]> git.saurik.com Git - apple/xnu.git/blame - libkern/kxld/kxld_dict.c
xnu-4903.221.2.tar.gz
[apple/xnu.git] / libkern / kxld / kxld_dict.c
CommitLineData
b0d623f7
A
1/*
2 * Copyright (c) 2007-2008 Apple Inc. All rights reserved.
3 *
4 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
5 *
6 * This file contains Original Code and/or Modifications of Original Code
7 * as defined in and that are subject to the Apple Public Source License
8 * Version 2.0 (the 'License'). You may not use this file except in
9 * compliance with the License. The rights granted to you under the License
10 * may not be used to create, or enable the creation or redistribution of,
11 * unlawful or unlicensed copies of an Apple operating system, or to
12 * circumvent, violate, or enable the circumvention or violation of, any
13 * terms of an Apple operating system software license agreement.
14 *
15 * Please obtain a copy of the License at
16 * http://www.opensource.apple.com/apsl/ and read it before using this file.
17 *
18 * The Original Code and all software distributed under the License are
19 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
20 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
21 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
22 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
23 * Please see the License for the specific language governing rights and
24 * limitations under the License.
25 *
26 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
27 */
28#include <string.h>
29#include <sys/types.h>
30
31#define DEBUG_ASSERT_COMPONENT_NAME_STRING "kxld"
32#include <AssertMacros.h>
33
34#include "kxld_dict.h"
35#include "kxld_util.h"
36
37/*******************************************************************************
38* Types and macros
39*******************************************************************************/
40
41/* Ratio of num_entries:num_buckets that will cause a resize */
42#define RESIZE_NUMER 7
43#define RESIZE_DENOM 10
44#define RESIZE_THRESHOLD(x) (((x)*RESIZE_NUMER) / RESIZE_DENOM)
45#define MIN_BUCKETS(x) (((x)*RESIZE_DENOM) / RESIZE_NUMER)
46
47/* Selected for good scaling qualities when resizing dictionary
48 * ... see: http://www.concentric.net/~ttwang/tech/hashsize.htm
49 */
50#define DEFAULT_DICT_SIZE 89
51
52typedef struct dict_entry DictEntry;
53
54typedef enum {
55 EMPTY = 0,
56 USED = 1,
57 DELETED = 2
58} DictEntryState;
59
60struct dict_entry {
61 const void *key;
62 void *value;
63 DictEntryState state;
64};
65
66/*******************************************************************************
67* Function prototypes
68*******************************************************************************/
69
70static kern_return_t get_locate_index(const KXLDDict *dict, const void *key,
71 u_int *idx);
72static kern_return_t get_insert_index(const KXLDDict *dict, const void *key,
73 u_int *idx);
74static kern_return_t resize_dict(KXLDDict *dict);
75
76/*******************************************************************************
77*******************************************************************************/
78kern_return_t
79kxld_dict_init(KXLDDict * dict, kxld_dict_hash hash, kxld_dict_cmp cmp,
80 u_int num_entries)
81{
82 kern_return_t rval = KERN_FAILURE;
83 u_int min_buckets = MIN_BUCKETS(num_entries);
84 u_int num_buckets = DEFAULT_DICT_SIZE;
85
86 check(dict);
87 check(hash);
88 check(cmp);
89
90 /* We want the number of allocated buckets to be at least twice that of the
91 * number to be inserted.
92 */
93 while (min_buckets > num_buckets) {
94 num_buckets *= 2;
95 num_buckets++;
96 }
97
98 /* Allocate enough buckets for the anticipated number of entries */
99 rval = kxld_array_init(&dict->buckets, sizeof(DictEntry), num_buckets);
100 require_noerr(rval, finish);
101
102 /* Initialize */
103 dict->hash = hash;
104 dict->cmp = cmp;
105 dict->num_entries = 0;
106 dict->resize_threshold = RESIZE_THRESHOLD(num_buckets);
107
108 rval = KERN_SUCCESS;
109
110finish:
111 return rval;
112}
113
114/*******************************************************************************
115*******************************************************************************/
116void
117kxld_dict_clear(KXLDDict *dict)
118{
119 check(dict);
120
121 dict->hash = NULL;
122 dict->cmp = NULL;
123 dict->num_entries = 0;
124 dict->resize_threshold = 0;
125 kxld_array_clear(&dict->buckets);
126 kxld_array_clear(&dict->resize_buckets);
127}
128
129/*******************************************************************************
130*******************************************************************************/
131void
132kxld_dict_iterator_init(KXLDDictIterator *iter, const KXLDDict *dict)
133{
134 check(iter);
135 check(dict);
136
137 iter->idx = 0;
138 iter->dict = dict;
139}
140
141/*******************************************************************************
142*******************************************************************************/
143void
144kxld_dict_deinit(KXLDDict *dict)
145{
146 check(dict);
147
148 kxld_array_deinit(&dict->buckets);
149 kxld_array_deinit(&dict->resize_buckets);
150}
151
152/*******************************************************************************
153*******************************************************************************/
154u_int
155kxld_dict_get_num_entries(const KXLDDict *dict)
156{
157 check(dict);
158
159 return dict->num_entries;
160}
161
162/*******************************************************************************
163*******************************************************************************/
164void *
165kxld_dict_find(const KXLDDict *dict, const void *key)
166{
167 kern_return_t rval = KERN_FAILURE;
168 DictEntry *entry = NULL;
169 u_int idx = 0;
170
171 check(dict);
172 check(key);
173
174 rval = get_locate_index(dict, key, &idx);
175 if (rval) return NULL;
176
177 entry = kxld_array_get_item(&dict->buckets, idx);
178
179 return entry->value;
180}
181
182/*******************************************************************************
183* This dictionary uses linear probing, which means that when there is a
184* collision, we just walk along the buckets until a free bucket shows up.
185* A consequence of this is that when looking up an item, items that lie between
186* its hash value and its actual bucket may have been deleted since it was
187* inserted. Thus, we should only stop a lookup when we've wrapped around the
188* dictionary or encountered an EMPTY bucket.
189********************************************************************************/
190static kern_return_t
191get_locate_index(const KXLDDict *dict, const void *key, u_int *_idx)
192{
193 kern_return_t rval = KERN_FAILURE;
194 DictEntry *entry = NULL;
195 u_int base, idx;
196
197 base = idx = dict->hash(dict, key);
198
199 /* Iterate until we match the key, wrap, or hit an empty bucket */
200 entry = kxld_array_get_item(&dict->buckets, idx);
201 while (!dict->cmp(entry->key, key)) {
202 if (entry->state == EMPTY) goto finish;
203
204 idx = (idx + 1) % dict->buckets.nitems;
205 if (idx == base) goto finish;
206
207 entry = kxld_array_get_item(&dict->buckets, idx);
208 }
209
210 check(idx < dict->buckets.nitems);
211
212 *_idx = idx;
213 rval = KERN_SUCCESS;
214
215finish:
216 return rval;
217}
218
219/*******************************************************************************
220*******************************************************************************/
221kern_return_t
222kxld_dict_insert(KXLDDict *dict, const void *key, void *value)
223{
224 kern_return_t rval = KERN_FAILURE;
225 DictEntry *entry = NULL;
226 u_int idx = 0;
227
228 check(dict);
229 check(key);
230 check(value);
231
232 /* Resize if we are greater than the capacity threshold.
233 * Note: this is expensive, but the dictionary can be sized correctly at
234 * construction to avoid ever having to do this.
235 */
236 while (dict->num_entries > dict->resize_threshold) {
237 rval = resize_dict(dict);
238 require_noerr(rval, finish);
239 }
240
241 /* If this function returns FULL after we've already resized appropriately
242 * something is very wrong and we should return an error.
243 */
244 rval = get_insert_index(dict, key, &idx);
245 require_noerr(rval, finish);
246
247 /* Insert the new key-value pair into the bucket, but only count it as a
248 * new entry if we are not overwriting an existing entry.
249 */
250 entry = kxld_array_get_item(&dict->buckets, idx);
251 if (entry->state != USED) {
252 dict->num_entries++;
253 entry->key = key;
254 entry->state = USED;
255 }
256 entry->value = value;
257
258 rval = KERN_SUCCESS;
259
260finish:
261 return rval;
262}
263
264/*******************************************************************************
265* Increases the hash table's capacity by 2N+1. Uses dictionary API. Not
266* fast; just correct.
267*******************************************************************************/
268static kern_return_t
269resize_dict(KXLDDict *dict)
270{
271 kern_return_t rval = KERN_FAILURE;
272 KXLDArray tmparray;
273 DictEntry *entry = NULL;
274 u_int nbuckets = (dict->buckets.nitems * 2 + 1);
275 u_int i = 0;
276
277 check(dict);
278
279 /* Initialize a new set of buckets to hold more entries */
280 rval = kxld_array_init(&dict->resize_buckets, sizeof(DictEntry), nbuckets);
281 require_noerr(rval, finish);
282
283 /* Swap the new buckets with the old buckets */
284 tmparray = dict->buckets;
285 dict->buckets = dict->resize_buckets;
286 dict->resize_buckets = tmparray;
287
288 /* Reset dictionary parameters */
289 dict->num_entries = 0;
290 dict->resize_threshold = RESIZE_THRESHOLD(dict->buckets.nitems);
291
292 /* Rehash all of the entries */
293 for (i = 0; i < dict->resize_buckets.nitems; ++i) {
294 entry = kxld_array_get_item(&dict->resize_buckets, i);
295 if (entry->state == USED) {
296 rval = kxld_dict_insert(dict, entry->key, entry->value);
297 require_noerr(rval, finish);
298 }
299 }
300
301 /* Clear the old buckets */
302 kxld_array_clear(&dict->resize_buckets);
303
304 rval = KERN_SUCCESS;
305
306finish:
307 return rval;
308}
309
310/*******************************************************************************
311* Simple function to find the first empty cell
312*******************************************************************************/
313static kern_return_t
314get_insert_index(const KXLDDict *dict, const void *key, u_int *r_index)
315{
316 kern_return_t rval = KERN_FAILURE;
317 DictEntry *entry = NULL;
318 u_int base, idx;
319
320 base = idx = dict->hash(dict, key);
321
322 /* Iterate through the buckets until we find an EMPTY bucket, a DELETED
323 * bucket, or a key match.
324 */
325 entry = kxld_array_get_item(&dict->buckets, idx);
326 while (entry->state == USED && !dict->cmp(entry->key, key)) {
327 idx = (idx + 1) % dict->buckets.nitems;
328 require_action(base != idx, finish, rval=KERN_FAILURE);
329 entry = kxld_array_get_item(&dict->buckets, idx);
330 }
331
332 *r_index = idx;
333 rval = KERN_SUCCESS;
334
335finish:
336 return rval;
337}
338
339/*******************************************************************************
340*******************************************************************************/
341void
342kxld_dict_remove(KXLDDict *dict, const void *key, void **value)
343{
344 kern_return_t rval = KERN_FAILURE;
345 DictEntry *entry = NULL;
346 u_int idx = 0;
347
348 check(dict);
349 check(key);
350
351 /* Find the item */
352 rval = get_locate_index(dict, key, &idx);
353 if (rval) {
354 if (value) *value = NULL;
355 return;
356 }
357
358 entry = kxld_array_get_item(&dict->buckets, idx);
359
360 /* Save the value if requested */
361 if (value) *value = entry->value;
362
363 /* Delete the item from the dictionary */
364 entry->key = NULL;
365 entry->value = NULL;
366 entry->state = DELETED;
367 dict->num_entries--;
368}
369
370/*******************************************************************************
371*******************************************************************************/
372void
373kxld_dict_iterator_get_next(KXLDDictIterator *iter, const void **key,
374 void **value)
375{
376 DictEntry *entry = NULL;
377
378 check(iter);
379 check(key);
380 check(value);
381
382 *key = NULL;
383 *value = NULL;
384
385 /* Walk over the dictionary looking for USED buckets */
386 for (; iter->idx < iter->dict->buckets.nitems; ++(iter->idx)) {
387 entry = kxld_array_get_item(&iter->dict->buckets, iter->idx);
388 if (entry->state == USED) {
389 *key = entry->key;
390 *value = entry->value;
391 ++(iter->idx);
392 break;
393 }
394 }
395}
396
397/*******************************************************************************
398*******************************************************************************/
399void
400kxld_dict_iterator_reset(KXLDDictIterator *iter)
401{
402 iter->idx = 0;
403}
404
405/*******************************************************************************
406* This is Daniel Bernstein's hash algorithm from comp.lang.c
407* It's fast and distributes well. Returns an idx into the symbol hash table.
408* NOTE: Will not check for a valid pointer - performance
409*******************************************************************************/
410u_int
411kxld_dict_string_hash(const KXLDDict *dict, const void *_key)
412{
413 const char *key = _key;
414 u_int c = 0;
415 u_int hash_val = 5381;
416
417 check(dict);
418 check(_key);
419
420 while ((c = *key++)) {
421 /* hash(i) = hash(i-1) *33 ^ name[i] */
422 hash_val = ((hash_val << 5) + hash_val) ^ c;
423 }
424
425 return (hash_val % dict->buckets.nitems);
426}
427
428u_int
429kxld_dict_uint32_hash(const KXLDDict *dict, const void *_key)
430{
431 uint32_t key = *(const uint32_t *) _key;
432
433 check(_key);
434
435 return (u_int) (key % dict->buckets.nitems);
436}
437
438u_int
439kxld_dict_kxldaddr_hash(const KXLDDict *dict, const void *_key)
440{
441 kxld_addr_t key = *(const kxld_addr_t *) _key;
442
443 check(_key);
444
445 return (u_int) (key % dict->buckets.nitems);
446}
447
448u_int
449kxld_dict_string_cmp(const void *key1, const void *key2)
450{
451 return streq(key1, key2);
452}
453
454u_int
455kxld_dict_uint32_cmp(const void *key1, const void *key2)
456{
457 const uint32_t *a = key1;
458 const uint32_t *b = key2;
459
460 return (a && b && (*a == *b));
461}
462
463u_int
464kxld_dict_kxldaddr_cmp(const void *key1, const void *key2)
465{
466 const kxld_addr_t *a = key1;
467 const kxld_addr_t *b = key2;
468
469 return (a && b && (*a == *b));
470}
471