net/core/bpf_sk_storage.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 /* Copyright (c) 2019 Facebook  */
   3 #include <linux/rculist.h>
   4 #include <linux/list.h>
   5 #include <linux/hash.h>
   6 #include <linux/types.h>
   7 #include <linux/spinlock.h>
   8 #include <linux/bpf.h>
   9 #include <linux/btf.h>
  10 #include <linux/btf_ids.h>
  11 #include <linux/bpf_local_storage.h>
  12 #include <net/bpf_sk_storage.h>
  13 #include <net/sock.h>
  14 #include <uapi/linux/sock_diag.h>
  15 #include <uapi/linux/btf.h>
  16 #include <linux/rcupdate_trace.h>
  17
  18 DEFINE_BPF_STORAGE_CACHE(sk_cache);
  19
  20 static struct bpf_local_storage_data *
  21 bpf_sk_storage_lookup(struct sock *sk, struct bpf_map *map, bool cacheit_lockit)
  22 {
  23         struct bpf_local_storage *sk_storage;
  24         struct bpf_local_storage_map *smap;
  25
  26         sk_storage =
  27                 rcu_dereference_check(sk->sk_bpf_storage, bpf_rcu_lock_held());
  28         if (!sk_storage)
  29                 return NULL;
  30
  31         smap = (struct bpf_local_storage_map *)map;
  32         return bpf_local_storage_lookup(sk_storage, smap, cacheit_lockit);
  33 }
  34
  35 static int bpf_sk_storage_del(struct sock *sk, struct bpf_map *map)
  36 {
  37         struct bpf_local_storage_data *sdata;
  38
  39         sdata = bpf_sk_storage_lookup(sk, map, false);
  40         if (!sdata)
  41                 return -ENOENT;
  42
  43         bpf_selem_unlink(SELEM(sdata), true);
  44
  45         return 0;
  46 }
  47
  48 /* Called by __sk_destruct() & bpf_sk_storage_clone() */
  49 void bpf_sk_storage_free(struct sock *sk)
  50 {
  51         struct bpf_local_storage_elem *selem;
  52         struct bpf_local_storage *sk_storage;
  53         bool free_sk_storage = false;
  54         struct hlist_node *n;
  55
  56         rcu_read_lock();
  57         sk_storage = rcu_dereference(sk->sk_bpf_storage);
  58         if (!sk_storage) {
  59                 rcu_read_unlock();
  60                 return;
  61         }
  62
  63         /* Netiher the bpf_prog nor the bpf-map's syscall
  64          * could be modifying the sk_storage->list now.
  65          * Thus, no elem can be added-to or deleted-from the
  66          * sk_storage->list by the bpf_prog or by the bpf-map's syscall.
  67          *
  68          * It is racing with bpf_local_storage_map_free() alone
  69          * when unlinking elem from the sk_storage->list and
  70          * the map's bucket->list.
  71          */
  72         raw_spin_lock_bh(&sk_storage->lock);
  73         hlist_for_each_entry_safe(selem, n, &sk_storage->list, snode) {
  74                 /* Always unlink from map before unlinking from
  75                  * sk_storage.
  76                  */
  77                 bpf_selem_unlink_map(selem);
  78                 free_sk_storage = bpf_selem_unlink_storage_nolock(
  79                         sk_storage, selem, true, false);
  80         }
  81         raw_spin_unlock_bh(&sk_storage->lock);
  82         rcu_read_unlock();
  83
  84         if (free_sk_storage)
  85                 kfree_rcu(sk_storage, rcu);
  86 }
  87
  88 static void bpf_sk_storage_map_free(struct bpf_map *map)
  89 {
  90         struct bpf_local_storage_map *smap;
  91
  92         smap = (struct bpf_local_storage_map *)map;
  93         bpf_local_storage_cache_idx_free(&sk_cache, smap->cache_idx);
  94         bpf_local_storage_map_free(smap, NULL);
  95 }
  96
  97 static struct bpf_map *bpf_sk_storage_map_alloc(union bpf_attr *attr)
  98 {
  99         struct bpf_local_storage_map *smap;
 100
 101         smap = bpf_local_storage_map_alloc(attr);
 102         if (IS_ERR(smap))
 103                 return ERR_CAST(smap);
 104
 105         smap->cache_idx = bpf_local_storage_cache_idx_get(&sk_cache);
 106         return &smap->map;
 107 }
 108
 109 static int notsupp_get_next_key(struct bpf_map *map, void *key,
 110                                 void *next_key)
 111 {
 112         return -ENOTSUPP;
 113 }
 114
 115 static void *bpf_fd_sk_storage_lookup_elem(struct bpf_map *map, void *key)
 116 {
 117         struct bpf_local_storage_data *sdata;
 118         struct socket *sock;
 119         int fd, err;
 120
 121         fd = *(int *)key;
 122         sock = sockfd_lookup(fd, &err);
 123         if (sock) {
 124                 sdata = bpf_sk_storage_lookup(sock->sk, map, true);
 125                 sockfd_put(sock);
 126                 return sdata ? sdata->data : NULL;
 127         }
 128
 129         return ERR_PTR(err);
 130 }
 131
 132 static int bpf_fd_sk_storage_update_elem(struct bpf_map *map, void *key,
 133                                          void *value, u64 map_flags)
 134 {
 135         struct bpf_local_storage_data *sdata;
 136         struct socket *sock;
 137         int fd, err;
 138
 139         fd = *(int *)key;
 140         sock = sockfd_lookup(fd, &err);
 141         if (sock) {
 142                 sdata = bpf_local_storage_update(
 143                         sock->sk, (struct bpf_local_storage_map *)map, value,
 144                         map_flags, GFP_ATOMIC);
 145                 sockfd_put(sock);
 146                 return PTR_ERR_OR_ZERO(sdata);
 147         }
 148
 149         return err;
 150 }
 151
 152 static int bpf_fd_sk_storage_delete_elem(struct bpf_map *map, void *key)
 153 {
 154         struct socket *sock;
 155         int fd, err;
 156
 157         fd = *(int *)key;
 158         sock = sockfd_lookup(fd, &err);
 159         if (sock) {
 160                 err = bpf_sk_storage_del(sock->sk, map);
 161                 sockfd_put(sock);
 162                 return err;
 163         }
 164
 165         return err;
 166 }
 167
 168 static struct bpf_local_storage_elem *
 169 bpf_sk_storage_clone_elem(struct sock *newsk,
 170                           struct bpf_local_storage_map *smap,
 171                           struct bpf_local_storage_elem *selem)
 172 {
 173         struct bpf_local_storage_elem *copy_selem;
 174
 175         copy_selem = bpf_selem_alloc(smap, newsk, NULL, true, GFP_ATOMIC);
 176         if (!copy_selem)
 177                 return NULL;
 178
 179         if (map_value_has_spin_lock(&smap->map))
 180                 copy_map_value_locked(&smap->map, SDATA(copy_selem)->data,
 181                                       SDATA(selem)->data, true);
 182         else
 183                 copy_map_value(&smap->map, SDATA(copy_selem)->data,
 184                                SDATA(selem)->data);
 185
 186         return copy_selem;
 187 }
 188
 189 int bpf_sk_storage_clone(const struct sock *sk, struct sock *newsk)
 190 {
 191         struct bpf_local_storage *new_sk_storage = NULL;
 192         struct bpf_local_storage *sk_storage;
 193         struct bpf_local_storage_elem *selem;
 194         int ret = 0;
 195
 196         RCU_INIT_POINTER(newsk->sk_bpf_storage, NULL);
 197
 198         rcu_read_lock();
 199         sk_storage = rcu_dereference(sk->sk_bpf_storage);
 200
 201         if (!sk_storage || hlist_empty(&sk_storage->list))
 202                 goto out;
 203
 204         hlist_for_each_entry_rcu(selem, &sk_storage->list, snode) {
 205                 struct bpf_local_storage_elem *copy_selem;
 206                 struct bpf_local_storage_map *smap;
 207                 struct bpf_map *map;
 208
 209                 smap = rcu_dereference(SDATA(selem)->smap);
 210                 if (!(smap->map.map_flags & BPF_F_CLONE))
 211                         continue;
 212
 213                 /* Note that for lockless listeners adding new element
 214                  * here can race with cleanup in bpf_local_storage_map_free.
 215                  * Try to grab map refcnt to make sure that it's still
 216                  * alive and prevent concurrent removal.
 217                  */
 218                 map = bpf_map_inc_not_zero(&smap->map);
 219                 if (IS_ERR(map))
 220                         continue;
 221
 222                 copy_selem = bpf_sk_storage_clone_elem(newsk, smap, selem);
 223                 if (!copy_selem) {
 224                         ret = -ENOMEM;
 225                         bpf_map_put(map);
 226                         goto out;
 227                 }
 228
 229                 if (new_sk_storage) {
 230                         bpf_selem_link_map(smap, copy_selem);
 231                         bpf_selem_link_storage_nolock(new_sk_storage, copy_selem);
 232                 } else {
 233                         ret = bpf_local_storage_alloc(newsk, smap, copy_selem, GFP_ATOMIC);
 234                         if (ret) {
 235                                 kfree(copy_selem);
 236                                 atomic_sub(smap->elem_size,
 237                                            &newsk->sk_omem_alloc);
 238                                 bpf_map_put(map);
 239                                 goto out;
 240                         }
 241
 242                         new_sk_storage =
 243                                 rcu_dereference(copy_selem->local_storage);
 244                 }
 245                 bpf_map_put(map);
 246         }
 247
 248 out:
 249         rcu_read_unlock();
 250
 251         /* In case of an error, don't free anything explicitly here, the
 252          * caller is responsible to call bpf_sk_storage_free.
 253          */
 254
 255         return ret;
 256 }
 257
 258 /* *gfp_flags* is a hidden argument provided by the verifier */
 259 BPF_CALL_5(bpf_sk_storage_get, struct bpf_map *, map, struct sock *, sk,
 260            void *, value, u64, flags, gfp_t, gfp_flags)
 261 {
 262         struct bpf_local_storage_data *sdata;
 263
 264         WARN_ON_ONCE(!bpf_rcu_lock_held());
 265         if (!sk || !sk_fullsock(sk) || flags > BPF_SK_STORAGE_GET_F_CREATE)
 266                 return (unsigned long)NULL;
 267
 268         sdata = bpf_sk_storage_lookup(sk, map, true);
 269         if (sdata)
 270                 return (unsigned long)sdata->data;
 271
 272         if (flags == BPF_SK_STORAGE_GET_F_CREATE &&
 273             /* Cannot add new elem to a going away sk.
 274              * Otherwise, the new elem may become a leak
 275              * (and also other memory issues during map
 276              *  destruction).
 277              */
 278             refcount_inc_not_zero(&sk->sk_refcnt)) {
 279                 sdata = bpf_local_storage_update(
 280                         sk, (struct bpf_local_storage_map *)map, value,
 281                         BPF_NOEXIST, gfp_flags);
 282                 /* sk must be a fullsock (guaranteed by verifier),
 283                  * so sock_gen_put() is unnecessary.
 284                  */
 285                 sock_put(sk);
 286                 return IS_ERR(sdata) ?
 287                         (unsigned long)NULL : (unsigned long)sdata->data;
 288         }
 289
 290         return (unsigned long)NULL;
 291 }
 292
 293 BPF_CALL_2(bpf_sk_storage_delete, struct bpf_map *, map, struct sock *, sk)
 294 {
 295         WARN_ON_ONCE(!bpf_rcu_lock_held());
 296         if (!sk || !sk_fullsock(sk))
 297                 return -EINVAL;
 298
 299         if (refcount_inc_not_zero(&sk->sk_refcnt)) {
 300                 int err;
 301
 302                 err = bpf_sk_storage_del(sk, map);
 303                 sock_put(sk);
 304                 return err;
 305         }
 306
 307         return -ENOENT;
 308 }
 309
 310 static int bpf_sk_storage_charge(struct bpf_local_storage_map *smap,
 311                                  void *owner, u32 size)
 312 {
 313         struct sock *sk = (struct sock *)owner;
 314
 315         /* same check as in sock_kmalloc() */
 316         if (size <= sysctl_optmem_max &&
 317             atomic_read(&sk->sk_omem_alloc) + size < sysctl_optmem_max) {
 318                 atomic_add(size, &sk->sk_omem_alloc);
 319                 return 0;
 320         }
 321
 322         return -ENOMEM;
 323 }
 324
 325 static void bpf_sk_storage_uncharge(struct bpf_local_storage_map *smap,
 326                                     void *owner, u32 size)
 327 {
 328         struct sock *sk = owner;
 329
 330         atomic_sub(size, &sk->sk_omem_alloc);
 331 }
 332
 333 static struct bpf_local_storage __rcu **
 334 bpf_sk_storage_ptr(void *owner)
 335 {
 336         struct sock *sk = owner;
 337
 338         return &sk->sk_bpf_storage;
 339 }
 340
 341 BTF_ID_LIST_SINGLE(sk_storage_map_btf_ids, struct, bpf_local_storage_map)
 342 const struct bpf_map_ops sk_storage_map_ops = {
 343         .map_meta_equal = bpf_map_meta_equal,
 344         .map_alloc_check = bpf_local_storage_map_alloc_check,
 345         .map_alloc = bpf_sk_storage_map_alloc,
 346         .map_free = bpf_sk_storage_map_free,
 347         .map_get_next_key = notsupp_get_next_key,
 348         .map_lookup_elem = bpf_fd_sk_storage_lookup_elem,
 349         .map_update_elem = bpf_fd_sk_storage_update_elem,
 350         .map_delete_elem = bpf_fd_sk_storage_delete_elem,
 351         .map_check_btf = bpf_local_storage_map_check_btf,
 352         .map_btf_id = &sk_storage_map_btf_ids[0],
 353         .map_local_storage_charge = bpf_sk_storage_charge,
 354         .map_local_storage_uncharge = bpf_sk_storage_uncharge,
 355         .map_owner_storage_ptr = bpf_sk_storage_ptr,
 356 };
 357
 358 const struct bpf_func_proto bpf_sk_storage_get_proto = {
 359         .func           = bpf_sk_storage_get,
 360         .gpl_only       = false,
 361         .ret_type       = RET_PTR_TO_MAP_VALUE_OR_NULL,
 362         .arg1_type      = ARG_CONST_MAP_PTR,
 363         .arg2_type      = ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 364         .arg3_type      = ARG_PTR_TO_MAP_VALUE_OR_NULL,
 365         .arg4_type      = ARG_ANYTHING,
 366 };
 367
 368 const struct bpf_func_proto bpf_sk_storage_get_cg_sock_proto = {
 369         .func           = bpf_sk_storage_get,
 370         .gpl_only       = false,
 371         .ret_type       = RET_PTR_TO_MAP_VALUE_OR_NULL,
 372         .arg1_type      = ARG_CONST_MAP_PTR,
 373         .arg2_type      = ARG_PTR_TO_CTX, /* context is 'struct sock' */
 374         .arg3_type      = ARG_PTR_TO_MAP_VALUE_OR_NULL,
 375         .arg4_type      = ARG_ANYTHING,
 376 };
 377
 378 const struct bpf_func_proto bpf_sk_storage_delete_proto = {
 379         .func           = bpf_sk_storage_delete,
 380         .gpl_only       = false,
 381         .ret_type       = RET_INTEGER,
 382         .arg1_type      = ARG_CONST_MAP_PTR,
 383         .arg2_type      = ARG_PTR_TO_BTF_ID_SOCK_COMMON,
 384 };
 385
 386 static bool bpf_sk_storage_tracing_allowed(const struct bpf_prog *prog)
 387 {
 388         const struct btf *btf_vmlinux;
 389         const struct btf_type *t;
 390         const char *tname;
 391         u32 btf_id;
 392
 393         if (prog->aux->dst_prog)
 394                 return false;
 395
 396         /* Ensure the tracing program is not tracing
 397          * any bpf_sk_storage*() function and also
 398          * use the bpf_sk_storage_(get|delete) helper.
 399          */
 400         switch (prog->expected_attach_type) {
 401         case BPF_TRACE_ITER:
 402         case BPF_TRACE_RAW_TP:
 403                 /* bpf_sk_storage has no trace point */
 404                 return true;
 405         case BPF_TRACE_FENTRY:
 406         case BPF_TRACE_FEXIT:
 407                 btf_vmlinux = bpf_get_btf_vmlinux();
 408                 if (IS_ERR_OR_NULL(btf_vmlinux))
 409                         return false;
 410                 btf_id = prog->aux->attach_btf_id;
 411                 t = btf_type_by_id(btf_vmlinux, btf_id);
 412                 tname = btf_name_by_offset(btf_vmlinux, t->name_off);
 413                 return !!strncmp(tname, "bpf_sk_storage",
 414                                  strlen("bpf_sk_storage"));
 415         default:
 416                 return false;
 417         }
 418
 419         return false;
 420 }
 421
 422 /* *gfp_flags* is a hidden argument provided by the verifier */
 423 BPF_CALL_5(bpf_sk_storage_get_tracing, struct bpf_map *, map, struct sock *, sk,
 424            void *, value, u64, flags, gfp_t, gfp_flags)
 425 {
 426         WARN_ON_ONCE(!bpf_rcu_lock_held());
 427         if (in_hardirq() || in_nmi())
 428                 return (unsigned long)NULL;
 429
 430         return (unsigned long)____bpf_sk_storage_get(map, sk, value, flags,
 431                                                      gfp_flags);
 432 }
 433
 434 BPF_CALL_2(bpf_sk_storage_delete_tracing, struct bpf_map *, map,
 435            struct sock *, sk)
 436 {
 437         WARN_ON_ONCE(!bpf_rcu_lock_held());
 438         if (in_hardirq() || in_nmi())
 439                 return -EPERM;
 440
 441         return ____bpf_sk_storage_delete(map, sk);
 442 }
 443
 444 const struct bpf_func_proto bpf_sk_storage_get_tracing_proto = {
 445         .func           = bpf_sk_storage_get_tracing,
 446         .gpl_only       = false,
 447         .ret_type       = RET_PTR_TO_MAP_VALUE_OR_NULL,
 448         .arg1_type      = ARG_CONST_MAP_PTR,
 449         .arg2_type      = ARG_PTR_TO_BTF_ID,
 450         .arg2_btf_id    = &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
 451         .arg3_type      = ARG_PTR_TO_MAP_VALUE_OR_NULL,
 452         .arg4_type      = ARG_ANYTHING,
 453         .allowed        = bpf_sk_storage_tracing_allowed,
 454 };
 455
 456 const struct bpf_func_proto bpf_sk_storage_delete_tracing_proto = {
 457         .func           = bpf_sk_storage_delete_tracing,
 458         .gpl_only       = false,
 459         .ret_type       = RET_INTEGER,
 460         .arg1_type      = ARG_CONST_MAP_PTR,
 461         .arg2_type      = ARG_PTR_TO_BTF_ID,
 462         .arg2_btf_id    = &btf_sock_ids[BTF_SOCK_TYPE_SOCK_COMMON],
 463         .allowed        = bpf_sk_storage_tracing_allowed,
 464 };
 465
 466 struct bpf_sk_storage_diag {
 467         u32 nr_maps;
 468         struct bpf_map *maps[];
 469 };
 470
 471 /* The reply will be like:
 472  * INET_DIAG_BPF_SK_STORAGES (nla_nest)
 473  *      SK_DIAG_BPF_STORAGE (nla_nest)
 474  *              SK_DIAG_BPF_STORAGE_MAP_ID (nla_put_u32)
 475  *              SK_DIAG_BPF_STORAGE_MAP_VALUE (nla_reserve_64bit)
 476  *      SK_DIAG_BPF_STORAGE (nla_nest)
 477  *              SK_DIAG_BPF_STORAGE_MAP_ID (nla_put_u32)
 478  *              SK_DIAG_BPF_STORAGE_MAP_VALUE (nla_reserve_64bit)
 479  *      ....
 480  */
 481 static int nla_value_size(u32 value_size)
 482 {
 483         /* SK_DIAG_BPF_STORAGE (nla_nest)
 484          *      SK_DIAG_BPF_STORAGE_MAP_ID (nla_put_u32)
 485          *      SK_DIAG_BPF_STORAGE_MAP_VALUE (nla_reserve_64bit)
 486          */
 487         return nla_total_size(0) + nla_total_size(sizeof(u32)) +
 488                 nla_total_size_64bit(value_size);
 489 }
 490
 491 void bpf_sk_storage_diag_free(struct bpf_sk_storage_diag *diag)
 492 {
 493         u32 i;
 494
 495         if (!diag)
 496                 return;
 497
 498         for (i = 0; i < diag->nr_maps; i++)
 499                 bpf_map_put(diag->maps[i]);
 500
 501         kfree(diag);
 502 }
 503 EXPORT_SYMBOL_GPL(bpf_sk_storage_diag_free);
 504
 505 static bool diag_check_dup(const struct bpf_sk_storage_diag *diag,
 506                            const struct bpf_map *map)
 507 {
 508         u32 i;
 509
 510         for (i = 0; i < diag->nr_maps; i++) {
 511                 if (diag->maps[i] == map)
 512                         return true;
 513         }
 514
 515         return false;
 516 }
 517
 518 struct bpf_sk_storage_diag *
 519 bpf_sk_storage_diag_alloc(const struct nlattr *nla_stgs)
 520 {
 521         struct bpf_sk_storage_diag *diag;
 522         struct nlattr *nla;
 523         u32 nr_maps = 0;
 524         int rem, err;
 525
 526         /* bpf_local_storage_map is currently limited to CAP_SYS_ADMIN as
 527          * the map_alloc_check() side also does.
 528          */
 529         if (!bpf_capable())
 530                 return ERR_PTR(-EPERM);
 531
 532         nla_for_each_nested(nla, nla_stgs, rem) {
 533                 if (nla_type(nla) == SK_DIAG_BPF_STORAGE_REQ_MAP_FD)
 534                         nr_maps++;
 535         }
 536
 537         diag = kzalloc(struct_size(diag, maps, nr_maps), GFP_KERNEL);
 538         if (!diag)
 539                 return ERR_PTR(-ENOMEM);
 540
 541         nla_for_each_nested(nla, nla_stgs, rem) {
 542                 struct bpf_map *map;
 543                 int map_fd;
 544
 545                 if (nla_type(nla) != SK_DIAG_BPF_STORAGE_REQ_MAP_FD)
 546                         continue;
 547
 548                 map_fd = nla_get_u32(nla);
 549                 map = bpf_map_get(map_fd);
 550                 if (IS_ERR(map)) {
 551                         err = PTR_ERR(map);
 552                         goto err_free;
 553                 }
 554                 if (map->map_type != BPF_MAP_TYPE_SK_STORAGE) {
 555                         bpf_map_put(map);
 556                         err = -EINVAL;
 557                         goto err_free;
 558                 }
 559                 if (diag_check_dup(diag, map)) {
 560                         bpf_map_put(map);
 561                         err = -EEXIST;
 562                         goto err_free;
 563                 }
 564                 diag->maps[diag->nr_maps++] = map;
 565         }
 566
 567         return diag;
 568
 569 err_free:
 570         bpf_sk_storage_diag_free(diag);
 571         return ERR_PTR(err);
 572 }
 573 EXPORT_SYMBOL_GPL(bpf_sk_storage_diag_alloc);
 574
 575 static int diag_get(struct bpf_local_storage_data *sdata, struct sk_buff *skb)
 576 {
 577         struct nlattr *nla_stg, *nla_value;
 578         struct bpf_local_storage_map *smap;
 579
 580         /* It cannot exceed max nlattr's payload */
 581         BUILD_BUG_ON(U16_MAX - NLA_HDRLEN < BPF_LOCAL_STORAGE_MAX_VALUE_SIZE);
 582
 583         nla_stg = nla_nest_start(skb, SK_DIAG_BPF_STORAGE);
 584         if (!nla_stg)
 585                 return -EMSGSIZE;
 586
 587         smap = rcu_dereference(sdata->smap);
 588         if (nla_put_u32(skb, SK_DIAG_BPF_STORAGE_MAP_ID, smap->map.id))
 589                 goto errout;
 590
 591         nla_value = nla_reserve_64bit(skb, SK_DIAG_BPF_STORAGE_MAP_VALUE,
 592                                       smap->map.value_size,
 593                                       SK_DIAG_BPF_STORAGE_PAD);
 594         if (!nla_value)
 595                 goto errout;
 596
 597         if (map_value_has_spin_lock(&smap->map))
 598                 copy_map_value_locked(&smap->map, nla_data(nla_value),
 599                                       sdata->data, true);
 600         else
 601                 copy_map_value(&smap->map, nla_data(nla_value), sdata->data);
 602
 603         nla_nest_end(skb, nla_stg);
 604         return 0;
 605
 606 errout:
 607         nla_nest_cancel(skb, nla_stg);
 608         return -EMSGSIZE;
 609 }
 610
 611 static int bpf_sk_storage_diag_put_all(struct sock *sk, struct sk_buff *skb,
 612                                        int stg_array_type,
 613                                        unsigned int *res_diag_size)
 614 {
 615         /* stg_array_type (e.g. INET_DIAG_BPF_SK_STORAGES) */
 616         unsigned int diag_size = nla_total_size(0);
 617         struct bpf_local_storage *sk_storage;
 618         struct bpf_local_storage_elem *selem;
 619         struct bpf_local_storage_map *smap;
 620         struct nlattr *nla_stgs;
 621         unsigned int saved_len;
 622         int err = 0;
 623
 624         rcu_read_lock();
 625
 626         sk_storage = rcu_dereference(sk->sk_bpf_storage);
 627         if (!sk_storage || hlist_empty(&sk_storage->list)) {
 628                 rcu_read_unlock();
 629                 return 0;
 630         }
 631
 632         nla_stgs = nla_nest_start(skb, stg_array_type);
 633         if (!nla_stgs)
 634                 /* Continue to learn diag_size */
 635                 err = -EMSGSIZE;
 636
 637         saved_len = skb->len;
 638         hlist_for_each_entry_rcu(selem, &sk_storage->list, snode) {
 639                 smap = rcu_dereference(SDATA(selem)->smap);
 640                 diag_size += nla_value_size(smap->map.value_size);
 641
 642                 if (nla_stgs && diag_get(SDATA(selem), skb))
 643                         /* Continue to learn diag_size */
 644                         err = -EMSGSIZE;
 645         }
 646
 647         rcu_read_unlock();
 648
 649         if (nla_stgs) {
 650                 if (saved_len == skb->len)
 651                         nla_nest_cancel(skb, nla_stgs);
 652                 else
 653                         nla_nest_end(skb, nla_stgs);
 654         }
 655
 656         if (diag_size == nla_total_size(0)) {
 657                 *res_diag_size = 0;
 658                 return 0;
 659         }
 660
 661         *res_diag_size = diag_size;
 662         return err;
 663 }
 664
 665 int bpf_sk_storage_diag_put(struct bpf_sk_storage_diag *diag,
 666                             struct sock *sk, struct sk_buff *skb,
 667                             int stg_array_type,
 668                             unsigned int *res_diag_size)
 669 {
 670         /* stg_array_type (e.g. INET_DIAG_BPF_SK_STORAGES) */
 671         unsigned int diag_size = nla_total_size(0);
 672         struct bpf_local_storage *sk_storage;
 673         struct bpf_local_storage_data *sdata;
 674         struct nlattr *nla_stgs;
 675         unsigned int saved_len;
 676         int err = 0;
 677         u32 i;
 678
 679         *res_diag_size = 0;
 680
 681         /* No map has been specified.  Dump all. */
 682         if (!diag->nr_maps)
 683                 return bpf_sk_storage_diag_put_all(sk, skb, stg_array_type,
 684                                                    res_diag_size);
 685
 686         rcu_read_lock();
 687         sk_storage = rcu_dereference(sk->sk_bpf_storage);
 688         if (!sk_storage || hlist_empty(&sk_storage->list)) {
 689                 rcu_read_unlock();
 690                 return 0;
 691         }
 692
 693         nla_stgs = nla_nest_start(skb, stg_array_type);
 694         if (!nla_stgs)
 695                 /* Continue to learn diag_size */
 696                 err = -EMSGSIZE;
 697
 698         saved_len = skb->len;
 699         for (i = 0; i < diag->nr_maps; i++) {
 700                 sdata = bpf_local_storage_lookup(sk_storage,
 701                                 (struct bpf_local_storage_map *)diag->maps[i],
 702                                 false);
 703
 704                 if (!sdata)
 705                         continue;
 706
 707                 diag_size += nla_value_size(diag->maps[i]->value_size);
 708
 709                 if (nla_stgs && diag_get(sdata, skb))
 710                         /* Continue to learn diag_size */
 711                         err = -EMSGSIZE;
 712         }
 713         rcu_read_unlock();
 714
 715         if (nla_stgs) {
 716                 if (saved_len == skb->len)
 717                         nla_nest_cancel(skb, nla_stgs);
 718                 else
 719                         nla_nest_end(skb, nla_stgs);
 720         }
 721
 722         if (diag_size == nla_total_size(0)) {
 723                 *res_diag_size = 0;
 724                 return 0;
 725         }
 726
 727         *res_diag_size = diag_size;
 728         return err;
 729 }
 730 EXPORT_SYMBOL_GPL(bpf_sk_storage_diag_put);
 731
 732 struct bpf_iter_seq_sk_storage_map_info {
 733         struct bpf_map *map;
 734         unsigned int bucket_id;
 735         unsigned skip_elems;
 736 };
 737
 738 static struct bpf_local_storage_elem *
 739 bpf_sk_storage_map_seq_find_next(struct bpf_iter_seq_sk_storage_map_info *info,
 740                                  struct bpf_local_storage_elem *prev_selem)
 741         __acquires(RCU) __releases(RCU)
 742 {
 743         struct bpf_local_storage *sk_storage;
 744         struct bpf_local_storage_elem *selem;
 745         u32 skip_elems = info->skip_elems;
 746         struct bpf_local_storage_map *smap;
 747         u32 bucket_id = info->bucket_id;
 748         u32 i, count, n_buckets;
 749         struct bpf_local_storage_map_bucket *b;
 750
 751         smap = (struct bpf_local_storage_map *)info->map;
 752         n_buckets = 1U << smap->bucket_log;
 753         if (bucket_id >= n_buckets)
 754                 return NULL;
 755
 756         /* try to find next selem in the same bucket */
 757         selem = prev_selem;
 758         count = 0;
 759         while (selem) {
 760                 selem = hlist_entry_safe(rcu_dereference(hlist_next_rcu(&selem->map_node)),
 761                                          struct bpf_local_storage_elem, map_node);
 762                 if (!selem) {
 763                         /* not found, unlock and go to the next bucket */
 764                         b = &smap->buckets[bucket_id++];
 765                         rcu_read_unlock();
 766                         skip_elems = 0;
 767                         break;
 768                 }
 769                 sk_storage = rcu_dereference(selem->local_storage);
 770                 if (sk_storage) {
 771                         info->skip_elems = skip_elems + count;
 772                         return selem;
 773                 }
 774                 count++;
 775         }
 776
 777         for (i = bucket_id; i < (1U << smap->bucket_log); i++) {
 778                 b = &smap->buckets[i];
 779                 rcu_read_lock();
 780                 count = 0;
 781                 hlist_for_each_entry_rcu(selem, &b->list, map_node) {
 782                         sk_storage = rcu_dereference(selem->local_storage);
 783                         if (sk_storage && count >= skip_elems) {
 784                                 info->bucket_id = i;
 785                                 info->skip_elems = count;
 786                                 return selem;
 787                         }
 788                         count++;
 789                 }
 790                 rcu_read_unlock();
 791                 skip_elems = 0;
 792         }
 793
 794         info->bucket_id = i;
 795         info->skip_elems = 0;
 796         return NULL;
 797 }
 798
 799 static void *bpf_sk_storage_map_seq_start(struct seq_file *seq, loff_t *pos)
 800 {
 801         struct bpf_local_storage_elem *selem;
 802
 803         selem = bpf_sk_storage_map_seq_find_next(seq->private, NULL);
 804         if (!selem)
 805                 return NULL;
 806
 807         if (*pos == 0)
 808                 ++*pos;
 809         return selem;
 810 }
 811
 812 static void *bpf_sk_storage_map_seq_next(struct seq_file *seq, void *v,
 813                                          loff_t *pos)
 814 {
 815         struct bpf_iter_seq_sk_storage_map_info *info = seq->private;
 816
 817         ++*pos;
 818         ++info->skip_elems;
 819         return bpf_sk_storage_map_seq_find_next(seq->private, v);
 820 }
 821
 822 struct bpf_iter__bpf_sk_storage_map {
 823         __bpf_md_ptr(struct bpf_iter_meta *, meta);
 824         __bpf_md_ptr(struct bpf_map *, map);
 825         __bpf_md_ptr(struct sock *, sk);
 826         __bpf_md_ptr(void *, value);
 827 };
 828
 829 DEFINE_BPF_ITER_FUNC(bpf_sk_storage_map, struct bpf_iter_meta *meta,
 830                      struct bpf_map *map, struct sock *sk,
 831                      void *value)
 832
 833 static int __bpf_sk_storage_map_seq_show(struct seq_file *seq,
 834                                          struct bpf_local_storage_elem *selem)
 835 {
 836         struct bpf_iter_seq_sk_storage_map_info *info = seq->private;
 837         struct bpf_iter__bpf_sk_storage_map ctx = {};
 838         struct bpf_local_storage *sk_storage;
 839         struct bpf_iter_meta meta;
 840         struct bpf_prog *prog;
 841         int ret = 0;
 842
 843         meta.seq = seq;
 844         prog = bpf_iter_get_info(&meta, selem == NULL);
 845         if (prog) {
 846                 ctx.meta = &meta;
 847                 ctx.map = info->map;
 848                 if (selem) {
 849                         sk_storage = rcu_dereference(selem->local_storage);
 850                         ctx.sk = sk_storage->owner;
 851                         ctx.value = SDATA(selem)->data;
 852                 }
 853                 ret = bpf_iter_run_prog(prog, &ctx);
 854         }
 855
 856         return ret;
 857 }
 858
 859 static int bpf_sk_storage_map_seq_show(struct seq_file *seq, void *v)
 860 {
 861         return __bpf_sk_storage_map_seq_show(seq, v);
 862 }
 863
 864 static void bpf_sk_storage_map_seq_stop(struct seq_file *seq, void *v)
 865         __releases(RCU)
 866 {
 867         if (!v)
 868                 (void)__bpf_sk_storage_map_seq_show(seq, v);
 869         else
 870                 rcu_read_unlock();
 871 }
 872
 873 static int bpf_iter_init_sk_storage_map(void *priv_data,
 874                                         struct bpf_iter_aux_info *aux)
 875 {
 876         struct bpf_iter_seq_sk_storage_map_info *seq_info = priv_data;
 877
 878         seq_info->map = aux->map;
 879         return 0;
 880 }
 881
 882 static int bpf_iter_attach_map(struct bpf_prog *prog,
 883                                union bpf_iter_link_info *linfo,
 884                                struct bpf_iter_aux_info *aux)
 885 {
 886         struct bpf_map *map;
 887         int err = -EINVAL;
 888
 889         if (!linfo->map.map_fd)
 890                 return -EBADF;
 891
 892         map = bpf_map_get_with_uref(linfo->map.map_fd);
 893         if (IS_ERR(map))
 894                 return PTR_ERR(map);
 895
 896         if (map->map_type != BPF_MAP_TYPE_SK_STORAGE)
 897                 goto put_map;
 898
 899         if (prog->aux->max_rdonly_access > map->value_size) {
 900                 err = -EACCES;
 901                 goto put_map;
 902         }
 903
 904         aux->map = map;
 905         return 0;
 906
 907 put_map:
 908         bpf_map_put_with_uref(map);
 909         return err;
 910 }
 911
 912 static void bpf_iter_detach_map(struct bpf_iter_aux_info *aux)
 913 {
 914         bpf_map_put_with_uref(aux->map);
 915 }
 916
 917 static const struct seq_operations bpf_sk_storage_map_seq_ops = {
 918         .start  = bpf_sk_storage_map_seq_start,
 919         .next   = bpf_sk_storage_map_seq_next,
 920         .stop   = bpf_sk_storage_map_seq_stop,
 921         .show   = bpf_sk_storage_map_seq_show,
 922 };
 923
 924 static const struct bpf_iter_seq_info iter_seq_info = {
 925         .seq_ops                = &bpf_sk_storage_map_seq_ops,
 926         .init_seq_private       = bpf_iter_init_sk_storage_map,
 927         .fini_seq_private       = NULL,
 928         .seq_priv_size          = sizeof(struct bpf_iter_seq_sk_storage_map_info),
 929 };
 930
 931 static struct bpf_iter_reg bpf_sk_storage_map_reg_info = {
 932         .target                 = "bpf_sk_storage_map",
 933         .attach_target          = bpf_iter_attach_map,
 934         .detach_target          = bpf_iter_detach_map,
 935         .show_fdinfo            = bpf_iter_map_show_fdinfo,
 936         .fill_link_info         = bpf_iter_map_fill_link_info,
 937         .ctx_arg_info_size      = 2,
 938         .ctx_arg_info           = {
 939                 { offsetof(struct bpf_iter__bpf_sk_storage_map, sk),
 940                   PTR_TO_BTF_ID_OR_NULL },
 941                 { offsetof(struct bpf_iter__bpf_sk_storage_map, value),
 942                   PTR_TO_BUF | PTR_MAYBE_NULL },
 943         },
 944         .seq_info               = &iter_seq_info,
 945 };
 946
 947 static int __init bpf_sk_storage_map_iter_init(void)
 948 {
 949         bpf_sk_storage_map_reg_info.ctx_arg_info[0].btf_id =
 950                 btf_sock_ids[BTF_SOCK_TYPE_SOCK];
 951         return bpf_iter_reg_target(&bpf_sk_storage_map_reg_info);
 952 }
 953 late_initcall(bpf_sk_storage_map_iter_init);