lib/rhashtable.c

   1 /*
   2  * Resizable, Scalable, Concurrent Hash Table
   3  *
   4  * Copyright (c) 2014 Thomas Graf <tgraf@suug.ch>
   5  * Copyright (c) 2008-2014 Patrick McHardy <kaber@trash.net>
   6  *
   7  * Based on the following paper:
   8  * https://www.usenix.org/legacy/event/atc11/tech/final_files/Triplett.pdf
   9  *
  10  * Code partially derived from nft_hash
  11  *
  12  * This program is free software; you can redistribute it and/or modify
  13  * it under the terms of the GNU General Public License version 2 as
  14  * published by the Free Software Foundation.
  15  */
  16
  17 #include <linux/kernel.h>
  18 #include <linux/init.h>
  19 #include <linux/log2.h>
  20 #include <linux/slab.h>
  21 #include <linux/vmalloc.h>
  22 #include <linux/mm.h>
  23 #include <linux/jhash.h>
  24 #include <linux/random.h>
  25 #include <linux/rhashtable.h>
  26
  27 #define HASH_DEFAULT_SIZE       64UL
  28 #define HASH_MIN_SIZE           4UL
  29 #define BUCKET_LOCKS_PER_CPU   128UL
  30
  31 /* Base bits plus 1 bit for nulls marker */
  32 #define HASH_RESERVED_SPACE     (RHT_BASE_BITS + 1)
  33
  34 enum {
  35         RHT_LOCK_NORMAL,
  36         RHT_LOCK_NESTED,
  37         RHT_LOCK_NESTED2,
  38 };
  39
  40 /* The bucket lock is selected based on the hash and protects mutations
  41  * on a group of hash buckets.
  42  *
  43  * IMPORTANT: When holding the bucket lock of both the old and new table
  44  * during expansions and shrinking, the old bucket lock must always be
  45  * acquired first.
  46  */
  47 static spinlock_t *bucket_lock(const struct bucket_table *tbl, u32 hash)
  48 {
  49         return &tbl->locks[hash & tbl->locks_mask];
  50 }
  51
  52 #define ASSERT_RHT_MUTEX(HT) BUG_ON(!lockdep_rht_mutex_is_held(HT))
  53 #define ASSERT_BUCKET_LOCK(TBL, HASH) \
  54         BUG_ON(!lockdep_rht_bucket_is_held(TBL, HASH))
  55
  56 #ifdef CONFIG_PROVE_LOCKING
  57 int lockdep_rht_mutex_is_held(struct rhashtable *ht)
  58 {
  59         return (debug_locks) ? lockdep_is_held(&ht->mutex) : 1;
  60 }
  61 EXPORT_SYMBOL_GPL(lockdep_rht_mutex_is_held);
  62
  63 int lockdep_rht_bucket_is_held(const struct bucket_table *tbl, u32 hash)
  64 {
  65         spinlock_t *lock = bucket_lock(tbl, hash);
  66
  67         return (debug_locks) ? lockdep_is_held(lock) : 1;
  68 }
  69 EXPORT_SYMBOL_GPL(lockdep_rht_bucket_is_held);
  70 #endif
  71
  72 static void *rht_obj(const struct rhashtable *ht, const struct rhash_head *he)
  73 {
  74         return (void *) he - ht->p.head_offset;
  75 }
  76
  77 static u32 rht_bucket_index(const struct bucket_table *tbl, u32 hash)
  78 {
  79         return hash & (tbl->size - 1);
  80 }
  81
  82 static u32 obj_raw_hashfn(const struct rhashtable *ht, const void *ptr)
  83 {
  84         u32 hash;
  85
  86         if (unlikely(!ht->p.key_len))
  87                 hash = ht->p.obj_hashfn(ptr, ht->p.hash_rnd);
  88         else
  89                 hash = ht->p.hashfn(ptr + ht->p.key_offset, ht->p.key_len,
  90                                     ht->p.hash_rnd);
  91
  92         return hash >> HASH_RESERVED_SPACE;
  93 }
  94
  95 static u32 key_hashfn(struct rhashtable *ht, const void *key, u32 len)
  96 {
  97         struct bucket_table *tbl = rht_dereference_rcu(ht->tbl, ht);
  98         u32 hash;
  99
 100         hash = ht->p.hashfn(key, len, ht->p.hash_rnd);
 101         hash >>= HASH_RESERVED_SPACE;
 102
 103         return rht_bucket_index(tbl, hash);
 104 }
 105
 106 static u32 head_hashfn(const struct rhashtable *ht,
 107                        const struct bucket_table *tbl,
 108                        const struct rhash_head *he)
 109 {
 110         return rht_bucket_index(tbl, obj_raw_hashfn(ht, rht_obj(ht, he)));
 111 }
 112
 113 static struct rhash_head __rcu **bucket_tail(struct bucket_table *tbl, u32 n)
 114 {
 115         struct rhash_head __rcu **pprev;
 116
 117         for (pprev = &tbl->buckets[n];
 118              !rht_is_a_nulls(rht_dereference_bucket(*pprev, tbl, n));
 119              pprev = &rht_dereference_bucket(*pprev, tbl, n)->next)
 120                 ;
 121
 122         return pprev;
 123 }
 124
 125 static int alloc_bucket_locks(struct rhashtable *ht, struct bucket_table *tbl)
 126 {
 127         unsigned int i, size;
 128 #if defined(CONFIG_PROVE_LOCKING)
 129         unsigned int nr_pcpus = 2;
 130 #else
 131         unsigned int nr_pcpus = num_possible_cpus();
 132 #endif
 133
 134         nr_pcpus = min_t(unsigned int, nr_pcpus, 32UL);
 135         size = roundup_pow_of_two(nr_pcpus * ht->p.locks_mul);
 136
 137         /* Never allocate more than one lock per bucket */
 138         size = min_t(unsigned int, size, tbl->size);
 139
 140         if (sizeof(spinlock_t) != 0) {
 141 #ifdef CONFIG_NUMA
 142                 if (size * sizeof(spinlock_t) > PAGE_SIZE)
 143                         tbl->locks = vmalloc(size * sizeof(spinlock_t));
 144                 else
 145 #endif
 146                 tbl->locks = kmalloc_array(size, sizeof(spinlock_t),
 147                                            GFP_KERNEL);
 148                 if (!tbl->locks)
 149                         return -ENOMEM;
 150                 for (i = 0; i < size; i++)
 151                         spin_lock_init(&tbl->locks[i]);
 152         }
 153         tbl->locks_mask = size - 1;
 154
 155         return 0;
 156 }
 157
 158 static void bucket_table_free(const struct bucket_table *tbl)
 159 {
 160         if (tbl)
 161                 kvfree(tbl->locks);
 162
 163         kvfree(tbl);
 164 }
 165
 166 static struct bucket_table *bucket_table_alloc(struct rhashtable *ht,
 167                                                size_t nbuckets)
 168 {
 169         struct bucket_table *tbl;
 170         size_t size;
 171         int i;
 172
 173         size = sizeof(*tbl) + nbuckets * sizeof(tbl->buckets[0]);
 174         tbl = kzalloc(size, GFP_KERNEL | __GFP_NOWARN);
 175         if (tbl == NULL)
 176                 tbl = vzalloc(size);
 177
 178         if (tbl == NULL)
 179                 return NULL;
 180
 181         tbl->size = nbuckets;
 182
 183         if (alloc_bucket_locks(ht, tbl) < 0) {
 184                 bucket_table_free(tbl);
 185                 return NULL;
 186         }
 187
 188         for (i = 0; i < nbuckets; i++)
 189                 INIT_RHT_NULLS_HEAD(tbl->buckets[i], ht, i);
 190
 191         return tbl;
 192 }
 193
 194 /**
 195  * rht_grow_above_75 - returns true if nelems > 0.75 * table-size
 196  * @ht:         hash table
 197  * @new_size:   new table size
 198  */
 199 bool rht_grow_above_75(const struct rhashtable *ht, size_t new_size)
 200 {
 201         /* Expand table when exceeding 75% load */
 202         return atomic_read(&ht->nelems) > (new_size / 4 * 3) &&
 203                (ht->p.max_shift && atomic_read(&ht->shift) < ht->p.max_shift);
 204 }
 205 EXPORT_SYMBOL_GPL(rht_grow_above_75);
 206
 207 /**
 208  * rht_shrink_below_30 - returns true if nelems < 0.3 * table-size
 209  * @ht:         hash table
 210  * @new_size:   new table size
 211  */
 212 bool rht_shrink_below_30(const struct rhashtable *ht, size_t new_size)
 213 {
 214         /* Shrink table beneath 30% load */
 215         return atomic_read(&ht->nelems) < (new_size * 3 / 10) &&
 216                (atomic_read(&ht->shift) > ht->p.min_shift);
 217 }
 218 EXPORT_SYMBOL_GPL(rht_shrink_below_30);
 219
 220 static void hashtable_chain_unzip(const struct rhashtable *ht,
 221                                   const struct bucket_table *new_tbl,
 222                                   struct bucket_table *old_tbl,
 223                                   size_t old_hash)
 224 {
 225         struct rhash_head *he, *p, *next;
 226         spinlock_t *new_bucket_lock, *new_bucket_lock2 = NULL;
 227         unsigned int new_hash, new_hash2;
 228
 229         ASSERT_BUCKET_LOCK(old_tbl, old_hash);
 230
 231         /* Old bucket empty, no work needed. */
 232         p = rht_dereference_bucket(old_tbl->buckets[old_hash], old_tbl,
 233                                    old_hash);
 234         if (rht_is_a_nulls(p))
 235                 return;
 236
 237         new_hash = new_hash2 = head_hashfn(ht, new_tbl, p);
 238         new_bucket_lock = bucket_lock(new_tbl, new_hash);
 239
 240         /* Advance the old bucket pointer one or more times until it
 241          * reaches a node that doesn't hash to the same bucket as the
 242          * previous node p. Call the previous node p;
 243          */
 244         rht_for_each_continue(he, p->next, old_tbl, old_hash) {
 245                 new_hash2 = head_hashfn(ht, new_tbl, he);
 246                 if (new_hash != new_hash2)
 247                         break;
 248                 p = he;
 249         }
 250         rcu_assign_pointer(old_tbl->buckets[old_hash], p->next);
 251
 252         spin_lock_bh_nested(new_bucket_lock, RHT_LOCK_NESTED);
 253
 254         /* If we have encountered an entry that maps to a different bucket in
 255          * the new table, lock down that bucket as well as we might cut off
 256          * the end of the chain.
 257          */
 258         new_bucket_lock2 = bucket_lock(new_tbl, new_hash);
 259         if (new_bucket_lock != new_bucket_lock2)
 260                 spin_lock_bh_nested(new_bucket_lock2, RHT_LOCK_NESTED2);
 261
 262         /* Find the subsequent node which does hash to the same
 263          * bucket as node P, or NULL if no such node exists.
 264          */
 265         INIT_RHT_NULLS_HEAD(next, ht, old_hash);
 266         if (!rht_is_a_nulls(he)) {
 267                 rht_for_each_continue(he, he->next, old_tbl, old_hash) {
 268                         if (head_hashfn(ht, new_tbl, he) == new_hash) {
 269                                 next = he;
 270                                 break;
 271                         }
 272                 }
 273         }
 274
 275         /* Set p's next pointer to that subsequent node pointer,
 276          * bypassing the nodes which do not hash to p's bucket
 277          */
 278         rcu_assign_pointer(p->next, next);
 279
 280         if (new_bucket_lock != new_bucket_lock2)
 281                 spin_unlock_bh(new_bucket_lock2);
 282         spin_unlock_bh(new_bucket_lock);
 283 }
 284
 285 static void link_old_to_new(struct bucket_table *new_tbl,
 286                             unsigned int new_hash, struct rhash_head *entry)
 287 {
 288         spinlock_t *new_bucket_lock;
 289
 290         new_bucket_lock = bucket_lock(new_tbl, new_hash);
 291
 292         spin_lock_bh_nested(new_bucket_lock, RHT_LOCK_NESTED);
 293         rcu_assign_pointer(*bucket_tail(new_tbl, new_hash), entry);
 294         spin_unlock_bh(new_bucket_lock);
 295 }
 296
 297 /**
 298  * rhashtable_expand - Expand hash table while allowing concurrent lookups
 299  * @ht:         the hash table to expand
 300  *
 301  * A secondary bucket array is allocated and the hash entries are migrated
 302  * while keeping them on both lists until the end of the RCU grace period.
 303  *
 304  * This function may only be called in a context where it is safe to call
 305  * synchronize_rcu(), e.g. not within a rcu_read_lock() section.
 306  *
 307  * The caller must ensure that no concurrent resizing occurs by holding
 308  * ht->mutex.
 309  *
 310  * It is valid to have concurrent insertions and deletions protected by per
 311  * bucket locks or concurrent RCU protected lookups and traversals.
 312  */
 313 int rhashtable_expand(struct rhashtable *ht)
 314 {
 315         struct bucket_table *new_tbl, *old_tbl = rht_dereference(ht->tbl, ht);
 316         struct rhash_head *he;
 317         spinlock_t *old_bucket_lock;
 318         unsigned int new_hash, old_hash;
 319         bool complete = false;
 320
 321         ASSERT_RHT_MUTEX(ht);
 322
 323         new_tbl = bucket_table_alloc(ht, old_tbl->size * 2);
 324         if (new_tbl == NULL)
 325                 return -ENOMEM;
 326
 327         atomic_inc(&ht->shift);
 328
 329         /* Make insertions go into the new, empty table right away. Deletions
 330          * and lookups will be attempted in both tables until we synchronize.
 331          * The synchronize_rcu() guarantees for the new table to be picked up
 332          * so no new additions go into the old table while we relink.
 333          */
 334         rcu_assign_pointer(ht->future_tbl, new_tbl);
 335         synchronize_rcu();
 336
 337         /* For each new bucket, search the corresponding old bucket for the
 338          * first entry that hashes to the new bucket, and link the end of
 339          * newly formed bucket chain (containing entries added to future
 340          * table) to that entry. Since all the entries which will end up in
 341          * the new bucket appear in the same old bucket, this constructs an
 342          * entirely valid new hash table, but with multiple buckets
 343          * "zipped" together into a single imprecise chain.
 344          */
 345         for (new_hash = 0; new_hash < new_tbl->size; new_hash++) {
 346                 old_hash = rht_bucket_index(old_tbl, new_hash);
 347                 old_bucket_lock = bucket_lock(old_tbl, old_hash);
 348
 349                 spin_lock_bh(old_bucket_lock);
 350                 rht_for_each(he, old_tbl, old_hash) {
 351                         if (head_hashfn(ht, new_tbl, he) == new_hash) {
 352                                 link_old_to_new(new_tbl, new_hash, he);
 353                                 break;
 354                         }
 355                 }
 356                 spin_unlock_bh(old_bucket_lock);
 357         }
 358
 359         /* Publish the new table pointer. Lookups may now traverse
 360          * the new table, but they will not benefit from any
 361          * additional efficiency until later steps unzip the buckets.
 362          */
 363         rcu_assign_pointer(ht->tbl, new_tbl);
 364
 365         /* Unzip interleaved hash chains */
 366         while (!complete && !ht->being_destroyed) {
 367                 /* Wait for readers. All new readers will see the new
 368                  * table, and thus no references to the old table will
 369                  * remain.
 370                  */
 371                 synchronize_rcu();
 372
 373                 /* For each bucket in the old table (each of which
 374                  * contains items from multiple buckets of the new
 375                  * table): ...
 376                  */
 377                 complete = true;
 378                 for (old_hash = 0; old_hash < old_tbl->size; old_hash++) {
 379                         struct rhash_head *head;
 380
 381                         old_bucket_lock = bucket_lock(old_tbl, old_hash);
 382                         spin_lock_bh(old_bucket_lock);
 383
 384                         hashtable_chain_unzip(ht, new_tbl, old_tbl, old_hash);
 385                         head = rht_dereference_bucket(old_tbl->buckets[old_hash],
 386                                                       old_tbl, old_hash);
 387                         if (!rht_is_a_nulls(head))
 388                                 complete = false;
 389
 390                         spin_unlock_bh(old_bucket_lock);
 391                 }
 392         }
 393
 394         bucket_table_free(old_tbl);
 395         return 0;
 396 }
 397 EXPORT_SYMBOL_GPL(rhashtable_expand);
 398
 399 /**
 400  * rhashtable_shrink - Shrink hash table while allowing concurrent lookups
 401  * @ht:         the hash table to shrink
 402  *
 403  * This function may only be called in a context where it is safe to call
 404  * synchronize_rcu(), e.g. not within a rcu_read_lock() section.
 405  *
 406  * The caller must ensure that no concurrent resizing occurs by holding
 407  * ht->mutex.
 408  *
 409  * The caller must ensure that no concurrent table mutations take place.
 410  * It is however valid to have concurrent lookups if they are RCU protected.
 411  *
 412  * It is valid to have concurrent insertions and deletions protected by per
 413  * bucket locks or concurrent RCU protected lookups and traversals.
 414  */
 415 int rhashtable_shrink(struct rhashtable *ht)
 416 {
 417         struct bucket_table *new_tbl, *tbl = rht_dereference(ht->tbl, ht);
 418         spinlock_t *new_bucket_lock, *old_bucket_lock1, *old_bucket_lock2;
 419         unsigned int new_hash;
 420
 421         ASSERT_RHT_MUTEX(ht);
 422
 423         new_tbl = bucket_table_alloc(ht, tbl->size / 2);
 424         if (new_tbl == NULL)
 425                 return -ENOMEM;
 426
 427         rcu_assign_pointer(ht->future_tbl, new_tbl);
 428         synchronize_rcu();
 429
 430         /* Link the first entry in the old bucket to the end of the
 431          * bucket in the new table. As entries are concurrently being
 432          * added to the new table, lock down the new bucket. As we
 433          * always divide the size in half when shrinking, each bucket
 434          * in the new table maps to exactly two buckets in the old
 435          * table.
 436          *
 437          * As removals can occur concurrently on the old table, we need
 438          * to lock down both matching buckets in the old table.
 439          */
 440         for (new_hash = 0; new_hash < new_tbl->size; new_hash++) {
 441                 old_bucket_lock1 = bucket_lock(tbl, new_hash);
 442                 old_bucket_lock2 = bucket_lock(tbl, new_hash + new_tbl->size);
 443                 new_bucket_lock = bucket_lock(new_tbl, new_hash);
 444
 445                 spin_lock_bh(old_bucket_lock1);
 446
 447                 /* Depending on the lock per buckets mapping, the bucket in
 448                  * the lower and upper region may map to the same lock.
 449                  */
 450                 if (old_bucket_lock1 != old_bucket_lock2) {
 451                         spin_lock_bh_nested(old_bucket_lock2, RHT_LOCK_NESTED);
 452                         spin_lock_bh_nested(new_bucket_lock, RHT_LOCK_NESTED2);
 453                 } else {
 454                         spin_lock_bh_nested(new_bucket_lock, RHT_LOCK_NESTED);
 455                 }
 456
 457                 rcu_assign_pointer(*bucket_tail(new_tbl, new_hash),
 458                                    tbl->buckets[new_hash]);
 459                 rcu_assign_pointer(*bucket_tail(new_tbl, new_hash),
 460                                    tbl->buckets[new_hash + new_tbl->size]);
 461
 462                 spin_unlock_bh(new_bucket_lock);
 463                 if (old_bucket_lock1 != old_bucket_lock2)
 464                         spin_unlock_bh(old_bucket_lock2);
 465                 spin_unlock_bh(old_bucket_lock1);
 466         }
 467
 468         /* Publish the new, valid hash table */
 469         rcu_assign_pointer(ht->tbl, new_tbl);
 470         atomic_dec(&ht->shift);
 471
 472         /* Wait for readers. No new readers will have references to the
 473          * old hash table.
 474          */
 475         synchronize_rcu();
 476
 477         bucket_table_free(tbl);
 478
 479         return 0;
 480 }
 481 EXPORT_SYMBOL_GPL(rhashtable_shrink);
 482
 483 static void rht_deferred_worker(struct work_struct *work)
 484 {
 485         struct rhashtable *ht;
 486         struct bucket_table *tbl;
 487
 488         ht = container_of(work, struct rhashtable, run_work);
 489         mutex_lock(&ht->mutex);
 490         tbl = rht_dereference(ht->tbl, ht);
 491
 492         if (ht->p.grow_decision && ht->p.grow_decision(ht, tbl->size))
 493                 rhashtable_expand(ht);
 494         else if (ht->p.shrink_decision && ht->p.shrink_decision(ht, tbl->size))
 495                 rhashtable_shrink(ht);
 496
 497         mutex_unlock(&ht->mutex);
 498 }
 499
 500 static void rhashtable_wakeup_worker(struct rhashtable *ht)
 501 {
 502         struct bucket_table *tbl = rht_dereference_rcu(ht->tbl, ht);
 503         struct bucket_table *new_tbl = rht_dereference_rcu(ht->future_tbl, ht);
 504         size_t size = tbl->size;
 505
 506         /* Only adjust the table if no resizing is currently in progress. */
 507         if (tbl == new_tbl &&
 508             ((ht->p.grow_decision && ht->p.grow_decision(ht, size)) ||
 509              (ht->p.shrink_decision && ht->p.shrink_decision(ht, size))))
 510                 schedule_work(&ht->run_work);
 511 }
 512
 513 static void __rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj,
 514                                 struct bucket_table *tbl, u32 hash)
 515 {
 516         struct rhash_head *head = rht_dereference_bucket(tbl->buckets[hash],
 517                                                          tbl, hash);
 518
 519         if (rht_is_a_nulls(head))
 520                 INIT_RHT_NULLS_HEAD(obj->next, ht, hash);
 521         else
 522                 RCU_INIT_POINTER(obj->next, head);
 523
 524         rcu_assign_pointer(tbl->buckets[hash], obj);
 525
 526         atomic_inc(&ht->nelems);
 527
 528         rhashtable_wakeup_worker(ht);
 529 }
 530
 531 /**
 532  * rhashtable_insert - insert object into hash table
 533  * @ht:         hash table
 534  * @obj:        pointer to hash head inside object
 535  *
 536  * Will take a per bucket spinlock to protect against mutual mutations
 537  * on the same bucket. Multiple insertions may occur in parallel unless
 538  * they map to the same bucket lock.
 539  *
 540  * It is safe to call this function from atomic context.
 541  *
 542  * Will trigger an automatic deferred table resizing if the size grows
 543  * beyond the watermark indicated by grow_decision() which can be passed
 544  * to rhashtable_init().
 545  */
 546 void rhashtable_insert(struct rhashtable *ht, struct rhash_head *obj)
 547 {
 548         struct bucket_table *tbl;
 549         spinlock_t *lock;
 550         unsigned hash;
 551
 552         rcu_read_lock();
 553
 554         tbl = rht_dereference_rcu(ht->future_tbl, ht);
 555         hash = head_hashfn(ht, tbl, obj);
 556         lock = bucket_lock(tbl, hash);
 557
 558         spin_lock_bh(lock);
 559         __rhashtable_insert(ht, obj, tbl, hash);
 560         spin_unlock_bh(lock);
 561
 562         rcu_read_unlock();
 563 }
 564 EXPORT_SYMBOL_GPL(rhashtable_insert);
 565
 566 /**
 567  * rhashtable_remove - remove object from hash table
 568  * @ht:         hash table
 569  * @obj:        pointer to hash head inside object
 570  *
 571  * Since the hash chain is single linked, the removal operation needs to
 572  * walk the bucket chain upon removal. The removal operation is thus
 573  * considerable slow if the hash table is not correctly sized.
 574  *
 575  * Will automatically shrink the table via rhashtable_expand() if the
 576  * shrink_decision function specified at rhashtable_init() returns true.
 577  *
 578  * The caller must ensure that no concurrent table mutations occur. It is
 579  * however valid to have concurrent lookups if they are RCU protected.
 580  */
 581 bool rhashtable_remove(struct rhashtable *ht, struct rhash_head *obj)
 582 {
 583         struct bucket_table *tbl;
 584         struct rhash_head __rcu **pprev;
 585         struct rhash_head *he;
 586         spinlock_t *lock;
 587         unsigned int hash;
 588         bool ret = false;
 589
 590         rcu_read_lock();
 591         tbl = rht_dereference_rcu(ht->tbl, ht);
 592         hash = head_hashfn(ht, tbl, obj);
 593
 594         lock = bucket_lock(tbl, hash);
 595         spin_lock_bh(lock);
 596
 597 restart:
 598         pprev = &tbl->buckets[hash];
 599         rht_for_each(he, tbl, hash) {
 600                 if (he != obj) {
 601                         pprev = &he->next;
 602                         continue;
 603                 }
 604
 605                 rcu_assign_pointer(*pprev, obj->next);
 606
 607                 ret = true;
 608                 break;
 609         }
 610
 611         /* The entry may be linked in either 'tbl', 'future_tbl', or both.
 612          * 'future_tbl' only exists for a short period of time during
 613          * resizing. Thus traversing both is fine and the added cost is
 614          * very rare.
 615          */
 616         if (tbl != rht_dereference_rcu(ht->future_tbl, ht)) {
 617                 spin_unlock_bh(lock);
 618
 619                 tbl = rht_dereference_rcu(ht->future_tbl, ht);
 620                 hash = head_hashfn(ht, tbl, obj);
 621
 622                 lock = bucket_lock(tbl, hash);
 623                 spin_lock_bh(lock);
 624                 goto restart;
 625         }
 626
 627         spin_unlock_bh(lock);
 628
 629         if (ret) {
 630                 atomic_dec(&ht->nelems);
 631                 rhashtable_wakeup_worker(ht);
 632         }
 633
 634         rcu_read_unlock();
 635
 636         return ret;
 637 }
 638 EXPORT_SYMBOL_GPL(rhashtable_remove);
 639
 640 struct rhashtable_compare_arg {
 641         struct rhashtable *ht;
 642         const void *key;
 643 };
 644
 645 static bool rhashtable_compare(void *ptr, void *arg)
 646 {
 647         struct rhashtable_compare_arg *x = arg;
 648         struct rhashtable *ht = x->ht;
 649
 650         return !memcmp(ptr + ht->p.key_offset, x->key, ht->p.key_len);
 651 }
 652
 653 /**
 654  * rhashtable_lookup - lookup key in hash table
 655  * @ht:         hash table
 656  * @key:        pointer to key
 657  *
 658  * Computes the hash value for the key and traverses the bucket chain looking
 659  * for a entry with an identical key. The first matching entry is returned.
 660  *
 661  * This lookup function may only be used for fixed key hash table (key_len
 662  * parameter set). It will BUG() if used inappropriately.
 663  *
 664  * Lookups may occur in parallel with hashtable mutations and resizing.
 665  */
 666 void *rhashtable_lookup(struct rhashtable *ht, const void *key)
 667 {
 668         struct rhashtable_compare_arg arg = {
 669                 .ht = ht,
 670                 .key = key,
 671         };
 672
 673         BUG_ON(!ht->p.key_len);
 674
 675         return rhashtable_lookup_compare(ht, key, &rhashtable_compare, &arg);
 676 }
 677 EXPORT_SYMBOL_GPL(rhashtable_lookup);
 678
 679 /**
 680  * rhashtable_lookup_compare - search hash table with compare function
 681  * @ht:         hash table
 682  * @key:        the pointer to the key
 683  * @compare:    compare function, must return true on match
 684  * @arg:        argument passed on to compare function
 685  *
 686  * Traverses the bucket chain behind the provided hash value and calls the
 687  * specified compare function for each entry.
 688  *
 689  * Lookups may occur in parallel with hashtable mutations and resizing.
 690  *
 691  * Returns the first entry on which the compare function returned true.
 692  */
 693 void *rhashtable_lookup_compare(struct rhashtable *ht, const void *key,
 694                                 bool (*compare)(void *, void *), void *arg)
 695 {
 696         const struct bucket_table *tbl, *old_tbl;
 697         struct rhash_head *he;
 698         u32 hash;
 699
 700         rcu_read_lock();
 701
 702         old_tbl = rht_dereference_rcu(ht->tbl, ht);
 703         tbl = rht_dereference_rcu(ht->future_tbl, ht);
 704         hash = key_hashfn(ht, key, ht->p.key_len);
 705 restart:
 706         rht_for_each_rcu(he, tbl, rht_bucket_index(tbl, hash)) {
 707                 if (!compare(rht_obj(ht, he), arg))
 708                         continue;
 709                 rcu_read_unlock();
 710                 return rht_obj(ht, he);
 711         }
 712
 713         if (unlikely(tbl != old_tbl)) {
 714                 tbl = old_tbl;
 715                 goto restart;
 716         }
 717         rcu_read_unlock();
 718
 719         return NULL;
 720 }
 721 EXPORT_SYMBOL_GPL(rhashtable_lookup_compare);
 722
 723 /**
 724  * rhashtable_lookup_insert - lookup and insert object into hash table
 725  * @ht:         hash table
 726  * @obj:        pointer to hash head inside object
 727  *
 728  * Locks down the bucket chain in both the old and new table if a resize
 729  * is in progress to ensure that writers can't remove from the old table
 730  * and can't insert to the new table during the atomic operation of search
 731  * and insertion. Searches for duplicates in both the old and new table if
 732  * a resize is in progress.
 733  *
 734  * This lookup function may only be used for fixed key hash table (key_len
 735  * parameter set). It will BUG() if used inappropriately.
 736  *
 737  * It is safe to call this function from atomic context.
 738  *
 739  * Will trigger an automatic deferred table resizing if the size grows
 740  * beyond the watermark indicated by grow_decision() which can be passed
 741  * to rhashtable_init().
 742  */
 743 bool rhashtable_lookup_insert(struct rhashtable *ht, struct rhash_head *obj)
 744 {
 745         struct rhashtable_compare_arg arg = {
 746                 .ht = ht,
 747                 .key = rht_obj(ht, obj) + ht->p.key_offset,
 748         };
 749
 750         BUG_ON(!ht->p.key_len);
 751
 752         return rhashtable_lookup_compare_insert(ht, obj, &rhashtable_compare,
 753                                                 &arg);
 754 }
 755 EXPORT_SYMBOL_GPL(rhashtable_lookup_insert);
 756
 757 /**
 758  * rhashtable_lookup_compare_insert - search and insert object to hash table
 759  *                                    with compare function
 760  * @ht:         hash table
 761  * @obj:        pointer to hash head inside object
 762  * @compare:    compare function, must return true on match
 763  * @arg:        argument passed on to compare function
 764  *
 765  * Locks down the bucket chain in both the old and new table if a resize
 766  * is in progress to ensure that writers can't remove from the old table
 767  * and can't insert to the new table during the atomic operation of search
 768  * and insertion. Searches for duplicates in both the old and new table if
 769  * a resize is in progress.
 770  *
 771  * Lookups may occur in parallel with hashtable mutations and resizing.
 772  *
 773  * Will trigger an automatic deferred table resizing if the size grows
 774  * beyond the watermark indicated by grow_decision() which can be passed
 775  * to rhashtable_init().
 776  */
 777 bool rhashtable_lookup_compare_insert(struct rhashtable *ht,
 778                                       struct rhash_head *obj,
 779                                       bool (*compare)(void *, void *),
 780                                       void *arg)
 781 {
 782         struct bucket_table *new_tbl, *old_tbl;
 783         spinlock_t *new_bucket_lock, *old_bucket_lock;
 784         u32 new_hash, old_hash;
 785         bool success = true;
 786
 787         BUG_ON(!ht->p.key_len);
 788
 789         rcu_read_lock();
 790
 791         old_tbl = rht_dereference_rcu(ht->tbl, ht);
 792         old_hash = head_hashfn(ht, old_tbl, obj);
 793         old_bucket_lock = bucket_lock(old_tbl, old_hash);
 794         spin_lock_bh(old_bucket_lock);
 795
 796         new_tbl = rht_dereference_rcu(ht->future_tbl, ht);
 797         new_hash = head_hashfn(ht, new_tbl, obj);
 798         new_bucket_lock = bucket_lock(new_tbl, new_hash);
 799         if (unlikely(old_tbl != new_tbl))
 800                 spin_lock_bh_nested(new_bucket_lock, RHT_LOCK_NESTED);
 801
 802         if (rhashtable_lookup_compare(ht, rht_obj(ht, obj) + ht->p.key_offset,
 803                                       compare, arg)) {
 804                 success = false;
 805                 goto exit;
 806         }
 807
 808         __rhashtable_insert(ht, obj, new_tbl, new_hash);
 809
 810 exit:
 811         if (unlikely(old_tbl != new_tbl))
 812                 spin_unlock_bh(new_bucket_lock);
 813         spin_unlock_bh(old_bucket_lock);
 814
 815         rcu_read_unlock();
 816
 817         return success;
 818 }
 819 EXPORT_SYMBOL_GPL(rhashtable_lookup_compare_insert);
 820
 821 static size_t rounded_hashtable_size(struct rhashtable_params *params)
 822 {
 823         return max(roundup_pow_of_two(params->nelem_hint * 4 / 3),
 824                    1UL << params->min_shift);
 825 }
 826
 827 /**
 828  * rhashtable_init - initialize a new hash table
 829  * @ht:         hash table to be initialized
 830  * @params:     configuration parameters
 831  *
 832  * Initializes a new hash table based on the provided configuration
 833  * parameters. A table can be configured either with a variable or
 834  * fixed length key:
 835  *
 836  * Configuration Example 1: Fixed length keys
 837  * struct test_obj {
 838  *      int                     key;
 839  *      void *                  my_member;
 840  *      struct rhash_head       node;
 841  * };
 842  *
 843  * struct rhashtable_params params = {
 844  *      .head_offset = offsetof(struct test_obj, node),
 845  *      .key_offset = offsetof(struct test_obj, key),
 846  *      .key_len = sizeof(int),
 847  *      .hashfn = jhash,
 848  *      .nulls_base = (1U << RHT_BASE_SHIFT),
 849  * };
 850  *
 851  * Configuration Example 2: Variable length keys
 852  * struct test_obj {
 853  *      [...]
 854  *      struct rhash_head       node;
 855  * };
 856  *
 857  * u32 my_hash_fn(const void *data, u32 seed)
 858  * {
 859  *      struct test_obj *obj = data;
 860  *
 861  *      return [... hash ...];
 862  * }
 863  *
 864  * struct rhashtable_params params = {
 865  *      .head_offset = offsetof(struct test_obj, node),
 866  *      .hashfn = jhash,
 867  *      .obj_hashfn = my_hash_fn,
 868  * };
 869  */
 870 int rhashtable_init(struct rhashtable *ht, struct rhashtable_params *params)
 871 {
 872         struct bucket_table *tbl;
 873         size_t size;
 874
 875         size = HASH_DEFAULT_SIZE;
 876
 877         if ((params->key_len && !params->hashfn) ||
 878             (!params->key_len && !params->obj_hashfn))
 879                 return -EINVAL;
 880
 881         if (params->nulls_base && params->nulls_base < (1U << RHT_BASE_SHIFT))
 882                 return -EINVAL;
 883
 884         params->min_shift = max_t(size_t, params->min_shift,
 885                                   ilog2(HASH_MIN_SIZE));
 886
 887         if (params->nelem_hint)
 888                 size = rounded_hashtable_size(params);
 889
 890         memset(ht, 0, sizeof(*ht));
 891         mutex_init(&ht->mutex);
 892         memcpy(&ht->p, params, sizeof(*params));
 893
 894         if (params->locks_mul)
 895                 ht->p.locks_mul = roundup_pow_of_two(params->locks_mul);
 896         else
 897                 ht->p.locks_mul = BUCKET_LOCKS_PER_CPU;
 898
 899         tbl = bucket_table_alloc(ht, size);
 900         if (tbl == NULL)
 901                 return -ENOMEM;
 902
 903         atomic_set(&ht->nelems, 0);
 904         atomic_set(&ht->shift, ilog2(tbl->size));
 905         RCU_INIT_POINTER(ht->tbl, tbl);
 906         RCU_INIT_POINTER(ht->future_tbl, tbl);
 907
 908         if (!ht->p.hash_rnd)
 909                 get_random_bytes(&ht->p.hash_rnd, sizeof(ht->p.hash_rnd));
 910
 911         if (ht->p.grow_decision || ht->p.shrink_decision)
 912                 INIT_WORK(&ht->run_work, rht_deferred_worker);
 913
 914         return 0;
 915 }
 916 EXPORT_SYMBOL_GPL(rhashtable_init);
 917
 918 /**
 919  * rhashtable_destroy - destroy hash table
 920  * @ht:         the hash table to destroy
 921  *
 922  * Frees the bucket array. This function is not rcu safe, therefore the caller
 923  * has to make sure that no resizing may happen by unpublishing the hashtable
 924  * and waiting for the quiescent cycle before releasing the bucket array.
 925  */
 926 void rhashtable_destroy(struct rhashtable *ht)
 927 {
 928         ht->being_destroyed = true;
 929
 930         if (ht->p.grow_decision || ht->p.shrink_decision)
 931                 cancel_work_sync(&ht->run_work);
 932
 933         mutex_lock(&ht->mutex);
 934         bucket_table_free(rht_dereference(ht->tbl, ht));
 935         mutex_unlock(&ht->mutex);
 936 }
 937 EXPORT_SYMBOL_GPL(rhashtable_destroy);