net/core/sock_map.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 /* Copyright (c) 2017 - 2018 Covalent IO, Inc. http://covalent.io */
   3
   4 #include <linux/bpf.h>
   5 #include <linux/filter.h>
   6 #include <linux/errno.h>
   7 #include <linux/file.h>
   8 #include <linux/net.h>
   9 #include <linux/workqueue.h>
  10 #include <linux/skmsg.h>
  11 #include <linux/list.h>
  12 #include <linux/jhash.h>
  13
  14 struct bpf_stab {
  15         struct bpf_map map;
  16         struct sock **sks;
  17         struct sk_psock_progs progs;
  18         raw_spinlock_t lock;
  19 };
  20
  21 #define SOCK_CREATE_FLAG_MASK                           \
  22         (BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY)
  23
  24 static struct bpf_map *sock_map_alloc(union bpf_attr *attr)
  25 {
  26         struct bpf_stab *stab;
  27         u64 cost;
  28         int err;
  29
  30         if (!capable(CAP_NET_ADMIN))
  31                 return ERR_PTR(-EPERM);
  32         if (attr->max_entries == 0 ||
  33             attr->key_size    != 4 ||
  34             attr->value_size  != 4 ||
  35             attr->map_flags & ~SOCK_CREATE_FLAG_MASK)
  36                 return ERR_PTR(-EINVAL);
  37
  38         stab = kzalloc(sizeof(*stab), GFP_USER);
  39         if (!stab)
  40                 return ERR_PTR(-ENOMEM);
  41
  42         bpf_map_init_from_attr(&stab->map, attr);
  43         raw_spin_lock_init(&stab->lock);
  44
  45         /* Make sure page count doesn't overflow. */
  46         cost = (u64) stab->map.max_entries * sizeof(struct sock *);
  47         err = bpf_map_charge_init(&stab->map.memory, cost);
  48         if (err)
  49                 goto free_stab;
  50
  51         stab->sks = bpf_map_area_alloc(stab->map.max_entries *
  52                                        sizeof(struct sock *),
  53                                        stab->map.numa_node);
  54         if (stab->sks)
  55                 return &stab->map;
  56         err = -ENOMEM;
  57         bpf_map_charge_finish(&stab->map.memory);
  58 free_stab:
  59         kfree(stab);
  60         return ERR_PTR(err);
  61 }
  62
  63 int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog *prog)
  64 {
  65         u32 ufd = attr->target_fd;
  66         struct bpf_map *map;
  67         struct fd f;
  68         int ret;
  69
  70         f = fdget(ufd);
  71         map = __bpf_map_get(f);
  72         if (IS_ERR(map))
  73                 return PTR_ERR(map);
  74         ret = sock_map_prog_update(map, prog, attr->attach_type);
  75         fdput(f);
  76         return ret;
  77 }
  78
  79 static void sock_map_sk_acquire(struct sock *sk)
  80         __acquires(&sk->sk_lock.slock)
  81 {
  82         lock_sock(sk);
  83         preempt_disable();
  84         rcu_read_lock();
  85 }
  86
  87 static void sock_map_sk_release(struct sock *sk)
  88         __releases(&sk->sk_lock.slock)
  89 {
  90         rcu_read_unlock();
  91         preempt_enable();
  92         release_sock(sk);
  93 }
  94
  95 static void sock_map_add_link(struct sk_psock *psock,
  96                               struct sk_psock_link *link,
  97                               struct bpf_map *map, void *link_raw)
  98 {
  99         link->link_raw = link_raw;
 100         link->map = map;
 101         spin_lock_bh(&psock->link_lock);
 102         list_add_tail(&link->list, &psock->link);
 103         spin_unlock_bh(&psock->link_lock);
 104 }
 105
 106 static void sock_map_del_link(struct sock *sk,
 107                               struct sk_psock *psock, void *link_raw)
 108 {
 109         struct sk_psock_link *link, *tmp;
 110         bool strp_stop = false;
 111
 112         spin_lock_bh(&psock->link_lock);
 113         list_for_each_entry_safe(link, tmp, &psock->link, list) {
 114                 if (link->link_raw == link_raw) {
 115                         struct bpf_map *map = link->map;
 116                         struct bpf_stab *stab = container_of(map, struct bpf_stab,
 117                                                              map);
 118                         if (psock->parser.enabled && stab->progs.skb_parser)
 119                                 strp_stop = true;
 120                         list_del(&link->list);
 121                         sk_psock_free_link(link);
 122                 }
 123         }
 124         spin_unlock_bh(&psock->link_lock);
 125         if (strp_stop) {
 126                 write_lock_bh(&sk->sk_callback_lock);
 127                 sk_psock_stop_strp(sk, psock);
 128                 write_unlock_bh(&sk->sk_callback_lock);
 129         }
 130 }
 131
 132 static void sock_map_unref(struct sock *sk, void *link_raw)
 133 {
 134         struct sk_psock *psock = sk_psock(sk);
 135
 136         if (likely(psock)) {
 137                 sock_map_del_link(sk, psock, link_raw);
 138                 sk_psock_put(sk, psock);
 139         }
 140 }
 141
 142 static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
 143                          struct sock *sk)
 144 {
 145         struct bpf_prog *msg_parser, *skb_parser, *skb_verdict;
 146         bool skb_progs, sk_psock_is_new = false;
 147         struct sk_psock *psock;
 148         int ret;
 149
 150         skb_verdict = READ_ONCE(progs->skb_verdict);
 151         skb_parser = READ_ONCE(progs->skb_parser);
 152         skb_progs = skb_parser && skb_verdict;
 153         if (skb_progs) {
 154                 skb_verdict = bpf_prog_inc_not_zero(skb_verdict);
 155                 if (IS_ERR(skb_verdict))
 156                         return PTR_ERR(skb_verdict);
 157                 skb_parser = bpf_prog_inc_not_zero(skb_parser);
 158                 if (IS_ERR(skb_parser)) {
 159                         bpf_prog_put(skb_verdict);
 160                         return PTR_ERR(skb_parser);
 161                 }
 162         }
 163
 164         msg_parser = READ_ONCE(progs->msg_parser);
 165         if (msg_parser) {
 166                 msg_parser = bpf_prog_inc_not_zero(msg_parser);
 167                 if (IS_ERR(msg_parser)) {
 168                         ret = PTR_ERR(msg_parser);
 169                         goto out;
 170                 }
 171         }
 172
 173         psock = sk_psock_get_checked(sk);
 174         if (IS_ERR(psock)) {
 175                 ret = PTR_ERR(psock);
 176                 goto out_progs;
 177         }
 178
 179         if (psock) {
 180                 if ((msg_parser && READ_ONCE(psock->progs.msg_parser)) ||
 181                     (skb_progs  && READ_ONCE(psock->progs.skb_parser))) {
 182                         sk_psock_put(sk, psock);
 183                         ret = -EBUSY;
 184                         goto out_progs;
 185                 }
 186         } else {
 187                 psock = sk_psock_init(sk, map->numa_node);
 188                 if (!psock) {
 189                         ret = -ENOMEM;
 190                         goto out_progs;
 191                 }
 192                 sk_psock_is_new = true;
 193         }
 194
 195         if (msg_parser)
 196                 psock_set_prog(&psock->progs.msg_parser, msg_parser);
 197         if (sk_psock_is_new) {
 198                 ret = tcp_bpf_init(sk);
 199                 if (ret < 0)
 200                         goto out_drop;
 201         } else {
 202                 tcp_bpf_reinit(sk);
 203         }
 204
 205         write_lock_bh(&sk->sk_callback_lock);
 206         if (skb_progs && !psock->parser.enabled) {
 207                 ret = sk_psock_init_strp(sk, psock);
 208                 if (ret) {
 209                         write_unlock_bh(&sk->sk_callback_lock);
 210                         goto out_drop;
 211                 }
 212                 psock_set_prog(&psock->progs.skb_verdict, skb_verdict);
 213                 psock_set_prog(&psock->progs.skb_parser, skb_parser);
 214                 sk_psock_start_strp(sk, psock);
 215         }
 216         write_unlock_bh(&sk->sk_callback_lock);
 217         return 0;
 218 out_drop:
 219         sk_psock_put(sk, psock);
 220 out_progs:
 221         if (msg_parser)
 222                 bpf_prog_put(msg_parser);
 223 out:
 224         if (skb_progs) {
 225                 bpf_prog_put(skb_verdict);
 226                 bpf_prog_put(skb_parser);
 227         }
 228         return ret;
 229 }
 230
 231 static void sock_map_free(struct bpf_map *map)
 232 {
 233         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 234         int i;
 235
 236         synchronize_rcu();
 237         rcu_read_lock();
 238         raw_spin_lock_bh(&stab->lock);
 239         for (i = 0; i < stab->map.max_entries; i++) {
 240                 struct sock **psk = &stab->sks[i];
 241                 struct sock *sk;
 242
 243                 sk = xchg(psk, NULL);
 244                 if (sk) {
 245                         lock_sock(sk);
 246                         sock_map_unref(sk, psk);
 247                         release_sock(sk);
 248                 }
 249         }
 250         raw_spin_unlock_bh(&stab->lock);
 251         rcu_read_unlock();
 252
 253         synchronize_rcu();
 254
 255         bpf_map_area_free(stab->sks);
 256         kfree(stab);
 257 }
 258
 259 static void sock_map_release_progs(struct bpf_map *map)
 260 {
 261         psock_progs_drop(&container_of(map, struct bpf_stab, map)->progs);
 262 }
 263
 264 static struct sock *__sock_map_lookup_elem(struct bpf_map *map, u32 key)
 265 {
 266         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 267
 268         WARN_ON_ONCE(!rcu_read_lock_held());
 269
 270         if (unlikely(key >= map->max_entries))
 271                 return NULL;
 272         return READ_ONCE(stab->sks[key]);
 273 }
 274
 275 static void *sock_map_lookup(struct bpf_map *map, void *key)
 276 {
 277         return ERR_PTR(-EOPNOTSUPP);
 278 }
 279
 280 static int __sock_map_delete(struct bpf_stab *stab, struct sock *sk_test,
 281                              struct sock **psk)
 282 {
 283         struct sock *sk;
 284         int err = 0;
 285
 286         raw_spin_lock_bh(&stab->lock);
 287         sk = *psk;
 288         if (!sk_test || sk_test == sk)
 289                 sk = xchg(psk, NULL);
 290
 291         if (likely(sk))
 292                 sock_map_unref(sk, psk);
 293         else
 294                 err = -EINVAL;
 295
 296         raw_spin_unlock_bh(&stab->lock);
 297         return err;
 298 }
 299
 300 static void sock_map_delete_from_link(struct bpf_map *map, struct sock *sk,
 301                                       void *link_raw)
 302 {
 303         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 304
 305         __sock_map_delete(stab, sk, link_raw);
 306 }
 307
 308 static int sock_map_delete_elem(struct bpf_map *map, void *key)
 309 {
 310         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 311         u32 i = *(u32 *)key;
 312         struct sock **psk;
 313
 314         if (unlikely(i >= map->max_entries))
 315                 return -EINVAL;
 316
 317         psk = &stab->sks[i];
 318         return __sock_map_delete(stab, NULL, psk);
 319 }
 320
 321 static int sock_map_get_next_key(struct bpf_map *map, void *key, void *next)
 322 {
 323         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 324         u32 i = key ? *(u32 *)key : U32_MAX;
 325         u32 *key_next = next;
 326
 327         if (i == stab->map.max_entries - 1)
 328                 return -ENOENT;
 329         if (i >= stab->map.max_entries)
 330                 *key_next = 0;
 331         else
 332                 *key_next = i + 1;
 333         return 0;
 334 }
 335
 336 static int sock_map_update_common(struct bpf_map *map, u32 idx,
 337                                   struct sock *sk, u64 flags)
 338 {
 339         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 340         struct inet_connection_sock *icsk = inet_csk(sk);
 341         struct sk_psock_link *link;
 342         struct sk_psock *psock;
 343         struct sock *osk;
 344         int ret;
 345
 346         WARN_ON_ONCE(!rcu_read_lock_held());
 347         if (unlikely(flags > BPF_EXIST))
 348                 return -EINVAL;
 349         if (unlikely(idx >= map->max_entries))
 350                 return -E2BIG;
 351         if (unlikely(rcu_access_pointer(icsk->icsk_ulp_data)))
 352                 return -EINVAL;
 353
 354         link = sk_psock_init_link();
 355         if (!link)
 356                 return -ENOMEM;
 357
 358         ret = sock_map_link(map, &stab->progs, sk);
 359         if (ret < 0)
 360                 goto out_free;
 361
 362         psock = sk_psock(sk);
 363         WARN_ON_ONCE(!psock);
 364
 365         raw_spin_lock_bh(&stab->lock);
 366         osk = stab->sks[idx];
 367         if (osk && flags == BPF_NOEXIST) {
 368                 ret = -EEXIST;
 369                 goto out_unlock;
 370         } else if (!osk && flags == BPF_EXIST) {
 371                 ret = -ENOENT;
 372                 goto out_unlock;
 373         }
 374
 375         sock_map_add_link(psock, link, map, &stab->sks[idx]);
 376         stab->sks[idx] = sk;
 377         if (osk)
 378                 sock_map_unref(osk, &stab->sks[idx]);
 379         raw_spin_unlock_bh(&stab->lock);
 380         return 0;
 381 out_unlock:
 382         raw_spin_unlock_bh(&stab->lock);
 383         if (psock)
 384                 sk_psock_put(sk, psock);
 385 out_free:
 386         sk_psock_free_link(link);
 387         return ret;
 388 }
 389
 390 static bool sock_map_op_okay(const struct bpf_sock_ops_kern *ops)
 391 {
 392         return ops->op == BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB ||
 393                ops->op == BPF_SOCK_OPS_ACTIVE_ESTABLISHED_CB;
 394 }
 395
 396 static bool sock_map_sk_is_suitable(const struct sock *sk)
 397 {
 398         return sk->sk_type == SOCK_STREAM &&
 399                sk->sk_protocol == IPPROTO_TCP;
 400 }
 401
 402 static int sock_map_update_elem(struct bpf_map *map, void *key,
 403                                 void *value, u64 flags)
 404 {
 405         u32 ufd = *(u32 *)value;
 406         u32 idx = *(u32 *)key;
 407         struct socket *sock;
 408         struct sock *sk;
 409         int ret;
 410
 411         sock = sockfd_lookup(ufd, &ret);
 412         if (!sock)
 413                 return ret;
 414         sk = sock->sk;
 415         if (!sk) {
 416                 ret = -EINVAL;
 417                 goto out;
 418         }
 419         if (!sock_map_sk_is_suitable(sk) ||
 420             sk->sk_state != TCP_ESTABLISHED) {
 421                 ret = -EOPNOTSUPP;
 422                 goto out;
 423         }
 424
 425         sock_map_sk_acquire(sk);
 426         ret = sock_map_update_common(map, idx, sk, flags);
 427         sock_map_sk_release(sk);
 428 out:
 429         fput(sock->file);
 430         return ret;
 431 }
 432
 433 BPF_CALL_4(bpf_sock_map_update, struct bpf_sock_ops_kern *, sops,
 434            struct bpf_map *, map, void *, key, u64, flags)
 435 {
 436         WARN_ON_ONCE(!rcu_read_lock_held());
 437
 438         if (likely(sock_map_sk_is_suitable(sops->sk) &&
 439                    sock_map_op_okay(sops)))
 440                 return sock_map_update_common(map, *(u32 *)key, sops->sk,
 441                                               flags);
 442         return -EOPNOTSUPP;
 443 }
 444
 445 const struct bpf_func_proto bpf_sock_map_update_proto = {
 446         .func           = bpf_sock_map_update,
 447         .gpl_only       = false,
 448         .pkt_access     = true,
 449         .ret_type       = RET_INTEGER,
 450         .arg1_type      = ARG_PTR_TO_CTX,
 451         .arg2_type      = ARG_CONST_MAP_PTR,
 452         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 453         .arg4_type      = ARG_ANYTHING,
 454 };
 455
 456 BPF_CALL_4(bpf_sk_redirect_map, struct sk_buff *, skb,
 457            struct bpf_map *, map, u32, key, u64, flags)
 458 {
 459         struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
 460
 461         if (unlikely(flags & ~(BPF_F_INGRESS)))
 462                 return SK_DROP;
 463         tcb->bpf.flags = flags;
 464         tcb->bpf.sk_redir = __sock_map_lookup_elem(map, key);
 465         if (!tcb->bpf.sk_redir)
 466                 return SK_DROP;
 467         return SK_PASS;
 468 }
 469
 470 const struct bpf_func_proto bpf_sk_redirect_map_proto = {
 471         .func           = bpf_sk_redirect_map,
 472         .gpl_only       = false,
 473         .ret_type       = RET_INTEGER,
 474         .arg1_type      = ARG_PTR_TO_CTX,
 475         .arg2_type      = ARG_CONST_MAP_PTR,
 476         .arg3_type      = ARG_ANYTHING,
 477         .arg4_type      = ARG_ANYTHING,
 478 };
 479
 480 BPF_CALL_4(bpf_msg_redirect_map, struct sk_msg *, msg,
 481            struct bpf_map *, map, u32, key, u64, flags)
 482 {
 483         if (unlikely(flags & ~(BPF_F_INGRESS)))
 484                 return SK_DROP;
 485         msg->flags = flags;
 486         msg->sk_redir = __sock_map_lookup_elem(map, key);
 487         if (!msg->sk_redir)
 488                 return SK_DROP;
 489         return SK_PASS;
 490 }
 491
 492 const struct bpf_func_proto bpf_msg_redirect_map_proto = {
 493         .func           = bpf_msg_redirect_map,
 494         .gpl_only       = false,
 495         .ret_type       = RET_INTEGER,
 496         .arg1_type      = ARG_PTR_TO_CTX,
 497         .arg2_type      = ARG_CONST_MAP_PTR,
 498         .arg3_type      = ARG_ANYTHING,
 499         .arg4_type      = ARG_ANYTHING,
 500 };
 501
 502 const struct bpf_map_ops sock_map_ops = {
 503         .map_alloc              = sock_map_alloc,
 504         .map_free               = sock_map_free,
 505         .map_get_next_key       = sock_map_get_next_key,
 506         .map_update_elem        = sock_map_update_elem,
 507         .map_delete_elem        = sock_map_delete_elem,
 508         .map_lookup_elem        = sock_map_lookup,
 509         .map_release_uref       = sock_map_release_progs,
 510         .map_check_btf          = map_check_no_btf,
 511 };
 512
 513 struct bpf_htab_elem {
 514         struct rcu_head rcu;
 515         u32 hash;
 516         struct sock *sk;
 517         struct hlist_node node;
 518         u8 key[0];
 519 };
 520
 521 struct bpf_htab_bucket {
 522         struct hlist_head head;
 523         raw_spinlock_t lock;
 524 };
 525
 526 struct bpf_htab {
 527         struct bpf_map map;
 528         struct bpf_htab_bucket *buckets;
 529         u32 buckets_num;
 530         u32 elem_size;
 531         struct sk_psock_progs progs;
 532         atomic_t count;
 533 };
 534
 535 static inline u32 sock_hash_bucket_hash(const void *key, u32 len)
 536 {
 537         return jhash(key, len, 0);
 538 }
 539
 540 static struct bpf_htab_bucket *sock_hash_select_bucket(struct bpf_htab *htab,
 541                                                        u32 hash)
 542 {
 543         return &htab->buckets[hash & (htab->buckets_num - 1)];
 544 }
 545
 546 static struct bpf_htab_elem *
 547 sock_hash_lookup_elem_raw(struct hlist_head *head, u32 hash, void *key,
 548                           u32 key_size)
 549 {
 550         struct bpf_htab_elem *elem;
 551
 552         hlist_for_each_entry_rcu(elem, head, node) {
 553                 if (elem->hash == hash &&
 554                     !memcmp(&elem->key, key, key_size))
 555                         return elem;
 556         }
 557
 558         return NULL;
 559 }
 560
 561 static struct sock *__sock_hash_lookup_elem(struct bpf_map *map, void *key)
 562 {
 563         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 564         u32 key_size = map->key_size, hash;
 565         struct bpf_htab_bucket *bucket;
 566         struct bpf_htab_elem *elem;
 567
 568         WARN_ON_ONCE(!rcu_read_lock_held());
 569
 570         hash = sock_hash_bucket_hash(key, key_size);
 571         bucket = sock_hash_select_bucket(htab, hash);
 572         elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
 573
 574         return elem ? elem->sk : NULL;
 575 }
 576
 577 static void sock_hash_free_elem(struct bpf_htab *htab,
 578                                 struct bpf_htab_elem *elem)
 579 {
 580         atomic_dec(&htab->count);
 581         kfree_rcu(elem, rcu);
 582 }
 583
 584 static void sock_hash_delete_from_link(struct bpf_map *map, struct sock *sk,
 585                                        void *link_raw)
 586 {
 587         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 588         struct bpf_htab_elem *elem_probe, *elem = link_raw;
 589         struct bpf_htab_bucket *bucket;
 590
 591         WARN_ON_ONCE(!rcu_read_lock_held());
 592         bucket = sock_hash_select_bucket(htab, elem->hash);
 593
 594         /* elem may be deleted in parallel from the map, but access here
 595          * is okay since it's going away only after RCU grace period.
 596          * However, we need to check whether it's still present.
 597          */
 598         raw_spin_lock_bh(&bucket->lock);
 599         elem_probe = sock_hash_lookup_elem_raw(&bucket->head, elem->hash,
 600                                                elem->key, map->key_size);
 601         if (elem_probe && elem_probe == elem) {
 602                 hlist_del_rcu(&elem->node);
 603                 sock_map_unref(elem->sk, elem);
 604                 sock_hash_free_elem(htab, elem);
 605         }
 606         raw_spin_unlock_bh(&bucket->lock);
 607 }
 608
 609 static int sock_hash_delete_elem(struct bpf_map *map, void *key)
 610 {
 611         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 612         u32 hash, key_size = map->key_size;
 613         struct bpf_htab_bucket *bucket;
 614         struct bpf_htab_elem *elem;
 615         int ret = -ENOENT;
 616
 617         hash = sock_hash_bucket_hash(key, key_size);
 618         bucket = sock_hash_select_bucket(htab, hash);
 619
 620         raw_spin_lock_bh(&bucket->lock);
 621         elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
 622         if (elem) {
 623                 hlist_del_rcu(&elem->node);
 624                 sock_map_unref(elem->sk, elem);
 625                 sock_hash_free_elem(htab, elem);
 626                 ret = 0;
 627         }
 628         raw_spin_unlock_bh(&bucket->lock);
 629         return ret;
 630 }
 631
 632 static struct bpf_htab_elem *sock_hash_alloc_elem(struct bpf_htab *htab,
 633                                                   void *key, u32 key_size,
 634                                                   u32 hash, struct sock *sk,
 635                                                   struct bpf_htab_elem *old)
 636 {
 637         struct bpf_htab_elem *new;
 638
 639         if (atomic_inc_return(&htab->count) > htab->map.max_entries) {
 640                 if (!old) {
 641                         atomic_dec(&htab->count);
 642                         return ERR_PTR(-E2BIG);
 643                 }
 644         }
 645
 646         new = kmalloc_node(htab->elem_size, GFP_ATOMIC | __GFP_NOWARN,
 647                            htab->map.numa_node);
 648         if (!new) {
 649                 atomic_dec(&htab->count);
 650                 return ERR_PTR(-ENOMEM);
 651         }
 652         memcpy(new->key, key, key_size);
 653         new->sk = sk;
 654         new->hash = hash;
 655         return new;
 656 }
 657
 658 static int sock_hash_update_common(struct bpf_map *map, void *key,
 659                                    struct sock *sk, u64 flags)
 660 {
 661         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 662         struct inet_connection_sock *icsk = inet_csk(sk);
 663         u32 key_size = map->key_size, hash;
 664         struct bpf_htab_elem *elem, *elem_new;
 665         struct bpf_htab_bucket *bucket;
 666         struct sk_psock_link *link;
 667         struct sk_psock *psock;
 668         int ret;
 669
 670         WARN_ON_ONCE(!rcu_read_lock_held());
 671         if (unlikely(flags > BPF_EXIST))
 672                 return -EINVAL;
 673         if (unlikely(icsk->icsk_ulp_data))
 674                 return -EINVAL;
 675
 676         link = sk_psock_init_link();
 677         if (!link)
 678                 return -ENOMEM;
 679
 680         ret = sock_map_link(map, &htab->progs, sk);
 681         if (ret < 0)
 682                 goto out_free;
 683
 684         psock = sk_psock(sk);
 685         WARN_ON_ONCE(!psock);
 686
 687         hash = sock_hash_bucket_hash(key, key_size);
 688         bucket = sock_hash_select_bucket(htab, hash);
 689
 690         raw_spin_lock_bh(&bucket->lock);
 691         elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
 692         if (elem && flags == BPF_NOEXIST) {
 693                 ret = -EEXIST;
 694                 goto out_unlock;
 695         } else if (!elem && flags == BPF_EXIST) {
 696                 ret = -ENOENT;
 697                 goto out_unlock;
 698         }
 699
 700         elem_new = sock_hash_alloc_elem(htab, key, key_size, hash, sk, elem);
 701         if (IS_ERR(elem_new)) {
 702                 ret = PTR_ERR(elem_new);
 703                 goto out_unlock;
 704         }
 705
 706         sock_map_add_link(psock, link, map, elem_new);
 707         /* Add new element to the head of the list, so that
 708          * concurrent search will find it before old elem.
 709          */
 710         hlist_add_head_rcu(&elem_new->node, &bucket->head);
 711         if (elem) {
 712                 hlist_del_rcu(&elem->node);
 713                 sock_map_unref(elem->sk, elem);
 714                 sock_hash_free_elem(htab, elem);
 715         }
 716         raw_spin_unlock_bh(&bucket->lock);
 717         return 0;
 718 out_unlock:
 719         raw_spin_unlock_bh(&bucket->lock);
 720         sk_psock_put(sk, psock);
 721 out_free:
 722         sk_psock_free_link(link);
 723         return ret;
 724 }
 725
 726 static int sock_hash_update_elem(struct bpf_map *map, void *key,
 727                                  void *value, u64 flags)
 728 {
 729         u32 ufd = *(u32 *)value;
 730         struct socket *sock;
 731         struct sock *sk;
 732         int ret;
 733
 734         sock = sockfd_lookup(ufd, &ret);
 735         if (!sock)
 736                 return ret;
 737         sk = sock->sk;
 738         if (!sk) {
 739                 ret = -EINVAL;
 740                 goto out;
 741         }
 742         if (!sock_map_sk_is_suitable(sk) ||
 743             sk->sk_state != TCP_ESTABLISHED) {
 744                 ret = -EOPNOTSUPP;
 745                 goto out;
 746         }
 747
 748         sock_map_sk_acquire(sk);
 749         ret = sock_hash_update_common(map, key, sk, flags);
 750         sock_map_sk_release(sk);
 751 out:
 752         fput(sock->file);
 753         return ret;
 754 }
 755
 756 static int sock_hash_get_next_key(struct bpf_map *map, void *key,
 757                                   void *key_next)
 758 {
 759         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 760         struct bpf_htab_elem *elem, *elem_next;
 761         u32 hash, key_size = map->key_size;
 762         struct hlist_head *head;
 763         int i = 0;
 764
 765         if (!key)
 766                 goto find_first_elem;
 767         hash = sock_hash_bucket_hash(key, key_size);
 768         head = &sock_hash_select_bucket(htab, hash)->head;
 769         elem = sock_hash_lookup_elem_raw(head, hash, key, key_size);
 770         if (!elem)
 771                 goto find_first_elem;
 772
 773         elem_next = hlist_entry_safe(rcu_dereference_raw(hlist_next_rcu(&elem->node)),
 774                                      struct bpf_htab_elem, node);
 775         if (elem_next) {
 776                 memcpy(key_next, elem_next->key, key_size);
 777                 return 0;
 778         }
 779
 780         i = hash & (htab->buckets_num - 1);
 781         i++;
 782 find_first_elem:
 783         for (; i < htab->buckets_num; i++) {
 784                 head = &sock_hash_select_bucket(htab, i)->head;
 785                 elem_next = hlist_entry_safe(rcu_dereference_raw(hlist_first_rcu(head)),
 786                                              struct bpf_htab_elem, node);
 787                 if (elem_next) {
 788                         memcpy(key_next, elem_next->key, key_size);
 789                         return 0;
 790                 }
 791         }
 792
 793         return -ENOENT;
 794 }
 795
 796 static struct bpf_map *sock_hash_alloc(union bpf_attr *attr)
 797 {
 798         struct bpf_htab *htab;
 799         int i, err;
 800         u64 cost;
 801
 802         if (!capable(CAP_NET_ADMIN))
 803                 return ERR_PTR(-EPERM);
 804         if (attr->max_entries == 0 ||
 805             attr->key_size    == 0 ||
 806             attr->value_size  != 4 ||
 807             attr->map_flags & ~SOCK_CREATE_FLAG_MASK)
 808                 return ERR_PTR(-EINVAL);
 809         if (attr->key_size > MAX_BPF_STACK)
 810                 return ERR_PTR(-E2BIG);
 811
 812         htab = kzalloc(sizeof(*htab), GFP_USER);
 813         if (!htab)
 814                 return ERR_PTR(-ENOMEM);
 815
 816         bpf_map_init_from_attr(&htab->map, attr);
 817
 818         htab->buckets_num = roundup_pow_of_two(htab->map.max_entries);
 819         htab->elem_size = sizeof(struct bpf_htab_elem) +
 820                           round_up(htab->map.key_size, 8);
 821         if (htab->buckets_num == 0 ||
 822             htab->buckets_num > U32_MAX / sizeof(struct bpf_htab_bucket)) {
 823                 err = -EINVAL;
 824                 goto free_htab;
 825         }
 826
 827         cost = (u64) htab->buckets_num * sizeof(struct bpf_htab_bucket) +
 828                (u64) htab->elem_size * htab->map.max_entries;
 829         if (cost >= U32_MAX - PAGE_SIZE) {
 830                 err = -EINVAL;
 831                 goto free_htab;
 832         }
 833
 834         htab->buckets = bpf_map_area_alloc(htab->buckets_num *
 835                                            sizeof(struct bpf_htab_bucket),
 836                                            htab->map.numa_node);
 837         if (!htab->buckets) {
 838                 err = -ENOMEM;
 839                 goto free_htab;
 840         }
 841
 842         for (i = 0; i < htab->buckets_num; i++) {
 843                 INIT_HLIST_HEAD(&htab->buckets[i].head);
 844                 raw_spin_lock_init(&htab->buckets[i].lock);
 845         }
 846
 847         return &htab->map;
 848 free_htab:
 849         kfree(htab);
 850         return ERR_PTR(err);
 851 }
 852
 853 static void sock_hash_free(struct bpf_map *map)
 854 {
 855         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 856         struct bpf_htab_bucket *bucket;
 857         struct bpf_htab_elem *elem;
 858         struct hlist_node *node;
 859         int i;
 860
 861         synchronize_rcu();
 862         rcu_read_lock();
 863         for (i = 0; i < htab->buckets_num; i++) {
 864                 bucket = sock_hash_select_bucket(htab, i);
 865                 raw_spin_lock_bh(&bucket->lock);
 866                 hlist_for_each_entry_safe(elem, node, &bucket->head, node) {
 867                         hlist_del_rcu(&elem->node);
 868                         lock_sock(elem->sk);
 869                         sock_map_unref(elem->sk, elem);
 870                         release_sock(elem->sk);
 871                 }
 872                 raw_spin_unlock_bh(&bucket->lock);
 873         }
 874         rcu_read_unlock();
 875
 876         bpf_map_area_free(htab->buckets);
 877         kfree(htab);
 878 }
 879
 880 static void sock_hash_release_progs(struct bpf_map *map)
 881 {
 882         psock_progs_drop(&container_of(map, struct bpf_htab, map)->progs);
 883 }
 884
 885 BPF_CALL_4(bpf_sock_hash_update, struct bpf_sock_ops_kern *, sops,
 886            struct bpf_map *, map, void *, key, u64, flags)
 887 {
 888         WARN_ON_ONCE(!rcu_read_lock_held());
 889
 890         if (likely(sock_map_sk_is_suitable(sops->sk) &&
 891                    sock_map_op_okay(sops)))
 892                 return sock_hash_update_common(map, key, sops->sk, flags);
 893         return -EOPNOTSUPP;
 894 }
 895
 896 const struct bpf_func_proto bpf_sock_hash_update_proto = {
 897         .func           = bpf_sock_hash_update,
 898         .gpl_only       = false,
 899         .pkt_access     = true,
 900         .ret_type       = RET_INTEGER,
 901         .arg1_type      = ARG_PTR_TO_CTX,
 902         .arg2_type      = ARG_CONST_MAP_PTR,
 903         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 904         .arg4_type      = ARG_ANYTHING,
 905 };
 906
 907 BPF_CALL_4(bpf_sk_redirect_hash, struct sk_buff *, skb,
 908            struct bpf_map *, map, void *, key, u64, flags)
 909 {
 910         struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
 911
 912         if (unlikely(flags & ~(BPF_F_INGRESS)))
 913                 return SK_DROP;
 914         tcb->bpf.flags = flags;
 915         tcb->bpf.sk_redir = __sock_hash_lookup_elem(map, key);
 916         if (!tcb->bpf.sk_redir)
 917                 return SK_DROP;
 918         return SK_PASS;
 919 }
 920
 921 const struct bpf_func_proto bpf_sk_redirect_hash_proto = {
 922         .func           = bpf_sk_redirect_hash,
 923         .gpl_only       = false,
 924         .ret_type       = RET_INTEGER,
 925         .arg1_type      = ARG_PTR_TO_CTX,
 926         .arg2_type      = ARG_CONST_MAP_PTR,
 927         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 928         .arg4_type      = ARG_ANYTHING,
 929 };
 930
 931 BPF_CALL_4(bpf_msg_redirect_hash, struct sk_msg *, msg,
 932            struct bpf_map *, map, void *, key, u64, flags)
 933 {
 934         if (unlikely(flags & ~(BPF_F_INGRESS)))
 935                 return SK_DROP;
 936         msg->flags = flags;
 937         msg->sk_redir = __sock_hash_lookup_elem(map, key);
 938         if (!msg->sk_redir)
 939                 return SK_DROP;
 940         return SK_PASS;
 941 }
 942
 943 const struct bpf_func_proto bpf_msg_redirect_hash_proto = {
 944         .func           = bpf_msg_redirect_hash,
 945         .gpl_only       = false,
 946         .ret_type       = RET_INTEGER,
 947         .arg1_type      = ARG_PTR_TO_CTX,
 948         .arg2_type      = ARG_CONST_MAP_PTR,
 949         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 950         .arg4_type      = ARG_ANYTHING,
 951 };
 952
 953 const struct bpf_map_ops sock_hash_ops = {
 954         .map_alloc              = sock_hash_alloc,
 955         .map_free               = sock_hash_free,
 956         .map_get_next_key       = sock_hash_get_next_key,
 957         .map_update_elem        = sock_hash_update_elem,
 958         .map_delete_elem        = sock_hash_delete_elem,
 959         .map_lookup_elem        = sock_map_lookup,
 960         .map_release_uref       = sock_hash_release_progs,
 961         .map_check_btf          = map_check_no_btf,
 962 };
 963
 964 static struct sk_psock_progs *sock_map_progs(struct bpf_map *map)
 965 {
 966         switch (map->map_type) {
 967         case BPF_MAP_TYPE_SOCKMAP:
 968                 return &container_of(map, struct bpf_stab, map)->progs;
 969         case BPF_MAP_TYPE_SOCKHASH:
 970                 return &container_of(map, struct bpf_htab, map)->progs;
 971         default:
 972                 break;
 973         }
 974
 975         return NULL;
 976 }
 977
 978 int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog,
 979                          u32 which)
 980 {
 981         struct sk_psock_progs *progs = sock_map_progs(map);
 982
 983         if (!progs)
 984                 return -EOPNOTSUPP;
 985
 986         switch (which) {
 987         case BPF_SK_MSG_VERDICT:
 988                 psock_set_prog(&progs->msg_parser, prog);
 989                 break;
 990         case BPF_SK_SKB_STREAM_PARSER:
 991                 psock_set_prog(&progs->skb_parser, prog);
 992                 break;
 993         case BPF_SK_SKB_STREAM_VERDICT:
 994                 psock_set_prog(&progs->skb_verdict, prog);
 995                 break;
 996         default:
 997                 return -EOPNOTSUPP;
 998         }
 999
1000         return 0;
1001 }
1002
1003 void sk_psock_unlink(struct sock *sk, struct sk_psock_link *link)
1004 {
1005         switch (link->map->map_type) {
1006         case BPF_MAP_TYPE_SOCKMAP:
1007                 return sock_map_delete_from_link(link->map, sk,
1008                                                  link->link_raw);
1009         case BPF_MAP_TYPE_SOCKHASH:
1010                 return sock_hash_delete_from_link(link->map, sk,
1011                                                   link->link_raw);
1012         default:
1013                 break;
1014         }
1015 }