1 // SPDX-License-Identifier: GPL-2.0-only 1
2 /*
3 * Resizable, Scalable, Concurrent Hash Table
4 *
5 * Copyright (c) 2015 Herbert Xu <herbert@gond
6 * Copyright (c) 2014-2015 Thomas Graf <tgraf@
7 * Copyright (c) 2008-2014 Patrick McHardy <ka
8 *
9 * Code partially derived from nft_hash
10 * Rewritten with rehash code from br_multicas
11 * pointer as suggested by Josh Triplett
12 */
13
14 #include <linux/atomic.h>
15 #include <linux/kernel.h>
16 #include <linux/init.h>
17 #include <linux/log2.h>
18 #include <linux/sched.h>
19 #include <linux/rculist.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 #include <linux/err.h>
27 #include <linux/export.h>
28
29 #define HASH_DEFAULT_SIZE 64UL
30 #define HASH_MIN_SIZE 4U
31
32 union nested_table {
33 union nested_table __rcu *table;
34 struct rhash_lock_head __rcu *bucket;
35 };
36
37 static u32 head_hashfn(struct rhashtable *ht,
38 const struct bucket_tab
39 const struct rhash_head
40 {
41 return rht_head_hashfn(ht, tbl, he, ht
42 }
43
44 #ifdef CONFIG_PROVE_LOCKING
45 #define ASSERT_RHT_MUTEX(HT) BUG_ON(!lockdep_r
46
47 int lockdep_rht_mutex_is_held(struct rhashtabl
48 {
49 return (debug_locks) ? lockdep_is_held
50 }
51 EXPORT_SYMBOL_GPL(lockdep_rht_mutex_is_held);
52
53 int lockdep_rht_bucket_is_held(const struct bu
54 {
55 if (!debug_locks)
56 return 1;
57 if (unlikely(tbl->nest))
58 return 1;
59 return bit_spin_is_locked(0, (unsigned
60 }
61 EXPORT_SYMBOL_GPL(lockdep_rht_bucket_is_held);
62 #else
63 #define ASSERT_RHT_MUTEX(HT)
64 #endif
65
66 static inline union nested_table *nested_table
67 const struct bucket_table *tbl)
68 {
69 /* The top-level bucket entry does not
70 * because it's set at the same time a
71 */
72 return (void *)rcu_dereference_protect
73 }
74
75 static void nested_table_free(union nested_tab
76 {
77 const unsigned int shift = PAGE_SHIFT
78 const unsigned int len = 1 << shift;
79 unsigned int i;
80
81 ntbl = rcu_dereference_protected(ntbl-
82 if (!ntbl)
83 return;
84
85 if (size > len) {
86 size >>= shift;
87 for (i = 0; i < len; i++)
88 nested_table_free(ntbl
89 }
90
91 kfree(ntbl);
92 }
93
94 static void nested_bucket_table_free(const str
95 {
96 unsigned int size = tbl->size >> tbl->
97 unsigned int len = 1 << tbl->nest;
98 union nested_table *ntbl;
99 unsigned int i;
100
101 ntbl = nested_table_top(tbl);
102
103 for (i = 0; i < len; i++)
104 nested_table_free(ntbl + i, si
105
106 kfree(ntbl);
107 }
108
109 static void bucket_table_free(const struct buc
110 {
111 if (tbl->nest)
112 nested_bucket_table_free(tbl);
113
114 kvfree(tbl);
115 }
116
117 static void bucket_table_free_rcu(struct rcu_h
118 {
119 bucket_table_free(container_of(head, s
120 }
121
122 static union nested_table *nested_table_alloc(
123
124
125 {
126 union nested_table *ntbl;
127 int i;
128
129 ntbl = rcu_dereference(*prev);
130 if (ntbl)
131 return ntbl;
132
133 ntbl = alloc_hooks_tag(ht->alloc_tag,
134 kmalloc_noprof(PAGE_SI
135
136 if (ntbl && leaf) {
137 for (i = 0; i < PAGE_SIZE / si
138 INIT_RHT_NULLS_HEAD(nt
139 }
140
141 if (cmpxchg((union nested_table **)pre
142 return ntbl;
143 /* Raced with another thread. */
144 kfree(ntbl);
145 return rcu_dereference(*prev);
146 }
147
148 static struct bucket_table *nested_bucket_tabl
149
150
151 {
152 const unsigned int shift = PAGE_SHIFT
153 struct bucket_table *tbl;
154 size_t size;
155
156 if (nbuckets < (1 << (shift + 1)))
157 return NULL;
158
159 size = sizeof(*tbl) + sizeof(tbl->buck
160
161 tbl = alloc_hooks_tag(ht->alloc_tag,
162 kmalloc_noprof(size, g
163 if (!tbl)
164 return NULL;
165
166 if (!nested_table_alloc(ht, (union nes
167 false)) {
168 kfree(tbl);
169 return NULL;
170 }
171
172 tbl->nest = (ilog2(nbuckets) - 1) % sh
173
174 return tbl;
175 }
176
177 static struct bucket_table *bucket_table_alloc
178
179
180 {
181 struct bucket_table *tbl = NULL;
182 size_t size;
183 int i;
184 static struct lock_class_key __key;
185
186 tbl = alloc_hooks_tag(ht->alloc_tag,
187 kvmalloc_node_noprof(s
188 g
189
190 size = nbuckets;
191
192 if (tbl == NULL && (gfp & ~__GFP_NOFAI
193 tbl = nested_bucket_table_allo
194 nbuckets = 0;
195 }
196
197 if (tbl == NULL)
198 return NULL;
199
200 lockdep_init_map(&tbl->dep_map, "rhash
201
202 tbl->size = size;
203
204 rcu_head_init(&tbl->rcu);
205 INIT_LIST_HEAD(&tbl->walkers);
206
207 tbl->hash_rnd = get_random_u32();
208
209 for (i = 0; i < nbuckets; i++)
210 INIT_RHT_NULLS_HEAD(tbl->bucke
211
212 return tbl;
213 }
214
215 static struct bucket_table *rhashtable_last_ta
216
217 {
218 struct bucket_table *new_tbl;
219
220 do {
221 new_tbl = tbl;
222 tbl = rht_dereference_rcu(tbl-
223 } while (tbl);
224
225 return new_tbl;
226 }
227
228 static int rhashtable_rehash_one(struct rhasht
229 struct rhash_
230 unsigned int
231 {
232 struct bucket_table *old_tbl = rht_der
233 struct bucket_table *new_tbl = rhashta
234 int err = -EAGAIN;
235 struct rhash_head *head, *next, *entry
236 struct rhash_head __rcu **pprev = NULL
237 unsigned int new_hash;
238 unsigned long flags;
239
240 if (new_tbl->nest)
241 goto out;
242
243 err = -ENOENT;
244
245 rht_for_each_from(entry, rht_ptr(bkt,
246 old_tbl, old_hash) {
247 err = 0;
248 next = rht_dereference_bucket(
249
250 if (rht_is_a_nulls(next))
251 break;
252
253 pprev = &entry->next;
254 }
255
256 if (err)
257 goto out;
258
259 new_hash = head_hashfn(ht, new_tbl, en
260
261 flags = rht_lock_nested(new_tbl, &new_
262 SINGLE_DEPTH_N
263
264 head = rht_ptr(new_tbl->buckets + new_
265
266 RCU_INIT_POINTER(entry->next, head);
267
268 rht_assign_unlock(new_tbl, &new_tbl->b
269
270 if (pprev)
271 rcu_assign_pointer(*pprev, nex
272 else
273 /* Need to preserved the bit l
274 rht_assign_locked(bkt, next);
275
276 out:
277 return err;
278 }
279
280 static int rhashtable_rehash_chain(struct rhas
281 unsigned i
282 {
283 struct bucket_table *old_tbl = rht_der
284 struct rhash_lock_head __rcu **bkt = r
285 unsigned long flags;
286 int err;
287
288 if (!bkt)
289 return 0;
290 flags = rht_lock(old_tbl, bkt);
291
292 while (!(err = rhashtable_rehash_one(h
293 ;
294
295 if (err == -ENOENT)
296 err = 0;
297 rht_unlock(old_tbl, bkt, flags);
298
299 return err;
300 }
301
302 static int rhashtable_rehash_attach(struct rha
303 struct buc
304 struct buc
305 {
306 /* Make insertions go into the new, em
307 * and lookups will be attempted in bo
308 * As cmpxchg() provides strong barrie
309 * rcu_assign_pointer().
310 */
311
312 if (cmpxchg((struct bucket_table **)&o
313 new_tbl) != NULL)
314 return -EEXIST;
315
316 return 0;
317 }
318
319 static int rhashtable_rehash_table(struct rhas
320 {
321 struct bucket_table *old_tbl = rht_der
322 struct bucket_table *new_tbl;
323 struct rhashtable_walker *walker;
324 unsigned int old_hash;
325 int err;
326
327 new_tbl = rht_dereference(old_tbl->fut
328 if (!new_tbl)
329 return 0;
330
331 for (old_hash = 0; old_hash < old_tbl-
332 err = rhashtable_rehash_chain(
333 if (err)
334 return err;
335 cond_resched();
336 }
337
338 /* Publish the new table pointer. */
339 rcu_assign_pointer(ht->tbl, new_tbl);
340
341 spin_lock(&ht->lock);
342 list_for_each_entry(walker, &old_tbl->
343 walker->tbl = NULL;
344
345 /* Wait for readers. All new readers w
346 * table, and thus no references to th
347 * remain.
348 * We do this inside the locked region
349 * rhashtable_walk_stop() can use rcu_
350 * to check if it should not re-link t
351 */
352 call_rcu(&old_tbl->rcu, bucket_table_f
353 spin_unlock(&ht->lock);
354
355 return rht_dereference(new_tbl->future
356 }
357
358 static int rhashtable_rehash_alloc(struct rhas
359 struct buck
360 unsigned in
361 {
362 struct bucket_table *new_tbl;
363 int err;
364
365 ASSERT_RHT_MUTEX(ht);
366
367 new_tbl = bucket_table_alloc(ht, size,
368 if (new_tbl == NULL)
369 return -ENOMEM;
370
371 err = rhashtable_rehash_attach(ht, old
372 if (err)
373 bucket_table_free(new_tbl);
374
375 return err;
376 }
377
378 /**
379 * rhashtable_shrink - Shrink hash table while
380 * @ht: the hash table to shrink
381 *
382 * This function shrinks the hash table to fit
383 * size would not cause it to expand right awa
384 *
385 * The caller must ensure that no concurrent r
386 * ht->mutex.
387 *
388 * The caller must ensure that no concurrent t
389 * It is however valid to have concurrent look
390 *
391 * It is valid to have concurrent insertions a
392 * bucket locks or concurrent RCU protected lo
393 */
394 static int rhashtable_shrink(struct rhashtable
395 {
396 struct bucket_table *old_tbl = rht_der
397 unsigned int nelems = atomic_read(&ht-
398 unsigned int size = 0;
399
400 if (nelems)
401 size = roundup_pow_of_two(nele
402 if (size < ht->p.min_size)
403 size = ht->p.min_size;
404
405 if (old_tbl->size <= size)
406 return 0;
407
408 if (rht_dereference(old_tbl->future_tb
409 return -EEXIST;
410
411 return rhashtable_rehash_alloc(ht, old
412 }
413
414 static void rht_deferred_worker(struct work_st
415 {
416 struct rhashtable *ht;
417 struct bucket_table *tbl;
418 int err = 0;
419
420 ht = container_of(work, struct rhashta
421 mutex_lock(&ht->mutex);
422
423 tbl = rht_dereference(ht->tbl, ht);
424 tbl = rhashtable_last_table(ht, tbl);
425
426 if (rht_grow_above_75(ht, tbl))
427 err = rhashtable_rehash_alloc(
428 else if (ht->p.automatic_shrinking &&
429 err = rhashtable_shrink(ht);
430 else if (tbl->nest)
431 err = rhashtable_rehash_alloc(
432
433 if (!err || err == -EEXIST) {
434 int nerr;
435
436 nerr = rhashtable_rehash_table
437 err = err ?: nerr;
438 }
439
440 mutex_unlock(&ht->mutex);
441
442 if (err)
443 schedule_work(&ht->run_work);
444 }
445
446 static int rhashtable_insert_rehash(struct rha
447 struct buc
448 {
449 struct bucket_table *old_tbl;
450 struct bucket_table *new_tbl;
451 unsigned int size;
452 int err;
453
454 old_tbl = rht_dereference_rcu(ht->tbl,
455
456 size = tbl->size;
457
458 err = -EBUSY;
459
460 if (rht_grow_above_75(ht, tbl))
461 size *= 2;
462 /* Do not schedule more than one rehas
463 else if (old_tbl != tbl)
464 goto fail;
465
466 err = -ENOMEM;
467
468 new_tbl = bucket_table_alloc(ht, size,
469 if (new_tbl == NULL)
470 goto fail;
471
472 err = rhashtable_rehash_attach(ht, tbl
473 if (err) {
474 bucket_table_free(new_tbl);
475 if (err == -EEXIST)
476 err = 0;
477 } else
478 schedule_work(&ht->run_work);
479
480 return err;
481
482 fail:
483 /* Do not fail the insert if someone e
484 if (likely(rcu_access_pointer(tbl->fut
485 return 0;
486
487 /* Schedule async rehash to retry allo
488 if (err == -ENOMEM)
489 schedule_work(&ht->run_work);
490
491 return err;
492 }
493
494 static void *rhashtable_lookup_one(struct rhas
495 struct rhas
496 struct buck
497 const void
498 {
499 struct rhashtable_compare_arg arg = {
500 .ht = ht,
501 .key = key,
502 };
503 struct rhash_head __rcu **pprev = NULL
504 struct rhash_head *head;
505 int elasticity;
506
507 elasticity = RHT_ELASTICITY;
508 rht_for_each_from(head, rht_ptr(bkt, t
509 struct rhlist_head *list;
510 struct rhlist_head *plist;
511
512 elasticity--;
513 if (!key ||
514 (ht->p.obj_cmpfn ?
515 ht->p.obj_cmpfn(&arg, rht
516 rhashtable_compare(&arg,
517 pprev = &head->next;
518 continue;
519 }
520
521 if (!ht->rhlist)
522 return rht_obj(ht, hea
523
524 list = container_of(obj, struc
525 plist = container_of(head, str
526
527 RCU_INIT_POINTER(list->next, p
528 head = rht_dereference_bucket(
529 RCU_INIT_POINTER(list->rhead.n
530 if (pprev)
531 rcu_assign_pointer(*pp
532 else
533 /* Need to preserve th
534 rht_assign_locked(bkt,
535
536 return NULL;
537 }
538
539 if (elasticity <= 0)
540 return ERR_PTR(-EAGAIN);
541
542 return ERR_PTR(-ENOENT);
543 }
544
545 static struct bucket_table *rhashtable_insert_
546 struct rhashtable *ht, struct rhash_lo
547 struct bucket_table *tbl, unsigned int
548 void *data)
549 {
550 struct bucket_table *new_tbl;
551 struct rhash_head *head;
552
553 if (!IS_ERR_OR_NULL(data))
554 return ERR_PTR(-EEXIST);
555
556 if (PTR_ERR(data) != -EAGAIN && PTR_ER
557 return ERR_CAST(data);
558
559 new_tbl = rht_dereference_rcu(tbl->fut
560 if (new_tbl)
561 return new_tbl;
562
563 if (PTR_ERR(data) != -ENOENT)
564 return ERR_CAST(data);
565
566 if (unlikely(rht_grow_above_max(ht, tb
567 return ERR_PTR(-E2BIG);
568
569 if (unlikely(rht_grow_above_100(ht, tb
570 return ERR_PTR(-EAGAIN);
571
572 head = rht_ptr(bkt, tbl, hash);
573
574 RCU_INIT_POINTER(obj->next, head);
575 if (ht->rhlist) {
576 struct rhlist_head *list;
577
578 list = container_of(obj, struc
579 RCU_INIT_POINTER(list->next, N
580 }
581
582 /* bkt is always the head of the list,
583 * the lock, which we need to preserve
584 */
585 rht_assign_locked(bkt, obj);
586
587 atomic_inc(&ht->nelems);
588 if (rht_grow_above_75(ht, tbl))
589 schedule_work(&ht->run_work);
590
591 return NULL;
592 }
593
594 static void *rhashtable_try_insert(struct rhas
595 struct rhas
596 {
597 struct bucket_table *new_tbl;
598 struct bucket_table *tbl;
599 struct rhash_lock_head __rcu **bkt;
600 unsigned long flags;
601 unsigned int hash;
602 void *data;
603
604 new_tbl = rcu_dereference(ht->tbl);
605
606 do {
607 tbl = new_tbl;
608 hash = rht_head_hashfn(ht, tbl
609 if (rcu_access_pointer(tbl->fu
610 /* Failure is OK */
611 bkt = rht_bucket_var(t
612 else
613 bkt = rht_bucket_inser
614 if (bkt == NULL) {
615 new_tbl = rht_derefere
616 data = ERR_PTR(-EAGAIN
617 } else {
618 flags = rht_lock(tbl,
619 data = rhashtable_look
620
621 new_tbl = rhashtable_i
622
623 if (PTR_ERR(new_tbl) !
624 data = ERR_CAS
625
626 rht_unlock(tbl, bkt, f
627 }
628 } while (!IS_ERR_OR_NULL(new_tbl));
629
630 if (PTR_ERR(data) == -EAGAIN)
631 data = ERR_PTR(rhashtable_inse
632 -EAGAIN);
633
634 return data;
635 }
636
637 void *rhashtable_insert_slow(struct rhashtable
638 struct rhash_head
639 {
640 void *data;
641
642 do {
643 rcu_read_lock();
644 data = rhashtable_try_insert(h
645 rcu_read_unlock();
646 } while (PTR_ERR(data) == -EAGAIN);
647
648 return data;
649 }
650 EXPORT_SYMBOL_GPL(rhashtable_insert_slow);
651
652 /**
653 * rhashtable_walk_enter - Initialise an itera
654 * @ht: Table to walk over
655 * @iter: Hash table Iterator
656 *
657 * This function prepares a hash table walk.
658 *
659 * Note that if you restart a walk after rhash
660 * may see the same object twice. Also, you m
661 * there are removals in between rhashtable_wa
662 * call to rhashtable_walk_start.
663 *
664 * For a completely stable walk you should con
665 * structure outside the hash table.
666 *
667 * This function may be called from any proces
668 * non-preemptable context, but cannot be call
669 * hardirq context.
670 *
671 * You must call rhashtable_walk_exit after th
672 */
673 void rhashtable_walk_enter(struct rhashtable *
674 {
675 iter->ht = ht;
676 iter->p = NULL;
677 iter->slot = 0;
678 iter->skip = 0;
679 iter->end_of_table = 0;
680
681 spin_lock(&ht->lock);
682 iter->walker.tbl =
683 rcu_dereference_protected(ht->
684 list_add(&iter->walker.list, &iter->wa
685 spin_unlock(&ht->lock);
686 }
687 EXPORT_SYMBOL_GPL(rhashtable_walk_enter);
688
689 /**
690 * rhashtable_walk_exit - Free an iterator
691 * @iter: Hash table Iterator
692 *
693 * This function frees resources allocated by
694 */
695 void rhashtable_walk_exit(struct rhashtable_it
696 {
697 spin_lock(&iter->ht->lock);
698 if (iter->walker.tbl)
699 list_del(&iter->walker.list);
700 spin_unlock(&iter->ht->lock);
701 }
702 EXPORT_SYMBOL_GPL(rhashtable_walk_exit);
703
704 /**
705 * rhashtable_walk_start_check - Start a hash
706 * @iter: Hash table iterator
707 *
708 * Start a hash table walk at the current iter
709 * the RCU lock in all cases including when we
710 * always call rhashtable_walk_stop to clean u
711 *
712 * Returns zero if successful.
713 *
714 * Returns -EAGAIN if resize event occurred.
715 * will rewind back to the beginning and you m
716 * by calling rhashtable_walk_next.
717 *
718 * rhashtable_walk_start is defined as an inli
719 * void. This is preferred in cases where the
720 * resize events and always continue.
721 */
722 int rhashtable_walk_start_check(struct rhashta
723 __acquires(RCU)
724 {
725 struct rhashtable *ht = iter->ht;
726 bool rhlist = ht->rhlist;
727
728 rcu_read_lock();
729
730 spin_lock(&ht->lock);
731 if (iter->walker.tbl)
732 list_del(&iter->walker.list);
733 spin_unlock(&ht->lock);
734
735 if (iter->end_of_table)
736 return 0;
737 if (!iter->walker.tbl) {
738 iter->walker.tbl = rht_derefer
739 iter->slot = 0;
740 iter->skip = 0;
741 return -EAGAIN;
742 }
743
744 if (iter->p && !rhlist) {
745 /*
746 * We need to validate that 'p
747 * if so, update 'skip'
748 */
749 struct rhash_head *p;
750 int skip = 0;
751 rht_for_each_rcu(p, iter->walk
752 skip++;
753 if (p == iter->p) {
754 iter->skip = s
755 goto found;
756 }
757 }
758 iter->p = NULL;
759 } else if (iter->p && rhlist) {
760 /* Need to validate that 'list
761 * if so, update 'skip' and 'p
762 */
763 struct rhash_head *p;
764 struct rhlist_head *list;
765 int skip = 0;
766 rht_for_each_rcu(p, iter->walk
767 for (list = container_
768 list;
769 list = rcu_derefe
770 skip++;
771 if (list == it
772 iter->
773 iter->
774 goto f
775 }
776 }
777 }
778 iter->p = NULL;
779 }
780 found:
781 return 0;
782 }
783 EXPORT_SYMBOL_GPL(rhashtable_walk_start_check)
784
785 /**
786 * __rhashtable_walk_find_next - Find the next
787 * one in case of a new walk).
788 *
789 * @iter: Hash table iterator
790 *
791 * Returns the found object or NULL when the e
792 *
793 * Returns -EAGAIN if resize event occurred.
794 */
795 static void *__rhashtable_walk_find_next(struc
796 {
797 struct bucket_table *tbl = iter->walke
798 struct rhlist_head *list = iter->list;
799 struct rhashtable *ht = iter->ht;
800 struct rhash_head *p = iter->p;
801 bool rhlist = ht->rhlist;
802
803 if (!tbl)
804 return NULL;
805
806 for (; iter->slot < tbl->size; iter->s
807 int skip = iter->skip;
808
809 rht_for_each_rcu(p, tbl, iter-
810 if (rhlist) {
811 list = contain
812
813 do {
814 if (!s
815
816 skip--
817 list =
818 } while (list)
819
820 continue;
821 }
822 if (!skip)
823 break;
824 skip--;
825 }
826
827 next:
828 if (!rht_is_a_nulls(p)) {
829 iter->skip++;
830 iter->p = p;
831 iter->list = list;
832 return rht_obj(ht, rhl
833 }
834
835 iter->skip = 0;
836 }
837
838 iter->p = NULL;
839
840 /* Ensure we see any new tables. */
841 smp_rmb();
842
843 iter->walker.tbl = rht_dereference_rcu
844 if (iter->walker.tbl) {
845 iter->slot = 0;
846 iter->skip = 0;
847 return ERR_PTR(-EAGAIN);
848 } else {
849 iter->end_of_table = true;
850 }
851
852 return NULL;
853 }
854
855 /**
856 * rhashtable_walk_next - Return the next obje
857 * @iter: Hash table iterator
858 *
859 * Note that you must call rhashtable_walk_sto
860 * with the walk.
861 *
862 * Returns the next object or NULL when the en
863 *
864 * Returns -EAGAIN if resize event occurred.
865 * will rewind back to the beginning and you m
866 */
867 void *rhashtable_walk_next(struct rhashtable_i
868 {
869 struct rhlist_head *list = iter->list;
870 struct rhashtable *ht = iter->ht;
871 struct rhash_head *p = iter->p;
872 bool rhlist = ht->rhlist;
873
874 if (p) {
875 if (!rhlist || !(list = rcu_de
876 p = rcu_dereference(p-
877 list = container_of(p,
878 }
879 if (!rht_is_a_nulls(p)) {
880 iter->skip++;
881 iter->p = p;
882 iter->list = list;
883 return rht_obj(ht, rhl
884 }
885
886 /* At the end of this slot, sw
887 * next entry from that point.
888 */
889 iter->skip = 0;
890 iter->slot++;
891 }
892
893 return __rhashtable_walk_find_next(ite
894 }
895 EXPORT_SYMBOL_GPL(rhashtable_walk_next);
896
897 /**
898 * rhashtable_walk_peek - Return the next obje
899 * @iter: Hash table iterator
900 *
901 * Returns the next object or NULL when the en
902 *
903 * Returns -EAGAIN if resize event occurred.
904 * will rewind back to the beginning and you m
905 */
906 void *rhashtable_walk_peek(struct rhashtable_i
907 {
908 struct rhlist_head *list = iter->list;
909 struct rhashtable *ht = iter->ht;
910 struct rhash_head *p = iter->p;
911
912 if (p)
913 return rht_obj(ht, ht->rhlist
914
915 /* No object found in current iter, fi
916
917 if (iter->skip) {
918 /* A nonzero skip value points
919 * beyond that last one that w
920 * we find the current value.
921 * will restore the original v
922 * the table hasn't changed.
923 */
924 iter->skip--;
925 }
926
927 return __rhashtable_walk_find_next(ite
928 }
929 EXPORT_SYMBOL_GPL(rhashtable_walk_peek);
930
931 /**
932 * rhashtable_walk_stop - Finish a hash table
933 * @iter: Hash table iterator
934 *
935 * Finish a hash table walk. Does not reset t
936 * hash table.
937 */
938 void rhashtable_walk_stop(struct rhashtable_it
939 __releases(RCU)
940 {
941 struct rhashtable *ht;
942 struct bucket_table *tbl = iter->walke
943
944 if (!tbl)
945 goto out;
946
947 ht = iter->ht;
948
949 spin_lock(&ht->lock);
950 if (rcu_head_after_call_rcu(&tbl->rcu,
951 /* This bucket table is being
952 iter->walker.tbl = NULL;
953 else
954 list_add(&iter->walker.list, &
955 spin_unlock(&ht->lock);
956
957 out:
958 rcu_read_unlock();
959 }
960 EXPORT_SYMBOL_GPL(rhashtable_walk_stop);
961
962 static size_t rounded_hashtable_size(const str
963 {
964 size_t retsize;
965
966 if (params->nelem_hint)
967 retsize = max(roundup_pow_of_t
968 (unsigned long)p
969 else
970 retsize = max(HASH_DEFAULT_SIZ
971 (unsigned long)p
972
973 return retsize;
974 }
975
976 static u32 rhashtable_jhash2(const void *key,
977 {
978 return jhash2(key, length, seed);
979 }
980
981 /**
982 * rhashtable_init - initialize a new hash tab
983 * @ht: hash table to be initialized
984 * @params: configuration parameters
985 *
986 * Initializes a new hash table based on the p
987 * parameters. A table can be configured eithe
988 * fixed length key:
989 *
990 * Configuration Example 1: Fixed length keys
991 * struct test_obj {
992 * int key;
993 * void * my_member;
994 * struct rhash_head node;
995 * };
996 *
997 * struct rhashtable_params params = {
998 * .head_offset = offsetof(struct test_ob
999 * .key_offset = offsetof(struct test_obj
1000 * .key_len = sizeof(int),
1001 * .hashfn = jhash,
1002 * };
1003 *
1004 * Configuration Example 2: Variable length k
1005 * struct test_obj {
1006 * [...]
1007 * struct rhash_head node;
1008 * };
1009 *
1010 * u32 my_hash_fn(const void *data, u32 len,
1011 * {
1012 * struct test_obj *obj = data;
1013 *
1014 * return [... hash ...];
1015 * }
1016 *
1017 * struct rhashtable_params params = {
1018 * .head_offset = offsetof(struct test_o
1019 * .hashfn = jhash,
1020 * .obj_hashfn = my_hash_fn,
1021 * };
1022 */
1023 int rhashtable_init_noprof(struct rhashtable
1024 const struct rhashtable_p
1025 {
1026 struct bucket_table *tbl;
1027 size_t size;
1028
1029 if ((!params->key_len && !params->obj
1030 (params->obj_hashfn && !params->o
1031 return -EINVAL;
1032
1033 memset(ht, 0, sizeof(*ht));
1034 mutex_init(&ht->mutex);
1035 spin_lock_init(&ht->lock);
1036 memcpy(&ht->p, params, sizeof(*params
1037
1038 alloc_tag_record(ht->alloc_tag);
1039
1040 if (params->min_size)
1041 ht->p.min_size = roundup_pow_
1042
1043 /* Cap total entries at 2^31 to avoid
1044 ht->max_elems = 1u << 31;
1045
1046 if (params->max_size) {
1047 ht->p.max_size = rounddown_po
1048 if (ht->p.max_size < ht->max_
1049 ht->max_elems = ht->p
1050 }
1051
1052 ht->p.min_size = max_t(u16, ht->p.min
1053
1054 size = rounded_hashtable_size(&ht->p)
1055
1056 ht->key_len = ht->p.key_len;
1057 if (!params->hashfn) {
1058 ht->p.hashfn = jhash;
1059
1060 if (!(ht->key_len & (sizeof(u
1061 ht->key_len /= sizeof
1062 ht->p.hashfn = rhasht
1063 }
1064 }
1065
1066 /*
1067 * This is api initialization and thu
1068 * initial rhashtable allocation. Upo
1069 * smallest possible size with __GFP_
1070 */
1071 tbl = bucket_table_alloc(ht, size, GF
1072 if (unlikely(tbl == NULL)) {
1073 size = max_t(u16, ht->p.min_s
1074 tbl = bucket_table_alloc(ht,
1075 }
1076
1077 atomic_set(&ht->nelems, 0);
1078
1079 RCU_INIT_POINTER(ht->tbl, tbl);
1080
1081 INIT_WORK(&ht->run_work, rht_deferred
1082
1083 return 0;
1084 }
1085 EXPORT_SYMBOL_GPL(rhashtable_init_noprof);
1086
1087 /**
1088 * rhltable_init - initialize a new hash list
1089 * @hlt: hash list table to be initial
1090 * @params: configuration parameters
1091 *
1092 * Initializes a new hash list table.
1093 *
1094 * See documentation for rhashtable_init.
1095 */
1096 int rhltable_init_noprof(struct rhltable *hlt
1097 {
1098 int err;
1099
1100 err = rhashtable_init_noprof(&hlt->ht
1101 hlt->ht.rhlist = true;
1102 return err;
1103 }
1104 EXPORT_SYMBOL_GPL(rhltable_init_noprof);
1105
1106 static void rhashtable_free_one(struct rhasht
1107 void (*free_f
1108 void *arg)
1109 {
1110 struct rhlist_head *list;
1111
1112 if (!ht->rhlist) {
1113 free_fn(rht_obj(ht, obj), arg
1114 return;
1115 }
1116
1117 list = container_of(obj, struct rhlis
1118 do {
1119 obj = &list->rhead;
1120 list = rht_dereference(list->
1121 free_fn(rht_obj(ht, obj), arg
1122 } while (list);
1123 }
1124
1125 /**
1126 * rhashtable_free_and_destroy - free element
1127 * @ht: the hash table to destroy
1128 * @free_fn: callback to release resources
1129 * @arg: pointer passed to free_fn
1130 *
1131 * Stops an eventual async resize. If defined
1132 * element to releasal resources. Please note
1133 * readers may still be accessing the element
1134 * must occur in a compatible manner. Then fr
1135 *
1136 * This function will eventually sleep to wai
1137 * to complete. The caller is responsible tha
1138 * occurs in parallel.
1139 */
1140 void rhashtable_free_and_destroy(struct rhash
1141 void (*free_
1142 void *arg)
1143 {
1144 struct bucket_table *tbl, *next_tbl;
1145 unsigned int i;
1146
1147 cancel_work_sync(&ht->run_work);
1148
1149 mutex_lock(&ht->mutex);
1150 tbl = rht_dereference(ht->tbl, ht);
1151 restart:
1152 if (free_fn) {
1153 for (i = 0; i < tbl->size; i+
1154 struct rhash_head *po
1155
1156 cond_resched();
1157 for (pos = rht_ptr_ex
1158 next = !rht_is_a
1159 rht_d
1160 !rht_is_a_nulls(
1161 pos = next,
1162 next = !rht_is_a
1163 rht_d
1164 rhashtable_fr
1165 }
1166 }
1167
1168 next_tbl = rht_dereference(tbl->futur
1169 bucket_table_free(tbl);
1170 if (next_tbl) {
1171 tbl = next_tbl;
1172 goto restart;
1173 }
1174 mutex_unlock(&ht->mutex);
1175 }
1176 EXPORT_SYMBOL_GPL(rhashtable_free_and_destroy
1177
1178 void rhashtable_destroy(struct rhashtable *ht
1179 {
1180 return rhashtable_free_and_destroy(ht
1181 }
1182 EXPORT_SYMBOL_GPL(rhashtable_destroy);
1183
1184 struct rhash_lock_head __rcu **__rht_bucket_n
1185 const struct bucket_table *tbl, unsig
1186 {
1187 const unsigned int shift = PAGE_SHIFT
1188 unsigned int index = hash & ((1 << tb
1189 unsigned int size = tbl->size >> tbl-
1190 unsigned int subhash = hash;
1191 union nested_table *ntbl;
1192
1193 ntbl = nested_table_top(tbl);
1194 ntbl = rht_dereference_bucket_rcu(ntb
1195 subhash >>= tbl->nest;
1196
1197 while (ntbl && size > (1 << shift)) {
1198 index = subhash & ((1 << shif
1199 ntbl = rht_dereference_bucket
1200
1201 size >>= shift;
1202 subhash >>= shift;
1203 }
1204
1205 if (!ntbl)
1206 return NULL;
1207
1208 return &ntbl[subhash].bucket;
1209
1210 }
1211 EXPORT_SYMBOL_GPL(__rht_bucket_nested);
1212
1213 struct rhash_lock_head __rcu **rht_bucket_nes
1214 const struct bucket_table *tbl, unsig
1215 {
1216 static struct rhash_lock_head __rcu *
1217
1218 if (!rhnull)
1219 INIT_RHT_NULLS_HEAD(rhnull);
1220 return __rht_bucket_nested(tbl, hash)
1221 }
1222 EXPORT_SYMBOL_GPL(rht_bucket_nested);
1223
1224 struct rhash_lock_head __rcu **rht_bucket_nes
1225 struct rhashtable *ht, struct bucket_
1226 {
1227 const unsigned int shift = PAGE_SHIFT
1228 unsigned int index = hash & ((1 << tb
1229 unsigned int size = tbl->size >> tbl-
1230 union nested_table *ntbl;
1231
1232 ntbl = nested_table_top(tbl);
1233 hash >>= tbl->nest;
1234 ntbl = nested_table_alloc(ht, &ntbl[i
1235 size <= (1
1236
1237 while (ntbl && size > (1 << shift)) {
1238 index = hash & ((1 << shift)
1239 size >>= shift;
1240 hash >>= shift;
1241 ntbl = nested_table_alloc(ht,
1242 siz
1243 }
1244
1245 if (!ntbl)
1246 return NULL;
1247
1248 return &ntbl[hash].bucket;
1249
1250 }
1251 EXPORT_SYMBOL_GPL(rht_bucket_nested_insert);
1252
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.