 277  */
 278 void tipc_bclink_update_link_state(struct tipc_node *n_ptr, u32 last_sent)
 279 {
 280         struct sk_buff *buf;
 281
 282         /* Ignore "stale" link state info */
 283
 284         if (less_eq(last_sent, n_ptr->bclink.last_in))
 285                 return;
 286
 287         /* Update link synchronization state; quit if in sync */
 288
 289         bclink_update_last_sent(n_ptr, last_sent);
 290
 291         if (n_ptr->bclink.last_sent == n_ptr->bclink.last_in)
 292                 return;
 293
 294         /* Update out-of-sync state; quit if loss is still unconfirmed */
 295
 296         if ((++n_ptr->bclink.oos_state) == 1) {
 297                 if (n_ptr->bclink.deferred_size < (TIPC_MIN_LINK_WIN / 2))
 298                         return;
 299                 n_ptr->bclink.oos_state++;
 300         }
 301
 302         /* Don't NACK if one has been recently sent (or seen) */
 303
 304         if (n_ptr->bclink.oos_state & 0x1)
 305                 return;
 306
 307         /* Send NACK */
 308
 309         buf = tipc_buf_acquire(INT_H_SIZE);
 310         if (buf) {
 311                 struct tipc_msg *msg = buf_msg(buf);
 312
 313                 tipc_msg_init(msg, BCAST_PROTOCOL, STATE_MSG,
 314                               INT_H_SIZE, n_ptr->addr);
 315                 msg_set_non_seq(msg, 1);
 316                 msg_set_mc_netid(msg, tipc_net_id);
 317                 msg_set_bcast_ack(msg, n_ptr->bclink.last_in);
 318                 msg_set_bcgap_after(msg, n_ptr->bclink.last_in);
 319                 msg_set_bcgap_to(msg, n_ptr->bclink.deferred_head
 320                                  ? buf_seqno(n_ptr->bclink.deferred_head) - 1
 321                                  : n_ptr->bclink.last_sent);
 322
 323                 spin_lock_bh(&bc_lock);
 324                 tipc_bearer_send(&bcbearer->bearer, buf, NULL);
 325                 bcl->stats.sent_nacks++;
 326                 spin_unlock_bh(&bc_lock);
 327                 kfree_skb(buf);
 328
 329                 n_ptr->bclink.oos_state++;
 330         }
 331 }
 332
 333 /**
 334  * bclink_peek_nack - monitor retransmission requests sent by other nodes
 335  *
 336  * Delay any upcoming NACK by this node if another node has already
 337  * requested the first message this node is going to ask for.
 338  *
 339  * Only tipc_net_lock set.
 340  */
 341 static void bclink_peek_nack(struct tipc_msg *msg)
 342 {
 343         struct tipc_node *n_ptr = tipc_node_find(msg_destnode(msg));
 344
 345         if (unlikely(!n_ptr))
 346                 return;
 347
 348         tipc_node_lock(n_ptr);
 349
 350         if (n_ptr->bclink.recv_permitted &&
 351             (n_ptr->bclink.last_in != n_ptr->bclink.last_sent) &&
 352             (n_ptr->bclink.last_in == msg_bcgap_after(msg)))
 353                 n_ptr->bclink.oos_state = 2;
 354
 355         tipc_node_unlock(n_ptr);
 356 }
 357
 358 /*
 359  * tipc_bclink_send_msg - broadcast a packet to all nodes in cluster
 360  */
 361 int tipc_bclink_send_msg(struct sk_buff *buf)
 362 {
 363         int res;
 364
 365         spin_lock_bh(&bc_lock);
 366
 367         if (!bclink->bcast_nodes.count) {
 368                 res = msg_data_sz(buf_msg(buf));
 369                 kfree_skb(buf);
 370                 goto exit;
 371         }
 372
 373         res = tipc_link_send_buf(bcl, buf);
 374         if (likely(res >= 0)) {
 375                 bclink_set_last_sent();
 376                 bcl->stats.queue_sz_counts++;
 377                 bcl->stats.accu_queue_sz += bcl->out_queue_size;
 378         }
 379 exit:
 380         spin_unlock_bh(&bc_lock);
 381         return res;
 382 }
 383
 384 /**
 385  * bclink_accept_pkt - accept an incoming, in-sequence broadcast packet
 386  *
 387  * Called with both sending node's lock and bc_lock taken.
 388  */
 389 static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
 390 {
 391         bclink_update_last_sent(node, seqno);
 392         node->bclink.last_in = seqno;
 393         node->bclink.oos_state = 0;
 394         bcl->stats.recv_info++;
 395
 396         /*
 397          * Unicast an ACK periodically, ensuring that
 398          * all nodes in the cluster don't ACK at the same time
 399          */
 400
 401         if (((seqno - tipc_own_addr) % TIPC_MIN_LINK_WIN) == 0) {
 402                 tipc_link_send_proto_msg(
 403                         node->active_links[node->addr & 1],
 404                         STATE_MSG, 0, 0, 0, 0, 0);
 405                 bcl->stats.sent_acks++;
 406         }
 407 }
 408
 409 /**
 410  * tipc_bclink_recv_pkt - receive a broadcast packet, and deliver upwards
 411  *
 412  * tipc_net_lock is read_locked, no other locks set
 413  */
 414 void tipc_bclink_recv_pkt(struct sk_buff *buf)
 415 {
 416         struct tipc_msg *msg = buf_msg(buf);
 417         struct tipc_node *node;
 418         u32 next_in;
 419         u32 seqno;
 420         int deferred;
 421
 422         /* Screen out unwanted broadcast messages */
 423
 424         if (msg_mc_netid(msg) != tipc_net_id)
 425                 goto exit;
 426
 427         node = tipc_node_find(msg_prevnode(msg));
 428         if (unlikely(!node))
 429                 goto exit;
 430
 431         tipc_node_lock(node);
 432         if (unlikely(!node->bclink.recv_permitted))
 433                 goto unlock;
 434
 435         /* Handle broadcast protocol message */
 436
 437         if (unlikely(msg_user(msg) == BCAST_PROTOCOL)) {
 438                 if (msg_type(msg) != STATE_MSG)
 439                         goto unlock;
 440                 if (msg_destnode(msg) == tipc_own_addr) {
 441                         tipc_bclink_acknowledge(node, msg_bcast_ack(msg));
 442                         tipc_node_unlock(node);
 443                         spin_lock_bh(&bc_lock);
 444                         bcl->stats.recv_nacks++;
 445                         bclink->retransmit_to = node;
 446                         bclink_retransmit_pkt(msg_bcgap_after(msg),
 447                                               msg_bcgap_to(msg));
 448                         spin_unlock_bh(&bc_lock);
 449                 } else {
 450                         tipc_node_unlock(node);
 451                         bclink_peek_nack(msg);
 452                 }
 453                 goto exit;
 454         }
 455
 456         /* Handle in-sequence broadcast message */
 457
 458         seqno = msg_seqno(msg);
 459         next_in = mod(node->bclink.last_in + 1);
 460
 461         if (likely(seqno == next_in)) {
 462 receive:
 463                 /* Deliver message to destination */
 464
 465                 if (likely(msg_isdata(msg))) {
 466                         spin_lock_bh(&bc_lock);
 467                         bclink_accept_pkt(node, seqno);
 468                         spin_unlock_bh(&bc_lock);
 469                         tipc_node_unlock(node);
 470                         if (likely(msg_mcast(msg)))
 471                                 tipc_port_recv_mcast(buf, NULL);
 472                         else
 473                                 kfree_skb(buf);
 474                 } else if (msg_user(msg) == MSG_BUNDLER) {
 475                         spin_lock_bh(&bc_lock);
 476                         bclink_accept_pkt(node, seqno);
 477                         bcl->stats.recv_bundles++;
 478                         bcl->stats.recv_bundled += msg_msgcnt(msg);
 479                         spin_unlock_bh(&bc_lock);
 480                         tipc_node_unlock(node);
 481                         tipc_link_recv_bundle(buf);
 482                 } else if (msg_user(msg) == MSG_FRAGMENTER) {
 483                         int ret;
 484                         ret = tipc_link_recv_fragment(&node->bclink.reasm_head,
 485                                                       &node->bclink.reasm_tail,
 486                                                       &buf);
 487                         if (ret == LINK_REASM_ERROR)
 488                                 goto unlock;
 489                         spin_lock_bh(&bc_lock);
 490                         bclink_accept_pkt(node, seqno);
 491                         bcl->stats.recv_fragments++;
 492                         if (ret == LINK_REASM_COMPLETE) {
 493                                 bcl->stats.recv_fragmented++;
 494                                 /* Point msg to inner header */
 495                                 msg = buf_msg(buf);
 496                                 spin_unlock_bh(&bc_lock);
 497                                 goto receive;
 498                         }
 499                         spin_unlock_bh(&bc_lock);
 500                         tipc_node_unlock(node);
 501                 } else if (msg_user(msg) == NAME_DISTRIBUTOR) {
 502                         spin_lock_bh(&bc_lock);
 503                         bclink_accept_pkt(node, seqno);
 504                         spin_unlock_bh(&bc_lock);
 505                         tipc_node_unlock(node);
 506                         tipc_named_recv(buf);
 507                 } else {
 508                         spin_lock_bh(&bc_lock);
 509                         bclink_accept_pkt(node, seqno);
 510                         spin_unlock_bh(&bc_lock);
 511                         tipc_node_unlock(node);
 512                         kfree_skb(buf);
 513                 }
 514                 buf = NULL;
 515
 516                 /* Determine new synchronization state */
 517
 518                 tipc_node_lock(node);
 519                 if (unlikely(!tipc_node_is_up(node)))
 520                         goto unlock;
 521
 522                 if (node->bclink.last_in == node->bclink.last_sent)
 523                         goto unlock;
 524
 525                 if (!node->bclink.deferred_head) {
 526                         node->bclink.oos_state = 1;
 527                         goto unlock;
 528                 }
 529
 530                 msg = buf_msg(node->bclink.deferred_head);
 531                 seqno = msg_seqno(msg);
 532                 next_in = mod(next_in + 1);
 533                 if (seqno != next_in)
 534                         goto unlock;
 535
 536                 /* Take in-sequence message from deferred queue & deliver it */
 537
 538                 buf = node->bclink.deferred_head;
 539                 node->bclink.deferred_head = buf->next;
 540                 node->bclink.deferred_size--;
 541                 goto receive;
 542         }
 543
 544         /* Handle out-of-sequence broadcast message */
 545
 546         if (less(next_in, seqno)) {
 547                 deferred = tipc_link_defer_pkt(&node->bclink.deferred_head,
 548                                                &node->bclink.deferred_tail,
 549                                                buf);
 550                 node->bclink.deferred_size += deferred;
 551                 bclink_update_last_sent(node, seqno);
 552                 buf = NULL;
 553         } else
 554                 deferred = 0;
 555
 556         spin_lock_bh(&bc_lock);
 557
 558         if (deferred)
 559                 bcl->stats.deferred_recv++;
 560         else
 561                 bcl->stats.duplicates++;
 562
 563         spin_unlock_bh(&bc_lock);
 564
 565 unlock:
 566         tipc_node_unlock(node);
 567 exit:
 568         kfree_skb(buf);
 569 }
 570
 571 u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr)
 572 {
 573         return (n_ptr->bclink.recv_permitted &&
 574                 (tipc_bclink_get_last_sent() != n_ptr->bclink.acked));
 575 }
 576
 577
 578 /**
 579  * tipc_bcbearer_send - send a packet through the broadcast pseudo-bearer
 580  *
 581  * Send packet over as many bearers as necessary to reach all nodes
 582  * that have joined the broadcast link.
 583  *
 584  * Returns 0 (packet sent successfully) under all circumstances,
 585  * since the broadcast link's pseudo-bearer never blocks
 586  */
 587 static int tipc_bcbearer_send(struct sk_buff *buf, struct tipc_bearer *unused1,
 588                               struct tipc_media_addr *unused2)
 589 {
 590         int bp_index;
 591
 592         /* Prepare broadcast link message for reliable transmission,
 593          * if first time trying to send it;
 594          * preparation is skipped for broadcast link protocol messages
 595          * since they are sent in an unreliable manner and don't need it
 596          */
 597         if (likely(!msg_non_seq(buf_msg(buf)))) {
 598                 struct tipc_msg *msg;
 599
 600                 bcbuf_set_acks(buf, bclink->bcast_nodes.count);
 601                 msg = buf_msg(buf);
 602                 msg_set_non_seq(msg, 1);
 603                 msg_set_mc_netid(msg, tipc_net_id);
 604                 bcl->stats.sent_info++;
 605
 606                 if (WARN_ON(!bclink->bcast_nodes.count)) {
 607                         dump_stack();
 608                         return 0;
 609                 }
 610         }
 611
 612         /* Send buffer over bearers until all targets reached */
 613         bcbearer->remains = bclink->bcast_nodes;
 614
 615         for (bp_index = 0; bp_index < MAX_BEARERS; bp_index++) {
 616                 struct tipc_bearer *p = bcbearer->bpairs[bp_index].primary;
 617                 struct tipc_bearer *s = bcbearer->bpairs[bp_index].secondary;
 618                 struct tipc_bearer *b = p;
 619                 struct sk_buff *tbuf;
 620
 621                 if (!p)
 622                         break; /* No more bearers to try */
 623
 624                 tipc_nmap_diff(&bcbearer->remains, &b->nodes,
 625                                &bcbearer->remains_new);
 626                 if (bcbearer->remains_new.count == bcbearer->remains.count)
 627                         continue; /* Nothing added by bearer pair */
 628
 629                 if (bp_index == 0) {
 630                         /* Use original buffer for first bearer */
 631                         tipc_bearer_send(b, buf, &b->bcast_addr);
 632                 } else {
 633                         /* Avoid concurrent buffer access */
 634                         tbuf = pskb_copy(buf, GFP_ATOMIC);
 635                         if (!tbuf)
 636                                 break;
 637                         tipc_bearer_send(b, tbuf, &b->bcast_addr);
 638                         kfree_skb(tbuf); /* Bearer keeps a clone */
 639                 }
 640
 641                 /* Swap bearers for next packet */
 642                 if (s) {
 643                         bcbearer->bpairs[bp_index].primary = s;
 644                         bcbearer->bpairs[bp_index].secondary = p;
 645                 }
 646
 647                 if (bcbearer->remains_new.count == 0)
 648                         break; /* All targets reached */
 649
 650                 bcbearer->remains = bcbearer->remains_new;
 651         }
 652
 653         return 0;
 654 }
 655
 656 /**
 657  * tipc_bcbearer_sort - create sets of bearer pairs used by broadcast bearer
 658  */
 659 void tipc_bcbearer_sort(void)
 660 {
 661         struct tipc_bcbearer_pair *bp_temp = bcbearer->bpairs_temp;
 662         struct tipc_bcbearer_pair *bp_curr;
 663         int b_index;
 664         int pri;
 665
 666         spin_lock_bh(&bc_lock);
 667
 668         /* Group bearers by priority (can assume max of two per priority) */
 669         memset(bp_temp, 0, sizeof(bcbearer->bpairs_temp));
 670
 671         for (b_index = 0; b_index < MAX_BEARERS; b_index++) {
 672                 struct tipc_bearer *b = &tipc_bearers[b_index];
 673
 674                 if (!b->active || !b->nodes.count)
 675                         continue;
 676
 677                 if (!bp_temp[b->priority].primary)
 678                         bp_temp[b->priority].primary = b;
 679                 else
 680                         bp_temp[b->priority].secondary = b;
 681         }
 682
 683         /* Create array of bearer pairs for broadcasting */
 684         bp_curr = bcbearer->bpairs;
 685         memset(bcbearer->bpairs, 0, sizeof(bcbearer->bpairs));
 686
 687         for (pri = TIPC_MAX_LINK_PRI; pri >= 0; pri--) {
 688
 689                 if (!bp_temp[pri].primary)
 690                         continue;
 691
 692                 bp_curr->primary = bp_temp[pri].primary;
 693
 694                 if (bp_temp[pri].secondary) {
 695                         if (tipc_nmap_equal(&bp_temp[pri].primary->nodes,
 696                                             &bp_temp[pri].secondary->nodes)) {
 697                                 bp_curr->secondary = bp_temp[pri].secondary;
 698                         } else {
 699                                 bp_curr++;
 700                                 bp_curr->primary = bp_temp[pri].secondary;
 701                         }
 702                 }
 703
 704                 bp_curr++;
 705         }
 706
 707         spin_unlock_bh(&bc_lock);
 708 }
 709
 710
 711 int tipc_bclink_stats(char *buf, const u32 buf_size)
 712 {
 713         int ret;
 714         struct tipc_stats *s;
 715
 716         if (!bcl)
 717                 return 0;
 718
 719         spin_lock_bh(&bc_lock);
 720
 721         s = &bcl->stats;
 722
 723         ret = tipc_snprintf(buf, buf_size, "Link <%s>\n"
 724                             "  Window:%u packets\n",
 725                             bcl->name, bcl->queue_limit[0]);
 726         ret += tipc_snprintf(buf + ret, buf_size - ret,
 727                              "  RX packets:%u fragments:%u/%u bundles:%u/%u\n",
 728                              s->recv_info, s->recv_fragments,
 729                              s->recv_fragmented, s->recv_bundles,
 730                              s->recv_bundled);
 731         ret += tipc_snprintf(buf + ret, buf_size - ret,
 732                              "  TX packets:%u fragments:%u/%u bundles:%u/%u\n",
 733                              s->sent_info, s->sent_fragments,
 734                              s->sent_fragmented, s->sent_bundles,
 735                              s->sent_bundled);
 736         ret += tipc_snprintf(buf + ret, buf_size - ret,
 737                              "  RX naks:%u defs:%u dups:%u\n",
 738                              s->recv_nacks, s->deferred_recv, s->duplicates);
 739         ret += tipc_snprintf(buf + ret, buf_size - ret,
 740                              "  TX naks:%u acks:%u dups:%u\n",
 741                              s->sent_nacks, s->sent_acks, s->retransmitted);
 742         ret += tipc_snprintf(buf + ret, buf_size - ret,
 743                              "  Congestion link:%u  Send queue max:%u avg:%u\n",
 744                              s->link_congs, s->max_queue_sz,
 745                              s->queue_sz_counts ?
 746                              (s->accu_queue_sz / s->queue_sz_counts) : 0);
 747
 748         spin_unlock_bh(&bc_lock);
 749         return ret;
 750 }
 751
 752 int tipc_bclink_reset_stats(void)
 753 {
 754         if (!bcl)
 755                 return -ENOPROTOOPT;
 756
 757         spin_lock_bh(&bc_lock);
 758         memset(&bcl->stats, 0, sizeof(bcl->stats));
 759         spin_unlock_bh(&bc_lock);
 760         return 0;
 761 }
 762
 763 int tipc_bclink_set_queue_limits(u32 limit)
 764 {
 765         if (!bcl)
 766                 return -ENOPROTOOPT;
 767         if ((limit < TIPC_MIN_LINK_WIN) || (limit > TIPC_MAX_LINK_WIN))
 768                 return -EINVAL;
 769
 770         spin_lock_bh(&bc_lock);
 771         tipc_link_set_queue_limits(bcl, limit);
 772         spin_unlock_bh(&bc_lock);
 773         return 0;
 774 }
 775
 776 void tipc_bclink_init(void)
 777 {
 778         bcbearer->bearer.media = &bcbearer->media;
 779         bcbearer->media.send_msg = tipc_bcbearer_send;
 780         sprintf(bcbearer->media.name, "tipc-broadcast");
 781
 782         INIT_LIST_HEAD(&bcl->waiting_ports);
 783         bcl->next_out_no = 1;
 784         spin_lock_init(&bclink->node.lock);
 785         bcl->owner = &bclink->node;
 786         bcl->max_pkt = MAX_PKT_DEFAULT_MCAST;
 787         tipc_link_set_queue_limits(bcl, BCLINK_WIN_DEFAULT);
 788         spin_lock_init(&bcbearer->bearer.lock);
 789         bcl->b_ptr = &bcbearer->bearer;
 790         bcl->state = WORKING_WORKING;
 791         strlcpy(bcl->name, tipc_bclink_name, TIPC_MAX_LINK_NAME);
 792 }
 793
 794 void tipc_bclink_stop(void)
 795 {
 796         spin_lock_bh(&bc_lock);
 797         tipc_link_stop(bcl);
 798         spin_unlock_bh(&bc_lock);
 799
 800         memset(bclink, 0, sizeof(*bclink));
 801         memset(bcbearer, 0, sizeof(*bcbearer));
 802 }
 803
 804
 805 /**
 806  * tipc_nmap_add - add a node to a node map
 807  */
 808 void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node)
 809 {
 810         int n = tipc_node(node);
 811         int w = n / WSIZE;
 812         u32 mask = (1 << (n % WSIZE));
 813
 814         if ((nm_ptr->map[w] & mask) == 0) {
 815                 nm_ptr->count++;
 816                 nm_ptr->map[w] |= mask;
 817         }
 818 }
 819
 820 /**
 821  * tipc_nmap_remove - remove a node from a node map
 822  */
 823 void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node)
 824 {
 825         int n = tipc_node(node);
 826         int w = n / WSIZE;
 827         u32 mask = (1 << (n % WSIZE));
 828
 829         if ((nm_ptr->map[w] & mask) != 0) {
 830                 nm_ptr->map[w] &= ~mask;
 831                 nm_ptr->count--;
 832         }
 833 }
 834
 835 /**
 836  * tipc_nmap_diff - find differences between node maps
 837  * @nm_a: input node map A
 838  * @nm_b: input node map B
 839  * @nm_diff: output node map A-B (i.e. nodes of A that are not in B)
 840  */
 841 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
 842                            struct tipc_node_map *nm_b,
 843                            struct tipc_node_map *nm_diff)
 844 {
 845         int stop = ARRAY_SIZE(nm_a->map);
 846         int w;
 847         int b;
 848         u32 map;
 849
 850         memset(nm_diff, 0, sizeof(*nm_diff));
 851         for (w = 0; w < stop; w++) {
 852                 map = nm_a->map[w] ^ (nm_a->map[w] & nm_b->map[w]);
 853                 nm_diff->map[w] = map;
 854                 if (map != 0) {
 855                         for (b = 0 ; b < WSIZE; b++) {
 856                                 if (map & (1 << b))
 857                                         nm_diff->count++;
 858                         }
 859                 }
 860         }
 861 }
 862
 863 /**
 864  * tipc_port_list_add - add a port to a port list, ensuring no duplicates
 865  */
 866 void tipc_port_list_add(struct tipc_port_list *pl_ptr, u32 port)
 867 {
 868         struct tipc_port_list *item = pl_ptr;
 869         int i;
 870         int item_sz = PLSIZE;
 871         int cnt = pl_ptr->count;
 872
 873         for (; ; cnt -= item_sz, item = item->next) {
 874                 if (cnt < PLSIZE)
 875                         item_sz = cnt;
 876                 for (i = 0; i < item_sz; i++)
 877                         if (item->ports[i] == port)
 878                                 return;
 879                 if (i < PLSIZE) {
 880                         item->ports[i] = port;
 881                         pl_ptr->count++;
 882                         return;
 883                 }
 884                 if (!item->next) {
 885                         item->next = kmalloc(sizeof(*item), GFP_ATOMIC);
 886                         if (!item->next) {
 887                                 pr_warn("Incomplete multicast delivery, no memory\n");
 888                                 return;
 889                         }
 890                         item->next->next = NULL;
 891                 }
 892         }
 893 }
 894
 895 /**
 896  * tipc_port_list_free - free dynamically created entries in port_list chain
 897  *
 898  */
 899 void tipc_port_list_free(struct tipc_port_list *pl_ptr)
 900 {
 901         struct tipc_port_list *item;
 902         struct tipc_port_list *next;
 903
 904         for (item = pl_ptr->next; item; item = next) {
 905                 next = item->next;
 906                 kfree(item);
 907         }
 908 }