2 * Copyright (C) 2012, 2013
6 * Permission is hereby granted, free of charge, to any person obtaining a copy of
7 * this software and associated documentation files (the "Software"), to deal in
8 * the Software without restriction, including without limitation the rights to
9 * use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
10 * of the Software, and to permit persons to whom the Software is furnished to do
11 * so, subject to the following conditions:
13 * The above copyright notice and this permission notice shall be included in all
14 * copies or substantial portions of the Software.
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
28 /* TODO: remove globals ... */
29 static uint64_t mem_ab = 0;
30 static uint64_t mem_db = 0;
31 static uint64_t mem_at = 0;
32 static uint64_t mem_dt = 0;
33 static uint64_t mem_pk = 0;
34 static uint64_t mem_hw = 0;
40 struct memblock_t *next;
41 struct memblock_t *prev;
46 if (mem_hw > mem_pk) \
50 static struct memblock_t *mem_start = NULL;
52 void *util_memory_a(size_t byte, unsigned int line, const char *file) {
53 struct memblock_t *info = (struct memblock_t*)malloc(sizeof(struct memblock_t) + byte);
54 void *data = (void*)(info+1);
55 if (!info) return NULL;
60 info->next = mem_start;
62 mem_start->prev = info;
74 void util_memory_d(void *ptrn) {
75 struct memblock_t *info = NULL;
78 info = ((struct memblock_t*)ptrn - 1);
85 info->prev->next = info->next;
87 info->next->prev = info->prev;
88 if (info == mem_start)
89 mem_start = info->next;
94 void *util_memory_r(void *ptrn, size_t byte, unsigned int line, const char *file) {
95 struct memblock_t *oldinfo = NULL;
97 struct memblock_t *newinfo;
100 return util_memory_a(byte, line, file);
106 oldinfo = ((struct memblock_t*)ptrn - 1);
107 newinfo = ((struct memblock_t*)malloc(sizeof(struct memblock_t) + byte));
111 util_memory_d(oldinfo+1);
116 memcpy(newinfo+1, oldinfo+1, oldinfo->byte);
120 oldinfo->prev->next = oldinfo->next;
122 oldinfo->next->prev = oldinfo->prev;
123 if (oldinfo == mem_start)
124 mem_start = oldinfo->next;
127 newinfo->line = line;
128 newinfo->byte = byte;
129 newinfo->file = file;
130 newinfo->prev = NULL;
131 newinfo->next = mem_start;
133 mem_start->prev = newinfo;
136 mem_ab -= oldinfo->byte;
137 mem_hw -= oldinfo->byte;
138 mem_ab += newinfo->byte;
139 mem_hw += newinfo->byte;
148 static void util_dumpmem(struct memblock_t *memory, uint16_t cols) {
150 for (i = 0; i < memory->byte + ((memory->byte % cols) ? (cols - memory->byte % cols) : 0); i++) {
151 if (i % cols == 0) con_out(" 0x%06X: ", i);
152 if (i < memory->byte) con_out("%02X " , 0xFF & ((char*)(memory + 1))[i]);
155 if ((uint16_t)(i % cols) == (cols - 1)) {
156 for (j = i - (cols - 1); j <= i; j++) {
160 : (isprint(((char*)(memory + 1))[j]))
161 ? 0xFF & ((char*)(memory + 1)) [j]
170 void util_meminfo() {
171 struct memblock_t *info;
173 if (OPTS_OPTION_BOOL(OPTION_DEBUG)) {
174 for (info = mem_start; info; info = info->next) {
175 con_out("lost: %u (bytes) at %s:%u\n",
180 util_dumpmem(info, OPTS_OPTION_U16(OPTION_MEMDUMPCOLS));
184 if (OPTS_OPTION_BOOL(OPTION_DEBUG) ||
185 OPTS_OPTION_BOOL(OPTION_MEMCHK)) {
186 con_out("Memory information:\n\
187 Total allocations: %llu\n\
188 Total deallocations: %llu\n\
189 Total allocated: %f (MB)\n\
190 Total deallocated: %f (MB)\n\
191 Total peak memory: %f (MB)\n\
192 Total leaked memory: %f (MB) in %llu allocations\n",
195 (float)(mem_ab) / 1048576.0f,
196 (float)(mem_db) / 1048576.0f,
197 (float)(mem_pk) / 1048576.0f,
198 (float)(mem_ab - mem_db) / 1048576.0f,
200 /* could be more clever */
207 * Some string utility functions, because strdup uses malloc, and we want
208 * to track all memory (without replacing malloc).
210 char *_util_Estrdup(const char *s, const char *file, size_t line) {
214 /* in case of -DNOTRACK */
221 if ((len = strlen(s)) && (ptr = (char*)mem_af(len+1, line, file))) {
228 char *_util_Estrdup_empty(const char *s, const char *file, size_t line) {
232 /* in case of -DNOTRACK */
240 if ((ptr = (char*)mem_af(len+1, line, file))) {
247 void util_debug(const char *area, const char *ms, ...) {
249 if (!OPTS_OPTION_BOOL(OPTION_DEBUG))
252 if (!strcmp(area, "MEM") && !OPTS_OPTION_BOOL(OPTION_MEMCHK))
256 con_out ("[%s] ", area);
262 * only required if big endian .. otherwise no need to swap
265 #if PLATFORM_BYTE_ORDER == GMQCC_BYTE_ORDER_BIG
266 static GMQCC_INLINE void util_swap16(uint16_t *d, size_t l) {
268 d[l] = (d[l] << 8) | (d[l] >> 8);
272 static GMQCC_INLINE void util_swap32(uint32_t *d, size_t l) {
275 v = ((d[l] << 8) & 0xFF00FF00) | ((d[l] >> 8) & 0x00FF00FF);
276 d[l] = (v << 16) | (v >> 16);
280 /* Some strange system doesn't like constants that big, AND doesn't recognize an ULL suffix
281 * so let's go the safe way
283 static GMQCC_INLINE void util_swap64(uint32_t *d, size_t l) {
287 v = ((d[l] << 8) & 0xFF00FF00FF00FF00) | ((d[l] >> 8) & 0x00FF00FF00FF00FF);
288 v = ((v << 16) & 0xFFFF0000FFFF0000) | ((v >> 16) & 0x0000FFFF0000FFFF);
289 d[l] = (v << 32) | (v >> 32);
293 for (i = 0; i < l; i += 2) {
302 void util_endianswap(void *_data, size_t length, unsigned int typesize) {
303 # if PLATFORM_BYTE_ORDER == -1 /* runtime check */
304 if (*((char*)&typesize))
307 /* prevent unused warnings */
312 # if PLATFORM_BYTE_ORDER == GMQCC_BYTE_ORDER_LITTLE
318 util_swap16((uint16_t*)_data, length>>1);
321 util_swap32((uint32_t*)_data, length>>2);
324 util_swap64((uint32_t*)_data, length>>3);
327 default: exit(EXIT_FAILURE); /* please blow the fuck up! */
334 * CRC algorithms vary in the width of the polynomial, the value of said polynomial,
335 * the initial value used for the register, weather the bits of each byte are reflected
336 * before being processed, weather the algorithm itself feeds input bytes through the
337 * register or XORs them with a byte from one end and then straight into the table, as
338 * well as (but not limited to the idea of reflected versions) where the final register
339 * value becomes reversed, and finally weather the value itself is used to XOR the final
340 * register value. AS such you can already imagine how painfully annoying CRCs are,
341 * of course we stand to target Quake, which expects it's certian set of rules for proper
342 * calculation of a CRC.
344 * In most traditional CRC algorithms on uses a reflected table driven method where a value
345 * or register is reflected if it's bits are swapped around it's center. For example:
346 * take the bits 0101 is the 4-bit reflection of 1010, and respectfully 0011 would be the
347 * reflection of 1100. Quake however expects a NON-Reflected CRC on the output, but still
348 * requires a final XOR on the values (0xFFFF and 0x0000) this is a standard CCITT CRC-16
349 * which I respectfully as a programmer don't agree with.
351 * So now you know what we target, and why we target it, despite how unsettling it may seem
352 * but those are what Quake seems to request.
355 static const uint16_t util_crc16_table[] = {
356 0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50A5,
357 0x60C6, 0x70E7, 0x8108, 0x9129, 0xA14A, 0xB16B,
358 0xC18C, 0xD1AD, 0xE1CE, 0xF1EF, 0x1231, 0x0210,
359 0x3273, 0x2252, 0x52B5, 0x4294, 0x72F7, 0x62D6,
360 0x9339, 0x8318, 0xB37B, 0xA35A, 0xD3BD, 0xC39C,
361 0xF3FF, 0xE3DE, 0x2462, 0x3443, 0x0420, 0x1401,
362 0x64E6, 0x74C7, 0x44A4, 0x5485, 0xA56A, 0xB54B,
363 0x8528, 0x9509, 0xE5EE, 0xF5CF, 0xC5AC, 0xD58D,
364 0x3653, 0x2672, 0x1611, 0x0630, 0x76D7, 0x66F6,
365 0x5695, 0x46B4, 0xB75B, 0xA77A, 0x9719, 0x8738,
366 0xF7DF, 0xE7FE, 0xD79D, 0xC7BC, 0x48C4, 0x58E5,
367 0x6886, 0x78A7, 0x0840, 0x1861, 0x2802, 0x3823,
368 0xC9CC, 0xD9ED, 0xE98E, 0xF9AF, 0x8948, 0x9969,
369 0xA90A, 0xB92B, 0x5AF5, 0x4AD4, 0x7AB7, 0x6A96,
370 0x1A71, 0x0A50, 0x3A33, 0x2A12, 0xDBFD, 0xCBDC,
371 0xFBBF, 0xEB9E, 0x9B79, 0x8B58, 0xBB3B, 0xAB1A,
372 0x6CA6, 0x7C87, 0x4CE4, 0x5CC5, 0x2C22, 0x3C03,
373 0x0C60, 0x1C41, 0xEDAE, 0xFD8F, 0xCDEC, 0xDDCD,
374 0xAD2A, 0xBD0B, 0x8D68, 0x9D49, 0x7E97, 0x6EB6,
375 0x5ED5, 0x4EF4, 0x3E13, 0x2E32, 0x1E51, 0x0E70,
376 0xFF9F, 0xEFBE, 0xDFDD, 0xCFFC, 0xBF1B, 0xAF3A,
377 0x9F59, 0x8F78, 0x9188, 0x81A9, 0xB1CA, 0xA1EB,
378 0xD10C, 0xC12D, 0xF14E, 0xE16F, 0x1080, 0x00A1,
379 0x30C2, 0x20E3, 0x5004, 0x4025, 0x7046, 0x6067,
380 0x83B9, 0x9398, 0xA3FB, 0xB3DA, 0xC33D, 0xD31C,
381 0xE37F, 0xF35E, 0x02B1, 0x1290, 0x22F3, 0x32D2,
382 0x4235, 0x5214, 0x6277, 0x7256, 0xB5EA, 0xA5CB,
383 0x95A8, 0x8589, 0xF56E, 0xE54F, 0xD52C, 0xC50D,
384 0x34E2, 0x24C3, 0x14A0, 0x0481, 0x7466, 0x6447,
385 0x5424, 0x4405, 0xA7DB, 0xB7FA, 0x8799, 0x97B8,
386 0xE75F, 0xF77E, 0xC71D, 0xD73C, 0x26D3, 0x36F2,
387 0x0691, 0x16B0, 0x6657, 0x7676, 0x4615, 0x5634,
388 0xD94C, 0xC96D, 0xF90E, 0xE92F, 0x99C8, 0x89E9,
389 0xB98A, 0xA9AB, 0x5844, 0x4865, 0x7806, 0x6827,
390 0x18C0, 0x08E1, 0x3882, 0x28A3, 0xCB7D, 0xDB5C,
391 0xEB3F, 0xFB1E, 0x8BF9, 0x9BD8, 0xABBB, 0xBB9A,
392 0x4A75, 0x5A54, 0x6A37, 0x7A16, 0x0AF1, 0x1AD0,
393 0x2AB3, 0x3A92, 0xFD2E, 0xED0F, 0xDD6C, 0xCD4D,
394 0xBDAA, 0xAD8B, 0x9DE8, 0x8DC9, 0x7C26, 0x6C07,
395 0x5C64, 0x4C45, 0x3CA2, 0x2C83, 0x1CE0, 0x0CC1,
396 0xEF1F, 0xFF3E, 0xCF5D, 0xDF7C, 0xAF9B, 0xBFBA,
397 0x8FD9, 0x9FF8, 0x6E17, 0x7E36, 0x4E55, 0x5E74,
398 0x2E93, 0x3EB2, 0x0ED1, 0x1EF0
401 /* Non - Reflected */
402 uint16_t util_crc16(uint16_t current, const char *k, size_t len) {
403 register uint16_t h = current;
404 for (; len; --len, ++k)
405 h = util_crc16_table[(h>>8)^((unsigned char)*k)]^(h<<8);
408 /* Reflective Varation (for reference) */
410 uint16_t util_crc16(const char *k, int len, const short clamp) {
411 register uint16_t h= (uint16_t)0xFFFFFFFF;
412 for (; len; --len, ++k)
413 h = util_crc16_table[(h^((unsigned char)*k))&0xFF]^(h>>8);
418 size_t util_strtocmd(const char *in, char *out, size_t outsz) {
420 for (; *in && sz < outsz; ++in, ++out, ++sz)
421 *out = (*in == '-') ? '_' : (isalpha(*in) && !isupper(*in)) ? *in + 'A' - 'a': *in;
426 size_t util_strtononcmd(const char *in, char *out, size_t outsz) {
428 for (; *in && sz < outsz; ++in, ++out, ++sz)
429 *out = (*in == '_') ? '-' : (isalpha(*in) && isupper(*in)) ? *in + 'a' - 'A' : *in;
434 /* TODO: rewrite ... when I redo the ve cleanup */
435 void _util_vec_grow(void **a, size_t i, size_t s) {
436 vector_t *d = vec_meta(*a);
437 size_t m = *a ? 2 * d->allocated +i : i+1;
438 void *p = mem_r((*a ? d : NULL), s * m + sizeof(vector_t));
441 ((vector_t*)p)->used = 0;
442 *a = (vector_t*)p + 1;
444 vec_meta(*a)->allocated = m;
448 * Hash table for generic data, based on dynamic memory allocations
449 * all around. This is the internal interface, please look for
450 * EXPOSED INTERFACE comment below
452 typedef struct hash_node_t {
453 char *key; /* the key for this node in table */
454 void *value; /* pointer to the data as void* */
455 struct hash_node_t *next; /* next node (linked list) */
458 GMQCC_INLINE size_t util_hthash(hash_table_t *ht, const char *key) {
459 const uint32_t mix = 0x5BD1E995;
460 const uint32_t rot = 24;
461 size_t size = strlen(key);
462 uint32_t hash = 0x1EF0 /* LICRC TAB */ ^ size;
464 const unsigned char *data = (const unsigned char*)key;
467 alias = (data[0] | (data[1] << 8) | (data[2] << 16) | (data[3] << 24));
469 alias ^= alias >> rot;
480 case 3: hash ^= data[2] << 16;
481 case 2: hash ^= data[1] << 8;
482 case 1: hash ^= data[0];
490 return (size_t) (hash % ht->size);
493 static hash_node_t *_util_htnewpair(const char *key, void *value) {
495 if (!(node = (hash_node_t*)mem_a(sizeof(hash_node_t))))
498 if (!(node->key = util_strdupe(key))) {
510 * EXPOSED INTERFACE for the hashtable implementation
511 * util_htnew(size) -- to make a new hashtable
512 * util_htset(table, key, value, sizeof(value)) -- to set something in the table
513 * util_htget(table, key) -- to get something from the table
514 * util_htdel(table) -- to delete the table
516 hash_table_t *util_htnew(size_t size) {
517 hash_table_t *hashtable = NULL;
521 if (!(hashtable = (hash_table_t*)mem_a(sizeof(hash_table_t))))
524 if (!(hashtable->table = (hash_node_t**)mem_a(sizeof(hash_node_t*) * size))) {
529 hashtable->size = size;
530 memset(hashtable->table, 0, sizeof(hash_node_t*) * size);
535 void util_htseth(hash_table_t *ht, const char *key, size_t bin, void *value) {
536 hash_node_t *newnode = NULL;
537 hash_node_t *next = NULL;
538 hash_node_t *last = NULL;
540 next = ht->table[bin];
542 while (next && next->key && strcmp(key, next->key) > 0)
543 last = next, next = next->next;
545 /* already in table, do a replace */
546 if (next && next->key && strcmp(key, next->key) == 0) {
549 /* not found, grow a pair man :P */
550 newnode = _util_htnewpair(key, value);
551 if (next == ht->table[bin]) {
552 newnode->next = next;
553 ht->table[bin] = newnode;
555 last->next = newnode;
557 newnode->next = next;
558 last->next = newnode;
563 void util_htset(hash_table_t *ht, const char *key, void *value) {
564 util_htseth(ht, key, util_hthash(ht, key), value);
567 void *util_htgeth(hash_table_t *ht, const char *key, size_t bin) {
568 hash_node_t *pair = ht->table[bin];
570 while (pair && pair->key && strcmp(key, pair->key) > 0)
573 if (!pair || !pair->key || strcmp(key, pair->key) != 0)
579 void *util_htget(hash_table_t *ht, const char *key) {
580 return util_htgeth(ht, key, util_hthash(ht, key));
583 void *code_util_str_htgeth(hash_table_t *ht, const char *key, size_t bin) {
588 keylen = strlen(key);
590 pair = ht->table[bin];
591 while (pair && pair->key) {
592 len = strlen(pair->key);
598 cmp = strcmp(key, pair->key);
606 cmp = strcmp(key, pair->key + len - keylen);
608 uintptr_t up = (uintptr_t)pair->value;
618 * Free all allocated data in a hashtable, this is quite the amount
621 void util_htrem(hash_table_t *ht, void (*callback)(void *data)) {
623 for (; i < ht->size; i++) {
624 hash_node_t *n = ht->table[i];
644 void util_htrmh(hash_table_t *ht, const char *key, size_t bin, void (*cb)(void*)) {
645 hash_node_t **pair = &ht->table[bin];
648 while (*pair && (*pair)->key && strcmp(key, (*pair)->key) > 0)
649 pair = &(*pair)->next;
652 if (!tmp || !tmp->key || strcmp(key, tmp->key) != 0)
663 void util_htrm(hash_table_t *ht, const char *key, void (*cb)(void*)) {
664 util_htrmh(ht, key, util_hthash(ht, key), cb);
667 void util_htdel(hash_table_t *ht) {
668 util_htrem(ht, NULL);
672 * Portable implementation of vasprintf/asprintf. Assumes vsnprintf
673 * exists, otherwise compiler error.
675 * TODO: fix for MSVC ....
677 int util_vasprintf(char **dat, const char *fmt, va_list args) {
683 * For visuals tido _vsnprintf doesn't tell you the length of a
684 * formatted string if it overflows. However there is a MSVC
685 * intrinsic (which is documented wrong) called _vcsprintf which
686 * will return the required amount to allocate.
689 if ((len = _vscprintf(fmt, args)) < 0) {
694 tmp = (char*)mem_a(len + 1);
695 if ((ret = _vsnprintf_s(tmp, len+1, len+1, fmt, args)) != len) {
704 * For everything else we have a decent conformint vsnprintf that
705 * returns the number of bytes needed. We give it a try though on
706 * a short buffer, since efficently speaking, it could be nice to
707 * above a second vsnprintf call.
712 len = vsnprintf(buf, sizeof(buf), fmt, cpy);
715 if (len < (int)sizeof(buf)) {
716 *dat = util_strdup(buf);
720 /* not large enough ... */
721 tmp = (char*)mem_a(len + 1);
722 if ((ret = vsnprintf(tmp, len + 1, fmt, args)) != len) {
732 int util_asprintf(char **ret, const char *fmt, ...) {
736 read = util_vasprintf(ret, fmt, args);
743 * These are various re-implementations (wrapping the real ones) of
744 * string functions that MSVC consideres unsafe. We wrap these up and
745 * use the safe varations on MSVC.
748 static char **util_strerror_allocated() {
749 static char **data = NULL;
753 static void util_strerror_cleanup(void) {
755 char **data = util_strerror_allocated();
756 for (i = 0; i < vec_size(data); i++)
761 const char *util_strerror(int num) {
762 char *allocated = NULL;
763 static bool install = false;
764 static size_t tries = 0;
765 char **vector = util_strerror_allocated();
767 /* try installing cleanup handler */
772 install = !atexit(&util_strerror_cleanup);
776 allocated = (char*)mem_a(4096); /* A page must be enough */
777 strerror_s(allocated, 4096, num);
779 vec_push(vector, allocated);
780 return (const char *)allocated;
783 int util_snprintf(char *src, size_t bytes, const char *format, ...) {
786 va_start(va, format);
788 rt = vsprintf_s(src, bytes, format, va);
794 char *util_strcat(char *dest, const char *src) {
795 strcat_s(dest, strlen(src), src);
799 char *util_strncpy(char *dest, const char *src, size_t num) {
800 strncpy_s(dest, num, src, num);
804 const char *util_strerror(int num) {
805 return strerror(num);
808 int util_snprintf(char *src, size_t bytes, const char *format, ...) {
811 va_start(va, format);
812 rt = vsnprintf(src, bytes, format, va);
818 char *util_strcat(char *dest, const char *src) {
819 return strcat(dest, src);
822 char *util_strncpy(char *dest, const char *src, size_t num) {
823 return strncpy(dest, src, num);
826 #endif /*! _MSC_VER */
829 * Implementation of the Mersenne twister PRNG (pseudo random numer
830 * generator). Implementation of MT19937. Has a period of 2^19937-1
831 * which is a Mersenne Prime (hence the name).
833 * Implemented from specification and original paper:
834 * http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/ARTICLES/mt.pdf
836 * This code is placed in the public domain by me personally
837 * (Dale Weiler, a.k.a graphitemaster).
841 #define MT_PERIOD 397
842 #define MT_SPACE (MT_SIZE - MT_PERIOD)
844 static uint32_t mt_state[MT_SIZE];
845 static size_t mt_index = 0;
847 static GMQCC_INLINE void mt_generate() {
849 * The loop has been unrolled here: the original paper and implemenation
850 * Called for the following code:
851 * for (register unsigned i = 0; i < MT_SIZE; ++i) {
852 * register uint32_t load;
853 * load = (0x80000000 & mt_state[i]) // most significant 32nd bit
854 * load |= (0x7FFFFFFF & mt_state[(i + 1) % MT_SIZE]) // least significant 31nd bit
856 * mt_state[i] = mt_state[(i + MT_PERIOD) % MT_SIZE] ^ (load >> 1);
858 * if (load & 1) mt_state[i] ^= 0x9908B0DF;
861 * This essentially is a waste: we have two modulus operations, and
862 * a branch that is executed every iteration from [0, MT_SIZE).
864 * Please see: http://www.quadibloc.com/crypto/co4814.htm for more
865 * information on how this clever trick works.
867 static const uint32_t matrix[2] = {
872 * This register gives up a little more speed by instructing the compiler
873 * to force these into CPU registers (they're counters for indexing mt_state
874 * which we can force the compiler to generate prefetch instructions for)
880 * Said loop has been unrolled for MT_SPACE (226 iterations), opposed
881 * to [0, MT_SIZE) (634 iterations).
883 for (i = 0; i < MT_SPACE; ++i) {
884 y = (0x80000000 & mt_state[i]) | (0x7FFFFFF & mt_state[i + 1]);
885 mt_state[i] = mt_state[i + MT_PERIOD] ^ (y >> 1) ^ matrix[y & 1];
887 i ++; /* loop unroll */
889 y = (0x80000000 & mt_state[i]) | (0x7FFFFFF & mt_state[i + 1]);
890 mt_state[i] = mt_state[i + MT_PERIOD] ^ (y >> 1) ^ matrix[y & 1];
894 * collapsing the walls unrolled (evenly dividing 396 [632-227 = 396
898 while (i < MT_SIZE - 1) {
900 * We expand this 11 times .. manually, no macros are required
901 * here. This all fits in the CPU cache.
903 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
904 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
906 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
907 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
909 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
910 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
912 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
913 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
915 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
916 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
918 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
919 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
921 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
922 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
924 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
925 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
927 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
928 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
930 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
931 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
933 y = (0x80000000 & mt_state[i]) | (0x7FFFFFFF & mt_state[i + 1]);
934 mt_state[i] = mt_state[i - MT_SPACE] ^ (y >> 1) ^ matrix[y & 1];
938 /* i = mt_state[623] */
939 y = (0x80000000 & mt_state[MT_SIZE - 1]) | (0x7FFFFFFF & mt_state[MT_SIZE - 1]);
940 mt_state[MT_SIZE - 1] = mt_state[MT_PERIOD - 1] ^ (y >> 1) ^ matrix[y & 1];
943 void util_seed(uint32_t value) {
945 * We seed the mt_state with a LCG (linear congruential generator)
946 * We're operating exactly on exactly m=32, so there is no need to
949 * The multipler of choice is 0x6C07865, also knows as the Borosh-
950 * Niederreiter multipler used for modulus 2^32. More can be read
951 * about this in Knuth's TAOCP Volume 2, page 106.
953 * If you don't own TAOCP something is wrong with you :-) .. so I
954 * also provided a link to the original paper by Borosh and
955 * Niederreiter. It's called "Optional Multipliers for PRNG by The
956 * Linear Congruential Method" (1983).
957 * http://en.wikipedia.org/wiki/Linear_congruential_generator
959 * From said page, it says the following:
960 * "A common Mersenne twister implementation, interestingly enough
961 * used an LCG to generate seed data."
964 * The data we're operating on is 32-bits for the mt_state array, so
965 * there is no masking required with 0xFFFFFFFF
970 for (i = 1; i < MT_SIZE; ++i)
971 mt_state[i] = 0x6C078965 * (mt_state[i - 1] ^ mt_state[i - 1] >> 30) + i;
974 uint32_t util_rand() {
978 * This is inlined with any sane compiler (I checked)
979 * for some reason though, SubC seems to be generating invalid
980 * code when it inlines this.
985 y = mt_state[mt_index];
987 /* Standard tempering */
988 y ^= y >> 11; /* +7 */
989 y ^= y << 7 & 0x9D2C5680; /* +4 */
990 y ^= y << 15 & 0xEFC60000; /* -4 */
991 y ^= y >> 18; /* -7 */
993 if(++mt_index == MT_SIZE)