lib/rhashtable.c

   1 /*
   2  * Resizable, Scalable, Concurrent Hash Table
   3  *
   4  * Copyright (c) 2015 Herbert Xu <herbert@gondor.apana.org.au>
   5  * Copyright (c) 2014-2015 Thomas Graf <tgraf@suug.ch>
   6  * Copyright (c) 2008-2014 Patrick McHardy <kaber@trash.net>
   7  *
   8  * Code partially derived from nft_hash
   9  * Rewritten with rehash code from br_multicast plus single list
  10  * pointer as suggested by Josh Triplett
  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/atomic.h>
  18 #include <linux/kernel.h>
  19 #include <linux/init.h>
  20 #include <linux/log2.h>
  21 #include <linux/sched.h>
  22 #include <linux/slab.h>
  23 #include <linux/vmalloc.h>
  24 #include <linux/mm.h>
  25 #include <linux/jhash.h>
  26 #include <linux/random.h>
  27 #include <linux/rhashtable.h>
  28 #include <linux/err.h>
  29 #include <linux/export.h>
  30
  31 #define HASH_DEFAULT_SIZE       64UL
  32 #define HASH_MIN_SIZE           4U
  33 #define BUCKET_LOCKS_PER_CPU    32UL
  34
  35 static u32 head_hashfn(struct rhashtable *ht,
  36                        const struct bucket_table *tbl,
  37                        const struct rhash_head *he)
  38 {
  39         return rht_head_hashfn(ht, tbl, he, ht->p);
  40 }
  41
  42 #ifdef CONFIG_PROVE_LOCKING
  43 #define ASSERT_RHT_MUTEX(HT) BUG_ON(!lockdep_rht_mutex_is_held(HT))
  44
  45 int lockdep_rht_mutex_is_held(struct rhashtable *ht)
  46 {
  47         return (debug_locks) ? lockdep_is_held(&ht->mutex) : 1;
  48 }
  49 EXPORT_SYMBOL_GPL(lockdep_rht_mutex_is_held);
  50
  51 int lockdep_rht_bucket_is_held(const struct bucket_table *tbl, u32 hash)
  52 {
  53         spinlock_t *lock = rht_bucket_lock(tbl, hash);
  54
  55         return (debug_locks) ? lockdep_is_held(lock) : 1;
  56 }
  57 EXPORT_SYMBOL_GPL(lockdep_rht_bucket_is_held);
  58 #else
  59 #define ASSERT_RHT_MUTEX(HT)
  60 #endif
  61
  62
  63 static int alloc_bucket_locks(struct rhashtable *ht, struct bucket_table *tbl,
  64                               gfp_t gfp)
  65 {
  66         unsigned int i, size;
  67 #if defined(CONFIG_PROVE_LOCKING)
  68         unsigned int nr_pcpus = 2;
  69 #else
  70         unsigned int nr_pcpus = num_possible_cpus();
  71 #endif
  72
  73         nr_pcpus = min_t(unsigned int, nr_pcpus, 64UL);
  74         size = roundup_pow_of_two(nr_pcpus * ht->p.locks_mul);
  75
  76         /* Never allocate more than 0.5 locks per bucket */
  77         size = min_t(unsigned int, size, tbl->size >> 1);
  78
  79         if (sizeof(spinlock_t) != 0) {
  80                 tbl->locks = NULL;
  81 #ifdef CONFIG_NUMA
  82                 if (size * sizeof(spinlock_t) > PAGE_SIZE &&
  83                     gfp == GFP_KERNEL)
  84                         tbl->locks = vmalloc(size * sizeof(spinlock_t));
  85 #endif
  86                 if (gfp != GFP_KERNEL)
  87                         gfp |= __GFP_NOWARN | __GFP_NORETRY;
  88
  89                 if (!tbl->locks)
  90                         tbl->locks = kmalloc_array(size, sizeof(spinlock_t),
  91                                                    gfp);
  92                 if (!tbl->locks)
  93                         return -ENOMEM;
  94                 for (i = 0; i < size; i++)
  95                         spin_lock_init(&tbl->locks[i]);
  96         }
  97         tbl->locks_mask = size - 1;
  98
  99         return 0;
 100 }
 101
 102 static void bucket_table_free(const struct bucket_table *tbl)
 103 {
 104         if (tbl)
 105                 kvfree(tbl->locks);
 106
 107         kvfree(tbl);
 108 }
 109
 110 static void bucket_table_free_rcu(struct rcu_head *head)
 111 {
 112         bucket_table_free(container_of(head, struct bucket_table, rcu));
 113 }
 114
 115 static struct bucket_table *bucket_table_alloc(struct rhashtable *ht,
 116                                                size_t nbuckets,
 117                                                gfp_t gfp)
 118 {
 119         struct bucket_table *tbl = NULL;
 120         size_t size;
 121         int i;
 122
 123         size = sizeof(*tbl) + nbuckets * sizeof(tbl->buckets[0]);
 124         if (size <= (PAGE_SIZE << PAGE_ALLOC_COSTLY_ORDER) ||
 125             gfp != GFP_KERNEL)
 126                 tbl = kzalloc(size, gfp | __GFP_NOWARN | __GFP_NORETRY);
 127         if (tbl == NULL && gfp == GFP_KERNEL)
 128                 tbl = vzalloc(size);
 129         if (tbl == NULL)
 130                 return NULL;
 131
 132         tbl->size = nbuckets;
 133
 134         if (alloc_bucket_locks(ht, tbl, gfp) < 0) {
 135                 bucket_table_free(tbl);
 136                 return NULL;
 137         }
 138
 139         INIT_LIST_HEAD(&tbl->walkers);
 140
 141         get_random_bytes(&tbl->hash_rnd, sizeof(tbl->hash_rnd));
 142
 143         for (i = 0; i < nbuckets; i++)
 144                 INIT_RHT_NULLS_HEAD(tbl->buckets[i], ht, i);
 145
 146         return tbl;
 147 }
 148
 149 static struct bucket_table *rhashtable_last_table(struct rhashtable *ht,
 150                                                   struct bucket_table *tbl)
 151 {
 152         struct bucket_table *new_tbl;
 153
 154         do {
 155                 new_tbl = tbl;
 156                 tbl = rht_dereference_rcu(tbl->future_tbl, ht);
 157         } while (tbl);
 158
 159         return new_tbl;
 160 }
 161
 162 static int rhashtable_rehash_one(struct rhashtable *ht, unsigned int old_hash)
 163 {
 164         struct bucket_table *old_tbl = rht_dereference(ht->tbl, ht);
 165         struct bucket_table *new_tbl = rhashtable_last_table(ht,
 166                 rht_dereference_rcu(old_tbl->future_tbl, ht));
 167         struct rhash_head __rcu **pprev = &old_tbl->buckets[old_hash];
 168         int err = -ENOENT;
 169         struct rhash_head *head, *next, *entry;
 170         spinlock_t *new_bucket_lock;
 171         unsigned int new_hash;
 172
 173         rht_for_each(entry, old_tbl, old_hash) {
 174                 err = 0;
 175                 next = rht_dereference_bucket(entry->next, old_tbl, old_hash);
 176
 177                 if (rht_is_a_nulls(next))
 178                         break;
 179
 180                 pprev = &entry->next;
 181         }
 182
 183         if (err)
 184                 goto out;
 185
 186         new_hash = head_hashfn(ht, new_tbl, entry);
 187
 188         new_bucket_lock = rht_bucket_lock(new_tbl, new_hash);
 189
 190         spin_lock_nested(new_bucket_lock, SINGLE_DEPTH_NESTING);
 191         head = rht_dereference_bucket(new_tbl->buckets[new_hash],
 192                                       new_tbl, new_hash);
 193
 194         RCU_INIT_POINTER(entry->next, head);
 195
 196         rcu_assign_pointer(new_tbl->buckets[new_hash], entry);
 197         spin_unlock(new_bucket_lock);
 198
 199         rcu_assign_pointer(*pprev, next);
 200
 201 out:
 202         return err;
 203 }
 204
 205 static void rhashtable_rehash_chain(struct rhashtable *ht,
 206                                     unsigned int old_hash)
 207 {
 208         struct bucket_table *old_tbl = rht_dereference(ht->tbl, ht);
 209         spinlock_t *old_bucket_lock;
 210
 211         old_bucket_lock = rht_bucket_lock(old_tbl, old_hash);
 212
 213         spin_lock_bh(old_bucket_lock);
 214         while (!rhashtable_rehash_one(ht, old_hash))
 215                 ;
 216         old_tbl->rehash++;
 217         spin_unlock_bh(old_bucket_lock);
 218 }
 219
 220 static int rhashtable_rehash_attach(struct rhashtable *ht,
 221                                     struct bucket_table *old_tbl,
 222                                     struct bucket_table *new_tbl)
 223 {
 224         /* Protect future_tbl using the first bucket lock. */
 225         spin_lock_bh(old_tbl->locks);
 226
 227         /* Did somebody beat us to it? */
 228         if (rcu_access_pointer(old_tbl->future_tbl)) {
 229                 spin_unlock_bh(old_tbl->locks);
 230                 return -EEXIST;
 231         }
 232
 233         /* Make insertions go into the new, empty table right away. Deletions
 234          * and lookups will be attempted in both tables until we synchronize.
 235          */
 236         rcu_assign_pointer(old_tbl->future_tbl, new_tbl);
 237
 238         spin_unlock_bh(old_tbl->locks);
 239
 240         return 0;
 241 }
 242
 243 static int rhashtable_rehash_table(struct rhashtable *ht)
 244 {
 245         struct bucket_table *old_tbl = rht_dereference(ht->tbl, ht);
 246         struct bucket_table *new_tbl;
 247         struct rhashtable_walker *walker;
 248         unsigned int old_hash;
 249
 250         new_tbl = rht_dereference(old_tbl->future_tbl, ht);
 251         if (!new_tbl)
 252                 return 0;
 253
 254         for (old_hash = 0; old_hash < old_tbl->size; old_hash++)
 255                 rhashtable_rehash_chain(ht, old_hash);
 256
 257         /* Publish the new table pointer. */
 258         rcu_assign_pointer(ht->tbl, new_tbl);
 259
 260         spin_lock(&ht->lock);
 261         list_for_each_entry(walker, &old_tbl->walkers, list)
 262                 walker->tbl = NULL;
 263         spin_unlock(&ht->lock);
 264
 265         /* Wait for readers. All new readers will see the new
 266          * table, and thus no references to the old table will
 267          * remain.
 268          */
 269         call_rcu(&old_tbl->rcu, bucket_table_free_rcu);
 270
 271         return rht_dereference(new_tbl->future_tbl, ht) ? -EAGAIN : 0;
 272 }
 273
 274 /**
 275  * rhashtable_expand - Expand hash table while allowing concurrent lookups
 276  * @ht:         the hash table to expand
 277  *
 278  * A secondary bucket array is allocated and the hash entries are migrated.
 279  *
 280  * This function may only be called in a context where it is safe to call
 281  * synchronize_rcu(), e.g. not within a rcu_read_lock() section.
 282  *
 283  * The caller must ensure that no concurrent resizing occurs by holding
 284  * ht->mutex.
 285  *
 286  * It is valid to have concurrent insertions and deletions protected by per
 287  * bucket locks or concurrent RCU protected lookups and traversals.
 288  */
 289 static int rhashtable_expand(struct rhashtable *ht)
 290 {
 291         struct bucket_table *new_tbl, *old_tbl = rht_dereference(ht->tbl, ht);
 292         int err;
 293
 294         ASSERT_RHT_MUTEX(ht);
 295
 296         old_tbl = rhashtable_last_table(ht, old_tbl);
 297
 298         new_tbl = bucket_table_alloc(ht, old_tbl->size * 2, GFP_KERNEL);
 299         if (new_tbl == NULL)
 300                 return -ENOMEM;
 301
 302         err = rhashtable_rehash_attach(ht, old_tbl, new_tbl);
 303         if (err)
 304                 bucket_table_free(new_tbl);
 305
 306         return err;
 307 }
 308
 309 /**
 310  * rhashtable_shrink - Shrink hash table while allowing concurrent lookups
 311  * @ht:         the hash table to shrink
 312  *
 313  * This function shrinks the hash table to fit, i.e., the smallest
 314  * size would not cause it to expand right away automatically.
 315  *
 316  * The caller must ensure that no concurrent resizing occurs by holding
 317  * ht->mutex.
 318  *
 319  * The caller must ensure that no concurrent table mutations take place.
 320  * It is however valid to have concurrent lookups if they are RCU protected.
 321  *
 322  * It is valid to have concurrent insertions and deletions protected by per
 323  * bucket locks or concurrent RCU protected lookups and traversals.
 324  */
 325 static int rhashtable_shrink(struct rhashtable *ht)
 326 {
 327         struct bucket_table *new_tbl, *old_tbl = rht_dereference(ht->tbl, ht);
 328         unsigned int nelems = atomic_read(&ht->nelems);
 329         unsigned int size = 0;
 330         int err;
 331
 332         ASSERT_RHT_MUTEX(ht);
 333
 334         if (nelems)
 335                 size = roundup_pow_of_two(nelems * 3 / 2);
 336         if (size < ht->p.min_size)
 337                 size = ht->p.min_size;
 338
 339         if (old_tbl->size <= size)
 340                 return 0;
 341
 342         if (rht_dereference(old_tbl->future_tbl, ht))
 343                 return -EEXIST;
 344
 345         new_tbl = bucket_table_alloc(ht, size, GFP_KERNEL);
 346         if (new_tbl == NULL)
 347                 return -ENOMEM;
 348
 349         err = rhashtable_rehash_attach(ht, old_tbl, new_tbl);
 350         if (err)
 351                 bucket_table_free(new_tbl);
 352
 353         return err;
 354 }
 355
 356 static void rht_deferred_worker(struct work_struct *work)
 357 {
 358         struct rhashtable *ht;
 359         struct bucket_table *tbl;
 360         int err = 0;
 361
 362         ht = container_of(work, struct rhashtable, run_work);
 363         mutex_lock(&ht->mutex);
 364
 365         tbl = rht_dereference(ht->tbl, ht);
 366         tbl = rhashtable_last_table(ht, tbl);
 367
 368         if (rht_grow_above_75(ht, tbl))
 369                 rhashtable_expand(ht);
 370         else if (ht->p.automatic_shrinking && rht_shrink_below_30(ht, tbl))
 371                 rhashtable_shrink(ht);
 372
 373         err = rhashtable_rehash_table(ht);
 374
 375         mutex_unlock(&ht->mutex);
 376
 377         if (err)
 378                 schedule_work(&ht->run_work);
 379 }
 380
 381 static int rhashtable_insert_rehash(struct rhashtable *ht,
 382                                     struct bucket_table *tbl)
 383 {
 384         struct bucket_table *old_tbl;
 385         struct bucket_table *new_tbl;
 386         unsigned int size;
 387         int err;
 388
 389         old_tbl = rht_dereference_rcu(ht->tbl, ht);
 390
 391         size = tbl->size;
 392
 393         err = -EBUSY;
 394
 395         if (rht_grow_above_75(ht, tbl))
 396                 size *= 2;
 397         /* Do not schedule more than one rehash */
 398         else if (old_tbl != tbl)
 399                 goto fail;
 400
 401         err = -ENOMEM;
 402
 403         new_tbl = bucket_table_alloc(ht, size, GFP_ATOMIC);
 404         if (new_tbl == NULL)
 405                 goto fail;
 406
 407         err = rhashtable_rehash_attach(ht, tbl, new_tbl);
 408         if (err) {
 409                 bucket_table_free(new_tbl);
 410                 if (err == -EEXIST)
 411                         err = 0;
 412         } else
 413                 schedule_work(&ht->run_work);
 414
 415         return err;
 416
 417 fail:
 418         /* Do not fail the insert if someone else did a rehash. */
 419         if (likely(rcu_dereference_raw(tbl->future_tbl)))
 420                 return 0;
 421
 422         /* Schedule async rehash to retry allocation in process context. */
 423         if (err == -ENOMEM)
 424                 schedule_work(&ht->run_work);
 425
 426         return err;
 427 }
 428
 429 static void *rhashtable_lookup_one(struct rhashtable *ht,
 430                                    struct bucket_table *tbl, unsigned int hash,
 431                                    const void *key, struct rhash_head *obj)
 432 {
 433         struct rhashtable_compare_arg arg = {
 434                 .ht = ht,
 435                 .key = key,
 436         };
 437         struct rhash_head __rcu **pprev;
 438         struct rhash_head *head;
 439         int elasticity;
 440
 441         elasticity = ht->elasticity;
 442         pprev = &tbl->buckets[hash];
 443         rht_for_each(head, tbl, hash) {
 444                 struct rhlist_head *list;
 445                 struct rhlist_head *plist;
 446
 447                 elasticity--;
 448                 if (!key ||
 449                     (ht->p.obj_cmpfn ?
 450                      ht->p.obj_cmpfn(&arg, rht_obj(ht, head)) :
 451                      rhashtable_compare(&arg, rht_obj(ht, head))))
 452                         continue;
 453
 454                 if (!ht->rhlist)
 455                         return rht_obj(ht, head);
 456
 457                 list = container_of(obj, struct rhlist_head, rhead);
 458                 plist = container_of(head, struct rhlist_head, rhead);
 459
 460                 RCU_INIT_POINTER(list->next, plist);
 461                 head = rht_dereference_bucket(head->next, tbl, hash);
 462                 RCU_INIT_POINTER(list->rhead.next, head);
 463                 rcu_assign_pointer(*pprev, obj);
 464
 465                 return NULL;
 466         }
 467
 468         if (elasticity <= 0)
 469                 return ERR_PTR(-EAGAIN);
 470
 471         return ERR_PTR(-ENOENT);
 472 }
 473
 474 static struct bucket_table *rhashtable_insert_one(struct rhashtable *ht,
 475                                                   struct bucket_table *tbl,
 476                                                   unsigned int hash,
 477                                                   struct rhash_head *obj,
 478                                                   void *data)
 479 {
 480         struct bucket_table *new_tbl;
 481         struct rhash_head *head;
 482
 483         if (!IS_ERR_OR_NULL(data))
 484                 return ERR_PTR(-EEXIST);
 485
 486         if (PTR_ERR(data) != -EAGAIN && PTR_ERR(data) != -ENOENT)
 487                 return ERR_CAST(data);
 488
 489         new_tbl = rcu_dereference(tbl->future_tbl);
 490         if (new_tbl)
 491                 return new_tbl;
 492
 493         if (PTR_ERR(data) != -ENOENT)
 494                 return ERR_CAST(data);
 495
 496         if (unlikely(rht_grow_above_max(ht, tbl)))
 497                 return ERR_PTR(-E2BIG);
 498
 499         if (unlikely(rht_grow_above_100(ht, tbl)))
 500                 return ERR_PTR(-EAGAIN);
 501
 502         head = rht_dereference_bucket(tbl->buckets[hash], tbl, hash);
 503
 504         RCU_INIT_POINTER(obj->next, head);
 505         if (ht->rhlist) {
 506                 struct rhlist_head *list;
 507
 508                 list = container_of(obj, struct rhlist_head, rhead);
 509                 RCU_INIT_POINTER(list->next, NULL);
 510         }
 511
 512         rcu_assign_pointer(tbl->buckets[hash], obj);
 513
 514         atomic_inc(&ht->nelems);
 515         if (rht_grow_above_75(ht, tbl))
 516                 schedule_work(&ht->run_work);
 517
 518         return NULL;
 519 }
 520
 521 static void *rhashtable_try_insert(struct rhashtable *ht, const void *key,
 522                                    struct rhash_head *obj)
 523 {
 524         struct bucket_table *new_tbl;
 525         struct bucket_table *tbl;
 526         unsigned int hash;
 527         spinlock_t *lock;
 528         void *data;
 529
 530         tbl = rcu_dereference(ht->tbl);
 531
 532         /* All insertions must grab the oldest table containing
 533          * the hashed bucket that is yet to be rehashed.
 534          */
 535         for (;;) {
 536                 hash = rht_head_hashfn(ht, tbl, obj, ht->p);
 537                 lock = rht_bucket_lock(tbl, hash);
 538                 spin_lock_bh(lock);
 539
 540                 if (tbl->rehash <= hash)
 541                         break;
 542
 543                 spin_unlock_bh(lock);
 544                 tbl = rcu_dereference(tbl->future_tbl);
 545         }
 546
 547         data = rhashtable_lookup_one(ht, tbl, hash, key, obj);
 548         new_tbl = rhashtable_insert_one(ht, tbl, hash, obj, data);
 549         if (PTR_ERR(new_tbl) != -EEXIST)
 550                 data = ERR_CAST(new_tbl);
 551
 552         while (!IS_ERR_OR_NULL(new_tbl)) {
 553                 tbl = new_tbl;
 554                 hash = rht_head_hashfn(ht, tbl, obj, ht->p);
 555                 spin_lock_nested(rht_bucket_lock(tbl, hash),
 556                                  SINGLE_DEPTH_NESTING);
 557
 558                 data = rhashtable_lookup_one(ht, tbl, hash, key, obj);
 559                 new_tbl = rhashtable_insert_one(ht, tbl, hash, obj, data);
 560                 if (PTR_ERR(new_tbl) != -EEXIST)
 561                         data = ERR_CAST(new_tbl);
 562
 563                 spin_unlock(rht_bucket_lock(tbl, hash));
 564         }
 565
 566         spin_unlock_bh(lock);
 567
 568         if (PTR_ERR(data) == -EAGAIN)
 569                 data = ERR_PTR(rhashtable_insert_rehash(ht, tbl) ?:
 570                                -EAGAIN);
 571
 572         return data;
 573 }
 574
 575 void *rhashtable_insert_slow(struct rhashtable *ht, const void *key,
 576                              struct rhash_head *obj)
 577 {
 578         void *data;
 579
 580         do {
 581                 rcu_read_lock();
 582                 data = rhashtable_try_insert(ht, key, obj);
 583                 rcu_read_unlock();
 584         } while (PTR_ERR(data) == -EAGAIN);
 585
 586         return data;
 587 }
 588 EXPORT_SYMBOL_GPL(rhashtable_insert_slow);
 589
 590 /**
 591  * rhashtable_walk_enter - Initialise an iterator
 592  * @ht:         Table to walk over
 593  * @iter:       Hash table Iterator
 594  *
 595  * This function prepares a hash table walk.
 596  *
 597  * Note that if you restart a walk after rhashtable_walk_stop you
 598  * may see the same object twice.  Also, you may miss objects if
 599  * there are removals in between rhashtable_walk_stop and the next
 600  * call to rhashtable_walk_start.
 601  *
 602  * For a completely stable walk you should construct your own data
 603  * structure outside the hash table.
 604  *
 605  * This function may sleep so you must not call it from interrupt
 606  * context or with spin locks held.
 607  *
 608  * You must call rhashtable_walk_exit after this function returns.
 609  */
 610 void rhashtable_walk_enter(struct rhashtable *ht, struct rhashtable_iter *iter)
 611 {
 612         iter->ht = ht;
 613         iter->p = NULL;
 614         iter->slot = 0;
 615         iter->skip = 0;
 616
 617         spin_lock(&ht->lock);
 618         iter->walker.tbl =
 619                 rcu_dereference_protected(ht->tbl, lockdep_is_held(&ht->lock));
 620         list_add(&iter->walker.list, &iter->walker.tbl->walkers);
 621         spin_unlock(&ht->lock);
 622 }
 623 EXPORT_SYMBOL_GPL(rhashtable_walk_enter);
 624
 625 /**
 626  * rhashtable_walk_exit - Free an iterator
 627  * @iter:       Hash table Iterator
 628  *
 629  * This function frees resources allocated by rhashtable_walk_init.
 630  */
 631 void rhashtable_walk_exit(struct rhashtable_iter *iter)
 632 {
 633         spin_lock(&iter->ht->lock);
 634         if (iter->walker.tbl)
 635                 list_del(&iter->walker.list);
 636         spin_unlock(&iter->ht->lock);
 637 }
 638 EXPORT_SYMBOL_GPL(rhashtable_walk_exit);
 639
 640 /**
 641  * rhashtable_walk_start - Start a hash table walk
 642  * @iter:       Hash table iterator
 643  *
 644  * Start a hash table walk.  Note that we take the RCU lock in all
 645  * cases including when we return an error.  So you must always call
 646  * rhashtable_walk_stop to clean up.
 647  *
 648  * Returns zero if successful.
 649  *
 650  * Returns -EAGAIN if resize event occured.  Note that the iterator
 651  * will rewind back to the beginning and you may use it immediately
 652  * by calling rhashtable_walk_next.
 653  */
 654 int rhashtable_walk_start(struct rhashtable_iter *iter)
 655         __acquires(RCU)
 656 {
 657         struct rhashtable *ht = iter->ht;
 658
 659         rcu_read_lock();
 660
 661         spin_lock(&ht->lock);
 662         if (iter->walker.tbl)
 663                 list_del(&iter->walker.list);
 664         spin_unlock(&ht->lock);
 665
 666         if (!iter->walker.tbl) {
 667                 iter->walker.tbl = rht_dereference_rcu(ht->tbl, ht);
 668                 return -EAGAIN;
 669         }
 670
 671         return 0;
 672 }
 673 EXPORT_SYMBOL_GPL(rhashtable_walk_start);
 674
 675 /**
 676  * rhashtable_walk_next - Return the next object and advance the iterator
 677  * @iter:       Hash table iterator
 678  *
 679  * Note that you must call rhashtable_walk_stop when you are finished
 680  * with the walk.
 681  *
 682  * Returns the next object or NULL when the end of the table is reached.
 683  *
 684  * Returns -EAGAIN if resize event occured.  Note that the iterator
 685  * will rewind back to the beginning and you may continue to use it.
 686  */
 687 void *rhashtable_walk_next(struct rhashtable_iter *iter)
 688 {
 689         struct bucket_table *tbl = iter->walker.tbl;
 690         struct rhlist_head *list = iter->list;
 691         struct rhashtable *ht = iter->ht;
 692         struct rhash_head *p = iter->p;
 693         bool rhlist = ht->rhlist;
 694
 695         if (p) {
 696                 if (!rhlist || !(list = rcu_dereference(list->next))) {
 697                         p = rcu_dereference(p->next);
 698                         list = container_of(p, struct rhlist_head, rhead);
 699                 }
 700                 goto next;
 701         }
 702
 703         for (; iter->slot < tbl->size; iter->slot++) {
 704                 int skip = iter->skip;
 705
 706                 rht_for_each_rcu(p, tbl, iter->slot) {
 707                         if (rhlist) {
 708                                 list = container_of(p, struct rhlist_head,
 709                                                     rhead);
 710                                 do {
 711                                         if (!skip)
 712                                                 goto next;
 713                                         skip--;
 714                                         list = rcu_dereference(list->next);
 715                                 } while (list);
 716
 717                                 continue;
 718                         }
 719                         if (!skip)
 720                                 break;
 721                         skip--;
 722                 }
 723
 724 next:
 725                 if (!rht_is_a_nulls(p)) {
 726                         iter->skip++;
 727                         iter->p = p;
 728                         iter->list = list;
 729                         return rht_obj(ht, rhlist ? &list->rhead : p);
 730                 }
 731
 732                 iter->skip = 0;
 733         }
 734
 735         iter->p = NULL;
 736
 737         /* Ensure we see any new tables. */
 738         smp_rmb();
 739
 740         iter->walker.tbl = rht_dereference_rcu(tbl->future_tbl, ht);
 741         if (iter->walker.tbl) {
 742                 iter->slot = 0;
 743                 iter->skip = 0;
 744                 return ERR_PTR(-EAGAIN);
 745         }
 746
 747         return NULL;
 748 }
 749 EXPORT_SYMBOL_GPL(rhashtable_walk_next);
 750
 751 /**
 752  * rhashtable_walk_stop - Finish a hash table walk
 753  * @iter:       Hash table iterator
 754  *
 755  * Finish a hash table walk.
 756  */
 757 void rhashtable_walk_stop(struct rhashtable_iter *iter)
 758         __releases(RCU)
 759 {
 760         struct rhashtable *ht;
 761         struct bucket_table *tbl = iter->walker.tbl;
 762
 763         if (!tbl)
 764                 goto out;
 765
 766         ht = iter->ht;
 767
 768         spin_lock(&ht->lock);
 769         if (tbl->rehash < tbl->size)
 770                 list_add(&iter->walker.list, &tbl->walkers);
 771         else
 772                 iter->walker.tbl = NULL;
 773         spin_unlock(&ht->lock);
 774
 775         iter->p = NULL;
 776
 777 out:
 778         rcu_read_unlock();
 779 }
 780 EXPORT_SYMBOL_GPL(rhashtable_walk_stop);
 781
 782 static size_t rounded_hashtable_size(const struct rhashtable_params *params)
 783 {
 784         return max(roundup_pow_of_two(params->nelem_hint * 4 / 3),
 785                    (unsigned long)params->min_size);
 786 }
 787
 788 static u32 rhashtable_jhash2(const void *key, u32 length, u32 seed)
 789 {
 790         return jhash2(key, length, seed);
 791 }
 792
 793 /**
 794  * rhashtable_init - initialize a new hash table
 795  * @ht:         hash table to be initialized
 796  * @params:     configuration parameters
 797  *
 798  * Initializes a new hash table based on the provided configuration
 799  * parameters. A table can be configured either with a variable or
 800  * fixed length key:
 801  *
 802  * Configuration Example 1: Fixed length keys
 803  * struct test_obj {
 804  *      int                     key;
 805  *      void *                  my_member;
 806  *      struct rhash_head       node;
 807  * };
 808  *
 809  * struct rhashtable_params params = {
 810  *      .head_offset = offsetof(struct test_obj, node),
 811  *      .key_offset = offsetof(struct test_obj, key),
 812  *      .key_len = sizeof(int),
 813  *      .hashfn = jhash,
 814  *      .nulls_base = (1U << RHT_BASE_SHIFT),
 815  * };
 816  *
 817  * Configuration Example 2: Variable length keys
 818  * struct test_obj {
 819  *      [...]
 820  *      struct rhash_head       node;
 821  * };
 822  *
 823  * u32 my_hash_fn(const void *data, u32 len, u32 seed)
 824  * {
 825  *      struct test_obj *obj = data;
 826  *
 827  *      return [... hash ...];
 828  * }
 829  *
 830  * struct rhashtable_params params = {
 831  *      .head_offset = offsetof(struct test_obj, node),
 832  *      .hashfn = jhash,
 833  *      .obj_hashfn = my_hash_fn,
 834  * };
 835  */
 836 int rhashtable_init(struct rhashtable *ht,
 837                     const struct rhashtable_params *params)
 838 {
 839         struct bucket_table *tbl;
 840         size_t size;
 841
 842         size = HASH_DEFAULT_SIZE;
 843
 844         if ((!params->key_len && !params->obj_hashfn) ||
 845             (params->obj_hashfn && !params->obj_cmpfn))
 846                 return -EINVAL;
 847
 848         if (params->nulls_base && params->nulls_base < (1U << RHT_BASE_SHIFT))
 849                 return -EINVAL;
 850
 851         memset(ht, 0, sizeof(*ht));
 852         mutex_init(&ht->mutex);
 853         spin_lock_init(&ht->lock);
 854         memcpy(&ht->p, params, sizeof(*params));
 855
 856         if (params->min_size)
 857                 ht->p.min_size = roundup_pow_of_two(params->min_size);
 858
 859         if (params->max_size)
 860                 ht->p.max_size = rounddown_pow_of_two(params->max_size);
 861
 862         if (params->insecure_max_entries)
 863                 ht->p.insecure_max_entries =
 864                         rounddown_pow_of_two(params->insecure_max_entries);
 865         else
 866                 ht->p.insecure_max_entries = ht->p.max_size * 2;
 867
 868         ht->p.min_size = max(ht->p.min_size, HASH_MIN_SIZE);
 869
 870         if (params->nelem_hint)
 871                 size = rounded_hashtable_size(&ht->p);
 872
 873         /* The maximum (not average) chain length grows with the
 874          * size of the hash table, at a rate of (log N)/(log log N).
 875          * The value of 16 is selected so that even if the hash
 876          * table grew to 2^32 you would not expect the maximum
 877          * chain length to exceed it unless we are under attack
 878          * (or extremely unlucky).
 879          *
 880          * As this limit is only to detect attacks, we don't need
 881          * to set it to a lower value as you'd need the chain
 882          * length to vastly exceed 16 to have any real effect
 883          * on the system.
 884          */
 885         if (!params->insecure_elasticity)
 886                 ht->elasticity = 16;
 887
 888         if (params->locks_mul)
 889                 ht->p.locks_mul = roundup_pow_of_two(params->locks_mul);
 890         else
 891                 ht->p.locks_mul = BUCKET_LOCKS_PER_CPU;
 892
 893         ht->key_len = ht->p.key_len;
 894         if (!params->hashfn) {
 895                 ht->p.hashfn = jhash;
 896
 897                 if (!(ht->key_len & (sizeof(u32) - 1))) {
 898                         ht->key_len /= sizeof(u32);
 899                         ht->p.hashfn = rhashtable_jhash2;
 900                 }
 901         }
 902
 903         tbl = bucket_table_alloc(ht, size, GFP_KERNEL);
 904         if (tbl == NULL)
 905                 return -ENOMEM;
 906
 907         atomic_set(&ht->nelems, 0);
 908
 909         RCU_INIT_POINTER(ht->tbl, tbl);
 910
 911         INIT_WORK(&ht->run_work, rht_deferred_worker);
 912
 913         return 0;
 914 }
 915 EXPORT_SYMBOL_GPL(rhashtable_init);
 916
 917 /**
 918  * rhltable_init - initialize a new hash list table
 919  * @hlt:        hash list table to be initialized
 920  * @params:     configuration parameters
 921  *
 922  * Initializes a new hash list table.
 923  *
 924  * See documentation for rhashtable_init.
 925  */
 926 int rhltable_init(struct rhltable *hlt, const struct rhashtable_params *params)
 927 {
 928         int err;
 929
 930         /* No rhlist NULLs marking for now. */
 931         if (params->nulls_base)
 932                 return -EINVAL;
 933
 934         err = rhashtable_init(&hlt->ht, params);
 935         hlt->ht.rhlist = true;
 936         return err;
 937 }
 938 EXPORT_SYMBOL_GPL(rhltable_init);
 939
 940 static void rhashtable_free_one(struct rhashtable *ht, struct rhash_head *obj,
 941                                 void (*free_fn)(void *ptr, void *arg),
 942                                 void *arg)
 943 {
 944         struct rhlist_head *list;
 945
 946         if (!ht->rhlist) {
 947                 free_fn(rht_obj(ht, obj), arg);
 948                 return;
 949         }
 950
 951         list = container_of(obj, struct rhlist_head, rhead);
 952         do {
 953                 obj = &list->rhead;
 954                 list = rht_dereference(list->next, ht);
 955                 free_fn(rht_obj(ht, obj), arg);
 956         } while (list);
 957 }
 958
 959 /**
 960  * rhashtable_free_and_destroy - free elements and destroy hash table
 961  * @ht:         the hash table to destroy
 962  * @free_fn:    callback to release resources of element
 963  * @arg:        pointer passed to free_fn
 964  *
 965  * Stops an eventual async resize. If defined, invokes free_fn for each
 966  * element to releasal resources. Please note that RCU protected
 967  * readers may still be accessing the elements. Releasing of resources
 968  * must occur in a compatible manner. Then frees the bucket array.
 969  *
 970  * This function will eventually sleep to wait for an async resize
 971  * to complete. The caller is responsible that no further write operations
 972  * occurs in parallel.
 973  */
 974 void rhashtable_free_and_destroy(struct rhashtable *ht,
 975                                  void (*free_fn)(void *ptr, void *arg),
 976                                  void *arg)
 977 {
 978         const struct bucket_table *tbl;
 979         unsigned int i;
 980
 981         cancel_work_sync(&ht->run_work);
 982
 983         mutex_lock(&ht->mutex);
 984         tbl = rht_dereference(ht->tbl, ht);
 985         if (free_fn) {
 986                 for (i = 0; i < tbl->size; i++) {
 987                         struct rhash_head *pos, *next;
 988
 989                         for (pos = rht_dereference(tbl->buckets[i], ht),
 990                              next = !rht_is_a_nulls(pos) ?
 991                                         rht_dereference(pos->next, ht) : NULL;
 992                              !rht_is_a_nulls(pos);
 993                              pos = next,
 994                              next = !rht_is_a_nulls(pos) ?
 995                                         rht_dereference(pos->next, ht) : NULL)
 996                                 rhashtable_free_one(ht, pos, free_fn, arg);
 997                 }
 998         }
 999
1000         bucket_table_free(tbl);
1001         mutex_unlock(&ht->mutex);
1002 }
1003 EXPORT_SYMBOL_GPL(rhashtable_free_and_destroy);
1004
1005 void rhashtable_destroy(struct rhashtable *ht)
1006 {
1007         return rhashtable_free_and_destroy(ht, NULL, NULL);
1008 }
1009 EXPORT_SYMBOL_GPL(rhashtable_destroy);