linux/net/tipc/name_distr.c
<<
>>
Prefs
   1/*
   2 * net/tipc/name_distr.c: TIPC name distribution code
   3 *
   4 * Copyright (c) 2000-2006, 2014-2019, Ericsson AB
   5 * Copyright (c) 2005, 2010-2011, Wind River Systems
   6 * Copyright (c) 2020-2021, Red Hat Inc
   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 "name_distr.h"
  41
  42int sysctl_tipc_named_timeout __read_mostly = 2000;
  43
  44struct distr_queue_item {
  45        struct distr_item i;
  46        u32 dtype;
  47        u32 node;
  48        unsigned long expires;
  49        struct list_head next;
  50};
  51
  52/**
  53 * publ_to_item - add publication info to a publication message
  54 * @p: publication info
  55 * @i: location of item in the message
  56 */
  57static void publ_to_item(struct distr_item *i, struct publication *p)
  58{
  59        i->type = htonl(p->sr.type);
  60        i->lower = htonl(p->sr.lower);
  61        i->upper = htonl(p->sr.upper);
  62        i->port = htonl(p->sk.ref);
  63        i->key = htonl(p->key);
  64}
  65
  66/**
  67 * named_prepare_buf - allocate & initialize a publication message
  68 * @net: the associated network namespace
  69 * @type: message type
  70 * @size: payload size
  71 * @dest: destination node
  72 *
  73 * The buffer returned is of size INT_H_SIZE + payload size
  74 */
  75static struct sk_buff *named_prepare_buf(struct net *net, u32 type, u32 size,
  76                                         u32 dest)
  77{
  78        struct sk_buff *buf = tipc_buf_acquire(INT_H_SIZE + size, GFP_ATOMIC);
  79        u32 self = tipc_own_addr(net);
  80        struct tipc_msg *msg;
  81
  82        if (buf != NULL) {
  83                msg = buf_msg(buf);
  84                tipc_msg_init(self, msg, NAME_DISTRIBUTOR,
  85                              type, INT_H_SIZE, dest);
  86                msg_set_size(msg, INT_H_SIZE + size);
  87        }
  88        return buf;
  89}
  90
  91/**
  92 * tipc_named_publish - tell other nodes about a new publication by this node
  93 * @net: the associated network namespace
  94 * @p: the new publication
  95 */
  96struct sk_buff *tipc_named_publish(struct net *net, struct publication *p)
  97{
  98        struct name_table *nt = tipc_name_table(net);
  99        struct distr_item *item;
 100        struct sk_buff *skb;
 101
 102        if (p->scope == TIPC_NODE_SCOPE) {
 103                list_add_tail_rcu(&p->binding_node, &nt->node_scope);
 104                return NULL;
 105        }
 106        write_lock_bh(&nt->cluster_scope_lock);
 107        list_add_tail(&p->binding_node, &nt->cluster_scope);
 108        write_unlock_bh(&nt->cluster_scope_lock);
 109        skb = named_prepare_buf(net, PUBLICATION, ITEM_SIZE, 0);
 110        if (!skb) {
 111                pr_warn("Publication distribution failure\n");
 112                return NULL;
 113        }
 114        msg_set_named_seqno(buf_msg(skb), nt->snd_nxt++);
 115        msg_set_non_legacy(buf_msg(skb));
 116        item = (struct distr_item *)msg_data(buf_msg(skb));
 117        publ_to_item(item, p);
 118        return skb;
 119}
 120
 121/**
 122 * tipc_named_withdraw - tell other nodes about a withdrawn publication by this node
 123 * @net: the associated network namespace
 124 * @p: the withdrawn publication
 125 */
 126struct sk_buff *tipc_named_withdraw(struct net *net, struct publication *p)
 127{
 128        struct name_table *nt = tipc_name_table(net);
 129        struct distr_item *item;
 130        struct sk_buff *skb;
 131
 132        write_lock_bh(&nt->cluster_scope_lock);
 133        list_del(&p->binding_node);
 134        write_unlock_bh(&nt->cluster_scope_lock);
 135        if (p->scope == TIPC_NODE_SCOPE)
 136                return NULL;
 137
 138        skb = named_prepare_buf(net, WITHDRAWAL, ITEM_SIZE, 0);
 139        if (!skb) {
 140                pr_warn("Withdrawal distribution failure\n");
 141                return NULL;
 142        }
 143        msg_set_named_seqno(buf_msg(skb), nt->snd_nxt++);
 144        msg_set_non_legacy(buf_msg(skb));
 145        item = (struct distr_item *)msg_data(buf_msg(skb));
 146        publ_to_item(item, p);
 147        return skb;
 148}
 149
 150/**
 151 * named_distribute - prepare name info for bulk distribution to another node
 152 * @net: the associated network namespace
 153 * @list: list of messages (buffers) to be returned from this function
 154 * @dnode: node to be updated
 155 * @pls: linked list of publication items to be packed into buffer chain
 156 * @seqno: sequence number for this message
 157 */
 158static void named_distribute(struct net *net, struct sk_buff_head *list,
 159                             u32 dnode, struct list_head *pls, u16 seqno)
 160{
 161        struct publication *publ;
 162        struct sk_buff *skb = NULL;
 163        struct distr_item *item = NULL;
 164        u32 msg_dsz = ((tipc_node_get_mtu(net, dnode, 0, false) - INT_H_SIZE) /
 165                        ITEM_SIZE) * ITEM_SIZE;
 166        u32 msg_rem = msg_dsz;
 167        struct tipc_msg *hdr;
 168
 169        list_for_each_entry(publ, pls, binding_node) {
 170                /* Prepare next buffer: */
 171                if (!skb) {
 172                        skb = named_prepare_buf(net, PUBLICATION, msg_rem,
 173                                                dnode);
 174                        if (!skb) {
 175                                pr_warn("Bulk publication failure\n");
 176                                return;
 177                        }
 178                        hdr = buf_msg(skb);
 179                        msg_set_bc_ack_invalid(hdr, true);
 180                        msg_set_bulk(hdr);
 181                        msg_set_non_legacy(hdr);
 182                        item = (struct distr_item *)msg_data(hdr);
 183                }
 184
 185                /* Pack publication into message: */
 186                publ_to_item(item, publ);
 187                item++;
 188                msg_rem -= ITEM_SIZE;
 189
 190                /* Append full buffer to list: */
 191                if (!msg_rem) {
 192                        __skb_queue_tail(list, skb);
 193                        skb = NULL;
 194                        msg_rem = msg_dsz;
 195                }
 196        }
 197        if (skb) {
 198                hdr = buf_msg(skb);
 199                msg_set_size(hdr, INT_H_SIZE + (msg_dsz - msg_rem));
 200                skb_trim(skb, INT_H_SIZE + (msg_dsz - msg_rem));
 201                __skb_queue_tail(list, skb);
 202        }
 203        hdr = buf_msg(skb_peek_tail(list));
 204        msg_set_last_bulk(hdr);
 205        msg_set_named_seqno(hdr, seqno);
 206}
 207
 208/**
 209 * tipc_named_node_up - tell specified node about all publications by this node
 210 * @net: the associated network namespace
 211 * @dnode: destination node
 212 * @capabilities: peer node's capabilities
 213 */
 214void tipc_named_node_up(struct net *net, u32 dnode, u16 capabilities)
 215{
 216        struct name_table *nt = tipc_name_table(net);
 217        struct tipc_net *tn = tipc_net(net);
 218        struct sk_buff_head head;
 219        u16 seqno;
 220
 221        __skb_queue_head_init(&head);
 222        spin_lock_bh(&tn->nametbl_lock);
 223        if (!(capabilities & TIPC_NAMED_BCAST))
 224                nt->rc_dests++;
 225        seqno = nt->snd_nxt;
 226        spin_unlock_bh(&tn->nametbl_lock);
 227
 228        read_lock_bh(&nt->cluster_scope_lock);
 229        named_distribute(net, &head, dnode, &nt->cluster_scope, seqno);
 230        tipc_node_xmit(net, &head, dnode, 0);
 231        read_unlock_bh(&nt->cluster_scope_lock);
 232}
 233
 234/**
 235 * tipc_publ_purge - remove publication associated with a failed node
 236 * @net: the associated network namespace
 237 * @p: the publication to remove
 238 * @addr: failed node's address
 239 *
 240 * Invoked for each publication issued by a newly failed node.
 241 * Removes publication structure from name table & deletes it.
 242 */
 243static void tipc_publ_purge(struct net *net, struct publication *p, u32 addr)
 244{
 245        struct tipc_net *tn = tipc_net(net);
 246        struct publication *_p;
 247        struct tipc_uaddr ua;
 248
 249        tipc_uaddr(&ua, TIPC_SERVICE_RANGE, p->scope, p->sr.type,
 250                   p->sr.lower, p->sr.upper);
 251        spin_lock_bh(&tn->nametbl_lock);
 252        _p = tipc_nametbl_remove_publ(net, &ua, &p->sk, p->key);
 253        if (_p)
 254                tipc_node_unsubscribe(net, &_p->binding_node, addr);
 255        spin_unlock_bh(&tn->nametbl_lock);
 256        if (_p)
 257                kfree_rcu(_p, rcu);
 258}
 259
 260void tipc_publ_notify(struct net *net, struct list_head *nsub_list,
 261                      u32 addr, u16 capabilities)
 262{
 263        struct name_table *nt = tipc_name_table(net);
 264        struct tipc_net *tn = tipc_net(net);
 265
 266        struct publication *publ, *tmp;
 267
 268        list_for_each_entry_safe(publ, tmp, nsub_list, binding_node)
 269                tipc_publ_purge(net, publ, addr);
 270        spin_lock_bh(&tn->nametbl_lock);
 271        if (!(capabilities & TIPC_NAMED_BCAST))
 272                nt->rc_dests--;
 273        spin_unlock_bh(&tn->nametbl_lock);
 274}
 275
 276/**
 277 * tipc_update_nametbl - try to process a nametable update and notify
 278 *                       subscribers
 279 * @net: the associated network namespace
 280 * @i: location of item in the message
 281 * @node: node address
 282 * @dtype: name distributor message type
 283 *
 284 * tipc_nametbl_lock must be held.
 285 * Return: the publication item if successful, otherwise NULL.
 286 */
 287static bool tipc_update_nametbl(struct net *net, struct distr_item *i,
 288                                u32 node, u32 dtype)
 289{
 290        struct publication *p = NULL;
 291        struct tipc_socket_addr sk;
 292        struct tipc_uaddr ua;
 293        u32 key = ntohl(i->key);
 294
 295        tipc_uaddr(&ua, TIPC_SERVICE_RANGE, TIPC_CLUSTER_SCOPE,
 296                   ntohl(i->type), ntohl(i->lower), ntohl(i->upper));
 297        sk.ref = ntohl(i->port);
 298        sk.node = node;
 299
 300        if (dtype == PUBLICATION) {
 301                p = tipc_nametbl_insert_publ(net, &ua, &sk, key);
 302                if (p) {
 303                        tipc_node_subscribe(net, &p->binding_node, node);
 304                        return true;
 305                }
 306        } else if (dtype == WITHDRAWAL) {
 307                p = tipc_nametbl_remove_publ(net, &ua, &sk, key);
 308                if (p) {
 309                        tipc_node_unsubscribe(net, &p->binding_node, node);
 310                        kfree_rcu(p, rcu);
 311                        return true;
 312                }
 313                pr_warn_ratelimited("Failed to remove binding %u,%u from %u\n",
 314                                    ua.sr.type, ua.sr.lower, node);
 315        } else {
 316                pr_warn("Unrecognized name table message received\n");
 317        }
 318        return false;
 319}
 320
 321static struct sk_buff *tipc_named_dequeue(struct sk_buff_head *namedq,
 322                                          u16 *rcv_nxt, bool *open)
 323{
 324        struct sk_buff *skb, *tmp;
 325        struct tipc_msg *hdr;
 326        u16 seqno;
 327
 328        spin_lock_bh(&namedq->lock);
 329        skb_queue_walk_safe(namedq, skb, tmp) {
 330                if (unlikely(skb_linearize(skb))) {
 331                        __skb_unlink(skb, namedq);
 332                        kfree_skb(skb);
 333                        continue;
 334                }
 335                hdr = buf_msg(skb);
 336                seqno = msg_named_seqno(hdr);
 337                if (msg_is_last_bulk(hdr)) {
 338                        *rcv_nxt = seqno;
 339                        *open = true;
 340                }
 341
 342                if (msg_is_bulk(hdr) || msg_is_legacy(hdr)) {
 343                        __skb_unlink(skb, namedq);
 344                        spin_unlock_bh(&namedq->lock);
 345                        return skb;
 346                }
 347
 348                if (*open && (*rcv_nxt == seqno)) {
 349                        (*rcv_nxt)++;
 350                        __skb_unlink(skb, namedq);
 351                        spin_unlock_bh(&namedq->lock);
 352                        return skb;
 353                }
 354
 355                if (less(seqno, *rcv_nxt)) {
 356                        __skb_unlink(skb, namedq);
 357                        kfree_skb(skb);
 358                        continue;
 359                }
 360        }
 361        spin_unlock_bh(&namedq->lock);
 362        return NULL;
 363}
 364
 365/**
 366 * tipc_named_rcv - process name table update messages sent by another node
 367 * @net: the associated network namespace
 368 * @namedq: queue to receive from
 369 * @rcv_nxt: store last received seqno here
 370 * @open: last bulk msg was received (FIXME)
 371 */
 372void tipc_named_rcv(struct net *net, struct sk_buff_head *namedq,
 373                    u16 *rcv_nxt, bool *open)
 374{
 375        struct tipc_net *tn = tipc_net(net);
 376        struct distr_item *item;
 377        struct tipc_msg *hdr;
 378        struct sk_buff *skb;
 379        u32 count, node;
 380
 381        spin_lock_bh(&tn->nametbl_lock);
 382        while ((skb = tipc_named_dequeue(namedq, rcv_nxt, open))) {
 383                hdr = buf_msg(skb);
 384                node = msg_orignode(hdr);
 385                item = (struct distr_item *)msg_data(hdr);
 386                count = msg_data_sz(hdr) / ITEM_SIZE;
 387                while (count--) {
 388                        tipc_update_nametbl(net, item, node, msg_type(hdr));
 389                        item++;
 390                }
 391                kfree_skb(skb);
 392        }
 393        spin_unlock_bh(&tn->nametbl_lock);
 394}
 395
 396/**
 397 * tipc_named_reinit - re-initialize local publications
 398 * @net: the associated network namespace
 399 *
 400 * This routine is called whenever TIPC networking is enabled.
 401 * All name table entries published by this node are updated to reflect
 402 * the node's new network address.
 403 */
 404void tipc_named_reinit(struct net *net)
 405{
 406        struct name_table *nt = tipc_name_table(net);
 407        struct tipc_net *tn = tipc_net(net);
 408        struct publication *p;
 409        u32 self = tipc_own_addr(net);
 410
 411        spin_lock_bh(&tn->nametbl_lock);
 412
 413        list_for_each_entry_rcu(p, &nt->node_scope, binding_node)
 414                p->sk.node = self;
 415        list_for_each_entry_rcu(p, &nt->cluster_scope, binding_node)
 416                p->sk.node = self;
 417        nt->rc_dests = 0;
 418        spin_unlock_bh(&tn->nametbl_lock);
 419}
 420