| 1 |
/* |
| 2 |
* Copyright (c) 2016-2020, Yann Collet, Facebook, Inc. |
| 3 |
* All rights reserved. |
| 4 |
* |
| 5 |
* This source code is licensed under both the BSD-style license (found in the |
| 6 |
* LICENSE file in the root directory of this source tree) and the GPLv2 (found |
| 7 |
* in the COPYING file in the root directory of this source tree). |
| 8 |
* You may select, at your option, one of the above-listed licenses. |
| 9 |
*/ |
| 10 |
|
| 11 |
/* This header contains definitions |
| 12 |
* that shall **only** be used by modules within lib/compress. |
| 13 |
*/ |
| 14 |
|
| 15 |
#ifndef ZSTD_COMPRESS_H |
| 16 |
#define ZSTD_COMPRESS_H |
| 17 |
|
| 18 |
/*-************************************* |
| 19 |
* Dependencies |
| 20 |
***************************************/ |
| 21 |
#include "../common/zstd_internal.h" |
| 22 |
#include "zstd_cwksp.h" |
| 23 |
#ifdef ZSTD_MULTITHREAD |
| 24 |
# include "zstdmt_compress.h" |
| 25 |
#endif |
| 26 |
|
| 27 |
#if defined (__cplusplus) |
| 28 |
extern "C" { |
| 29 |
#endif |
| 30 |
|
| 31 |
|
| 32 |
/*-************************************* |
| 33 |
* Constants |
| 34 |
***************************************/ |
| 35 |
#define kSearchStrength 8 |
| 36 |
#define HASH_READ_SIZE 8 |
| 37 |
#define ZSTD_DUBT_UNSORTED_MARK 1 /* For btlazy2 strategy, index ZSTD_DUBT_UNSORTED_MARK==1 means "unsorted". |
| 38 |
It could be confused for a real successor at index "1", if sorted as larger than its predecessor. |
| 39 |
It's not a big deal though : candidate will just be sorted again. |
| 40 |
Additionally, candidate position 1 will be lost. |
| 41 |
But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss. |
| 42 |
The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy. |
| 43 |
This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */ |
| 44 |
|
| 45 |
|
| 46 |
/*-************************************* |
| 47 |
* Context memory management |
| 48 |
***************************************/ |
| 49 |
typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e; |
| 50 |
typedef enum { zcss_init=0, zcss_load, zcss_flush } ZSTD_cStreamStage; |
| 51 |
|
| 52 |
typedef struct ZSTD_prefixDict_s { |
| 53 |
const void* dict; |
| 54 |
size_t dictSize; |
| 55 |
ZSTD_dictContentType_e dictContentType; |
| 56 |
} ZSTD_prefixDict; |
| 57 |
|
| 58 |
typedef struct { |
| 59 |
void* dictBuffer; |
| 60 |
void const* dict; |
| 61 |
size_t dictSize; |
| 62 |
ZSTD_dictContentType_e dictContentType; |
| 63 |
ZSTD_CDict* cdict; |
| 64 |
} ZSTD_localDict; |
| 65 |
|
| 66 |
typedef struct { |
| 67 |
U32 CTable[HUF_CTABLE_SIZE_U32(255)]; |
| 68 |
HUF_repeat repeatMode; |
| 69 |
} ZSTD_hufCTables_t; |
| 70 |
|
| 71 |
typedef struct { |
| 72 |
FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)]; |
| 73 |
FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)]; |
| 74 |
FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)]; |
| 75 |
FSE_repeat offcode_repeatMode; |
| 76 |
FSE_repeat matchlength_repeatMode; |
| 77 |
FSE_repeat litlength_repeatMode; |
| 78 |
} ZSTD_fseCTables_t; |
| 79 |
|
| 80 |
typedef struct { |
| 81 |
ZSTD_hufCTables_t huf; |
| 82 |
ZSTD_fseCTables_t fse; |
| 83 |
} ZSTD_entropyCTables_t; |
| 84 |
|
| 85 |
typedef struct { |
| 86 |
U32 off; |
| 87 |
U32 len; |
| 88 |
} ZSTD_match_t; |
| 89 |
|
| 90 |
typedef struct { |
| 91 |
int price; |
| 92 |
U32 off; |
| 93 |
U32 mlen; |
| 94 |
U32 litlen; |
| 95 |
U32 rep[ZSTD_REP_NUM]; |
| 96 |
} ZSTD_optimal_t; |
| 97 |
|
| 98 |
typedef enum { zop_dynamic=0, zop_predef } ZSTD_OptPrice_e; |
| 99 |
|
| 100 |
typedef struct { |
| 101 |
/* All tables are allocated inside cctx->workspace by ZSTD_resetCCtx_internal() */ |
| 102 |
unsigned* litFreq; /* table of literals statistics, of size 256 */ |
| 103 |
unsigned* litLengthFreq; /* table of litLength statistics, of size (MaxLL+1) */ |
| 104 |
unsigned* matchLengthFreq; /* table of matchLength statistics, of size (MaxML+1) */ |
| 105 |
unsigned* offCodeFreq; /* table of offCode statistics, of size (MaxOff+1) */ |
| 106 |
ZSTD_match_t* matchTable; /* list of found matches, of size ZSTD_OPT_NUM+1 */ |
| 107 |
ZSTD_optimal_t* priceTable; /* All positions tracked by optimal parser, of size ZSTD_OPT_NUM+1 */ |
| 108 |
|
| 109 |
U32 litSum; /* nb of literals */ |
| 110 |
U32 litLengthSum; /* nb of litLength codes */ |
| 111 |
U32 matchLengthSum; /* nb of matchLength codes */ |
| 112 |
U32 offCodeSum; /* nb of offset codes */ |
| 113 |
U32 litSumBasePrice; /* to compare to log2(litfreq) */ |
| 114 |
U32 litLengthSumBasePrice; /* to compare to log2(llfreq) */ |
| 115 |
U32 matchLengthSumBasePrice;/* to compare to log2(mlfreq) */ |
| 116 |
U32 offCodeSumBasePrice; /* to compare to log2(offreq) */ |
| 117 |
ZSTD_OptPrice_e priceType; /* prices can be determined dynamically, or follow a pre-defined cost structure */ |
| 118 |
const ZSTD_entropyCTables_t* symbolCosts; /* pre-calculated dictionary statistics */ |
| 119 |
ZSTD_literalCompressionMode_e literalCompressionMode; |
| 120 |
} optState_t; |
| 121 |
|
| 122 |
typedef struct { |
| 123 |
ZSTD_entropyCTables_t entropy; |
| 124 |
U32 rep[ZSTD_REP_NUM]; |
| 125 |
} ZSTD_compressedBlockState_t; |
| 126 |
|
| 127 |
typedef struct { |
| 128 |
BYTE const* nextSrc; /* next block here to continue on current prefix */ |
| 129 |
BYTE const* base; /* All regular indexes relative to this position */ |
| 130 |
BYTE const* dictBase; /* extDict indexes relative to this position */ |
| 131 |
U32 dictLimit; /* below that point, need extDict */ |
| 132 |
U32 lowLimit; /* below that point, no more valid data */ |
| 133 |
} ZSTD_window_t; |
| 134 |
|
| 135 |
typedef struct ZSTD_matchState_t ZSTD_matchState_t; |
| 136 |
struct ZSTD_matchState_t { |
| 137 |
ZSTD_window_t window; /* State for window round buffer management */ |
| 138 |
U32 loadedDictEnd; /* index of end of dictionary, within context's referential. |
| 139 |
* When loadedDictEnd != 0, a dictionary is in use, and still valid. |
| 140 |
* This relies on a mechanism to set loadedDictEnd=0 when dictionary is no longer within distance. |
| 141 |
* Such mechanism is provided within ZSTD_window_enforceMaxDist() and ZSTD_checkDictValidity(). |
| 142 |
* When dict referential is copied into active context (i.e. not attached), |
| 143 |
* loadedDictEnd == dictSize, since referential starts from zero. |
| 144 |
*/ |
| 145 |
U32 nextToUpdate; /* index from which to continue table update */ |
| 146 |
U32 hashLog3; /* dispatch table for matches of len==3 : larger == faster, more memory */ |
| 147 |
U32* hashTable; |
| 148 |
U32* hashTable3; |
| 149 |
U32* chainTable; |
| 150 |
optState_t opt; /* optimal parser state */ |
| 151 |
const ZSTD_matchState_t* dictMatchState; |
| 152 |
ZSTD_compressionParameters cParams; |
| 153 |
}; |
| 154 |
|
| 155 |
typedef struct { |
| 156 |
ZSTD_compressedBlockState_t* prevCBlock; |
| 157 |
ZSTD_compressedBlockState_t* nextCBlock; |
| 158 |
ZSTD_matchState_t matchState; |
| 159 |
} ZSTD_blockState_t; |
| 160 |
|
| 161 |
typedef struct { |
| 162 |
U32 offset; |
| 163 |
U32 checksum; |
| 164 |
} ldmEntry_t; |
| 165 |
|
| 166 |
typedef struct { |
| 167 |
ZSTD_window_t window; /* State for the window round buffer management */ |
| 168 |
ldmEntry_t* hashTable; |
| 169 |
U32 loadedDictEnd; |
| 170 |
BYTE* bucketOffsets; /* Next position in bucket to insert entry */ |
| 171 |
U64 hashPower; /* Used to compute the rolling hash. |
| 172 |
* Depends on ldmParams.minMatchLength */ |
| 173 |
} ldmState_t; |
| 174 |
|
| 175 |
typedef struct { |
| 176 |
U32 enableLdm; /* 1 if enable long distance matching */ |
| 177 |
U32 hashLog; /* Log size of hashTable */ |
| 178 |
U32 bucketSizeLog; /* Log bucket size for collision resolution, at most 8 */ |
| 179 |
U32 minMatchLength; /* Minimum match length */ |
| 180 |
U32 hashRateLog; /* Log number of entries to skip */ |
| 181 |
U32 windowLog; /* Window log for the LDM */ |
| 182 |
} ldmParams_t; |
| 183 |
|
| 184 |
typedef struct { |
| 185 |
U32 offset; |
| 186 |
U32 litLength; |
| 187 |
U32 matchLength; |
| 188 |
} rawSeq; |
| 189 |
|
| 190 |
typedef struct { |
| 191 |
rawSeq* seq; /* The start of the sequences */ |
| 192 |
size_t pos; /* The position where reading stopped. <= size. */ |
| 193 |
size_t size; /* The number of sequences. <= capacity. */ |
| 194 |
size_t capacity; /* The capacity starting from `seq` pointer */ |
| 195 |
} rawSeqStore_t; |
| 196 |
|
| 197 |
typedef struct { |
| 198 |
int collectSequences; |
| 199 |
ZSTD_Sequence* seqStart; |
| 200 |
size_t seqIndex; |
| 201 |
size_t maxSequences; |
| 202 |
} SeqCollector; |
| 203 |
|
| 204 |
struct ZSTD_CCtx_params_s { |
| 205 |
ZSTD_format_e format; |
| 206 |
ZSTD_compressionParameters cParams; |
| 207 |
ZSTD_frameParameters fParams; |
| 208 |
|
| 209 |
int compressionLevel; |
| 210 |
int forceWindow; /* force back-references to respect limit of |
| 211 |
* 1<<wLog, even for dictionary */ |
| 212 |
size_t targetCBlockSize; /* Tries to fit compressed block size to be around targetCBlockSize. |
| 213 |
* No target when targetCBlockSize == 0. |
| 214 |
* There is no guarantee on compressed block size */ |
| 215 |
int srcSizeHint; /* User's best guess of source size. |
| 216 |
* Hint is not valid when srcSizeHint == 0. |
| 217 |
* There is no guarantee that hint is close to actual source size */ |
| 218 |
|
| 219 |
ZSTD_dictAttachPref_e attachDictPref; |
| 220 |
ZSTD_literalCompressionMode_e literalCompressionMode; |
| 221 |
|
| 222 |
/* Multithreading: used to pass parameters to mtctx */ |
| 223 |
int nbWorkers; |
| 224 |
size_t jobSize; |
| 225 |
int overlapLog; |
| 226 |
int rsyncable; |
| 227 |
|
| 228 |
/* Long distance matching parameters */ |
| 229 |
ldmParams_t ldmParams; |
| 230 |
|
| 231 |
/* Internal use, for createCCtxParams() and freeCCtxParams() only */ |
| 232 |
ZSTD_customMem customMem; |
| 233 |
}; /* typedef'd to ZSTD_CCtx_params within "zstd.h" */ |
| 234 |
|
| 235 |
struct ZSTD_CCtx_s { |
| 236 |
ZSTD_compressionStage_e stage; |
| 237 |
int cParamsChanged; /* == 1 if cParams(except wlog) or compression level are changed in requestedParams. Triggers transmission of new params to ZSTDMT (if available) then reset to 0. */ |
| 238 |
int bmi2; /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */ |
| 239 |
ZSTD_CCtx_params requestedParams; |
| 240 |
ZSTD_CCtx_params appliedParams; |
| 241 |
U32 dictID; |
| 242 |
|
| 243 |
ZSTD_cwksp workspace; /* manages buffer for dynamic allocations */ |
| 244 |
size_t blockSize; |
| 245 |
unsigned long long pledgedSrcSizePlusOne; /* this way, 0 (default) == unknown */ |
| 246 |
unsigned long long consumedSrcSize; |
| 247 |
unsigned long long producedCSize; |
| 248 |
XXH64_state_t xxhState; |
| 249 |
ZSTD_customMem customMem; |
| 250 |
size_t staticSize; |
| 251 |
SeqCollector seqCollector; |
| 252 |
int isFirstBlock; |
| 253 |
int initialized; |
| 254 |
|
| 255 |
seqStore_t seqStore; /* sequences storage ptrs */ |
| 256 |
ldmState_t ldmState; /* long distance matching state */ |
| 257 |
rawSeq* ldmSequences; /* Storage for the ldm output sequences */ |
| 258 |
size_t maxNbLdmSequences; |
| 259 |
rawSeqStore_t externSeqStore; /* Mutable reference to external sequences */ |
| 260 |
ZSTD_blockState_t blockState; |
| 261 |
U32* entropyWorkspace; /* entropy workspace of HUF_WORKSPACE_SIZE bytes */ |
| 262 |
|
| 263 |
/* streaming */ |
| 264 |
char* inBuff; |
| 265 |
size_t inBuffSize; |
| 266 |
size_t inToCompress; |
| 267 |
size_t inBuffPos; |
| 268 |
size_t inBuffTarget; |
| 269 |
char* outBuff; |
| 270 |
size_t outBuffSize; |
| 271 |
size_t outBuffContentSize; |
| 272 |
size_t outBuffFlushedSize; |
| 273 |
ZSTD_cStreamStage streamStage; |
| 274 |
U32 frameEnded; |
| 275 |
|
| 276 |
/* Dictionary */ |
| 277 |
ZSTD_localDict localDict; |
| 278 |
const ZSTD_CDict* cdict; |
| 279 |
ZSTD_prefixDict prefixDict; /* single-usage dictionary */ |
| 280 |
|
| 281 |
/* Multi-threading */ |
| 282 |
#ifdef ZSTD_MULTITHREAD |
| 283 |
ZSTDMT_CCtx* mtctx; |
| 284 |
#endif |
| 285 |
}; |
| 286 |
|
| 287 |
typedef enum { ZSTD_dtlm_fast, ZSTD_dtlm_full } ZSTD_dictTableLoadMethod_e; |
| 288 |
|
| 289 |
typedef enum { ZSTD_noDict = 0, ZSTD_extDict = 1, ZSTD_dictMatchState = 2 } ZSTD_dictMode_e; |
| 290 |
|
| 291 |
|
| 292 |
typedef size_t (*ZSTD_blockCompressor) ( |
| 293 |
ZSTD_matchState_t* bs, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], |
| 294 |
void const* src, size_t srcSize); |
| 295 |
ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, ZSTD_dictMode_e dictMode); |
| 296 |
|
| 297 |
|
| 298 |
MEM_STATIC U32 ZSTD_LLcode(U32 litLength) |
| 299 |
{ |
| 300 |
static const BYTE LL_Code[64] = { 0, 1, 2, 3, 4, 5, 6, 7, |
| 301 |
8, 9, 10, 11, 12, 13, 14, 15, |
| 302 |
16, 16, 17, 17, 18, 18, 19, 19, |
| 303 |
20, 20, 20, 20, 21, 21, 21, 21, |
| 304 |
22, 22, 22, 22, 22, 22, 22, 22, |
| 305 |
23, 23, 23, 23, 23, 23, 23, 23, |
| 306 |
24, 24, 24, 24, 24, 24, 24, 24, |
| 307 |
24, 24, 24, 24, 24, 24, 24, 24 }; |
| 308 |
static const U32 LL_deltaCode = 19; |
| 309 |
return (litLength > 63) ? ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength]; |
| 310 |
} |
| 311 |
|
| 312 |
/* ZSTD_MLcode() : |
| 313 |
* note : mlBase = matchLength - MINMATCH; |
| 314 |
* because it's the format it's stored in seqStore->sequences */ |
| 315 |
MEM_STATIC U32 ZSTD_MLcode(U32 mlBase) |
| 316 |
{ |
| 317 |
static const BYTE ML_Code[128] = { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, |
| 318 |
16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, |
| 319 |
32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37, |
| 320 |
38, 38, 38, 38, 38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39, |
| 321 |
40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, |
| 322 |
41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, |
| 323 |
42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, |
| 324 |
42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42 }; |
| 325 |
static const U32 ML_deltaCode = 36; |
| 326 |
return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase]; |
| 327 |
} |
| 328 |
|
| 329 |
typedef struct repcodes_s { |
| 330 |
U32 rep[3]; |
| 331 |
} repcodes_t; |
| 332 |
|
| 333 |
MEM_STATIC repcodes_t ZSTD_updateRep(U32 const rep[3], U32 const offset, U32 const ll0) |
| 334 |
{ |
| 335 |
repcodes_t newReps; |
| 336 |
if (offset >= ZSTD_REP_NUM) { /* full offset */ |
| 337 |
newReps.rep[2] = rep[1]; |
| 338 |
newReps.rep[1] = rep[0]; |
| 339 |
newReps.rep[0] = offset - ZSTD_REP_MOVE; |
| 340 |
} else { /* repcode */ |
| 341 |
U32 const repCode = offset + ll0; |
| 342 |
if (repCode > 0) { /* note : if repCode==0, no change */ |
| 343 |
U32 const currentOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode]; |
| 344 |
newReps.rep[2] = (repCode >= 2) ? rep[1] : rep[2]; |
| 345 |
newReps.rep[1] = rep[0]; |
| 346 |
newReps.rep[0] = currentOffset; |
| 347 |
} else { /* repCode == 0 */ |
| 348 |
memcpy(&newReps, rep, sizeof(newReps)); |
| 349 |
} |
| 350 |
} |
| 351 |
return newReps; |
| 352 |
} |
| 353 |
|
| 354 |
/* ZSTD_cParam_withinBounds: |
| 355 |
* @return 1 if value is within cParam bounds, |
| 356 |
* 0 otherwise */ |
| 357 |
MEM_STATIC int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value) |
| 358 |
{ |
| 359 |
ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam); |
| 360 |
if (ZSTD_isError(bounds.error)) return 0; |
| 361 |
if (value < bounds.lowerBound) return 0; |
| 362 |
if (value > bounds.upperBound) return 0; |
| 363 |
return 1; |
| 364 |
} |
| 365 |
|
| 366 |
/* ZSTD_noCompressBlock() : |
| 367 |
* Writes uncompressed block to dst buffer from given src. |
| 368 |
* Returns the size of the block */ |
| 369 |
MEM_STATIC size_t ZSTD_noCompressBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize, U32 lastBlock) |
| 370 |
{ |
| 371 |
U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(srcSize << 3); |
| 372 |
RETURN_ERROR_IF(srcSize + ZSTD_blockHeaderSize > dstCapacity, |
| 373 |
dstSize_tooSmall, "dst buf too small for uncompressed block"); |
| 374 |
MEM_writeLE24(dst, cBlockHeader24); |
| 375 |
memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize); |
| 376 |
return ZSTD_blockHeaderSize + srcSize; |
| 377 |
} |
| 378 |
|
| 379 |
MEM_STATIC size_t ZSTD_rleCompressBlock (void* dst, size_t dstCapacity, BYTE src, size_t srcSize, U32 lastBlock) |
| 380 |
{ |
| 381 |
BYTE* const op = (BYTE*)dst; |
| 382 |
U32 const cBlockHeader = lastBlock + (((U32)bt_rle)<<1) + (U32)(srcSize << 3); |
| 383 |
RETURN_ERROR_IF(dstCapacity < 4, dstSize_tooSmall, ""); |
| 384 |
MEM_writeLE24(op, cBlockHeader); |
| 385 |
op[3] = src; |
| 386 |
return 4; |
| 387 |
} |
| 388 |
|
| 389 |
|
| 390 |
/* ZSTD_minGain() : |
| 391 |
* minimum compression required |
| 392 |
* to generate a compress block or a compressed literals section. |
| 393 |
* note : use same formula for both situations */ |
| 394 |
MEM_STATIC size_t ZSTD_minGain(size_t srcSize, ZSTD_strategy strat) |
| 395 |
{ |
| 396 |
U32 const minlog = (strat>=ZSTD_btultra) ? (U32)(strat) - 1 : 6; |
| 397 |
ZSTD_STATIC_ASSERT(ZSTD_btultra == 8); |
| 398 |
assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat)); |
| 399 |
return (srcSize >> minlog) + 2; |
| 400 |
} |
| 401 |
|
| 402 |
MEM_STATIC int ZSTD_disableLiteralsCompression(const ZSTD_CCtx_params* cctxParams) |
| 403 |
{ |
| 404 |
switch (cctxParams->literalCompressionMode) { |
| 405 |
case ZSTD_lcm_huffman: |
| 406 |
return 0; |
| 407 |
case ZSTD_lcm_uncompressed: |
| 408 |
return 1; |
| 409 |
default: |
| 410 |
assert(0 /* impossible: pre-validated */); |
| 411 |
/* fall-through */ |
| 412 |
case ZSTD_lcm_auto: |
| 413 |
return (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0); |
| 414 |
} |
| 415 |
} |
| 416 |
|
| 417 |
/*! ZSTD_safecopyLiterals() : |
| 418 |
* memcpy() function that won't read beyond more than WILDCOPY_OVERLENGTH bytes past ilimit_w. |
| 419 |
* Only called when the sequence ends past ilimit_w, so it only needs to be optimized for single |
| 420 |
* large copies. |
| 421 |
*/ |
| 422 |
static void ZSTD_safecopyLiterals(BYTE* op, BYTE const* ip, BYTE const* const iend, BYTE const* ilimit_w) { |
| 423 |
assert(iend > ilimit_w); |
| 424 |
if (ip <= ilimit_w) { |
| 425 |
ZSTD_wildcopy(op, ip, ilimit_w - ip, ZSTD_no_overlap); |
| 426 |
op += ilimit_w - ip; |
| 427 |
ip = ilimit_w; |
| 428 |
} |
| 429 |
while (ip < iend) *op++ = *ip++; |
| 430 |
} |
| 431 |
|
| 432 |
/*! ZSTD_storeSeq() : |
| 433 |
* Store a sequence (litlen, litPtr, offCode and mlBase) into seqStore_t. |
| 434 |
* `offCode` : distance to match + ZSTD_REP_MOVE (values <= ZSTD_REP_MOVE are repCodes). |
| 435 |
* `mlBase` : matchLength - MINMATCH |
| 436 |
* Allowed to overread literals up to litLimit. |
| 437 |
*/ |
| 438 |
HINT_INLINE UNUSED_ATTR |
| 439 |
void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offCode, size_t mlBase) |
| 440 |
{ |
| 441 |
BYTE const* const litLimit_w = litLimit - WILDCOPY_OVERLENGTH; |
| 442 |
BYTE const* const litEnd = literals + litLength; |
| 443 |
#if defined(DEBUGLEVEL) && (DEBUGLEVEL >= 6) |
| 444 |
static const BYTE* g_start = NULL; |
| 445 |
if (g_start==NULL) g_start = (const BYTE*)literals; /* note : index only works for compression within a single segment */ |
| 446 |
{ U32 const pos = (U32)((const BYTE*)literals - g_start); |
| 447 |
DEBUGLOG(6, "Cpos%7u :%3u literals, match%4u bytes at offCode%7u", |
| 448 |
pos, (U32)litLength, (U32)mlBase+MINMATCH, (U32)offCode); |
| 449 |
} |
| 450 |
#endif |
| 451 |
assert((size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart) < seqStorePtr->maxNbSeq); |
| 452 |
/* copy Literals */ |
| 453 |
assert(seqStorePtr->maxNbLit <= 128 KB); |
| 454 |
assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit); |
| 455 |
assert(literals + litLength <= litLimit); |
| 456 |
if (litEnd <= litLimit_w) { |
| 457 |
/* Common case we can use wildcopy. |
| 458 |
* First copy 16 bytes, because literals are likely short. |
| 459 |
*/ |
| 460 |
assert(WILDCOPY_OVERLENGTH >= 16); |
| 461 |
ZSTD_copy16(seqStorePtr->lit, literals); |
| 462 |
if (litLength > 16) { |
| 463 |
ZSTD_wildcopy(seqStorePtr->lit+16, literals+16, (ptrdiff_t)litLength-16, ZSTD_no_overlap); |
| 464 |
} |
| 465 |
} else { |
| 466 |
ZSTD_safecopyLiterals(seqStorePtr->lit, literals, litEnd, litLimit_w); |
| 467 |
} |
| 468 |
seqStorePtr->lit += litLength; |
| 469 |
|
| 470 |
/* literal Length */ |
| 471 |
if (litLength>0xFFFF) { |
| 472 |
assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */ |
| 473 |
seqStorePtr->longLengthID = 1; |
| 474 |
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); |
| 475 |
} |
| 476 |
seqStorePtr->sequences[0].litLength = (U16)litLength; |
| 477 |
|
| 478 |
/* match offset */ |
| 479 |
seqStorePtr->sequences[0].offset = offCode + 1; |
| 480 |
|
| 481 |
/* match Length */ |
| 482 |
if (mlBase>0xFFFF) { |
| 483 |
assert(seqStorePtr->longLengthID == 0); /* there can only be a single long length */ |
| 484 |
seqStorePtr->longLengthID = 2; |
| 485 |
seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart); |
| 486 |
} |
| 487 |
seqStorePtr->sequences[0].matchLength = (U16)mlBase; |
| 488 |
|
| 489 |
seqStorePtr->sequences++; |
| 490 |
} |
| 491 |
|
| 492 |
|
| 493 |
/*-************************************* |
| 494 |
* Match length counter |
| 495 |
***************************************/ |
| 496 |
static unsigned ZSTD_NbCommonBytes (size_t val) |
| 497 |
{ |
| 498 |
if (MEM_isLittleEndian()) { |
| 499 |
if (MEM_64bits()) { |
| 500 |
# if defined(_MSC_VER) && defined(_WIN64) |
| 501 |
unsigned long r = 0; |
| 502 |
return _BitScanForward64( &r, (U64)val ) ? (unsigned)(r >> 3) : 0; |
| 503 |
# elif defined(__GNUC__) && (__GNUC__ >= 4) |
| 504 |
return (__builtin_ctzll((U64)val) >> 3); |
| 505 |
# else |
| 506 |
static const int DeBruijnBytePos[64] = { 0, 0, 0, 0, 0, 1, 1, 2, |
| 507 |
0, 3, 1, 3, 1, 4, 2, 7, |
| 508 |
0, 2, 3, 6, 1, 5, 3, 5, |
| 509 |
1, 3, 4, 4, 2, 5, 6, 7, |
| 510 |
7, 0, 1, 2, 3, 3, 4, 6, |
| 511 |
2, 6, 5, 5, 3, 4, 5, 6, |
| 512 |
7, 1, 2, 4, 6, 4, 4, 5, |
| 513 |
7, 2, 6, 5, 7, 6, 7, 7 }; |
| 514 |
return DeBruijnBytePos[((U64)((val & -(long long)val) * 0x0218A392CDABBD3FULL)) >> 58]; |
| 515 |
# endif |
| 516 |
} else { /* 32 bits */ |
| 517 |
# if defined(_MSC_VER) |
| 518 |
unsigned long r=0; |
| 519 |
return _BitScanForward( &r, (U32)val ) ? (unsigned)(r >> 3) : 0; |
| 520 |
# elif defined(__GNUC__) && (__GNUC__ >= 3) |
| 521 |
return (__builtin_ctz((U32)val) >> 3); |
| 522 |
# else |
| 523 |
static const int DeBruijnBytePos[32] = { 0, 0, 3, 0, 3, 1, 3, 0, |
| 524 |
3, 2, 2, 1, 3, 2, 0, 1, |
| 525 |
3, 3, 1, 2, 2, 2, 2, 0, |
| 526 |
3, 1, 2, 0, 1, 0, 1, 1 }; |
| 527 |
return DeBruijnBytePos[((U32)((val & -(S32)val) * 0x077CB531U)) >> 27]; |
| 528 |
# endif |
| 529 |
} |
| 530 |
} else { /* Big Endian CPU */ |
| 531 |
if (MEM_64bits()) { |
| 532 |
# if defined(_MSC_VER) && defined(_WIN64) |
| 533 |
unsigned long r = 0; |
| 534 |
return _BitScanReverse64( &r, val ) ? (unsigned)(r >> 3) : 0; |
| 535 |
# elif defined(__GNUC__) && (__GNUC__ >= 4) |
| 536 |
return (__builtin_clzll(val) >> 3); |
| 537 |
# else |
| 538 |
unsigned r; |
| 539 |
const unsigned n32 = sizeof(size_t)*4; /* calculate this way due to compiler complaining in 32-bits mode */ |
| 540 |
if (!(val>>n32)) { r=4; } else { r=0; val>>=n32; } |
| 541 |
if (!(val>>16)) { r+=2; val>>=8; } else { val>>=24; } |
| 542 |
r += (!val); |
| 543 |
return r; |
| 544 |
# endif |
| 545 |
} else { /* 32 bits */ |
| 546 |
# if defined(_MSC_VER) |
| 547 |
unsigned long r = 0; |
| 548 |
return _BitScanReverse( &r, (unsigned long)val ) ? (unsigned)(r >> 3) : 0; |
| 549 |
# elif defined(__GNUC__) && (__GNUC__ >= 3) |
| 550 |
return (__builtin_clz((U32)val) >> 3); |
| 551 |
# else |
| 552 |
unsigned r; |
| 553 |
if (!(val>>16)) { r=2; val>>=8; } else { r=0; val>>=24; } |
| 554 |
r += (!val); |
| 555 |
return r; |
| 556 |
# endif |
| 557 |
} } |
| 558 |
} |
| 559 |
|
| 560 |
|
| 561 |
MEM_STATIC size_t ZSTD_count(const BYTE* pIn, const BYTE* pMatch, const BYTE* const pInLimit) |
| 562 |
{ |
| 563 |
const BYTE* const pStart = pIn; |
| 564 |
const BYTE* const pInLoopLimit = pInLimit - (sizeof(size_t)-1); |
| 565 |
|
| 566 |
if (pIn < pInLoopLimit) { |
| 567 |
{ size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn); |
| 568 |
if (diff) return ZSTD_NbCommonBytes(diff); } |
| 569 |
pIn+=sizeof(size_t); pMatch+=sizeof(size_t); |
| 570 |
while (pIn < pInLoopLimit) { |
| 571 |
size_t const diff = MEM_readST(pMatch) ^ MEM_readST(pIn); |
| 572 |
if (!diff) { pIn+=sizeof(size_t); pMatch+=sizeof(size_t); continue; } |
| 573 |
pIn += ZSTD_NbCommonBytes(diff); |
| 574 |
return (size_t)(pIn - pStart); |
| 575 |
} } |
| 576 |
if (MEM_64bits() && (pIn<(pInLimit-3)) && (MEM_read32(pMatch) == MEM_read32(pIn))) { pIn+=4; pMatch+=4; } |
| 577 |
if ((pIn<(pInLimit-1)) && (MEM_read16(pMatch) == MEM_read16(pIn))) { pIn+=2; pMatch+=2; } |
| 578 |
if ((pIn<pInLimit) && (*pMatch == *pIn)) pIn++; |
| 579 |
return (size_t)(pIn - pStart); |
| 580 |
} |
| 581 |
|
| 582 |
/** ZSTD_count_2segments() : |
| 583 |
* can count match length with `ip` & `match` in 2 different segments. |
| 584 |
* convention : on reaching mEnd, match count continue starting from iStart |
| 585 |
*/ |
| 586 |
MEM_STATIC size_t |
| 587 |
ZSTD_count_2segments(const BYTE* ip, const BYTE* match, |
| 588 |
const BYTE* iEnd, const BYTE* mEnd, const BYTE* iStart) |
| 589 |
{ |
| 590 |
const BYTE* const vEnd = MIN( ip + (mEnd - match), iEnd); |
| 591 |
size_t const matchLength = ZSTD_count(ip, match, vEnd); |
| 592 |
if (match + matchLength != mEnd) return matchLength; |
| 593 |
DEBUGLOG(7, "ZSTD_count_2segments: found a 2-parts match (current length==%zu)", matchLength); |
| 594 |
DEBUGLOG(7, "distance from match beginning to end dictionary = %zi", mEnd - match); |
| 595 |
DEBUGLOG(7, "distance from current pos to end buffer = %zi", iEnd - ip); |
| 596 |
DEBUGLOG(7, "next byte : ip==%02X, istart==%02X", ip[matchLength], *iStart); |
| 597 |
DEBUGLOG(7, "final match length = %zu", matchLength + ZSTD_count(ip+matchLength, iStart, iEnd)); |
| 598 |
return matchLength + ZSTD_count(ip+matchLength, iStart, iEnd); |
| 599 |
} |
| 600 |
|
| 601 |
|
| 602 |
/*-************************************* |
| 603 |
* Hashes |
| 604 |
***************************************/ |
| 605 |
static const U32 prime3bytes = 506832829U; |
| 606 |
static U32 ZSTD_hash3(U32 u, U32 h) { return ((u << (32-24)) * prime3bytes) >> (32-h) ; } |
| 607 |
MEM_STATIC size_t ZSTD_hash3Ptr(const void* ptr, U32 h) { return ZSTD_hash3(MEM_readLE32(ptr), h); } /* only in zstd_opt.h */ |
| 608 |
|
| 609 |
static const U32 prime4bytes = 2654435761U; |
| 610 |
static U32 ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32-h) ; } |
| 611 |
static size_t ZSTD_hash4Ptr(const void* ptr, U32 h) { return ZSTD_hash4(MEM_read32(ptr), h); } |
| 612 |
|
| 613 |
static const U64 prime5bytes = 889523592379ULL; |
| 614 |
static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u << (64-40)) * prime5bytes) >> (64-h)) ; } |
| 615 |
static size_t ZSTD_hash5Ptr(const void* p, U32 h) { return ZSTD_hash5(MEM_readLE64(p), h); } |
| 616 |
|
| 617 |
static const U64 prime6bytes = 227718039650203ULL; |
| 618 |
static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64-48)) * prime6bytes) >> (64-h)) ; } |
| 619 |
static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); } |
| 620 |
|
| 621 |
static const U64 prime7bytes = 58295818150454627ULL; |
| 622 |
static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u << (64-56)) * prime7bytes) >> (64-h)) ; } |
| 623 |
static size_t ZSTD_hash7Ptr(const void* p, U32 h) { return ZSTD_hash7(MEM_readLE64(p), h); } |
| 624 |
|
| 625 |
static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL; |
| 626 |
static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; } |
| 627 |
static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); } |
| 628 |
|
| 629 |
MEM_STATIC size_t ZSTD_hashPtr(const void* p, U32 hBits, U32 mls) |
| 630 |
{ |
| 631 |
switch(mls) |
| 632 |
{ |
| 633 |
default: |
| 634 |
case 4: return ZSTD_hash4Ptr(p, hBits); |
| 635 |
case 5: return ZSTD_hash5Ptr(p, hBits); |
| 636 |
case 6: return ZSTD_hash6Ptr(p, hBits); |
| 637 |
case 7: return ZSTD_hash7Ptr(p, hBits); |
| 638 |
case 8: return ZSTD_hash8Ptr(p, hBits); |
| 639 |
} |
| 640 |
} |
| 641 |
|
| 642 |
/** ZSTD_ipow() : |
| 643 |
* Return base^exponent. |
| 644 |
*/ |
| 645 |
static U64 ZSTD_ipow(U64 base, U64 exponent) |
| 646 |
{ |
| 647 |
U64 power = 1; |
| 648 |
while (exponent) { |
| 649 |
if (exponent & 1) power *= base; |
| 650 |
exponent >>= 1; |
| 651 |
base *= base; |
| 652 |
} |
| 653 |
return power; |
| 654 |
} |
| 655 |
|
| 656 |
#define ZSTD_ROLL_HASH_CHAR_OFFSET 10 |
| 657 |
|
| 658 |
/** ZSTD_rollingHash_append() : |
| 659 |
* Add the buffer to the hash value. |
| 660 |
*/ |
| 661 |
static U64 ZSTD_rollingHash_append(U64 hash, void const* buf, size_t size) |
| 662 |
{ |
| 663 |
BYTE const* istart = (BYTE const*)buf; |
| 664 |
size_t pos; |
| 665 |
for (pos = 0; pos < size; ++pos) { |
| 666 |
hash *= prime8bytes; |
| 667 |
hash += istart[pos] + ZSTD_ROLL_HASH_CHAR_OFFSET; |
| 668 |
} |
| 669 |
return hash; |
| 670 |
} |
| 671 |
|
| 672 |
/** ZSTD_rollingHash_compute() : |
| 673 |
* Compute the rolling hash value of the buffer. |
| 674 |
*/ |
| 675 |
MEM_STATIC U64 ZSTD_rollingHash_compute(void const* buf, size_t size) |
| 676 |
{ |
| 677 |
return ZSTD_rollingHash_append(0, buf, size); |
| 678 |
} |
| 679 |
|
| 680 |
/** ZSTD_rollingHash_primePower() : |
| 681 |
* Compute the primePower to be passed to ZSTD_rollingHash_rotate() for a hash |
| 682 |
* over a window of length bytes. |
| 683 |
*/ |
| 684 |
MEM_STATIC U64 ZSTD_rollingHash_primePower(U32 length) |
| 685 |
{ |
| 686 |
return ZSTD_ipow(prime8bytes, length - 1); |
| 687 |
} |
| 688 |
|
| 689 |
/** ZSTD_rollingHash_rotate() : |
| 690 |
* Rotate the rolling hash by one byte. |
| 691 |
*/ |
| 692 |
MEM_STATIC U64 ZSTD_rollingHash_rotate(U64 hash, BYTE toRemove, BYTE toAdd, U64 primePower) |
| 693 |
{ |
| 694 |
hash -= (toRemove + ZSTD_ROLL_HASH_CHAR_OFFSET) * primePower; |
| 695 |
hash *= prime8bytes; |
| 696 |
hash += toAdd + ZSTD_ROLL_HASH_CHAR_OFFSET; |
| 697 |
return hash; |
| 698 |
} |
| 699 |
|
| 700 |
/*-************************************* |
| 701 |
* Round buffer management |
| 702 |
***************************************/ |
| 703 |
#if (ZSTD_WINDOWLOG_MAX_64 > 31) |
| 704 |
# error "ZSTD_WINDOWLOG_MAX is too large : would overflow ZSTD_CURRENT_MAX" |
| 705 |
#endif |
| 706 |
/* Max current allowed */ |
| 707 |
#define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX)) |
| 708 |
/* Maximum chunk size before overflow correction needs to be called again */ |
| 709 |
#define ZSTD_CHUNKSIZE_MAX \ |
| 710 |
( ((U32)-1) /* Maximum ending current index */ \ |
| 711 |
- ZSTD_CURRENT_MAX) /* Maximum beginning lowLimit */ |
| 712 |
|
| 713 |
/** |
| 714 |
* ZSTD_window_clear(): |
| 715 |
* Clears the window containing the history by simply setting it to empty. |
| 716 |
*/ |
| 717 |
MEM_STATIC void ZSTD_window_clear(ZSTD_window_t* window) |
| 718 |
{ |
| 719 |
size_t const endT = (size_t)(window->nextSrc - window->base); |
| 720 |
U32 const end = (U32)endT; |
| 721 |
|
| 722 |
window->lowLimit = end; |
| 723 |
window->dictLimit = end; |
| 724 |
} |
| 725 |
|
| 726 |
/** |
| 727 |
* ZSTD_window_hasExtDict(): |
| 728 |
* Returns non-zero if the window has a non-empty extDict. |
| 729 |
*/ |
| 730 |
MEM_STATIC U32 ZSTD_window_hasExtDict(ZSTD_window_t const window) |
| 731 |
{ |
| 732 |
return window.lowLimit < window.dictLimit; |
| 733 |
} |
| 734 |
|
| 735 |
/** |
| 736 |
* ZSTD_matchState_dictMode(): |
| 737 |
* Inspects the provided matchState and figures out what dictMode should be |
| 738 |
* passed to the compressor. |
| 739 |
*/ |
| 740 |
MEM_STATIC ZSTD_dictMode_e ZSTD_matchState_dictMode(const ZSTD_matchState_t *ms) |
| 741 |
{ |
| 742 |
return ZSTD_window_hasExtDict(ms->window) ? |
| 743 |
ZSTD_extDict : |
| 744 |
ms->dictMatchState != NULL ? |
| 745 |
ZSTD_dictMatchState : |
| 746 |
ZSTD_noDict; |
| 747 |
} |
| 748 |
|
| 749 |
/** |
| 750 |
* ZSTD_window_needOverflowCorrection(): |
| 751 |
* Returns non-zero if the indices are getting too large and need overflow |
| 752 |
* protection. |
| 753 |
*/ |
| 754 |
MEM_STATIC U32 ZSTD_window_needOverflowCorrection(ZSTD_window_t const window, |
| 755 |
void const* srcEnd) |
| 756 |
{ |
| 757 |
U32 const current = (U32)((BYTE const*)srcEnd - window.base); |
| 758 |
return current > ZSTD_CURRENT_MAX; |
| 759 |
} |
| 760 |
|
| 761 |
/** |
| 762 |
* ZSTD_window_correctOverflow(): |
| 763 |
* Reduces the indices to protect from index overflow. |
| 764 |
* Returns the correction made to the indices, which must be applied to every |
| 765 |
* stored index. |
| 766 |
* |
| 767 |
* The least significant cycleLog bits of the indices must remain the same, |
| 768 |
* which may be 0. Every index up to maxDist in the past must be valid. |
| 769 |
* NOTE: (maxDist & cycleMask) must be zero. |
| 770 |
*/ |
| 771 |
MEM_STATIC U32 ZSTD_window_correctOverflow(ZSTD_window_t* window, U32 cycleLog, |
| 772 |
U32 maxDist, void const* src) |
| 773 |
{ |
| 774 |
/* preemptive overflow correction: |
| 775 |
* 1. correction is large enough: |
| 776 |
* lowLimit > (3<<29) ==> current > 3<<29 + 1<<windowLog |
| 777 |
* 1<<windowLog <= newCurrent < 1<<chainLog + 1<<windowLog |
| 778 |
* |
| 779 |
* current - newCurrent |
| 780 |
* > (3<<29 + 1<<windowLog) - (1<<windowLog + 1<<chainLog) |
| 781 |
* > (3<<29) - (1<<chainLog) |
| 782 |
* > (3<<29) - (1<<30) (NOTE: chainLog <= 30) |
| 783 |
* > 1<<29 |
| 784 |
* |
| 785 |
* 2. (ip+ZSTD_CHUNKSIZE_MAX - cctx->base) doesn't overflow: |
| 786 |
* After correction, current is less than (1<<chainLog + 1<<windowLog). |
| 787 |
* In 64-bit mode we are safe, because we have 64-bit ptrdiff_t. |
| 788 |
* In 32-bit mode we are safe, because (chainLog <= 29), so |
| 789 |
* ip+ZSTD_CHUNKSIZE_MAX - cctx->base < 1<<32. |
| 790 |
* 3. (cctx->lowLimit + 1<<windowLog) < 1<<32: |
| 791 |
* windowLog <= 31 ==> 3<<29 + 1<<windowLog < 7<<29 < 1<<32. |
| 792 |
*/ |
| 793 |
U32 const cycleMask = (1U << cycleLog) - 1; |
| 794 |
U32 const current = (U32)((BYTE const*)src - window->base); |
| 795 |
U32 const currentCycle0 = current & cycleMask; |
| 796 |
/* Exclude zero so that newCurrent - maxDist >= 1. */ |
| 797 |
U32 const currentCycle1 = currentCycle0 == 0 ? (1U << cycleLog) : currentCycle0; |
| 798 |
U32 const newCurrent = currentCycle1 + maxDist; |
| 799 |
U32 const correction = current - newCurrent; |
| 800 |
assert((maxDist & cycleMask) == 0); |
| 801 |
assert(current > newCurrent); |
| 802 |
/* Loose bound, should be around 1<<29 (see above) */ |
| 803 |
assert(correction > 1<<28); |
| 804 |
|
| 805 |
window->base += correction; |
| 806 |
window->dictBase += correction; |
| 807 |
if (window->lowLimit <= correction) window->lowLimit = 1; |
| 808 |
else window->lowLimit -= correction; |
| 809 |
if (window->dictLimit <= correction) window->dictLimit = 1; |
| 810 |
else window->dictLimit -= correction; |
| 811 |
|
| 812 |
/* Ensure we can still reference the full window. */ |
| 813 |
assert(newCurrent >= maxDist); |
| 814 |
assert(newCurrent - maxDist >= 1); |
| 815 |
/* Ensure that lowLimit and dictLimit didn't underflow. */ |
| 816 |
assert(window->lowLimit <= newCurrent); |
| 817 |
assert(window->dictLimit <= newCurrent); |
| 818 |
|
| 819 |
DEBUGLOG(4, "Correction of 0x%x bytes to lowLimit=0x%x", correction, |
| 820 |
window->lowLimit); |
| 821 |
return correction; |
| 822 |
} |
| 823 |
|
| 824 |
/** |
| 825 |
* ZSTD_window_enforceMaxDist(): |
| 826 |
* Updates lowLimit so that: |
| 827 |
* (srcEnd - base) - lowLimit == maxDist + loadedDictEnd |
| 828 |
* |
| 829 |
* It ensures index is valid as long as index >= lowLimit. |
| 830 |
* This must be called before a block compression call. |
| 831 |
* |
| 832 |
* loadedDictEnd is only defined if a dictionary is in use for current compression. |
| 833 |
* As the name implies, loadedDictEnd represents the index at end of dictionary. |
| 834 |
* The value lies within context's referential, it can be directly compared to blockEndIdx. |
| 835 |
* |
| 836 |
* If loadedDictEndPtr is NULL, no dictionary is in use, and we use loadedDictEnd == 0. |
| 837 |
* If loadedDictEndPtr is not NULL, we set it to zero after updating lowLimit. |
| 838 |
* This is because dictionaries are allowed to be referenced fully |
| 839 |
* as long as the last byte of the dictionary is in the window. |
| 840 |
* Once input has progressed beyond window size, dictionary cannot be referenced anymore. |
| 841 |
* |
| 842 |
* In normal dict mode, the dictionary lies between lowLimit and dictLimit. |
| 843 |
* In dictMatchState mode, lowLimit and dictLimit are the same, |
| 844 |
* and the dictionary is below them. |
| 845 |
* forceWindow and dictMatchState are therefore incompatible. |
| 846 |
*/ |
| 847 |
MEM_STATIC void |
| 848 |
ZSTD_window_enforceMaxDist(ZSTD_window_t* window, |
| 849 |
const void* blockEnd, |
| 850 |
U32 maxDist, |
| 851 |
U32* loadedDictEndPtr, |
| 852 |
const ZSTD_matchState_t** dictMatchStatePtr) |
| 853 |
{ |
| 854 |
U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base); |
| 855 |
U32 const loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0; |
| 856 |
DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u", |
| 857 |
(unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd); |
| 858 |
|
| 859 |
/* - When there is no dictionary : loadedDictEnd == 0. |
| 860 |
In which case, the test (blockEndIdx > maxDist) is merely to avoid |
| 861 |
overflowing next operation `newLowLimit = blockEndIdx - maxDist`. |
| 862 |
- When there is a standard dictionary : |
| 863 |
Index referential is copied from the dictionary, |
| 864 |
which means it starts from 0. |
| 865 |
In which case, loadedDictEnd == dictSize, |
| 866 |
and it makes sense to compare `blockEndIdx > maxDist + dictSize` |
| 867 |
since `blockEndIdx` also starts from zero. |
| 868 |
- When there is an attached dictionary : |
| 869 |
loadedDictEnd is expressed within the referential of the context, |
| 870 |
so it can be directly compared against blockEndIdx. |
| 871 |
*/ |
| 872 |
if (blockEndIdx > maxDist + loadedDictEnd) { |
| 873 |
U32 const newLowLimit = blockEndIdx - maxDist; |
| 874 |
if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit; |
| 875 |
if (window->dictLimit < window->lowLimit) { |
| 876 |
DEBUGLOG(5, "Update dictLimit to match lowLimit, from %u to %u", |
| 877 |
(unsigned)window->dictLimit, (unsigned)window->lowLimit); |
| 878 |
window->dictLimit = window->lowLimit; |
| 879 |
} |
| 880 |
/* On reaching window size, dictionaries are invalidated */ |
| 881 |
if (loadedDictEndPtr) *loadedDictEndPtr = 0; |
| 882 |
if (dictMatchStatePtr) *dictMatchStatePtr = NULL; |
| 883 |
} |
| 884 |
} |
| 885 |
|
| 886 |
/* Similar to ZSTD_window_enforceMaxDist(), |
| 887 |
* but only invalidates dictionary |
| 888 |
* when input progresses beyond window size. |
| 889 |
* assumption : loadedDictEndPtr and dictMatchStatePtr are valid (non NULL) |
| 890 |
* loadedDictEnd uses same referential as window->base |
| 891 |
* maxDist is the window size */ |
| 892 |
MEM_STATIC void |
| 893 |
ZSTD_checkDictValidity(const ZSTD_window_t* window, |
| 894 |
const void* blockEnd, |
| 895 |
U32 maxDist, |
| 896 |
U32* loadedDictEndPtr, |
| 897 |
const ZSTD_matchState_t** dictMatchStatePtr) |
| 898 |
{ |
| 899 |
assert(loadedDictEndPtr != NULL); |
| 900 |
assert(dictMatchStatePtr != NULL); |
| 901 |
{ U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base); |
| 902 |
U32 const loadedDictEnd = *loadedDictEndPtr; |
| 903 |
DEBUGLOG(5, "ZSTD_checkDictValidity: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u", |
| 904 |
(unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd); |
| 905 |
assert(blockEndIdx >= loadedDictEnd); |
| 906 |
|
| 907 |
if (blockEndIdx > loadedDictEnd + maxDist) { |
| 908 |
/* On reaching window size, dictionaries are invalidated. |
| 909 |
* For simplification, if window size is reached anywhere within next block, |
| 910 |
* the dictionary is invalidated for the full block. |
| 911 |
*/ |
| 912 |
DEBUGLOG(6, "invalidating dictionary for current block (distance > windowSize)"); |
| 913 |
*loadedDictEndPtr = 0; |
| 914 |
*dictMatchStatePtr = NULL; |
| 915 |
} else { |
| 916 |
if (*loadedDictEndPtr != 0) { |
| 917 |
DEBUGLOG(6, "dictionary considered valid for current block"); |
| 918 |
} } } |
| 919 |
} |
| 920 |
|
| 921 |
MEM_STATIC void ZSTD_window_init(ZSTD_window_t* window) { |
| 922 |
memset(window, 0, sizeof(*window)); |
| 923 |
window->base = (BYTE const*)""; |
| 924 |
window->dictBase = (BYTE const*)""; |
| 925 |
window->dictLimit = 1; /* start from 1, so that 1st position is valid */ |
| 926 |
window->lowLimit = 1; /* it ensures first and later CCtx usages compress the same */ |
| 927 |
window->nextSrc = window->base + 1; /* see issue #1241 */ |
| 928 |
} |
| 929 |
|
| 930 |
/** |
| 931 |
* ZSTD_window_update(): |
| 932 |
* Updates the window by appending [src, src + srcSize) to the window. |
| 933 |
* If it is not contiguous, the current prefix becomes the extDict, and we |
| 934 |
* forget about the extDict. Handles overlap of the prefix and extDict. |
| 935 |
* Returns non-zero if the segment is contiguous. |
| 936 |
*/ |
| 937 |
MEM_STATIC U32 ZSTD_window_update(ZSTD_window_t* window, |
| 938 |
void const* src, size_t srcSize) |
| 939 |
{ |
| 940 |
BYTE const* const ip = (BYTE const*)src; |
| 941 |
U32 contiguous = 1; |
| 942 |
DEBUGLOG(5, "ZSTD_window_update"); |
| 943 |
if (srcSize == 0) |
| 944 |
return contiguous; |
| 945 |
assert(window->base != NULL); |
| 946 |
assert(window->dictBase != NULL); |
| 947 |
/* Check if blocks follow each other */ |
| 948 |
if (src != window->nextSrc) { |
| 949 |
/* not contiguous */ |
| 950 |
size_t const distanceFromBase = (size_t)(window->nextSrc - window->base); |
| 951 |
DEBUGLOG(5, "Non contiguous blocks, new segment starts at %u", window->dictLimit); |
| 952 |
window->lowLimit = window->dictLimit; |
| 953 |
assert(distanceFromBase == (size_t)(U32)distanceFromBase); /* should never overflow */ |
| 954 |
window->dictLimit = (U32)distanceFromBase; |
| 955 |
window->dictBase = window->base; |
| 956 |
window->base = ip - distanceFromBase; |
| 957 |
/* ms->nextToUpdate = window->dictLimit; */ |
| 958 |
if (window->dictLimit - window->lowLimit < HASH_READ_SIZE) window->lowLimit = window->dictLimit; /* too small extDict */ |
| 959 |
contiguous = 0; |
| 960 |
} |
| 961 |
window->nextSrc = ip + srcSize; |
| 962 |
/* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */ |
| 963 |
if ( (ip+srcSize > window->dictBase + window->lowLimit) |
| 964 |
& (ip < window->dictBase + window->dictLimit)) { |
| 965 |
ptrdiff_t const highInputIdx = (ip + srcSize) - window->dictBase; |
| 966 |
U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)window->dictLimit) ? window->dictLimit : (U32)highInputIdx; |
| 967 |
window->lowLimit = lowLimitMax; |
| 968 |
DEBUGLOG(5, "Overlapping extDict and input : new lowLimit = %u", window->lowLimit); |
| 969 |
} |
| 970 |
return contiguous; |
| 971 |
} |
| 972 |
|
| 973 |
/** |
| 974 |
* Returns the lowest allowed match index. It may either be in the ext-dict or the prefix. |
| 975 |
*/ |
| 976 |
MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 current, unsigned windowLog) |
| 977 |
{ |
| 978 |
U32 const maxDistance = 1U << windowLog; |
| 979 |
U32 const lowestValid = ms->window.lowLimit; |
| 980 |
U32 const withinWindow = (current - lowestValid > maxDistance) ? current - maxDistance : lowestValid; |
| 981 |
U32 const isDictionary = (ms->loadedDictEnd != 0); |
| 982 |
U32 const matchLowest = isDictionary ? lowestValid : withinWindow; |
| 983 |
return matchLowest; |
| 984 |
} |
| 985 |
|
| 986 |
/** |
| 987 |
* Returns the lowest allowed match index in the prefix. |
| 988 |
*/ |
| 989 |
MEM_STATIC U32 ZSTD_getLowestPrefixIndex(const ZSTD_matchState_t* ms, U32 current, unsigned windowLog) |
| 990 |
{ |
| 991 |
U32 const maxDistance = 1U << windowLog; |
| 992 |
U32 const lowestValid = ms->window.dictLimit; |
| 993 |
U32 const withinWindow = (current - lowestValid > maxDistance) ? current - maxDistance : lowestValid; |
| 994 |
U32 const isDictionary = (ms->loadedDictEnd != 0); |
| 995 |
U32 const matchLowest = isDictionary ? lowestValid : withinWindow; |
| 996 |
return matchLowest; |
| 997 |
} |
| 998 |
|
| 999 |
|
| 1000 |
|
| 1001 |
/* debug functions */ |
| 1002 |
#if (DEBUGLEVEL>=2) |
| 1003 |
|
| 1004 |
MEM_STATIC double ZSTD_fWeight(U32 rawStat) |
| 1005 |
{ |
| 1006 |
U32 const fp_accuracy = 8; |
| 1007 |
U32 const fp_multiplier = (1 << fp_accuracy); |
| 1008 |
U32 const newStat = rawStat + 1; |
| 1009 |
U32 const hb = ZSTD_highbit32(newStat); |
| 1010 |
U32 const BWeight = hb * fp_multiplier; |
| 1011 |
U32 const FWeight = (newStat << fp_accuracy) >> hb; |
| 1012 |
U32 const weight = BWeight + FWeight; |
| 1013 |
assert(hb + fp_accuracy < 31); |
| 1014 |
return (double)weight / fp_multiplier; |
| 1015 |
} |
| 1016 |
|
| 1017 |
/* display a table content, |
| 1018 |
* listing each element, its frequency, and its predicted bit cost */ |
| 1019 |
MEM_STATIC void ZSTD_debugTable(const U32* table, U32 max) |
| 1020 |
{ |
| 1021 |
unsigned u, sum; |
| 1022 |
for (u=0, sum=0; u<=max; u++) sum += table[u]; |
| 1023 |
DEBUGLOG(2, "total nb elts: %u", sum); |
| 1024 |
for (u=0; u<=max; u++) { |
| 1025 |
DEBUGLOG(2, "%2u: %5u (%.2f)", |
| 1026 |
u, table[u], ZSTD_fWeight(sum) - ZSTD_fWeight(table[u]) ); |
| 1027 |
} |
| 1028 |
} |
| 1029 |
|
| 1030 |
#endif |
| 1031 |
|
| 1032 |
|
| 1033 |
#if defined (__cplusplus) |
| 1034 |
} |
| 1035 |
#endif |
| 1036 |
|
| 1037 |
/* =============================================================== |
| 1038 |
* Shared internal declarations |
| 1039 |
* These prototypes may be called from sources not in lib/compress |
| 1040 |
* =============================================================== */ |
| 1041 |
|
| 1042 |
/* ZSTD_loadCEntropy() : |
| 1043 |
* dict : must point at beginning of a valid zstd dictionary. |
| 1044 |
* return : size of dictionary header (size of magic number + dict ID + entropy tables) |
| 1045 |
* assumptions : magic number supposed already checked |
| 1046 |
* and dictSize >= 8 */ |
| 1047 |
size_t ZSTD_loadCEntropy(ZSTD_compressedBlockState_t* bs, void* workspace, |
| 1048 |
short* offcodeNCount, unsigned* offcodeMaxValue, |
| 1049 |
const void* const dict, size_t dictSize); |
| 1050 |
|
| 1051 |
void ZSTD_reset_compressedBlockState(ZSTD_compressedBlockState_t* bs); |
| 1052 |
|
| 1053 |
/* ============================================================== |
| 1054 |
* Private declarations |
| 1055 |
* These prototypes shall only be called from within lib/compress |
| 1056 |
* ============================================================== */ |
| 1057 |
|
| 1058 |
/* ZSTD_getCParamsFromCCtxParams() : |
| 1059 |
* cParams are built depending on compressionLevel, src size hints, |
| 1060 |
* LDM and manually set compression parameters. |
| 1061 |
* Note: srcSizeHint == 0 means 0! |
| 1062 |
*/ |
| 1063 |
ZSTD_compressionParameters ZSTD_getCParamsFromCCtxParams( |
| 1064 |
const ZSTD_CCtx_params* CCtxParams, U64 srcSizeHint, size_t dictSize); |
| 1065 |
|
| 1066 |
/*! ZSTD_initCStream_internal() : |
| 1067 |
* Private use only. Init streaming operation. |
| 1068 |
* expects params to be valid. |
| 1069 |
* must receive dict, or cdict, or none, but not both. |
| 1070 |
* @return : 0, or an error code */ |
| 1071 |
size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs, |
| 1072 |
const void* dict, size_t dictSize, |
| 1073 |
const ZSTD_CDict* cdict, |
| 1074 |
const ZSTD_CCtx_params* params, unsigned long long pledgedSrcSize); |
| 1075 |
|
| 1076 |
void ZSTD_resetSeqStore(seqStore_t* ssPtr); |
| 1077 |
|
| 1078 |
/*! ZSTD_getCParamsFromCDict() : |
| 1079 |
* as the name implies */ |
| 1080 |
ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict); |
| 1081 |
|
| 1082 |
/* ZSTD_compressBegin_advanced_internal() : |
| 1083 |
* Private use only. To be called from zstdmt_compress.c. */ |
| 1084 |
size_t ZSTD_compressBegin_advanced_internal(ZSTD_CCtx* cctx, |
| 1085 |
const void* dict, size_t dictSize, |
| 1086 |
ZSTD_dictContentType_e dictContentType, |
| 1087 |
ZSTD_dictTableLoadMethod_e dtlm, |
| 1088 |
const ZSTD_CDict* cdict, |
| 1089 |
const ZSTD_CCtx_params* params, |
| 1090 |
unsigned long long pledgedSrcSize); |
| 1091 |
|
| 1092 |
/* ZSTD_compress_advanced_internal() : |
| 1093 |
* Private use only. To be called from zstdmt_compress.c. */ |
| 1094 |
size_t ZSTD_compress_advanced_internal(ZSTD_CCtx* cctx, |
| 1095 |
void* dst, size_t dstCapacity, |
| 1096 |
const void* src, size_t srcSize, |
| 1097 |
const void* dict,size_t dictSize, |
| 1098 |
const ZSTD_CCtx_params* params); |
| 1099 |
|
| 1100 |
|
| 1101 |
/* ZSTD_writeLastEmptyBlock() : |
| 1102 |
* output an empty Block with end-of-frame mark to complete a frame |
| 1103 |
* @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h)) |
| 1104 |
* or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize) |
| 1105 |
*/ |
| 1106 |
size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity); |
| 1107 |
|
| 1108 |
|
| 1109 |
/* ZSTD_referenceExternalSequences() : |
| 1110 |
* Must be called before starting a compression operation. |
| 1111 |
* seqs must parse a prefix of the source. |
| 1112 |
* This cannot be used when long range matching is enabled. |
| 1113 |
* Zstd will use these sequences, and pass the literals to a secondary block |
| 1114 |
* compressor. |
| 1115 |
* @return : An error code on failure. |
| 1116 |
* NOTE: seqs are not verified! Invalid sequences can cause out-of-bounds memory |
| 1117 |
* access and data corruption. |
| 1118 |
*/ |
| 1119 |
size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq); |
| 1120 |
|
| 1121 |
/** ZSTD_cycleLog() : |
| 1122 |
* condition for correct operation : hashLog > 1 */ |
| 1123 |
U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat); |
| 1124 |
|
| 1125 |
#endif /* ZSTD_COMPRESS_H */ |