kernel/bpf/hashtab.c

   1 // SPDX-License-Identifier: GPL-2.0-only
   2 /* Copyright (c) 2011-2014 PLUMgrid, http://plumgrid.com
   3  * Copyright (c) 2016 Facebook
   4  */
   5 #include <linux/bpf.h>
   6 #include <linux/btf.h>
   7 #include <linux/jhash.h>
   8 #include <linux/filter.h>
   9 #include <linux/rculist_nulls.h>
  10 #include <linux/random.h>
  11 #include <uapi/linux/btf.h>
  12 #include <linux/rcupdate_trace.h>
  13 #include "percpu_freelist.h"
  14 #include "bpf_lru_list.h"
  15 #include "map_in_map.h"
  16
  17 #define HTAB_CREATE_FLAG_MASK                                           \
  18         (BPF_F_NO_PREALLOC | BPF_F_NO_COMMON_LRU | BPF_F_NUMA_NODE |    \
  19          BPF_F_ACCESS_MASK | BPF_F_ZERO_SEED)
  20
  21 #define BATCH_OPS(_name)                        \
  22         .map_lookup_batch =                     \
  23         _name##_map_lookup_batch,               \
  24         .map_lookup_and_delete_batch =          \
  25         _name##_map_lookup_and_delete_batch,    \
  26         .map_update_batch =                     \
  27         generic_map_update_batch,               \
  28         .map_delete_batch =                     \
  29         generic_map_delete_batch
  30
  31 /*
  32  * The bucket lock has two protection scopes:
  33  *
  34  * 1) Serializing concurrent operations from BPF programs on differrent
  35  *    CPUs
  36  *
  37  * 2) Serializing concurrent operations from BPF programs and sys_bpf()
  38  *
  39  * BPF programs can execute in any context including perf, kprobes and
  40  * tracing. As there are almost no limits where perf, kprobes and tracing
  41  * can be invoked from the lock operations need to be protected against
  42  * deadlocks. Deadlocks can be caused by recursion and by an invocation in
  43  * the lock held section when functions which acquire this lock are invoked
  44  * from sys_bpf(). BPF recursion is prevented by incrementing the per CPU
  45  * variable bpf_prog_active, which prevents BPF programs attached to perf
  46  * events, kprobes and tracing to be invoked before the prior invocation
  47  * from one of these contexts completed. sys_bpf() uses the same mechanism
  48  * by pinning the task to the current CPU and incrementing the recursion
  49  * protection accross the map operation.
  50  *
  51  * This has subtle implications on PREEMPT_RT. PREEMPT_RT forbids certain
  52  * operations like memory allocations (even with GFP_ATOMIC) from atomic
  53  * contexts. This is required because even with GFP_ATOMIC the memory
  54  * allocator calls into code pathes which acquire locks with long held lock
  55  * sections. To ensure the deterministic behaviour these locks are regular
  56  * spinlocks, which are converted to 'sleepable' spinlocks on RT. The only
  57  * true atomic contexts on an RT kernel are the low level hardware
  58  * handling, scheduling, low level interrupt handling, NMIs etc. None of
  59  * these contexts should ever do memory allocations.
  60  *
  61  * As regular device interrupt handlers and soft interrupts are forced into
  62  * thread context, the existing code which does
  63  *   spin_lock*(); alloc(GPF_ATOMIC); spin_unlock*();
  64  * just works.
  65  *
  66  * In theory the BPF locks could be converted to regular spinlocks as well,
  67  * but the bucket locks and percpu_freelist locks can be taken from
  68  * arbitrary contexts (perf, kprobes, tracepoints) which are required to be
  69  * atomic contexts even on RT. These mechanisms require preallocated maps,
  70  * so there is no need to invoke memory allocations within the lock held
  71  * sections.
  72  *
  73  * BPF maps which need dynamic allocation are only used from (forced)
  74  * thread context on RT and can therefore use regular spinlocks which in
  75  * turn allows to invoke memory allocations from the lock held section.
  76  *
  77  * On a non RT kernel this distinction is neither possible nor required.
  78  * spinlock maps to raw_spinlock and the extra code is optimized out by the
  79  * compiler.
  80  */
  81 struct bucket {
  82         struct hlist_nulls_head head;
  83         union {
  84                 raw_spinlock_t raw_lock;
  85                 spinlock_t     lock;
  86         };
  87 };
  88
  89 #define HASHTAB_MAP_LOCK_COUNT 8
  90 #define HASHTAB_MAP_LOCK_MASK (HASHTAB_MAP_LOCK_COUNT - 1)
  91
  92 struct bpf_htab {
  93         struct bpf_map map;
  94         struct bucket *buckets;
  95         void *elems;
  96         union {
  97                 struct pcpu_freelist freelist;
  98                 struct bpf_lru lru;
  99         };
 100         struct htab_elem *__percpu *extra_elems;
 101         atomic_t count; /* number of elements in this hashtable */
 102         u32 n_buckets;  /* number of hash buckets */
 103         u32 elem_size;  /* size of each element in bytes */
 104         u32 hashrnd;
 105         struct lock_class_key lockdep_key;
 106         int __percpu *map_locked[HASHTAB_MAP_LOCK_COUNT];
 107 };
 108
 109 /* each htab element is struct htab_elem + key + value */
 110 struct htab_elem {
 111         union {
 112                 struct hlist_nulls_node hash_node;
 113                 struct {
 114                         void *padding;
 115                         union {
 116                                 struct bpf_htab *htab;
 117                                 struct pcpu_freelist_node fnode;
 118                                 struct htab_elem *batch_flink;
 119                         };
 120                 };
 121         };
 122         union {
 123                 struct rcu_head rcu;
 124                 struct bpf_lru_node lru_node;
 125         };
 126         u32 hash;
 127         char key[] __aligned(8);
 128 };
 129
 130 static inline bool htab_is_prealloc(const struct bpf_htab *htab)
 131 {
 132         return !(htab->map.map_flags & BPF_F_NO_PREALLOC);
 133 }
 134
 135 static inline bool htab_use_raw_lock(const struct bpf_htab *htab)
 136 {
 137         return (!IS_ENABLED(CONFIG_PREEMPT_RT) || htab_is_prealloc(htab));
 138 }
 139
 140 static void htab_init_buckets(struct bpf_htab *htab)
 141 {
 142         unsigned i;
 143
 144         for (i = 0; i < htab->n_buckets; i++) {
 145                 INIT_HLIST_NULLS_HEAD(&htab->buckets[i].head, i);
 146                 if (htab_use_raw_lock(htab)) {
 147                         raw_spin_lock_init(&htab->buckets[i].raw_lock);
 148                         lockdep_set_class(&htab->buckets[i].raw_lock,
 149                                           &htab->lockdep_key);
 150                 } else {
 151                         spin_lock_init(&htab->buckets[i].lock);
 152                         lockdep_set_class(&htab->buckets[i].lock,
 153                                           &htab->lockdep_key);
 154                 }
 155         }
 156 }
 157
 158 static inline int htab_lock_bucket(const struct bpf_htab *htab,
 159                                    struct bucket *b, u32 hash,
 160                                    unsigned long *pflags)
 161 {
 162         unsigned long flags;
 163
 164         hash = hash & HASHTAB_MAP_LOCK_MASK;
 165
 166         migrate_disable();
 167         if (unlikely(__this_cpu_inc_return(*(htab->map_locked[hash])) != 1)) {
 168                 __this_cpu_dec(*(htab->map_locked[hash]));
 169                 migrate_enable();
 170                 return -EBUSY;
 171         }
 172
 173         if (htab_use_raw_lock(htab))
 174                 raw_spin_lock_irqsave(&b->raw_lock, flags);
 175         else
 176                 spin_lock_irqsave(&b->lock, flags);
 177         *pflags = flags;
 178
 179         return 0;
 180 }
 181
 182 static inline void htab_unlock_bucket(const struct bpf_htab *htab,
 183                                       struct bucket *b, u32 hash,
 184                                       unsigned long flags)
 185 {
 186         hash = hash & HASHTAB_MAP_LOCK_MASK;
 187         if (htab_use_raw_lock(htab))
 188                 raw_spin_unlock_irqrestore(&b->raw_lock, flags);
 189         else
 190                 spin_unlock_irqrestore(&b->lock, flags);
 191         __this_cpu_dec(*(htab->map_locked[hash]));
 192         migrate_enable();
 193 }
 194
 195 static bool htab_lru_map_delete_node(void *arg, struct bpf_lru_node *node);
 196
 197 static bool htab_is_lru(const struct bpf_htab *htab)
 198 {
 199         return htab->map.map_type == BPF_MAP_TYPE_LRU_HASH ||
 200                 htab->map.map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH;
 201 }
 202
 203 static bool htab_is_percpu(const struct bpf_htab *htab)
 204 {
 205         return htab->map.map_type == BPF_MAP_TYPE_PERCPU_HASH ||
 206                 htab->map.map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH;
 207 }
 208
 209 static inline void htab_elem_set_ptr(struct htab_elem *l, u32 key_size,
 210                                      void __percpu *pptr)
 211 {
 212         *(void __percpu **)(l->key + key_size) = pptr;
 213 }
 214
 215 static inline void __percpu *htab_elem_get_ptr(struct htab_elem *l, u32 key_size)
 216 {
 217         return *(void __percpu **)(l->key + key_size);
 218 }
 219
 220 static void *fd_htab_map_get_ptr(const struct bpf_map *map, struct htab_elem *l)
 221 {
 222         return *(void **)(l->key + roundup(map->key_size, 8));
 223 }
 224
 225 static struct htab_elem *get_htab_elem(struct bpf_htab *htab, int i)
 226 {
 227         return (struct htab_elem *) (htab->elems + i * htab->elem_size);
 228 }
 229
 230 static void htab_free_elems(struct bpf_htab *htab)
 231 {
 232         int i;
 233
 234         if (!htab_is_percpu(htab))
 235                 goto free_elems;
 236
 237         for (i = 0; i < htab->map.max_entries; i++) {
 238                 void __percpu *pptr;
 239
 240                 pptr = htab_elem_get_ptr(get_htab_elem(htab, i),
 241                                          htab->map.key_size);
 242                 free_percpu(pptr);
 243                 cond_resched();
 244         }
 245 free_elems:
 246         bpf_map_area_free(htab->elems);
 247 }
 248
 249 /* The LRU list has a lock (lru_lock). Each htab bucket has a lock
 250  * (bucket_lock). If both locks need to be acquired together, the lock
 251  * order is always lru_lock -> bucket_lock and this only happens in
 252  * bpf_lru_list.c logic. For example, certain code path of
 253  * bpf_lru_pop_free(), which is called by function prealloc_lru_pop(),
 254  * will acquire lru_lock first followed by acquiring bucket_lock.
 255  *
 256  * In hashtab.c, to avoid deadlock, lock acquisition of
 257  * bucket_lock followed by lru_lock is not allowed. In such cases,
 258  * bucket_lock needs to be released first before acquiring lru_lock.
 259  */
 260 static struct htab_elem *prealloc_lru_pop(struct bpf_htab *htab, void *key,
 261                                           u32 hash)
 262 {
 263         struct bpf_lru_node *node = bpf_lru_pop_free(&htab->lru, hash);
 264         struct htab_elem *l;
 265
 266         if (node) {
 267                 l = container_of(node, struct htab_elem, lru_node);
 268                 memcpy(l->key, key, htab->map.key_size);
 269                 return l;
 270         }
 271
 272         return NULL;
 273 }
 274
 275 static int prealloc_init(struct bpf_htab *htab)
 276 {
 277         u32 num_entries = htab->map.max_entries;
 278         int err = -ENOMEM, i;
 279
 280         if (!htab_is_percpu(htab) && !htab_is_lru(htab))
 281                 num_entries += num_possible_cpus();
 282
 283         htab->elems = bpf_map_area_alloc(htab->elem_size * num_entries,
 284                                          htab->map.numa_node);
 285         if (!htab->elems)
 286                 return -ENOMEM;
 287
 288         if (!htab_is_percpu(htab))
 289                 goto skip_percpu_elems;
 290
 291         for (i = 0; i < num_entries; i++) {
 292                 u32 size = round_up(htab->map.value_size, 8);
 293                 void __percpu *pptr;
 294
 295                 pptr = __alloc_percpu_gfp(size, 8, GFP_USER | __GFP_NOWARN);
 296                 if (!pptr)
 297                         goto free_elems;
 298                 htab_elem_set_ptr(get_htab_elem(htab, i), htab->map.key_size,
 299                                   pptr);
 300                 cond_resched();
 301         }
 302
 303 skip_percpu_elems:
 304         if (htab_is_lru(htab))
 305                 err = bpf_lru_init(&htab->lru,
 306                                    htab->map.map_flags & BPF_F_NO_COMMON_LRU,
 307                                    offsetof(struct htab_elem, hash) -
 308                                    offsetof(struct htab_elem, lru_node),
 309                                    htab_lru_map_delete_node,
 310                                    htab);
 311         else
 312                 err = pcpu_freelist_init(&htab->freelist);
 313
 314         if (err)
 315                 goto free_elems;
 316
 317         if (htab_is_lru(htab))
 318                 bpf_lru_populate(&htab->lru, htab->elems,
 319                                  offsetof(struct htab_elem, lru_node),
 320                                  htab->elem_size, num_entries);
 321         else
 322                 pcpu_freelist_populate(&htab->freelist,
 323                                        htab->elems + offsetof(struct htab_elem, fnode),
 324                                        htab->elem_size, num_entries);
 325
 326         return 0;
 327
 328 free_elems:
 329         htab_free_elems(htab);
 330         return err;
 331 }
 332
 333 static void prealloc_destroy(struct bpf_htab *htab)
 334 {
 335         htab_free_elems(htab);
 336
 337         if (htab_is_lru(htab))
 338                 bpf_lru_destroy(&htab->lru);
 339         else
 340                 pcpu_freelist_destroy(&htab->freelist);
 341 }
 342
 343 static int alloc_extra_elems(struct bpf_htab *htab)
 344 {
 345         struct htab_elem *__percpu *pptr, *l_new;
 346         struct pcpu_freelist_node *l;
 347         int cpu;
 348
 349         pptr = __alloc_percpu_gfp(sizeof(struct htab_elem *), 8,
 350                                   GFP_USER | __GFP_NOWARN);
 351         if (!pptr)
 352                 return -ENOMEM;
 353
 354         for_each_possible_cpu(cpu) {
 355                 l = pcpu_freelist_pop(&htab->freelist);
 356                 /* pop will succeed, since prealloc_init()
 357                  * preallocated extra num_possible_cpus elements
 358                  */
 359                 l_new = container_of(l, struct htab_elem, fnode);
 360                 *per_cpu_ptr(pptr, cpu) = l_new;
 361         }
 362         htab->extra_elems = pptr;
 363         return 0;
 364 }
 365
 366 /* Called from syscall */
 367 static int htab_map_alloc_check(union bpf_attr *attr)
 368 {
 369         bool percpu = (attr->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
 370                        attr->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH);
 371         bool lru = (attr->map_type == BPF_MAP_TYPE_LRU_HASH ||
 372                     attr->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH);
 373         /* percpu_lru means each cpu has its own LRU list.
 374          * it is different from BPF_MAP_TYPE_PERCPU_HASH where
 375          * the map's value itself is percpu.  percpu_lru has
 376          * nothing to do with the map's value.
 377          */
 378         bool percpu_lru = (attr->map_flags & BPF_F_NO_COMMON_LRU);
 379         bool prealloc = !(attr->map_flags & BPF_F_NO_PREALLOC);
 380         bool zero_seed = (attr->map_flags & BPF_F_ZERO_SEED);
 381         int numa_node = bpf_map_attr_numa_node(attr);
 382
 383         BUILD_BUG_ON(offsetof(struct htab_elem, htab) !=
 384                      offsetof(struct htab_elem, hash_node.pprev));
 385         BUILD_BUG_ON(offsetof(struct htab_elem, fnode.next) !=
 386                      offsetof(struct htab_elem, hash_node.pprev));
 387
 388         if (lru && !bpf_capable())
 389                 /* LRU implementation is much complicated than other
 390                  * maps.  Hence, limit to CAP_BPF.
 391                  */
 392                 return -EPERM;
 393
 394         if (zero_seed && !capable(CAP_SYS_ADMIN))
 395                 /* Guard against local DoS, and discourage production use. */
 396                 return -EPERM;
 397
 398         if (attr->map_flags & ~HTAB_CREATE_FLAG_MASK ||
 399             !bpf_map_flags_access_ok(attr->map_flags))
 400                 return -EINVAL;
 401
 402         if (!lru && percpu_lru)
 403                 return -EINVAL;
 404
 405         if (lru && !prealloc)
 406                 return -ENOTSUPP;
 407
 408         if (numa_node != NUMA_NO_NODE && (percpu || percpu_lru))
 409                 return -EINVAL;
 410
 411         /* check sanity of attributes.
 412          * value_size == 0 may be allowed in the future to use map as a set
 413          */
 414         if (attr->max_entries == 0 || attr->key_size == 0 ||
 415             attr->value_size == 0)
 416                 return -EINVAL;
 417
 418         if ((u64)attr->key_size + attr->value_size >= KMALLOC_MAX_SIZE -
 419            sizeof(struct htab_elem))
 420                 /* if key_size + value_size is bigger, the user space won't be
 421                  * able to access the elements via bpf syscall. This check
 422                  * also makes sure that the elem_size doesn't overflow and it's
 423                  * kmalloc-able later in htab_map_update_elem()
 424                  */
 425                 return -E2BIG;
 426
 427         return 0;
 428 }
 429
 430 static struct bpf_map *htab_map_alloc(union bpf_attr *attr)
 431 {
 432         bool percpu = (attr->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
 433                        attr->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH);
 434         bool lru = (attr->map_type == BPF_MAP_TYPE_LRU_HASH ||
 435                     attr->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH);
 436         /* percpu_lru means each cpu has its own LRU list.
 437          * it is different from BPF_MAP_TYPE_PERCPU_HASH where
 438          * the map's value itself is percpu.  percpu_lru has
 439          * nothing to do with the map's value.
 440          */
 441         bool percpu_lru = (attr->map_flags & BPF_F_NO_COMMON_LRU);
 442         bool prealloc = !(attr->map_flags & BPF_F_NO_PREALLOC);
 443         struct bpf_htab *htab;
 444         int err, i;
 445         u64 cost;
 446
 447         htab = kzalloc(sizeof(*htab), GFP_USER);
 448         if (!htab)
 449                 return ERR_PTR(-ENOMEM);
 450
 451         lockdep_register_key(&htab->lockdep_key);
 452
 453         bpf_map_init_from_attr(&htab->map, attr);
 454
 455         if (percpu_lru) {
 456                 /* ensure each CPU's lru list has >=1 elements.
 457                  * since we are at it, make each lru list has the same
 458                  * number of elements.
 459                  */
 460                 htab->map.max_entries = roundup(attr->max_entries,
 461                                                 num_possible_cpus());
 462                 if (htab->map.max_entries < attr->max_entries)
 463                         htab->map.max_entries = rounddown(attr->max_entries,
 464                                                           num_possible_cpus());
 465         }
 466
 467         /* hash table size must be power of 2 */
 468         htab->n_buckets = roundup_pow_of_two(htab->map.max_entries);
 469
 470         htab->elem_size = sizeof(struct htab_elem) +
 471                           round_up(htab->map.key_size, 8);
 472         if (percpu)
 473                 htab->elem_size += sizeof(void *);
 474         else
 475                 htab->elem_size += round_up(htab->map.value_size, 8);
 476
 477         err = -E2BIG;
 478         /* prevent zero size kmalloc and check for u32 overflow */
 479         if (htab->n_buckets == 0 ||
 480             htab->n_buckets > U32_MAX / sizeof(struct bucket))
 481                 goto free_htab;
 482
 483         cost = (u64) htab->n_buckets * sizeof(struct bucket) +
 484                (u64) htab->elem_size * htab->map.max_entries;
 485
 486         if (percpu)
 487                 cost += (u64) round_up(htab->map.value_size, 8) *
 488                         num_possible_cpus() * htab->map.max_entries;
 489         else
 490                cost += (u64) htab->elem_size * num_possible_cpus();
 491
 492         /* if map size is larger than memlock limit, reject it */
 493         err = bpf_map_charge_init(&htab->map.memory, cost);
 494         if (err)
 495                 goto free_htab;
 496
 497         err = -ENOMEM;
 498         htab->buckets = bpf_map_area_alloc(htab->n_buckets *
 499                                            sizeof(struct bucket),
 500                                            htab->map.numa_node);
 501         if (!htab->buckets)
 502                 goto free_charge;
 503
 504         for (i = 0; i < HASHTAB_MAP_LOCK_COUNT; i++) {
 505                 htab->map_locked[i] = __alloc_percpu_gfp(sizeof(int),
 506                                                          sizeof(int), GFP_USER);
 507                 if (!htab->map_locked[i])
 508                         goto free_map_locked;
 509         }
 510
 511         if (htab->map.map_flags & BPF_F_ZERO_SEED)
 512                 htab->hashrnd = 0;
 513         else
 514                 htab->hashrnd = get_random_int();
 515
 516         htab_init_buckets(htab);
 517
 518         if (prealloc) {
 519                 err = prealloc_init(htab);
 520                 if (err)
 521                         goto free_map_locked;
 522
 523                 if (!percpu && !lru) {
 524                         /* lru itself can remove the least used element, so
 525                          * there is no need for an extra elem during map_update.
 526                          */
 527                         err = alloc_extra_elems(htab);
 528                         if (err)
 529                                 goto free_prealloc;
 530                 }
 531         }
 532
 533         return &htab->map;
 534
 535 free_prealloc:
 536         prealloc_destroy(htab);
 537 free_map_locked:
 538         for (i = 0; i < HASHTAB_MAP_LOCK_COUNT; i++)
 539                 free_percpu(htab->map_locked[i]);
 540         bpf_map_area_free(htab->buckets);
 541 free_charge:
 542         bpf_map_charge_finish(&htab->map.memory);
 543 free_htab:
 544         lockdep_unregister_key(&htab->lockdep_key);
 545         kfree(htab);
 546         return ERR_PTR(err);
 547 }
 548
 549 static inline u32 htab_map_hash(const void *key, u32 key_len, u32 hashrnd)
 550 {
 551         return jhash(key, key_len, hashrnd);
 552 }
 553
 554 static inline struct bucket *__select_bucket(struct bpf_htab *htab, u32 hash)
 555 {
 556         return &htab->buckets[hash & (htab->n_buckets - 1)];
 557 }
 558
 559 static inline struct hlist_nulls_head *select_bucket(struct bpf_htab *htab, u32 hash)
 560 {
 561         return &__select_bucket(htab, hash)->head;
 562 }
 563
 564 /* this lookup function can only be called with bucket lock taken */
 565 static struct htab_elem *lookup_elem_raw(struct hlist_nulls_head *head, u32 hash,
 566                                          void *key, u32 key_size)
 567 {
 568         struct hlist_nulls_node *n;
 569         struct htab_elem *l;
 570
 571         hlist_nulls_for_each_entry_rcu(l, n, head, hash_node)
 572                 if (l->hash == hash && !memcmp(&l->key, key, key_size))
 573                         return l;
 574
 575         return NULL;
 576 }
 577
 578 /* can be called without bucket lock. it will repeat the loop in
 579  * the unlikely event when elements moved from one bucket into another
 580  * while link list is being walked
 581  */
 582 static struct htab_elem *lookup_nulls_elem_raw(struct hlist_nulls_head *head,
 583                                                u32 hash, void *key,
 584                                                u32 key_size, u32 n_buckets)
 585 {
 586         struct hlist_nulls_node *n;
 587         struct htab_elem *l;
 588
 589 again:
 590         hlist_nulls_for_each_entry_rcu(l, n, head, hash_node)
 591                 if (l->hash == hash && !memcmp(&l->key, key, key_size))
 592                         return l;
 593
 594         if (unlikely(get_nulls_value(n) != (hash & (n_buckets - 1))))
 595                 goto again;
 596
 597         return NULL;
 598 }
 599
 600 /* Called from syscall or from eBPF program directly, so
 601  * arguments have to match bpf_map_lookup_elem() exactly.
 602  * The return value is adjusted by BPF instructions
 603  * in htab_map_gen_lookup().
 604  */
 605 static void *__htab_map_lookup_elem(struct bpf_map *map, void *key)
 606 {
 607         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 608         struct hlist_nulls_head *head;
 609         struct htab_elem *l;
 610         u32 hash, key_size;
 611
 612         WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_trace_held());
 613
 614         key_size = map->key_size;
 615
 616         hash = htab_map_hash(key, key_size, htab->hashrnd);
 617
 618         head = select_bucket(htab, hash);
 619
 620         l = lookup_nulls_elem_raw(head, hash, key, key_size, htab->n_buckets);
 621
 622         return l;
 623 }
 624
 625 static void *htab_map_lookup_elem(struct bpf_map *map, void *key)
 626 {
 627         struct htab_elem *l = __htab_map_lookup_elem(map, key);
 628
 629         if (l)
 630                 return l->key + round_up(map->key_size, 8);
 631
 632         return NULL;
 633 }
 634
 635 /* inline bpf_map_lookup_elem() call.
 636  * Instead of:
 637  * bpf_prog
 638  *   bpf_map_lookup_elem
 639  *     map->ops->map_lookup_elem
 640  *       htab_map_lookup_elem
 641  *         __htab_map_lookup_elem
 642  * do:
 643  * bpf_prog
 644  *   __htab_map_lookup_elem
 645  */
 646 static int htab_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf)
 647 {
 648         struct bpf_insn *insn = insn_buf;
 649         const int ret = BPF_REG_0;
 650
 651         BUILD_BUG_ON(!__same_type(&__htab_map_lookup_elem,
 652                      (void *(*)(struct bpf_map *map, void *key))NULL));
 653         *insn++ = BPF_EMIT_CALL(BPF_CAST_CALL(__htab_map_lookup_elem));
 654         *insn++ = BPF_JMP_IMM(BPF_JEQ, ret, 0, 1);
 655         *insn++ = BPF_ALU64_IMM(BPF_ADD, ret,
 656                                 offsetof(struct htab_elem, key) +
 657                                 round_up(map->key_size, 8));
 658         return insn - insn_buf;
 659 }
 660
 661 static __always_inline void *__htab_lru_map_lookup_elem(struct bpf_map *map,
 662                                                         void *key, const bool mark)
 663 {
 664         struct htab_elem *l = __htab_map_lookup_elem(map, key);
 665
 666         if (l) {
 667                 if (mark)
 668                         bpf_lru_node_set_ref(&l->lru_node);
 669                 return l->key + round_up(map->key_size, 8);
 670         }
 671
 672         return NULL;
 673 }
 674
 675 static void *htab_lru_map_lookup_elem(struct bpf_map *map, void *key)
 676 {
 677         return __htab_lru_map_lookup_elem(map, key, true);
 678 }
 679
 680 static void *htab_lru_map_lookup_elem_sys(struct bpf_map *map, void *key)
 681 {
 682         return __htab_lru_map_lookup_elem(map, key, false);
 683 }
 684
 685 static int htab_lru_map_gen_lookup(struct bpf_map *map,
 686                                    struct bpf_insn *insn_buf)
 687 {
 688         struct bpf_insn *insn = insn_buf;
 689         const int ret = BPF_REG_0;
 690         const int ref_reg = BPF_REG_1;
 691
 692         BUILD_BUG_ON(!__same_type(&__htab_map_lookup_elem,
 693                      (void *(*)(struct bpf_map *map, void *key))NULL));
 694         *insn++ = BPF_EMIT_CALL(BPF_CAST_CALL(__htab_map_lookup_elem));
 695         *insn++ = BPF_JMP_IMM(BPF_JEQ, ret, 0, 4);
 696         *insn++ = BPF_LDX_MEM(BPF_B, ref_reg, ret,
 697                               offsetof(struct htab_elem, lru_node) +
 698                               offsetof(struct bpf_lru_node, ref));
 699         *insn++ = BPF_JMP_IMM(BPF_JNE, ref_reg, 0, 1);
 700         *insn++ = BPF_ST_MEM(BPF_B, ret,
 701                              offsetof(struct htab_elem, lru_node) +
 702                              offsetof(struct bpf_lru_node, ref),
 703                              1);
 704         *insn++ = BPF_ALU64_IMM(BPF_ADD, ret,
 705                                 offsetof(struct htab_elem, key) +
 706                                 round_up(map->key_size, 8));
 707         return insn - insn_buf;
 708 }
 709
 710 /* It is called from the bpf_lru_list when the LRU needs to delete
 711  * older elements from the htab.
 712  */
 713 static bool htab_lru_map_delete_node(void *arg, struct bpf_lru_node *node)
 714 {
 715         struct bpf_htab *htab = (struct bpf_htab *)arg;
 716         struct htab_elem *l = NULL, *tgt_l;
 717         struct hlist_nulls_head *head;
 718         struct hlist_nulls_node *n;
 719         unsigned long flags;
 720         struct bucket *b;
 721         int ret;
 722
 723         tgt_l = container_of(node, struct htab_elem, lru_node);
 724         b = __select_bucket(htab, tgt_l->hash);
 725         head = &b->head;
 726
 727         ret = htab_lock_bucket(htab, b, tgt_l->hash, &flags);
 728         if (ret)
 729                 return false;
 730
 731         hlist_nulls_for_each_entry_rcu(l, n, head, hash_node)
 732                 if (l == tgt_l) {
 733                         hlist_nulls_del_rcu(&l->hash_node);
 734                         break;
 735                 }
 736
 737         htab_unlock_bucket(htab, b, tgt_l->hash, flags);
 738
 739         return l == tgt_l;
 740 }
 741
 742 /* Called from syscall */
 743 static int htab_map_get_next_key(struct bpf_map *map, void *key, void *next_key)
 744 {
 745         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 746         struct hlist_nulls_head *head;
 747         struct htab_elem *l, *next_l;
 748         u32 hash, key_size;
 749         int i = 0;
 750
 751         WARN_ON_ONCE(!rcu_read_lock_held());
 752
 753         key_size = map->key_size;
 754
 755         if (!key)
 756                 goto find_first_elem;
 757
 758         hash = htab_map_hash(key, key_size, htab->hashrnd);
 759
 760         head = select_bucket(htab, hash);
 761
 762         /* lookup the key */
 763         l = lookup_nulls_elem_raw(head, hash, key, key_size, htab->n_buckets);
 764
 765         if (!l)
 766                 goto find_first_elem;
 767
 768         /* key was found, get next key in the same bucket */
 769         next_l = hlist_nulls_entry_safe(rcu_dereference_raw(hlist_nulls_next_rcu(&l->hash_node)),
 770                                   struct htab_elem, hash_node);
 771
 772         if (next_l) {
 773                 /* if next elem in this hash list is non-zero, just return it */
 774                 memcpy(next_key, next_l->key, key_size);
 775                 return 0;
 776         }
 777
 778         /* no more elements in this hash list, go to the next bucket */
 779         i = hash & (htab->n_buckets - 1);
 780         i++;
 781
 782 find_first_elem:
 783         /* iterate over buckets */
 784         for (; i < htab->n_buckets; i++) {
 785                 head = select_bucket(htab, i);
 786
 787                 /* pick first element in the bucket */
 788                 next_l = hlist_nulls_entry_safe(rcu_dereference_raw(hlist_nulls_first_rcu(head)),
 789                                           struct htab_elem, hash_node);
 790                 if (next_l) {
 791                         /* if it's not empty, just return it */
 792                         memcpy(next_key, next_l->key, key_size);
 793                         return 0;
 794                 }
 795         }
 796
 797         /* iterated over all buckets and all elements */
 798         return -ENOENT;
 799 }
 800
 801 static void htab_elem_free(struct bpf_htab *htab, struct htab_elem *l)
 802 {
 803         if (htab->map.map_type == BPF_MAP_TYPE_PERCPU_HASH)
 804                 free_percpu(htab_elem_get_ptr(l, htab->map.key_size));
 805         kfree(l);
 806 }
 807
 808 static void htab_elem_free_rcu(struct rcu_head *head)
 809 {
 810         struct htab_elem *l = container_of(head, struct htab_elem, rcu);
 811         struct bpf_htab *htab = l->htab;
 812
 813         htab_elem_free(htab, l);
 814 }
 815
 816 static void htab_put_fd_value(struct bpf_htab *htab, struct htab_elem *l)
 817 {
 818         struct bpf_map *map = &htab->map;
 819         void *ptr;
 820
 821         if (map->ops->map_fd_put_ptr) {
 822                 ptr = fd_htab_map_get_ptr(map, l);
 823                 map->ops->map_fd_put_ptr(ptr);
 824         }
 825 }
 826
 827 static void free_htab_elem(struct bpf_htab *htab, struct htab_elem *l)
 828 {
 829         htab_put_fd_value(htab, l);
 830
 831         if (htab_is_prealloc(htab)) {
 832                 __pcpu_freelist_push(&htab->freelist, &l->fnode);
 833         } else {
 834                 atomic_dec(&htab->count);
 835                 l->htab = htab;
 836                 call_rcu(&l->rcu, htab_elem_free_rcu);
 837         }
 838 }
 839
 840 static void pcpu_copy_value(struct bpf_htab *htab, void __percpu *pptr,
 841                             void *value, bool onallcpus)
 842 {
 843         if (!onallcpus) {
 844                 /* copy true value_size bytes */
 845                 memcpy(this_cpu_ptr(pptr), value, htab->map.value_size);
 846         } else {
 847                 u32 size = round_up(htab->map.value_size, 8);
 848                 int off = 0, cpu;
 849
 850                 for_each_possible_cpu(cpu) {
 851                         bpf_long_memcpy(per_cpu_ptr(pptr, cpu),
 852                                         value + off, size);
 853                         off += size;
 854                 }
 855         }
 856 }
 857
 858 static void pcpu_init_value(struct bpf_htab *htab, void __percpu *pptr,
 859                             void *value, bool onallcpus)
 860 {
 861         /* When using prealloc and not setting the initial value on all cpus,
 862          * zero-fill element values for other cpus (just as what happens when
 863          * not using prealloc). Otherwise, bpf program has no way to ensure
 864          * known initial values for cpus other than current one
 865          * (onallcpus=false always when coming from bpf prog).
 866          */
 867         if (htab_is_prealloc(htab) && !onallcpus) {
 868                 u32 size = round_up(htab->map.value_size, 8);
 869                 int current_cpu = raw_smp_processor_id();
 870                 int cpu;
 871
 872                 for_each_possible_cpu(cpu) {
 873                         if (cpu == current_cpu)
 874                                 bpf_long_memcpy(per_cpu_ptr(pptr, cpu), value,
 875                                                 size);
 876                         else
 877                                 memset(per_cpu_ptr(pptr, cpu), 0, size);
 878                 }
 879         } else {
 880                 pcpu_copy_value(htab, pptr, value, onallcpus);
 881         }
 882 }
 883
 884 static bool fd_htab_map_needs_adjust(const struct bpf_htab *htab)
 885 {
 886         return htab->map.map_type == BPF_MAP_TYPE_HASH_OF_MAPS &&
 887                BITS_PER_LONG == 64;
 888 }
 889
 890 static struct htab_elem *alloc_htab_elem(struct bpf_htab *htab, void *key,
 891                                          void *value, u32 key_size, u32 hash,
 892                                          bool percpu, bool onallcpus,
 893                                          struct htab_elem *old_elem)
 894 {
 895         u32 size = htab->map.value_size;
 896         bool prealloc = htab_is_prealloc(htab);
 897         struct htab_elem *l_new, **pl_new;
 898         void __percpu *pptr;
 899
 900         if (prealloc) {
 901                 if (old_elem) {
 902                         /* if we're updating the existing element,
 903                          * use per-cpu extra elems to avoid freelist_pop/push
 904                          */
 905                         pl_new = this_cpu_ptr(htab->extra_elems);
 906                         l_new = *pl_new;
 907                         htab_put_fd_value(htab, old_elem);
 908                         *pl_new = old_elem;
 909                 } else {
 910                         struct pcpu_freelist_node *l;
 911
 912                         l = __pcpu_freelist_pop(&htab->freelist);
 913                         if (!l)
 914                                 return ERR_PTR(-E2BIG);
 915                         l_new = container_of(l, struct htab_elem, fnode);
 916                 }
 917         } else {
 918                 if (atomic_inc_return(&htab->count) > htab->map.max_entries)
 919                         if (!old_elem) {
 920                                 /* when map is full and update() is replacing
 921                                  * old element, it's ok to allocate, since
 922                                  * old element will be freed immediately.
 923                                  * Otherwise return an error
 924                                  */
 925                                 l_new = ERR_PTR(-E2BIG);
 926                                 goto dec_count;
 927                         }
 928                 l_new = kmalloc_node(htab->elem_size, GFP_ATOMIC | __GFP_NOWARN,
 929                                      htab->map.numa_node);
 930                 if (!l_new) {
 931                         l_new = ERR_PTR(-ENOMEM);
 932                         goto dec_count;
 933                 }
 934                 check_and_init_map_lock(&htab->map,
 935                                         l_new->key + round_up(key_size, 8));
 936         }
 937
 938         memcpy(l_new->key, key, key_size);
 939         if (percpu) {
 940                 size = round_up(size, 8);
 941                 if (prealloc) {
 942                         pptr = htab_elem_get_ptr(l_new, key_size);
 943                 } else {
 944                         /* alloc_percpu zero-fills */
 945                         pptr = __alloc_percpu_gfp(size, 8,
 946                                                   GFP_ATOMIC | __GFP_NOWARN);
 947                         if (!pptr) {
 948                                 kfree(l_new);
 949                                 l_new = ERR_PTR(-ENOMEM);
 950                                 goto dec_count;
 951                         }
 952                 }
 953
 954                 pcpu_init_value(htab, pptr, value, onallcpus);
 955
 956                 if (!prealloc)
 957                         htab_elem_set_ptr(l_new, key_size, pptr);
 958         } else if (fd_htab_map_needs_adjust(htab)) {
 959                 size = round_up(size, 8);
 960                 memcpy(l_new->key + round_up(key_size, 8), value, size);
 961         } else {
 962                 copy_map_value(&htab->map,
 963                                l_new->key + round_up(key_size, 8),
 964                                value);
 965         }
 966
 967         l_new->hash = hash;
 968         return l_new;
 969 dec_count:
 970         atomic_dec(&htab->count);
 971         return l_new;
 972 }
 973
 974 static int check_flags(struct bpf_htab *htab, struct htab_elem *l_old,
 975                        u64 map_flags)
 976 {
 977         if (l_old && (map_flags & ~BPF_F_LOCK) == BPF_NOEXIST)
 978                 /* elem already exists */
 979                 return -EEXIST;
 980
 981         if (!l_old && (map_flags & ~BPF_F_LOCK) == BPF_EXIST)
 982                 /* elem doesn't exist, cannot update it */
 983                 return -ENOENT;
 984
 985         return 0;
 986 }
 987
 988 /* Called from syscall or from eBPF program */
 989 static int htab_map_update_elem(struct bpf_map *map, void *key, void *value,
 990                                 u64 map_flags)
 991 {
 992         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
 993         struct htab_elem *l_new = NULL, *l_old;
 994         struct hlist_nulls_head *head;
 995         unsigned long flags;
 996         struct bucket *b;
 997         u32 key_size, hash;
 998         int ret;
 999
1000         if (unlikely((map_flags & ~BPF_F_LOCK) > BPF_EXIST))
1001                 /* unknown flags */
1002                 return -EINVAL;
1003
1004         WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_trace_held());
1005
1006         key_size = map->key_size;
1007
1008         hash = htab_map_hash(key, key_size, htab->hashrnd);
1009
1010         b = __select_bucket(htab, hash);
1011         head = &b->head;
1012
1013         if (unlikely(map_flags & BPF_F_LOCK)) {
1014                 if (unlikely(!map_value_has_spin_lock(map)))
1015                         return -EINVAL;
1016                 /* find an element without taking the bucket lock */
1017                 l_old = lookup_nulls_elem_raw(head, hash, key, key_size,
1018                                               htab->n_buckets);
1019                 ret = check_flags(htab, l_old, map_flags);
1020                 if (ret)
1021                         return ret;
1022                 if (l_old) {
1023                         /* grab the element lock and update value in place */
1024                         copy_map_value_locked(map,
1025                                               l_old->key + round_up(key_size, 8),
1026                                               value, false);
1027                         return 0;
1028                 }
1029                 /* fall through, grab the bucket lock and lookup again.
1030                  * 99.9% chance that the element won't be found,
1031                  * but second lookup under lock has to be done.
1032                  */
1033         }
1034
1035         ret = htab_lock_bucket(htab, b, hash, &flags);
1036         if (ret)
1037                 return ret;
1038
1039         l_old = lookup_elem_raw(head, hash, key, key_size);
1040
1041         ret = check_flags(htab, l_old, map_flags);
1042         if (ret)
1043                 goto err;
1044
1045         if (unlikely(l_old && (map_flags & BPF_F_LOCK))) {
1046                 /* first lookup without the bucket lock didn't find the element,
1047                  * but second lookup with the bucket lock found it.
1048                  * This case is highly unlikely, but has to be dealt with:
1049                  * grab the element lock in addition to the bucket lock
1050                  * and update element in place
1051                  */
1052                 copy_map_value_locked(map,
1053                                       l_old->key + round_up(key_size, 8),
1054                                       value, false);
1055                 ret = 0;
1056                 goto err;
1057         }
1058
1059         l_new = alloc_htab_elem(htab, key, value, key_size, hash, false, false,
1060                                 l_old);
1061         if (IS_ERR(l_new)) {
1062                 /* all pre-allocated elements are in use or memory exhausted */
1063                 ret = PTR_ERR(l_new);
1064                 goto err;
1065         }
1066
1067         /* add new element to the head of the list, so that
1068          * concurrent search will find it before old elem
1069          */
1070         hlist_nulls_add_head_rcu(&l_new->hash_node, head);
1071         if (l_old) {
1072                 hlist_nulls_del_rcu(&l_old->hash_node);
1073                 if (!htab_is_prealloc(htab))
1074                         free_htab_elem(htab, l_old);
1075         }
1076         ret = 0;
1077 err:
1078         htab_unlock_bucket(htab, b, hash, flags);
1079         return ret;
1080 }
1081
1082 static int htab_lru_map_update_elem(struct bpf_map *map, void *key, void *value,
1083                                     u64 map_flags)
1084 {
1085         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1086         struct htab_elem *l_new, *l_old = NULL;
1087         struct hlist_nulls_head *head;
1088         unsigned long flags;
1089         struct bucket *b;
1090         u32 key_size, hash;
1091         int ret;
1092
1093         if (unlikely(map_flags > BPF_EXIST))
1094                 /* unknown flags */
1095                 return -EINVAL;
1096
1097         WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_trace_held());
1098
1099         key_size = map->key_size;
1100
1101         hash = htab_map_hash(key, key_size, htab->hashrnd);
1102
1103         b = __select_bucket(htab, hash);
1104         head = &b->head;
1105
1106         /* For LRU, we need to alloc before taking bucket's
1107          * spinlock because getting free nodes from LRU may need
1108          * to remove older elements from htab and this removal
1109          * operation will need a bucket lock.
1110          */
1111         l_new = prealloc_lru_pop(htab, key, hash);
1112         if (!l_new)
1113                 return -ENOMEM;
1114         memcpy(l_new->key + round_up(map->key_size, 8), value, map->value_size);
1115
1116         ret = htab_lock_bucket(htab, b, hash, &flags);
1117         if (ret)
1118                 return ret;
1119
1120         l_old = lookup_elem_raw(head, hash, key, key_size);
1121
1122         ret = check_flags(htab, l_old, map_flags);
1123         if (ret)
1124                 goto err;
1125
1126         /* add new element to the head of the list, so that
1127          * concurrent search will find it before old elem
1128          */
1129         hlist_nulls_add_head_rcu(&l_new->hash_node, head);
1130         if (l_old) {
1131                 bpf_lru_node_set_ref(&l_new->lru_node);
1132                 hlist_nulls_del_rcu(&l_old->hash_node);
1133         }
1134         ret = 0;
1135
1136 err:
1137         htab_unlock_bucket(htab, b, hash, flags);
1138
1139         if (ret)
1140                 bpf_lru_push_free(&htab->lru, &l_new->lru_node);
1141         else if (l_old)
1142                 bpf_lru_push_free(&htab->lru, &l_old->lru_node);
1143
1144         return ret;
1145 }
1146
1147 static int __htab_percpu_map_update_elem(struct bpf_map *map, void *key,
1148                                          void *value, u64 map_flags,
1149                                          bool onallcpus)
1150 {
1151         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1152         struct htab_elem *l_new = NULL, *l_old;
1153         struct hlist_nulls_head *head;
1154         unsigned long flags;
1155         struct bucket *b;
1156         u32 key_size, hash;
1157         int ret;
1158
1159         if (unlikely(map_flags > BPF_EXIST))
1160                 /* unknown flags */
1161                 return -EINVAL;
1162
1163         WARN_ON_ONCE(!rcu_read_lock_held());
1164
1165         key_size = map->key_size;
1166
1167         hash = htab_map_hash(key, key_size, htab->hashrnd);
1168
1169         b = __select_bucket(htab, hash);
1170         head = &b->head;
1171
1172         ret = htab_lock_bucket(htab, b, hash, &flags);
1173         if (ret)
1174                 return ret;
1175
1176         l_old = lookup_elem_raw(head, hash, key, key_size);
1177
1178         ret = check_flags(htab, l_old, map_flags);
1179         if (ret)
1180                 goto err;
1181
1182         if (l_old) {
1183                 /* per-cpu hash map can update value in-place */
1184                 pcpu_copy_value(htab, htab_elem_get_ptr(l_old, key_size),
1185                                 value, onallcpus);
1186         } else {
1187                 l_new = alloc_htab_elem(htab, key, value, key_size,
1188                                         hash, true, onallcpus, NULL);
1189                 if (IS_ERR(l_new)) {
1190                         ret = PTR_ERR(l_new);
1191                         goto err;
1192                 }
1193                 hlist_nulls_add_head_rcu(&l_new->hash_node, head);
1194         }
1195         ret = 0;
1196 err:
1197         htab_unlock_bucket(htab, b, hash, flags);
1198         return ret;
1199 }
1200
1201 static int __htab_lru_percpu_map_update_elem(struct bpf_map *map, void *key,
1202                                              void *value, u64 map_flags,
1203                                              bool onallcpus)
1204 {
1205         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1206         struct htab_elem *l_new = NULL, *l_old;
1207         struct hlist_nulls_head *head;
1208         unsigned long flags;
1209         struct bucket *b;
1210         u32 key_size, hash;
1211         int ret;
1212
1213         if (unlikely(map_flags > BPF_EXIST))
1214                 /* unknown flags */
1215                 return -EINVAL;
1216
1217         WARN_ON_ONCE(!rcu_read_lock_held());
1218
1219         key_size = map->key_size;
1220
1221         hash = htab_map_hash(key, key_size, htab->hashrnd);
1222
1223         b = __select_bucket(htab, hash);
1224         head = &b->head;
1225
1226         /* For LRU, we need to alloc before taking bucket's
1227          * spinlock because LRU's elem alloc may need
1228          * to remove older elem from htab and this removal
1229          * operation will need a bucket lock.
1230          */
1231         if (map_flags != BPF_EXIST) {
1232                 l_new = prealloc_lru_pop(htab, key, hash);
1233                 if (!l_new)
1234                         return -ENOMEM;
1235         }
1236
1237         ret = htab_lock_bucket(htab, b, hash, &flags);
1238         if (ret)
1239                 return ret;
1240
1241         l_old = lookup_elem_raw(head, hash, key, key_size);
1242
1243         ret = check_flags(htab, l_old, map_flags);
1244         if (ret)
1245                 goto err;
1246
1247         if (l_old) {
1248                 bpf_lru_node_set_ref(&l_old->lru_node);
1249
1250                 /* per-cpu hash map can update value in-place */
1251                 pcpu_copy_value(htab, htab_elem_get_ptr(l_old, key_size),
1252                                 value, onallcpus);
1253         } else {
1254                 pcpu_init_value(htab, htab_elem_get_ptr(l_new, key_size),
1255                                 value, onallcpus);
1256                 hlist_nulls_add_head_rcu(&l_new->hash_node, head);
1257                 l_new = NULL;
1258         }
1259         ret = 0;
1260 err:
1261         htab_unlock_bucket(htab, b, hash, flags);
1262         if (l_new)
1263                 bpf_lru_push_free(&htab->lru, &l_new->lru_node);
1264         return ret;
1265 }
1266
1267 static int htab_percpu_map_update_elem(struct bpf_map *map, void *key,
1268                                        void *value, u64 map_flags)
1269 {
1270         return __htab_percpu_map_update_elem(map, key, value, map_flags, false);
1271 }
1272
1273 static int htab_lru_percpu_map_update_elem(struct bpf_map *map, void *key,
1274                                            void *value, u64 map_flags)
1275 {
1276         return __htab_lru_percpu_map_update_elem(map, key, value, map_flags,
1277                                                  false);
1278 }
1279
1280 /* Called from syscall or from eBPF program */
1281 static int htab_map_delete_elem(struct bpf_map *map, void *key)
1282 {
1283         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1284         struct hlist_nulls_head *head;
1285         struct bucket *b;
1286         struct htab_elem *l;
1287         unsigned long flags;
1288         u32 hash, key_size;
1289         int ret;
1290
1291         WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_trace_held());
1292
1293         key_size = map->key_size;
1294
1295         hash = htab_map_hash(key, key_size, htab->hashrnd);
1296         b = __select_bucket(htab, hash);
1297         head = &b->head;
1298
1299         ret = htab_lock_bucket(htab, b, hash, &flags);
1300         if (ret)
1301                 return ret;
1302
1303         l = lookup_elem_raw(head, hash, key, key_size);
1304
1305         if (l) {
1306                 hlist_nulls_del_rcu(&l->hash_node);
1307                 free_htab_elem(htab, l);
1308         } else {
1309                 ret = -ENOENT;
1310         }
1311
1312         htab_unlock_bucket(htab, b, hash, flags);
1313         return ret;
1314 }
1315
1316 static int htab_lru_map_delete_elem(struct bpf_map *map, void *key)
1317 {
1318         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1319         struct hlist_nulls_head *head;
1320         struct bucket *b;
1321         struct htab_elem *l;
1322         unsigned long flags;
1323         u32 hash, key_size;
1324         int ret;
1325
1326         WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_trace_held());
1327
1328         key_size = map->key_size;
1329
1330         hash = htab_map_hash(key, key_size, htab->hashrnd);
1331         b = __select_bucket(htab, hash);
1332         head = &b->head;
1333
1334         ret = htab_lock_bucket(htab, b, hash, &flags);
1335         if (ret)
1336                 return ret;
1337
1338         l = lookup_elem_raw(head, hash, key, key_size);
1339
1340         if (l)
1341                 hlist_nulls_del_rcu(&l->hash_node);
1342         else
1343                 ret = -ENOENT;
1344
1345         htab_unlock_bucket(htab, b, hash, flags);
1346         if (l)
1347                 bpf_lru_push_free(&htab->lru, &l->lru_node);
1348         return ret;
1349 }
1350
1351 static void delete_all_elements(struct bpf_htab *htab)
1352 {
1353         int i;
1354
1355         for (i = 0; i < htab->n_buckets; i++) {
1356                 struct hlist_nulls_head *head = select_bucket(htab, i);
1357                 struct hlist_nulls_node *n;
1358                 struct htab_elem *l;
1359
1360                 hlist_nulls_for_each_entry_safe(l, n, head, hash_node) {
1361                         hlist_nulls_del_rcu(&l->hash_node);
1362                         htab_elem_free(htab, l);
1363                 }
1364         }
1365 }
1366
1367 /* Called when map->refcnt goes to zero, either from workqueue or from syscall */
1368 static void htab_map_free(struct bpf_map *map)
1369 {
1370         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1371         int i;
1372
1373         /* bpf_free_used_maps() or close(map_fd) will trigger this map_free callback.
1374          * bpf_free_used_maps() is called after bpf prog is no longer executing.
1375          * There is no need to synchronize_rcu() here to protect map elements.
1376          */
1377
1378         /* some of free_htab_elem() callbacks for elements of this map may
1379          * not have executed. Wait for them.
1380          */
1381         rcu_barrier();
1382         if (!htab_is_prealloc(htab))
1383                 delete_all_elements(htab);
1384         else
1385                 prealloc_destroy(htab);
1386
1387         free_percpu(htab->extra_elems);
1388         bpf_map_area_free(htab->buckets);
1389         for (i = 0; i < HASHTAB_MAP_LOCK_COUNT; i++)
1390                 free_percpu(htab->map_locked[i]);
1391         lockdep_unregister_key(&htab->lockdep_key);
1392         kfree(htab);
1393 }
1394
1395 static void htab_map_seq_show_elem(struct bpf_map *map, void *key,
1396                                    struct seq_file *m)
1397 {
1398         void *value;
1399
1400         rcu_read_lock();
1401
1402         value = htab_map_lookup_elem(map, key);
1403         if (!value) {
1404                 rcu_read_unlock();
1405                 return;
1406         }
1407
1408         btf_type_seq_show(map->btf, map->btf_key_type_id, key, m);
1409         seq_puts(m, ": ");
1410         btf_type_seq_show(map->btf, map->btf_value_type_id, value, m);
1411         seq_puts(m, "\n");
1412
1413         rcu_read_unlock();
1414 }
1415
1416 static int
1417 __htab_map_lookup_and_delete_batch(struct bpf_map *map,
1418                                    const union bpf_attr *attr,
1419                                    union bpf_attr __user *uattr,
1420                                    bool do_delete, bool is_lru_map,
1421                                    bool is_percpu)
1422 {
1423         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1424         u32 bucket_cnt, total, key_size, value_size, roundup_key_size;
1425         void *keys = NULL, *values = NULL, *value, *dst_key, *dst_val;
1426         void __user *uvalues = u64_to_user_ptr(attr->batch.values);
1427         void __user *ukeys = u64_to_user_ptr(attr->batch.keys);
1428         void *ubatch = u64_to_user_ptr(attr->batch.in_batch);
1429         u32 batch, max_count, size, bucket_size;
1430         struct htab_elem *node_to_free = NULL;
1431         u64 elem_map_flags, map_flags;
1432         struct hlist_nulls_head *head;
1433         struct hlist_nulls_node *n;
1434         unsigned long flags = 0;
1435         bool locked = false;
1436         struct htab_elem *l;
1437         struct bucket *b;
1438         int ret = 0;
1439
1440         elem_map_flags = attr->batch.elem_flags;
1441         if ((elem_map_flags & ~BPF_F_LOCK) ||
1442             ((elem_map_flags & BPF_F_LOCK) && !map_value_has_spin_lock(map)))
1443                 return -EINVAL;
1444
1445         map_flags = attr->batch.flags;
1446         if (map_flags)
1447                 return -EINVAL;
1448
1449         max_count = attr->batch.count;
1450         if (!max_count)
1451                 return 0;
1452
1453         if (put_user(0, &uattr->batch.count))
1454                 return -EFAULT;
1455
1456         batch = 0;
1457         if (ubatch && copy_from_user(&batch, ubatch, sizeof(batch)))
1458                 return -EFAULT;
1459
1460         if (batch >= htab->n_buckets)
1461                 return -ENOENT;
1462
1463         key_size = htab->map.key_size;
1464         roundup_key_size = round_up(htab->map.key_size, 8);
1465         value_size = htab->map.value_size;
1466         size = round_up(value_size, 8);
1467         if (is_percpu)
1468                 value_size = size * num_possible_cpus();
1469         total = 0;
1470         /* while experimenting with hash tables with sizes ranging from 10 to
1471          * 1000, it was observed that a bucket can have upto 5 entries.
1472          */
1473         bucket_size = 5;
1474
1475 alloc:
1476         /* We cannot do copy_from_user or copy_to_user inside
1477          * the rcu_read_lock. Allocate enough space here.
1478          */
1479         keys = kvmalloc(key_size * bucket_size, GFP_USER | __GFP_NOWARN);
1480         values = kvmalloc(value_size * bucket_size, GFP_USER | __GFP_NOWARN);
1481         if (!keys || !values) {
1482                 ret = -ENOMEM;
1483                 goto after_loop;
1484         }
1485
1486 again:
1487         bpf_disable_instrumentation();
1488         rcu_read_lock();
1489 again_nocopy:
1490         dst_key = keys;
1491         dst_val = values;
1492         b = &htab->buckets[batch];
1493         head = &b->head;
1494         /* do not grab the lock unless need it (bucket_cnt > 0). */
1495         if (locked) {
1496                 ret = htab_lock_bucket(htab, b, batch, &flags);
1497                 if (ret)
1498                         goto next_batch;
1499         }
1500
1501         bucket_cnt = 0;
1502         hlist_nulls_for_each_entry_rcu(l, n, head, hash_node)
1503                 bucket_cnt++;
1504
1505         if (bucket_cnt && !locked) {
1506                 locked = true;
1507                 goto again_nocopy;
1508         }
1509
1510         if (bucket_cnt > (max_count - total)) {
1511                 if (total == 0)
1512                         ret = -ENOSPC;
1513                 /* Note that since bucket_cnt > 0 here, it is implicit
1514                  * that the locked was grabbed, so release it.
1515                  */
1516                 htab_unlock_bucket(htab, b, batch, flags);
1517                 rcu_read_unlock();
1518                 bpf_enable_instrumentation();
1519                 goto after_loop;
1520         }
1521
1522         if (bucket_cnt > bucket_size) {
1523                 bucket_size = bucket_cnt;
1524                 /* Note that since bucket_cnt > 0 here, it is implicit
1525                  * that the locked was grabbed, so release it.
1526                  */
1527                 htab_unlock_bucket(htab, b, batch, flags);
1528                 rcu_read_unlock();
1529                 bpf_enable_instrumentation();
1530                 kvfree(keys);
1531                 kvfree(values);
1532                 goto alloc;
1533         }
1534
1535         /* Next block is only safe to run if you have grabbed the lock */
1536         if (!locked)
1537                 goto next_batch;
1538
1539         hlist_nulls_for_each_entry_safe(l, n, head, hash_node) {
1540                 memcpy(dst_key, l->key, key_size);
1541
1542                 if (is_percpu) {
1543                         int off = 0, cpu;
1544                         void __percpu *pptr;
1545
1546                         pptr = htab_elem_get_ptr(l, map->key_size);
1547                         for_each_possible_cpu(cpu) {
1548                                 bpf_long_memcpy(dst_val + off,
1549                                                 per_cpu_ptr(pptr, cpu), size);
1550                                 off += size;
1551                         }
1552                 } else {
1553                         value = l->key + roundup_key_size;
1554                         if (elem_map_flags & BPF_F_LOCK)
1555                                 copy_map_value_locked(map, dst_val, value,
1556                                                       true);
1557                         else
1558                                 copy_map_value(map, dst_val, value);
1559                         check_and_init_map_lock(map, dst_val);
1560                 }
1561                 if (do_delete) {
1562                         hlist_nulls_del_rcu(&l->hash_node);
1563
1564                         /* bpf_lru_push_free() will acquire lru_lock, which
1565                          * may cause deadlock. See comments in function
1566                          * prealloc_lru_pop(). Let us do bpf_lru_push_free()
1567                          * after releasing the bucket lock.
1568                          */
1569                         if (is_lru_map) {
1570                                 l->batch_flink = node_to_free;
1571                                 node_to_free = l;
1572                         } else {
1573                                 free_htab_elem(htab, l);
1574                         }
1575                 }
1576                 dst_key += key_size;
1577                 dst_val += value_size;
1578         }
1579
1580         htab_unlock_bucket(htab, b, batch, flags);
1581         locked = false;
1582
1583         while (node_to_free) {
1584                 l = node_to_free;
1585                 node_to_free = node_to_free->batch_flink;
1586                 bpf_lru_push_free(&htab->lru, &l->lru_node);
1587         }
1588
1589 next_batch:
1590         /* If we are not copying data, we can go to next bucket and avoid
1591          * unlocking the rcu.
1592          */
1593         if (!bucket_cnt && (batch + 1 < htab->n_buckets)) {
1594                 batch++;
1595                 goto again_nocopy;
1596         }
1597
1598         rcu_read_unlock();
1599         bpf_enable_instrumentation();
1600         if (bucket_cnt && (copy_to_user(ukeys + total * key_size, keys,
1601             key_size * bucket_cnt) ||
1602             copy_to_user(uvalues + total * value_size, values,
1603             value_size * bucket_cnt))) {
1604                 ret = -EFAULT;
1605                 goto after_loop;
1606         }
1607
1608         total += bucket_cnt;
1609         batch++;
1610         if (batch >= htab->n_buckets) {
1611                 ret = -ENOENT;
1612                 goto after_loop;
1613         }
1614         goto again;
1615
1616 after_loop:
1617         if (ret == -EFAULT)
1618                 goto out;
1619
1620         /* copy # of entries and next batch */
1621         ubatch = u64_to_user_ptr(attr->batch.out_batch);
1622         if (copy_to_user(ubatch, &batch, sizeof(batch)) ||
1623             put_user(total, &uattr->batch.count))
1624                 ret = -EFAULT;
1625
1626 out:
1627         kvfree(keys);
1628         kvfree(values);
1629         return ret;
1630 }
1631
1632 static int
1633 htab_percpu_map_lookup_batch(struct bpf_map *map, const union bpf_attr *attr,
1634                              union bpf_attr __user *uattr)
1635 {
1636         return __htab_map_lookup_and_delete_batch(map, attr, uattr, false,
1637                                                   false, true);
1638 }
1639
1640 static int
1641 htab_percpu_map_lookup_and_delete_batch(struct bpf_map *map,
1642                                         const union bpf_attr *attr,
1643                                         union bpf_attr __user *uattr)
1644 {
1645         return __htab_map_lookup_and_delete_batch(map, attr, uattr, true,
1646                                                   false, true);
1647 }
1648
1649 static int
1650 htab_map_lookup_batch(struct bpf_map *map, const union bpf_attr *attr,
1651                       union bpf_attr __user *uattr)
1652 {
1653         return __htab_map_lookup_and_delete_batch(map, attr, uattr, false,
1654                                                   false, false);
1655 }
1656
1657 static int
1658 htab_map_lookup_and_delete_batch(struct bpf_map *map,
1659                                  const union bpf_attr *attr,
1660                                  union bpf_attr __user *uattr)
1661 {
1662         return __htab_map_lookup_and_delete_batch(map, attr, uattr, true,
1663                                                   false, false);
1664 }
1665
1666 static int
1667 htab_lru_percpu_map_lookup_batch(struct bpf_map *map,
1668                                  const union bpf_attr *attr,
1669                                  union bpf_attr __user *uattr)
1670 {
1671         return __htab_map_lookup_and_delete_batch(map, attr, uattr, false,
1672                                                   true, true);
1673 }
1674
1675 static int
1676 htab_lru_percpu_map_lookup_and_delete_batch(struct bpf_map *map,
1677                                             const union bpf_attr *attr,
1678                                             union bpf_attr __user *uattr)
1679 {
1680         return __htab_map_lookup_and_delete_batch(map, attr, uattr, true,
1681                                                   true, true);
1682 }
1683
1684 static int
1685 htab_lru_map_lookup_batch(struct bpf_map *map, const union bpf_attr *attr,
1686                           union bpf_attr __user *uattr)
1687 {
1688         return __htab_map_lookup_and_delete_batch(map, attr, uattr, false,
1689                                                   true, false);
1690 }
1691
1692 static int
1693 htab_lru_map_lookup_and_delete_batch(struct bpf_map *map,
1694                                      const union bpf_attr *attr,
1695                                      union bpf_attr __user *uattr)
1696 {
1697         return __htab_map_lookup_and_delete_batch(map, attr, uattr, true,
1698                                                   true, false);
1699 }
1700
1701 struct bpf_iter_seq_hash_map_info {
1702         struct bpf_map *map;
1703         struct bpf_htab *htab;
1704         void *percpu_value_buf; // non-zero means percpu hash
1705         u32 bucket_id;
1706         u32 skip_elems;
1707 };
1708
1709 static struct htab_elem *
1710 bpf_hash_map_seq_find_next(struct bpf_iter_seq_hash_map_info *info,
1711                            struct htab_elem *prev_elem)
1712 {
1713         const struct bpf_htab *htab = info->htab;
1714         u32 skip_elems = info->skip_elems;
1715         u32 bucket_id = info->bucket_id;
1716         struct hlist_nulls_head *head;
1717         struct hlist_nulls_node *n;
1718         struct htab_elem *elem;
1719         struct bucket *b;
1720         u32 i, count;
1721
1722         if (bucket_id >= htab->n_buckets)
1723                 return NULL;
1724
1725         /* try to find next elem in the same bucket */
1726         if (prev_elem) {
1727                 /* no update/deletion on this bucket, prev_elem should be still valid
1728                  * and we won't skip elements.
1729                  */
1730                 n = rcu_dereference_raw(hlist_nulls_next_rcu(&prev_elem->hash_node));
1731                 elem = hlist_nulls_entry_safe(n, struct htab_elem, hash_node);
1732                 if (elem)
1733                         return elem;
1734
1735                 /* not found, unlock and go to the next bucket */
1736                 b = &htab->buckets[bucket_id++];
1737                 rcu_read_unlock();
1738                 skip_elems = 0;
1739         }
1740
1741         for (i = bucket_id; i < htab->n_buckets; i++) {
1742                 b = &htab->buckets[i];
1743                 rcu_read_lock();
1744
1745                 count = 0;
1746                 head = &b->head;
1747                 hlist_nulls_for_each_entry_rcu(elem, n, head, hash_node) {
1748                         if (count >= skip_elems) {
1749                                 info->bucket_id = i;
1750                                 info->skip_elems = count;
1751                                 return elem;
1752                         }
1753                         count++;
1754                 }
1755
1756                 rcu_read_unlock();
1757                 skip_elems = 0;
1758         }
1759
1760         info->bucket_id = i;
1761         info->skip_elems = 0;
1762         return NULL;
1763 }
1764
1765 static void *bpf_hash_map_seq_start(struct seq_file *seq, loff_t *pos)
1766 {
1767         struct bpf_iter_seq_hash_map_info *info = seq->private;
1768         struct htab_elem *elem;
1769
1770         elem = bpf_hash_map_seq_find_next(info, NULL);
1771         if (!elem)
1772                 return NULL;
1773
1774         if (*pos == 0)
1775                 ++*pos;
1776         return elem;
1777 }
1778
1779 static void *bpf_hash_map_seq_next(struct seq_file *seq, void *v, loff_t *pos)
1780 {
1781         struct bpf_iter_seq_hash_map_info *info = seq->private;
1782
1783         ++*pos;
1784         ++info->skip_elems;
1785         return bpf_hash_map_seq_find_next(info, v);
1786 }
1787
1788 static int __bpf_hash_map_seq_show(struct seq_file *seq, struct htab_elem *elem)
1789 {
1790         struct bpf_iter_seq_hash_map_info *info = seq->private;
1791         u32 roundup_key_size, roundup_value_size;
1792         struct bpf_iter__bpf_map_elem ctx = {};
1793         struct bpf_map *map = info->map;
1794         struct bpf_iter_meta meta;
1795         int ret = 0, off = 0, cpu;
1796         struct bpf_prog *prog;
1797         void __percpu *pptr;
1798
1799         meta.seq = seq;
1800         prog = bpf_iter_get_info(&meta, elem == NULL);
1801         if (prog) {
1802                 ctx.meta = &meta;
1803                 ctx.map = info->map;
1804                 if (elem) {
1805                         roundup_key_size = round_up(map->key_size, 8);
1806                         ctx.key = elem->key;
1807                         if (!info->percpu_value_buf) {
1808                                 ctx.value = elem->key + roundup_key_size;
1809                         } else {
1810                                 roundup_value_size = round_up(map->value_size, 8);
1811                                 pptr = htab_elem_get_ptr(elem, map->key_size);
1812                                 for_each_possible_cpu(cpu) {
1813                                         bpf_long_memcpy(info->percpu_value_buf + off,
1814                                                         per_cpu_ptr(pptr, cpu),
1815                                                         roundup_value_size);
1816                                         off += roundup_value_size;
1817                                 }
1818                                 ctx.value = info->percpu_value_buf;
1819                         }
1820                 }
1821                 ret = bpf_iter_run_prog(prog, &ctx);
1822         }
1823
1824         return ret;
1825 }
1826
1827 static int bpf_hash_map_seq_show(struct seq_file *seq, void *v)
1828 {
1829         return __bpf_hash_map_seq_show(seq, v);
1830 }
1831
1832 static void bpf_hash_map_seq_stop(struct seq_file *seq, void *v)
1833 {
1834         if (!v)
1835                 (void)__bpf_hash_map_seq_show(seq, NULL);
1836         else
1837                 rcu_read_unlock();
1838 }
1839
1840 static int bpf_iter_init_hash_map(void *priv_data,
1841                                   struct bpf_iter_aux_info *aux)
1842 {
1843         struct bpf_iter_seq_hash_map_info *seq_info = priv_data;
1844         struct bpf_map *map = aux->map;
1845         void *value_buf;
1846         u32 buf_size;
1847
1848         if (map->map_type == BPF_MAP_TYPE_PERCPU_HASH ||
1849             map->map_type == BPF_MAP_TYPE_LRU_PERCPU_HASH) {
1850                 buf_size = round_up(map->value_size, 8) * num_possible_cpus();
1851                 value_buf = kmalloc(buf_size, GFP_USER | __GFP_NOWARN);
1852                 if (!value_buf)
1853                         return -ENOMEM;
1854
1855                 seq_info->percpu_value_buf = value_buf;
1856         }
1857
1858         seq_info->map = map;
1859         seq_info->htab = container_of(map, struct bpf_htab, map);
1860         return 0;
1861 }
1862
1863 static void bpf_iter_fini_hash_map(void *priv_data)
1864 {
1865         struct bpf_iter_seq_hash_map_info *seq_info = priv_data;
1866
1867         kfree(seq_info->percpu_value_buf);
1868 }
1869
1870 static const struct seq_operations bpf_hash_map_seq_ops = {
1871         .start  = bpf_hash_map_seq_start,
1872         .next   = bpf_hash_map_seq_next,
1873         .stop   = bpf_hash_map_seq_stop,
1874         .show   = bpf_hash_map_seq_show,
1875 };
1876
1877 static const struct bpf_iter_seq_info iter_seq_info = {
1878         .seq_ops                = &bpf_hash_map_seq_ops,
1879         .init_seq_private       = bpf_iter_init_hash_map,
1880         .fini_seq_private       = bpf_iter_fini_hash_map,
1881         .seq_priv_size          = sizeof(struct bpf_iter_seq_hash_map_info),
1882 };
1883
1884 static int htab_map_btf_id;
1885 const struct bpf_map_ops htab_map_ops = {
1886         .map_meta_equal = bpf_map_meta_equal,
1887         .map_alloc_check = htab_map_alloc_check,
1888         .map_alloc = htab_map_alloc,
1889         .map_free = htab_map_free,
1890         .map_get_next_key = htab_map_get_next_key,
1891         .map_lookup_elem = htab_map_lookup_elem,
1892         .map_update_elem = htab_map_update_elem,
1893         .map_delete_elem = htab_map_delete_elem,
1894         .map_gen_lookup = htab_map_gen_lookup,
1895         .map_seq_show_elem = htab_map_seq_show_elem,
1896         BATCH_OPS(htab),
1897         .map_btf_name = "bpf_htab",
1898         .map_btf_id = &htab_map_btf_id,
1899         .iter_seq_info = &iter_seq_info,
1900 };
1901
1902 static int htab_lru_map_btf_id;
1903 const struct bpf_map_ops htab_lru_map_ops = {
1904         .map_meta_equal = bpf_map_meta_equal,
1905         .map_alloc_check = htab_map_alloc_check,
1906         .map_alloc = htab_map_alloc,
1907         .map_free = htab_map_free,
1908         .map_get_next_key = htab_map_get_next_key,
1909         .map_lookup_elem = htab_lru_map_lookup_elem,
1910         .map_lookup_elem_sys_only = htab_lru_map_lookup_elem_sys,
1911         .map_update_elem = htab_lru_map_update_elem,
1912         .map_delete_elem = htab_lru_map_delete_elem,
1913         .map_gen_lookup = htab_lru_map_gen_lookup,
1914         .map_seq_show_elem = htab_map_seq_show_elem,
1915         BATCH_OPS(htab_lru),
1916         .map_btf_name = "bpf_htab",
1917         .map_btf_id = &htab_lru_map_btf_id,
1918         .iter_seq_info = &iter_seq_info,
1919 };
1920
1921 /* Called from eBPF program */
1922 static void *htab_percpu_map_lookup_elem(struct bpf_map *map, void *key)
1923 {
1924         struct htab_elem *l = __htab_map_lookup_elem(map, key);
1925
1926         if (l)
1927                 return this_cpu_ptr(htab_elem_get_ptr(l, map->key_size));
1928         else
1929                 return NULL;
1930 }
1931
1932 static void *htab_lru_percpu_map_lookup_elem(struct bpf_map *map, void *key)
1933 {
1934         struct htab_elem *l = __htab_map_lookup_elem(map, key);
1935
1936         if (l) {
1937                 bpf_lru_node_set_ref(&l->lru_node);
1938                 return this_cpu_ptr(htab_elem_get_ptr(l, map->key_size));
1939         }
1940
1941         return NULL;
1942 }
1943
1944 int bpf_percpu_hash_copy(struct bpf_map *map, void *key, void *value)
1945 {
1946         struct htab_elem *l;
1947         void __percpu *pptr;
1948         int ret = -ENOENT;
1949         int cpu, off = 0;
1950         u32 size;
1951
1952         /* per_cpu areas are zero-filled and bpf programs can only
1953          * access 'value_size' of them, so copying rounded areas
1954          * will not leak any kernel data
1955          */
1956         size = round_up(map->value_size, 8);
1957         rcu_read_lock();
1958         l = __htab_map_lookup_elem(map, key);
1959         if (!l)
1960                 goto out;
1961         /* We do not mark LRU map element here in order to not mess up
1962          * eviction heuristics when user space does a map walk.
1963          */
1964         pptr = htab_elem_get_ptr(l, map->key_size);
1965         for_each_possible_cpu(cpu) {
1966                 bpf_long_memcpy(value + off,
1967                                 per_cpu_ptr(pptr, cpu), size);
1968                 off += size;
1969         }
1970         ret = 0;
1971 out:
1972         rcu_read_unlock();
1973         return ret;
1974 }
1975
1976 int bpf_percpu_hash_update(struct bpf_map *map, void *key, void *value,
1977                            u64 map_flags)
1978 {
1979         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
1980         int ret;
1981
1982         rcu_read_lock();
1983         if (htab_is_lru(htab))
1984                 ret = __htab_lru_percpu_map_update_elem(map, key, value,
1985                                                         map_flags, true);
1986         else
1987                 ret = __htab_percpu_map_update_elem(map, key, value, map_flags,
1988                                                     true);
1989         rcu_read_unlock();
1990
1991         return ret;
1992 }
1993
1994 static void htab_percpu_map_seq_show_elem(struct bpf_map *map, void *key,
1995                                           struct seq_file *m)
1996 {
1997         struct htab_elem *l;
1998         void __percpu *pptr;
1999         int cpu;
2000
2001         rcu_read_lock();
2002
2003         l = __htab_map_lookup_elem(map, key);
2004         if (!l) {
2005                 rcu_read_unlock();
2006                 return;
2007         }
2008
2009         btf_type_seq_show(map->btf, map->btf_key_type_id, key, m);
2010         seq_puts(m, ": {\n");
2011         pptr = htab_elem_get_ptr(l, map->key_size);
2012         for_each_possible_cpu(cpu) {
2013                 seq_printf(m, "\tcpu%d: ", cpu);
2014                 btf_type_seq_show(map->btf, map->btf_value_type_id,
2015                                   per_cpu_ptr(pptr, cpu), m);
2016                 seq_puts(m, "\n");
2017         }
2018         seq_puts(m, "}\n");
2019
2020         rcu_read_unlock();
2021 }
2022
2023 static int htab_percpu_map_btf_id;
2024 const struct bpf_map_ops htab_percpu_map_ops = {
2025         .map_meta_equal = bpf_map_meta_equal,
2026         .map_alloc_check = htab_map_alloc_check,
2027         .map_alloc = htab_map_alloc,
2028         .map_free = htab_map_free,
2029         .map_get_next_key = htab_map_get_next_key,
2030         .map_lookup_elem = htab_percpu_map_lookup_elem,
2031         .map_update_elem = htab_percpu_map_update_elem,
2032         .map_delete_elem = htab_map_delete_elem,
2033         .map_seq_show_elem = htab_percpu_map_seq_show_elem,
2034         BATCH_OPS(htab_percpu),
2035         .map_btf_name = "bpf_htab",
2036         .map_btf_id = &htab_percpu_map_btf_id,
2037         .iter_seq_info = &iter_seq_info,
2038 };
2039
2040 static int htab_lru_percpu_map_btf_id;
2041 const struct bpf_map_ops htab_lru_percpu_map_ops = {
2042         .map_meta_equal = bpf_map_meta_equal,
2043         .map_alloc_check = htab_map_alloc_check,
2044         .map_alloc = htab_map_alloc,
2045         .map_free = htab_map_free,
2046         .map_get_next_key = htab_map_get_next_key,
2047         .map_lookup_elem = htab_lru_percpu_map_lookup_elem,
2048         .map_update_elem = htab_lru_percpu_map_update_elem,
2049         .map_delete_elem = htab_lru_map_delete_elem,
2050         .map_seq_show_elem = htab_percpu_map_seq_show_elem,
2051         BATCH_OPS(htab_lru_percpu),
2052         .map_btf_name = "bpf_htab",
2053         .map_btf_id = &htab_lru_percpu_map_btf_id,
2054         .iter_seq_info = &iter_seq_info,
2055 };
2056
2057 static int fd_htab_map_alloc_check(union bpf_attr *attr)
2058 {
2059         if (attr->value_size != sizeof(u32))
2060                 return -EINVAL;
2061         return htab_map_alloc_check(attr);
2062 }
2063
2064 static void fd_htab_map_free(struct bpf_map *map)
2065 {
2066         struct bpf_htab *htab = container_of(map, struct bpf_htab, map);
2067         struct hlist_nulls_node *n;
2068         struct hlist_nulls_head *head;
2069         struct htab_elem *l;
2070         int i;
2071
2072         for (i = 0; i < htab->n_buckets; i++) {
2073                 head = select_bucket(htab, i);
2074
2075                 hlist_nulls_for_each_entry_safe(l, n, head, hash_node) {
2076                         void *ptr = fd_htab_map_get_ptr(map, l);
2077
2078                         map->ops->map_fd_put_ptr(ptr);
2079                 }
2080         }
2081
2082         htab_map_free(map);
2083 }
2084
2085 /* only called from syscall */
2086 int bpf_fd_htab_map_lookup_elem(struct bpf_map *map, void *key, u32 *value)
2087 {
2088         void **ptr;
2089         int ret = 0;
2090
2091         if (!map->ops->map_fd_sys_lookup_elem)
2092                 return -ENOTSUPP;
2093
2094         rcu_read_lock();
2095         ptr = htab_map_lookup_elem(map, key);
2096         if (ptr)
2097                 *value = map->ops->map_fd_sys_lookup_elem(READ_ONCE(*ptr));
2098         else
2099                 ret = -ENOENT;
2100         rcu_read_unlock();
2101
2102         return ret;
2103 }
2104
2105 /* only called from syscall */
2106 int bpf_fd_htab_map_update_elem(struct bpf_map *map, struct file *map_file,
2107                                 void *key, void *value, u64 map_flags)
2108 {
2109         void *ptr;
2110         int ret;
2111         u32 ufd = *(u32 *)value;
2112
2113         ptr = map->ops->map_fd_get_ptr(map, map_file, ufd);
2114         if (IS_ERR(ptr))
2115                 return PTR_ERR(ptr);
2116
2117         ret = htab_map_update_elem(map, key, &ptr, map_flags);
2118         if (ret)
2119                 map->ops->map_fd_put_ptr(ptr);
2120
2121         return ret;
2122 }
2123
2124 static struct bpf_map *htab_of_map_alloc(union bpf_attr *attr)
2125 {
2126         struct bpf_map *map, *inner_map_meta;
2127
2128         inner_map_meta = bpf_map_meta_alloc(attr->inner_map_fd);
2129         if (IS_ERR(inner_map_meta))
2130                 return inner_map_meta;
2131
2132         map = htab_map_alloc(attr);
2133         if (IS_ERR(map)) {
2134                 bpf_map_meta_free(inner_map_meta);
2135                 return map;
2136         }
2137
2138         map->inner_map_meta = inner_map_meta;
2139
2140         return map;
2141 }
2142
2143 static void *htab_of_map_lookup_elem(struct bpf_map *map, void *key)
2144 {
2145         struct bpf_map **inner_map  = htab_map_lookup_elem(map, key);
2146
2147         if (!inner_map)
2148                 return NULL;
2149
2150         return READ_ONCE(*inner_map);
2151 }
2152
2153 static int htab_of_map_gen_lookup(struct bpf_map *map,
2154                                   struct bpf_insn *insn_buf)
2155 {
2156         struct bpf_insn *insn = insn_buf;
2157         const int ret = BPF_REG_0;
2158
2159         BUILD_BUG_ON(!__same_type(&__htab_map_lookup_elem,
2160                      (void *(*)(struct bpf_map *map, void *key))NULL));
2161         *insn++ = BPF_EMIT_CALL(BPF_CAST_CALL(__htab_map_lookup_elem));
2162         *insn++ = BPF_JMP_IMM(BPF_JEQ, ret, 0, 2);
2163         *insn++ = BPF_ALU64_IMM(BPF_ADD, ret,
2164                                 offsetof(struct htab_elem, key) +
2165                                 round_up(map->key_size, 8));
2166         *insn++ = BPF_LDX_MEM(BPF_DW, ret, ret, 0);
2167
2168         return insn - insn_buf;
2169 }
2170
2171 static void htab_of_map_free(struct bpf_map *map)
2172 {
2173         bpf_map_meta_free(map->inner_map_meta);
2174         fd_htab_map_free(map);
2175 }
2176
2177 static int htab_of_maps_map_btf_id;
2178 const struct bpf_map_ops htab_of_maps_map_ops = {
2179         .map_alloc_check = fd_htab_map_alloc_check,
2180         .map_alloc = htab_of_map_alloc,
2181         .map_free = htab_of_map_free,
2182         .map_get_next_key = htab_map_get_next_key,
2183         .map_lookup_elem = htab_of_map_lookup_elem,
2184         .map_delete_elem = htab_map_delete_elem,
2185         .map_fd_get_ptr = bpf_map_fd_get_ptr,
2186         .map_fd_put_ptr = bpf_map_fd_put_ptr,
2187         .map_fd_sys_lookup_elem = bpf_map_fd_sys_lookup_elem,
2188         .map_gen_lookup = htab_of_map_gen_lookup,
2189         .map_check_btf = map_check_no_btf,
2190         .map_btf_name = "bpf_htab",
2191         .map_btf_id = &htab_of_maps_map_btf_id,
2192 };