mbcache.c

   1 // SPDX-License-Identifier: GPL-2.0-only
   2 #include <linux/spinlock.h>
   3 #include <linux/slab.h>
   4 #include <linux/list.h>
   5 #include <linux/list_bl.h>
   6 #include <linux/module.h>
   7 #include <linux/sched.h>
   8 #include <linux/workqueue.h>
   9 #include <linux/mbcache.h>
  10
  11 /*
  12  * Mbcache is a simple key-value store. Keys need not be unique, however
  13  * key-value pairs are expected to be unique (we use this fact in
  14  * mb_cache_entry_delete()).
  15  *
  16  * Ext2 and ext4 use this cache for deduplication of extended attribute blocks.
  17  * Ext4 also uses it for deduplication of xattr values stored in inodes.
  18  * They use hash of data as a key and provide a value that may represent a
  19  * block or inode number. That's why keys need not be unique (hash of different
  20  * data may be the same). However user provided value always uniquely
  21  * identifies a cache entry.
  22  *
  23  * We provide functions for creation and removal of entries, search by key,
  24  * and a special "delete entry with given key-value pair" operation. Fixed
  25  * size hash table is used for fast key lookups.
  26  */
  27
  28 struct mb_cache {
  29         /* Hash table of entries */
  30         struct hlist_bl_head    *c_hash;
  31         /* log2 of hash table size */
  32         int                     c_bucket_bits;
  33         /* Maximum entries in cache to avoid degrading hash too much */
  34         unsigned long           c_max_entries;
  35         /* Protects c_list, c_entry_count */
  36         spinlock_t              c_list_lock;
  37         struct list_head        c_list;
  38         /* Number of entries in cache */
  39         unsigned long           c_entry_count;
  40         struct shrinker         c_shrink;
  41         /* Work for shrinking when the cache has too many entries */
  42         struct work_struct      c_shrink_work;
  43 };
  44
  45 static struct kmem_cache *mb_entry_cache;
  46
  47 static unsigned long mb_cache_shrink(struct mb_cache *cache,
  48                                      unsigned long nr_to_scan);
  49
  50 static inline struct hlist_bl_head *mb_cache_entry_head(struct mb_cache *cache,
  51                                                         u32 key)
  52 {
  53         return &cache->c_hash[hash_32(key, cache->c_bucket_bits)];
  54 }
  55
  56 /*
  57  * Number of entries to reclaim synchronously when there are too many entries
  58  * in cache
  59  */
  60 #define SYNC_SHRINK_BATCH 64
  61
  62 /*
  63  * mb_cache_entry_create - create entry in cache
  64  * @cache - cache where the entry should be created
  65  * @mask - gfp mask with which the entry should be allocated
  66  * @key - key of the entry
  67  * @value - value of the entry
  68  * @reusable - is the entry reusable by others?
  69  *
  70  * Creates entry in @cache with key @key and value @value. The function returns
  71  * -EBUSY if entry with the same key and value already exists in cache.
  72  * Otherwise 0 is returned.
  73  */
  74 int mb_cache_entry_create(struct mb_cache *cache, gfp_t mask, u32 key,
  75                           u64 value, bool reusable)
  76 {
  77         struct mb_cache_entry *entry, *dup;
  78         struct hlist_bl_node *dup_node;
  79         struct hlist_bl_head *head;
  80
  81         /* Schedule background reclaim if there are too many entries */
  82         if (cache->c_entry_count >= cache->c_max_entries)
  83                 schedule_work(&cache->c_shrink_work);
  84         /* Do some sync reclaim if background reclaim cannot keep up */
  85         if (cache->c_entry_count >= 2*cache->c_max_entries)
  86                 mb_cache_shrink(cache, SYNC_SHRINK_BATCH);
  87
  88         entry = kmem_cache_alloc(mb_entry_cache, mask);
  89         if (!entry)
  90                 return -ENOMEM;
  91
  92         INIT_LIST_HEAD(&entry->e_list);
  93         /* One ref for hash, one ref returned */
  94         atomic_set(&entry->e_refcnt, 1);
  95         entry->e_key = key;
  96         entry->e_value = value;
  97         entry->e_reusable = reusable;
  98         entry->e_referenced = 0;
  99         head = mb_cache_entry_head(cache, key);
 100         hlist_bl_lock(head);
 101         hlist_bl_for_each_entry(dup, dup_node, head, e_hash_list) {
 102                 if (dup->e_key == key && dup->e_value == value) {
 103                         hlist_bl_unlock(head);
 104                         kmem_cache_free(mb_entry_cache, entry);
 105                         return -EBUSY;
 106                 }
 107         }
 108         hlist_bl_add_head(&entry->e_hash_list, head);
 109         hlist_bl_unlock(head);
 110
 111         spin_lock(&cache->c_list_lock);
 112         list_add_tail(&entry->e_list, &cache->c_list);
 113         /* Grab ref for LRU list */
 114         atomic_inc(&entry->e_refcnt);
 115         cache->c_entry_count++;
 116         spin_unlock(&cache->c_list_lock);
 117
 118         return 0;
 119 }
 120 EXPORT_SYMBOL(mb_cache_entry_create);
 121
 122 void __mb_cache_entry_free(struct mb_cache_entry *entry)
 123 {
 124         kmem_cache_free(mb_entry_cache, entry);
 125 }
 126 EXPORT_SYMBOL(__mb_cache_entry_free);
 127
 128 static struct mb_cache_entry *__entry_find(struct mb_cache *cache,
 129                                            struct mb_cache_entry *entry,
 130                                            u32 key)
 131 {
 132         struct mb_cache_entry *old_entry = entry;
 133         struct hlist_bl_node *node;
 134         struct hlist_bl_head *head;
 135
 136         head = mb_cache_entry_head(cache, key);
 137         hlist_bl_lock(head);
 138         if (entry && !hlist_bl_unhashed(&entry->e_hash_list))
 139                 node = entry->e_hash_list.next;
 140         else
 141                 node = hlist_bl_first(head);
 142         while (node) {
 143                 entry = hlist_bl_entry(node, struct mb_cache_entry,
 144                                        e_hash_list);
 145                 if (entry->e_key == key && entry->e_reusable) {
 146                         atomic_inc(&entry->e_refcnt);
 147                         goto out;
 148                 }
 149                 node = node->next;
 150         }
 151         entry = NULL;
 152 out:
 153         hlist_bl_unlock(head);
 154         if (old_entry)
 155                 mb_cache_entry_put(cache, old_entry);
 156
 157         return entry;
 158 }
 159
 160 /*
 161  * mb_cache_entry_find_first - find the first reusable entry with the given key
 162  * @cache: cache where we should search
 163  * @key: key to look for
 164  *
 165  * Search in @cache for a reusable entry with key @key. Grabs reference to the
 166  * first reusable entry found and returns the entry.
 167  */
 168 struct mb_cache_entry *mb_cache_entry_find_first(struct mb_cache *cache,
 169                                                  u32 key)
 170 {
 171         return __entry_find(cache, NULL, key);
 172 }
 173 EXPORT_SYMBOL(mb_cache_entry_find_first);
 174
 175 /*
 176  * mb_cache_entry_find_next - find next reusable entry with the same key
 177  * @cache: cache where we should search
 178  * @entry: entry to start search from
 179  *
 180  * Finds next reusable entry in the hash chain which has the same key as @entry.
 181  * If @entry is unhashed (which can happen when deletion of entry races with the
 182  * search), finds the first reusable entry in the hash chain. The function drops
 183  * reference to @entry and returns with a reference to the found entry.
 184  */
 185 struct mb_cache_entry *mb_cache_entry_find_next(struct mb_cache *cache,
 186                                                 struct mb_cache_entry *entry)
 187 {
 188         return __entry_find(cache, entry, entry->e_key);
 189 }
 190 EXPORT_SYMBOL(mb_cache_entry_find_next);
 191
 192 /*
 193  * mb_cache_entry_get - get a cache entry by value (and key)
 194  * @cache - cache we work with
 195  * @key - key
 196  * @value - value
 197  */
 198 struct mb_cache_entry *mb_cache_entry_get(struct mb_cache *cache, u32 key,
 199                                           u64 value)
 200 {
 201         struct hlist_bl_node *node;
 202         struct hlist_bl_head *head;
 203         struct mb_cache_entry *entry;
 204
 205         head = mb_cache_entry_head(cache, key);
 206         hlist_bl_lock(head);
 207         hlist_bl_for_each_entry(entry, node, head, e_hash_list) {
 208                 if (entry->e_key == key && entry->e_value == value) {
 209                         atomic_inc(&entry->e_refcnt);
 210                         goto out;
 211                 }
 212         }
 213         entry = NULL;
 214 out:
 215         hlist_bl_unlock(head);
 216         return entry;
 217 }
 218 EXPORT_SYMBOL(mb_cache_entry_get);
 219
 220 /* mb_cache_entry_delete - remove a cache entry
 221  * @cache - cache we work with
 222  * @key - key
 223  * @value - value
 224  *
 225  * Remove entry from cache @cache with key @key and value @value.
 226  */
 227 void mb_cache_entry_delete(struct mb_cache *cache, u32 key, u64 value)
 228 {
 229         struct hlist_bl_node *node;
 230         struct hlist_bl_head *head;
 231         struct mb_cache_entry *entry;
 232
 233         head = mb_cache_entry_head(cache, key);
 234         hlist_bl_lock(head);
 235         hlist_bl_for_each_entry(entry, node, head, e_hash_list) {
 236                 if (entry->e_key == key && entry->e_value == value) {
 237                         /* We keep hash list reference to keep entry alive */
 238                         hlist_bl_del_init(&entry->e_hash_list);
 239                         hlist_bl_unlock(head);
 240                         spin_lock(&cache->c_list_lock);
 241                         if (!list_empty(&entry->e_list)) {
 242                                 list_del_init(&entry->e_list);
 243                                 if (!WARN_ONCE(cache->c_entry_count == 0,
 244                 "mbcache: attempt to decrement c_entry_count past zero"))
 245                                         cache->c_entry_count--;
 246                                 atomic_dec(&entry->e_refcnt);
 247                         }
 248                         spin_unlock(&cache->c_list_lock);
 249                         mb_cache_entry_put(cache, entry);
 250                         return;
 251                 }
 252         }
 253         hlist_bl_unlock(head);
 254 }
 255 EXPORT_SYMBOL(mb_cache_entry_delete);
 256
 257 /* mb_cache_entry_touch - cache entry got used
 258  * @cache - cache the entry belongs to
 259  * @entry - entry that got used
 260  *
 261  * Marks entry as used to give hit higher chances of surviving in cache.
 262  */
 263 void mb_cache_entry_touch(struct mb_cache *cache,
 264                           struct mb_cache_entry *entry)
 265 {
 266         entry->e_referenced = 1;
 267 }
 268 EXPORT_SYMBOL(mb_cache_entry_touch);
 269
 270 static unsigned long mb_cache_count(struct shrinker *shrink,
 271                                     struct shrink_control *sc)
 272 {
 273         struct mb_cache *cache = container_of(shrink, struct mb_cache,
 274                                               c_shrink);
 275
 276         return cache->c_entry_count;
 277 }
 278
 279 /* Shrink number of entries in cache */
 280 static unsigned long mb_cache_shrink(struct mb_cache *cache,
 281                                      unsigned long nr_to_scan)
 282 {
 283         struct mb_cache_entry *entry;
 284         struct hlist_bl_head *head;
 285         unsigned long shrunk = 0;
 286
 287         spin_lock(&cache->c_list_lock);
 288         while (nr_to_scan-- && !list_empty(&cache->c_list)) {
 289                 entry = list_first_entry(&cache->c_list,
 290                                          struct mb_cache_entry, e_list);
 291                 if (entry->e_referenced) {
 292                         entry->e_referenced = 0;
 293                         list_move_tail(&entry->e_list, &cache->c_list);
 294                         continue;
 295                 }
 296                 list_del_init(&entry->e_list);
 297                 cache->c_entry_count--;
 298                 /*
 299                  * We keep LRU list reference so that entry doesn't go away
 300                  * from under us.
 301                  */
 302                 spin_unlock(&cache->c_list_lock);
 303                 head = mb_cache_entry_head(cache, entry->e_key);
 304                 hlist_bl_lock(head);
 305                 if (!hlist_bl_unhashed(&entry->e_hash_list)) {
 306                         hlist_bl_del_init(&entry->e_hash_list);
 307                         atomic_dec(&entry->e_refcnt);
 308                 }
 309                 hlist_bl_unlock(head);
 310                 if (mb_cache_entry_put(cache, entry))
 311                         shrunk++;
 312                 cond_resched();
 313                 spin_lock(&cache->c_list_lock);
 314         }
 315         spin_unlock(&cache->c_list_lock);
 316
 317         return shrunk;
 318 }
 319
 320 static unsigned long mb_cache_scan(struct shrinker *shrink,
 321                                    struct shrink_control *sc)
 322 {
 323         struct mb_cache *cache = container_of(shrink, struct mb_cache,
 324                                               c_shrink);
 325         return mb_cache_shrink(cache, sc->nr_to_scan);
 326 }
 327
 328 /* We shrink 1/X of the cache when we have too many entries in it */
 329 #define SHRINK_DIVISOR 16
 330
 331 static void mb_cache_shrink_worker(struct work_struct *work)
 332 {
 333         struct mb_cache *cache = container_of(work, struct mb_cache,
 334                                               c_shrink_work);
 335         mb_cache_shrink(cache, cache->c_max_entries / SHRINK_DIVISOR);
 336 }
 337
 338 /*
 339  * mb_cache_create - create cache
 340  * @bucket_bits: log2 of the hash table size
 341  *
 342  * Create cache for keys with 2^bucket_bits hash entries.
 343  */
 344 struct mb_cache *mb_cache_create(int bucket_bits)
 345 {
 346         struct mb_cache *cache;
 347         unsigned long bucket_count = 1UL << bucket_bits;
 348         unsigned long i;
 349
 350         cache = kzalloc(sizeof(struct mb_cache), GFP_KERNEL);
 351         if (!cache)
 352                 goto err_out;
 353         cache->c_bucket_bits = bucket_bits;
 354         cache->c_max_entries = bucket_count << 4;
 355         INIT_LIST_HEAD(&cache->c_list);
 356         spin_lock_init(&cache->c_list_lock);
 357         cache->c_hash = kmalloc_array(bucket_count,
 358                                       sizeof(struct hlist_bl_head),
 359                                       GFP_KERNEL);
 360         if (!cache->c_hash) {
 361                 kfree(cache);
 362                 goto err_out;
 363         }
 364         for (i = 0; i < bucket_count; i++)
 365                 INIT_HLIST_BL_HEAD(&cache->c_hash[i]);
 366
 367         cache->c_shrink.count_objects = mb_cache_count;
 368         cache->c_shrink.scan_objects = mb_cache_scan;
 369         cache->c_shrink.seeks = DEFAULT_SEEKS;
 370         if (register_shrinker(&cache->c_shrink)) {
 371                 kfree(cache->c_hash);
 372                 kfree(cache);
 373                 goto err_out;
 374         }
 375
 376         INIT_WORK(&cache->c_shrink_work, mb_cache_shrink_worker);
 377
 378         return cache;
 379
 380 err_out:
 381         return NULL;
 382 }
 383 EXPORT_SYMBOL(mb_cache_create);
 384
 385 /*
 386  * mb_cache_destroy - destroy cache
 387  * @cache: the cache to destroy
 388  *
 389  * Free all entries in cache and cache itself. Caller must make sure nobody
 390  * (except shrinker) can reach @cache when calling this.
 391  */
 392 void mb_cache_destroy(struct mb_cache *cache)
 393 {
 394         struct mb_cache_entry *entry, *next;
 395
 396         unregister_shrinker(&cache->c_shrink);
 397
 398         /*
 399          * We don't bother with any locking. Cache must not be used at this
 400          * point.
 401          */
 402         list_for_each_entry_safe(entry, next, &cache->c_list, e_list) {
 403                 if (!hlist_bl_unhashed(&entry->e_hash_list)) {
 404                         hlist_bl_del_init(&entry->e_hash_list);
 405                         atomic_dec(&entry->e_refcnt);
 406                 } else
 407                         WARN_ON(1);
 408                 list_del(&entry->e_list);
 409                 WARN_ON(atomic_read(&entry->e_refcnt) != 1);
 410                 mb_cache_entry_put(cache, entry);
 411         }
 412         kfree(cache->c_hash);
 413         kfree(cache);
 414 }
 415 EXPORT_SYMBOL(mb_cache_destroy);
 416
 417 static int __init mbcache_init(void)
 418 {
 419         mb_entry_cache = kmem_cache_create("mbcache",
 420                                 sizeof(struct mb_cache_entry), 0,
 421                                 SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL);
 422         if (!mb_entry_cache)
 423                 return -ENOMEM;
 424         return 0;
 425 }
 426
 427 static void __exit mbcache_exit(void)
 428 {
 429         kmem_cache_destroy(mb_entry_cache);
 430 }
 431
 432 module_init(mbcache_init)
 433 module_exit(mbcache_exit)
 434
 435 MODULE_AUTHOR("Jan Kara <jack@suse.cz>");
 436 MODULE_DESCRIPTION("Meta block cache (for extended attributes)");
 437 MODULE_LICENSE("GPL");