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                         sock_map_unref(sk, psk);
 246         }
 247         raw_spin_unlock_bh(&stab->lock);
 248         rcu_read_unlock();
 249
 250         bpf_map_area_free(stab->sks);
 251         kfree(stab);
 252 }
 253
 254 static void sock_map_release_progs(struct bpf_map *map)
 255 {
 256         psock_progs_drop(&container_of(map, struct bpf_stab, map)->progs);
 257 }
 258
 259 static struct sock *__sock_map_lookup_elem(struct bpf_map *map, u32 key)
 260 {
 261         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 262
 263         WARN_ON_ONCE(!rcu_read_lock_held());
 264
 265         if (unlikely(key >= map->max_entries))
 266                 return NULL;
 267         return READ_ONCE(stab->sks[key]);
 268 }
 269
 270 static void *sock_map_lookup(struct bpf_map *map, void *key)
 271 {
 272         return ERR_PTR(-EOPNOTSUPP);
 273 }
 274
 275 static int __sock_map_delete(struct bpf_stab *stab, struct sock *sk_test,
 276                              struct sock **psk)
 277 {
 278         struct sock *sk;
 279
 280         raw_spin_lock_bh(&stab->lock);
 281         sk = *psk;
 282         if (!sk_test || sk_test == sk)
 283                 *psk = NULL;
 284         raw_spin_unlock_bh(&stab->lock);
 285         if (unlikely(!sk))
 286                 return -EINVAL;
 287         sock_map_unref(sk, psk);
 288         return 0;
 289 }
 290
 291 static void sock_map_delete_from_link(struct bpf_map *map, struct sock *sk,
 292                                       void *link_raw)
 293 {
 294         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 295
 296         __sock_map_delete(stab, sk, link_raw);
 297 }
 298
 299 static int sock_map_delete_elem(struct bpf_map *map, void *key)
 300 {
 301         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 302         u32 i = *(u32 *)key;
 303         struct sock **psk;
 304
 305         if (unlikely(i >= map->max_entries))
 306                 return -EINVAL;
 307
 308         psk = &stab->sks[i];
 309         return __sock_map_delete(stab, NULL, psk);
 310 }
 311
 312 static int sock_map_get_next_key(struct bpf_map *map, void *key, void *next)
 313 {
 314         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 315         u32 i = key ? *(u32 *)key : U32_MAX;
 316         u32 *key_next = next;
 317
 318         if (i == stab->map.max_entries - 1)
 319                 return -ENOENT;
 320         if (i >= stab->map.max_entries)
 321                 *key_next = 0;
 322         else
 323                 *key_next = i + 1;
 324         return 0;
 325 }
 326
 327 static int sock_map_update_common(struct bpf_map *map, u32 idx,
 328                                   struct sock *sk, u64 flags)
 329 {
 330         struct bpf_stab *stab = container_of(map, struct bpf_stab, map);
 331         struct sk_psock_link *link;
 332         struct sk_psock *psock;
 333         struct sock *osk;
 334         int ret;
 335
 336         WARN_ON_ONCE(!rcu_read_lock_held());
 337         if (unlikely(flags > BPF_EXIST))
 338                 return -EINVAL;
 339         if (unlikely(idx >= map->max_entries))
 340                 return -E2BIG;
 341
 342         link = sk_psock_init_link();
 343         if (!link)
 344                 return -ENOMEM;
 345
 346         ret = sock_map_link(map, &stab->progs, sk);
 347         if (ret < 0)
 348                 goto out_free;
 349
 350         psock = sk_psock(sk);
 351         WARN_ON_ONCE(!psock);
 352
 353         raw_spin_lock_bh(&stab->lock);
 354         osk = stab->sks[idx];
 355         if (osk && flags == BPF_NOEXIST) {
 356                 ret = -EEXIST;
 357                 goto out_unlock;
 358         } else if (!osk && flags == BPF_EXIST) {
 359                 ret = -ENOENT;
 360                 goto out_unlock;
 361         }
 362
 363         sock_map_add_link(psock, link, map, &stab->sks[idx]);
 364         stab->sks[idx] = sk;
 365         if (osk)
 366                 sock_map_unref(osk, &stab->sks[idx]);
 367         raw_spin_unlock_bh(&stab->lock);
 368         return 0;
 369 out_unlock:
 370         raw_spin_unlock_bh(&stab->lock);
 371         if (psock)
 372                 sk_psock_put(sk, psock);
 373 out_free:
 374         sk_psock_free_link(link);
 375         return ret;
 376 }
 377
 378 static bool sock_map_op_okay(const struct bpf_sock_ops_kern *ops)
 379 {
 380         return ops->op == BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB ||
 381                ops->op == BPF_SOCK_OPS_ACTIVE_ESTABLISHED_CB;
 382 }
 383
 384 static bool sock_map_sk_is_suitable(const struct sock *sk)
 385 {
 386         return sk->sk_type == SOCK_STREAM &&
 387                sk->sk_protocol == IPPROTO_TCP;
 388 }
 389
 390 static int sock_map_update_elem(struct bpf_map *map, void *key,
 391                                 void *value, u64 flags)
 392 {
 393         u32 ufd = *(u32 *)value;
 394         u32 idx = *(u32 *)key;
 395         struct socket *sock;
 396         struct sock *sk;
 397         int ret;
 398
 399         sock = sockfd_lookup(ufd, &ret);
 400         if (!sock)
 401                 return ret;
 402         sk = sock->sk;
 403         if (!sk) {
 404                 ret = -EINVAL;
 405                 goto out;
 406         }
 407         if (!sock_map_sk_is_suitable(sk) ||
 408             sk->sk_state != TCP_ESTABLISHED) {
 409                 ret = -EOPNOTSUPP;
 410                 goto out;
 411         }
 412
 413         sock_map_sk_acquire(sk);
 414         ret = sock_map_update_common(map, idx, sk, flags);
 415         sock_map_sk_release(sk);
 416 out:
 417         fput(sock->file);
 418         return ret;
 419 }
 420
 421 BPF_CALL_4(bpf_sock_map_update, struct bpf_sock_ops_kern *, sops,
 422            struct bpf_map *, map, void *, key, u64, flags)
 423 {
 424         WARN_ON_ONCE(!rcu_read_lock_held());
 425
 426         if (likely(sock_map_sk_is_suitable(sops->sk) &&
 427                    sock_map_op_okay(sops)))
 428                 return sock_map_update_common(map, *(u32 *)key, sops->sk,
 429                                               flags);
 430         return -EOPNOTSUPP;
 431 }
 432
 433 const struct bpf_func_proto bpf_sock_map_update_proto = {
 434         .func           = bpf_sock_map_update,
 435         .gpl_only       = false,
 436         .pkt_access     = true,
 437         .ret_type       = RET_INTEGER,
 438         .arg1_type      = ARG_PTR_TO_CTX,
 439         .arg2_type      = ARG_CONST_MAP_PTR,
 440         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 441         .arg4_type      = ARG_ANYTHING,
 442 };
 443
 444 BPF_CALL_4(bpf_sk_redirect_map, struct sk_buff *, skb,
 445            struct bpf_map *, map, u32, key, u64, flags)
 446 {
 447         struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
 448
 449         if (unlikely(flags & ~(BPF_F_INGRESS)))
 450                 return SK_DROP;
 451         tcb->bpf.flags = flags;
 452         tcb->bpf.sk_redir = __sock_map_lookup_elem(map, key);
 453         if (!tcb->bpf.sk_redir)
 454                 return SK_DROP;
 455         return SK_PASS;
 456 }
 457
 458 const struct bpf_func_proto bpf_sk_redirect_map_proto = {
 459         .func           = bpf_sk_redirect_map,
 460         .gpl_only       = false,
 461         .ret_type       = RET_INTEGER,
 462         .arg1_type      = ARG_PTR_TO_CTX,
 463         .arg2_type      = ARG_CONST_MAP_PTR,
 464         .arg3_type      = ARG_ANYTHING,
 465         .arg4_type      = ARG_ANYTHING,
 466 };
 467
 468 BPF_CALL_4(bpf_msg_redirect_map, struct sk_msg *, msg,
 469            struct bpf_map *, map, u32, key, u64, flags)
 470 {
 471         if (unlikely(flags & ~(BPF_F_INGRESS)))
 472                 return SK_DROP;
 473         msg->flags = flags;
 474         msg->sk_redir = __sock_map_lookup_elem(map, key);
 475         if (!msg->sk_redir)
 476                 return SK_DROP;
 477         return SK_PASS;
 478 }
 479
 480 const struct bpf_func_proto bpf_msg_redirect_map_proto = {
 481         .func           = bpf_msg_redirect_map,
 482         .gpl_only       = false,
 483         .ret_type       = RET_INTEGER,
 484         .arg1_type      = ARG_PTR_TO_CTX,
 485         .arg2_type      = ARG_CONST_MAP_PTR,
 486         .arg3_type      = ARG_ANYTHING,
 487         .arg4_type      = ARG_ANYTHING,
 488 };
 489
 490 const struct bpf_map_ops sock_map_ops = {
 491         .map_alloc              = sock_map_alloc,
 492         .map_free               = sock_map_free,
 493         .map_get_next_key       = sock_map_get_next_key,
 494         .map_update_elem        = sock_map_update_elem,
 495         .map_delete_elem        = sock_map_delete_elem,
 496         .map_lookup_elem        = sock_map_lookup,
 497         .map_release_uref       = sock_map_release_progs,
 498         .map_check_btf          = map_check_no_btf,
 499 };
 500
 501 struct bpf_htab_elem {
 502         struct rcu_head rcu;
 503         u32 hash;
 504         struct sock *sk;
 505         struct hlist_node node;
 506         u8 key[0];
 507 };
 508
 509 struct bpf_htab_bucket {
 510         struct hlist_head head;
 511         raw_spinlock_t lock;
 512 };
 513
 514 struct bpf_htab {
 515         struct bpf_map map;
 516         struct bpf_htab_bucket *buckets;
 517         u32 buckets_num;
 518         u32 elem_size;
 519         struct sk_psock_progs progs;
 520         atomic_t count;
 521 };
 522
 523 static inline u32 sock_hash_bucket_hash(const void *key, u32 len)
 524 {
 525         return jhash(key, len, 0);
 526 }
 527
 528 static struct bpf_htab_bucket *sock_hash_select_bucket(struct bpf_htab *htab,
 529                                                        u32 hash)
 530 {
 531         return &htab->buckets[hash & (htab->buckets_num - 1)];
 532 }
 533
 534 static struct bpf_htab_elem *
 535 sock_hash_lookup_elem_raw(struct hlist_head *head, u32 hash, void *key,
 536                           u32 key_size)
 537 {
 538         struct bpf_htab_elem *elem;
 539
 540         hlist_for_each_entry_rcu(elem, head, node) {
 541                 if (elem->hash == hash &&
 542                     !memcmp(&elem->key, key, key_size))
 543                         return elem;
 544         }
 545
 546         return NULL;
 547 }
 548
 549 static struct sock *__sock_hash_lookup_elem(struct bpf_map *map, void *key)
 550 {
 551         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 552         u32 key_size = map->key_size, hash;
 553         struct bpf_htab_bucket *bucket;
 554         struct bpf_htab_elem *elem;
 555
 556         WARN_ON_ONCE(!rcu_read_lock_held());
 557
 558         hash = sock_hash_bucket_hash(key, key_size);
 559         bucket = sock_hash_select_bucket(htab, hash);
 560         elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
 561
 562         return elem ? elem->sk : NULL;
 563 }
 564
 565 static void sock_hash_free_elem(struct bpf_htab *htab,
 566                                 struct bpf_htab_elem *elem)
 567 {
 568         atomic_dec(&htab->count);
 569         kfree_rcu(elem, rcu);
 570 }
 571
 572 static void sock_hash_delete_from_link(struct bpf_map *map, struct sock *sk,
 573                                        void *link_raw)
 574 {
 575         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 576         struct bpf_htab_elem *elem_probe, *elem = link_raw;
 577         struct bpf_htab_bucket *bucket;
 578
 579         WARN_ON_ONCE(!rcu_read_lock_held());
 580         bucket = sock_hash_select_bucket(htab, elem->hash);
 581
 582         /* elem may be deleted in parallel from the map, but access here
 583          * is okay since it's going away only after RCU grace period.
 584          * However, we need to check whether it's still present.
 585          */
 586         raw_spin_lock_bh(&bucket->lock);
 587         elem_probe = sock_hash_lookup_elem_raw(&bucket->head, elem->hash,
 588                                                elem->key, map->key_size);
 589         if (elem_probe && elem_probe == elem) {
 590                 hlist_del_rcu(&elem->node);
 591                 sock_map_unref(elem->sk, elem);
 592                 sock_hash_free_elem(htab, elem);
 593         }
 594         raw_spin_unlock_bh(&bucket->lock);
 595 }
 596
 597 static int sock_hash_delete_elem(struct bpf_map *map, void *key)
 598 {
 599         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 600         u32 hash, key_size = map->key_size;
 601         struct bpf_htab_bucket *bucket;
 602         struct bpf_htab_elem *elem;
 603         int ret = -ENOENT;
 604
 605         hash = sock_hash_bucket_hash(key, key_size);
 606         bucket = sock_hash_select_bucket(htab, hash);
 607
 608         raw_spin_lock_bh(&bucket->lock);
 609         elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
 610         if (elem) {
 611                 hlist_del_rcu(&elem->node);
 612                 sock_map_unref(elem->sk, elem);
 613                 sock_hash_free_elem(htab, elem);
 614                 ret = 0;
 615         }
 616         raw_spin_unlock_bh(&bucket->lock);
 617         return ret;
 618 }
 619
 620 static struct bpf_htab_elem *sock_hash_alloc_elem(struct bpf_htab *htab,
 621                                                   void *key, u32 key_size,
 622                                                   u32 hash, struct sock *sk,
 623                                                   struct bpf_htab_elem *old)
 624 {
 625         struct bpf_htab_elem *new;
 626
 627         if (atomic_inc_return(&htab->count) > htab->map.max_entries) {
 628                 if (!old) {
 629                         atomic_dec(&htab->count);
 630                         return ERR_PTR(-E2BIG);
 631                 }
 632         }
 633
 634         new = kmalloc_node(htab->elem_size, GFP_ATOMIC | __GFP_NOWARN,
 635                            htab->map.numa_node);
 636         if (!new) {
 637                 atomic_dec(&htab->count);
 638                 return ERR_PTR(-ENOMEM);
 639         }
 640         memcpy(new->key, key, key_size);
 641         new->sk = sk;
 642         new->hash = hash;
 643         return new;
 644 }
 645
 646 static int sock_hash_update_common(struct bpf_map *map, void *key,
 647                                    struct sock *sk, u64 flags)
 648 {
 649         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 650         u32 key_size = map->key_size, hash;
 651         struct bpf_htab_elem *elem, *elem_new;
 652         struct bpf_htab_bucket *bucket;
 653         struct sk_psock_link *link;
 654         struct sk_psock *psock;
 655         int ret;
 656
 657         WARN_ON_ONCE(!rcu_read_lock_held());
 658         if (unlikely(flags > BPF_EXIST))
 659                 return -EINVAL;
 660
 661         link = sk_psock_init_link();
 662         if (!link)
 663                 return -ENOMEM;
 664
 665         ret = sock_map_link(map, &htab->progs, sk);
 666         if (ret < 0)
 667                 goto out_free;
 668
 669         psock = sk_psock(sk);
 670         WARN_ON_ONCE(!psock);
 671
 672         hash = sock_hash_bucket_hash(key, key_size);
 673         bucket = sock_hash_select_bucket(htab, hash);
 674
 675         raw_spin_lock_bh(&bucket->lock);
 676         elem = sock_hash_lookup_elem_raw(&bucket->head, hash, key, key_size);
 677         if (elem && flags == BPF_NOEXIST) {
 678                 ret = -EEXIST;
 679                 goto out_unlock;
 680         } else if (!elem && flags == BPF_EXIST) {
 681                 ret = -ENOENT;
 682                 goto out_unlock;
 683         }
 684
 685         elem_new = sock_hash_alloc_elem(htab, key, key_size, hash, sk, elem);
 686         if (IS_ERR(elem_new)) {
 687                 ret = PTR_ERR(elem_new);
 688                 goto out_unlock;
 689         }
 690
 691         sock_map_add_link(psock, link, map, elem_new);
 692         /* Add new element to the head of the list, so that
 693          * concurrent search will find it before old elem.
 694          */
 695         hlist_add_head_rcu(&elem_new->node, &bucket->head);
 696         if (elem) {
 697                 hlist_del_rcu(&elem->node);
 698                 sock_map_unref(elem->sk, elem);
 699                 sock_hash_free_elem(htab, elem);
 700         }
 701         raw_spin_unlock_bh(&bucket->lock);
 702         return 0;
 703 out_unlock:
 704         raw_spin_unlock_bh(&bucket->lock);
 705         sk_psock_put(sk, psock);
 706 out_free:
 707         sk_psock_free_link(link);
 708         return ret;
 709 }
 710
 711 static int sock_hash_update_elem(struct bpf_map *map, void *key,
 712                                  void *value, u64 flags)
 713 {
 714         u32 ufd = *(u32 *)value;
 715         struct socket *sock;
 716         struct sock *sk;
 717         int ret;
 718
 719         sock = sockfd_lookup(ufd, &ret);
 720         if (!sock)
 721                 return ret;
 722         sk = sock->sk;
 723         if (!sk) {
 724                 ret = -EINVAL;
 725                 goto out;
 726         }
 727         if (!sock_map_sk_is_suitable(sk) ||
 728             sk->sk_state != TCP_ESTABLISHED) {
 729                 ret = -EOPNOTSUPP;
 730                 goto out;
 731         }
 732
 733         sock_map_sk_acquire(sk);
 734         ret = sock_hash_update_common(map, key, sk, flags);
 735         sock_map_sk_release(sk);
 736 out:
 737         fput(sock->file);
 738         return ret;
 739 }
 740
 741 static int sock_hash_get_next_key(struct bpf_map *map, void *key,
 742                                   void *key_next)
 743 {
 744         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 745         struct bpf_htab_elem *elem, *elem_next;
 746         u32 hash, key_size = map->key_size;
 747         struct hlist_head *head;
 748         int i = 0;
 749
 750         if (!key)
 751                 goto find_first_elem;
 752         hash = sock_hash_bucket_hash(key, key_size);
 753         head = &sock_hash_select_bucket(htab, hash)->head;
 754         elem = sock_hash_lookup_elem_raw(head, hash, key, key_size);
 755         if (!elem)
 756                 goto find_first_elem;
 757
 758         elem_next = hlist_entry_safe(rcu_dereference_raw(hlist_next_rcu(&elem->node)),
 759                                      struct bpf_htab_elem, node);
 760         if (elem_next) {
 761                 memcpy(key_next, elem_next->key, key_size);
 762                 return 0;
 763         }
 764
 765         i = hash & (htab->buckets_num - 1);
 766         i++;
 767 find_first_elem:
 768         for (; i < htab->buckets_num; i++) {
 769                 head = &sock_hash_select_bucket(htab, i)->head;
 770                 elem_next = hlist_entry_safe(rcu_dereference_raw(hlist_first_rcu(head)),
 771                                              struct bpf_htab_elem, node);
 772                 if (elem_next) {
 773                         memcpy(key_next, elem_next->key, key_size);
 774                         return 0;
 775                 }
 776         }
 777
 778         return -ENOENT;
 779 }
 780
 781 static struct bpf_map *sock_hash_alloc(union bpf_attr *attr)
 782 {
 783         struct bpf_htab *htab;
 784         int i, err;
 785         u64 cost;
 786
 787         if (!capable(CAP_NET_ADMIN))
 788                 return ERR_PTR(-EPERM);
 789         if (attr->max_entries == 0 ||
 790             attr->key_size    == 0 ||
 791             attr->value_size  != 4 ||
 792             attr->map_flags & ~SOCK_CREATE_FLAG_MASK)
 793                 return ERR_PTR(-EINVAL);
 794         if (attr->key_size > MAX_BPF_STACK)
 795                 return ERR_PTR(-E2BIG);
 796
 797         htab = kzalloc(sizeof(*htab), GFP_USER);
 798         if (!htab)
 799                 return ERR_PTR(-ENOMEM);
 800
 801         bpf_map_init_from_attr(&htab->map, attr);
 802
 803         htab->buckets_num = roundup_pow_of_two(htab->map.max_entries);
 804         htab->elem_size = sizeof(struct bpf_htab_elem) +
 805                           round_up(htab->map.key_size, 8);
 806         if (htab->buckets_num == 0 ||
 807             htab->buckets_num > U32_MAX / sizeof(struct bpf_htab_bucket)) {
 808                 err = -EINVAL;
 809                 goto free_htab;
 810         }
 811
 812         cost = (u64) htab->buckets_num * sizeof(struct bpf_htab_bucket) +
 813                (u64) htab->elem_size * htab->map.max_entries;
 814         if (cost >= U32_MAX - PAGE_SIZE) {
 815                 err = -EINVAL;
 816                 goto free_htab;
 817         }
 818
 819         htab->buckets = bpf_map_area_alloc(htab->buckets_num *
 820                                            sizeof(struct bpf_htab_bucket),
 821                                            htab->map.numa_node);
 822         if (!htab->buckets) {
 823                 err = -ENOMEM;
 824                 goto free_htab;
 825         }
 826
 827         for (i = 0; i < htab->buckets_num; i++) {
 828                 INIT_HLIST_HEAD(&htab->buckets[i].head);
 829                 raw_spin_lock_init(&htab->buckets[i].lock);
 830         }
 831
 832         return &htab->map;
 833 free_htab:
 834         kfree(htab);
 835         return ERR_PTR(err);
 836 }
 837
 838 static void sock_hash_free(struct bpf_map *map)
 839 {
 840         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 841         struct bpf_htab_bucket *bucket;
 842         struct bpf_htab_elem *elem;
 843         struct hlist_node *node;
 844         int i;
 845
 846         synchronize_rcu();
 847         rcu_read_lock();
 848         for (i = 0; i < htab->buckets_num; i++) {
 849                 bucket = sock_hash_select_bucket(htab, i);
 850                 raw_spin_lock_bh(&bucket->lock);
 851                 hlist_for_each_entry_safe(elem, node, &bucket->head, node) {
 852                         hlist_del_rcu(&elem->node);
 853                         sock_map_unref(elem->sk, elem);
 854                 }
 855                 raw_spin_unlock_bh(&bucket->lock);
 856         }
 857         rcu_read_unlock();
 858
 859         bpf_map_area_free(htab->buckets);
 860         kfree(htab);
 861 }
 862
 863 static void sock_hash_release_progs(struct bpf_map *map)
 864 {
 865         psock_progs_drop(&container_of(map, struct bpf_htab, map)->progs);
 866 }
 867
 868 BPF_CALL_4(bpf_sock_hash_update, struct bpf_sock_ops_kern *, sops,
 869            struct bpf_map *, map, void *, key, u64, flags)
 870 {
 871         WARN_ON_ONCE(!rcu_read_lock_held());
 872
 873         if (likely(sock_map_sk_is_suitable(sops->sk) &&
 874                    sock_map_op_okay(sops)))
 875                 return sock_hash_update_common(map, key, sops->sk, flags);
 876         return -EOPNOTSUPP;
 877 }
 878
 879 const struct bpf_func_proto bpf_sock_hash_update_proto = {
 880         .func           = bpf_sock_hash_update,
 881         .gpl_only       = false,
 882         .pkt_access     = true,
 883         .ret_type       = RET_INTEGER,
 884         .arg1_type      = ARG_PTR_TO_CTX,
 885         .arg2_type      = ARG_CONST_MAP_PTR,
 886         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 887         .arg4_type      = ARG_ANYTHING,
 888 };
 889
 890 BPF_CALL_4(bpf_sk_redirect_hash, struct sk_buff *, skb,
 891            struct bpf_map *, map, void *, key, u64, flags)
 892 {
 893         struct tcp_skb_cb *tcb = TCP_SKB_CB(skb);
 894
 895         if (unlikely(flags & ~(BPF_F_INGRESS)))
 896                 return SK_DROP;
 897         tcb->bpf.flags = flags;
 898         tcb->bpf.sk_redir = __sock_hash_lookup_elem(map, key);
 899         if (!tcb->bpf.sk_redir)
 900                 return SK_DROP;
 901         return SK_PASS;
 902 }
 903
 904 const struct bpf_func_proto bpf_sk_redirect_hash_proto = {
 905         .func           = bpf_sk_redirect_hash,
 906         .gpl_only       = false,
 907         .ret_type       = RET_INTEGER,
 908         .arg1_type      = ARG_PTR_TO_CTX,
 909         .arg2_type      = ARG_CONST_MAP_PTR,
 910         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 911         .arg4_type      = ARG_ANYTHING,
 912 };
 913
 914 BPF_CALL_4(bpf_msg_redirect_hash, struct sk_msg *, msg,
 915            struct bpf_map *, map, void *, key, u64, flags)
 916 {
 917         if (unlikely(flags & ~(BPF_F_INGRESS)))
 918                 return SK_DROP;
 919         msg->flags = flags;
 920         msg->sk_redir = __sock_hash_lookup_elem(map, key);
 921         if (!msg->sk_redir)
 922                 return SK_DROP;
 923         return SK_PASS;
 924 }
 925
 926 const struct bpf_func_proto bpf_msg_redirect_hash_proto = {
 927         .func           = bpf_msg_redirect_hash,
 928         .gpl_only       = false,
 929         .ret_type       = RET_INTEGER,
 930         .arg1_type      = ARG_PTR_TO_CTX,
 931         .arg2_type      = ARG_CONST_MAP_PTR,
 932         .arg3_type      = ARG_PTR_TO_MAP_KEY,
 933         .arg4_type      = ARG_ANYTHING,
 934 };
 935
 936 const struct bpf_map_ops sock_hash_ops = {
 937         .map_alloc              = sock_hash_alloc,
 938         .map_free               = sock_hash_free,
 939         .map_get_next_key       = sock_hash_get_next_key,
 940         .map_update_elem        = sock_hash_update_elem,
 941         .map_delete_elem        = sock_hash_delete_elem,
 942         .map_lookup_elem        = sock_map_lookup,
 943         .map_release_uref       = sock_hash_release_progs,
 944         .map_check_btf          = map_check_no_btf,
 945 };
 946
 947 static struct sk_psock_progs *sock_map_progs(struct bpf_map *map)
 948 {
 949         switch (map->map_type) {
 950         case BPF_MAP_TYPE_SOCKMAP:
 951                 return &container_of(map, struct bpf_stab, map)->progs;
 952         case BPF_MAP_TYPE_SOCKHASH:
 953                 return &container_of(map, struct bpf_htab, map)->progs;
 954         default:
 955                 break;
 956         }
 957
 958         return NULL;
 959 }
 960
 961 int sock_map_prog_update(struct bpf_map *map, struct bpf_prog *prog,
 962                          u32 which)
 963 {
 964         struct sk_psock_progs *progs = sock_map_progs(map);
 965
 966         if (!progs)
 967                 return -EOPNOTSUPP;
 968
 969         switch (which) {
 970         case BPF_SK_MSG_VERDICT:
 971                 psock_set_prog(&progs->msg_parser, prog);
 972                 break;
 973         case BPF_SK_SKB_STREAM_PARSER:
 974                 psock_set_prog(&progs->skb_parser, prog);
 975                 break;
 976         case BPF_SK_SKB_STREAM_VERDICT:
 977                 psock_set_prog(&progs->skb_verdict, prog);
 978                 break;
 979         default:
 980                 return -EOPNOTSUPP;
 981         }
 982
 983         return 0;
 984 }
 985
 986 void sk_psock_unlink(struct sock *sk, struct sk_psock_link *link)
 987 {
 988         switch (link->map->map_type) {
 989         case BPF_MAP_TYPE_SOCKMAP:
 990                 return sock_map_delete_from_link(link->map, sk,
 991                                                  link->link_raw);
 992         case BPF_MAP_TYPE_SOCKHASH:
 993                 return sock_hash_delete_from_link(link->map, sk,
 994                                                   link->link_raw);
 995         default:
 996                 break;
 997         }
 998 }