Line | Count | Source (jump to first uncovered line) |
1 | | /* |
2 | | * dict.c: dictionary of reusable strings, just used to avoid allocation |
3 | | * and freeing operations. |
4 | | * |
5 | | * Copyright (C) 2003-2012 Daniel Veillard. |
6 | | * |
7 | | * Permission to use, copy, modify, and distribute this software for any |
8 | | * purpose with or without fee is hereby granted, provided that the above |
9 | | * copyright notice and this permission notice appear in all copies. |
10 | | * |
11 | | * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED |
12 | | * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF |
13 | | * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE AUTHORS AND |
14 | | * CONTRIBUTORS ACCEPT NO RESPONSIBILITY IN ANY CONCEIVABLE MANNER. |
15 | | * |
16 | | * Author: daniel@veillard.com |
17 | | */ |
18 | | |
19 | | #define IN_LIBXML |
20 | | #include "libxml.h" |
21 | | |
22 | | #include <errno.h> |
23 | | #include <limits.h> |
24 | | #include <stdlib.h> |
25 | | #include <string.h> |
26 | | #include <time.h> |
27 | | #ifdef HAVE_SYS_RANDOM_H |
28 | | #include <sys/random.h> |
29 | | #endif |
30 | | |
31 | | #ifdef _WIN32 |
32 | | #define WIN32_LEAN_AND_MEAN |
33 | | #include <windows.h> |
34 | | #include <bcrypt.h> |
35 | | #endif |
36 | | |
37 | | #include "private/dict.h" |
38 | | #include "private/globals.h" |
39 | | #include "private/threads.h" |
40 | | |
41 | | #include <libxml/parser.h> |
42 | | #include <libxml/dict.h> |
43 | | #include <libxml/xmlmemory.h> |
44 | | #include <libxml/xmlstring.h> |
45 | | |
46 | | #ifndef SIZE_MAX |
47 | 1.47M | #define SIZE_MAX ((size_t) -1) |
48 | | #endif |
49 | | |
50 | 849k | #define MAX_FILL_NUM 7 |
51 | 849k | #define MAX_FILL_DENOM 8 |
52 | 85.2k | #define MIN_HASH_SIZE 8 |
53 | 10.2M | #define MAX_HASH_SIZE (1u << 31) |
54 | | |
55 | | typedef struct _xmlDictStrings xmlDictStrings; |
56 | | typedef xmlDictStrings *xmlDictStringsPtr; |
57 | | struct _xmlDictStrings { |
58 | | xmlDictStringsPtr next; |
59 | | xmlChar *free; |
60 | | xmlChar *end; |
61 | | size_t size; |
62 | | size_t nbStrings; |
63 | | xmlChar array[1]; |
64 | | }; |
65 | | |
66 | | typedef xmlHashedString xmlDictEntry; |
67 | | |
68 | | /* |
69 | | * The entire dictionary |
70 | | */ |
71 | | struct _xmlDict { |
72 | | int ref_counter; |
73 | | |
74 | | xmlDictEntry *table; |
75 | | size_t size; |
76 | | unsigned int nbElems; |
77 | | xmlDictStringsPtr strings; |
78 | | |
79 | | struct _xmlDict *subdict; |
80 | | /* used for randomization */ |
81 | | unsigned seed; |
82 | | /* used to impose a limit on size */ |
83 | | size_t limit; |
84 | | }; |
85 | | |
86 | | /* |
87 | | * A mutex for modifying the reference counter for shared |
88 | | * dictionaries. |
89 | | */ |
90 | | static xmlMutex xmlDictMutex; |
91 | | |
92 | | /** |
93 | | * xmlInitializeDict: |
94 | | * |
95 | | * DEPRECATED: Alias for xmlInitParser. |
96 | | * |
97 | | * Returns 0. |
98 | | */ |
99 | | int |
100 | 0 | xmlInitializeDict(void) { |
101 | 0 | xmlInitParser(); |
102 | 0 | return(0); |
103 | 0 | } |
104 | | |
105 | | /** |
106 | | * xmlInitDictInternal: |
107 | | * |
108 | | * Initialize mutex. |
109 | | */ |
110 | | void |
111 | 2 | xmlInitDictInternal(void) { |
112 | 2 | xmlInitMutex(&xmlDictMutex); |
113 | 2 | } |
114 | | |
115 | | /** |
116 | | * xmlDictCleanup: |
117 | | * |
118 | | * DEPRECATED: This function is a no-op. Call xmlCleanupParser |
119 | | * to free global state but see the warnings there. xmlCleanupParser |
120 | | * should be only called once at program exit. In most cases, you don't |
121 | | * have call cleanup functions at all. |
122 | | */ |
123 | | void |
124 | 0 | xmlDictCleanup(void) { |
125 | 0 | } |
126 | | |
127 | | /** |
128 | | * xmlCleanupDictInternal: |
129 | | * |
130 | | * Free the dictionary mutex. |
131 | | */ |
132 | | void |
133 | 0 | xmlCleanupDictInternal(void) { |
134 | 0 | xmlCleanupMutex(&xmlDictMutex); |
135 | 0 | } |
136 | | |
137 | | /* |
138 | | * xmlDictAddString: |
139 | | * @dict: the dictionary |
140 | | * @name: the name of the userdata |
141 | | * @len: the length of the name |
142 | | * |
143 | | * Add the string to the array[s] |
144 | | * |
145 | | * Returns the pointer of the local string, or NULL in case of error. |
146 | | */ |
147 | | static const xmlChar * |
148 | 838k | xmlDictAddString(xmlDictPtr dict, const xmlChar *name, unsigned int namelen) { |
149 | 838k | xmlDictStringsPtr pool; |
150 | 838k | const xmlChar *ret; |
151 | 838k | size_t size = 0; /* + sizeof(_xmlDictStrings) == 1024 */ |
152 | 838k | size_t limit = 0; |
153 | | |
154 | 838k | pool = dict->strings; |
155 | 842k | while (pool != NULL) { |
156 | 753k | if ((size_t)(pool->end - pool->free) > namelen) |
157 | 749k | goto found_pool; |
158 | 4.34k | if (pool->size > size) size = pool->size; |
159 | 4.34k | limit += pool->size; |
160 | 4.34k | pool = pool->next; |
161 | 4.34k | } |
162 | | /* |
163 | | * Not found, need to allocate |
164 | | */ |
165 | 88.6k | if (pool == NULL) { |
166 | 88.6k | if ((dict->limit > 0) && (limit > dict->limit)) { |
167 | 4 | return(NULL); |
168 | 4 | } |
169 | | |
170 | 88.6k | if (size == 0) { |
171 | 85.3k | size = 1000; |
172 | 85.3k | } else { |
173 | 3.32k | if (size < (SIZE_MAX - sizeof(xmlDictStrings)) / 4) |
174 | 3.32k | size *= 4; /* exponential growth */ |
175 | 0 | else |
176 | 0 | size = SIZE_MAX - sizeof(xmlDictStrings); |
177 | 3.32k | } |
178 | 88.6k | if (size / 4 < namelen) { |
179 | 1.25k | if ((size_t) namelen + 0 < (SIZE_MAX - sizeof(xmlDictStrings)) / 4) |
180 | 1.25k | size = 4 * (size_t) namelen; /* just in case ! */ |
181 | 0 | else |
182 | 0 | return(NULL); |
183 | 1.25k | } |
184 | 88.6k | pool = (xmlDictStringsPtr) xmlMalloc(sizeof(xmlDictStrings) + size); |
185 | 88.6k | if (pool == NULL) |
186 | 401 | return(NULL); |
187 | 88.2k | pool->size = size; |
188 | 88.2k | pool->nbStrings = 0; |
189 | 88.2k | pool->free = &pool->array[0]; |
190 | 88.2k | pool->end = &pool->array[size]; |
191 | 88.2k | pool->next = dict->strings; |
192 | 88.2k | dict->strings = pool; |
193 | 88.2k | } |
194 | 837k | found_pool: |
195 | 837k | ret = pool->free; |
196 | 837k | memcpy(pool->free, name, namelen); |
197 | 837k | pool->free += namelen; |
198 | 837k | *(pool->free++) = 0; |
199 | 837k | pool->nbStrings++; |
200 | 837k | return(ret); |
201 | 88.6k | } |
202 | | |
203 | | /* |
204 | | * xmlDictAddQString: |
205 | | * @dict: the dictionary |
206 | | * @prefix: the prefix of the userdata |
207 | | * @plen: the prefix length |
208 | | * @name: the name of the userdata |
209 | | * @len: the length of the name |
210 | | * |
211 | | * Add the QName to the array[s] |
212 | | * |
213 | | * Returns the pointer of the local string, or NULL in case of error. |
214 | | */ |
215 | | static const xmlChar * |
216 | | xmlDictAddQString(xmlDictPtr dict, const xmlChar *prefix, unsigned int plen, |
217 | | const xmlChar *name, unsigned int namelen) |
218 | 10.6k | { |
219 | 10.6k | xmlDictStringsPtr pool; |
220 | 10.6k | const xmlChar *ret; |
221 | 10.6k | size_t size = 0; /* + sizeof(_xmlDictStrings) == 1024 */ |
222 | 10.6k | size_t limit = 0; |
223 | | |
224 | 10.6k | pool = dict->strings; |
225 | 10.8k | while (pool != NULL) { |
226 | 10.7k | if ((size_t)(pool->end - pool->free) > namelen + plen + 1) |
227 | 10.5k | goto found_pool; |
228 | 225 | if (pool->size > size) size = pool->size; |
229 | 225 | limit += pool->size; |
230 | 225 | pool = pool->next; |
231 | 225 | } |
232 | | /* |
233 | | * Not found, need to allocate |
234 | | */ |
235 | 97 | if (pool == NULL) { |
236 | 97 | if ((dict->limit > 0) && (limit > dict->limit)) { |
237 | 0 | return(NULL); |
238 | 0 | } |
239 | | |
240 | 97 | if (size == 0) size = 1000; |
241 | 97 | else size *= 4; /* exponential growth */ |
242 | 97 | if (size < 4 * (namelen + plen + 1)) |
243 | 0 | size = 4 * (namelen + plen + 1); /* just in case ! */ |
244 | 97 | pool = (xmlDictStringsPtr) xmlMalloc(sizeof(xmlDictStrings) + size); |
245 | 97 | if (pool == NULL) |
246 | 6 | return(NULL); |
247 | 91 | pool->size = size; |
248 | 91 | pool->nbStrings = 0; |
249 | 91 | pool->free = &pool->array[0]; |
250 | 91 | pool->end = &pool->array[size]; |
251 | 91 | pool->next = dict->strings; |
252 | 91 | dict->strings = pool; |
253 | 91 | } |
254 | 10.6k | found_pool: |
255 | 10.6k | ret = pool->free; |
256 | 10.6k | memcpy(pool->free, prefix, plen); |
257 | 10.6k | pool->free += plen; |
258 | 10.6k | *(pool->free++) = ':'; |
259 | 10.6k | memcpy(pool->free, name, namelen); |
260 | 10.6k | pool->free += namelen; |
261 | 10.6k | *(pool->free++) = 0; |
262 | 10.6k | pool->nbStrings++; |
263 | 10.6k | return(ret); |
264 | 97 | } |
265 | | |
266 | | /** |
267 | | * xmlDictCreate: |
268 | | * |
269 | | * Create a new dictionary |
270 | | * |
271 | | * Returns the newly created dictionary, or NULL if an error occurred. |
272 | | */ |
273 | | xmlDictPtr |
274 | 86.4k | xmlDictCreate(void) { |
275 | 86.4k | xmlDictPtr dict; |
276 | | |
277 | 86.4k | xmlInitParser(); |
278 | | |
279 | 86.4k | dict = xmlMalloc(sizeof(xmlDict)); |
280 | 86.4k | if (dict == NULL) |
281 | 14 | return(NULL); |
282 | 86.4k | dict->ref_counter = 1; |
283 | 86.4k | dict->limit = 0; |
284 | | |
285 | 86.4k | dict->size = 0; |
286 | 86.4k | dict->nbElems = 0; |
287 | 86.4k | dict->table = NULL; |
288 | 86.4k | dict->strings = NULL; |
289 | 86.4k | dict->subdict = NULL; |
290 | 86.4k | dict->seed = xmlRandom(); |
291 | 86.4k | #ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION |
292 | 86.4k | dict->seed = 0; |
293 | 86.4k | #endif |
294 | 86.4k | return(dict); |
295 | 86.4k | } |
296 | | |
297 | | /** |
298 | | * xmlDictCreateSub: |
299 | | * @sub: an existing dictionary |
300 | | * |
301 | | * Create a new dictionary, inheriting strings from the read-only |
302 | | * dictionary @sub. On lookup, strings are first searched in the |
303 | | * new dictionary, then in @sub, and if not found are created in the |
304 | | * new dictionary. |
305 | | * |
306 | | * Returns the newly created dictionary, or NULL if an error occurred. |
307 | | */ |
308 | | xmlDictPtr |
309 | 0 | xmlDictCreateSub(xmlDictPtr sub) { |
310 | 0 | xmlDictPtr dict = xmlDictCreate(); |
311 | |
|
312 | 0 | if ((dict != NULL) && (sub != NULL)) { |
313 | 0 | dict->seed = sub->seed; |
314 | 0 | dict->subdict = sub; |
315 | 0 | xmlDictReference(dict->subdict); |
316 | 0 | } |
317 | 0 | return(dict); |
318 | 0 | } |
319 | | |
320 | | /** |
321 | | * xmlDictReference: |
322 | | * @dict: the dictionary |
323 | | * |
324 | | * Increment the reference counter of a dictionary |
325 | | * |
326 | | * Returns 0 in case of success and -1 in case of error |
327 | | */ |
328 | | int |
329 | 132k | xmlDictReference(xmlDictPtr dict) { |
330 | 132k | if (dict == NULL) return -1; |
331 | 110k | xmlMutexLock(&xmlDictMutex); |
332 | 110k | dict->ref_counter++; |
333 | 110k | xmlMutexUnlock(&xmlDictMutex); |
334 | 110k | return(0); |
335 | 132k | } |
336 | | |
337 | | /** |
338 | | * xmlDictFree: |
339 | | * @dict: the dictionary |
340 | | * |
341 | | * Free the hash @dict and its contents. The userdata is |
342 | | * deallocated with @f if provided. |
343 | | */ |
344 | | void |
345 | 197k | xmlDictFree(xmlDictPtr dict) { |
346 | 197k | xmlDictStringsPtr pool, nextp; |
347 | | |
348 | 197k | if (dict == NULL) |
349 | 0 | return; |
350 | | |
351 | | /* decrement the counter, it may be shared by a parser and docs */ |
352 | 197k | xmlMutexLock(&xmlDictMutex); |
353 | 197k | dict->ref_counter--; |
354 | 197k | if (dict->ref_counter > 0) { |
355 | 110k | xmlMutexUnlock(&xmlDictMutex); |
356 | 110k | return; |
357 | 110k | } |
358 | | |
359 | 86.4k | xmlMutexUnlock(&xmlDictMutex); |
360 | | |
361 | 86.4k | if (dict->subdict != NULL) { |
362 | 0 | xmlDictFree(dict->subdict); |
363 | 0 | } |
364 | | |
365 | 86.4k | if (dict->table) { |
366 | 84.9k | xmlFree(dict->table); |
367 | 84.9k | } |
368 | 86.4k | pool = dict->strings; |
369 | 174k | while (pool != NULL) { |
370 | 88.3k | nextp = pool->next; |
371 | 88.3k | xmlFree(pool); |
372 | 88.3k | pool = nextp; |
373 | 88.3k | } |
374 | 86.4k | xmlFree(dict); |
375 | 86.4k | } |
376 | | |
377 | | /** |
378 | | * xmlDictOwns: |
379 | | * @dict: the dictionary |
380 | | * @str: the string |
381 | | * |
382 | | * check if a string is owned by the dictionary |
383 | | * |
384 | | * Returns 1 if true, 0 if false and -1 in case of error |
385 | | * -1 in case of error |
386 | | */ |
387 | | int |
388 | 8.44M | xmlDictOwns(xmlDictPtr dict, const xmlChar *str) { |
389 | 8.44M | xmlDictStringsPtr pool; |
390 | | |
391 | 8.44M | if ((dict == NULL) || (str == NULL)) |
392 | 0 | return(-1); |
393 | 8.44M | pool = dict->strings; |
394 | 15.7M | while (pool != NULL) { |
395 | 10.9M | if ((str >= &pool->array[0]) && (str <= pool->free)) |
396 | 3.69M | return(1); |
397 | 7.27M | pool = pool->next; |
398 | 7.27M | } |
399 | 4.75M | if (dict->subdict) |
400 | 0 | return(xmlDictOwns(dict->subdict, str)); |
401 | 4.75M | return(0); |
402 | 4.75M | } |
403 | | |
404 | | /** |
405 | | * xmlDictSize: |
406 | | * @dict: the dictionary |
407 | | * |
408 | | * Query the number of elements installed in the hash @dict. |
409 | | * |
410 | | * Returns the number of elements in the dictionary or |
411 | | * -1 in case of error |
412 | | */ |
413 | | int |
414 | 0 | xmlDictSize(xmlDictPtr dict) { |
415 | 0 | if (dict == NULL) |
416 | 0 | return(-1); |
417 | 0 | if (dict->subdict) |
418 | 0 | return(dict->nbElems + dict->subdict->nbElems); |
419 | 0 | return(dict->nbElems); |
420 | 0 | } |
421 | | |
422 | | /** |
423 | | * xmlDictSetLimit: |
424 | | * @dict: the dictionary |
425 | | * @limit: the limit in bytes |
426 | | * |
427 | | * Set a size limit for the dictionary |
428 | | * Added in 2.9.0 |
429 | | * |
430 | | * Returns the previous limit of the dictionary or 0 |
431 | | */ |
432 | | size_t |
433 | 113k | xmlDictSetLimit(xmlDictPtr dict, size_t limit) { |
434 | 113k | size_t ret; |
435 | | |
436 | 113k | if (dict == NULL) |
437 | 0 | return(0); |
438 | 113k | ret = dict->limit; |
439 | 113k | dict->limit = limit; |
440 | 113k | return(ret); |
441 | 113k | } |
442 | | |
443 | | /** |
444 | | * xmlDictGetUsage: |
445 | | * @dict: the dictionary |
446 | | * |
447 | | * Get how much memory is used by a dictionary for strings |
448 | | * Added in 2.9.0 |
449 | | * |
450 | | * Returns the amount of strings allocated |
451 | | */ |
452 | | size_t |
453 | 0 | xmlDictGetUsage(xmlDictPtr dict) { |
454 | 0 | xmlDictStringsPtr pool; |
455 | 0 | size_t limit = 0; |
456 | |
|
457 | 0 | if (dict == NULL) |
458 | 0 | return(0); |
459 | 0 | pool = dict->strings; |
460 | 0 | while (pool != NULL) { |
461 | 0 | limit += pool->size; |
462 | 0 | pool = pool->next; |
463 | 0 | } |
464 | 0 | return(limit); |
465 | 0 | } |
466 | | |
467 | | /***************************************************************** |
468 | | * |
469 | | * The code below was rewritten and is additionally licensed under |
470 | | * the main license in file 'Copyright'. |
471 | | * |
472 | | *****************************************************************/ |
473 | | |
474 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
475 | | static unsigned |
476 | | xmlDictHashName(unsigned seed, const xmlChar* data, size_t maxLen, |
477 | 10.0M | size_t *plen) { |
478 | 10.0M | unsigned h1, h2; |
479 | 10.0M | size_t i; |
480 | | |
481 | 10.0M | HASH_INIT(h1, h2, seed); |
482 | | |
483 | 634M | for (i = 0; i < maxLen && data[i]; i++) { |
484 | 624M | HASH_UPDATE(h1, h2, data[i]); |
485 | 624M | } |
486 | | |
487 | 10.0M | HASH_FINISH(h1, h2); |
488 | | |
489 | 10.0M | *plen = i; |
490 | 10.0M | return(h2 | MAX_HASH_SIZE); |
491 | 10.0M | } |
492 | | |
493 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
494 | | static unsigned |
495 | | xmlDictHashQName(unsigned seed, const xmlChar *prefix, const xmlChar *name, |
496 | 90.9k | size_t *pplen, size_t *plen) { |
497 | 90.9k | unsigned h1, h2; |
498 | 90.9k | size_t i; |
499 | | |
500 | 90.9k | HASH_INIT(h1, h2, seed); |
501 | | |
502 | 933k | for (i = 0; prefix[i] != 0; i++) { |
503 | 842k | HASH_UPDATE(h1, h2, prefix[i]); |
504 | 842k | } |
505 | 90.9k | *pplen = i; |
506 | | |
507 | 90.9k | HASH_UPDATE(h1, h2, ':'); |
508 | | |
509 | 1.13M | for (i = 0; name[i] != 0; i++) { |
510 | 1.04M | HASH_UPDATE(h1, h2, name[i]); |
511 | 1.04M | } |
512 | 90.9k | *plen = i; |
513 | | |
514 | 90.9k | HASH_FINISH(h1, h2); |
515 | | |
516 | | /* |
517 | | * Always set the upper bit of hash values since 0 means an unoccupied |
518 | | * bucket. |
519 | | */ |
520 | 90.9k | return(h2 | MAX_HASH_SIZE); |
521 | 90.9k | } |
522 | | |
523 | | unsigned |
524 | 185k | xmlDictComputeHash(const xmlDict *dict, const xmlChar *string) { |
525 | 185k | size_t len; |
526 | 185k | return(xmlDictHashName(dict->seed, string, SIZE_MAX, &len)); |
527 | 185k | } |
528 | | |
529 | 1.09M | #define HASH_ROL31(x,n) ((x) << (n) | ((x) & 0x7FFFFFFF) >> (31 - (n))) |
530 | | |
531 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
532 | | unsigned |
533 | 1.09M | xmlDictCombineHash(unsigned v1, unsigned v2) { |
534 | | /* |
535 | | * The upper bit of hash values is always set, so we have to operate on |
536 | | * 31-bit hashes here. |
537 | | */ |
538 | 1.09M | v1 ^= v2; |
539 | 1.09M | v1 += HASH_ROL31(v2, 5); |
540 | | |
541 | 1.09M | return((v1 & 0xFFFFFFFF) | 0x80000000); |
542 | 1.09M | } |
543 | | |
544 | | /** |
545 | | * xmlDictFindEntry: |
546 | | * @dict: dict |
547 | | * @prefix: optional QName prefix |
548 | | * @name: string |
549 | | * @len: length of string |
550 | | * @hashValue: valid hash value of string |
551 | | * @pfound: result of search |
552 | | * |
553 | | * Try to find a matching hash table entry. If an entry was found, set |
554 | | * @found to 1 and return the entry. Otherwise, set @found to 0 and return |
555 | | * the location where a new entry should be inserted. |
556 | | */ |
557 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
558 | | static xmlDictEntry * |
559 | | xmlDictFindEntry(const xmlDict *dict, const xmlChar *prefix, |
560 | | const xmlChar *name, int len, unsigned hashValue, |
561 | 9.89M | int *pfound) { |
562 | 9.89M | xmlDictEntry *entry; |
563 | 9.89M | unsigned mask, pos, displ; |
564 | 9.89M | int found = 0; |
565 | | |
566 | 9.89M | mask = dict->size - 1; |
567 | 9.89M | pos = hashValue & mask; |
568 | 9.89M | entry = &dict->table[pos]; |
569 | | |
570 | 9.89M | if (entry->hashValue != 0) { |
571 | | /* |
572 | | * Robin hood hashing: abort if the displacement of the entry |
573 | | * is smaller than the displacement of the key we look for. |
574 | | * This also stops at the correct position when inserting. |
575 | | */ |
576 | 9.47M | displ = 0; |
577 | | |
578 | 16.5M | do { |
579 | 16.5M | if (entry->hashValue == hashValue) { |
580 | 9.12M | if (prefix == NULL) { |
581 | | /* |
582 | | * name is not necessarily null-terminated. |
583 | | */ |
584 | 9.04M | if ((strncmp((const char *) entry->name, |
585 | 9.04M | (const char *) name, len) == 0) && |
586 | 9.04M | (entry->name[len] == 0)) { |
587 | 9.04M | found = 1; |
588 | 9.04M | break; |
589 | 9.04M | } |
590 | 9.04M | } else { |
591 | 80.3k | if (xmlStrQEqual(prefix, name, entry->name)) { |
592 | 80.3k | found = 1; |
593 | 80.3k | break; |
594 | 80.3k | } |
595 | 80.3k | } |
596 | 9.12M | } |
597 | | |
598 | 7.42M | displ++; |
599 | 7.42M | pos++; |
600 | 7.42M | entry++; |
601 | 7.42M | if ((pos & mask) == 0) |
602 | 374k | entry = dict->table; |
603 | 7.42M | } while ((entry->hashValue != 0) && |
604 | 7.42M | (((pos - entry->hashValue) & mask) >= displ)); |
605 | 9.47M | } |
606 | | |
607 | 0 | *pfound = found; |
608 | 9.89M | return(entry); |
609 | 9.89M | } |
610 | | |
611 | | /** |
612 | | * xmlDictGrow: |
613 | | * @dict: dictionary |
614 | | * @size: new size of the dictionary |
615 | | * |
616 | | * Resize the dictionary hash table. |
617 | | * |
618 | | * Returns 0 in case of success, -1 if a memory allocation failed. |
619 | | */ |
620 | | static int |
621 | 131k | xmlDictGrow(xmlDictPtr dict, unsigned size) { |
622 | 131k | const xmlDictEntry *oldentry, *oldend, *end; |
623 | 131k | xmlDictEntry *table; |
624 | 131k | unsigned oldsize, i; |
625 | | |
626 | | /* Add 0 to avoid spurious -Wtype-limits warning on 64-bit GCC */ |
627 | 131k | if ((size_t) size + 0 > SIZE_MAX / sizeof(table[0])) |
628 | 0 | return(-1); |
629 | 131k | table = xmlMalloc(size * sizeof(table[0])); |
630 | 131k | if (table == NULL) |
631 | 453 | return(-1); |
632 | 131k | memset(table, 0, size * sizeof(table[0])); |
633 | | |
634 | 131k | oldsize = dict->size; |
635 | 131k | if (oldsize == 0) |
636 | 84.9k | goto done; |
637 | | |
638 | 46.1k | oldend = &dict->table[oldsize]; |
639 | 46.1k | end = &table[size]; |
640 | | |
641 | | /* |
642 | | * Robin Hood sorting order is maintained if we |
643 | | * |
644 | | * - compute dict indices with modulo |
645 | | * - resize by an integer factor |
646 | | * - start to copy from the beginning of a probe sequence |
647 | | */ |
648 | 46.1k | oldentry = dict->table; |
649 | 305k | while (oldentry->hashValue != 0) { |
650 | 259k | if (++oldentry >= oldend) |
651 | 0 | oldentry = dict->table; |
652 | 259k | } |
653 | | |
654 | 767k | for (i = 0; i < oldsize; i++) { |
655 | 720k | if (oldentry->hashValue != 0) { |
656 | 630k | xmlDictEntry *entry = &table[oldentry->hashValue & (size - 1)]; |
657 | | |
658 | 828k | while (entry->hashValue != 0) { |
659 | 198k | if (++entry >= end) |
660 | 15.3k | entry = table; |
661 | 198k | } |
662 | 630k | *entry = *oldentry; |
663 | 630k | } |
664 | | |
665 | 720k | if (++oldentry >= oldend) |
666 | 46.1k | oldentry = dict->table; |
667 | 720k | } |
668 | | |
669 | 46.1k | xmlFree(dict->table); |
670 | | |
671 | 131k | done: |
672 | 131k | dict->table = table; |
673 | 131k | dict->size = size; |
674 | | |
675 | 131k | return(0); |
676 | 46.1k | } |
677 | | |
678 | | /** |
679 | | * xmlDictLookupInternal: |
680 | | * @dict: dict |
681 | | * @prefix: optional QName prefix |
682 | | * @name: string |
683 | | * @maybeLen: length of string or -1 if unknown |
684 | | * @update: whether the string should be added |
685 | | * |
686 | | * Internal lookup and update function. |
687 | | */ |
688 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
689 | | static const xmlDictEntry * |
690 | | xmlDictLookupInternal(xmlDictPtr dict, const xmlChar *prefix, |
691 | 9.97M | const xmlChar *name, int maybeLen, int update) { |
692 | 9.97M | xmlDictEntry *entry = NULL; |
693 | 9.97M | const xmlChar *ret; |
694 | 9.97M | unsigned hashValue; |
695 | 9.97M | size_t maxLen, len, plen, klen; |
696 | 9.97M | int found = 0; |
697 | | |
698 | 9.97M | if ((dict == NULL) || (name == NULL)) |
699 | 0 | return(NULL); |
700 | | |
701 | 9.97M | maxLen = (maybeLen < 0) ? SIZE_MAX : (size_t) maybeLen; |
702 | | |
703 | 9.97M | if (prefix == NULL) { |
704 | 9.88M | hashValue = xmlDictHashName(dict->seed, name, maxLen, &len); |
705 | 9.88M | if (len > INT_MAX / 2) |
706 | 0 | return(NULL); |
707 | 9.88M | klen = len; |
708 | 9.88M | } else { |
709 | 90.9k | hashValue = xmlDictHashQName(dict->seed, prefix, name, &plen, &len); |
710 | 90.9k | if ((len > INT_MAX / 2) || (plen >= INT_MAX / 2 - len)) |
711 | 0 | return(NULL); |
712 | 90.9k | klen = plen + 1 + len; |
713 | 90.9k | } |
714 | | |
715 | 9.97M | if ((dict->limit > 0) && (klen >= dict->limit)) |
716 | 0 | return(NULL); |
717 | | |
718 | | /* |
719 | | * Check for an existing entry |
720 | | */ |
721 | 9.97M | if (dict->size > 0) |
722 | 9.89M | entry = xmlDictFindEntry(dict, prefix, name, klen, hashValue, &found); |
723 | 9.97M | if (found) |
724 | 9.12M | return(entry); |
725 | | |
726 | 849k | if ((dict->subdict != NULL) && (dict->subdict->size > 0)) { |
727 | 0 | xmlDictEntry *subEntry; |
728 | 0 | unsigned subHashValue; |
729 | |
|
730 | 0 | if (prefix == NULL) |
731 | 0 | subHashValue = xmlDictHashName(dict->subdict->seed, name, len, |
732 | 0 | &len); |
733 | 0 | else |
734 | 0 | subHashValue = xmlDictHashQName(dict->subdict->seed, prefix, name, |
735 | 0 | &plen, &len); |
736 | 0 | subEntry = xmlDictFindEntry(dict->subdict, prefix, name, klen, |
737 | 0 | subHashValue, &found); |
738 | 0 | if (found) |
739 | 0 | return(subEntry); |
740 | 0 | } |
741 | | |
742 | 849k | if (!update) |
743 | 0 | return(NULL); |
744 | | |
745 | | /* |
746 | | * Grow the hash table if needed |
747 | | */ |
748 | 849k | if (dict->nbElems + 1 > dict->size / MAX_FILL_DENOM * MAX_FILL_NUM) { |
749 | 131k | unsigned newSize, mask, displ, pos; |
750 | | |
751 | 131k | if (dict->size == 0) { |
752 | 85.2k | newSize = MIN_HASH_SIZE; |
753 | 85.2k | } else { |
754 | 46.3k | if (dict->size >= MAX_HASH_SIZE) |
755 | 0 | return(NULL); |
756 | 46.3k | newSize = dict->size * 2; |
757 | 46.3k | } |
758 | 131k | if (xmlDictGrow(dict, newSize) != 0) |
759 | 453 | return(NULL); |
760 | | |
761 | | /* |
762 | | * Find new entry |
763 | | */ |
764 | 131k | mask = dict->size - 1; |
765 | 131k | displ = 0; |
766 | 131k | pos = hashValue & mask; |
767 | 131k | entry = &dict->table[pos]; |
768 | | |
769 | 154k | while ((entry->hashValue != 0) && |
770 | 154k | ((pos - entry->hashValue) & mask) >= displ) { |
771 | 23.3k | displ++; |
772 | 23.3k | pos++; |
773 | 23.3k | entry++; |
774 | 23.3k | if ((pos & mask) == 0) |
775 | 2.18k | entry = dict->table; |
776 | 23.3k | } |
777 | 131k | } |
778 | | |
779 | 848k | if (prefix == NULL) |
780 | 838k | ret = xmlDictAddString(dict, name, len); |
781 | 10.6k | else |
782 | 10.6k | ret = xmlDictAddQString(dict, prefix, plen, name, len); |
783 | 848k | if (ret == NULL) |
784 | 411 | return(NULL); |
785 | | |
786 | | /* |
787 | | * Shift the remainder of the probe sequence to the right |
788 | | */ |
789 | 848k | if (entry->hashValue != 0) { |
790 | 167k | const xmlDictEntry *end = &dict->table[dict->size]; |
791 | 167k | const xmlDictEntry *cur = entry; |
792 | | |
793 | 716k | do { |
794 | 716k | cur++; |
795 | 716k | if (cur >= end) |
796 | 27.5k | cur = dict->table; |
797 | 716k | } while (cur->hashValue != 0); |
798 | | |
799 | 167k | if (cur < entry) { |
800 | | /* |
801 | | * If we traversed the end of the buffer, handle the part |
802 | | * at the start of the buffer. |
803 | | */ |
804 | 27.5k | memmove(&dict->table[1], dict->table, |
805 | 27.5k | (char *) cur - (char *) dict->table); |
806 | 27.5k | cur = end - 1; |
807 | 27.5k | dict->table[0] = *cur; |
808 | 27.5k | } |
809 | | |
810 | 167k | memmove(&entry[1], entry, (char *) cur - (char *) entry); |
811 | 167k | } |
812 | | |
813 | | /* |
814 | | * Populate entry |
815 | | */ |
816 | 848k | entry->hashValue = hashValue; |
817 | 848k | entry->name = ret; |
818 | | |
819 | 848k | dict->nbElems++; |
820 | | |
821 | 848k | return(entry); |
822 | 848k | } |
823 | | |
824 | | /** |
825 | | * xmlDictLookup: |
826 | | * @dict: dictionary |
827 | | * @name: string key |
828 | | * @len: length of the key, if -1 it is recomputed |
829 | | * |
830 | | * Lookup a string and add it to the dictionary if it wasn't found. |
831 | | * |
832 | | * Returns the interned copy of the string or NULL if a memory allocation |
833 | | * failed. |
834 | | */ |
835 | | const xmlChar * |
836 | 6.96M | xmlDictLookup(xmlDictPtr dict, const xmlChar *name, int len) { |
837 | 6.96M | const xmlDictEntry *entry; |
838 | | |
839 | 6.96M | entry = xmlDictLookupInternal(dict, NULL, name, len, 1); |
840 | 6.96M | if (entry == NULL) |
841 | 751 | return(NULL); |
842 | 6.96M | return(entry->name); |
843 | 6.96M | } |
844 | | |
845 | | /** |
846 | | * xmlDictLookupHashed: |
847 | | * @dict: dictionary |
848 | | * @name: string key |
849 | | * @len: length of the key, if -1 it is recomputed |
850 | | * |
851 | | * Lookup a dictionary entry and add the string to the dictionary if |
852 | | * it wasn't found. |
853 | | * |
854 | | * Returns the dictionary entry. |
855 | | */ |
856 | | xmlHashedString |
857 | 2.92M | xmlDictLookupHashed(xmlDictPtr dict, const xmlChar *name, int len) { |
858 | 2.92M | const xmlDictEntry *entry; |
859 | 2.92M | xmlHashedString ret; |
860 | | |
861 | 2.92M | entry = xmlDictLookupInternal(dict, NULL, name, len, 1); |
862 | | |
863 | 2.92M | if (entry == NULL) { |
864 | 93 | ret.name = NULL; |
865 | 93 | ret.hashValue = 0; |
866 | 2.92M | } else { |
867 | 2.92M | ret = *entry; |
868 | 2.92M | } |
869 | | |
870 | 2.92M | return(ret); |
871 | 2.92M | } |
872 | | |
873 | | /** |
874 | | * xmlDictExists: |
875 | | * @dict: the dictionary |
876 | | * @name: the name of the userdata |
877 | | * @len: the length of the name, if -1 it is recomputed |
878 | | * |
879 | | * Check if a string exists in the dictionary. |
880 | | * |
881 | | * Returns the internal copy of the name or NULL if not found. |
882 | | */ |
883 | | const xmlChar * |
884 | 0 | xmlDictExists(xmlDictPtr dict, const xmlChar *name, int len) { |
885 | 0 | const xmlDictEntry *entry; |
886 | |
|
887 | 0 | entry = xmlDictLookupInternal(dict, NULL, name, len, 0); |
888 | 0 | if (entry == NULL) |
889 | 0 | return(NULL); |
890 | 0 | return(entry->name); |
891 | 0 | } |
892 | | |
893 | | /** |
894 | | * xmlDictQLookup: |
895 | | * @dict: the dictionary |
896 | | * @prefix: the prefix |
897 | | * @name: the name |
898 | | * |
899 | | * Lookup the QName @prefix:@name and add it to the dictionary if |
900 | | * it wasn't found. |
901 | | * |
902 | | * Returns the interned copy of the string or NULL if a memory allocation |
903 | | * failed. |
904 | | */ |
905 | | const xmlChar * |
906 | 90.9k | xmlDictQLookup(xmlDictPtr dict, const xmlChar *prefix, const xmlChar *name) { |
907 | 90.9k | const xmlDictEntry *entry; |
908 | | |
909 | 90.9k | entry = xmlDictLookupInternal(dict, prefix, name, -1, 1); |
910 | 90.9k | if (entry == NULL) |
911 | 20 | return(NULL); |
912 | 90.9k | return(entry->name); |
913 | 90.9k | } |
914 | | |
915 | | /* |
916 | | * Pseudo-random generator |
917 | | */ |
918 | | |
919 | | static xmlMutex xmlRngMutex; |
920 | | |
921 | | static unsigned globalRngState[2]; |
922 | | |
923 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
924 | | void |
925 | 2 | xmlInitRandom(void) { |
926 | 2 | xmlInitMutex(&xmlRngMutex); |
927 | | |
928 | 2 | { |
929 | | #ifdef _WIN32 |
930 | | NTSTATUS status; |
931 | | |
932 | | status = BCryptGenRandom(NULL, (unsigned char *) globalRngState, |
933 | | sizeof(globalRngState), |
934 | | BCRYPT_USE_SYSTEM_PREFERRED_RNG); |
935 | | if (!BCRYPT_SUCCESS(status)) { |
936 | | fprintf(stderr, "libxml2: BCryptGenRandom failed with " |
937 | | "error code %lu\n", GetLastError()); |
938 | | abort(); |
939 | | } |
940 | | #elif defined(HAVE_GETENTROPY) |
941 | 2 | while (1) { |
942 | 2 | if (getentropy(globalRngState, sizeof(globalRngState)) == 0) |
943 | 2 | break; |
944 | | |
945 | 0 | if (errno != EINTR) { |
946 | 0 | fprintf(stderr, "libxml2: getentropy failed with " |
947 | 0 | "error code %d\n", errno); |
948 | 0 | abort(); |
949 | 0 | } |
950 | 0 | } |
951 | | #else |
952 | | int var; |
953 | | |
954 | | globalRngState[0] = |
955 | | (unsigned) time(NULL) ^ |
956 | | HASH_ROL((unsigned) ((size_t) &xmlInitRandom & 0xFFFFFFFF), 8); |
957 | | globalRngState[1] = |
958 | | HASH_ROL((unsigned) ((size_t) &xmlRngMutex & 0xFFFFFFFF), 16) ^ |
959 | | HASH_ROL((unsigned) ((size_t) &var & 0xFFFFFFFF), 24); |
960 | | #endif |
961 | 2 | } |
962 | 2 | } |
963 | | |
964 | | void |
965 | 0 | xmlCleanupRandom(void) { |
966 | 0 | xmlCleanupMutex(&xmlRngMutex); |
967 | 0 | } |
968 | | |
969 | | ATTRIBUTE_NO_SANITIZE_INTEGER |
970 | | static unsigned |
971 | 189k | xoroshiro64ss(unsigned *s) { |
972 | 189k | unsigned s0 = s[0]; |
973 | 189k | unsigned s1 = s[1]; |
974 | 189k | unsigned result = HASH_ROL(s0 * 0x9E3779BB, 5) * 5; |
975 | | |
976 | 189k | s1 ^= s0; |
977 | 189k | s[0] = HASH_ROL(s0, 26) ^ s1 ^ (s1 << 9); |
978 | 189k | s[1] = HASH_ROL(s1, 13); |
979 | | |
980 | 189k | return(result & 0xFFFFFFFF); |
981 | 189k | } |
982 | | |
983 | | unsigned |
984 | 4 | xmlGlobalRandom(void) { |
985 | 4 | unsigned ret; |
986 | | |
987 | 4 | xmlMutexLock(&xmlRngMutex); |
988 | 4 | ret = xoroshiro64ss(globalRngState); |
989 | 4 | xmlMutexUnlock(&xmlRngMutex); |
990 | | |
991 | 4 | return(ret); |
992 | 4 | } |
993 | | |
994 | | unsigned |
995 | 189k | xmlRandom(void) { |
996 | 189k | #ifdef LIBXML_THREAD_ENABLED |
997 | 189k | return(xoroshiro64ss(xmlGetLocalRngState())); |
998 | | #else |
999 | | return(xmlGlobalRandom()); |
1000 | | #endif |
1001 | 189k | } |
1002 | | |