• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2 xxHash - Fast Hash algorithm
3 Copyright (C) 2012-2014, Yann Collet.
4 BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
5 
6 Redistribution and use in source and binary forms, with or without
7 modification, are permitted provided that the following conditions are
8 met:
9 
10 * Redistributions of source code must retain the above copyright
11 notice, this list of conditions and the following disclaimer.
12 * Redistributions in binary form must reproduce the above
13 copyright notice, this list of conditions and the following disclaimer
14 in the documentation and/or other materials provided with the
15 distribution.
16 
17 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
21 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
22 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
23 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
27 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 
29 You can contact the author at :
30 - xxHash source repository : http://code.google.com/p/xxhash/
31 */
32 
33 
34 //**************************************
35 // Tuning parameters
36 //**************************************
37 // Unaligned memory access is automatically enabled for "common" CPU, such as x86.
38 // For others CPU, the compiler will be more cautious, and insert extra code to ensure aligned access is respected.
39 // If you know your target CPU supports unaligned memory access, you want to force this option manually to improve performance.
40 // You can also enable this parameter if you know your input data will always be aligned (boundaries of 4, for uint32_t).
41 #if defined(__ARM_FEATURE_UNALIGNED) || defined(__i386) || defined(_M_IX86) || defined(__x86_64__) || defined(_M_X64)
42 #  define XXH_USE_UNALIGNED_ACCESS 1
43 #endif
44 
45 // XXH_ACCEPT_NULL_INPUT_POINTER :
46 // If the input pointer is a null pointer, xxHash default behavior is to trigger a memory access error, since it is a bad pointer.
47 // When this option is enabled, xxHash output for null input pointers will be the same as a null-length input.
48 // This option has a very small performance cost (only measurable on small inputs).
49 // By default, this option is disabled. To enable it, uncomment below define :
50 //#define XXH_ACCEPT_NULL_INPUT_POINTER 1
51 
52 // XXH_FORCE_NATIVE_FORMAT :
53 // By default, xxHash library provides endian-independant Hash values, based on little-endian convention.
54 // Results are therefore identical for little-endian and big-endian CPU.
55 // This comes at a performance cost for big-endian CPU, since some swapping is required to emulate little-endian format.
56 // Should endian-independance be of no importance for your application, you may set the #define below to 1.
57 // It will improve speed for Big-endian CPU.
58 // This option has no impact on Little_Endian CPU.
59 #define XXH_FORCE_NATIVE_FORMAT 0
60 
61 
62 //**************************************
63 // Includes & Memory related functions
64 //**************************************
65 #include "xxhash.h"
66 #include <stdlib.h>
67 #include <string.h>
68 
69 
70 #if defined(__GNUC__)  && !defined(XXH_USE_UNALIGNED_ACCESS)
71 #  define _PACKED __attribute__ ((packed))
72 #else
73 #  define _PACKED
74 #endif
75 
76 #if !defined(XXH_USE_UNALIGNED_ACCESS) && !defined(__GNUC__)
77 #  ifdef __IBMC__
78 #    pragma pack(1)
79 #  else
80 #    pragma pack(push, 1)
81 #  endif
82 #endif
83 
84 typedef struct _uint32_t_S { uint32_t v; } _PACKED uint32_t_S;
85 
86 #if !defined(XXH_USE_UNALIGNED_ACCESS) && !defined(__GNUC__)
87 #  pragma pack(pop)
88 #endif
89 
90 #define A32(x) (((uint32_t_S *)(x))->v)
91 
92 
93 //***************************************
94 // Compiler-specific Functions and Macros
95 //***************************************
96 #define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
97 
98 // Note : although _rotl exists for minGW (GCC under windows), performance seems poor
99 #if defined(_MSC_VER)
100 #  define XXH_rotl32(x,r) _rotl(x,r)
101 #else
102 #  define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r)))
103 #endif
104 
105 #if defined(_MSC_VER)     // Visual Studio
106 #  define XXH_swap32 _byteswap_ulong
107 #elif GCC_VERSION >= 403
108 #  define XXH_swap32 __builtin_bswap32
109 #else
XXH_swap32(uint32_t x)110 static inline uint32_t XXH_swap32 (uint32_t x)
111 {
112     return  ((x << 24) & 0xff000000 ) |
113         ((x <<  8) & 0x00ff0000 ) |
114         ((x >>  8) & 0x0000ff00 ) |
115         ((x >> 24) & 0x000000ff );
116 }
117 #endif
118 
119 
120 //**************************************
121 // Constants
122 //**************************************
123 #define PRIME32_1   2654435761U
124 #define PRIME32_2   2246822519U
125 #define PRIME32_3   3266489917U
126 #define PRIME32_4    668265263U
127 #define PRIME32_5    374761393U
128 
129 
130 //**************************************
131 // Architecture Macros
132 //**************************************
133 typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
134 #ifndef XXH_CPU_LITTLE_ENDIAN   // It is possible to define XXH_CPU_LITTLE_ENDIAN externally, for example using a compiler switch
135     static const int one = 1;
136 #   define XXH_CPU_LITTLE_ENDIAN   (*(char*)(&one))
137 #endif
138 
139 
140 //**************************************
141 // Macros
142 //**************************************
143 #define XXH_STATIC_ASSERT(c)   { enum { XXH_static_assert = 1/(!!(c)) }; }    // use only *after* variable declarations
144 
145 
146 //****************************
147 // Memory reads
148 //****************************
149 typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment;
150 
XXH_readLE32_align(const uint32_t * ptr,XXH_endianess endian,XXH_alignment align)151 static uint32_t XXH_readLE32_align(const uint32_t* ptr, XXH_endianess endian, XXH_alignment align)
152 {
153     if (align==XXH_unaligned)
154         return endian==XXH_littleEndian ? A32(ptr) : XXH_swap32(A32(ptr));
155     else
156         return endian==XXH_littleEndian ? *ptr : XXH_swap32(*ptr);
157 }
158 
XXH_readLE32(const uint32_t * ptr,XXH_endianess endian)159 static uint32_t XXH_readLE32(const uint32_t* ptr, XXH_endianess endian) { return XXH_readLE32_align(ptr, endian, XXH_unaligned); }
160 
161 
162 //****************************
163 // Simple Hash Functions
164 //****************************
XXH32_endian_align(const void * input,int len,uint32_t seed,XXH_endianess endian,XXH_alignment align)165 static uint32_t XXH32_endian_align(const void* input, int len, uint32_t seed, XXH_endianess endian, XXH_alignment align)
166 {
167     const uint8_t *p = (const uint8_t *)input;
168     const uint8_t * const bEnd = p + len;
169     uint32_t h32;
170 
171 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER
172     if (p==NULL) { len=0; p=(const uint8_t *)(size_t)16; }
173 #endif
174 
175     if (len>=16)
176     {
177         const uint8_t * const limit = bEnd - 16;
178         uint32_t v1 = seed + PRIME32_1 + PRIME32_2;
179         uint32_t v2 = seed + PRIME32_2;
180         uint32_t v3 = seed + 0;
181         uint32_t v4 = seed - PRIME32_1;
182 
183         do
184         {
185             v1 += XXH_readLE32_align((const uint32_t*)p, endian, align) * PRIME32_2; v1 = XXH_rotl32(v1, 13); v1 *= PRIME32_1; p+=4;
186             v2 += XXH_readLE32_align((const uint32_t*)p, endian, align) * PRIME32_2; v2 = XXH_rotl32(v2, 13); v2 *= PRIME32_1; p+=4;
187             v3 += XXH_readLE32_align((const uint32_t*)p, endian, align) * PRIME32_2; v3 = XXH_rotl32(v3, 13); v3 *= PRIME32_1; p+=4;
188             v4 += XXH_readLE32_align((const uint32_t*)p, endian, align) * PRIME32_2; v4 = XXH_rotl32(v4, 13); v4 *= PRIME32_1; p+=4;
189         } while (p<=limit);
190 
191         h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18);
192     }
193     else
194     {
195         h32  = seed + PRIME32_5;
196     }
197 
198     h32 += (uint32_t) len;
199 
200     while (p<=bEnd-4)
201     {
202         h32 += XXH_readLE32_align((const uint32_t*)p, endian, align) * PRIME32_3;
203         h32  = XXH_rotl32(h32, 17) * PRIME32_4 ;
204         p+=4;
205     }
206 
207     while (p<bEnd)
208     {
209         h32 += (*p) * PRIME32_5;
210         h32 = XXH_rotl32(h32, 11) * PRIME32_1 ;
211         p++;
212     }
213 
214     h32 ^= h32 >> 15;
215     h32 *= PRIME32_2;
216     h32 ^= h32 >> 13;
217     h32 *= PRIME32_3;
218     h32 ^= h32 >> 16;
219 
220     return h32;
221 }
222 
223 
XXH32(const void * input,int len,uint32_t seed)224 uint32_t XXH32(const void* input, int len, uint32_t seed)
225 {
226 #if 0
227     // Simple version, good for code maintenance, but unfortunately slow for small inputs
228     void* state = XXH32_init(seed);
229     XXH32_update(state, input, len);
230     return XXH32_digest(state);
231 #else
232     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
233 
234 #  if !defined(XXH_USE_UNALIGNED_ACCESS)
235     if ((((size_t)input) & 3))   // Input is aligned, let's leverage the speed advantage
236     {
237         if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
238             return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
239         else
240             return XXH32_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned);
241     }
242 #  endif
243 
244     if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
245         return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned);
246     else
247         return XXH32_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned);
248 #endif
249 }
250 
251 
252 //****************************
253 // Advanced Hash Functions
254 //****************************
255 
XXH32_sizeofState(void)256 int XXH32_sizeofState(void)
257 {
258     XXH_STATIC_ASSERT(XXH32_SIZEOFSTATE >= sizeof(struct XXH_state32_t));   // A compilation error here means XXH32_SIZEOFSTATE is not large enough
259     return sizeof(struct XXH_state32_t);
260 }
261 
262 
XXH32_resetState(void * state_in,uint32_t seed)263 XXH_errorcode XXH32_resetState(void* state_in, uint32_t seed)
264 {
265     struct XXH_state32_t * state = (struct XXH_state32_t *) state_in;
266     state->seed = seed;
267     state->v1 = seed + PRIME32_1 + PRIME32_2;
268     state->v2 = seed + PRIME32_2;
269     state->v3 = seed + 0;
270     state->v4 = seed - PRIME32_1;
271     state->total_len = 0;
272     state->memsize = 0;
273     return XXH_OK;
274 }
275 
276 
XXH32_init(uint32_t seed)277 void* XXH32_init (uint32_t seed)
278 {
279     void *state = malloc (sizeof(struct XXH_state32_t));
280     XXH32_resetState(state, seed);
281     return state;
282 }
283 
284 
XXH32_update_endian(void * state_in,const void * input,int len,XXH_endianess endian)285 static XXH_errorcode XXH32_update_endian (void* state_in, const void* input, int len, XXH_endianess endian)
286 {
287     struct XXH_state32_t * state = (struct XXH_state32_t *) state_in;
288     const uint8_t *p = (const uint8_t *)input;
289     const uint8_t * const bEnd = p + len;
290 
291 #ifdef XXH_ACCEPT_NULL_INPUT_POINTER
292     if (input==NULL) return XXH_ERROR;
293 #endif
294 
295     state->total_len += len;
296 
297     if (state->memsize + len < 16)   // fill in tmp buffer
298     {
299         memcpy(state->memory + state->memsize, input, len);
300         state->memsize +=  len;
301         return XXH_OK;
302     }
303 
304     if (state->memsize)   // some data left from previous update
305     {
306         memcpy(state->memory + state->memsize, input, 16-state->memsize);
307         {
308             const uint32_t* p32 = (const uint32_t*)state->memory;
309             state->v1 += XXH_readLE32(p32, endian) * PRIME32_2; state->v1 = XXH_rotl32(state->v1, 13); state->v1 *= PRIME32_1; p32++;
310             state->v2 += XXH_readLE32(p32, endian) * PRIME32_2; state->v2 = XXH_rotl32(state->v2, 13); state->v2 *= PRIME32_1; p32++;
311             state->v3 += XXH_readLE32(p32, endian) * PRIME32_2; state->v3 = XXH_rotl32(state->v3, 13); state->v3 *= PRIME32_1; p32++;
312             state->v4 += XXH_readLE32(p32, endian) * PRIME32_2; state->v4 = XXH_rotl32(state->v4, 13); state->v4 *= PRIME32_1; p32++;
313         }
314         p += 16-state->memsize;
315         state->memsize = 0;
316     }
317 
318     if (p <= bEnd-16)
319     {
320         const uint8_t * const limit = bEnd - 16;
321         uint32_t v1 = state->v1;
322         uint32_t v2 = state->v2;
323         uint32_t v3 = state->v3;
324         uint32_t v4 = state->v4;
325 
326         do
327         {
328             v1 += XXH_readLE32((const uint32_t*)p, endian) * PRIME32_2; v1 = XXH_rotl32(v1, 13); v1 *= PRIME32_1; p+=4;
329             v2 += XXH_readLE32((const uint32_t*)p, endian) * PRIME32_2; v2 = XXH_rotl32(v2, 13); v2 *= PRIME32_1; p+=4;
330             v3 += XXH_readLE32((const uint32_t*)p, endian) * PRIME32_2; v3 = XXH_rotl32(v3, 13); v3 *= PRIME32_1; p+=4;
331             v4 += XXH_readLE32((const uint32_t*)p, endian) * PRIME32_2; v4 = XXH_rotl32(v4, 13); v4 *= PRIME32_1; p+=4;
332         } while (p<=limit);
333 
334         state->v1 = v1;
335         state->v2 = v2;
336         state->v3 = v3;
337         state->v4 = v4;
338     }
339 
340     if (p < bEnd)
341     {
342         memcpy(state->memory, p, bEnd-p);
343         state->memsize = (int)(bEnd-p);
344     }
345 
346     return XXH_OK;
347 }
348 
XXH32_update(void * state_in,const void * input,int len)349 XXH_errorcode XXH32_update (void* state_in, const void* input, int len)
350 {
351     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
352 
353     if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
354         return XXH32_update_endian(state_in, input, len, XXH_littleEndian);
355     else
356         return XXH32_update_endian(state_in, input, len, XXH_bigEndian);
357 }
358 
359 
360 
XXH32_intermediateDigest_endian(void * state_in,XXH_endianess endian)361 static uint32_t XXH32_intermediateDigest_endian (void* state_in, XXH_endianess endian)
362 {
363     struct XXH_state32_t * state = (struct XXH_state32_t *) state_in;
364     const uint8_t *p = (const uint8_t *)state->memory;
365     uint8_t * bEnd = (uint8_t *)state->memory + state->memsize;
366     uint32_t h32;
367 
368     if (state->total_len >= 16)
369     {
370         h32 = XXH_rotl32(state->v1, 1) + XXH_rotl32(state->v2, 7) + XXH_rotl32(state->v3, 12) + XXH_rotl32(state->v4, 18);
371     }
372     else
373     {
374         h32  = state->seed + PRIME32_5;
375     }
376 
377     h32 += (uint32_t) state->total_len;
378 
379     while (p<=bEnd-4)
380     {
381         h32 += XXH_readLE32((const uint32_t*)p, endian) * PRIME32_3;
382         h32  = XXH_rotl32(h32, 17) * PRIME32_4;
383         p+=4;
384     }
385 
386     while (p<bEnd)
387     {
388         h32 += (*p) * PRIME32_5;
389         h32 = XXH_rotl32(h32, 11) * PRIME32_1;
390         p++;
391     }
392 
393     h32 ^= h32 >> 15;
394     h32 *= PRIME32_2;
395     h32 ^= h32 >> 13;
396     h32 *= PRIME32_3;
397     h32 ^= h32 >> 16;
398 
399     return h32;
400 }
401 
402 
XXH32_intermediateDigest(void * state_in)403 uint32_t XXH32_intermediateDigest (void* state_in)
404 {
405     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
406 
407     if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
408         return XXH32_intermediateDigest_endian(state_in, XXH_littleEndian);
409     else
410         return XXH32_intermediateDigest_endian(state_in, XXH_bigEndian);
411 }
412 
413 
XXH32_digest(void * state_in)414 uint32_t XXH32_digest (void* state_in)
415 {
416     uint32_t h32 = XXH32_intermediateDigest(state_in);
417 
418     free(state_in);
419 
420     return h32;
421 }
422