diff options
Diffstat (limited to 'LzFind.c')
| -rw-r--r-- | LzFind.c | 629 |
1 files changed, 629 insertions, 0 deletions
diff --git a/LzFind.c b/LzFind.c new file mode 100644 index 0000000..c312927 --- /dev/null +++ b/LzFind.c @@ -0,0 +1,629 @@ +/* LzFind.c -- Match finder for LZ algorithms +2009-04-22 : Igor Pavlov : Public domain */ + +#define _FILE_OFFSET_BITS 64 + +#include <errno.h> +#include <stdbool.h> +#include <stddef.h> +#include <stdint.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#include "lzip.h" +#include "LzFind.h" + +#define kHash2Size (1 << 10) +#define kHash3Size (1 << 16) +#define kHash4Size (1 << 20) + +#define kFix3HashSize (kHash2Size) +#define kFix4HashSize (kHash2Size + kHash3Size) + +#define HASH2_CALC hashValue = cur[0] | ((uint32_t)cur[1] << 8); + +#define HASH3_CALC { \ + uint32_t temp = crc32[cur[0]] ^ cur[1]; \ + hash2Value = temp & (kHash2Size - 1); \ + hashValue = (temp ^ ((uint32_t)cur[2] << 8)) & p->hashMask; } + +#define HASH4_CALC { \ + uint32_t temp = crc32[cur[0]] ^ cur[1]; \ + hash2Value = temp & (kHash2Size - 1); \ + hash3Value = (temp ^ ((uint32_t)cur[2] << 8)) & (kHash3Size - 1); \ + hashValue = (temp ^ ((uint32_t)cur[2] << 8) ^ (crc32[cur[3]] << 5)) & p->hashMask; } + +#define kEmptyHashValue 0 +#define kMaxValForNormalize ((uint32_t)0xFFFFFFFF) +#define kNormalizeStepMin (1 << 10) /* it must be power of 2 */ +#define kNormalizeMask (~(kNormalizeStepMin - 1)) + +#define kStartMaxLen 3 + + +static void Mf_ReadBlock(CMatchFinder *p) +{ + if (p->streamEndWasReached || p->result != SZ_OK) + return; + for (;;) + { + uint8_t * const dest = p->buffer + (p->streamPos - p->pos); + const int size = (p->bufferBase + p->blockSize - dest); + int rd; + if (size == 0) + return; + rd = readblock( p->infd, dest, size ); + if (rd != size && errno) + { p->result = SZ_ERROR_READ; return; } + if (rd == 0) + { + p->streamEndWasReached = true; + return; + } + CRC32_update_buf( &p->crc, dest, rd ); + p->streamPos += rd; + if (p->streamPos - p->pos > p->keepSizeAfter) + return; + } +} + + +static void Mf_CheckAndMoveAndRead(CMatchFinder *p) +{ + if ((uint32_t)(p->bufferBase + p->blockSize - p->buffer) <= p->keepSizeAfter) + { + memmove(p->bufferBase, + p->buffer - p->keepSizeBefore, + p->streamPos - p->pos + p->keepSizeBefore); + p->buffer = p->bufferBase + p->keepSizeBefore; + } + Mf_ReadBlock(p); +} + + +void Mf_Free(CMatchFinder *p) +{ + free(p->hash); + p->hash = 0; + free(p->bufferBase); + p->bufferBase = 0; +} + +static CLzRef* AllocRefs(uint32_t num) +{ + uint32_t sizeInBytes = num * sizeof(CLzRef); + if (sizeInBytes / sizeof(CLzRef) != num) + return 0; + return (CLzRef *)malloc(sizeInBytes); +} + +static void Mf_SetLimits(CMatchFinder *p) +{ + uint32_t limit = kMaxValForNormalize - p->pos; + uint32_t limit2 = p->cyclicBufferSize - p->cyclicBufferPos; + if (limit2 < limit) + limit = limit2; + limit2 = p->streamPos - p->pos; + if (limit2 <= p->keepSizeAfter) + { + if (limit2 > 0) + limit2 = 1; + } + else + limit2 -= p->keepSizeAfter; + if (limit2 < limit) + limit = limit2; + { + uint32_t lenLimit = p->streamPos - p->pos; + if (lenLimit > p->matchMaxLen) + lenLimit = p->matchMaxLen; + p->lenLimit = lenLimit; + } + p->posLimit = p->pos + limit; +} + + +int Mf_Init(CMatchFinder *p, const int ifd, const int mc, uint32_t historySize, + uint32_t keepAddBufferBefore, uint32_t matchMaxLen, uint32_t keepAddBufferAfter) +{ + const uint32_t sizeReserv = ( historySize >> 1 ) + + (keepAddBufferBefore + matchMaxLen + keepAddBufferAfter) / 2 + (1 << 19); + + p->hash = 0; + p->cutValue = mc; + p->infd = ifd; + p->btMode = true; + p->numHashBytes = 4; + p->crc = 0xFFFFFFFFU; + p->keepSizeBefore = historySize + keepAddBufferBefore + 1; + p->keepSizeAfter = matchMaxLen + keepAddBufferAfter; + /* we need one additional byte, since we use MoveBlock after pos++ and before dictionary using */ + /* keepSizeBefore + keepSizeAfter + sizeReserv must be < 4G) */ + p->blockSize = p->keepSizeBefore + p->keepSizeAfter + sizeReserv; + p->buffer = p->bufferBase = (uint8_t *)malloc(p->blockSize); + if( p->bufferBase ) + { + uint32_t newCyclicBufferSize = historySize + 1; + uint32_t hs; + p->matchMaxLen = matchMaxLen; + { + if (p->numHashBytes == 2) + hs = (1 << 16) - 1; + else + { + hs = historySize - 1; + hs |= (hs >> 1); + hs |= (hs >> 2); + hs |= (hs >> 4); + hs |= (hs >> 8); + hs >>= 1; + hs |= 0xFFFF; /* don't change it! It's required for Deflate */ + if (hs > (1 << 24)) + { + if (p->numHashBytes == 3) + hs = (1 << 24) - 1; + else + hs >>= 1; + } + } + p->hashMask = hs; + hs++; + if (p->numHashBytes > 2) hs += kHash2Size; + if (p->numHashBytes > 3) hs += kHash3Size; + if (p->numHashBytes > 4) hs += kHash4Size; + } + + { + uint32_t newSize; + p->historySize = historySize; + p->hashSizeSum = hs; + p->cyclicBufferSize = newCyclicBufferSize; + p->numSons = (p->btMode ? newCyclicBufferSize * 2 : newCyclicBufferSize); + newSize = p->hashSizeSum + p->numSons; + p->hash = AllocRefs(newSize); + if (p->hash != 0) + { + uint32_t i; + p->son = p->hash + p->hashSizeSum; + for (i = 0; i < p->hashSizeSum; i++) + p->hash[i] = kEmptyHashValue; + p->cyclicBufferPos = 0; + p->pos = p->streamPos = p->cyclicBufferSize; + p->result = SZ_OK; + p->streamEndWasReached = false; + Mf_ReadBlock(p); + Mf_SetLimits(p); + return 1; + } + } + } + Mf_Free(p); + return 0; +} + +static void Mf_Normalize3(uint32_t subValue, CLzRef *items, uint32_t numItems) +{ + uint32_t i; + for (i = 0; i < numItems; i++) + { + uint32_t value = items[i]; + if (value <= subValue) + value = kEmptyHashValue; + else + value -= subValue; + items[i] = value; + } +} + +static void Mf_Normalize(CMatchFinder *p) +{ + uint32_t subValue = (p->pos - p->historySize - 1) & kNormalizeMask; + Mf_Normalize3(subValue, p->hash, p->hashSizeSum + p->numSons); + p->posLimit -= subValue; + p->pos -= subValue; + p->streamPos -= subValue; +} + +static void Mf_CheckLimits(CMatchFinder *p) +{ + if (p->pos == kMaxValForNormalize) + Mf_Normalize(p); + if (!p->streamEndWasReached && p->keepSizeAfter == p->streamPos - p->pos) + Mf_CheckAndMoveAndRead(p); + if (p->cyclicBufferPos == p->cyclicBufferSize) + p->cyclicBufferPos = 0; + Mf_SetLimits(p); +} + +static uint32_t * Hc_GetMatchesSpec(uint32_t lenLimit, uint32_t curMatch, uint32_t pos, const uint8_t *cur, CLzRef *son, + uint32_t _cyclicBufferPos, uint32_t _cyclicBufferSize, uint32_t cutValue, + uint32_t *distances, uint32_t maxLen) +{ + son[_cyclicBufferPos] = curMatch; + for (;;) + { + uint32_t delta = pos - curMatch; + if (cutValue-- == 0 || delta >= _cyclicBufferSize) + return distances; + { + const uint8_t *pb = cur - delta; + curMatch = son[_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)]; + if (pb[maxLen] == cur[maxLen] && *pb == *cur) + { + uint32_t len = 0; + while (++len != lenLimit) + if (pb[len] != cur[len]) + break; + if (maxLen < len) + { + *distances++ = maxLen = len; + *distances++ = delta - 1; + if (len == lenLimit) + return distances; + } + } + } + } +} + + +static uint32_t * GetMatchesSpec1( uint32_t lenLimit, uint32_t curMatch, + uint32_t pos, const uint8_t *cur, CLzRef *son, + uint32_t _cyclicBufferPos, uint32_t _cyclicBufferSize, uint32_t cutValue, + uint32_t *distances, uint32_t maxLen ) +{ + CLzRef *ptr0 = son + (_cyclicBufferPos << 1) + 1; + CLzRef *ptr1 = son + (_cyclicBufferPos << 1); + uint32_t len0 = 0, len1 = 0; + for (;;) + { + uint32_t delta = pos - curMatch; + if (cutValue-- == 0 || delta >= _cyclicBufferSize) + { + *ptr0 = *ptr1 = kEmptyHashValue; + return distances; + } + { + CLzRef *pair = son + ((_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); + const uint8_t *pb = cur - delta; + uint32_t len = (len0 < len1 ? len0 : len1); + if (pb[len] == cur[len]) + { + if (++len != lenLimit && pb[len] == cur[len]) + while (++len != lenLimit) + if (pb[len] != cur[len]) + break; + if (maxLen < len) + { + *distances++ = maxLen = len; + *distances++ = delta - 1; + if (len == lenLimit) + { + *ptr1 = pair[0]; + *ptr0 = pair[1]; + return distances; + } + } + } + if (pb[len] < cur[len]) + { + *ptr1 = curMatch; + ptr1 = pair + 1; + curMatch = *ptr1; + len1 = len; + } + else + { + *ptr0 = curMatch; + ptr0 = pair; + curMatch = *ptr0; + len0 = len; + } + } + } +} + +static void SkipMatchesSpec(uint32_t lenLimit, uint32_t curMatch, uint32_t pos, const uint8_t *cur, CLzRef *son, + uint32_t _cyclicBufferPos, uint32_t _cyclicBufferSize, uint32_t cutValue) +{ + CLzRef *ptr0 = son + (_cyclicBufferPos << 1) + 1; + CLzRef *ptr1 = son + (_cyclicBufferPos << 1); + uint32_t len0 = 0, len1 = 0; + for (;;) + { + uint32_t delta = pos - curMatch; + if (cutValue-- == 0 || delta >= _cyclicBufferSize) + { + *ptr0 = *ptr1 = kEmptyHashValue; + return; + } + { + CLzRef *pair = son + ((_cyclicBufferPos - delta + ((delta > _cyclicBufferPos) ? _cyclicBufferSize : 0)) << 1); + const uint8_t *pb = cur - delta; + uint32_t len = (len0 < len1 ? len0 : len1); + if (pb[len] == cur[len]) + { + while (++len != lenLimit) + if (pb[len] != cur[len]) + break; + { + if (len == lenLimit) + { + *ptr1 = pair[0]; + *ptr0 = pair[1]; + return; + } + } + } + if (pb[len] < cur[len]) + { + *ptr1 = curMatch; + ptr1 = pair + 1; + curMatch = *ptr1; + len1 = len; + } + else + { + *ptr0 = curMatch; + ptr0 = pair; + curMatch = *ptr0; + len0 = len; + } + } + } +} + +#define MOVE_POS \ + ++p->cyclicBufferPos; \ + p->buffer++; \ + if (++p->pos == p->posLimit) Mf_CheckLimits(p); + +#define MOVE_POS_RET MOVE_POS return offset; + +static void Mf_MovePos(CMatchFinder *p) { MOVE_POS; } + +#define GET_MATCHES_HEADER2(minLen, ret_op) \ + uint32_t lenLimit; uint32_t hashValue; const uint8_t *cur; uint32_t curMatch; \ + lenLimit = p->lenLimit; { if (lenLimit < minLen) { Mf_MovePos(p); ret_op; }} \ + cur = p->buffer; + +#define GET_MATCHES_HEADER(minLen) GET_MATCHES_HEADER2(minLen, return 0) +#define SKIP_HEADER(minLen) GET_MATCHES_HEADER2(minLen, continue) + +#define MF_PARAMS(p) p->pos, p->buffer, p->son, p->cyclicBufferPos, p->cyclicBufferSize, p->cutValue + +#define GET_MATCHES_FOOTER(offset, maxLen) \ + offset = (uint32_t)(GetMatchesSpec1(lenLimit, curMatch, MF_PARAMS(p), \ + distances + offset, maxLen) - distances); MOVE_POS_RET; + +#define SKIP_FOOTER \ + SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); MOVE_POS; + +static uint32_t Bt2_MatchFinder_GetMatches(CMatchFinder *p, uint32_t *distances) +{ + uint32_t offset; + GET_MATCHES_HEADER(2) + HASH2_CALC; + curMatch = p->hash[hashValue]; + p->hash[hashValue] = p->pos; + offset = 0; + GET_MATCHES_FOOTER(offset, 1) +} + +static uint32_t Bt3_MatchFinder_GetMatches(CMatchFinder *p, uint32_t *distances) +{ + uint32_t hash2Value, delta2, maxLen, offset; + GET_MATCHES_HEADER(3) + + HASH3_CALC; + + delta2 = p->pos - p->hash[hash2Value]; + curMatch = p->hash[kFix3HashSize + hashValue]; + + p->hash[hash2Value] = + p->hash[kFix3HashSize + hashValue] = p->pos; + + + maxLen = 2; + offset = 0; + if (delta2 < p->cyclicBufferSize && *(cur - delta2) == *cur) + { + for (; maxLen != lenLimit; maxLen++) + if (cur[(ptrdiff_t)maxLen - delta2] != cur[maxLen]) + break; + distances[0] = maxLen; + distances[1] = delta2 - 1; + offset = 2; + if (maxLen == lenLimit) + { + SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); + MOVE_POS_RET; + } + } + GET_MATCHES_FOOTER(offset, maxLen) +} + +static uint32_t Bt4_MatchFinder_GetMatches(CMatchFinder *p, uint32_t *distances) +{ + uint32_t hash2Value, hash3Value, delta2, delta3, maxLen, offset; + GET_MATCHES_HEADER(4) + + HASH4_CALC; + + delta2 = p->pos - p->hash[ hash2Value]; + delta3 = p->pos - p->hash[kFix3HashSize + hash3Value]; + curMatch = p->hash[kFix4HashSize + hashValue]; + + p->hash[ hash2Value] = + p->hash[kFix3HashSize + hash3Value] = + p->hash[kFix4HashSize + hashValue] = p->pos; + + maxLen = 1; + offset = 0; + if (delta2 < p->cyclicBufferSize && *(cur - delta2) == *cur) + { + distances[0] = maxLen = 2; + distances[1] = delta2 - 1; + offset = 2; + } + if (delta2 != delta3 && delta3 < p->cyclicBufferSize && *(cur - delta3) == *cur) + { + maxLen = 3; + distances[offset + 1] = delta3 - 1; + offset += 2; + delta2 = delta3; + } + if (offset != 0) + { + for (; maxLen != lenLimit; maxLen++) + if (cur[(ptrdiff_t)maxLen - delta2] != cur[maxLen]) + break; + distances[offset - 2] = maxLen; + if (maxLen == lenLimit) + { + SkipMatchesSpec(lenLimit, curMatch, MF_PARAMS(p)); + MOVE_POS_RET; + } + } + if (maxLen < 3) + maxLen = 3; + GET_MATCHES_FOOTER(offset, maxLen) +} + +static uint32_t Hc4_MatchFinder_GetMatches(CMatchFinder *p, uint32_t *distances) +{ + uint32_t hash2Value, hash3Value, delta2, delta3, maxLen, offset; + GET_MATCHES_HEADER(4) + + HASH4_CALC; + + delta2 = p->pos - p->hash[ hash2Value]; + delta3 = p->pos - p->hash[kFix3HashSize + hash3Value]; + curMatch = p->hash[kFix4HashSize + hashValue]; + + p->hash[ hash2Value] = + p->hash[kFix3HashSize + hash3Value] = + p->hash[kFix4HashSize + hashValue] = p->pos; + + maxLen = 1; + offset = 0; + if (delta2 < p->cyclicBufferSize && *(cur - delta2) == *cur) + { + distances[0] = maxLen = 2; + distances[1] = delta2 - 1; + offset = 2; + } + if (delta2 != delta3 && delta3 < p->cyclicBufferSize && *(cur - delta3) == *cur) + { + maxLen = 3; + distances[offset + 1] = delta3 - 1; + offset += 2; + delta2 = delta3; + } + if (offset != 0) + { + for (; maxLen != lenLimit; maxLen++) + if (cur[(ptrdiff_t)maxLen - delta2] != cur[maxLen]) + break; + distances[offset - 2] = maxLen; + if (maxLen == lenLimit) + { + p->son[p->cyclicBufferPos] = curMatch; + MOVE_POS_RET; + } + } + if (maxLen < 3) + maxLen = 3; + offset = (uint32_t)(Hc_GetMatchesSpec(lenLimit, curMatch, MF_PARAMS(p), + distances + offset, maxLen) - (distances)); + MOVE_POS_RET +} + + +static void Bt2_MatchFinder_Skip(CMatchFinder *p, uint32_t num) +{ + do + { + SKIP_HEADER(2) + HASH2_CALC; + curMatch = p->hash[hashValue]; + p->hash[hashValue] = p->pos; + SKIP_FOOTER + } + while (--num != 0); +} + + +static void Bt3_MatchFinder_Skip(CMatchFinder *p, uint32_t num) +{ + do + { + uint32_t hash2Value; + SKIP_HEADER(3) + HASH3_CALC; + curMatch = p->hash[kFix3HashSize + hashValue]; + p->hash[hash2Value] = + p->hash[kFix3HashSize + hashValue] = p->pos; + SKIP_FOOTER + } + while (--num != 0); +} + +static void Bt4_MatchFinder_Skip(CMatchFinder *p, uint32_t num) +{ + do + { + uint32_t hash2Value, hash3Value; + SKIP_HEADER(4) + HASH4_CALC; + curMatch = p->hash[kFix4HashSize + hashValue]; + p->hash[ hash2Value] = + p->hash[kFix3HashSize + hash3Value] = p->pos; + p->hash[kFix4HashSize + hashValue] = p->pos; + SKIP_FOOTER + } + while (--num != 0); +} + +static void Hc4_MatchFinder_Skip(CMatchFinder *p, uint32_t num) +{ + do + { + uint32_t hash2Value, hash3Value; + SKIP_HEADER(4) + HASH4_CALC; + curMatch = p->hash[kFix4HashSize + hashValue]; + p->hash[ hash2Value] = + p->hash[kFix3HashSize + hash3Value] = + p->hash[kFix4HashSize + hashValue] = p->pos; + p->son[p->cyclicBufferPos] = curMatch; + MOVE_POS + } + while (--num != 0); +} + + +void Mf_CreateVTable(CMatchFinder *p, IMatchFinder *vTable) +{ + if (!p->btMode) + { + vTable->GetMatches = (Mf_GetMatches_Func)Hc4_MatchFinder_GetMatches; + vTable->Skip = (Mf_Skip_Func)Hc4_MatchFinder_Skip; + } + else if (p->numHashBytes == 2) + { + vTable->GetMatches = (Mf_GetMatches_Func)Bt2_MatchFinder_GetMatches; + vTable->Skip = (Mf_Skip_Func)Bt2_MatchFinder_Skip; + } + else if (p->numHashBytes == 3) + { + vTable->GetMatches = (Mf_GetMatches_Func)Bt3_MatchFinder_GetMatches; + vTable->Skip = (Mf_Skip_Func)Bt3_MatchFinder_Skip; + } + else + { + vTable->GetMatches = (Mf_GetMatches_Func)Bt4_MatchFinder_GetMatches; + vTable->Skip = (Mf_Skip_Func)Bt4_MatchFinder_Skip; + } +} |