net/mptcp/pm_netlink.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 /* Multipath TCP
   3  *
   4  * Copyright (c) 2020, Red Hat, Inc.
   5  */
   6
   7 #define pr_fmt(fmt) "MPTCP: " fmt
   8
   9 #include <linux/inet.h>
  10 #include <linux/kernel.h>
  11 #include <net/tcp.h>
  12 #include <net/netns/generic.h>
  13 #include <net/mptcp.h>
  14 #include <net/genetlink.h>
  15 #include <uapi/linux/mptcp.h>
  16
  17 #include "protocol.h"
  18 #include "mib.h"
  19
  20 /* forward declaration */
  21 static struct genl_family mptcp_genl_family;
  22
  23 static int pm_nl_pernet_id;
  24
  25 struct mptcp_pm_addr_entry {
  26         struct list_head        list;
  27         struct mptcp_addr_info  addr;
  28         u8                      flags;
  29         int                     ifindex;
  30         struct socket           *lsk;
  31 };
  32
  33 struct mptcp_pm_add_entry {
  34         struct list_head        list;
  35         struct mptcp_addr_info  addr;
  36         struct timer_list       add_timer;
  37         struct mptcp_sock       *sock;
  38         u8                      retrans_times;
  39 };
  40
  41 #define MAX_ADDR_ID             255
  42 #define BITMAP_SZ DIV_ROUND_UP(MAX_ADDR_ID + 1, BITS_PER_LONG)
  43
  44 struct pm_nl_pernet {
  45         /* protects pernet updates */
  46         spinlock_t              lock;
  47         struct list_head        local_addr_list;
  48         unsigned int            addrs;
  49         unsigned int            stale_loss_cnt;
  50         unsigned int            add_addr_signal_max;
  51         unsigned int            add_addr_accept_max;
  52         unsigned int            local_addr_max;
  53         unsigned int            subflows_max;
  54         unsigned int            next_id;
  55         unsigned long           id_bitmap[BITMAP_SZ];
  56 };
  57
  58 #define MPTCP_PM_ADDR_MAX       8
  59 #define ADD_ADDR_RETRANS_MAX    3
  60
  61 static bool addresses_equal(const struct mptcp_addr_info *a,
  62                             struct mptcp_addr_info *b, bool use_port)
  63 {
  64         bool addr_equals = false;
  65
  66         if (a->family == b->family) {
  67                 if (a->family == AF_INET)
  68                         addr_equals = a->addr.s_addr == b->addr.s_addr;
  69 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
  70                 else
  71                         addr_equals = !ipv6_addr_cmp(&a->addr6, &b->addr6);
  72         } else if (a->family == AF_INET) {
  73                 if (ipv6_addr_v4mapped(&b->addr6))
  74                         addr_equals = a->addr.s_addr == b->addr6.s6_addr32[3];
  75         } else if (b->family == AF_INET) {
  76                 if (ipv6_addr_v4mapped(&a->addr6))
  77                         addr_equals = a->addr6.s6_addr32[3] == b->addr.s_addr;
  78 #endif
  79         }
  80
  81         if (!addr_equals)
  82                 return false;
  83         if (!use_port)
  84                 return true;
  85
  86         return a->port == b->port;
  87 }
  88
  89 static bool address_zero(const struct mptcp_addr_info *addr)
  90 {
  91         struct mptcp_addr_info zero;
  92
  93         memset(&zero, 0, sizeof(zero));
  94         zero.family = addr->family;
  95
  96         return addresses_equal(addr, &zero, true);
  97 }
  98
  99 static void local_address(const struct sock_common *skc,
 100                           struct mptcp_addr_info *addr)
 101 {
 102         addr->family = skc->skc_family;
 103         addr->port = htons(skc->skc_num);
 104         if (addr->family == AF_INET)
 105                 addr->addr.s_addr = skc->skc_rcv_saddr;
 106 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
 107         else if (addr->family == AF_INET6)
 108                 addr->addr6 = skc->skc_v6_rcv_saddr;
 109 #endif
 110 }
 111
 112 static void remote_address(const struct sock_common *skc,
 113                            struct mptcp_addr_info *addr)
 114 {
 115         addr->family = skc->skc_family;
 116         addr->port = skc->skc_dport;
 117         if (addr->family == AF_INET)
 118                 addr->addr.s_addr = skc->skc_daddr;
 119 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
 120         else if (addr->family == AF_INET6)
 121                 addr->addr6 = skc->skc_v6_daddr;
 122 #endif
 123 }
 124
 125 static bool lookup_subflow_by_saddr(const struct list_head *list,
 126                                     struct mptcp_addr_info *saddr)
 127 {
 128         struct mptcp_subflow_context *subflow;
 129         struct mptcp_addr_info cur;
 130         struct sock_common *skc;
 131
 132         list_for_each_entry(subflow, list, node) {
 133                 skc = (struct sock_common *)mptcp_subflow_tcp_sock(subflow);
 134
 135                 local_address(skc, &cur);
 136                 if (addresses_equal(&cur, saddr, saddr->port))
 137                         return true;
 138         }
 139
 140         return false;
 141 }
 142
 143 static bool lookup_subflow_by_daddr(const struct list_head *list,
 144                                     struct mptcp_addr_info *daddr)
 145 {
 146         struct mptcp_subflow_context *subflow;
 147         struct mptcp_addr_info cur;
 148         struct sock_common *skc;
 149
 150         list_for_each_entry(subflow, list, node) {
 151                 skc = (struct sock_common *)mptcp_subflow_tcp_sock(subflow);
 152
 153                 remote_address(skc, &cur);
 154                 if (addresses_equal(&cur, daddr, daddr->port))
 155                         return true;
 156         }
 157
 158         return false;
 159 }
 160
 161 static struct mptcp_pm_addr_entry *
 162 select_local_address(const struct pm_nl_pernet *pernet,
 163                      struct mptcp_sock *msk)
 164 {
 165         struct mptcp_pm_addr_entry *entry, *ret = NULL;
 166         struct sock *sk = (struct sock *)msk;
 167
 168         msk_owned_by_me(msk);
 169
 170         rcu_read_lock();
 171         __mptcp_flush_join_list(msk);
 172         list_for_each_entry_rcu(entry, &pernet->local_addr_list, list) {
 173                 if (!(entry->flags & MPTCP_PM_ADDR_FLAG_SUBFLOW))
 174                         continue;
 175
 176                 if (entry->addr.family != sk->sk_family) {
 177 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
 178                         if ((entry->addr.family == AF_INET &&
 179                              !ipv6_addr_v4mapped(&sk->sk_v6_daddr)) ||
 180                             (sk->sk_family == AF_INET &&
 181                              !ipv6_addr_v4mapped(&entry->addr.addr6)))
 182 #endif
 183                                 continue;
 184                 }
 185
 186                 /* avoid any address already in use by subflows and
 187                  * pending join
 188                  */
 189                 if (!lookup_subflow_by_saddr(&msk->conn_list, &entry->addr)) {
 190                         ret = entry;
 191                         break;
 192                 }
 193         }
 194         rcu_read_unlock();
 195         return ret;
 196 }
 197
 198 static struct mptcp_pm_addr_entry *
 199 select_signal_address(struct pm_nl_pernet *pernet, unsigned int pos)
 200 {
 201         struct mptcp_pm_addr_entry *entry, *ret = NULL;
 202         int i = 0;
 203
 204         rcu_read_lock();
 205         /* do not keep any additional per socket state, just signal
 206          * the address list in order.
 207          * Note: removal from the local address list during the msk life-cycle
 208          * can lead to additional addresses not being announced.
 209          */
 210         list_for_each_entry_rcu(entry, &pernet->local_addr_list, list) {
 211                 if (!(entry->flags & MPTCP_PM_ADDR_FLAG_SIGNAL))
 212                         continue;
 213                 if (i++ == pos) {
 214                         ret = entry;
 215                         break;
 216                 }
 217         }
 218         rcu_read_unlock();
 219         return ret;
 220 }
 221
 222 unsigned int mptcp_pm_get_add_addr_signal_max(struct mptcp_sock *msk)
 223 {
 224         struct pm_nl_pernet *pernet;
 225
 226         pernet = net_generic(sock_net((struct sock *)msk), pm_nl_pernet_id);
 227         return READ_ONCE(pernet->add_addr_signal_max);
 228 }
 229 EXPORT_SYMBOL_GPL(mptcp_pm_get_add_addr_signal_max);
 230
 231 unsigned int mptcp_pm_get_add_addr_accept_max(struct mptcp_sock *msk)
 232 {
 233         struct pm_nl_pernet *pernet;
 234
 235         pernet = net_generic(sock_net((struct sock *)msk), pm_nl_pernet_id);
 236         return READ_ONCE(pernet->add_addr_accept_max);
 237 }
 238 EXPORT_SYMBOL_GPL(mptcp_pm_get_add_addr_accept_max);
 239
 240 unsigned int mptcp_pm_get_subflows_max(struct mptcp_sock *msk)
 241 {
 242         struct pm_nl_pernet *pernet;
 243
 244         pernet = net_generic(sock_net((struct sock *)msk), pm_nl_pernet_id);
 245         return READ_ONCE(pernet->subflows_max);
 246 }
 247 EXPORT_SYMBOL_GPL(mptcp_pm_get_subflows_max);
 248
 249 unsigned int mptcp_pm_get_local_addr_max(struct mptcp_sock *msk)
 250 {
 251         struct pm_nl_pernet *pernet;
 252
 253         pernet = net_generic(sock_net((struct sock *)msk), pm_nl_pernet_id);
 254         return READ_ONCE(pernet->local_addr_max);
 255 }
 256 EXPORT_SYMBOL_GPL(mptcp_pm_get_local_addr_max);
 257
 258 static void check_work_pending(struct mptcp_sock *msk)
 259 {
 260         if (msk->pm.add_addr_signaled == mptcp_pm_get_add_addr_signal_max(msk) &&
 261             (msk->pm.local_addr_used == mptcp_pm_get_local_addr_max(msk) ||
 262              msk->pm.subflows == mptcp_pm_get_subflows_max(msk)))
 263                 WRITE_ONCE(msk->pm.work_pending, false);
 264 }
 265
 266 struct mptcp_pm_add_entry *
 267 mptcp_lookup_anno_list_by_saddr(struct mptcp_sock *msk,
 268                                 struct mptcp_addr_info *addr)
 269 {
 270         struct mptcp_pm_add_entry *entry;
 271
 272         lockdep_assert_held(&msk->pm.lock);
 273
 274         list_for_each_entry(entry, &msk->pm.anno_list, list) {
 275                 if (addresses_equal(&entry->addr, addr, true))
 276                         return entry;
 277         }
 278
 279         return NULL;
 280 }
 281
 282 bool mptcp_pm_sport_in_anno_list(struct mptcp_sock *msk, const struct sock *sk)
 283 {
 284         struct mptcp_pm_add_entry *entry;
 285         struct mptcp_addr_info saddr;
 286         bool ret = false;
 287
 288         local_address((struct sock_common *)sk, &saddr);
 289
 290         spin_lock_bh(&msk->pm.lock);
 291         list_for_each_entry(entry, &msk->pm.anno_list, list) {
 292                 if (addresses_equal(&entry->addr, &saddr, true)) {
 293                         ret = true;
 294                         goto out;
 295                 }
 296         }
 297
 298 out:
 299         spin_unlock_bh(&msk->pm.lock);
 300         return ret;
 301 }
 302
 303 static void mptcp_pm_add_timer(struct timer_list *timer)
 304 {
 305         struct mptcp_pm_add_entry *entry = from_timer(entry, timer, add_timer);
 306         struct mptcp_sock *msk = entry->sock;
 307         struct sock *sk = (struct sock *)msk;
 308
 309         pr_debug("msk=%p", msk);
 310
 311         if (!msk)
 312                 return;
 313
 314         if (inet_sk_state_load(sk) == TCP_CLOSE)
 315                 return;
 316
 317         if (!entry->addr.id)
 318                 return;
 319
 320         if (mptcp_pm_should_add_signal_addr(msk)) {
 321                 sk_reset_timer(sk, timer, jiffies + TCP_RTO_MAX / 8);
 322                 goto out;
 323         }
 324
 325         spin_lock_bh(&msk->pm.lock);
 326
 327         if (!mptcp_pm_should_add_signal_addr(msk)) {
 328                 pr_debug("retransmit ADD_ADDR id=%d", entry->addr.id);
 329                 mptcp_pm_announce_addr(msk, &entry->addr, false);
 330                 mptcp_pm_add_addr_send_ack(msk);
 331                 entry->retrans_times++;
 332         }
 333
 334         if (entry->retrans_times < ADD_ADDR_RETRANS_MAX)
 335                 sk_reset_timer(sk, timer,
 336                                jiffies + mptcp_get_add_addr_timeout(sock_net(sk)));
 337
 338         spin_unlock_bh(&msk->pm.lock);
 339
 340         if (entry->retrans_times == ADD_ADDR_RETRANS_MAX)
 341                 mptcp_pm_subflow_established(msk);
 342
 343 out:
 344         __sock_put(sk);
 345 }
 346
 347 struct mptcp_pm_add_entry *
 348 mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 349                        struct mptcp_addr_info *addr, bool check_id)
 350 {
 351         struct mptcp_pm_add_entry *entry;
 352         struct sock *sk = (struct sock *)msk;
 353
 354         spin_lock_bh(&msk->pm.lock);
 355         entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
 356         if (entry && (!check_id || entry->addr.id == addr->id))
 357                 entry->retrans_times = ADD_ADDR_RETRANS_MAX;
 358         spin_unlock_bh(&msk->pm.lock);
 359
 360         if (entry && (!check_id || entry->addr.id == addr->id))
 361                 sk_stop_timer_sync(sk, &entry->add_timer);
 362
 363         return entry;
 364 }
 365
 366 static bool mptcp_pm_alloc_anno_list(struct mptcp_sock *msk,
 367                                      struct mptcp_pm_addr_entry *entry)
 368 {
 369         struct mptcp_pm_add_entry *add_entry = NULL;
 370         struct sock *sk = (struct sock *)msk;
 371         struct net *net = sock_net(sk);
 372
 373         lockdep_assert_held(&msk->pm.lock);
 374
 375         if (mptcp_lookup_anno_list_by_saddr(msk, &entry->addr))
 376                 return false;
 377
 378         add_entry = kmalloc(sizeof(*add_entry), GFP_ATOMIC);
 379         if (!add_entry)
 380                 return false;
 381
 382         list_add(&add_entry->list, &msk->pm.anno_list);
 383
 384         add_entry->addr = entry->addr;
 385         add_entry->sock = msk;
 386         add_entry->retrans_times = 0;
 387
 388         timer_setup(&add_entry->add_timer, mptcp_pm_add_timer, 0);
 389         sk_reset_timer(sk, &add_entry->add_timer,
 390                        jiffies + mptcp_get_add_addr_timeout(net));
 391
 392         return true;
 393 }
 394
 395 void mptcp_pm_free_anno_list(struct mptcp_sock *msk)
 396 {
 397         struct mptcp_pm_add_entry *entry, *tmp;
 398         struct sock *sk = (struct sock *)msk;
 399         LIST_HEAD(free_list);
 400
 401         pr_debug("msk=%p", msk);
 402
 403         spin_lock_bh(&msk->pm.lock);
 404         list_splice_init(&msk->pm.anno_list, &free_list);
 405         spin_unlock_bh(&msk->pm.lock);
 406
 407         list_for_each_entry_safe(entry, tmp, &free_list, list) {
 408                 sk_stop_timer_sync(sk, &entry->add_timer);
 409                 kfree(entry);
 410         }
 411 }
 412
 413 static bool lookup_address_in_vec(struct mptcp_addr_info *addrs, unsigned int nr,
 414                                   struct mptcp_addr_info *addr)
 415 {
 416         int i;
 417
 418         for (i = 0; i < nr; i++) {
 419                 if (addresses_equal(&addrs[i], addr, addr->port))
 420                         return true;
 421         }
 422
 423         return false;
 424 }
 425
 426 /* Fill all the remote addresses into the array addrs[],
 427  * and return the array size.
 428  */
 429 static unsigned int fill_remote_addresses_vec(struct mptcp_sock *msk, bool fullmesh,
 430                                               struct mptcp_addr_info *addrs)
 431 {
 432         struct sock *sk = (struct sock *)msk, *ssk;
 433         struct mptcp_subflow_context *subflow;
 434         struct mptcp_addr_info remote = { 0 };
 435         unsigned int subflows_max;
 436         int i = 0;
 437
 438         subflows_max = mptcp_pm_get_subflows_max(msk);
 439
 440         /* Non-fullmesh endpoint, fill in the single entry
 441          * corresponding to the primary MPC subflow remote address
 442          */
 443         if (!fullmesh) {
 444                 remote_address((struct sock_common *)sk, &remote);
 445                 msk->pm.subflows++;
 446                 addrs[i++] = remote;
 447         } else {
 448                 mptcp_for_each_subflow(msk, subflow) {
 449                         ssk = mptcp_subflow_tcp_sock(subflow);
 450                         remote_address((struct sock_common *)ssk, &remote);
 451                         if (!lookup_address_in_vec(addrs, i, &remote) &&
 452                             msk->pm.subflows < subflows_max) {
 453                                 msk->pm.subflows++;
 454                                 addrs[i++] = remote;
 455                         }
 456                 }
 457         }
 458
 459         return i;
 460 }
 461
 462 static void mptcp_pm_create_subflow_or_signal_addr(struct mptcp_sock *msk)
 463 {
 464         struct sock *sk = (struct sock *)msk;
 465         struct mptcp_pm_addr_entry *local;
 466         unsigned int add_addr_signal_max;
 467         unsigned int local_addr_max;
 468         struct pm_nl_pernet *pernet;
 469         unsigned int subflows_max;
 470
 471         pernet = net_generic(sock_net(sk), pm_nl_pernet_id);
 472
 473         add_addr_signal_max = mptcp_pm_get_add_addr_signal_max(msk);
 474         local_addr_max = mptcp_pm_get_local_addr_max(msk);
 475         subflows_max = mptcp_pm_get_subflows_max(msk);
 476
 477         pr_debug("local %d:%d signal %d:%d subflows %d:%d\n",
 478                  msk->pm.local_addr_used, local_addr_max,
 479                  msk->pm.add_addr_signaled, add_addr_signal_max,
 480                  msk->pm.subflows, subflows_max);
 481
 482         /* check first for announce */
 483         if (msk->pm.add_addr_signaled < add_addr_signal_max) {
 484                 local = select_signal_address(pernet,
 485                                               msk->pm.add_addr_signaled);
 486
 487                 if (local) {
 488                         if (mptcp_pm_alloc_anno_list(msk, local)) {
 489                                 msk->pm.add_addr_signaled++;
 490                                 mptcp_pm_announce_addr(msk, &local->addr, false);
 491                                 mptcp_pm_nl_addr_send_ack(msk);
 492                         }
 493                 } else {
 494                         /* pick failed, avoid fourther attempts later */
 495                         msk->pm.local_addr_used = add_addr_signal_max;
 496                 }
 497
 498                 check_work_pending(msk);
 499         }
 500
 501         /* check if should create a new subflow */
 502         if (msk->pm.local_addr_used < local_addr_max &&
 503             msk->pm.subflows < subflows_max &&
 504             !READ_ONCE(msk->pm.remote_deny_join_id0)) {
 505                 local = select_local_address(pernet, msk);
 506                 if (local) {
 507                         bool fullmesh = !!(local->flags & MPTCP_PM_ADDR_FLAG_FULLMESH);
 508                         struct mptcp_addr_info addrs[MPTCP_PM_ADDR_MAX];
 509                         int i, nr;
 510
 511                         msk->pm.local_addr_used++;
 512                         check_work_pending(msk);
 513                         nr = fill_remote_addresses_vec(msk, fullmesh, addrs);
 514                         spin_unlock_bh(&msk->pm.lock);
 515                         for (i = 0; i < nr; i++)
 516                                 __mptcp_subflow_connect(sk, &local->addr, &addrs[i]);
 517                         spin_lock_bh(&msk->pm.lock);
 518                         return;
 519                 }
 520
 521                 /* lookup failed, avoid fourther attempts later */
 522                 msk->pm.local_addr_used = local_addr_max;
 523                 check_work_pending(msk);
 524         }
 525 }
 526
 527 static void mptcp_pm_nl_fully_established(struct mptcp_sock *msk)
 528 {
 529         mptcp_pm_create_subflow_or_signal_addr(msk);
 530 }
 531
 532 static void mptcp_pm_nl_subflow_established(struct mptcp_sock *msk)
 533 {
 534         mptcp_pm_create_subflow_or_signal_addr(msk);
 535 }
 536
 537 /* Fill all the local addresses into the array addrs[],
 538  * and return the array size.
 539  */
 540 static unsigned int fill_local_addresses_vec(struct mptcp_sock *msk,
 541                                              struct mptcp_addr_info *addrs)
 542 {
 543         struct sock *sk = (struct sock *)msk;
 544         struct mptcp_pm_addr_entry *entry;
 545         struct mptcp_addr_info local;
 546         struct pm_nl_pernet *pernet;
 547         unsigned int subflows_max;
 548         int i = 0;
 549
 550         pernet = net_generic(sock_net(sk), pm_nl_pernet_id);
 551         subflows_max = mptcp_pm_get_subflows_max(msk);
 552
 553         rcu_read_lock();
 554         __mptcp_flush_join_list(msk);
 555         list_for_each_entry_rcu(entry, &pernet->local_addr_list, list) {
 556                 if (!(entry->flags & MPTCP_PM_ADDR_FLAG_FULLMESH))
 557                         continue;
 558
 559                 if (entry->addr.family != sk->sk_family) {
 560 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
 561                         if ((entry->addr.family == AF_INET &&
 562                              !ipv6_addr_v4mapped(&sk->sk_v6_daddr)) ||
 563                             (sk->sk_family == AF_INET &&
 564                              !ipv6_addr_v4mapped(&entry->addr.addr6)))
 565 #endif
 566                                 continue;
 567                 }
 568
 569                 if (msk->pm.subflows < subflows_max) {
 570                         msk->pm.subflows++;
 571                         addrs[i++] = entry->addr;
 572                 }
 573         }
 574         rcu_read_unlock();
 575
 576         /* If the array is empty, fill in the single
 577          * 'IPADDRANY' local address
 578          */
 579         if (!i) {
 580                 memset(&local, 0, sizeof(local));
 581                 local.family = msk->pm.remote.family;
 582
 583                 msk->pm.subflows++;
 584                 addrs[i++] = local;
 585         }
 586
 587         return i;
 588 }
 589
 590 static void mptcp_pm_nl_add_addr_received(struct mptcp_sock *msk)
 591 {
 592         struct mptcp_addr_info addrs[MPTCP_PM_ADDR_MAX];
 593         struct sock *sk = (struct sock *)msk;
 594         unsigned int add_addr_accept_max;
 595         struct mptcp_addr_info remote;
 596         unsigned int subflows_max;
 597         int i, nr;
 598
 599         add_addr_accept_max = mptcp_pm_get_add_addr_accept_max(msk);
 600         subflows_max = mptcp_pm_get_subflows_max(msk);
 601
 602         pr_debug("accepted %d:%d remote family %d",
 603                  msk->pm.add_addr_accepted, add_addr_accept_max,
 604                  msk->pm.remote.family);
 605
 606         if (lookup_subflow_by_daddr(&msk->conn_list, &msk->pm.remote))
 607                 goto add_addr_echo;
 608
 609         /* connect to the specified remote address, using whatever
 610          * local address the routing configuration will pick.
 611          */
 612         remote = msk->pm.remote;
 613         if (!remote.port)
 614                 remote.port = sk->sk_dport;
 615         nr = fill_local_addresses_vec(msk, addrs);
 616
 617         msk->pm.add_addr_accepted++;
 618         if (msk->pm.add_addr_accepted >= add_addr_accept_max ||
 619             msk->pm.subflows >= subflows_max)
 620                 WRITE_ONCE(msk->pm.accept_addr, false);
 621
 622         spin_unlock_bh(&msk->pm.lock);
 623         for (i = 0; i < nr; i++)
 624                 __mptcp_subflow_connect(sk, &addrs[i], &remote);
 625         spin_lock_bh(&msk->pm.lock);
 626
 627 add_addr_echo:
 628         mptcp_pm_announce_addr(msk, &msk->pm.remote, true);
 629         mptcp_pm_nl_addr_send_ack(msk);
 630 }
 631
 632 void mptcp_pm_nl_addr_send_ack(struct mptcp_sock *msk)
 633 {
 634         struct mptcp_subflow_context *subflow;
 635
 636         msk_owned_by_me(msk);
 637         lockdep_assert_held(&msk->pm.lock);
 638
 639         if (!mptcp_pm_should_add_signal(msk) &&
 640             !mptcp_pm_should_rm_signal(msk))
 641                 return;
 642
 643         __mptcp_flush_join_list(msk);
 644         subflow = list_first_entry_or_null(&msk->conn_list, typeof(*subflow), node);
 645         if (subflow) {
 646                 struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 647
 648                 spin_unlock_bh(&msk->pm.lock);
 649                 pr_debug("send ack for %s",
 650                          mptcp_pm_should_add_signal(msk) ? "add_addr" : "rm_addr");
 651
 652                 mptcp_subflow_send_ack(ssk);
 653                 spin_lock_bh(&msk->pm.lock);
 654         }
 655 }
 656
 657 int mptcp_pm_nl_mp_prio_send_ack(struct mptcp_sock *msk,
 658                                  struct mptcp_addr_info *addr,
 659                                  u8 bkup)
 660 {
 661         struct mptcp_subflow_context *subflow;
 662
 663         pr_debug("bkup=%d", bkup);
 664
 665         mptcp_for_each_subflow(msk, subflow) {
 666                 struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 667                 struct sock *sk = (struct sock *)msk;
 668                 struct mptcp_addr_info local;
 669
 670                 local_address((struct sock_common *)ssk, &local);
 671                 if (!addresses_equal(&local, addr, addr->port))
 672                         continue;
 673
 674                 subflow->backup = bkup;
 675                 subflow->send_mp_prio = 1;
 676                 subflow->request_bkup = bkup;
 677                 __MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_MPPRIOTX);
 678
 679                 spin_unlock_bh(&msk->pm.lock);
 680                 pr_debug("send ack for mp_prio");
 681                 mptcp_subflow_send_ack(ssk);
 682                 spin_lock_bh(&msk->pm.lock);
 683
 684                 return 0;
 685         }
 686
 687         return -EINVAL;
 688 }
 689
 690 static void mptcp_pm_nl_rm_addr_or_subflow(struct mptcp_sock *msk,
 691                                            const struct mptcp_rm_list *rm_list,
 692                                            enum linux_mptcp_mib_field rm_type)
 693 {
 694         struct mptcp_subflow_context *subflow, *tmp;
 695         struct sock *sk = (struct sock *)msk;
 696         u8 i;
 697
 698         pr_debug("%s rm_list_nr %d",
 699                  rm_type == MPTCP_MIB_RMADDR ? "address" : "subflow", rm_list->nr);
 700
 701         msk_owned_by_me(msk);
 702
 703         if (!rm_list->nr)
 704                 return;
 705
 706         if (list_empty(&msk->conn_list))
 707                 return;
 708
 709         for (i = 0; i < rm_list->nr; i++) {
 710                 list_for_each_entry_safe(subflow, tmp, &msk->conn_list, node) {
 711                         struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 712                         int how = RCV_SHUTDOWN | SEND_SHUTDOWN;
 713                         u8 id = subflow->local_id;
 714
 715                         if (rm_type == MPTCP_MIB_RMADDR)
 716                                 id = subflow->remote_id;
 717
 718                         if (rm_list->ids[i] != id)
 719                                 continue;
 720
 721                         pr_debug(" -> %s rm_list_ids[%d]=%u local_id=%u remote_id=%u",
 722                                  rm_type == MPTCP_MIB_RMADDR ? "address" : "subflow",
 723                                  i, rm_list->ids[i], subflow->local_id, subflow->remote_id);
 724                         spin_unlock_bh(&msk->pm.lock);
 725                         mptcp_subflow_shutdown(sk, ssk, how);
 726                         mptcp_close_ssk(sk, ssk, subflow);
 727                         spin_lock_bh(&msk->pm.lock);
 728
 729                         if (rm_type == MPTCP_MIB_RMADDR) {
 730                                 msk->pm.add_addr_accepted--;
 731                                 WRITE_ONCE(msk->pm.accept_addr, true);
 732                         } else if (rm_type == MPTCP_MIB_RMSUBFLOW) {
 733                                 msk->pm.local_addr_used--;
 734                         }
 735                         msk->pm.subflows--;
 736                         __MPTCP_INC_STATS(sock_net(sk), rm_type);
 737                 }
 738         }
 739 }
 740
 741 static void mptcp_pm_nl_rm_addr_received(struct mptcp_sock *msk)
 742 {
 743         mptcp_pm_nl_rm_addr_or_subflow(msk, &msk->pm.rm_list_rx, MPTCP_MIB_RMADDR);
 744 }
 745
 746 void mptcp_pm_nl_rm_subflow_received(struct mptcp_sock *msk,
 747                                      const struct mptcp_rm_list *rm_list)
 748 {
 749         mptcp_pm_nl_rm_addr_or_subflow(msk, rm_list, MPTCP_MIB_RMSUBFLOW);
 750 }
 751
 752 void mptcp_pm_nl_work(struct mptcp_sock *msk)
 753 {
 754         struct mptcp_pm_data *pm = &msk->pm;
 755
 756         msk_owned_by_me(msk);
 757
 758         spin_lock_bh(&msk->pm.lock);
 759
 760         pr_debug("msk=%p status=%x", msk, pm->status);
 761         if (pm->status & BIT(MPTCP_PM_ADD_ADDR_RECEIVED)) {
 762                 pm->status &= ~BIT(MPTCP_PM_ADD_ADDR_RECEIVED);
 763                 mptcp_pm_nl_add_addr_received(msk);
 764         }
 765         if (pm->status & BIT(MPTCP_PM_ADD_ADDR_SEND_ACK)) {
 766                 pm->status &= ~BIT(MPTCP_PM_ADD_ADDR_SEND_ACK);
 767                 mptcp_pm_nl_addr_send_ack(msk);
 768         }
 769         if (pm->status & BIT(MPTCP_PM_RM_ADDR_RECEIVED)) {
 770                 pm->status &= ~BIT(MPTCP_PM_RM_ADDR_RECEIVED);
 771                 mptcp_pm_nl_rm_addr_received(msk);
 772         }
 773         if (pm->status & BIT(MPTCP_PM_ESTABLISHED)) {
 774                 pm->status &= ~BIT(MPTCP_PM_ESTABLISHED);
 775                 mptcp_pm_nl_fully_established(msk);
 776         }
 777         if (pm->status & BIT(MPTCP_PM_SUBFLOW_ESTABLISHED)) {
 778                 pm->status &= ~BIT(MPTCP_PM_SUBFLOW_ESTABLISHED);
 779                 mptcp_pm_nl_subflow_established(msk);
 780         }
 781
 782         spin_unlock_bh(&msk->pm.lock);
 783 }
 784
 785 static bool address_use_port(struct mptcp_pm_addr_entry *entry)
 786 {
 787         return (entry->flags &
 788                 (MPTCP_PM_ADDR_FLAG_SIGNAL | MPTCP_PM_ADDR_FLAG_SUBFLOW)) ==
 789                 MPTCP_PM_ADDR_FLAG_SIGNAL;
 790 }
 791
 792 static int mptcp_pm_nl_append_new_local_addr(struct pm_nl_pernet *pernet,
 793                                              struct mptcp_pm_addr_entry *entry)
 794 {
 795         struct mptcp_pm_addr_entry *cur;
 796         unsigned int addr_max;
 797         int ret = -EINVAL;
 798
 799         spin_lock_bh(&pernet->lock);
 800         /* to keep the code simple, don't do IDR-like allocation for address ID,
 801          * just bail when we exceed limits
 802          */
 803         if (pernet->next_id == MAX_ADDR_ID)
 804                 pernet->next_id = 1;
 805         if (pernet->addrs >= MPTCP_PM_ADDR_MAX)
 806                 goto out;
 807         if (test_bit(entry->addr.id, pernet->id_bitmap))
 808                 goto out;
 809
 810         /* do not insert duplicate address, differentiate on port only
 811          * singled addresses
 812          */
 813         list_for_each_entry(cur, &pernet->local_addr_list, list) {
 814                 if (addresses_equal(&cur->addr, &entry->addr,
 815                                     address_use_port(entry) &&
 816                                     address_use_port(cur)))
 817                         goto out;
 818         }
 819
 820         if (!entry->addr.id) {
 821 find_next:
 822                 entry->addr.id = find_next_zero_bit(pernet->id_bitmap,
 823                                                     MAX_ADDR_ID + 1,
 824                                                     pernet->next_id);
 825                 if ((!entry->addr.id || entry->addr.id > MAX_ADDR_ID) &&
 826                     pernet->next_id != 1) {
 827                         pernet->next_id = 1;
 828                         goto find_next;
 829                 }
 830         }
 831
 832         if (!entry->addr.id || entry->addr.id > MAX_ADDR_ID)
 833                 goto out;
 834
 835         __set_bit(entry->addr.id, pernet->id_bitmap);
 836         if (entry->addr.id > pernet->next_id)
 837                 pernet->next_id = entry->addr.id;
 838
 839         if (entry->flags & MPTCP_PM_ADDR_FLAG_SIGNAL) {
 840                 addr_max = pernet->add_addr_signal_max;
 841                 WRITE_ONCE(pernet->add_addr_signal_max, addr_max + 1);
 842         }
 843         if (entry->flags & MPTCP_PM_ADDR_FLAG_SUBFLOW) {
 844                 addr_max = pernet->local_addr_max;
 845                 WRITE_ONCE(pernet->local_addr_max, addr_max + 1);
 846         }
 847
 848         pernet->addrs++;
 849         list_add_tail_rcu(&entry->list, &pernet->local_addr_list);
 850         ret = entry->addr.id;
 851
 852 out:
 853         spin_unlock_bh(&pernet->lock);
 854         return ret;
 855 }
 856
 857 static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
 858                                             struct mptcp_pm_addr_entry *entry)
 859 {
 860         struct sockaddr_storage addr;
 861         struct mptcp_sock *msk;
 862         struct socket *ssock;
 863         int backlog = 1024;
 864         int err;
 865
 866         err = sock_create_kern(sock_net(sk), entry->addr.family,
 867                                SOCK_STREAM, IPPROTO_MPTCP, &entry->lsk);
 868         if (err)
 869                 return err;
 870
 871         msk = mptcp_sk(entry->lsk->sk);
 872         if (!msk) {
 873                 err = -EINVAL;
 874                 goto out;
 875         }
 876
 877         ssock = __mptcp_nmpc_socket(msk);
 878         if (!ssock) {
 879                 err = -EINVAL;
 880                 goto out;
 881         }
 882
 883         mptcp_info2sockaddr(&entry->addr, &addr, entry->addr.family);
 884         err = kernel_bind(ssock, (struct sockaddr *)&addr,
 885                           sizeof(struct sockaddr_in));
 886         if (err) {
 887                 pr_warn("kernel_bind error, err=%d", err);
 888                 goto out;
 889         }
 890
 891         err = kernel_listen(ssock, backlog);
 892         if (err) {
 893                 pr_warn("kernel_listen error, err=%d", err);
 894                 goto out;
 895         }
 896
 897         return 0;
 898
 899 out:
 900         sock_release(entry->lsk);
 901         return err;
 902 }
 903
 904 int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
 905 {
 906         struct mptcp_pm_addr_entry *entry;
 907         struct mptcp_addr_info skc_local;
 908         struct mptcp_addr_info msk_local;
 909         struct pm_nl_pernet *pernet;
 910         int ret = -1;
 911
 912         if (WARN_ON_ONCE(!msk))
 913                 return -1;
 914
 915         /* The 0 ID mapping is defined by the first subflow, copied into the msk
 916          * addr
 917          */
 918         local_address((struct sock_common *)msk, &msk_local);
 919         local_address((struct sock_common *)skc, &skc_local);
 920         if (addresses_equal(&msk_local, &skc_local, false))
 921                 return 0;
 922
 923         if (address_zero(&skc_local))
 924                 return 0;
 925
 926         pernet = net_generic(sock_net((struct sock *)msk), pm_nl_pernet_id);
 927
 928         rcu_read_lock();
 929         list_for_each_entry_rcu(entry, &pernet->local_addr_list, list) {
 930                 if (addresses_equal(&entry->addr, &skc_local, entry->addr.port)) {
 931                         ret = entry->addr.id;
 932                         break;
 933                 }
 934         }
 935         rcu_read_unlock();
 936         if (ret >= 0)
 937                 return ret;
 938
 939         /* address not found, add to local list */
 940         entry = kmalloc(sizeof(*entry), GFP_ATOMIC);
 941         if (!entry)
 942                 return -ENOMEM;
 943
 944         entry->addr = skc_local;
 945         entry->addr.id = 0;
 946         entry->addr.port = 0;
 947         entry->ifindex = 0;
 948         entry->flags = 0;
 949         entry->lsk = NULL;
 950         ret = mptcp_pm_nl_append_new_local_addr(pernet, entry);
 951         if (ret < 0)
 952                 kfree(entry);
 953
 954         return ret;
 955 }
 956
 957 void mptcp_pm_nl_data_init(struct mptcp_sock *msk)
 958 {
 959         struct mptcp_pm_data *pm = &msk->pm;
 960         bool subflows;
 961
 962         subflows = !!mptcp_pm_get_subflows_max(msk);
 963         WRITE_ONCE(pm->work_pending, (!!mptcp_pm_get_local_addr_max(msk) && subflows) ||
 964                    !!mptcp_pm_get_add_addr_signal_max(msk));
 965         WRITE_ONCE(pm->accept_addr, !!mptcp_pm_get_add_addr_accept_max(msk) && subflows);
 966         WRITE_ONCE(pm->accept_subflow, subflows);
 967 }
 968
 969 #define MPTCP_PM_CMD_GRP_OFFSET       0
 970 #define MPTCP_PM_EV_GRP_OFFSET        1
 971
 972 static const struct genl_multicast_group mptcp_pm_mcgrps[] = {
 973         [MPTCP_PM_CMD_GRP_OFFSET]       = { .name = MPTCP_PM_CMD_GRP_NAME, },
 974         [MPTCP_PM_EV_GRP_OFFSET]        = { .name = MPTCP_PM_EV_GRP_NAME,
 975                                             .flags = GENL_UNS_ADMIN_PERM,
 976                                           },
 977 };
 978
 979 static const struct nla_policy
 980 mptcp_pm_addr_policy[MPTCP_PM_ADDR_ATTR_MAX + 1] = {
 981         [MPTCP_PM_ADDR_ATTR_FAMILY]     = { .type       = NLA_U16,      },
 982         [MPTCP_PM_ADDR_ATTR_ID]         = { .type       = NLA_U8,       },
 983         [MPTCP_PM_ADDR_ATTR_ADDR4]      = { .type       = NLA_U32,      },
 984         [MPTCP_PM_ADDR_ATTR_ADDR6]      =
 985                 NLA_POLICY_EXACT_LEN(sizeof(struct in6_addr)),
 986         [MPTCP_PM_ADDR_ATTR_PORT]       = { .type       = NLA_U16       },
 987         [MPTCP_PM_ADDR_ATTR_FLAGS]      = { .type       = NLA_U32       },
 988         [MPTCP_PM_ADDR_ATTR_IF_IDX]     = { .type       = NLA_S32       },
 989 };
 990
 991 static const struct nla_policy mptcp_pm_policy[MPTCP_PM_ATTR_MAX + 1] = {
 992         [MPTCP_PM_ATTR_ADDR]            =
 993                                         NLA_POLICY_NESTED(mptcp_pm_addr_policy),
 994         [MPTCP_PM_ATTR_RCV_ADD_ADDRS]   = { .type       = NLA_U32,      },
 995         [MPTCP_PM_ATTR_SUBFLOWS]        = { .type       = NLA_U32,      },
 996 };
 997
 998 void mptcp_pm_nl_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
 999 {
1000         struct mptcp_subflow_context *iter, *subflow = mptcp_subflow_ctx(ssk);
1001         struct sock *sk = (struct sock *)msk;
1002         unsigned int active_max_loss_cnt;
1003         struct net *net = sock_net(sk);
1004         unsigned int stale_loss_cnt;
1005         bool slow;
1006
1007         stale_loss_cnt = mptcp_stale_loss_cnt(net);
1008         if (subflow->stale || !stale_loss_cnt || subflow->stale_count <= stale_loss_cnt)
1009                 return;
1010
1011         /* look for another available subflow not in loss state */
1012         active_max_loss_cnt = max_t(int, stale_loss_cnt - 1, 1);
1013         mptcp_for_each_subflow(msk, iter) {
1014                 if (iter != subflow && mptcp_subflow_active(iter) &&
1015                     iter->stale_count < active_max_loss_cnt) {
1016                         /* we have some alternatives, try to mark this subflow as idle ...*/
1017                         slow = lock_sock_fast(ssk);
1018                         if (!tcp_rtx_and_write_queues_empty(ssk)) {
1019                                 subflow->stale = 1;
1020                                 __mptcp_retransmit_pending_data(sk);
1021                                 MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_SUBFLOWSTALE);
1022                         }
1023                         unlock_sock_fast(ssk, slow);
1024
1025                         /* always try to push the pending data regarless of re-injections:
1026                          * we can possibly use backup subflows now, and subflow selection
1027                          * is cheap under the msk socket lock
1028                          */
1029                         __mptcp_push_pending(sk, 0);
1030                         return;
1031                 }
1032         }
1033 }
1034
1035 static int mptcp_pm_family_to_addr(int family)
1036 {
1037 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
1038         if (family == AF_INET6)
1039                 return MPTCP_PM_ADDR_ATTR_ADDR6;
1040 #endif
1041         return MPTCP_PM_ADDR_ATTR_ADDR4;
1042 }
1043
1044 static int mptcp_pm_parse_addr(struct nlattr *attr, struct genl_info *info,
1045                                bool require_family,
1046                                struct mptcp_pm_addr_entry *entry)
1047 {
1048         struct nlattr *tb[MPTCP_PM_ADDR_ATTR_MAX + 1];
1049         int err, addr_addr;
1050
1051         if (!attr) {
1052                 GENL_SET_ERR_MSG(info, "missing address info");
1053                 return -EINVAL;
1054         }
1055
1056         /* no validation needed - was already done via nested policy */
1057         err = nla_parse_nested_deprecated(tb, MPTCP_PM_ADDR_ATTR_MAX, attr,
1058                                           mptcp_pm_addr_policy, info->extack);
1059         if (err)
1060                 return err;
1061
1062         memset(entry, 0, sizeof(*entry));
1063         if (!tb[MPTCP_PM_ADDR_ATTR_FAMILY]) {
1064                 if (!require_family)
1065                         goto skip_family;
1066
1067                 NL_SET_ERR_MSG_ATTR(info->extack, attr,
1068                                     "missing family");
1069                 return -EINVAL;
1070         }
1071
1072         entry->addr.family = nla_get_u16(tb[MPTCP_PM_ADDR_ATTR_FAMILY]);
1073         if (entry->addr.family != AF_INET
1074 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
1075             && entry->addr.family != AF_INET6
1076 #endif
1077             ) {
1078                 NL_SET_ERR_MSG_ATTR(info->extack, attr,
1079                                     "unknown address family");
1080                 return -EINVAL;
1081         }
1082         addr_addr = mptcp_pm_family_to_addr(entry->addr.family);
1083         if (!tb[addr_addr]) {
1084                 NL_SET_ERR_MSG_ATTR(info->extack, attr,
1085                                     "missing address data");
1086                 return -EINVAL;
1087         }
1088
1089 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
1090         if (entry->addr.family == AF_INET6)
1091                 entry->addr.addr6 = nla_get_in6_addr(tb[addr_addr]);
1092         else
1093 #endif
1094                 entry->addr.addr.s_addr = nla_get_in_addr(tb[addr_addr]);
1095
1096 skip_family:
1097         if (tb[MPTCP_PM_ADDR_ATTR_IF_IDX]) {
1098                 u32 val = nla_get_s32(tb[MPTCP_PM_ADDR_ATTR_IF_IDX]);
1099
1100                 entry->ifindex = val;
1101         }
1102
1103         if (tb[MPTCP_PM_ADDR_ATTR_ID])
1104                 entry->addr.id = nla_get_u8(tb[MPTCP_PM_ADDR_ATTR_ID]);
1105
1106         if (tb[MPTCP_PM_ADDR_ATTR_FLAGS])
1107                 entry->flags = nla_get_u32(tb[MPTCP_PM_ADDR_ATTR_FLAGS]);
1108
1109         if (tb[MPTCP_PM_ADDR_ATTR_PORT]) {
1110                 if (!(entry->flags & MPTCP_PM_ADDR_FLAG_SIGNAL)) {
1111                         NL_SET_ERR_MSG_ATTR(info->extack, attr,
1112                                             "flags must have signal when using port");
1113                         return -EINVAL;
1114                 }
1115                 entry->addr.port = htons(nla_get_u16(tb[MPTCP_PM_ADDR_ATTR_PORT]));
1116         }
1117
1118         return 0;
1119 }
1120
1121 static struct pm_nl_pernet *genl_info_pm_nl(struct genl_info *info)
1122 {
1123         return net_generic(genl_info_net(info), pm_nl_pernet_id);
1124 }
1125
1126 static int mptcp_nl_add_subflow_or_signal_addr(struct net *net)
1127 {
1128         struct mptcp_sock *msk;
1129         long s_slot = 0, s_num = 0;
1130
1131         while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
1132                 struct sock *sk = (struct sock *)msk;
1133
1134                 if (!READ_ONCE(msk->fully_established))
1135                         goto next;
1136
1137                 lock_sock(sk);
1138                 spin_lock_bh(&msk->pm.lock);
1139                 mptcp_pm_create_subflow_or_signal_addr(msk);
1140                 spin_unlock_bh(&msk->pm.lock);
1141                 release_sock(sk);
1142
1143 next:
1144                 sock_put(sk);
1145                 cond_resched();
1146         }
1147
1148         return 0;
1149 }
1150
1151 static int mptcp_nl_cmd_add_addr(struct sk_buff *skb, struct genl_info *info)
1152 {
1153         struct nlattr *attr = info->attrs[MPTCP_PM_ATTR_ADDR];
1154         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1155         struct mptcp_pm_addr_entry addr, *entry;
1156         int ret;
1157
1158         ret = mptcp_pm_parse_addr(attr, info, true, &addr);
1159         if (ret < 0)
1160                 return ret;
1161
1162         entry = kmalloc(sizeof(*entry), GFP_KERNEL);
1163         if (!entry) {
1164                 GENL_SET_ERR_MSG(info, "can't allocate addr");
1165                 return -ENOMEM;
1166         }
1167
1168         *entry = addr;
1169         if (entry->addr.port) {
1170                 ret = mptcp_pm_nl_create_listen_socket(skb->sk, entry);
1171                 if (ret) {
1172                         GENL_SET_ERR_MSG(info, "create listen socket error");
1173                         kfree(entry);
1174                         return ret;
1175                 }
1176         }
1177         ret = mptcp_pm_nl_append_new_local_addr(pernet, entry);
1178         if (ret < 0) {
1179                 GENL_SET_ERR_MSG(info, "too many addresses or duplicate one");
1180                 if (entry->lsk)
1181                         sock_release(entry->lsk);
1182                 kfree(entry);
1183                 return ret;
1184         }
1185
1186         mptcp_nl_add_subflow_or_signal_addr(sock_net(skb->sk));
1187
1188         return 0;
1189 }
1190
1191 static struct mptcp_pm_addr_entry *
1192 __lookup_addr_by_id(struct pm_nl_pernet *pernet, unsigned int id)
1193 {
1194         struct mptcp_pm_addr_entry *entry;
1195
1196         list_for_each_entry(entry, &pernet->local_addr_list, list) {
1197                 if (entry->addr.id == id)
1198                         return entry;
1199         }
1200         return NULL;
1201 }
1202
1203 int mptcp_pm_get_flags_and_ifindex_by_id(struct net *net, unsigned int id,
1204                                          u8 *flags, int *ifindex)
1205 {
1206         struct mptcp_pm_addr_entry *entry;
1207
1208         *flags = 0;
1209         *ifindex = 0;
1210
1211         if (id) {
1212                 rcu_read_lock();
1213                 entry = __lookup_addr_by_id(net_generic(net, pm_nl_pernet_id), id);
1214                 if (entry) {
1215                         *flags = entry->flags;
1216                         *ifindex = entry->ifindex;
1217                 }
1218                 rcu_read_unlock();
1219         }
1220
1221         return 0;
1222 }
1223
1224 static bool remove_anno_list_by_saddr(struct mptcp_sock *msk,
1225                                       struct mptcp_addr_info *addr)
1226 {
1227         struct mptcp_pm_add_entry *entry;
1228
1229         entry = mptcp_pm_del_add_timer(msk, addr, false);
1230         if (entry) {
1231                 list_del(&entry->list);
1232                 kfree(entry);
1233                 return true;
1234         }
1235
1236         return false;
1237 }
1238
1239 static bool mptcp_pm_remove_anno_addr(struct mptcp_sock *msk,
1240                                       struct mptcp_addr_info *addr,
1241                                       bool force)
1242 {
1243         struct mptcp_rm_list list = { .nr = 0 };
1244         bool ret;
1245
1246         list.ids[list.nr++] = addr->id;
1247
1248         ret = remove_anno_list_by_saddr(msk, addr);
1249         if (ret || force) {
1250                 spin_lock_bh(&msk->pm.lock);
1251                 mptcp_pm_remove_addr(msk, &list);
1252                 spin_unlock_bh(&msk->pm.lock);
1253         }
1254         return ret;
1255 }
1256
1257 static int mptcp_nl_remove_subflow_and_signal_addr(struct net *net,
1258                                                    struct mptcp_addr_info *addr)
1259 {
1260         struct mptcp_sock *msk;
1261         long s_slot = 0, s_num = 0;
1262         struct mptcp_rm_list list = { .nr = 0 };
1263
1264         pr_debug("remove_id=%d", addr->id);
1265
1266         list.ids[list.nr++] = addr->id;
1267
1268         while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
1269                 struct sock *sk = (struct sock *)msk;
1270                 bool remove_subflow;
1271
1272                 if (list_empty(&msk->conn_list)) {
1273                         mptcp_pm_remove_anno_addr(msk, addr, false);
1274                         goto next;
1275                 }
1276
1277                 lock_sock(sk);
1278                 remove_subflow = lookup_subflow_by_saddr(&msk->conn_list, addr);
1279                 mptcp_pm_remove_anno_addr(msk, addr, remove_subflow);
1280                 if (remove_subflow)
1281                         mptcp_pm_remove_subflow(msk, &list);
1282                 release_sock(sk);
1283
1284 next:
1285                 sock_put(sk);
1286                 cond_resched();
1287         }
1288
1289         return 0;
1290 }
1291
1292 /* caller must ensure the RCU grace period is already elapsed */
1293 static void __mptcp_pm_release_addr_entry(struct mptcp_pm_addr_entry *entry)
1294 {
1295         if (entry->lsk)
1296                 sock_release(entry->lsk);
1297         kfree(entry);
1298 }
1299
1300 static int mptcp_nl_remove_id_zero_address(struct net *net,
1301                                            struct mptcp_addr_info *addr)
1302 {
1303         struct mptcp_rm_list list = { .nr = 0 };
1304         long s_slot = 0, s_num = 0;
1305         struct mptcp_sock *msk;
1306
1307         list.ids[list.nr++] = 0;
1308
1309         while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
1310                 struct sock *sk = (struct sock *)msk;
1311                 struct mptcp_addr_info msk_local;
1312
1313                 if (list_empty(&msk->conn_list))
1314                         goto next;
1315
1316                 local_address((struct sock_common *)msk, &msk_local);
1317                 if (!addresses_equal(&msk_local, addr, addr->port))
1318                         goto next;
1319
1320                 lock_sock(sk);
1321                 spin_lock_bh(&msk->pm.lock);
1322                 mptcp_pm_remove_addr(msk, &list);
1323                 mptcp_pm_nl_rm_subflow_received(msk, &list);
1324                 spin_unlock_bh(&msk->pm.lock);
1325                 release_sock(sk);
1326
1327 next:
1328                 sock_put(sk);
1329                 cond_resched();
1330         }
1331
1332         return 0;
1333 }
1334
1335 static int mptcp_nl_cmd_del_addr(struct sk_buff *skb, struct genl_info *info)
1336 {
1337         struct nlattr *attr = info->attrs[MPTCP_PM_ATTR_ADDR];
1338         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1339         struct mptcp_pm_addr_entry addr, *entry;
1340         unsigned int addr_max;
1341         int ret;
1342
1343         ret = mptcp_pm_parse_addr(attr, info, false, &addr);
1344         if (ret < 0)
1345                 return ret;
1346
1347         /* the zero id address is special: the first address used by the msk
1348          * always gets such an id, so different subflows can have different zero
1349          * id addresses. Additionally zero id is not accounted for in id_bitmap.
1350          * Let's use an 'mptcp_rm_list' instead of the common remove code.
1351          */
1352         if (addr.addr.id == 0)
1353                 return mptcp_nl_remove_id_zero_address(sock_net(skb->sk), &addr.addr);
1354
1355         spin_lock_bh(&pernet->lock);
1356         entry = __lookup_addr_by_id(pernet, addr.addr.id);
1357         if (!entry) {
1358                 GENL_SET_ERR_MSG(info, "address not found");
1359                 spin_unlock_bh(&pernet->lock);
1360                 return -EINVAL;
1361         }
1362         if (entry->flags & MPTCP_PM_ADDR_FLAG_SIGNAL) {
1363                 addr_max = pernet->add_addr_signal_max;
1364                 WRITE_ONCE(pernet->add_addr_signal_max, addr_max - 1);
1365         }
1366         if (entry->flags & MPTCP_PM_ADDR_FLAG_SUBFLOW) {
1367                 addr_max = pernet->local_addr_max;
1368                 WRITE_ONCE(pernet->local_addr_max, addr_max - 1);
1369         }
1370
1371         pernet->addrs--;
1372         list_del_rcu(&entry->list);
1373         __clear_bit(entry->addr.id, pernet->id_bitmap);
1374         spin_unlock_bh(&pernet->lock);
1375
1376         mptcp_nl_remove_subflow_and_signal_addr(sock_net(skb->sk), &entry->addr);
1377         synchronize_rcu();
1378         __mptcp_pm_release_addr_entry(entry);
1379
1380         return ret;
1381 }
1382
1383 static void mptcp_pm_remove_addrs_and_subflows(struct mptcp_sock *msk,
1384                                                struct list_head *rm_list)
1385 {
1386         struct mptcp_rm_list alist = { .nr = 0 }, slist = { .nr = 0 };
1387         struct mptcp_pm_addr_entry *entry;
1388
1389         list_for_each_entry(entry, rm_list, list) {
1390                 if (lookup_subflow_by_saddr(&msk->conn_list, &entry->addr) &&
1391                     alist.nr < MPTCP_RM_IDS_MAX &&
1392                     slist.nr < MPTCP_RM_IDS_MAX) {
1393                         alist.ids[alist.nr++] = entry->addr.id;
1394                         slist.ids[slist.nr++] = entry->addr.id;
1395                 } else if (remove_anno_list_by_saddr(msk, &entry->addr) &&
1396                          alist.nr < MPTCP_RM_IDS_MAX) {
1397                         alist.ids[alist.nr++] = entry->addr.id;
1398                 }
1399         }
1400
1401         if (alist.nr) {
1402                 spin_lock_bh(&msk->pm.lock);
1403                 mptcp_pm_remove_addr(msk, &alist);
1404                 spin_unlock_bh(&msk->pm.lock);
1405         }
1406         if (slist.nr)
1407                 mptcp_pm_remove_subflow(msk, &slist);
1408 }
1409
1410 static void mptcp_nl_remove_addrs_list(struct net *net,
1411                                        struct list_head *rm_list)
1412 {
1413         long s_slot = 0, s_num = 0;
1414         struct mptcp_sock *msk;
1415
1416         if (list_empty(rm_list))
1417                 return;
1418
1419         while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
1420                 struct sock *sk = (struct sock *)msk;
1421
1422                 lock_sock(sk);
1423                 mptcp_pm_remove_addrs_and_subflows(msk, rm_list);
1424                 release_sock(sk);
1425
1426                 sock_put(sk);
1427                 cond_resched();
1428         }
1429 }
1430
1431 /* caller must ensure the RCU grace period is already elapsed */
1432 static void __flush_addrs(struct list_head *list)
1433 {
1434         while (!list_empty(list)) {
1435                 struct mptcp_pm_addr_entry *cur;
1436
1437                 cur = list_entry(list->next,
1438                                  struct mptcp_pm_addr_entry, list);
1439                 list_del_rcu(&cur->list);
1440                 __mptcp_pm_release_addr_entry(cur);
1441         }
1442 }
1443
1444 static void __reset_counters(struct pm_nl_pernet *pernet)
1445 {
1446         WRITE_ONCE(pernet->add_addr_signal_max, 0);
1447         WRITE_ONCE(pernet->add_addr_accept_max, 0);
1448         WRITE_ONCE(pernet->local_addr_max, 0);
1449         pernet->addrs = 0;
1450 }
1451
1452 static int mptcp_nl_cmd_flush_addrs(struct sk_buff *skb, struct genl_info *info)
1453 {
1454         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1455         LIST_HEAD(free_list);
1456
1457         spin_lock_bh(&pernet->lock);
1458         list_splice_init(&pernet->local_addr_list, &free_list);
1459         __reset_counters(pernet);
1460         pernet->next_id = 1;
1461         bitmap_zero(pernet->id_bitmap, MAX_ADDR_ID + 1);
1462         spin_unlock_bh(&pernet->lock);
1463         mptcp_nl_remove_addrs_list(sock_net(skb->sk), &free_list);
1464         synchronize_rcu();
1465         __flush_addrs(&free_list);
1466         return 0;
1467 }
1468
1469 static int mptcp_nl_fill_addr(struct sk_buff *skb,
1470                               struct mptcp_pm_addr_entry *entry)
1471 {
1472         struct mptcp_addr_info *addr = &entry->addr;
1473         struct nlattr *attr;
1474
1475         attr = nla_nest_start(skb, MPTCP_PM_ATTR_ADDR);
1476         if (!attr)
1477                 return -EMSGSIZE;
1478
1479         if (nla_put_u16(skb, MPTCP_PM_ADDR_ATTR_FAMILY, addr->family))
1480                 goto nla_put_failure;
1481         if (nla_put_u16(skb, MPTCP_PM_ADDR_ATTR_PORT, ntohs(addr->port)))
1482                 goto nla_put_failure;
1483         if (nla_put_u8(skb, MPTCP_PM_ADDR_ATTR_ID, addr->id))
1484                 goto nla_put_failure;
1485         if (nla_put_u32(skb, MPTCP_PM_ADDR_ATTR_FLAGS, entry->flags))
1486                 goto nla_put_failure;
1487         if (entry->ifindex &&
1488             nla_put_s32(skb, MPTCP_PM_ADDR_ATTR_IF_IDX, entry->ifindex))
1489                 goto nla_put_failure;
1490
1491         if (addr->family == AF_INET &&
1492             nla_put_in_addr(skb, MPTCP_PM_ADDR_ATTR_ADDR4,
1493                             addr->addr.s_addr))
1494                 goto nla_put_failure;
1495 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
1496         else if (addr->family == AF_INET6 &&
1497                  nla_put_in6_addr(skb, MPTCP_PM_ADDR_ATTR_ADDR6, &addr->addr6))
1498                 goto nla_put_failure;
1499 #endif
1500         nla_nest_end(skb, attr);
1501         return 0;
1502
1503 nla_put_failure:
1504         nla_nest_cancel(skb, attr);
1505         return -EMSGSIZE;
1506 }
1507
1508 static int mptcp_nl_cmd_get_addr(struct sk_buff *skb, struct genl_info *info)
1509 {
1510         struct nlattr *attr = info->attrs[MPTCP_PM_ATTR_ADDR];
1511         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1512         struct mptcp_pm_addr_entry addr, *entry;
1513         struct sk_buff *msg;
1514         void *reply;
1515         int ret;
1516
1517         ret = mptcp_pm_parse_addr(attr, info, false, &addr);
1518         if (ret < 0)
1519                 return ret;
1520
1521         msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
1522         if (!msg)
1523                 return -ENOMEM;
1524
1525         reply = genlmsg_put_reply(msg, info, &mptcp_genl_family, 0,
1526                                   info->genlhdr->cmd);
1527         if (!reply) {
1528                 GENL_SET_ERR_MSG(info, "not enough space in Netlink message");
1529                 ret = -EMSGSIZE;
1530                 goto fail;
1531         }
1532
1533         spin_lock_bh(&pernet->lock);
1534         entry = __lookup_addr_by_id(pernet, addr.addr.id);
1535         if (!entry) {
1536                 GENL_SET_ERR_MSG(info, "address not found");
1537                 ret = -EINVAL;
1538                 goto unlock_fail;
1539         }
1540
1541         ret = mptcp_nl_fill_addr(msg, entry);
1542         if (ret)
1543                 goto unlock_fail;
1544
1545         genlmsg_end(msg, reply);
1546         ret = genlmsg_reply(msg, info);
1547         spin_unlock_bh(&pernet->lock);
1548         return ret;
1549
1550 unlock_fail:
1551         spin_unlock_bh(&pernet->lock);
1552
1553 fail:
1554         nlmsg_free(msg);
1555         return ret;
1556 }
1557
1558 static int mptcp_nl_cmd_dump_addrs(struct sk_buff *msg,
1559                                    struct netlink_callback *cb)
1560 {
1561         struct net *net = sock_net(msg->sk);
1562         struct mptcp_pm_addr_entry *entry;
1563         struct pm_nl_pernet *pernet;
1564         int id = cb->args[0];
1565         void *hdr;
1566         int i;
1567
1568         pernet = net_generic(net, pm_nl_pernet_id);
1569
1570         spin_lock_bh(&pernet->lock);
1571         for (i = id; i < MAX_ADDR_ID + 1; i++) {
1572                 if (test_bit(i, pernet->id_bitmap)) {
1573                         entry = __lookup_addr_by_id(pernet, i);
1574                         if (!entry)
1575                                 break;
1576
1577                         if (entry->addr.id <= id)
1578                                 continue;
1579
1580                         hdr = genlmsg_put(msg, NETLINK_CB(cb->skb).portid,
1581                                           cb->nlh->nlmsg_seq, &mptcp_genl_family,
1582                                           NLM_F_MULTI, MPTCP_PM_CMD_GET_ADDR);
1583                         if (!hdr)
1584                                 break;
1585
1586                         if (mptcp_nl_fill_addr(msg, entry) < 0) {
1587                                 genlmsg_cancel(msg, hdr);
1588                                 break;
1589                         }
1590
1591                         id = entry->addr.id;
1592                         genlmsg_end(msg, hdr);
1593                 }
1594         }
1595         spin_unlock_bh(&pernet->lock);
1596
1597         cb->args[0] = id;
1598         return msg->len;
1599 }
1600
1601 static int parse_limit(struct genl_info *info, int id, unsigned int *limit)
1602 {
1603         struct nlattr *attr = info->attrs[id];
1604
1605         if (!attr)
1606                 return 0;
1607
1608         *limit = nla_get_u32(attr);
1609         if (*limit > MPTCP_PM_ADDR_MAX) {
1610                 GENL_SET_ERR_MSG(info, "limit greater than maximum");
1611                 return -EINVAL;
1612         }
1613         return 0;
1614 }
1615
1616 static int
1617 mptcp_nl_cmd_set_limits(struct sk_buff *skb, struct genl_info *info)
1618 {
1619         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1620         unsigned int rcv_addrs, subflows;
1621         int ret;
1622
1623         spin_lock_bh(&pernet->lock);
1624         rcv_addrs = pernet->add_addr_accept_max;
1625         ret = parse_limit(info, MPTCP_PM_ATTR_RCV_ADD_ADDRS, &rcv_addrs);
1626         if (ret)
1627                 goto unlock;
1628
1629         subflows = pernet->subflows_max;
1630         ret = parse_limit(info, MPTCP_PM_ATTR_SUBFLOWS, &subflows);
1631         if (ret)
1632                 goto unlock;
1633
1634         WRITE_ONCE(pernet->add_addr_accept_max, rcv_addrs);
1635         WRITE_ONCE(pernet->subflows_max, subflows);
1636
1637 unlock:
1638         spin_unlock_bh(&pernet->lock);
1639         return ret;
1640 }
1641
1642 static int
1643 mptcp_nl_cmd_get_limits(struct sk_buff *skb, struct genl_info *info)
1644 {
1645         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1646         struct sk_buff *msg;
1647         void *reply;
1648
1649         msg = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
1650         if (!msg)
1651                 return -ENOMEM;
1652
1653         reply = genlmsg_put_reply(msg, info, &mptcp_genl_family, 0,
1654                                   MPTCP_PM_CMD_GET_LIMITS);
1655         if (!reply)
1656                 goto fail;
1657
1658         if (nla_put_u32(msg, MPTCP_PM_ATTR_RCV_ADD_ADDRS,
1659                         READ_ONCE(pernet->add_addr_accept_max)))
1660                 goto fail;
1661
1662         if (nla_put_u32(msg, MPTCP_PM_ATTR_SUBFLOWS,
1663                         READ_ONCE(pernet->subflows_max)))
1664                 goto fail;
1665
1666         genlmsg_end(msg, reply);
1667         return genlmsg_reply(msg, info);
1668
1669 fail:
1670         GENL_SET_ERR_MSG(info, "not enough space in Netlink message");
1671         nlmsg_free(msg);
1672         return -EMSGSIZE;
1673 }
1674
1675 static int mptcp_nl_addr_backup(struct net *net,
1676                                 struct mptcp_addr_info *addr,
1677                                 u8 bkup)
1678 {
1679         long s_slot = 0, s_num = 0;
1680         struct mptcp_sock *msk;
1681         int ret = -EINVAL;
1682
1683         while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
1684                 struct sock *sk = (struct sock *)msk;
1685
1686                 if (list_empty(&msk->conn_list))
1687                         goto next;
1688
1689                 lock_sock(sk);
1690                 spin_lock_bh(&msk->pm.lock);
1691                 ret = mptcp_pm_nl_mp_prio_send_ack(msk, addr, bkup);
1692                 spin_unlock_bh(&msk->pm.lock);
1693                 release_sock(sk);
1694
1695 next:
1696                 sock_put(sk);
1697                 cond_resched();
1698         }
1699
1700         return ret;
1701 }
1702
1703 static int mptcp_nl_cmd_set_flags(struct sk_buff *skb, struct genl_info *info)
1704 {
1705         struct nlattr *attr = info->attrs[MPTCP_PM_ATTR_ADDR];
1706         struct pm_nl_pernet *pernet = genl_info_pm_nl(info);
1707         struct mptcp_pm_addr_entry addr, *entry;
1708         struct net *net = sock_net(skb->sk);
1709         u8 bkup = 0;
1710         int ret;
1711
1712         ret = mptcp_pm_parse_addr(attr, info, true, &addr);
1713         if (ret < 0)
1714                 return ret;
1715
1716         if (addr.flags & MPTCP_PM_ADDR_FLAG_BACKUP)
1717                 bkup = 1;
1718
1719         list_for_each_entry(entry, &pernet->local_addr_list, list) {
1720                 if (addresses_equal(&entry->addr, &addr.addr, true)) {
1721                         ret = mptcp_nl_addr_backup(net, &entry->addr, bkup);
1722                         if (ret)
1723                                 return ret;
1724
1725                         if (bkup)
1726                                 entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
1727                         else
1728                                 entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
1729                 }
1730         }
1731
1732         return 0;
1733 }
1734
1735 static void mptcp_nl_mcast_send(struct net *net, struct sk_buff *nlskb, gfp_t gfp)
1736 {
1737         genlmsg_multicast_netns(&mptcp_genl_family, net,
1738                                 nlskb, 0, MPTCP_PM_EV_GRP_OFFSET, gfp);
1739 }
1740
1741 static int mptcp_event_add_subflow(struct sk_buff *skb, const struct sock *ssk)
1742 {
1743         const struct inet_sock *issk = inet_sk(ssk);
1744         const struct mptcp_subflow_context *sf;
1745
1746         if (nla_put_u16(skb, MPTCP_ATTR_FAMILY, ssk->sk_family))
1747                 return -EMSGSIZE;
1748
1749         switch (ssk->sk_family) {
1750         case AF_INET:
1751                 if (nla_put_in_addr(skb, MPTCP_ATTR_SADDR4, issk->inet_saddr))
1752                         return -EMSGSIZE;
1753                 if (nla_put_in_addr(skb, MPTCP_ATTR_DADDR4, issk->inet_daddr))
1754                         return -EMSGSIZE;
1755                 break;
1756 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
1757         case AF_INET6: {
1758                 const struct ipv6_pinfo *np = inet6_sk(ssk);
1759
1760                 if (nla_put_in6_addr(skb, MPTCP_ATTR_SADDR6, &np->saddr))
1761                         return -EMSGSIZE;
1762                 if (nla_put_in6_addr(skb, MPTCP_ATTR_DADDR6, &ssk->sk_v6_daddr))
1763                         return -EMSGSIZE;
1764                 break;
1765         }
1766 #endif
1767         default:
1768                 WARN_ON_ONCE(1);
1769                 return -EMSGSIZE;
1770         }
1771
1772         if (nla_put_be16(skb, MPTCP_ATTR_SPORT, issk->inet_sport))
1773                 return -EMSGSIZE;
1774         if (nla_put_be16(skb, MPTCP_ATTR_DPORT, issk->inet_dport))
1775                 return -EMSGSIZE;
1776
1777         sf = mptcp_subflow_ctx(ssk);
1778         if (WARN_ON_ONCE(!sf))
1779                 return -EINVAL;
1780
1781         if (nla_put_u8(skb, MPTCP_ATTR_LOC_ID, sf->local_id))
1782                 return -EMSGSIZE;
1783
1784         if (nla_put_u8(skb, MPTCP_ATTR_REM_ID, sf->remote_id))
1785                 return -EMSGSIZE;
1786
1787         return 0;
1788 }
1789
1790 static int mptcp_event_put_token_and_ssk(struct sk_buff *skb,
1791                                          const struct mptcp_sock *msk,
1792                                          const struct sock *ssk)
1793 {
1794         const struct sock *sk = (const struct sock *)msk;
1795         const struct mptcp_subflow_context *sf;
1796         u8 sk_err;
1797
1798         if (nla_put_u32(skb, MPTCP_ATTR_TOKEN, msk->token))
1799                 return -EMSGSIZE;
1800
1801         if (mptcp_event_add_subflow(skb, ssk))
1802                 return -EMSGSIZE;
1803
1804         sf = mptcp_subflow_ctx(ssk);
1805         if (WARN_ON_ONCE(!sf))
1806                 return -EINVAL;
1807
1808         if (nla_put_u8(skb, MPTCP_ATTR_BACKUP, sf->backup))
1809                 return -EMSGSIZE;
1810
1811         if (ssk->sk_bound_dev_if &&
1812             nla_put_s32(skb, MPTCP_ATTR_IF_IDX, ssk->sk_bound_dev_if))
1813                 return -EMSGSIZE;
1814
1815         sk_err = ssk->sk_err;
1816         if (sk_err && sk->sk_state == TCP_ESTABLISHED &&
1817             nla_put_u8(skb, MPTCP_ATTR_ERROR, sk_err))
1818                 return -EMSGSIZE;
1819
1820         return 0;
1821 }
1822
1823 static int mptcp_event_sub_established(struct sk_buff *skb,
1824                                        const struct mptcp_sock *msk,
1825                                        const struct sock *ssk)
1826 {
1827         return mptcp_event_put_token_and_ssk(skb, msk, ssk);
1828 }
1829
1830 static int mptcp_event_sub_closed(struct sk_buff *skb,
1831                                   const struct mptcp_sock *msk,
1832                                   const struct sock *ssk)
1833 {
1834         const struct mptcp_subflow_context *sf;
1835
1836         if (mptcp_event_put_token_and_ssk(skb, msk, ssk))
1837                 return -EMSGSIZE;
1838
1839         sf = mptcp_subflow_ctx(ssk);
1840         if (!sf->reset_seen)
1841                 return 0;
1842
1843         if (nla_put_u32(skb, MPTCP_ATTR_RESET_REASON, sf->reset_reason))
1844                 return -EMSGSIZE;
1845
1846         if (nla_put_u32(skb, MPTCP_ATTR_RESET_FLAGS, sf->reset_transient))
1847                 return -EMSGSIZE;
1848
1849         return 0;
1850 }
1851
1852 static int mptcp_event_created(struct sk_buff *skb,
1853                                const struct mptcp_sock *msk,
1854                                const struct sock *ssk)
1855 {
1856         int err = nla_put_u32(skb, MPTCP_ATTR_TOKEN, msk->token);
1857
1858         if (err)
1859                 return err;
1860
1861         return mptcp_event_add_subflow(skb, ssk);
1862 }
1863
1864 void mptcp_event_addr_removed(const struct mptcp_sock *msk, uint8_t id)
1865 {
1866         struct net *net = sock_net((const struct sock *)msk);
1867         struct nlmsghdr *nlh;
1868         struct sk_buff *skb;
1869
1870         if (!genl_has_listeners(&mptcp_genl_family, net, MPTCP_PM_EV_GRP_OFFSET))
1871                 return;
1872
1873         skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_ATOMIC);
1874         if (!skb)
1875                 return;
1876
1877         nlh = genlmsg_put(skb, 0, 0, &mptcp_genl_family, 0, MPTCP_EVENT_REMOVED);
1878         if (!nlh)
1879                 goto nla_put_failure;
1880
1881         if (nla_put_u32(skb, MPTCP_ATTR_TOKEN, msk->token))
1882                 goto nla_put_failure;
1883
1884         if (nla_put_u8(skb, MPTCP_ATTR_REM_ID, id))
1885                 goto nla_put_failure;
1886
1887         genlmsg_end(skb, nlh);
1888         mptcp_nl_mcast_send(net, skb, GFP_ATOMIC);
1889         return;
1890
1891 nla_put_failure:
1892         kfree_skb(skb);
1893 }
1894
1895 void mptcp_event_addr_announced(const struct mptcp_sock *msk,
1896                                 const struct mptcp_addr_info *info)
1897 {
1898         struct net *net = sock_net((const struct sock *)msk);
1899         struct nlmsghdr *nlh;
1900         struct sk_buff *skb;
1901
1902         if (!genl_has_listeners(&mptcp_genl_family, net, MPTCP_PM_EV_GRP_OFFSET))
1903                 return;
1904
1905         skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_ATOMIC);
1906         if (!skb)
1907                 return;
1908
1909         nlh = genlmsg_put(skb, 0, 0, &mptcp_genl_family, 0,
1910                           MPTCP_EVENT_ANNOUNCED);
1911         if (!nlh)
1912                 goto nla_put_failure;
1913
1914         if (nla_put_u32(skb, MPTCP_ATTR_TOKEN, msk->token))
1915                 goto nla_put_failure;
1916
1917         if (nla_put_u8(skb, MPTCP_ATTR_REM_ID, info->id))
1918                 goto nla_put_failure;
1919
1920         if (nla_put_be16(skb, MPTCP_ATTR_DPORT, info->port))
1921                 goto nla_put_failure;
1922
1923         switch (info->family) {
1924         case AF_INET:
1925                 if (nla_put_in_addr(skb, MPTCP_ATTR_DADDR4, info->addr.s_addr))
1926                         goto nla_put_failure;
1927                 break;
1928 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
1929         case AF_INET6:
1930                 if (nla_put_in6_addr(skb, MPTCP_ATTR_DADDR6, &info->addr6))
1931                         goto nla_put_failure;
1932                 break;
1933 #endif
1934         default:
1935                 WARN_ON_ONCE(1);
1936                 goto nla_put_failure;
1937         }
1938
1939         genlmsg_end(skb, nlh);
1940         mptcp_nl_mcast_send(net, skb, GFP_ATOMIC);
1941         return;
1942
1943 nla_put_failure:
1944         kfree_skb(skb);
1945 }
1946
1947 void mptcp_event(enum mptcp_event_type type, const struct mptcp_sock *msk,
1948                  const struct sock *ssk, gfp_t gfp)
1949 {
1950         struct net *net = sock_net((const struct sock *)msk);
1951         struct nlmsghdr *nlh;
1952         struct sk_buff *skb;
1953
1954         if (!genl_has_listeners(&mptcp_genl_family, net, MPTCP_PM_EV_GRP_OFFSET))
1955                 return;
1956
1957         skb = nlmsg_new(NLMSG_DEFAULT_SIZE, gfp);
1958         if (!skb)
1959                 return;
1960
1961         nlh = genlmsg_put(skb, 0, 0, &mptcp_genl_family, 0, type);
1962         if (!nlh)
1963                 goto nla_put_failure;
1964
1965         switch (type) {
1966         case MPTCP_EVENT_UNSPEC:
1967                 WARN_ON_ONCE(1);
1968                 break;
1969         case MPTCP_EVENT_CREATED:
1970         case MPTCP_EVENT_ESTABLISHED:
1971                 if (mptcp_event_created(skb, msk, ssk) < 0)
1972                         goto nla_put_failure;
1973                 break;
1974         case MPTCP_EVENT_CLOSED:
1975                 if (nla_put_u32(skb, MPTCP_ATTR_TOKEN, msk->token) < 0)
1976                         goto nla_put_failure;
1977                 break;
1978         case MPTCP_EVENT_ANNOUNCED:
1979         case MPTCP_EVENT_REMOVED:
1980                 /* call mptcp_event_addr_announced()/removed instead */
1981                 WARN_ON_ONCE(1);
1982                 break;
1983         case MPTCP_EVENT_SUB_ESTABLISHED:
1984         case MPTCP_EVENT_SUB_PRIORITY:
1985                 if (mptcp_event_sub_established(skb, msk, ssk) < 0)
1986                         goto nla_put_failure;
1987                 break;
1988         case MPTCP_EVENT_SUB_CLOSED:
1989                 if (mptcp_event_sub_closed(skb, msk, ssk) < 0)
1990                         goto nla_put_failure;
1991                 break;
1992         }
1993
1994         genlmsg_end(skb, nlh);
1995         mptcp_nl_mcast_send(net, skb, gfp);
1996         return;
1997
1998 nla_put_failure:
1999         kfree_skb(skb);
2000 }
2001
2002 static const struct genl_small_ops mptcp_pm_ops[] = {
2003         {
2004                 .cmd    = MPTCP_PM_CMD_ADD_ADDR,
2005                 .doit   = mptcp_nl_cmd_add_addr,
2006                 .flags  = GENL_ADMIN_PERM,
2007         },
2008         {
2009                 .cmd    = MPTCP_PM_CMD_DEL_ADDR,
2010                 .doit   = mptcp_nl_cmd_del_addr,
2011                 .flags  = GENL_ADMIN_PERM,
2012         },
2013         {
2014                 .cmd    = MPTCP_PM_CMD_FLUSH_ADDRS,
2015                 .doit   = mptcp_nl_cmd_flush_addrs,
2016                 .flags  = GENL_ADMIN_PERM,
2017         },
2018         {
2019                 .cmd    = MPTCP_PM_CMD_GET_ADDR,
2020                 .doit   = mptcp_nl_cmd_get_addr,
2021                 .dumpit   = mptcp_nl_cmd_dump_addrs,
2022         },
2023         {
2024                 .cmd    = MPTCP_PM_CMD_SET_LIMITS,
2025                 .doit   = mptcp_nl_cmd_set_limits,
2026                 .flags  = GENL_ADMIN_PERM,
2027         },
2028         {
2029                 .cmd    = MPTCP_PM_CMD_GET_LIMITS,
2030                 .doit   = mptcp_nl_cmd_get_limits,
2031         },
2032         {
2033                 .cmd    = MPTCP_PM_CMD_SET_FLAGS,
2034                 .doit   = mptcp_nl_cmd_set_flags,
2035                 .flags  = GENL_ADMIN_PERM,
2036         },
2037 };
2038
2039 static struct genl_family mptcp_genl_family __ro_after_init = {
2040         .name           = MPTCP_PM_NAME,
2041         .version        = MPTCP_PM_VER,
2042         .maxattr        = MPTCP_PM_ATTR_MAX,
2043         .policy         = mptcp_pm_policy,
2044         .netnsok        = true,
2045         .module         = THIS_MODULE,
2046         .small_ops      = mptcp_pm_ops,
2047         .n_small_ops    = ARRAY_SIZE(mptcp_pm_ops),
2048         .mcgrps         = mptcp_pm_mcgrps,
2049         .n_mcgrps       = ARRAY_SIZE(mptcp_pm_mcgrps),
2050 };
2051
2052 static int __net_init pm_nl_init_net(struct net *net)
2053 {
2054         struct pm_nl_pernet *pernet = net_generic(net, pm_nl_pernet_id);
2055
2056         INIT_LIST_HEAD_RCU(&pernet->local_addr_list);
2057         pernet->next_id = 1;
2058         pernet->stale_loss_cnt = 4;
2059         spin_lock_init(&pernet->lock);
2060
2061         /* No need to initialize other pernet fields, the struct is zeroed at
2062          * allocation time.
2063          */
2064
2065         return 0;
2066 }
2067
2068 static void __net_exit pm_nl_exit_net(struct list_head *net_list)
2069 {
2070         struct net *net;
2071
2072         list_for_each_entry(net, net_list, exit_list) {
2073                 struct pm_nl_pernet *pernet = net_generic(net, pm_nl_pernet_id);
2074
2075                 /* net is removed from namespace list, can't race with
2076                  * other modifiers, also netns core already waited for a
2077                  * RCU grace period.
2078                  */
2079                 __flush_addrs(&pernet->local_addr_list);
2080         }
2081 }
2082
2083 static struct pernet_operations mptcp_pm_pernet_ops = {
2084         .init = pm_nl_init_net,
2085         .exit_batch = pm_nl_exit_net,
2086         .id = &pm_nl_pernet_id,
2087         .size = sizeof(struct pm_nl_pernet),
2088 };
2089
2090 void __init mptcp_pm_nl_init(void)
2091 {
2092         if (register_pernet_subsys(&mptcp_pm_pernet_ops) < 0)
2093                 panic("Failed to register MPTCP PM pernet subsystem.\n");
2094
2095         if (genl_register_family(&mptcp_genl_family))
2096                 panic("Failed to register MPTCP PM netlink family\n");
2097 }