net/tipc/bcast.c

   1 /*
   2  * net/tipc/bcast.c: TIPC broadcast code
   3  *
   4  * Copyright (c) 2004-2006, 2014-2015, 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 "socket.h"
  39 #include "msg.h"
  40 #include "bcast.h"
  41 #include "name_distr.h"
  42 #include "core.h"
  43
  44 #define MAX_PKT_DEFAULT_MCAST   1500    /* bcast link max packet size (fixed) */
  45 #define BCLINK_WIN_DEFAULT      20      /* bcast link window size (default) */
  46
  47 const char tipc_bclink_name[] = "broadcast-link";
  48
  49 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
  50                            struct tipc_node_map *nm_b,
  51                            struct tipc_node_map *nm_diff);
  52 static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node);
  53 static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node);
  54
  55 static void tipc_bclink_lock(struct net *net)
  56 {
  57         struct tipc_net *tn = net_generic(net, tipc_net_id);
  58
  59         spin_lock_bh(&tn->bclink->lock);
  60 }
  61
  62 static void tipc_bclink_unlock(struct net *net)
  63 {
  64         struct tipc_net *tn = net_generic(net, tipc_net_id);
  65         struct tipc_node *node = NULL;
  66
  67         if (likely(!tn->bclink->flags)) {
  68                 spin_unlock_bh(&tn->bclink->lock);
  69                 return;
  70         }
  71
  72         if (tn->bclink->flags & TIPC_BCLINK_RESET) {
  73                 tn->bclink->flags &= ~TIPC_BCLINK_RESET;
  74                 node = tipc_bclink_retransmit_to(net);
  75         }
  76         spin_unlock_bh(&tn->bclink->lock);
  77
  78         if (node)
  79                 tipc_link_reset_all(node);
  80 }
  81
  82 void tipc_bclink_input(struct net *net)
  83 {
  84         struct tipc_net *tn = net_generic(net, tipc_net_id);
  85
  86         tipc_sk_mcast_rcv(net, &tn->bclink->arrvq, &tn->bclink->inputq);
  87 }
  88
  89 uint  tipc_bclink_get_mtu(void)
  90 {
  91         return MAX_PKT_DEFAULT_MCAST;
  92 }
  93
  94 void tipc_bclink_set_flags(struct net *net, unsigned int flags)
  95 {
  96         struct tipc_net *tn = net_generic(net, tipc_net_id);
  97
  98         tn->bclink->flags |= flags;
  99 }
 100
 101 static u32 bcbuf_acks(struct sk_buff *buf)
 102 {
 103         return (u32)(unsigned long)TIPC_SKB_CB(buf)->handle;
 104 }
 105
 106 static void bcbuf_set_acks(struct sk_buff *buf, u32 acks)
 107 {
 108         TIPC_SKB_CB(buf)->handle = (void *)(unsigned long)acks;
 109 }
 110
 111 static void bcbuf_decr_acks(struct sk_buff *buf)
 112 {
 113         bcbuf_set_acks(buf, bcbuf_acks(buf) - 1);
 114 }
 115
 116 void tipc_bclink_add_node(struct net *net, u32 addr)
 117 {
 118         struct tipc_net *tn = net_generic(net, tipc_net_id);
 119
 120         tipc_bclink_lock(net);
 121         tipc_nmap_add(&tn->bclink->bcast_nodes, addr);
 122         tipc_bclink_unlock(net);
 123 }
 124
 125 void tipc_bclink_remove_node(struct net *net, u32 addr)
 126 {
 127         struct tipc_net *tn = net_generic(net, tipc_net_id);
 128
 129         tipc_bclink_lock(net);
 130         tipc_nmap_remove(&tn->bclink->bcast_nodes, addr);
 131         tipc_bclink_unlock(net);
 132 }
 133
 134 static void bclink_set_last_sent(struct net *net)
 135 {
 136         struct tipc_net *tn = net_generic(net, tipc_net_id);
 137         struct tipc_link *bcl = tn->bcl;
 138
 139         if (bcl->next_out)
 140                 bcl->fsm_msg_cnt = mod(buf_seqno(bcl->next_out) - 1);
 141         else
 142                 bcl->fsm_msg_cnt = mod(bcl->next_out_no - 1);
 143 }
 144
 145 u32 tipc_bclink_get_last_sent(struct net *net)
 146 {
 147         struct tipc_net *tn = net_generic(net, tipc_net_id);
 148
 149         return tn->bcl->fsm_msg_cnt;
 150 }
 151
 152 static void bclink_update_last_sent(struct tipc_node *node, u32 seqno)
 153 {
 154         node->bclink.last_sent = less_eq(node->bclink.last_sent, seqno) ?
 155                                                 seqno : node->bclink.last_sent;
 156 }
 157
 158
 159 /**
 160  * tipc_bclink_retransmit_to - get most recent node to request retransmission
 161  *
 162  * Called with bclink_lock locked
 163  */
 164 struct tipc_node *tipc_bclink_retransmit_to(struct net *net)
 165 {
 166         struct tipc_net *tn = net_generic(net, tipc_net_id);
 167
 168         return tn->bclink->retransmit_to;
 169 }
 170
 171 /**
 172  * bclink_retransmit_pkt - retransmit broadcast packets
 173  * @after: sequence number of last packet to *not* retransmit
 174  * @to: sequence number of last packet to retransmit
 175  *
 176  * Called with bclink_lock locked
 177  */
 178 static void bclink_retransmit_pkt(struct tipc_net *tn, u32 after, u32 to)
 179 {
 180         struct sk_buff *skb;
 181         struct tipc_link *bcl = tn->bcl;
 182
 183         skb_queue_walk(&bcl->outqueue, skb) {
 184                 if (more(buf_seqno(skb), after)) {
 185                         tipc_link_retransmit(bcl, skb, mod(to - after));
 186                         break;
 187                 }
 188         }
 189 }
 190
 191 /**
 192  * tipc_bclink_wakeup_users - wake up pending users
 193  *
 194  * Called with no locks taken
 195  */
 196 void tipc_bclink_wakeup_users(struct net *net)
 197 {
 198         struct tipc_net *tn = net_generic(net, tipc_net_id);
 199
 200         tipc_sk_rcv(net, &tn->bclink->link.wakeupq);
 201 }
 202
 203 /**
 204  * tipc_bclink_acknowledge - handle acknowledgement of broadcast packets
 205  * @n_ptr: node that sent acknowledgement info
 206  * @acked: broadcast sequence # that has been acknowledged
 207  *
 208  * Node is locked, bclink_lock unlocked.
 209  */
 210 void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked)
 211 {
 212         struct sk_buff *skb, *tmp;
 213         struct sk_buff *next;
 214         unsigned int released = 0;
 215         struct net *net = n_ptr->net;
 216         struct tipc_net *tn = net_generic(net, tipc_net_id);
 217
 218         tipc_bclink_lock(net);
 219         /* Bail out if tx queue is empty (no clean up is required) */
 220         skb = skb_peek(&tn->bcl->outqueue);
 221         if (!skb)
 222                 goto exit;
 223
 224         /* Determine which messages need to be acknowledged */
 225         if (acked == INVALID_LINK_SEQ) {
 226                 /*
 227                  * Contact with specified node has been lost, so need to
 228                  * acknowledge sent messages only (if other nodes still exist)
 229                  * or both sent and unsent messages (otherwise)
 230                  */
 231                 if (tn->bclink->bcast_nodes.count)
 232                         acked = tn->bcl->fsm_msg_cnt;
 233                 else
 234                         acked = tn->bcl->next_out_no;
 235         } else {
 236                 /*
 237                  * Bail out if specified sequence number does not correspond
 238                  * to a message that has been sent and not yet acknowledged
 239                  */
 240                 if (less(acked, buf_seqno(skb)) ||
 241                     less(tn->bcl->fsm_msg_cnt, acked) ||
 242                     less_eq(acked, n_ptr->bclink.acked))
 243                         goto exit;
 244         }
 245
 246         /* Skip over packets that node has previously acknowledged */
 247         skb_queue_walk(&tn->bcl->outqueue, skb) {
 248                 if (more(buf_seqno(skb), n_ptr->bclink.acked))
 249                         break;
 250         }
 251
 252         /* Update packets that node is now acknowledging */
 253         skb_queue_walk_from_safe(&tn->bcl->outqueue, skb, tmp) {
 254                 if (more(buf_seqno(skb), acked))
 255                         break;
 256
 257                 next = tipc_skb_queue_next(&tn->bcl->outqueue, skb);
 258                 if (skb != tn->bcl->next_out) {
 259                         bcbuf_decr_acks(skb);
 260                 } else {
 261                         bcbuf_set_acks(skb, 0);
 262                         tn->bcl->next_out = next;
 263                         bclink_set_last_sent(net);
 264                 }
 265
 266                 if (bcbuf_acks(skb) == 0) {
 267                         __skb_unlink(skb, &tn->bcl->outqueue);
 268                         kfree_skb(skb);
 269                         released = 1;
 270                 }
 271         }
 272         n_ptr->bclink.acked = acked;
 273
 274         /* Try resolving broadcast link congestion, if necessary */
 275         if (unlikely(tn->bcl->next_out)) {
 276                 tipc_link_push_packets(tn->bcl);
 277                 bclink_set_last_sent(net);
 278         }
 279         if (unlikely(released && !skb_queue_empty(&tn->bcl->wakeupq)))
 280                 n_ptr->action_flags |= TIPC_WAKEUP_BCAST_USERS;
 281 exit:
 282         tipc_bclink_unlock(net);
 283 }
 284
 285 /**
 286  * tipc_bclink_update_link_state - update broadcast link state
 287  *
 288  * RCU and node lock set
 289  */
 290 void tipc_bclink_update_link_state(struct tipc_node *n_ptr,
 291                                    u32 last_sent)
 292 {
 293         struct sk_buff *buf;
 294         struct net *net = n_ptr->net;
 295         struct tipc_net *tn = net_generic(net, tipc_net_id);
 296
 297         /* Ignore "stale" link state info */
 298         if (less_eq(last_sent, n_ptr->bclink.last_in))
 299                 return;
 300
 301         /* Update link synchronization state; quit if in sync */
 302         bclink_update_last_sent(n_ptr, last_sent);
 303
 304         if (n_ptr->bclink.last_sent == n_ptr->bclink.last_in)
 305                 return;
 306
 307         /* Update out-of-sync state; quit if loss is still unconfirmed */
 308         if ((++n_ptr->bclink.oos_state) == 1) {
 309                 if (n_ptr->bclink.deferred_size < (TIPC_MIN_LINK_WIN / 2))
 310                         return;
 311                 n_ptr->bclink.oos_state++;
 312         }
 313
 314         /* Don't NACK if one has been recently sent (or seen) */
 315         if (n_ptr->bclink.oos_state & 0x1)
 316                 return;
 317
 318         /* Send NACK */
 319         buf = tipc_buf_acquire(INT_H_SIZE);
 320         if (buf) {
 321                 struct tipc_msg *msg = buf_msg(buf);
 322                 struct sk_buff *skb = skb_peek(&n_ptr->bclink.deferred_queue);
 323                 u32 to = skb ? buf_seqno(skb) - 1 : n_ptr->bclink.last_sent;
 324
 325                 tipc_msg_init(tn->own_addr, msg, BCAST_PROTOCOL, STATE_MSG,
 326                               INT_H_SIZE, n_ptr->addr);
 327                 msg_set_non_seq(msg, 1);
 328                 msg_set_mc_netid(msg, tn->net_id);
 329                 msg_set_bcast_ack(msg, n_ptr->bclink.last_in);
 330                 msg_set_bcgap_after(msg, n_ptr->bclink.last_in);
 331                 msg_set_bcgap_to(msg, to);
 332
 333                 tipc_bclink_lock(net);
 334                 tipc_bearer_send(net, MAX_BEARERS, buf, NULL);
 335                 tn->bcl->stats.sent_nacks++;
 336                 tipc_bclink_unlock(net);
 337                 kfree_skb(buf);
 338
 339                 n_ptr->bclink.oos_state++;
 340         }
 341 }
 342
 343 /**
 344  * bclink_peek_nack - monitor retransmission requests sent by other nodes
 345  *
 346  * Delay any upcoming NACK by this node if another node has already
 347  * requested the first message this node is going to ask for.
 348  */
 349 static void bclink_peek_nack(struct net *net, struct tipc_msg *msg)
 350 {
 351         struct tipc_node *n_ptr = tipc_node_find(net, msg_destnode(msg));
 352
 353         if (unlikely(!n_ptr))
 354                 return;
 355
 356         tipc_node_lock(n_ptr);
 357
 358         if (n_ptr->bclink.recv_permitted &&
 359             (n_ptr->bclink.last_in != n_ptr->bclink.last_sent) &&
 360             (n_ptr->bclink.last_in == msg_bcgap_after(msg)))
 361                 n_ptr->bclink.oos_state = 2;
 362
 363         tipc_node_unlock(n_ptr);
 364 }
 365
 366 /* tipc_bclink_xmit - deliver buffer chain to all nodes in cluster
 367  *                    and to identified node local sockets
 368  * @net: the applicable net namespace
 369  * @list: chain of buffers containing message
 370  * Consumes the buffer chain, except when returning -ELINKCONG
 371  * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE
 372  */
 373 int tipc_bclink_xmit(struct net *net, struct sk_buff_head *list)
 374 {
 375         struct tipc_net *tn = net_generic(net, tipc_net_id);
 376         struct tipc_link *bcl = tn->bcl;
 377         struct tipc_bclink *bclink = tn->bclink;
 378         int rc = 0;
 379         int bc = 0;
 380         struct sk_buff *skb;
 381         struct sk_buff_head arrvq;
 382         struct sk_buff_head inputq;
 383
 384         /* Prepare clone of message for local node */
 385         skb = tipc_msg_reassemble(list);
 386         if (unlikely(!skb)) {
 387                 __skb_queue_purge(list);
 388                 return -EHOSTUNREACH;
 389         }
 390
 391         /* Broadcast to all nodes */
 392         if (likely(bclink)) {
 393                 tipc_bclink_lock(net);
 394                 if (likely(bclink->bcast_nodes.count)) {
 395                         rc = __tipc_link_xmit(net, bcl, list);
 396                         if (likely(!rc)) {
 397                                 u32 len = skb_queue_len(&bcl->outqueue);
 398
 399                                 bclink_set_last_sent(net);
 400                                 bcl->stats.queue_sz_counts++;
 401                                 bcl->stats.accu_queue_sz += len;
 402                         }
 403                         bc = 1;
 404                 }
 405                 tipc_bclink_unlock(net);
 406         }
 407
 408         if (unlikely(!bc))
 409                 __skb_queue_purge(list);
 410
 411         if (unlikely(rc)) {
 412                 kfree_skb(skb);
 413                 return rc;
 414         }
 415         /* Deliver message clone */
 416         __skb_queue_head_init(&arrvq);
 417         skb_queue_head_init(&inputq);
 418         __skb_queue_tail(&arrvq, skb);
 419         tipc_sk_mcast_rcv(net, &arrvq, &inputq);
 420         return rc;
 421 }
 422
 423 /**
 424  * bclink_accept_pkt - accept an incoming, in-sequence broadcast packet
 425  *
 426  * Called with both sending node's lock and bclink_lock taken.
 427  */
 428 static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
 429 {
 430         struct tipc_net *tn = net_generic(node->net, tipc_net_id);
 431
 432         bclink_update_last_sent(node, seqno);
 433         node->bclink.last_in = seqno;
 434         node->bclink.oos_state = 0;
 435         tn->bcl->stats.recv_info++;
 436
 437         /*
 438          * Unicast an ACK periodically, ensuring that
 439          * all nodes in the cluster don't ACK at the same time
 440          */
 441         if (((seqno - tn->own_addr) % TIPC_MIN_LINK_WIN) == 0) {
 442                 tipc_link_proto_xmit(node->active_links[node->addr & 1],
 443                                      STATE_MSG, 0, 0, 0, 0, 0);
 444                 tn->bcl->stats.sent_acks++;
 445         }
 446 }
 447
 448 /**
 449  * tipc_bclink_rcv - receive a broadcast packet, and deliver upwards
 450  *
 451  * RCU is locked, no other locks set
 452  */
 453 void tipc_bclink_rcv(struct net *net, struct sk_buff *buf)
 454 {
 455         struct tipc_net *tn = net_generic(net, tipc_net_id);
 456         struct tipc_link *bcl = tn->bcl;
 457         struct tipc_msg *msg = buf_msg(buf);
 458         struct tipc_node *node;
 459         u32 next_in;
 460         u32 seqno;
 461         int deferred = 0;
 462         int pos = 0;
 463         struct sk_buff *iskb;
 464         struct sk_buff_head *arrvq, *inputq;
 465
 466         /* Screen out unwanted broadcast messages */
 467         if (msg_mc_netid(msg) != tn->net_id)
 468                 goto exit;
 469
 470         node = tipc_node_find(net, msg_prevnode(msg));
 471         if (unlikely(!node))
 472                 goto exit;
 473
 474         tipc_node_lock(node);
 475         if (unlikely(!node->bclink.recv_permitted))
 476                 goto unlock;
 477
 478         /* Handle broadcast protocol message */
 479         if (unlikely(msg_user(msg) == BCAST_PROTOCOL)) {
 480                 if (msg_type(msg) != STATE_MSG)
 481                         goto unlock;
 482                 if (msg_destnode(msg) == tn->own_addr) {
 483                         tipc_bclink_acknowledge(node, msg_bcast_ack(msg));
 484                         tipc_node_unlock(node);
 485                         tipc_bclink_lock(net);
 486                         bcl->stats.recv_nacks++;
 487                         tn->bclink->retransmit_to = node;
 488                         bclink_retransmit_pkt(tn, msg_bcgap_after(msg),
 489                                               msg_bcgap_to(msg));
 490                         tipc_bclink_unlock(net);
 491                 } else {
 492                         tipc_node_unlock(node);
 493                         bclink_peek_nack(net, msg);
 494                 }
 495                 goto exit;
 496         }
 497
 498         /* Handle in-sequence broadcast message */
 499         seqno = msg_seqno(msg);
 500         next_in = mod(node->bclink.last_in + 1);
 501         arrvq = &tn->bclink->arrvq;
 502         inputq = &tn->bclink->inputq;
 503
 504         if (likely(seqno == next_in)) {
 505 receive:
 506                 /* Deliver message to destination */
 507                 if (likely(msg_isdata(msg))) {
 508                         tipc_bclink_lock(net);
 509                         bclink_accept_pkt(node, seqno);
 510                         spin_lock_bh(&inputq->lock);
 511                         __skb_queue_tail(arrvq, buf);
 512                         spin_unlock_bh(&inputq->lock);
 513                         node->action_flags |= TIPC_BCAST_MSG_EVT;
 514                         tipc_bclink_unlock(net);
 515                         tipc_node_unlock(node);
 516                 } else if (msg_user(msg) == MSG_BUNDLER) {
 517                         tipc_bclink_lock(net);
 518                         bclink_accept_pkt(node, seqno);
 519                         bcl->stats.recv_bundles++;
 520                         bcl->stats.recv_bundled += msg_msgcnt(msg);
 521                         pos = 0;
 522                         while (tipc_msg_extract(buf, &iskb, &pos)) {
 523                                 spin_lock_bh(&inputq->lock);
 524                                 __skb_queue_tail(arrvq, iskb);
 525                                 spin_unlock_bh(&inputq->lock);
 526                         }
 527                         node->action_flags |= TIPC_BCAST_MSG_EVT;
 528                         tipc_bclink_unlock(net);
 529                         tipc_node_unlock(node);
 530                 } else if (msg_user(msg) == MSG_FRAGMENTER) {
 531                         tipc_buf_append(&node->bclink.reasm_buf, &buf);
 532                         if (unlikely(!buf && !node->bclink.reasm_buf))
 533                                 goto unlock;
 534                         tipc_bclink_lock(net);
 535                         bclink_accept_pkt(node, seqno);
 536                         bcl->stats.recv_fragments++;
 537                         if (buf) {
 538                                 bcl->stats.recv_fragmented++;
 539                                 msg = buf_msg(buf);
 540                                 tipc_bclink_unlock(net);
 541                                 goto receive;
 542                         }
 543                         tipc_bclink_unlock(net);
 544                         tipc_node_unlock(node);
 545                 } else {
 546                         tipc_bclink_lock(net);
 547                         bclink_accept_pkt(node, seqno);
 548                         tipc_bclink_unlock(net);
 549                         tipc_node_unlock(node);
 550                         kfree_skb(buf);
 551                 }
 552                 buf = NULL;
 553
 554                 /* Determine new synchronization state */
 555                 tipc_node_lock(node);
 556                 if (unlikely(!tipc_node_is_up(node)))
 557                         goto unlock;
 558
 559                 if (node->bclink.last_in == node->bclink.last_sent)
 560                         goto unlock;
 561
 562                 if (skb_queue_empty(&node->bclink.deferred_queue)) {
 563                         node->bclink.oos_state = 1;
 564                         goto unlock;
 565                 }
 566
 567                 msg = buf_msg(skb_peek(&node->bclink.deferred_queue));
 568                 seqno = msg_seqno(msg);
 569                 next_in = mod(next_in + 1);
 570                 if (seqno != next_in)
 571                         goto unlock;
 572
 573                 /* Take in-sequence message from deferred queue & deliver it */
 574                 buf = __skb_dequeue(&node->bclink.deferred_queue);
 575                 goto receive;
 576         }
 577
 578         /* Handle out-of-sequence broadcast message */
 579         if (less(next_in, seqno)) {
 580                 deferred = tipc_link_defer_pkt(&node->bclink.deferred_queue,
 581                                                buf);
 582                 bclink_update_last_sent(node, seqno);
 583                 buf = NULL;
 584         }
 585
 586         tipc_bclink_lock(net);
 587
 588         if (deferred)
 589                 bcl->stats.deferred_recv++;
 590         else
 591                 bcl->stats.duplicates++;
 592
 593         tipc_bclink_unlock(net);
 594
 595 unlock:
 596         tipc_node_unlock(node);
 597 exit:
 598         kfree_skb(buf);
 599 }
 600
 601 u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr)
 602 {
 603         return (n_ptr->bclink.recv_permitted &&
 604                 (tipc_bclink_get_last_sent(n_ptr->net) != n_ptr->bclink.acked));
 605 }
 606
 607
 608 /**
 609  * tipc_bcbearer_send - send a packet through the broadcast pseudo-bearer
 610  *
 611  * Send packet over as many bearers as necessary to reach all nodes
 612  * that have joined the broadcast link.
 613  *
 614  * Returns 0 (packet sent successfully) under all circumstances,
 615  * since the broadcast link's pseudo-bearer never blocks
 616  */
 617 static int tipc_bcbearer_send(struct net *net, struct sk_buff *buf,
 618                               struct tipc_bearer *unused1,
 619                               struct tipc_media_addr *unused2)
 620 {
 621         int bp_index;
 622         struct tipc_msg *msg = buf_msg(buf);
 623         struct tipc_net *tn = net_generic(net, tipc_net_id);
 624         struct tipc_bcbearer *bcbearer = tn->bcbearer;
 625         struct tipc_bclink *bclink = tn->bclink;
 626
 627         /* Prepare broadcast link message for reliable transmission,
 628          * if first time trying to send it;
 629          * preparation is skipped for broadcast link protocol messages
 630          * since they are sent in an unreliable manner and don't need it
 631          */
 632         if (likely(!msg_non_seq(buf_msg(buf)))) {
 633                 bcbuf_set_acks(buf, bclink->bcast_nodes.count);
 634                 msg_set_non_seq(msg, 1);
 635                 msg_set_mc_netid(msg, tn->net_id);
 636                 tn->bcl->stats.sent_info++;
 637
 638                 if (WARN_ON(!bclink->bcast_nodes.count)) {
 639                         dump_stack();
 640                         return 0;
 641                 }
 642         }
 643
 644         /* Send buffer over bearers until all targets reached */
 645         bcbearer->remains = bclink->bcast_nodes;
 646
 647         for (bp_index = 0; bp_index < MAX_BEARERS; bp_index++) {
 648                 struct tipc_bearer *p = bcbearer->bpairs[bp_index].primary;
 649                 struct tipc_bearer *s = bcbearer->bpairs[bp_index].secondary;
 650                 struct tipc_bearer *bp[2] = {p, s};
 651                 struct tipc_bearer *b = bp[msg_link_selector(msg)];
 652                 struct sk_buff *tbuf;
 653
 654                 if (!p)
 655                         break; /* No more bearers to try */
 656                 if (!b)
 657                         b = p;
 658                 tipc_nmap_diff(&bcbearer->remains, &b->nodes,
 659                                &bcbearer->remains_new);
 660                 if (bcbearer->remains_new.count == bcbearer->remains.count)
 661                         continue; /* Nothing added by bearer pair */
 662
 663                 if (bp_index == 0) {
 664                         /* Use original buffer for first bearer */
 665                         tipc_bearer_send(net, b->identity, buf, &b->bcast_addr);
 666                 } else {
 667                         /* Avoid concurrent buffer access */
 668                         tbuf = pskb_copy_for_clone(buf, GFP_ATOMIC);
 669                         if (!tbuf)
 670                                 break;
 671                         tipc_bearer_send(net, b->identity, tbuf,
 672                                          &b->bcast_addr);
 673                         kfree_skb(tbuf); /* Bearer keeps a clone */
 674                 }
 675                 if (bcbearer->remains_new.count == 0)
 676                         break; /* All targets reached */
 677
 678                 bcbearer->remains = bcbearer->remains_new;
 679         }
 680
 681         return 0;
 682 }
 683
 684 /**
 685  * tipc_bcbearer_sort - create sets of bearer pairs used by broadcast bearer
 686  */
 687 void tipc_bcbearer_sort(struct net *net, struct tipc_node_map *nm_ptr,
 688                         u32 node, bool action)
 689 {
 690         struct tipc_net *tn = net_generic(net, tipc_net_id);
 691         struct tipc_bcbearer *bcbearer = tn->bcbearer;
 692         struct tipc_bcbearer_pair *bp_temp = bcbearer->bpairs_temp;
 693         struct tipc_bcbearer_pair *bp_curr;
 694         struct tipc_bearer *b;
 695         int b_index;
 696         int pri;
 697
 698         tipc_bclink_lock(net);
 699
 700         if (action)
 701                 tipc_nmap_add(nm_ptr, node);
 702         else
 703                 tipc_nmap_remove(nm_ptr, node);
 704
 705         /* Group bearers by priority (can assume max of two per priority) */
 706         memset(bp_temp, 0, sizeof(bcbearer->bpairs_temp));
 707
 708         rcu_read_lock();
 709         for (b_index = 0; b_index < MAX_BEARERS; b_index++) {
 710                 b = rcu_dereference_rtnl(tn->bearer_list[b_index]);
 711                 if (!b || !b->nodes.count)
 712                         continue;
 713
 714                 if (!bp_temp[b->priority].primary)
 715                         bp_temp[b->priority].primary = b;
 716                 else
 717                         bp_temp[b->priority].secondary = b;
 718         }
 719         rcu_read_unlock();
 720
 721         /* Create array of bearer pairs for broadcasting */
 722         bp_curr = bcbearer->bpairs;
 723         memset(bcbearer->bpairs, 0, sizeof(bcbearer->bpairs));
 724
 725         for (pri = TIPC_MAX_LINK_PRI; pri >= 0; pri--) {
 726
 727                 if (!bp_temp[pri].primary)
 728                         continue;
 729
 730                 bp_curr->primary = bp_temp[pri].primary;
 731
 732                 if (bp_temp[pri].secondary) {
 733                         if (tipc_nmap_equal(&bp_temp[pri].primary->nodes,
 734                                             &bp_temp[pri].secondary->nodes)) {
 735                                 bp_curr->secondary = bp_temp[pri].secondary;
 736                         } else {
 737                                 bp_curr++;
 738                                 bp_curr->primary = bp_temp[pri].secondary;
 739                         }
 740                 }
 741
 742                 bp_curr++;
 743         }
 744
 745         tipc_bclink_unlock(net);
 746 }
 747
 748 static int __tipc_nl_add_bc_link_stat(struct sk_buff *skb,
 749                                       struct tipc_stats *stats)
 750 {
 751         int i;
 752         struct nlattr *nest;
 753
 754         struct nla_map {
 755                 __u32 key;
 756                 __u32 val;
 757         };
 758
 759         struct nla_map map[] = {
 760                 {TIPC_NLA_STATS_RX_INFO, stats->recv_info},
 761                 {TIPC_NLA_STATS_RX_FRAGMENTS, stats->recv_fragments},
 762                 {TIPC_NLA_STATS_RX_FRAGMENTED, stats->recv_fragmented},
 763                 {TIPC_NLA_STATS_RX_BUNDLES, stats->recv_bundles},
 764                 {TIPC_NLA_STATS_RX_BUNDLED, stats->recv_bundled},
 765                 {TIPC_NLA_STATS_TX_INFO, stats->sent_info},
 766                 {TIPC_NLA_STATS_TX_FRAGMENTS, stats->sent_fragments},
 767                 {TIPC_NLA_STATS_TX_FRAGMENTED, stats->sent_fragmented},
 768                 {TIPC_NLA_STATS_TX_BUNDLES, stats->sent_bundles},
 769                 {TIPC_NLA_STATS_TX_BUNDLED, stats->sent_bundled},
 770                 {TIPC_NLA_STATS_RX_NACKS, stats->recv_nacks},
 771                 {TIPC_NLA_STATS_RX_DEFERRED, stats->deferred_recv},
 772                 {TIPC_NLA_STATS_TX_NACKS, stats->sent_nacks},
 773                 {TIPC_NLA_STATS_TX_ACKS, stats->sent_acks},
 774                 {TIPC_NLA_STATS_RETRANSMITTED, stats->retransmitted},
 775                 {TIPC_NLA_STATS_DUPLICATES, stats->duplicates},
 776                 {TIPC_NLA_STATS_LINK_CONGS, stats->link_congs},
 777                 {TIPC_NLA_STATS_MAX_QUEUE, stats->max_queue_sz},
 778                 {TIPC_NLA_STATS_AVG_QUEUE, stats->queue_sz_counts ?
 779                         (stats->accu_queue_sz / stats->queue_sz_counts) : 0}
 780         };
 781
 782         nest = nla_nest_start(skb, TIPC_NLA_LINK_STATS);
 783         if (!nest)
 784                 return -EMSGSIZE;
 785
 786         for (i = 0; i <  ARRAY_SIZE(map); i++)
 787                 if (nla_put_u32(skb, map[i].key, map[i].val))
 788                         goto msg_full;
 789
 790         nla_nest_end(skb, nest);
 791
 792         return 0;
 793 msg_full:
 794         nla_nest_cancel(skb, nest);
 795
 796         return -EMSGSIZE;
 797 }
 798
 799 int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg)
 800 {
 801         int err;
 802         void *hdr;
 803         struct nlattr *attrs;
 804         struct nlattr *prop;
 805         struct tipc_net *tn = net_generic(net, tipc_net_id);
 806         struct tipc_link *bcl = tn->bcl;
 807
 808         if (!bcl)
 809                 return 0;
 810
 811         tipc_bclink_lock(net);
 812
 813         hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
 814                           NLM_F_MULTI, TIPC_NL_LINK_GET);
 815         if (!hdr)
 816                 return -EMSGSIZE;
 817
 818         attrs = nla_nest_start(msg->skb, TIPC_NLA_LINK);
 819         if (!attrs)
 820                 goto msg_full;
 821
 822         /* The broadcast link is always up */
 823         if (nla_put_flag(msg->skb, TIPC_NLA_LINK_UP))
 824                 goto attr_msg_full;
 825
 826         if (nla_put_flag(msg->skb, TIPC_NLA_LINK_BROADCAST))
 827                 goto attr_msg_full;
 828         if (nla_put_string(msg->skb, TIPC_NLA_LINK_NAME, bcl->name))
 829                 goto attr_msg_full;
 830         if (nla_put_u32(msg->skb, TIPC_NLA_LINK_RX, bcl->next_in_no))
 831                 goto attr_msg_full;
 832         if (nla_put_u32(msg->skb, TIPC_NLA_LINK_TX, bcl->next_out_no))
 833                 goto attr_msg_full;
 834
 835         prop = nla_nest_start(msg->skb, TIPC_NLA_LINK_PROP);
 836         if (!prop)
 837                 goto attr_msg_full;
 838         if (nla_put_u32(msg->skb, TIPC_NLA_PROP_WIN, bcl->queue_limit[0]))
 839                 goto prop_msg_full;
 840         nla_nest_end(msg->skb, prop);
 841
 842         err = __tipc_nl_add_bc_link_stat(msg->skb, &bcl->stats);
 843         if (err)
 844                 goto attr_msg_full;
 845
 846         tipc_bclink_unlock(net);
 847         nla_nest_end(msg->skb, attrs);
 848         genlmsg_end(msg->skb, hdr);
 849
 850         return 0;
 851
 852 prop_msg_full:
 853         nla_nest_cancel(msg->skb, prop);
 854 attr_msg_full:
 855         nla_nest_cancel(msg->skb, attrs);
 856 msg_full:
 857         tipc_bclink_unlock(net);
 858         genlmsg_cancel(msg->skb, hdr);
 859
 860         return -EMSGSIZE;
 861 }
 862
 863 int tipc_bclink_reset_stats(struct net *net)
 864 {
 865         struct tipc_net *tn = net_generic(net, tipc_net_id);
 866         struct tipc_link *bcl = tn->bcl;
 867
 868         if (!bcl)
 869                 return -ENOPROTOOPT;
 870
 871         tipc_bclink_lock(net);
 872         memset(&bcl->stats, 0, sizeof(bcl->stats));
 873         tipc_bclink_unlock(net);
 874         return 0;
 875 }
 876
 877 int tipc_bclink_set_queue_limits(struct net *net, u32 limit)
 878 {
 879         struct tipc_net *tn = net_generic(net, tipc_net_id);
 880         struct tipc_link *bcl = tn->bcl;
 881
 882         if (!bcl)
 883                 return -ENOPROTOOPT;
 884         if ((limit < TIPC_MIN_LINK_WIN) || (limit > TIPC_MAX_LINK_WIN))
 885                 return -EINVAL;
 886
 887         tipc_bclink_lock(net);
 888         tipc_link_set_queue_limits(bcl, limit);
 889         tipc_bclink_unlock(net);
 890         return 0;
 891 }
 892
 893 int tipc_bclink_init(struct net *net)
 894 {
 895         struct tipc_net *tn = net_generic(net, tipc_net_id);
 896         struct tipc_bcbearer *bcbearer;
 897         struct tipc_bclink *bclink;
 898         struct tipc_link *bcl;
 899
 900         bcbearer = kzalloc(sizeof(*bcbearer), GFP_ATOMIC);
 901         if (!bcbearer)
 902                 return -ENOMEM;
 903
 904         bclink = kzalloc(sizeof(*bclink), GFP_ATOMIC);
 905         if (!bclink) {
 906                 kfree(bcbearer);
 907                 return -ENOMEM;
 908         }
 909
 910         bcl = &bclink->link;
 911         bcbearer->bearer.media = &bcbearer->media;
 912         bcbearer->media.send_msg = tipc_bcbearer_send;
 913         sprintf(bcbearer->media.name, "tipc-broadcast");
 914
 915         spin_lock_init(&bclink->lock);
 916         __skb_queue_head_init(&bcl->outqueue);
 917         __skb_queue_head_init(&bcl->deferred_queue);
 918         skb_queue_head_init(&bcl->wakeupq);
 919         bcl->next_out_no = 1;
 920         spin_lock_init(&bclink->node.lock);
 921         __skb_queue_head_init(&bclink->arrvq);
 922         skb_queue_head_init(&bclink->inputq);
 923         bcl->owner = &bclink->node;
 924         bcl->owner->net = net;
 925         bcl->max_pkt = MAX_PKT_DEFAULT_MCAST;
 926         tipc_link_set_queue_limits(bcl, BCLINK_WIN_DEFAULT);
 927         bcl->bearer_id = MAX_BEARERS;
 928         rcu_assign_pointer(tn->bearer_list[MAX_BEARERS], &bcbearer->bearer);
 929         bcl->state = WORKING_WORKING;
 930         bcl->pmsg = (struct tipc_msg *)&bcl->proto_msg;
 931         msg_set_prevnode(bcl->pmsg, tn->own_addr);
 932         strlcpy(bcl->name, tipc_bclink_name, TIPC_MAX_LINK_NAME);
 933         tn->bcbearer = bcbearer;
 934         tn->bclink = bclink;
 935         tn->bcl = bcl;
 936         return 0;
 937 }
 938
 939 void tipc_bclink_stop(struct net *net)
 940 {
 941         struct tipc_net *tn = net_generic(net, tipc_net_id);
 942
 943         tipc_bclink_lock(net);
 944         tipc_link_purge_queues(tn->bcl);
 945         tipc_bclink_unlock(net);
 946
 947         RCU_INIT_POINTER(tn->bearer_list[BCBEARER], NULL);
 948         synchronize_net();
 949         kfree(tn->bcbearer);
 950         kfree(tn->bclink);
 951 }
 952
 953 /**
 954  * tipc_nmap_add - add a node to a node map
 955  */
 956 static void tipc_nmap_add(struct tipc_node_map *nm_ptr, u32 node)
 957 {
 958         int n = tipc_node(node);
 959         int w = n / WSIZE;
 960         u32 mask = (1 << (n % WSIZE));
 961
 962         if ((nm_ptr->map[w] & mask) == 0) {
 963                 nm_ptr->count++;
 964                 nm_ptr->map[w] |= mask;
 965         }
 966 }
 967
 968 /**
 969  * tipc_nmap_remove - remove a node from a node map
 970  */
 971 static void tipc_nmap_remove(struct tipc_node_map *nm_ptr, u32 node)
 972 {
 973         int n = tipc_node(node);
 974         int w = n / WSIZE;
 975         u32 mask = (1 << (n % WSIZE));
 976
 977         if ((nm_ptr->map[w] & mask) != 0) {
 978                 nm_ptr->map[w] &= ~mask;
 979                 nm_ptr->count--;
 980         }
 981 }
 982
 983 /**
 984  * tipc_nmap_diff - find differences between node maps
 985  * @nm_a: input node map A
 986  * @nm_b: input node map B
 987  * @nm_diff: output node map A-B (i.e. nodes of A that are not in B)
 988  */
 989 static void tipc_nmap_diff(struct tipc_node_map *nm_a,
 990                            struct tipc_node_map *nm_b,
 991                            struct tipc_node_map *nm_diff)
 992 {
 993         int stop = ARRAY_SIZE(nm_a->map);
 994         int w;
 995         int b;
 996         u32 map;
 997
 998         memset(nm_diff, 0, sizeof(*nm_diff));
 999         for (w = 0; w < stop; w++) {
1000                 map = nm_a->map[w] ^ (nm_a->map[w] & nm_b->map[w]);
1001                 nm_diff->map[w] = map;
1002                 if (map != 0) {
1003                         for (b = 0 ; b < WSIZE; b++) {
1004                                 if (map & (1 << b))
1005                                         nm_diff->count++;
1006                         }
1007                 }
1008         }
1009 }