[409] | 1 | /* |
---|
| 2 | ------------------------------------------------------------------------------- |
---|
| 3 | lookup3.c, by Bob Jenkins, May 2006, Public Domain. |
---|
| 4 | Original: http://burtleburtle.net/bob/c/lookup3.c |
---|
| 5 | Modified by Russ Rew for adaption in netCDF. |
---|
| 6 | - Make use of Paul Hsieh's pstdint.h, if stdint.h not available. |
---|
| 7 | - Declare unused functions static to keep global namespace clean. |
---|
| 8 | - Provide function hash_fast() that uses either hashlittle() or |
---|
| 9 | hashbig(), depending on endianness. |
---|
| 10 | - Because portability is more important than speed for netCDF use, |
---|
| 11 | we define VALGRIND to skip "#ifndef VALGRIND" code, so reads of |
---|
| 12 | strings don't access extra bytes after end of string. This may |
---|
| 13 | slow it down enough to justify a simpler hash, but blame me, not |
---|
| 14 | original author! |
---|
| 15 | |
---|
| 16 | These are functions for producing 32-bit hashes for hash table lookup. |
---|
| 17 | hashword(), hashlittle(), hashlittle2(), hashbig(), mix(), and final() |
---|
| 18 | are externally useful functions. Routines to test the hash are included |
---|
| 19 | if SELF_TEST is defined. You can use this free for any purpose. It's in |
---|
| 20 | the public domain. It has no warranty. |
---|
| 21 | |
---|
| 22 | You probably want to use hashlittle(). hashlittle() and hashbig() |
---|
| 23 | hash byte arrays. hashlittle() is is faster than hashbig() on |
---|
| 24 | little-endian machines. Intel and AMD are little-endian machines. |
---|
| 25 | On second thought, you probably want hashlittle2(), which is identical to |
---|
| 26 | hashlittle() except it returns two 32-bit hashes for the price of one. |
---|
| 27 | You could implement hashbig2() if you wanted but I haven't bothered here. |
---|
| 28 | |
---|
| 29 | If you want to find a hash of, say, exactly 7 integers, do |
---|
| 30 | a = i1; b = i2; c = i3; |
---|
| 31 | mix(a,b,c); |
---|
| 32 | a += i4; b += i5; c += i6; |
---|
| 33 | mix(a,b,c); |
---|
| 34 | a += i7; |
---|
| 35 | final(a,b,c); |
---|
| 36 | then use c as the hash value. If you have a variable length array of |
---|
| 37 | 4-byte integers to hash, use hashword(). If you have a byte array (like |
---|
| 38 | a character string), use hashlittle(). If you have several byte arrays, or |
---|
| 39 | a mix of things, see the comments above hashlittle(). |
---|
| 40 | |
---|
| 41 | Why is this so big? I read 12 bytes at a time into 3 4-byte integers, |
---|
| 42 | then mix those integers. This is fast (you can do a lot more thorough |
---|
| 43 | mixing with 12*3 instructions on 3 integers than you can with 3 instructions |
---|
| 44 | on 1 byte), but shoehorning those bytes into integers efficiently is messy. |
---|
| 45 | ------------------------------------------------------------------------------- |
---|
| 46 | */ |
---|
| 47 | /* #define SELF_TEST 1 */ |
---|
| 48 | |
---|
| 49 | #include <config.h> |
---|
| 50 | #include <stdio.h> /* defines printf for tests */ |
---|
| 51 | #include <time.h> /* defines time_t for timings in the test */ |
---|
| 52 | #ifndef HAVE_STDINT_H |
---|
| 53 | # include "pstdint.h" /* attempts to define uint32_t etc portably */ |
---|
| 54 | #else |
---|
| 55 | # include <stdint.h> |
---|
| 56 | #endif /* HAVE_STDINT_H */ |
---|
| 57 | #ifdef HAVE_SYS_PARAM_H |
---|
| 58 | #include <sys/param.h> /* attempt to define endianness */ |
---|
| 59 | #endif /* HAVE_SYS_PARAM_H */ |
---|
| 60 | #ifdef linux |
---|
| 61 | # include <endian.h> /* attempt to define endianness */ |
---|
| 62 | #endif |
---|
| 63 | |
---|
| 64 | #define VALGRIND /* added by Russ Rew, for portability over speed */ |
---|
| 65 | |
---|
| 66 | #ifndef WORDS_BIGENDIAN /* from config.h */ |
---|
| 67 | #define HASH_LITTLE_ENDIAN 1 |
---|
| 68 | #define HASH_BIG_ENDIAN 0 |
---|
| 69 | #else |
---|
| 70 | #define HASH_LITTLE_ENDIAN 0 |
---|
| 71 | #define HASH_BIG_ENDIAN 1 |
---|
| 72 | #endif |
---|
| 73 | |
---|
| 74 | #define hashsize(n) ((uint32_t)1<<(n)) |
---|
| 75 | #define hashmask(n) (hashsize(n)-1) |
---|
| 76 | #define rot(x,k) (((x)<<(k)) | ((x)>>(32-(k)))) |
---|
| 77 | |
---|
| 78 | /* |
---|
| 79 | ------------------------------------------------------------------------------- |
---|
| 80 | mix -- mix 3 32-bit values reversibly. |
---|
| 81 | |
---|
| 82 | This is reversible, so any information in (a,b,c) before mix() is |
---|
| 83 | still in (a,b,c) after mix(). |
---|
| 84 | |
---|
| 85 | If four pairs of (a,b,c) inputs are run through mix(), or through |
---|
| 86 | mix() in reverse, there are at least 32 bits of the output that |
---|
| 87 | are sometimes the same for one pair and different for another pair. |
---|
| 88 | This was tested for: |
---|
| 89 | * pairs that differed by one bit, by two bits, in any combination |
---|
| 90 | of top bits of (a,b,c), or in any combination of bottom bits of |
---|
| 91 | (a,b,c). |
---|
| 92 | * "differ" is defined as +, -, ^, or ~^. For + and -, I transformed |
---|
| 93 | the output delta to a Gray code (a^(a>>1)) so a string of 1's (as |
---|
| 94 | is commonly produced by subtraction) look like a single 1-bit |
---|
| 95 | difference. |
---|
| 96 | * the base values were pseudorandom, all zero but one bit set, or |
---|
| 97 | all zero plus a counter that starts at zero. |
---|
| 98 | |
---|
| 99 | Some k values for my "a-=c; a^=rot(c,k); c+=b;" arrangement that |
---|
| 100 | satisfy this are |
---|
| 101 | 4 6 8 16 19 4 |
---|
| 102 | 9 15 3 18 27 15 |
---|
| 103 | 14 9 3 7 17 3 |
---|
| 104 | Well, "9 15 3 18 27 15" didn't quite get 32 bits diffing |
---|
| 105 | for "differ" defined as + with a one-bit base and a two-bit delta. I |
---|
| 106 | used http://burtleburtle.net/bob/hash/avalanche.html to choose |
---|
| 107 | the operations, constants, and arrangements of the variables. |
---|
| 108 | |
---|
| 109 | This does not achieve avalanche. There are input bits of (a,b,c) |
---|
| 110 | that fail to affect some output bits of (a,b,c), especially of a. The |
---|
| 111 | most thoroughly mixed value is c, but it doesn't really even achieve |
---|
| 112 | avalanche in c. |
---|
| 113 | |
---|
| 114 | This allows some parallelism. Read-after-writes are good at doubling |
---|
| 115 | the number of bits affected, so the goal of mixing pulls in the opposite |
---|
| 116 | direction as the goal of parallelism. I did what I could. Rotates |
---|
| 117 | seem to cost as much as shifts on every machine I could lay my hands |
---|
| 118 | on, and rotates are much kinder to the top and bottom bits, so I used |
---|
| 119 | rotates. |
---|
| 120 | ------------------------------------------------------------------------------- |
---|
| 121 | */ |
---|
| 122 | #define mix(a,b,c) \ |
---|
| 123 | { \ |
---|
| 124 | a -= c; a ^= rot(c, 4); c += b; \ |
---|
| 125 | b -= a; b ^= rot(a, 6); a += c; \ |
---|
| 126 | c -= b; c ^= rot(b, 8); b += a; \ |
---|
| 127 | a -= c; a ^= rot(c,16); c += b; \ |
---|
| 128 | b -= a; b ^= rot(a,19); a += c; \ |
---|
| 129 | c -= b; c ^= rot(b, 4); b += a; \ |
---|
| 130 | } |
---|
| 131 | |
---|
| 132 | /* |
---|
| 133 | ------------------------------------------------------------------------------- |
---|
| 134 | final -- final mixing of 3 32-bit values (a,b,c) into c |
---|
| 135 | |
---|
| 136 | Pairs of (a,b,c) values differing in only a few bits will usually |
---|
| 137 | produce values of c that look totally different. This was tested for |
---|
| 138 | * pairs that differed by one bit, by two bits, in any combination |
---|
| 139 | of top bits of (a,b,c), or in any combination of bottom bits of |
---|
| 140 | (a,b,c). |
---|
| 141 | * "differ" is defined as +, -, ^, or ~^. For + and -, I transformed |
---|
| 142 | the output delta to a Gray code (a^(a>>1)) so a string of 1's (as |
---|
| 143 | is commonly produced by subtraction) look like a single 1-bit |
---|
| 144 | difference. |
---|
| 145 | * the base values were pseudorandom, all zero but one bit set, or |
---|
| 146 | all zero plus a counter that starts at zero. |
---|
| 147 | |
---|
| 148 | These constants passed: |
---|
| 149 | 14 11 25 16 4 14 24 |
---|
| 150 | 12 14 25 16 4 14 24 |
---|
| 151 | and these came close: |
---|
| 152 | 4 8 15 26 3 22 24 |
---|
| 153 | 10 8 15 26 3 22 24 |
---|
| 154 | 11 8 15 26 3 22 24 |
---|
| 155 | ------------------------------------------------------------------------------- |
---|
| 156 | */ |
---|
| 157 | #define final(a,b,c) \ |
---|
| 158 | { \ |
---|
| 159 | c ^= b; c -= rot(b,14); \ |
---|
| 160 | a ^= c; a -= rot(c,11); \ |
---|
| 161 | b ^= a; b -= rot(a,25); \ |
---|
| 162 | c ^= b; c -= rot(b,16); \ |
---|
| 163 | a ^= c; a -= rot(c,4); \ |
---|
| 164 | b ^= a; b -= rot(a,14); \ |
---|
| 165 | c ^= b; c -= rot(b,24); \ |
---|
| 166 | } |
---|
| 167 | |
---|
| 168 | /* |
---|
| 169 | -------------------------------------------------------------------- |
---|
| 170 | This works on all machines. To be useful, it requires |
---|
| 171 | -- that the key be an array of uint32_t's, and |
---|
| 172 | -- that the length be the number of uint32_t's in the key |
---|
| 173 | |
---|
| 174 | The function hashword() is identical to hashlittle() on little-endian |
---|
| 175 | machines, and identical to hashbig() on big-endian machines, |
---|
| 176 | except that the length has to be measured in uint32_ts rather than in |
---|
| 177 | bytes. hashlittle() is more complicated than hashword() only because |
---|
| 178 | hashlittle() has to dance around fitting the key bytes into registers. |
---|
| 179 | -------------------------------------------------------------------- |
---|
| 180 | */ |
---|
| 181 | #ifdef SELF_TEST |
---|
| 182 | static |
---|
| 183 | uint32_t hashword( |
---|
| 184 | const uint32_t *k, /* the key, an array of uint32_t values */ |
---|
| 185 | size_t length, /* the length of the key, in uint32_ts */ |
---|
| 186 | uint32_t initval) /* the previous hash, or an arbitrary value */ |
---|
| 187 | { |
---|
| 188 | uint32_t a,b,c; |
---|
| 189 | |
---|
| 190 | /* Set up the internal state */ |
---|
| 191 | a = b = c = 0xdeadbeef + (((uint32_t)length)<<2) + initval; |
---|
| 192 | |
---|
| 193 | /*------------------------------------------------- handle most of the key */ |
---|
| 194 | while (length > 3) |
---|
| 195 | { |
---|
| 196 | a += k[0]; |
---|
| 197 | b += k[1]; |
---|
| 198 | c += k[2]; |
---|
| 199 | mix(a,b,c); |
---|
| 200 | length -= 3; |
---|
| 201 | k += 3; |
---|
| 202 | } |
---|
| 203 | |
---|
| 204 | /*------------------------------------------- handle the last 3 uint32_t's */ |
---|
| 205 | switch(length) /* all the case statements fall through */ |
---|
| 206 | { |
---|
| 207 | case 3 : c+=k[2]; |
---|
| 208 | case 2 : b+=k[1]; |
---|
| 209 | case 1 : a+=k[0]; |
---|
| 210 | final(a,b,c); |
---|
| 211 | case 0: /* case 0: nothing left to add */ |
---|
| 212 | break; |
---|
| 213 | } |
---|
| 214 | /*------------------------------------------------------ report the result */ |
---|
| 215 | return c; |
---|
| 216 | } |
---|
| 217 | |
---|
| 218 | /* |
---|
| 219 | -------------------------------------------------------------------- |
---|
| 220 | hashword2() -- same as hashword(), but take two seeds and return two |
---|
| 221 | 32-bit values. pc and pb must both be nonnull, and *pc and *pb must |
---|
| 222 | both be initialized with seeds. If you pass in (*pb)==0, the output |
---|
| 223 | (*pc) will be the same as the return value from hashword(). |
---|
| 224 | -------------------------------------------------------------------- |
---|
| 225 | */ |
---|
| 226 | static |
---|
| 227 | void hashword2 ( |
---|
| 228 | const uint32_t *k, /* the key, an array of uint32_t values */ |
---|
| 229 | size_t length, /* the length of the key, in uint32_ts */ |
---|
| 230 | uint32_t *pc, /* IN: seed OUT: primary hash value */ |
---|
| 231 | uint32_t *pb) /* IN: more seed OUT: secondary hash value */ |
---|
| 232 | { |
---|
| 233 | uint32_t a,b,c; |
---|
| 234 | |
---|
| 235 | /* Set up the internal state */ |
---|
| 236 | a = b = c = 0xdeadbeef + ((uint32_t)(length<<2)) + *pc; |
---|
| 237 | c += *pb; |
---|
| 238 | |
---|
| 239 | /*------------------------------------------------- handle most of the key */ |
---|
| 240 | while (length > 3) |
---|
| 241 | { |
---|
| 242 | a += k[0]; |
---|
| 243 | b += k[1]; |
---|
| 244 | c += k[2]; |
---|
| 245 | mix(a,b,c); |
---|
| 246 | length -= 3; |
---|
| 247 | k += 3; |
---|
| 248 | } |
---|
| 249 | |
---|
| 250 | /*------------------------------------------- handle the last 3 uint32_t's */ |
---|
| 251 | switch(length) /* all the case statements fall through */ |
---|
| 252 | { |
---|
| 253 | case 3 : c+=k[2]; |
---|
| 254 | case 2 : b+=k[1]; |
---|
| 255 | case 1 : a+=k[0]; |
---|
| 256 | final(a,b,c); |
---|
| 257 | case 0: /* case 0: nothing left to add */ |
---|
| 258 | break; |
---|
| 259 | } |
---|
| 260 | /*------------------------------------------------------ report the result */ |
---|
| 261 | *pc=c; *pb=b; |
---|
| 262 | } |
---|
| 263 | |
---|
| 264 | /* |
---|
| 265 | * hashlittle2: return 2 32-bit hash values |
---|
| 266 | * |
---|
| 267 | * This is identical to hashlittle(), except it returns two 32-bit hash |
---|
| 268 | * values instead of just one. This is good enough for hash table |
---|
| 269 | * lookup with 2^^64 buckets, or if you want a second hash if you're not |
---|
| 270 | * happy with the first, or if you want a probably-unique 64-bit ID for |
---|
| 271 | * the key. *pc is better mixed than *pb, so use *pc first. If you want |
---|
| 272 | * a 64-bit value do something like "*pc + (((uint64_t)*pb)<<32)". |
---|
| 273 | */ |
---|
| 274 | static void |
---|
| 275 | hashlittle2( |
---|
| 276 | const void *key, /* the key to hash */ |
---|
| 277 | size_t length, /* length of the key */ |
---|
| 278 | uint32_t *pc, /* IN: primary initval, OUT: primary hash */ |
---|
| 279 | uint32_t *pb) /* IN: secondary initval, OUT: secondary hash */ |
---|
| 280 | { |
---|
| 281 | uint32_t a,b,c; /* internal state */ |
---|
| 282 | union { const void *ptr; size_t i; } u; /* needed for Mac Powerbook G4 */ |
---|
| 283 | |
---|
| 284 | /* Set up the internal state */ |
---|
| 285 | a = b = c = 0xdeadbeef + ((uint32_t)length) + *pc; |
---|
| 286 | c += *pb; |
---|
| 287 | |
---|
| 288 | u.ptr = key; |
---|
| 289 | if (HASH_LITTLE_ENDIAN && ((u.i & 0x3) == 0)) { |
---|
| 290 | const uint32_t *k = (const uint32_t *)key; /* read 32-bit chunks */ |
---|
| 291 | const uint8_t *k8; |
---|
| 292 | |
---|
| 293 | /*------ all but last block: aligned reads and affect 32 bits of (a,b,c) */ |
---|
| 294 | while (length > 12) |
---|
| 295 | { |
---|
| 296 | a += k[0]; |
---|
| 297 | b += k[1]; |
---|
| 298 | c += k[2]; |
---|
| 299 | mix(a,b,c); |
---|
| 300 | length -= 12; |
---|
| 301 | k += 3; |
---|
| 302 | } |
---|
| 303 | |
---|
| 304 | /*----------------------------- handle the last (probably partial) block */ |
---|
| 305 | /* |
---|
| 306 | * "k[2]&0xffffff" actually reads beyond the end of the string, but |
---|
| 307 | * then masks off the part it's not allowed to read. Because the |
---|
| 308 | * string is aligned, the masked-off tail is in the same word as the |
---|
| 309 | * rest of the string. Every machine with memory protection I've seen |
---|
| 310 | * does it on word boundaries, so is OK with this. But VALGRIND will |
---|
| 311 | * still catch it and complain. The masking trick does make the hash |
---|
| 312 | * noticably faster for short strings (like English words). |
---|
| 313 | */ |
---|
| 314 | #ifndef VALGRIND |
---|
| 315 | |
---|
| 316 | switch(length) |
---|
| 317 | { |
---|
| 318 | case 12: c+=k[2]; b+=k[1]; a+=k[0]; break; |
---|
| 319 | case 11: c+=k[2]&0xffffff; b+=k[1]; a+=k[0]; break; |
---|
| 320 | case 10: c+=k[2]&0xffff; b+=k[1]; a+=k[0]; break; |
---|
| 321 | case 9 : c+=k[2]&0xff; b+=k[1]; a+=k[0]; break; |
---|
| 322 | case 8 : b+=k[1]; a+=k[0]; break; |
---|
| 323 | case 7 : b+=k[1]&0xffffff; a+=k[0]; break; |
---|
| 324 | case 6 : b+=k[1]&0xffff; a+=k[0]; break; |
---|
| 325 | case 5 : b+=k[1]&0xff; a+=k[0]; break; |
---|
| 326 | case 4 : a+=k[0]; break; |
---|
| 327 | case 3 : a+=k[0]&0xffffff; break; |
---|
| 328 | case 2 : a+=k[0]&0xffff; break; |
---|
| 329 | case 1 : a+=k[0]&0xff; break; |
---|
| 330 | case 0 : *pc=c; *pb=b; return; /* zero length strings require no mixing */ |
---|
| 331 | } |
---|
| 332 | |
---|
| 333 | #else /* make valgrind happy */ |
---|
| 334 | |
---|
| 335 | k8 = (const uint8_t *)k; |
---|
| 336 | switch(length) |
---|
| 337 | { |
---|
| 338 | case 12: c+=k[2]; b+=k[1]; a+=k[0]; break; |
---|
| 339 | case 11: c+=((uint32_t)k8[10])<<16; /* fall through */ |
---|
| 340 | case 10: c+=((uint32_t)k8[9])<<8; /* fall through */ |
---|
| 341 | case 9 : c+=k8[8]; /* fall through */ |
---|
| 342 | case 8 : b+=k[1]; a+=k[0]; break; |
---|
| 343 | case 7 : b+=((uint32_t)k8[6])<<16; /* fall through */ |
---|
| 344 | case 6 : b+=((uint32_t)k8[5])<<8; /* fall through */ |
---|
| 345 | case 5 : b+=k8[4]; /* fall through */ |
---|
| 346 | case 4 : a+=k[0]; break; |
---|
| 347 | case 3 : a+=((uint32_t)k8[2])<<16; /* fall through */ |
---|
| 348 | case 2 : a+=((uint32_t)k8[1])<<8; /* fall through */ |
---|
| 349 | case 1 : a+=k8[0]; break; |
---|
| 350 | case 0 : *pc=c; *pb=b; return; /* zero length strings require no mixing */ |
---|
| 351 | } |
---|
| 352 | |
---|
| 353 | #endif /* !valgrind */ |
---|
| 354 | |
---|
| 355 | } else if (HASH_LITTLE_ENDIAN && ((u.i & 0x1) == 0)) { |
---|
| 356 | const uint16_t *k = (const uint16_t *)key; /* read 16-bit chunks */ |
---|
| 357 | const uint8_t *k8; |
---|
| 358 | |
---|
| 359 | /*--------------- all but last block: aligned reads and different mixing */ |
---|
| 360 | while (length > 12) |
---|
| 361 | { |
---|
| 362 | a += k[0] + (((uint32_t)k[1])<<16); |
---|
| 363 | b += k[2] + (((uint32_t)k[3])<<16); |
---|
| 364 | c += k[4] + (((uint32_t)k[5])<<16); |
---|
| 365 | mix(a,b,c); |
---|
| 366 | length -= 12; |
---|
| 367 | k += 6; |
---|
| 368 | } |
---|
| 369 | |
---|
| 370 | /*----------------------------- handle the last (probably partial) block */ |
---|
| 371 | k8 = (const uint8_t *)k; |
---|
| 372 | switch(length) |
---|
| 373 | { |
---|
| 374 | case 12: c+=k[4]+(((uint32_t)k[5])<<16); |
---|
| 375 | b+=k[2]+(((uint32_t)k[3])<<16); |
---|
| 376 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 377 | break; |
---|
| 378 | case 11: c+=((uint32_t)k8[10])<<16; /* fall through */ |
---|
| 379 | case 10: c+=k[4]; |
---|
| 380 | b+=k[2]+(((uint32_t)k[3])<<16); |
---|
| 381 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 382 | break; |
---|
| 383 | case 9 : c+=k8[8]; /* fall through */ |
---|
| 384 | case 8 : b+=k[2]+(((uint32_t)k[3])<<16); |
---|
| 385 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 386 | break; |
---|
| 387 | case 7 : b+=((uint32_t)k8[6])<<16; /* fall through */ |
---|
| 388 | case 6 : b+=k[2]; |
---|
| 389 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 390 | break; |
---|
| 391 | case 5 : b+=k8[4]; /* fall through */ |
---|
| 392 | case 4 : a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 393 | break; |
---|
| 394 | case 3 : a+=((uint32_t)k8[2])<<16; /* fall through */ |
---|
| 395 | case 2 : a+=k[0]; |
---|
| 396 | break; |
---|
| 397 | case 1 : a+=k8[0]; |
---|
| 398 | break; |
---|
| 399 | case 0 : *pc=c; *pb=b; return; /* zero length strings require no mixing */ |
---|
| 400 | } |
---|
| 401 | |
---|
| 402 | } else { /* need to read the key one byte at a time */ |
---|
| 403 | const uint8_t *k = (const uint8_t *)key; |
---|
| 404 | |
---|
| 405 | /*--------------- all but the last block: affect some 32 bits of (a,b,c) */ |
---|
| 406 | while (length > 12) |
---|
| 407 | { |
---|
| 408 | a += k[0]; |
---|
| 409 | a += ((uint32_t)k[1])<<8; |
---|
| 410 | a += ((uint32_t)k[2])<<16; |
---|
| 411 | a += ((uint32_t)k[3])<<24; |
---|
| 412 | b += k[4]; |
---|
| 413 | b += ((uint32_t)k[5])<<8; |
---|
| 414 | b += ((uint32_t)k[6])<<16; |
---|
| 415 | b += ((uint32_t)k[7])<<24; |
---|
| 416 | c += k[8]; |
---|
| 417 | c += ((uint32_t)k[9])<<8; |
---|
| 418 | c += ((uint32_t)k[10])<<16; |
---|
| 419 | c += ((uint32_t)k[11])<<24; |
---|
| 420 | mix(a,b,c); |
---|
| 421 | length -= 12; |
---|
| 422 | k += 12; |
---|
| 423 | } |
---|
| 424 | |
---|
| 425 | /*-------------------------------- last block: affect all 32 bits of (c) */ |
---|
| 426 | switch(length) /* all the case statements fall through */ |
---|
| 427 | { |
---|
| 428 | case 12: c+=((uint32_t)k[11])<<24; |
---|
| 429 | case 11: c+=((uint32_t)k[10])<<16; |
---|
| 430 | case 10: c+=((uint32_t)k[9])<<8; |
---|
| 431 | case 9 : c+=k[8]; |
---|
| 432 | case 8 : b+=((uint32_t)k[7])<<24; |
---|
| 433 | case 7 : b+=((uint32_t)k[6])<<16; |
---|
| 434 | case 6 : b+=((uint32_t)k[5])<<8; |
---|
| 435 | case 5 : b+=k[4]; |
---|
| 436 | case 4 : a+=((uint32_t)k[3])<<24; |
---|
| 437 | case 3 : a+=((uint32_t)k[2])<<16; |
---|
| 438 | case 2 : a+=((uint32_t)k[1])<<8; |
---|
| 439 | case 1 : a+=k[0]; |
---|
| 440 | break; |
---|
| 441 | case 0 : *pc=c; *pb=b; return; /* zero length strings require no mixing */ |
---|
| 442 | } |
---|
| 443 | } |
---|
| 444 | |
---|
| 445 | final(a,b,c); |
---|
| 446 | *pc=c; *pb=b; |
---|
| 447 | } |
---|
| 448 | #endif /*SELF_TEST*/ |
---|
| 449 | |
---|
| 450 | |
---|
| 451 | #ifdef WORDS_BIGENDIAN |
---|
| 452 | /* |
---|
| 453 | * hashbig(): |
---|
| 454 | * This is the same as hashword() on big-endian machines. It is different |
---|
| 455 | * from hashlittle() on all machines. hashbig() takes advantage of |
---|
| 456 | * big-endian byte ordering. |
---|
| 457 | */ |
---|
| 458 | static uint32_t |
---|
| 459 | hashbig( const void *key, size_t length, uint32_t initval) |
---|
| 460 | { |
---|
| 461 | uint32_t a,b,c; |
---|
| 462 | union { const void *ptr; size_t i; } u; /* to cast key to (size_t) happily */ |
---|
| 463 | |
---|
| 464 | /* Set up the internal state */ |
---|
| 465 | a = b = c = 0xdeadbeef + ((uint32_t)length) + initval; |
---|
| 466 | |
---|
| 467 | u.ptr = key; |
---|
| 468 | if (HASH_BIG_ENDIAN && ((u.i & 0x3) == 0)) { |
---|
| 469 | const uint32_t *k = (const uint32_t *)key; /* read 32-bit chunks */ |
---|
| 470 | const uint8_t *k8; |
---|
| 471 | |
---|
| 472 | /*------ all but last block: aligned reads and affect 32 bits of (a,b,c) */ |
---|
| 473 | while (length > 12) |
---|
| 474 | { |
---|
| 475 | a += k[0]; |
---|
| 476 | b += k[1]; |
---|
| 477 | c += k[2]; |
---|
| 478 | mix(a,b,c); |
---|
| 479 | length -= 12; |
---|
| 480 | k += 3; |
---|
| 481 | } |
---|
| 482 | |
---|
| 483 | /*----------------------------- handle the last (probably partial) block */ |
---|
| 484 | /* |
---|
| 485 | * "k[2]<<8" actually reads beyond the end of the string, but |
---|
| 486 | * then shifts out the part it's not allowed to read. Because the |
---|
| 487 | * string is aligned, the illegal read is in the same word as the |
---|
| 488 | * rest of the string. Every machine with memory protection I've seen |
---|
| 489 | * does it on word boundaries, so is OK with this. But VALGRIND will |
---|
| 490 | * still catch it and complain. The masking trick does make the hash |
---|
| 491 | * noticably faster for short strings (like English words). |
---|
| 492 | */ |
---|
| 493 | #ifndef VALGRIND |
---|
| 494 | |
---|
| 495 | switch(length) |
---|
| 496 | { |
---|
| 497 | case 12: c+=k[2]; b+=k[1]; a+=k[0]; break; |
---|
| 498 | case 11: c+=k[2]&0xffffff00; b+=k[1]; a+=k[0]; break; |
---|
| 499 | case 10: c+=k[2]&0xffff0000; b+=k[1]; a+=k[0]; break; |
---|
| 500 | case 9 : c+=k[2]&0xff000000; b+=k[1]; a+=k[0]; break; |
---|
| 501 | case 8 : b+=k[1]; a+=k[0]; break; |
---|
| 502 | case 7 : b+=k[1]&0xffffff00; a+=k[0]; break; |
---|
| 503 | case 6 : b+=k[1]&0xffff0000; a+=k[0]; break; |
---|
| 504 | case 5 : b+=k[1]&0xff000000; a+=k[0]; break; |
---|
| 505 | case 4 : a+=k[0]; break; |
---|
| 506 | case 3 : a+=k[0]&0xffffff00; break; |
---|
| 507 | case 2 : a+=k[0]&0xffff0000; break; |
---|
| 508 | case 1 : a+=k[0]&0xff000000; break; |
---|
| 509 | case 0 : return c; /* zero length strings require no mixing */ |
---|
| 510 | } |
---|
| 511 | |
---|
| 512 | #else /* make valgrind happy */ |
---|
| 513 | |
---|
| 514 | k8 = (const uint8_t *)k; |
---|
| 515 | switch(length) /* all the case statements fall through */ |
---|
| 516 | { |
---|
| 517 | case 12: c+=k[2]; b+=k[1]; a+=k[0]; break; |
---|
| 518 | case 11: c+=((uint32_t)k8[10])<<8; /* fall through */ |
---|
| 519 | case 10: c+=((uint32_t)k8[9])<<16; /* fall through */ |
---|
| 520 | case 9 : c+=((uint32_t)k8[8])<<24; /* fall through */ |
---|
| 521 | case 8 : b+=k[1]; a+=k[0]; break; |
---|
| 522 | case 7 : b+=((uint32_t)k8[6])<<8; /* fall through */ |
---|
| 523 | case 6 : b+=((uint32_t)k8[5])<<16; /* fall through */ |
---|
| 524 | case 5 : b+=((uint32_t)k8[4])<<24; /* fall through */ |
---|
| 525 | case 4 : a+=k[0]; break; |
---|
| 526 | case 3 : a+=((uint32_t)k8[2])<<8; /* fall through */ |
---|
| 527 | case 2 : a+=((uint32_t)k8[1])<<16; /* fall through */ |
---|
| 528 | case 1 : a+=((uint32_t)k8[0])<<24; break; |
---|
| 529 | case 0 : return c; |
---|
| 530 | } |
---|
| 531 | |
---|
| 532 | #endif /* !VALGRIND */ |
---|
| 533 | |
---|
| 534 | } else { /* need to read the key one byte at a time */ |
---|
| 535 | const uint8_t *k = (const uint8_t *)key; |
---|
| 536 | |
---|
| 537 | /*--------------- all but the last block: affect some 32 bits of (a,b,c) */ |
---|
| 538 | while (length > 12) |
---|
| 539 | { |
---|
| 540 | a += ((uint32_t)k[0])<<24; |
---|
| 541 | a += ((uint32_t)k[1])<<16; |
---|
| 542 | a += ((uint32_t)k[2])<<8; |
---|
| 543 | a += ((uint32_t)k[3]); |
---|
| 544 | b += ((uint32_t)k[4])<<24; |
---|
| 545 | b += ((uint32_t)k[5])<<16; |
---|
| 546 | b += ((uint32_t)k[6])<<8; |
---|
| 547 | b += ((uint32_t)k[7]); |
---|
| 548 | c += ((uint32_t)k[8])<<24; |
---|
| 549 | c += ((uint32_t)k[9])<<16; |
---|
| 550 | c += ((uint32_t)k[10])<<8; |
---|
| 551 | c += ((uint32_t)k[11]); |
---|
| 552 | mix(a,b,c); |
---|
| 553 | length -= 12; |
---|
| 554 | k += 12; |
---|
| 555 | } |
---|
| 556 | |
---|
| 557 | /*-------------------------------- last block: affect all 32 bits of (c) */ |
---|
| 558 | switch(length) /* all the case statements fall through */ |
---|
| 559 | { |
---|
| 560 | case 12: c+=k[11]; |
---|
| 561 | case 11: c+=((uint32_t)k[10])<<8; |
---|
| 562 | case 10: c+=((uint32_t)k[9])<<16; |
---|
| 563 | case 9 : c+=((uint32_t)k[8])<<24; |
---|
| 564 | case 8 : b+=k[7]; |
---|
| 565 | case 7 : b+=((uint32_t)k[6])<<8; |
---|
| 566 | case 6 : b+=((uint32_t)k[5])<<16; |
---|
| 567 | case 5 : b+=((uint32_t)k[4])<<24; |
---|
| 568 | case 4 : a+=k[3]; |
---|
| 569 | case 3 : a+=((uint32_t)k[2])<<8; |
---|
| 570 | case 2 : a+=((uint32_t)k[1])<<16; |
---|
| 571 | case 1 : a+=((uint32_t)k[0])<<24; |
---|
| 572 | break; |
---|
| 573 | case 0 : return c; |
---|
| 574 | } |
---|
| 575 | } |
---|
| 576 | |
---|
| 577 | final(a,b,c); |
---|
| 578 | return c; |
---|
| 579 | } |
---|
| 580 | #endif /*WORDS_BIGENDIAN*/ |
---|
| 581 | |
---|
| 582 | /* |
---|
| 583 | ------------------------------------------------------------------------------- |
---|
| 584 | hashlittle() -- hash a variable-length key into a 32-bit value |
---|
| 585 | k : the key (the unaligned variable-length array of bytes) |
---|
| 586 | length : the length of the key, counting by bytes |
---|
| 587 | initval : can be any 4-byte value |
---|
| 588 | Returns a 32-bit value. Every bit of the key affects every bit of |
---|
| 589 | the return value. Two keys differing by one or two bits will have |
---|
| 590 | totally different hash values. |
---|
| 591 | |
---|
| 592 | The best hash table sizes are powers of 2. There is no need to do |
---|
| 593 | mod a prime (mod is sooo slow!). If you need less than 32 bits, |
---|
| 594 | use a bitmask. For example, if you need only 10 bits, do |
---|
| 595 | h = (h & hashmask(10)); |
---|
| 596 | In which case, the hash table should have hashsize(10) elements. |
---|
| 597 | |
---|
| 598 | If you are hashing n strings (uint8_t **)k, do it like this: |
---|
| 599 | for (i=0, h=0; i<n; ++i) h = hashlittle( k[i], len[i], h); |
---|
| 600 | |
---|
| 601 | By Bob Jenkins, 2006. bob_jenkins@burtleburtle.net. You may use this |
---|
| 602 | code any way you wish, private, educational, or commercial. It's free. |
---|
| 603 | |
---|
| 604 | Use for hash table lookup, or anything where one collision in 2^^32 is |
---|
| 605 | acceptable. Do NOT use for cryptographic purposes. |
---|
| 606 | ------------------------------------------------------------------------------- |
---|
| 607 | */ |
---|
| 608 | |
---|
| 609 | static uint32_t |
---|
| 610 | hashlittle( const void *key, size_t length, uint32_t initval) |
---|
| 611 | { |
---|
| 612 | uint32_t a,b,c; /* internal state */ |
---|
| 613 | union { const void *ptr; size_t i; } u; /* needed for Mac Powerbook G4 */ |
---|
| 614 | |
---|
| 615 | /* Set up the internal state */ |
---|
| 616 | a = b = c = 0xdeadbeef + ((uint32_t)length) + initval; |
---|
| 617 | |
---|
| 618 | u.ptr = key; |
---|
| 619 | if (HASH_LITTLE_ENDIAN && ((u.i & 0x3) == 0)) { |
---|
| 620 | const uint32_t *k = (const uint32_t *)key; /* read 32-bit chunks */ |
---|
| 621 | const uint8_t *k8; |
---|
| 622 | |
---|
| 623 | /*------ all but last block: aligned reads and affect 32 bits of (a,b,c) */ |
---|
| 624 | while (length > 12) |
---|
| 625 | { |
---|
| 626 | a += k[0]; |
---|
| 627 | b += k[1]; |
---|
| 628 | c += k[2]; |
---|
| 629 | mix(a,b,c); |
---|
| 630 | length -= 12; |
---|
| 631 | k += 3; |
---|
| 632 | } |
---|
| 633 | |
---|
| 634 | /*----------------------------- handle the last (probably partial) block */ |
---|
| 635 | /* |
---|
| 636 | * "k[2]&0xffffff" actually reads beyond the end of the string, but |
---|
| 637 | * then masks off the part it's not allowed to read. Because the |
---|
| 638 | * string is aligned, the masked-off tail is in the same word as the |
---|
| 639 | * rest of the string. Every machine with memory protection I've seen |
---|
| 640 | * does it on word boundaries, so is OK with this. But VALGRIND will |
---|
| 641 | * still catch it and complain. The masking trick does make the hash |
---|
| 642 | * noticably faster for short strings (like English words). |
---|
| 643 | */ |
---|
| 644 | #ifndef VALGRIND |
---|
| 645 | |
---|
| 646 | switch(length) |
---|
| 647 | { |
---|
| 648 | case 12: c+=k[2]; b+=k[1]; a+=k[0]; break; |
---|
| 649 | case 11: c+=k[2]&0xffffff; b+=k[1]; a+=k[0]; break; |
---|
| 650 | case 10: c+=k[2]&0xffff; b+=k[1]; a+=k[0]; break; |
---|
| 651 | case 9 : c+=k[2]&0xff; b+=k[1]; a+=k[0]; break; |
---|
| 652 | case 8 : b+=k[1]; a+=k[0]; break; |
---|
| 653 | case 7 : b+=k[1]&0xffffff; a+=k[0]; break; |
---|
| 654 | case 6 : b+=k[1]&0xffff; a+=k[0]; break; |
---|
| 655 | case 5 : b+=k[1]&0xff; a+=k[0]; break; |
---|
| 656 | case 4 : a+=k[0]; break; |
---|
| 657 | case 3 : a+=k[0]&0xffffff; break; |
---|
| 658 | case 2 : a+=k[0]&0xffff; break; |
---|
| 659 | case 1 : a+=k[0]&0xff; break; |
---|
| 660 | case 0 : return c; /* zero length strings require no mixing */ |
---|
| 661 | } |
---|
| 662 | |
---|
| 663 | #else /* make valgrind happy */ |
---|
| 664 | |
---|
| 665 | k8 = (const uint8_t *)k; |
---|
| 666 | switch(length) |
---|
| 667 | { |
---|
| 668 | case 12: c+=k[2]; b+=k[1]; a+=k[0]; break; |
---|
| 669 | case 11: c+=((uint32_t)k8[10])<<16; /* fall through */ |
---|
| 670 | case 10: c+=((uint32_t)k8[9])<<8; /* fall through */ |
---|
| 671 | case 9 : c+=k8[8]; /* fall through */ |
---|
| 672 | case 8 : b+=k[1]; a+=k[0]; break; |
---|
| 673 | case 7 : b+=((uint32_t)k8[6])<<16; /* fall through */ |
---|
| 674 | case 6 : b+=((uint32_t)k8[5])<<8; /* fall through */ |
---|
| 675 | case 5 : b+=k8[4]; /* fall through */ |
---|
| 676 | case 4 : a+=k[0]; break; |
---|
| 677 | case 3 : a+=((uint32_t)k8[2])<<16; /* fall through */ |
---|
| 678 | case 2 : a+=((uint32_t)k8[1])<<8; /* fall through */ |
---|
| 679 | case 1 : a+=k8[0]; break; |
---|
| 680 | case 0 : return c; |
---|
| 681 | } |
---|
| 682 | |
---|
| 683 | #endif /* !valgrind */ |
---|
| 684 | |
---|
| 685 | } else if (HASH_LITTLE_ENDIAN && ((u.i & 0x1) == 0)) { |
---|
| 686 | const uint16_t *k = (const uint16_t *)key; /* read 16-bit chunks */ |
---|
| 687 | const uint8_t *k8; |
---|
| 688 | |
---|
| 689 | /*--------------- all but last block: aligned reads and different mixing */ |
---|
| 690 | while (length > 12) |
---|
| 691 | { |
---|
| 692 | a += k[0] + (((uint32_t)k[1])<<16); |
---|
| 693 | b += k[2] + (((uint32_t)k[3])<<16); |
---|
| 694 | c += k[4] + (((uint32_t)k[5])<<16); |
---|
| 695 | mix(a,b,c); |
---|
| 696 | length -= 12; |
---|
| 697 | k += 6; |
---|
| 698 | } |
---|
| 699 | |
---|
| 700 | /*----------------------------- handle the last (probably partial) block */ |
---|
| 701 | k8 = (const uint8_t *)k; |
---|
| 702 | switch(length) |
---|
| 703 | { |
---|
| 704 | case 12: c+=k[4]+(((uint32_t)k[5])<<16); |
---|
| 705 | b+=k[2]+(((uint32_t)k[3])<<16); |
---|
| 706 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 707 | break; |
---|
| 708 | case 11: c+=((uint32_t)k8[10])<<16; /* fall through */ |
---|
| 709 | case 10: c+=k[4]; |
---|
| 710 | b+=k[2]+(((uint32_t)k[3])<<16); |
---|
| 711 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 712 | break; |
---|
| 713 | case 9 : c+=k8[8]; /* fall through */ |
---|
| 714 | case 8 : b+=k[2]+(((uint32_t)k[3])<<16); |
---|
| 715 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 716 | break; |
---|
| 717 | case 7 : b+=((uint32_t)k8[6])<<16; /* fall through */ |
---|
| 718 | case 6 : b+=k[2]; |
---|
| 719 | a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 720 | break; |
---|
| 721 | case 5 : b+=k8[4]; /* fall through */ |
---|
| 722 | case 4 : a+=k[0]+(((uint32_t)k[1])<<16); |
---|
| 723 | break; |
---|
| 724 | case 3 : a+=((uint32_t)k8[2])<<16; /* fall through */ |
---|
| 725 | case 2 : a+=k[0]; |
---|
| 726 | break; |
---|
| 727 | case 1 : a+=k8[0]; |
---|
| 728 | break; |
---|
| 729 | case 0 : return c; /* zero length requires no mixing */ |
---|
| 730 | } |
---|
| 731 | |
---|
| 732 | } else { /* need to read the key one byte at a time */ |
---|
| 733 | const uint8_t *k = (const uint8_t *)key; |
---|
| 734 | |
---|
| 735 | /*--------------- all but the last block: affect some 32 bits of (a,b,c) */ |
---|
| 736 | while (length > 12) |
---|
| 737 | { |
---|
| 738 | a += k[0]; |
---|
| 739 | a += ((uint32_t)k[1])<<8; |
---|
| 740 | a += ((uint32_t)k[2])<<16; |
---|
| 741 | a += ((uint32_t)k[3])<<24; |
---|
| 742 | b += k[4]; |
---|
| 743 | b += ((uint32_t)k[5])<<8; |
---|
| 744 | b += ((uint32_t)k[6])<<16; |
---|
| 745 | b += ((uint32_t)k[7])<<24; |
---|
| 746 | c += k[8]; |
---|
| 747 | c += ((uint32_t)k[9])<<8; |
---|
| 748 | c += ((uint32_t)k[10])<<16; |
---|
| 749 | c += ((uint32_t)k[11])<<24; |
---|
| 750 | mix(a,b,c); |
---|
| 751 | length -= 12; |
---|
| 752 | k += 12; |
---|
| 753 | } |
---|
| 754 | |
---|
| 755 | /*-------------------------------- last block: affect all 32 bits of (c) */ |
---|
| 756 | switch(length) /* all the case statements fall through */ |
---|
| 757 | { |
---|
| 758 | case 12: c+=((uint32_t)k[11])<<24; |
---|
| 759 | case 11: c+=((uint32_t)k[10])<<16; |
---|
| 760 | case 10: c+=((uint32_t)k[9])<<8; |
---|
| 761 | case 9 : c+=k[8]; |
---|
| 762 | case 8 : b+=((uint32_t)k[7])<<24; |
---|
| 763 | case 7 : b+=((uint32_t)k[6])<<16; |
---|
| 764 | case 6 : b+=((uint32_t)k[5])<<8; |
---|
| 765 | case 5 : b+=k[4]; |
---|
| 766 | case 4 : a+=((uint32_t)k[3])<<24; |
---|
| 767 | case 3 : a+=((uint32_t)k[2])<<16; |
---|
| 768 | case 2 : a+=((uint32_t)k[1])<<8; |
---|
| 769 | case 1 : a+=k[0]; |
---|
| 770 | break; |
---|
| 771 | case 0 : return c; |
---|
| 772 | } |
---|
| 773 | } |
---|
| 774 | |
---|
| 775 | final(a,b,c); |
---|
| 776 | return c; |
---|
| 777 | } |
---|
| 778 | |
---|
| 779 | |
---|
| 780 | /* |
---|
| 781 | * hash_fast(key, length, initval) |
---|
| 782 | * Wrapper that calls either hashlittle or hashbig, depending on endianness. |
---|
| 783 | */ |
---|
| 784 | uint32_t |
---|
| 785 | hash_fast( const void *key, size_t length) { |
---|
| 786 | #define NC_ARBITRARY_UINT (992099683U) |
---|
| 787 | #ifndef WORDS_BIGENDIAN |
---|
| 788 | return hashlittle(key, length, NC_ARBITRARY_UINT); |
---|
| 789 | #else |
---|
| 790 | return hashbig(key, length, NC_ARBITRARY_UINT); |
---|
| 791 | #endif |
---|
| 792 | } |
---|
| 793 | |
---|
| 794 | #ifdef SELF_TEST |
---|
| 795 | /* used for timings */ |
---|
| 796 | void driver1() |
---|
| 797 | { |
---|
| 798 | uint8_t buf[256]; |
---|
| 799 | uint32_t i; |
---|
| 800 | uint32_t h=0; |
---|
| 801 | time_t a,z; |
---|
| 802 | |
---|
| 803 | time(&a); |
---|
| 804 | for (i=0; i<256; ++i) buf[i] = 'x'; |
---|
| 805 | for (i=0; i<1; ++i) |
---|
| 806 | { |
---|
| 807 | h = hashlittle(&buf[0],1,h); |
---|
| 808 | } |
---|
| 809 | time(&z); |
---|
| 810 | if (z-a > 0) printf("time %d %.8x\n", z-a, h); |
---|
| 811 | } |
---|
| 812 | |
---|
| 813 | /* check that every input bit changes every output bit half the time */ |
---|
| 814 | #define HASHSTATE 1 |
---|
| 815 | #define HASHLEN 1 |
---|
| 816 | #define MAXPAIR 60 |
---|
| 817 | #define MAXLEN 70 |
---|
| 818 | void driver2() |
---|
| 819 | { |
---|
| 820 | uint8_t qa[MAXLEN+1], qb[MAXLEN+2], *a = &qa[0], *b = &qb[1]; |
---|
| 821 | uint32_t c[HASHSTATE], d[HASHSTATE], i=0, j=0, k, l, m=0, z; |
---|
| 822 | uint32_t e[HASHSTATE],f[HASHSTATE],g[HASHSTATE],h[HASHSTATE]; |
---|
| 823 | uint32_t x[HASHSTATE],y[HASHSTATE]; |
---|
| 824 | uint32_t hlen; |
---|
| 825 | |
---|
| 826 | printf("No more than %d trials should ever be needed \n",MAXPAIR/2); |
---|
| 827 | for (hlen=0; hlen < MAXLEN; ++hlen) |
---|
| 828 | { |
---|
| 829 | z=0; |
---|
| 830 | for (i=0; i<hlen; ++i) /*----------------------- for each input byte, */ |
---|
| 831 | { |
---|
| 832 | for (j=0; j<8; ++j) /*------------------------ for each input bit, */ |
---|
| 833 | { |
---|
| 834 | for (m=1; m<8; ++m) /*------------ for serveral possible initvals, */ |
---|
| 835 | { |
---|
| 836 | for (l=0; l<HASHSTATE; ++l) |
---|
| 837 | e[l]=f[l]=g[l]=h[l]=x[l]=y[l]=~((uint32_t)0); |
---|
| 838 | |
---|
| 839 | /*---- check that every output bit is affected by that input bit */ |
---|
| 840 | for (k=0; k<MAXPAIR; k+=2) |
---|
| 841 | { |
---|
| 842 | uint32_t finished=1; |
---|
| 843 | /* keys have one bit different */ |
---|
| 844 | for (l=0; l<hlen+1; ++l) {a[l] = b[l] = (uint8_t)0;} |
---|
| 845 | /* have a and b be two keys differing in only one bit */ |
---|
| 846 | a[i] ^= (k<<j); |
---|
| 847 | a[i] ^= (k>>(8-j)); |
---|
| 848 | c[0] = hashlittle(a, hlen, m); |
---|
| 849 | b[i] ^= ((k+1)<<j); |
---|
| 850 | b[i] ^= ((k+1)>>(8-j)); |
---|
| 851 | d[0] = hashlittle(b, hlen, m); |
---|
| 852 | /* check every bit is 1, 0, set, and not set at least once */ |
---|
| 853 | for (l=0; l<HASHSTATE; ++l) |
---|
| 854 | { |
---|
| 855 | e[l] &= (c[l]^d[l]); |
---|
| 856 | f[l] &= ~(c[l]^d[l]); |
---|
| 857 | g[l] &= c[l]; |
---|
| 858 | h[l] &= ~c[l]; |
---|
| 859 | x[l] &= d[l]; |
---|
| 860 | y[l] &= ~d[l]; |
---|
| 861 | if (e[l]|f[l]|g[l]|h[l]|x[l]|y[l]) finished=0; |
---|
| 862 | } |
---|
| 863 | if (finished) break; |
---|
| 864 | } |
---|
| 865 | if (k>z) z=k; |
---|
| 866 | if (k==MAXPAIR) |
---|
| 867 | { |
---|
| 868 | printf("Some bit didn't change: "); |
---|
| 869 | printf("%.8x %.8x %.8x %.8x %.8x %.8x ", |
---|
| 870 | e[0],f[0],g[0],h[0],x[0],y[0]); |
---|
| 871 | printf("i %d j %d m %d len %d\n", i, j, m, hlen); |
---|
| 872 | } |
---|
| 873 | if (z==MAXPAIR) goto done; |
---|
| 874 | } |
---|
| 875 | } |
---|
| 876 | } |
---|
| 877 | done: |
---|
| 878 | if (z < MAXPAIR) |
---|
| 879 | { |
---|
| 880 | printf("Mix success %2d bytes %2d initvals ",i,m); |
---|
| 881 | printf("required %d trials\n", z/2); |
---|
| 882 | } |
---|
| 883 | } |
---|
| 884 | printf("\n"); |
---|
| 885 | } |
---|
| 886 | |
---|
| 887 | /* Check for reading beyond the end of the buffer and alignment problems */ |
---|
| 888 | void driver3() |
---|
| 889 | { |
---|
| 890 | uint8_t buf[MAXLEN+20], *b; |
---|
| 891 | uint32_t len; |
---|
| 892 | uint8_t q[] = "This is the time for all good men to come to the aid of their country..."; |
---|
| 893 | uint32_t h; |
---|
| 894 | uint8_t qq[] = "xThis is the time for all good men to come to the aid of their country..."; |
---|
| 895 | uint32_t i; |
---|
| 896 | uint8_t qqq[] = "xxThis is the time for all good men to come to the aid of their country..."; |
---|
| 897 | uint32_t j; |
---|
| 898 | uint8_t qqqq[] = "xxxThis is the time for all good men to come to the aid of their country..."; |
---|
| 899 | uint32_t ref,x,y; |
---|
| 900 | uint8_t *p; |
---|
| 901 | |
---|
| 902 | printf("Endianness. These lines should all be the same (for values filled in):\n"); |
---|
| 903 | printf("%.8x %.8x %.8x\n", |
---|
| 904 | hashword((const uint32_t *)q, (sizeof(q)-1)/4, 13), |
---|
| 905 | hashword((const uint32_t *)q, (sizeof(q)-5)/4, 13), |
---|
| 906 | hashword((const uint32_t *)q, (sizeof(q)-9)/4, 13)); |
---|
| 907 | p = q; |
---|
| 908 | printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n", |
---|
| 909 | hashlittle(p, sizeof(q)-1, 13), hashlittle(p, sizeof(q)-2, 13), |
---|
| 910 | hashlittle(p, sizeof(q)-3, 13), hashlittle(p, sizeof(q)-4, 13), |
---|
| 911 | hashlittle(p, sizeof(q)-5, 13), hashlittle(p, sizeof(q)-6, 13), |
---|
| 912 | hashlittle(p, sizeof(q)-7, 13), hashlittle(p, sizeof(q)-8, 13), |
---|
| 913 | hashlittle(p, sizeof(q)-9, 13), hashlittle(p, sizeof(q)-10, 13), |
---|
| 914 | hashlittle(p, sizeof(q)-11, 13), hashlittle(p, sizeof(q)-12, 13)); |
---|
| 915 | p = &qq[1]; |
---|
| 916 | printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n", |
---|
| 917 | hashlittle(p, sizeof(q)-1, 13), hashlittle(p, sizeof(q)-2, 13), |
---|
| 918 | hashlittle(p, sizeof(q)-3, 13), hashlittle(p, sizeof(q)-4, 13), |
---|
| 919 | hashlittle(p, sizeof(q)-5, 13), hashlittle(p, sizeof(q)-6, 13), |
---|
| 920 | hashlittle(p, sizeof(q)-7, 13), hashlittle(p, sizeof(q)-8, 13), |
---|
| 921 | hashlittle(p, sizeof(q)-9, 13), hashlittle(p, sizeof(q)-10, 13), |
---|
| 922 | hashlittle(p, sizeof(q)-11, 13), hashlittle(p, sizeof(q)-12, 13)); |
---|
| 923 | p = &qqq[2]; |
---|
| 924 | printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n", |
---|
| 925 | hashlittle(p, sizeof(q)-1, 13), hashlittle(p, sizeof(q)-2, 13), |
---|
| 926 | hashlittle(p, sizeof(q)-3, 13), hashlittle(p, sizeof(q)-4, 13), |
---|
| 927 | hashlittle(p, sizeof(q)-5, 13), hashlittle(p, sizeof(q)-6, 13), |
---|
| 928 | hashlittle(p, sizeof(q)-7, 13), hashlittle(p, sizeof(q)-8, 13), |
---|
| 929 | hashlittle(p, sizeof(q)-9, 13), hashlittle(p, sizeof(q)-10, 13), |
---|
| 930 | hashlittle(p, sizeof(q)-11, 13), hashlittle(p, sizeof(q)-12, 13)); |
---|
| 931 | p = &qqqq[3]; |
---|
| 932 | printf("%.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x %.8x\n", |
---|
| 933 | hashlittle(p, sizeof(q)-1, 13), hashlittle(p, sizeof(q)-2, 13), |
---|
| 934 | hashlittle(p, sizeof(q)-3, 13), hashlittle(p, sizeof(q)-4, 13), |
---|
| 935 | hashlittle(p, sizeof(q)-5, 13), hashlittle(p, sizeof(q)-6, 13), |
---|
| 936 | hashlittle(p, sizeof(q)-7, 13), hashlittle(p, sizeof(q)-8, 13), |
---|
| 937 | hashlittle(p, sizeof(q)-9, 13), hashlittle(p, sizeof(q)-10, 13), |
---|
| 938 | hashlittle(p, sizeof(q)-11, 13), hashlittle(p, sizeof(q)-12, 13)); |
---|
| 939 | printf("\n"); |
---|
| 940 | |
---|
| 941 | /* check that hashlittle2 and hashlittle produce the same results */ |
---|
| 942 | i=47; j=0; |
---|
| 943 | hashlittle2(q, sizeof(q), &i, &j); |
---|
| 944 | if (hashlittle(q, sizeof(q), 47) != i) |
---|
| 945 | printf("hashlittle2 and hashlittle mismatch\n"); |
---|
| 946 | |
---|
| 947 | /* check that hashword2 and hashword produce the same results */ |
---|
| 948 | len = 0xdeadbeef; |
---|
| 949 | i=47, j=0; |
---|
| 950 | hashword2(&len, 1, &i, &j); |
---|
| 951 | if (hashword(&len, 1, 47) != i) |
---|
| 952 | printf("hashword2 and hashword mismatch %x %x\n", |
---|
| 953 | i, hashword(&len, 1, 47)); |
---|
| 954 | |
---|
| 955 | /* check hashlittle doesn't read before or after the ends of the string */ |
---|
| 956 | for (h=0, b=buf+1; h<8; ++h, ++b) |
---|
| 957 | { |
---|
| 958 | for (i=0; i<MAXLEN; ++i) |
---|
| 959 | { |
---|
| 960 | len = i; |
---|
| 961 | for (j=0; j<i; ++j) *(b+j)=0; |
---|
| 962 | |
---|
| 963 | /* these should all be equal */ |
---|
| 964 | ref = hashlittle(b, len, (uint32_t)1); |
---|
| 965 | *(b+i)=(uint8_t)~0; |
---|
| 966 | *(b-1)=(uint8_t)~0; |
---|
| 967 | x = hashlittle(b, len, (uint32_t)1); |
---|
| 968 | y = hashlittle(b, len, (uint32_t)1); |
---|
| 969 | if ((ref != x) || (ref != y)) |
---|
| 970 | { |
---|
| 971 | printf("alignment error: %.8x %.8x %.8x %d %d\n",ref,x,y, |
---|
| 972 | h, i); |
---|
| 973 | } |
---|
| 974 | } |
---|
| 975 | } |
---|
| 976 | } |
---|
| 977 | |
---|
| 978 | /* check for problems with nulls */ |
---|
| 979 | void driver4() |
---|
| 980 | { |
---|
| 981 | uint8_t buf[1]; |
---|
| 982 | uint32_t h,i,state[HASHSTATE]; |
---|
| 983 | |
---|
| 984 | |
---|
| 985 | buf[0] = ~0; |
---|
| 986 | for (i=0; i<HASHSTATE; ++i) state[i] = 1; |
---|
| 987 | printf("These should all be different\n"); |
---|
| 988 | for (i=0, h=0; i<8; ++i) |
---|
| 989 | { |
---|
| 990 | h = hashlittle(buf, 0, h); |
---|
| 991 | printf("%2ld 0-byte strings, hash is %.8x\n", i, h); |
---|
| 992 | } |
---|
| 993 | } |
---|
| 994 | |
---|
| 995 | void driver5() |
---|
| 996 | { |
---|
| 997 | uint32_t b,c; |
---|
| 998 | b=0, c=0, hashlittle2("", 0, &c, &b); |
---|
| 999 | printf("hash is %.8lx %.8lx\n", c, b); /* deadbeef deadbeef */ |
---|
| 1000 | b=0xdeadbeef, c=0, hashlittle2("", 0, &c, &b); |
---|
| 1001 | printf("hash is %.8lx %.8lx\n", c, b); /* bd5b7dde deadbeef */ |
---|
| 1002 | b=0xdeadbeef, c=0xdeadbeef, hashlittle2("", 0, &c, &b); |
---|
| 1003 | printf("hash is %.8lx %.8lx\n", c, b); /* 9c093ccd bd5b7dde */ |
---|
| 1004 | b=0, c=0, hashlittle2("Four score and seven years ago", 30, &c, &b); |
---|
| 1005 | printf("hash is %.8lx %.8lx\n", c, b); /* 17770551 ce7226e6 */ |
---|
| 1006 | b=1, c=0, hashlittle2("Four score and seven years ago", 30, &c, &b); |
---|
| 1007 | printf("hash is %.8lx %.8lx\n", c, b); /* e3607cae bd371de4 */ |
---|
| 1008 | b=0, c=1, hashlittle2("Four score and seven years ago", 30, &c, &b); |
---|
| 1009 | printf("hash is %.8lx %.8lx\n", c, b); /* cd628161 6cbea4b3 */ |
---|
| 1010 | c = hashlittle("Four score and seven years ago", 30, 0); |
---|
| 1011 | printf("hash is %.8lx\n", c); /* 17770551 */ |
---|
| 1012 | c = hashlittle("Four score and seven years ago", 30, 1); |
---|
| 1013 | printf("hash is %.8lx\n", c); /* cd628161 */ |
---|
| 1014 | } |
---|
| 1015 | |
---|
| 1016 | |
---|
| 1017 | #endif /* SELF_TEST */ |
---|