1 /* Copyright (c) 2013, Vsevolod Stakhov
2 * All rights reserved.
3 *
4 * Redistribution and use in source and binary forms, with or without
5 * modification, are permitted provided that the following conditions are met:
6 * * Redistributions of source code must retain the above copyright
7 * notice, this list of conditions and the following disclaimer.
8 * * Redistributions in binary form must reproduce the above copyright
9 * notice, this list of conditions and the following disclaimer in the
10 * documentation and/or other materials provided with the distribution.
11 *
12 * THIS SOFTWARE IS PROVIDED ''AS IS'' AND ANY
13 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
14 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
15 * DISCLAIMED. IN NO EVENT SHALL AUTHOR BE LIABLE FOR ANY
16 * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
17 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
18 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
19 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
20 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
21 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
22 */
23
24 #include "ucl_internal.h"
25 #include "ucl_hash.h"
26 #include "khash.h"
27 #include "kvec.h"
28 #include "mum.h"
29
30 #include <time.h>
31 #include <limits.h>
32
33 struct ucl_hash_elt {
34 const ucl_object_t *obj;
35 size_t ar_idx;
36 };
37
38 struct ucl_hash_struct {
39 void *hash;
40 kvec_t(const ucl_object_t *) ar;
41 bool caseless;
42 };
43
44 static uint64_t
ucl_hash_seed(void)45 ucl_hash_seed (void)
46 {
47 static uint64_t seed;
48
49 if (seed == 0) {
50 #ifdef UCL_RANDOM_FUNCTION
51 seed = UCL_RANDOM_FUNCTION;
52 #else
53 /* Not very random but can be useful for our purposes */
54 seed = time (NULL);
55 #endif
56 }
57
58 return seed;
59 }
60
61 static const unsigned char lc_map[256] = {
62 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
63 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
64 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
65 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,
66 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27,
67 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f,
68 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
69 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f,
70 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
71 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f,
72 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77,
73 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f,
74 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
75 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f,
76 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77,
77 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f,
78 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87,
79 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f,
80 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97,
81 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f,
82 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7,
83 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf,
84 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7,
85 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf,
86 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7,
87 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf,
88 0xd0, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7,
89 0xd8, 0xd9, 0xda, 0xdb, 0xdc, 0xdd, 0xde, 0xdf,
90 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7,
91 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef,
92 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7,
93 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff
94 };
95
96 #if (defined(WORD_BIT) && WORD_BIT == 64) || \
97 (defined(__WORDSIZE) && __WORDSIZE == 64) || \
98 defined(__x86_64__) || \
99 defined(__amd64__)
100 #define UCL64_BIT_HASH 1
101 #endif
102
103 static inline uint32_t
ucl_hash_func(const ucl_object_t * o)104 ucl_hash_func (const ucl_object_t *o)
105 {
106 return mum_hash (o->key, o->keylen, ucl_hash_seed ());
107 }
108 static inline int
ucl_hash_equal(const ucl_object_t * k1,const ucl_object_t * k2)109 ucl_hash_equal (const ucl_object_t *k1, const ucl_object_t *k2)
110 {
111 if (k1->keylen == k2->keylen) {
112 return memcmp (k1->key, k2->key, k1->keylen) == 0;
113 }
114
115 return 0;
116 }
117
118 KHASH_INIT (ucl_hash_node, const ucl_object_t *, struct ucl_hash_elt, 1,
119 ucl_hash_func, ucl_hash_equal)
120
121 static inline uint32_t
ucl_hash_caseless_func(const ucl_object_t * o)122 ucl_hash_caseless_func (const ucl_object_t *o)
123 {
124 unsigned len = o->keylen;
125 unsigned leftover = o->keylen % 8;
126 unsigned fp, i;
127 const uint8_t* s = (const uint8_t*)o->key;
128 union {
129 struct {
130 unsigned char c1, c2, c3, c4, c5, c6, c7, c8;
131 } c;
132 uint64_t pp;
133 } u;
134 uint64_t r;
135
136 fp = len - leftover;
137 r = ucl_hash_seed ();
138
139 for (i = 0; i != fp; i += 8) {
140 u.c.c1 = s[i], u.c.c2 = s[i + 1], u.c.c3 = s[i + 2], u.c.c4 = s[i + 3];
141 u.c.c5 = s[i + 4], u.c.c6 = s[i + 5], u.c.c7 = s[i + 6], u.c.c8 = s[i + 7];
142 u.c.c1 = lc_map[u.c.c1];
143 u.c.c2 = lc_map[u.c.c2];
144 u.c.c3 = lc_map[u.c.c3];
145 u.c.c4 = lc_map[u.c.c4];
146 u.c.c5 = lc_map[u.c.c5];
147 u.c.c6 = lc_map[u.c.c6];
148 u.c.c7 = lc_map[u.c.c7];
149 u.c.c8 = lc_map[u.c.c8];
150 r = mum_hash_step (r, u.pp);
151 }
152
153 u.pp = 0;
154 switch (leftover) {
155 case 7:
156 u.c.c7 = lc_map[(unsigned char)s[i++]];
157 /* FALLTHRU */
158 case 6:
159 u.c.c6 = lc_map[(unsigned char)s[i++]];
160 /* FALLTHRU */
161 case 5:
162 u.c.c5 = lc_map[(unsigned char)s[i++]];
163 /* FALLTHRU */
164 case 4:
165 u.c.c4 = lc_map[(unsigned char)s[i++]];
166 /* FALLTHRU */
167 case 3:
168 u.c.c3 = lc_map[(unsigned char)s[i++]];
169 /* FALLTHRU */
170 case 2:
171 u.c.c2 = lc_map[(unsigned char)s[i++]];
172 /* FALLTHRU */
173 case 1:
174 u.c.c1 = lc_map[(unsigned char)s[i]];
175 r = mum_hash_step (r, u.pp);
176 break;
177 }
178
179 return mum_hash_finish (r);
180 }
181
182 static inline int
ucl_hash_caseless_equal(const ucl_object_t * k1,const ucl_object_t * k2)183 ucl_hash_caseless_equal (const ucl_object_t *k1, const ucl_object_t *k2)
184 {
185 if (k1->keylen == k2->keylen) {
186 unsigned fp, i;
187 const char *s = k1->key, *d = k2->key;
188 unsigned char c1, c2, c3, c4;
189 union {
190 unsigned char c[4];
191 uint32_t n;
192 } cmp1, cmp2;
193 size_t leftover = k1->keylen % 4;
194
195 fp = k1->keylen - leftover;
196
197 for (i = 0; i != fp; i += 4) {
198 c1 = s[i], c2 = s[i + 1], c3 = s[i + 2], c4 = s[i + 3];
199 cmp1.c[0] = lc_map[c1];
200 cmp1.c[1] = lc_map[c2];
201 cmp1.c[2] = lc_map[c3];
202 cmp1.c[3] = lc_map[c4];
203
204 c1 = d[i], c2 = d[i + 1], c3 = d[i + 2], c4 = d[i + 3];
205 cmp2.c[0] = lc_map[c1];
206 cmp2.c[1] = lc_map[c2];
207 cmp2.c[2] = lc_map[c3];
208 cmp2.c[3] = lc_map[c4];
209
210 if (cmp1.n != cmp2.n) {
211 return 0;
212 }
213 }
214
215 while (leftover > 0) {
216 if (lc_map[(unsigned char)s[i]] != lc_map[(unsigned char)d[i]]) {
217 return 0;
218 }
219
220 leftover--;
221 i++;
222 }
223
224 return 1;
225 }
226
227 return 0;
228 }
229
230 KHASH_INIT (ucl_hash_caseless_node, const ucl_object_t *, struct ucl_hash_elt, 1,
231 ucl_hash_caseless_func, ucl_hash_caseless_equal)
232
233 ucl_hash_t*
ucl_hash_create(bool ignore_case)234 ucl_hash_create (bool ignore_case)
235 {
236 ucl_hash_t *new;
237
238 new = UCL_ALLOC (sizeof (ucl_hash_t));
239 if (new != NULL) {
240 kv_init (new->ar);
241
242 new->caseless = ignore_case;
243 if (ignore_case) {
244 khash_t(ucl_hash_caseless_node) *h = kh_init (ucl_hash_caseless_node);
245 new->hash = (void *)h;
246 }
247 else {
248 khash_t(ucl_hash_node) *h = kh_init (ucl_hash_node);
249 new->hash = (void *)h;
250 }
251 }
252 return new;
253 }
254
ucl_hash_destroy(ucl_hash_t * hashlin,ucl_hash_free_func func)255 void ucl_hash_destroy (ucl_hash_t* hashlin, ucl_hash_free_func func)
256 {
257 const ucl_object_t *cur, *tmp;
258
259 if (hashlin == NULL) {
260 return;
261 }
262
263 if (func != NULL) {
264 /* Iterate over the hash first */
265 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
266 hashlin->hash;
267 khiter_t k;
268
269 for (k = kh_begin (h); k != kh_end (h); ++k) {
270 if (kh_exist (h, k)) {
271 cur = (kh_value (h, k)).obj;
272 while (cur != NULL) {
273 tmp = cur->next;
274 func (__DECONST (ucl_object_t *, cur));
275 cur = tmp;
276 }
277 }
278 }
279 }
280
281 if (hashlin->caseless) {
282 khash_t(ucl_hash_caseless_node) *h = (khash_t(ucl_hash_caseless_node) *)
283 hashlin->hash;
284 kh_destroy (ucl_hash_caseless_node, h);
285 }
286 else {
287 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
288 hashlin->hash;
289 kh_destroy (ucl_hash_node, h);
290 }
291
292 kv_destroy (hashlin->ar);
293 UCL_FREE (sizeof (*hashlin), hashlin);
294 }
295
296 void
ucl_hash_insert(ucl_hash_t * hashlin,const ucl_object_t * obj,const char * key,unsigned keylen)297 ucl_hash_insert (ucl_hash_t* hashlin, const ucl_object_t *obj,
298 const char *key, unsigned keylen)
299 {
300 khiter_t k;
301 int ret;
302 struct ucl_hash_elt *elt;
303
304 if (hashlin == NULL) {
305 return;
306 }
307
308 if (hashlin->caseless) {
309 khash_t(ucl_hash_caseless_node) *h = (khash_t(ucl_hash_caseless_node) *)
310 hashlin->hash;
311 k = kh_put (ucl_hash_caseless_node, h, obj, &ret);
312 if (ret > 0) {
313 elt = &kh_value (h, k);
314 kv_push (const ucl_object_t *, hashlin->ar, obj);
315 elt->obj = obj;
316 elt->ar_idx = kv_size (hashlin->ar) - 1;
317 }
318 }
319 else {
320 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
321 hashlin->hash;
322 k = kh_put (ucl_hash_node, h, obj, &ret);
323 if (ret > 0) {
324 elt = &kh_value (h, k);
325 kv_push (const ucl_object_t *, hashlin->ar, obj);
326 elt->obj = obj;
327 elt->ar_idx = kv_size (hashlin->ar) - 1;
328 }
329 }
330 }
331
ucl_hash_replace(ucl_hash_t * hashlin,const ucl_object_t * old,const ucl_object_t * new)332 void ucl_hash_replace (ucl_hash_t* hashlin, const ucl_object_t *old,
333 const ucl_object_t *new)
334 {
335 khiter_t k;
336 int ret;
337 struct ucl_hash_elt elt, *pelt;
338
339 if (hashlin == NULL) {
340 return;
341 }
342
343 if (hashlin->caseless) {
344 khash_t(ucl_hash_caseless_node) *h = (khash_t(ucl_hash_caseless_node) *)
345 hashlin->hash;
346 k = kh_put (ucl_hash_caseless_node, h, old, &ret);
347 if (ret == 0) {
348 elt = kh_value (h, k);
349 kh_del (ucl_hash_caseless_node, h, k);
350 k = kh_put (ucl_hash_caseless_node, h, new, &ret);
351 pelt = &kh_value (h, k);
352 pelt->obj = new;
353 pelt->ar_idx = elt.ar_idx;
354 kv_A (hashlin->ar, elt.ar_idx) = new;
355 }
356 }
357 else {
358 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
359 hashlin->hash;
360 k = kh_put (ucl_hash_node, h, old, &ret);
361 if (ret == 0) {
362 elt = kh_value (h, k);
363 kh_del (ucl_hash_node, h, k);
364 k = kh_put (ucl_hash_node, h, new, &ret);
365 pelt = &kh_value (h, k);
366 pelt->obj = new;
367 pelt->ar_idx = elt.ar_idx;
368 kv_A (hashlin->ar, elt.ar_idx) = new;
369 }
370 }
371 }
372
373 struct ucl_hash_real_iter {
374 const ucl_object_t **cur;
375 const ucl_object_t **end;
376 };
377
378 const void*
ucl_hash_iterate(ucl_hash_t * hashlin,ucl_hash_iter_t * iter)379 ucl_hash_iterate (ucl_hash_t *hashlin, ucl_hash_iter_t *iter)
380 {
381 struct ucl_hash_real_iter *it = (struct ucl_hash_real_iter *)(*iter);
382 const ucl_object_t *ret = NULL;
383
384 if (hashlin == NULL) {
385 return NULL;
386 }
387
388 if (it == NULL) {
389 it = UCL_ALLOC (sizeof (*it));
390
391 if (it == NULL) {
392 return NULL;
393 }
394
395 it->cur = &hashlin->ar.a[0];
396 it->end = it->cur + hashlin->ar.n;
397 }
398
399 if (it->cur < it->end) {
400 ret = *it->cur++;
401 }
402 else {
403 UCL_FREE (sizeof (*it), it);
404 *iter = NULL;
405 return NULL;
406 }
407
408 *iter = it;
409
410 return ret;
411 }
412
413 bool
ucl_hash_iter_has_next(ucl_hash_t * hashlin,ucl_hash_iter_t iter)414 ucl_hash_iter_has_next (ucl_hash_t *hashlin, ucl_hash_iter_t iter)
415 {
416 struct ucl_hash_real_iter *it = (struct ucl_hash_real_iter *)(iter);
417
418 return it->cur < it->end - 1;
419 }
420
421
422 const ucl_object_t*
ucl_hash_search(ucl_hash_t * hashlin,const char * key,unsigned keylen)423 ucl_hash_search (ucl_hash_t* hashlin, const char *key, unsigned keylen)
424 {
425 khiter_t k;
426 const ucl_object_t *ret = NULL;
427 ucl_object_t search;
428 struct ucl_hash_elt *elt;
429
430 search.key = key;
431 search.keylen = keylen;
432
433 if (hashlin == NULL) {
434 return NULL;
435 }
436
437 if (hashlin->caseless) {
438 khash_t(ucl_hash_caseless_node) *h = (khash_t(ucl_hash_caseless_node) *)
439 hashlin->hash;
440
441 k = kh_get (ucl_hash_caseless_node, h, &search);
442 if (k != kh_end (h)) {
443 elt = &kh_value (h, k);
444 ret = elt->obj;
445 }
446 }
447 else {
448 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
449 hashlin->hash;
450 k = kh_get (ucl_hash_node, h, &search);
451 if (k != kh_end (h)) {
452 elt = &kh_value (h, k);
453 ret = elt->obj;
454 }
455 }
456
457 return ret;
458 }
459
460 void
ucl_hash_delete(ucl_hash_t * hashlin,const ucl_object_t * obj)461 ucl_hash_delete (ucl_hash_t* hashlin, const ucl_object_t *obj)
462 {
463 khiter_t k;
464 struct ucl_hash_elt *elt;
465 size_t i;
466
467 if (hashlin == NULL) {
468 return;
469 }
470
471 if (hashlin->caseless) {
472 khash_t(ucl_hash_caseless_node) *h = (khash_t(ucl_hash_caseless_node) *)
473 hashlin->hash;
474
475 k = kh_get (ucl_hash_caseless_node, h, obj);
476 if (k != kh_end (h)) {
477 elt = &kh_value (h, k);
478 i = elt->ar_idx;
479 kv_del (const ucl_object_t *, hashlin->ar, elt->ar_idx);
480 kh_del (ucl_hash_caseless_node, h, k);
481
482 /* Update subsequent elts */
483 for (; i < hashlin->ar.n; i ++) {
484 elt = &kh_value (h, i);
485 elt->ar_idx --;
486 }
487 }
488 }
489 else {
490 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
491 hashlin->hash;
492 k = kh_get (ucl_hash_node, h, obj);
493 if (k != kh_end (h)) {
494 elt = &kh_value (h, k);
495 i = elt->ar_idx;
496 kv_del (const ucl_object_t *, hashlin->ar, elt->ar_idx);
497 kh_del (ucl_hash_node, h, k);
498
499 /* Update subsequent elts */
500 for (; i < hashlin->ar.n; i ++) {
501 elt = &kh_value (h, i);
502 elt->ar_idx --;
503 }
504 }
505 }
506 }
507
ucl_hash_reserve(ucl_hash_t * hashlin,size_t sz)508 void ucl_hash_reserve (ucl_hash_t *hashlin, size_t sz)
509 {
510 if (hashlin == NULL) {
511 return;
512 }
513
514 if (sz > hashlin->ar.m) {
515 kv_resize (const ucl_object_t *, hashlin->ar, sz);
516
517 if (hashlin->caseless) {
518 khash_t(ucl_hash_caseless_node) *h = (khash_t(
519 ucl_hash_caseless_node) *)
520 hashlin->hash;
521 kh_resize (ucl_hash_caseless_node, h, sz * 2);
522 } else {
523 khash_t(ucl_hash_node) *h = (khash_t(ucl_hash_node) *)
524 hashlin->hash;
525 kh_resize (ucl_hash_node, h, sz * 2);
526 }
527 }
528 }