teambeef/raze
0
0
Fork 0
mirror of https://github.com/DrBeef/Raze.git synced 2024-11-16 01:11:28 +00:00
Code Issues Projects Releases Packages Wiki Activity

Update xxhash to r36

Browse source 
git-svn-id: https://svn.eduke32.com/eduke32@4599 1a8010ca-5511-0410-912e-c29ae57300e0
This commit is contained in:
terminx 2014-09-30 04:04:12 +00:00
parent 76ad1ce07e
commit 514f556a32
2 changed files with 411 additions and 76 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

80
polymer/eduke32/build/include/xxhash.h
View file

@ -1,5 +1,5 @@
/* /*
xxHash - Fast Hash algorithm xxHash - Extremely Fast Hash algorithm
Header File Header File
Copyright (C) 2012-2014, Yann Collet. Copyright (C) 2012-2014, Yann Collet.
BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
@ -64,18 +64,19 @@ extern "C" {
#endif #endif
//**************************** /*****************************
// Type Type
//**************************** *****************************/
typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode; typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode;
//**************************** /*****************************
// Simple Hash Functions Simple Hash Functions
//**************************** *****************************/
unsigned int XXH32 (const void* input, int len, unsigned int seed); unsigned int XXH32 (const void* input, unsigned int len, unsigned int seed);
unsigned long long XXH64 (const void* input, unsigned int len, unsigned long long seed);
/* /*
XXH32() : XXH32() :
@ -86,79 +87,82 @@ XXH32() :
Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s
Note that "len" is type "int", which means it is limited to 2^31-1. Note that "len" is type "int", which means it is limited to 2^31-1.
If your data is larger, use the advanced functions below. If your data is larger, use the advanced functions below.
XXH64() :
Calculate the 64-bits hash of sequence of length "len" stored at memory address "input".
*/ */
//**************************** /*****************************
// Advanced Hash Functions Advanced Hash Functions
//**************************** *****************************/
void* XXH32_init (unsigned int seed); void* XXH32_init (unsigned int seed);
XXH_errorcode XXH32_update (void* state, const void* input, int len); XXH_errorcode XXH32_update (void* state, const void* input, unsigned int len);
unsigned int XXH32_digest (void* state); unsigned int XXH32_digest (void* state);
void* XXH64_init (unsigned long long seed);
XXH_errorcode XXH64_update (void* state, const void* input, unsigned int len);
unsigned long long XXH64_digest (void* state);
/* /*
These functions calculate the xxhash of an input provided in several small packets, These functions calculate the xxhash of an input provided in several small packets,
as opposed to an input provided as a single block. as opposed to an input provided as a single block.
It must be started with : It must be started with :
void* XXH32_init() void* XXHnn_init()
The function returns a pointer which holds the state of calculation. The function returns a pointer which holds the state of calculation.
If the pointer is NULL, allocation has failed, so no state can be tracked.
This pointer must be provided as "void* state" parameter for XXH32_update(). The state pointer must be provided as "void* state" parameter for XXHnn_update().
XXH32_update() can be called as many times as necessary. XXHnn_update() can be called as many times as necessary.
The user must provide a valid (allocated) input. The user must provide a valid (allocated) input.
The function returns an error code, with 0 meaning OK, and any other value meaning there is an error. The function returns an error code, with 0 meaning OK, and any other value meaning there is an error.
Note that "len" is type "int", which means it is limited to 2^31-1. Note that "len" is type "int", which means it is limited to 2^31-1.
If your data is larger, it is recommended to chunk your data into blocks If your data is larger, it is recommended to chunk your data into blocks
of size for example 2^30 (1GB) to avoid any "int" overflow issue. of size for example 2^30 (1GB) to avoid any "int" overflow issue.
Finally, you can end the calculation anytime, by using XXH32_digest(). Finally, you can end the calculation anytime, by using XXHnn_digest().
This function returns the final 32-bits hash. This function returns the final nn-bits hash.
You must provide the same "void* state" parameter created by XXH32_init(). You must provide the same "void* state" parameter created by XXHnn_init().
Memory will be freed by XXH32_digest(). Memory will be freed by XXHnn_digest().
*/ */
int XXH32_sizeofState(); int XXH32_sizeofState(void);
XXH_errorcode XXH32_resetState(void* state, unsigned int seed); XXH_errorcode XXH32_resetState(void* state, unsigned int seed);
#define XXH32_SIZEOFSTATE 48 #define XXH32_SIZEOFSTATE 48
typedef struct { long long ll[(XXH32_SIZEOFSTATE+(sizeof(long long)-1))/sizeof(long long)]; } XXH32_stateSpace_t; typedef struct { long long ll[(XXH32_SIZEOFSTATE+(sizeof(long long)-1))/sizeof(long long)]; } XXH32_stateSpace_t;
int XXH64_sizeofState(void);
XXH_errorcode XXH64_resetState(void* state, unsigned long long seed);
#define XXH64_SIZEOFSTATE 88
typedef struct { long long ll[(XXH64_SIZEOFSTATE+(sizeof(long long)-1))/sizeof(long long)]; } XXH64_stateSpace_t;
/* /*
These functions allow user application to make its own allocation for state. These functions allow user application to make its own allocation for state.
XXH32_sizeofState() is used to know how much space must be allocated for the xxHash 32-bits state. XXHnn_sizeofState() is used to know how much space must be allocated for the xxHash nn-bits state.
Note that the state must be aligned to access 'long long' fields. Memory must be allocated and referenced by a pointer. Note that the state must be aligned to access 'long long' fields. Memory must be allocated and referenced by a pointer.
This pointer must then be provided as 'state' into XXH32_resetState(), which initializes the state. This pointer must then be provided as 'state' into XXHnn_resetState(), which initializes the state.
For static allocation purposes (such as allocation on stack, or freestanding systems without malloc()), For static allocation purposes (such as allocation on stack, or freestanding systems without malloc()),
use the structure XXH32_stateSpace_t, which will ensure that memory space is large enough and correctly aligned to access 'long long' fields. use the structure XXHnn_stateSpace_t, which will ensure that memory space is large enough and correctly aligned to access 'long long' fields.
*/ */
unsigned int XXH32_intermediateDigest (void* state); unsigned int XXH32_intermediateDigest (void* state);
unsigned long long XXH64_intermediateDigest (void* state);
/* /*
This function does the same as XXH32_digest(), generating a 32-bit hash, These functions do the same as XXHnn_digest(), generating a nn-bit hash,
but preserve memory context. but preserve memory context.
This way, it becomes possible to generate intermediate hashes, and then continue feeding data with XXH32_update(). This way, it becomes possible to generate intermediate hashes, and then continue feeding data with XXHnn_update().
To free memory context, use XXH32_digest(), or free(). To free memory context, use XXHnn_digest(), or free().
*/ */
//****************************
// Deprecated function names
//****************************
// The following translations are provided to ease code transition
// You are encouraged to no longer this function names
#define XXH32_feed XXH32_update
#define XXH32_result XXH32_digest
#define XXH32_getIntermediateResult XXH32_intermediateDigest
#if defined (__cplusplus) #if defined (__cplusplus)
} }
#endif #endif

367
polymer/eduke32/build/src/xxhash.c
View file

@ -58,7 +58,6 @@ You can contact the author at :
// This option has no impact on Little_Endian CPU. // This option has no impact on Little_Endian CPU.
#define XXH_FORCE_NATIVE_FORMAT 0 #define XXH_FORCE_NATIVE_FORMAT 0
//************************************** //**************************************
// Compiler Specific Options // Compiler Specific Options
//************************************** //**************************************
@ -77,7 +76,6 @@ You can contact the author at :
# endif # endif
#endif #endif
//************************************** //**************************************
// Includes & Memory related functions // Includes & Memory related functions
//************************************** //**************************************
@ -126,12 +124,14 @@ FORCE_INLINE void* XXH_memcpy(void* dest, const void* src, size_t size) { return
#endif #endif
typedef struct _U32_S { U32 v; } _PACKED U32_S; typedef struct _U32_S { U32 v; } _PACKED U32_S;
typedef struct _U64_S { U64 v; } _PACKED U64_S;
#if !defined(XXH_USE_UNALIGNED_ACCESS) && !defined(__GNUC__) #if !defined(XXH_USE_UNALIGNED_ACCESS) && !defined(__GNUC__)
# pragma pack(pop) # pragma pack(pop)
#endif #endif
#define A32(x) (((U32_S *)(x))->v) #define A32(x) (((U32_S *)(x))->v)
#define A64(x) (((U64_S *)(x))->v)
//*************************************** //***************************************
@ -142,20 +142,33 @@ typedef struct _U32_S { U32 v; } _PACKED U32_S;
// Note : although _rotl exists for minGW (GCC under windows), performance seems poor // Note : although _rotl exists for minGW (GCC under windows), performance seems poor
#if defined(_MSC_VER) #if defined(_MSC_VER)
# define XXH_rotl32(x,r) _rotl(x,r) # define XXH_rotl32(x,r) _rotl(x,r)
# define XXH_rotl64(x,r) _rotl64(x,r)
#else #else
# define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r))) # define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r)))
# define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r)))
#endif #endif
#if defined(_MSC_VER) // Visual Studio #if defined(_MSC_VER) // Visual Studio
# define XXH_swap32 _byteswap_ulong # define XXH_swap32 _byteswap_ulong
# define XXH_swap64 _byteswap_uint64
#elif GCC_VERSION >= 403 #elif GCC_VERSION >= 403
# define XXH_swap32 __builtin_bswap32 # define XXH_swap32 __builtin_bswap32
# define XXH_swap64 __builtin_bswap64
#else #else
static inline U32 XXH_swap32 (U32 x) { static inline U32 XXH_swap32 (U32 x) {
return ((x << 24) & 0xff000000 ) | return ((x << 24) & 0xff000000 ) |
((x << 8) & 0x00ff0000 ) | ((x << 8) & 0x00ff0000 ) |
((x >> 8) & 0x0000ff00 ) | ((x >> 8) & 0x0000ff00 ) |
((x >> 24) & 0x000000ff );} ((x >> 24) & 0x000000ff );}
static inline U64 XXH_swap64 (U64 x) {
return ((x << 56) & 0xff00000000000000ULL) |
((x << 40) & 0x00ff000000000000ULL) |
((x << 24) & 0x0000ff0000000000ULL) |
((x << 8) & 0x000000ff00000000ULL) |
((x >> 8) & 0x00000000ff000000ULL) |
((x >> 24) & 0x0000000000ff0000ULL) |
((x >> 40) & 0x000000000000ff00ULL) |
((x >> 56) & 0x00000000000000ffULL);}
#endif #endif
@ -168,6 +181,11 @@ static inline U32 XXH_swap32 (U32 x) {
#define PRIME32_4 668265263U #define PRIME32_4 668265263U
#define PRIME32_5 374761393U #define PRIME32_5 374761393U
#define PRIME64_1 11400714785074694791ULL
#define PRIME64_2 14029467366897019727ULL
#define PRIME64_3 1609587929392839161ULL
#define PRIME64_4 9650029242287828579ULL
#define PRIME64_5 2870177450012600261ULL
//************************************** //**************************************
// Architecture Macros // Architecture Macros
@ -182,7 +200,7 @@ typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
//************************************** //**************************************
// Macros // Macros
//************************************** //**************************************
#define XXH_STATIC_ASSERT(c) { enum { XXH_static_assert = 1/(!!(c)) }; } // use only *after* variable declarations #define XXH_STATIC_ASSERT(c) { enum { XXH_static_assert = 1/(int)(!!(c)) }; } // use only *after* variable declarations
//**************************** //****************************
@ -200,18 +218,29 @@ FORCE_INLINE U32 XXH_readLE32_align(const U32* ptr, XXH_endianess endian, XXH_al
FORCE_INLINE U32 XXH_readLE32(const U32* ptr, XXH_endianess endian) { return XXH_readLE32_align(ptr, endian, XXH_unaligned); } FORCE_INLINE U32 XXH_readLE32(const U32* ptr, XXH_endianess endian) { return XXH_readLE32_align(ptr, endian, XXH_unaligned); }
FORCE_INLINE U64 XXH_readLE64_align(const U64* ptr, XXH_endianess endian, XXH_alignment align)
{
if (align==XXH_unaligned)
return endian==XXH_littleEndian ? A64(ptr) : XXH_swap64(A64(ptr));
else
return endian==XXH_littleEndian ? *ptr : XXH_swap64(*ptr);
}
FORCE_INLINE U64 XXH_readLE64(const U64* ptr, XXH_endianess endian) { return XXH_readLE64_align(ptr, endian, XXH_unaligned); }
//**************************** //****************************
// Simple Hash Functions // Simple Hash Functions
//**************************** //****************************
FORCE_INLINE U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_endianess endian, XXH_alignment align) FORCE_INLINE U32 XXH32_endian_align(const void* input, unsigned int len, U32 seed, XXH_endianess endian, XXH_alignment align)
{ {
const BYTE* p = (const BYTE*)input; const BYTE* p = (const BYTE*)input;
const BYTE* const bEnd = p + len; const BYTE* bEnd = p + len;
U32 h32; U32 h32;
#define XXH_get32bits(p) XXH_readLE32_align((const U32*)p, endian, align)
#ifdef XXH_ACCEPT_NULL_INPUT_POINTER #ifdef XXH_ACCEPT_NULL_INPUT_POINTER
if (p==NULL) { len=0; p=(const BYTE*)(size_t)16; } if (p==NULL) { len=0; bEnd=p=(const BYTE*)(size_t)16; }
#endif #endif
if (len>=16) if (len>=16)
@ -224,10 +253,10 @@ FORCE_INLINE U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_en
do do
{ {
v1 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v1 = XXH_rotl32(v1, 13); v1 *= PRIME32_1; p+=4; v1 += XXH_get32bits(p) * PRIME32_2; v1 = XXH_rotl32(v1, 13); v1 *= PRIME32_1; p+=4;
v2 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v2 = XXH_rotl32(v2, 13); v2 *= PRIME32_1; p+=4; v2 += XXH_get32bits(p) * PRIME32_2; v2 = XXH_rotl32(v2, 13); v2 *= PRIME32_1; p+=4;
v3 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v3 = XXH_rotl32(v3, 13); v3 *= PRIME32_1; p+=4; v3 += XXH_get32bits(p) * PRIME32_2; v3 = XXH_rotl32(v3, 13); v3 *= PRIME32_1; p+=4;
v4 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_2; v4 = XXH_rotl32(v4, 13); v4 *= PRIME32_1; p+=4; v4 += XXH_get32bits(p) * PRIME32_2; v4 = XXH_rotl32(v4, 13); v4 *= PRIME32_1; p+=4;
} while (p<=limit); } while (p<=limit);
h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18); h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18);
@ -239,9 +268,9 @@ FORCE_INLINE U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_en
h32 += (U32) len; h32 += (U32) len;
while (p<=bEnd-4) while (p+4<=bEnd)
{ {
h32 += XXH_readLE32_align((const U32*)p, endian, align) * PRIME32_3; h32 += XXH_get32bits(p) * PRIME32_3;
h32 = XXH_rotl32(h32, 17) * PRIME32_4 ; h32 = XXH_rotl32(h32, 17) * PRIME32_4 ;
p+=4; p+=4;
} }
@ -263,7 +292,7 @@ FORCE_INLINE U32 XXH32_endian_align(const void* input, int len, U32 seed, XXH_en
} }
U32 XXH32(const void* input, int len, U32 seed) U32 XXH32(const void* input, unsigned int len, U32 seed)
{ {
#if 0 #if 0
// Simple version, good for code maintenance, but unfortunately slow for small inputs // Simple version, good for code maintenance, but unfortunately slow for small inputs
@ -274,7 +303,7 @@ U32 XXH32(const void* input, int len, U32 seed)
XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
# if !defined(XXH_USE_UNALIGNED_ACCESS) # if !defined(XXH_USE_UNALIGNED_ACCESS)
if ((((size_t)input) & 3)) // Input is aligned, let's leverage the speed advantage if ((((size_t)input) & 3) == 0) // Input is aligned, let's leverage the speed advantage
{ {
if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT) if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned); return XXH32_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
@ -290,6 +319,105 @@ U32 XXH32(const void* input, int len, U32 seed)
#endif #endif
} }
FORCE_INLINE U64 XXH64_endian_align(const void* input, unsigned int len, U64 seed, XXH_endianess endian, XXH_alignment align)
{
const BYTE* p = (const BYTE*)input;
const BYTE* bEnd = p + len;
U64 h64;
#define XXH_get64bits(p) XXH_readLE64_align((const U64*)p, endian, align)
#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
if (p==NULL) { len=0; bEnd=p=(const BYTE*)(size_t)32; }
#endif
if (len>=32)
{
const BYTE* const limit = bEnd - 32;
U64 v1 = seed + PRIME64_1 + PRIME64_2;
U64 v2 = seed + PRIME64_2;
U64 v3 = seed + 0;
U64 v4 = seed - PRIME64_1;
do
{
v1 += XXH_get64bits(p) * PRIME64_2; p+=8; v1 = XXH_rotl64(v1, 31); v1 *= PRIME64_1;
v2 += XXH_get64bits(p) * PRIME64_2; p+=8; v2 = XXH_rotl64(v2, 31); v2 *= PRIME64_1;
v3 += XXH_get64bits(p) * PRIME64_2; p+=8; v3 = XXH_rotl64(v3, 31); v3 *= PRIME64_1;
v4 += XXH_get64bits(p) * PRIME64_2; p+=8; v4 = XXH_rotl64(v4, 31); v4 *= PRIME64_1;
} while (p<=limit);
h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
v1 *= PRIME64_2; v1 = XXH_rotl64(v1, 31); v1 *= PRIME64_1; h64 ^= v1;
h64 = h64 * PRIME64_1 + PRIME64_4;
v2 *= PRIME64_2; v2 = XXH_rotl64(v2, 31); v2 *= PRIME64_1; h64 ^= v2;
h64 = h64 * PRIME64_1 + PRIME64_4;
v3 *= PRIME64_2; v3 = XXH_rotl64(v3, 31); v3 *= PRIME64_1; h64 ^= v3;
h64 = h64 * PRIME64_1 + PRIME64_4;
v4 *= PRIME64_2; v4 = XXH_rotl64(v4, 31); v4 *= PRIME64_1; h64 ^= v4;
h64 = h64 * PRIME64_1 + PRIME64_4;
}
else
{
h64 = seed + PRIME64_5;
}
h64 += (U64) len;
while (p+8<=bEnd)
{
U64 k1 = XXH_get64bits(p);
k1 *= PRIME64_2; k1 = XXH_rotl64(k1,31); k1 *= PRIME64_1; h64 ^= k1;
h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4;
p+=8;
}
if (p+4<=bEnd)
{
h64 ^= (U64)(XXH_get32bits(p)) * PRIME64_1;
h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
p+=4;
}
while (p<bEnd)
{
h64 ^= (*p) * PRIME64_5;
h64 = XXH_rotl64(h64, 11) * PRIME64_1;
p++;
}
h64 ^= h64 >> 33;
h64 *= PRIME64_2;
h64 ^= h64 >> 29;
h64 *= PRIME64_3;
h64 ^= h64 >> 32;
return h64;
}
unsigned long long XXH64(const void* input, unsigned int len, unsigned long long seed)
{
XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
# if !defined(XXH_USE_UNALIGNED_ACCESS)
if ((((size_t)input) & 7)==0) // Input is aligned, let's leverage the speed advantage
{
if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
else
return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned);
}
# endif
if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned);
else
return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned);
}
//**************************** //****************************
// Advanced Hash Functions // Advanced Hash Functions
@ -307,13 +435,31 @@ struct XXH_state32_t
char memory[16]; char memory[16];
}; };
struct XXH_state64_t
{
U64 total_len;
U64 seed;
U64 v1;
U64 v2;
U64 v3;
U64 v4;
int memsize;
char memory[32];
};
int XXH32_sizeofState()
int XXH32_sizeofState(void)
{ {
XXH_STATIC_ASSERT(XXH32_SIZEOFSTATE >= sizeof(struct XXH_state32_t)); // A compilation error here means XXH32_SIZEOFSTATE is not large enough XXH_STATIC_ASSERT(XXH32_SIZEOFSTATE >= sizeof(struct XXH_state32_t)); // A compilation error here means XXH32_SIZEOFSTATE is not large enough
return sizeof(struct XXH_state32_t); return sizeof(struct XXH_state32_t);
} }
int XXH64_sizeofState(void)
{
XXH_STATIC_ASSERT(XXH64_SIZEOFSTATE >= sizeof(struct XXH_state64_t)); // A compilation error here means XXH64_SIZEOFSTATE is not large enough
return sizeof(struct XXH_state64_t);
}
XXH_errorcode XXH32_resetState(void* state_in, U32 seed) XXH_errorcode XXH32_resetState(void* state_in, U32 seed)
{ {
@ -328,11 +474,31 @@ XXH_errorcode XXH32_resetState(void* state_in, U32 seed)
return XXH_OK; return XXH_OK;
} }
XXH_errorcode XXH64_resetState(void* state_in, unsigned long long seed)
{
struct XXH_state64_t * state = (struct XXH_state64_t *) state_in;
state->seed = seed;
state->v1 = seed + PRIME64_1 + PRIME64_2;
state->v2 = seed + PRIME64_2;
state->v3 = seed + 0;
state->v4 = seed - PRIME64_1;
state->total_len = 0;
state->memsize = 0;
return XXH_OK;
}
void* XXH32_init (U32 seed) void* XXH32_init (U32 seed)
{ {
void* state = XXH_malloc (sizeof(struct XXH_state32_t)); void* state = XXH_malloc (sizeof(struct XXH_state32_t));
XXH32_resetState(state, seed); if (state != NULL) XXH32_resetState(state, seed);
return state;
}
void* XXH64_init (unsigned long long seed)
{
void* state = XXH_malloc (sizeof(struct XXH_state64_t));
if (state != NULL) XXH64_resetState(state, seed);
return state; return state;
} }
@ -401,7 +567,7 @@ FORCE_INLINE XXH_errorcode XXH32_update_endian (void* state_in, const void* inpu
return XXH_OK; return XXH_OK;
} }
XXH_errorcode XXH32_update (void* state_in, const void* input, int len) XXH_errorcode XXH32_update (void* state_in, const void* input, unsigned int len)
{ {
XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN; XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
@ -431,7 +597,7 @@ FORCE_INLINE U32 XXH32_intermediateDigest_endian (void* state_in, XXH_endianess
h32 += (U32) state->total_len; h32 += (U32) state->total_len;
while (p<=bEnd-4) while (p+4<=bEnd)
{ {
h32 += XXH_readLE32((const U32*)p, endian) * PRIME32_3; h32 += XXH_readLE32((const U32*)p, endian) * PRIME32_3;
h32 = XXH_rotl32(h32, 17) * PRIME32_4; h32 = XXH_rotl32(h32, 17) * PRIME32_4;
@ -474,3 +640,168 @@ U32 XXH32_digest (void* state_in)
return h32; return h32;
} }
FORCE_INLINE XXH_errorcode XXH64_update_endian (void* state_in, const void* input, int len, XXH_endianess endian)
{
struct XXH_state64_t * state = (struct XXH_state64_t *) state_in;
const BYTE* p = (const BYTE*)input;
const BYTE* const bEnd = p + len;
#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
if (input==NULL) return XXH_ERROR;
#endif
state->total_len += len;
if (state->memsize + len < 32) // fill in tmp buffer
{
XXH_memcpy(state->memory + state->memsize, input, len);
state->memsize += len;
return XXH_OK;
}
if (state->memsize) // some data left from previous update
{
XXH_memcpy(state->memory + state->memsize, input, 32-state->memsize);
{
const U64* p64 = (const U64*)state->memory;
state->v1 += XXH_readLE64(p64, endian) * PRIME64_2; state->v1 = XXH_rotl64(state->v1, 31); state->v1 *= PRIME64_1; p64++;
state->v2 += XXH_readLE64(p64, endian) * PRIME64_2; state->v2 = XXH_rotl64(state->v2, 31); state->v2 *= PRIME64_1; p64++;
state->v3 += XXH_readLE64(p64, endian) * PRIME64_2; state->v3 = XXH_rotl64(state->v3, 31); state->v3 *= PRIME64_1; p64++;
state->v4 += XXH_readLE64(p64, endian) * PRIME64_2; state->v4 = XXH_rotl64(state->v4, 31); state->v4 *= PRIME64_1; p64++;
}
p += 32-state->memsize;
state->memsize = 0;
}
if (p+32 <= bEnd)
{
const BYTE* const limit = bEnd - 32;
U64 v1 = state->v1;
U64 v2 = state->v2;
U64 v3 = state->v3;
U64 v4 = state->v4;
do
{
v1 += XXH_readLE64((const U64*)p, endian) * PRIME64_2; v1 = XXH_rotl64(v1, 31); v1 *= PRIME64_1; p+=8;
v2 += XXH_readLE64((const U64*)p, endian) * PRIME64_2; v2 = XXH_rotl64(v2, 31); v2 *= PRIME64_1; p+=8;
v3 += XXH_readLE64((const U64*)p, endian) * PRIME64_2; v3 = XXH_rotl64(v3, 31); v3 *= PRIME64_1; p+=8;
v4 += XXH_readLE64((const U64*)p, endian) * PRIME64_2; v4 = XXH_rotl64(v4, 31); v4 *= PRIME64_1; p+=8;
} while (p<=limit);
state->v1 = v1;
state->v2 = v2;
state->v3 = v3;
state->v4 = v4;
}
if (p < bEnd)
{
XXH_memcpy(state->memory, p, bEnd-p);
state->memsize = (int)(bEnd-p);
}
return XXH_OK;
}
XXH_errorcode XXH64_update (void* state_in, const void* input, unsigned int len)
{
XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
return XXH64_update_endian(state_in, input, len, XXH_littleEndian);
else
return XXH64_update_endian(state_in, input, len, XXH_bigEndian);
}
FORCE_INLINE U64 XXH64_intermediateDigest_endian (void* state_in, XXH_endianess endian)
{
struct XXH_state64_t * state = (struct XXH_state64_t *) state_in;
const BYTE * p = (const BYTE*)state->memory;
BYTE* bEnd = (BYTE*)state->memory + state->memsize;
U64 h64;
if (state->total_len >= 32)
{
U64 v1 = state->v1;
U64 v2 = state->v2;
U64 v3 = state->v3;
U64 v4 = state->v4;
h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
v1 *= PRIME64_2; v1 = XXH_rotl64(v1, 31); v1 *= PRIME64_1; h64 ^= v1;
h64 = h64*PRIME64_1 + PRIME64_4;
v2 *= PRIME64_2; v2 = XXH_rotl64(v2, 31); v2 *= PRIME64_1; h64 ^= v2;
h64 = h64*PRIME64_1 + PRIME64_4;
v3 *= PRIME64_2; v3 = XXH_rotl64(v3, 31); v3 *= PRIME64_1; h64 ^= v3;
h64 = h64*PRIME64_1 + PRIME64_4;
v4 *= PRIME64_2; v4 = XXH_rotl64(v4, 31); v4 *= PRIME64_1; h64 ^= v4;
h64 = h64*PRIME64_1 + PRIME64_4;
}
else
{
h64 = state->seed + PRIME64_5;
}
h64 += (U64) state->total_len;
while (p+8<=bEnd)
{
U64 k1 = XXH_readLE64((const U64*)p, endian);
k1 *= PRIME64_2; k1 = XXH_rotl64(k1,31); k1 *= PRIME64_1; h64 ^= k1;
h64 = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4;
p+=8;
}
if (p+4<=bEnd)
{
h64 ^= (U64)(XXH_readLE32((const U32*)p, endian)) * PRIME64_1;
h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
p+=4;
}
while (p<bEnd)
{
h64 ^= (*p) * PRIME64_5;
h64 = XXH_rotl64(h64, 11) * PRIME64_1;
p++;
}
h64 ^= h64 >> 33;
h64 *= PRIME64_2;
h64 ^= h64 >> 29;
h64 *= PRIME64_3;
h64 ^= h64 >> 32;
return h64;
}
unsigned long long XXH64_intermediateDigest (void* state_in)
{
XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
return XXH64_intermediateDigest_endian(state_in, XXH_littleEndian);
else
return XXH64_intermediateDigest_endian(state_in, XXH_bigEndian);
}
unsigned long long XXH64_digest (void* state_in)
{
U64 h64 = XXH64_intermediateDigest(state_in);
XXH_free(state_in);
return h64;
}