void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
ZSTD_compressionParameters const* cParams,
void const* end, ZSTD_dictTableLoadMethod_e dtlm)
{
U32* const hashTable = ms->hashTable;
U32 const hBits = cParams->hashLog;
U32 const mls = cParams->searchLength;
const BYTE* const base = ms->window.base;
const BYTE* ip = base + ms->nextToUpdate;
const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
const U32 fastHashFillStep = 3;
/* Always insert every fastHashFillStep position into the hash table.
* Insert the other positions if their hash entry is empty.
*/
for (; ip + fastHashFillStep - 1 <= iend; ip += fastHashFillStep) {
U32 const current = (U32)(ip - base);
U32 i;
for (i = 0; i < fastHashFillStep; ++i) {
size_t const hash = ZSTD_hashPtr(ip + i, hBits, mls);
if (i == 0 || hashTable[hash] == 0)
hashTable[hash] = current + i;
/* Only load extra positions for ZSTD_dtlm_full */
if (dtlm == ZSTD_dtlm_fast)
break;
}
}
}
void ZSTD_fillDoubleHashTable(ZSTD_CCtx* cctx, const void* end, const U32 mls)
{
U32* const hashLarge = cctx->hashTable;
U32 const hBitsL = cctx->appliedParams.cParams.hashLog;
U32* const hashSmall = cctx->chainTable;
U32 const hBitsS = cctx->appliedParams.cParams.chainLog;
const BYTE* const base = cctx->base;
const BYTE* ip = base + cctx->nextToUpdate;
const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
const size_t fastHashFillStep = 3;
while(ip <= iend) {
hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
hashLarge[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
ip += fastHashFillStep;
}
}
/* Update chains up to ip (excluded)
Assumption : always within prefix (i.e. not within extDict) */
U32 ZSTD_insertAndFindFirstIndex (ZSTD_CCtx* zc, const BYTE* ip, U32 mls)
{
U32* const hashTable = zc->hashTable;
const U32 hashLog = zc->appliedParams.cParams.hashLog;
U32* const chainTable = zc->chainTable;
const U32 chainMask = (1 << zc->appliedParams.cParams.chainLog) - 1;
const BYTE* const base = zc->base;
const U32 target = (U32)(ip - base);
U32 idx = zc->nextToUpdate;
while(idx < target) { /* catch up */
size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);
NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
hashTable[h] = idx;
idx++;
}
zc->nextToUpdate = target;
return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
}
/* Update chains up to ip (excluded)
Assumption : always within prefix (i.e. not within extDict) */
static U32 ZSTD_insertAndFindFirstIndex_internal(
ZSTD_matchState_t* ms,
const ZSTD_compressionParameters* const cParams,
const BYTE* ip, U32 const mls)
{
U32* const hashTable = ms->hashTable;
const U32 hashLog = cParams->hashLog;
U32* const chainTable = ms->chainTable;
const U32 chainMask = (1 << cParams->chainLog) - 1;
const BYTE* const base = ms->window.base;
const U32 target = (U32)(ip - base);
U32 idx = ms->nextToUpdate;
while(idx < target) { /* catch up */
size_t const h = ZSTD_hashPtr(base+idx, hashLog, mls);
NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
hashTable[h] = idx;
idx++;
}
ms->nextToUpdate = target;
return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
}
static void
ZSTD_updateDUBT(ZSTD_matchState_t* ms,
const BYTE* ip, const BYTE* iend,
U32 mls)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable;
U32 const hashLog = cParams->hashLog;
U32* const bt = ms->chainTable;
U32 const btLog = cParams->chainLog - 1;
U32 const btMask = (1 << btLog) - 1;
const BYTE* const base = ms->window.base;
U32 const target = (U32)(ip - base);
U32 idx = ms->nextToUpdate;
if (idx != target)
DEBUGLOG(7, "ZSTD_updateDUBT, from %u to %u (dictLimit:%u)",
idx, target, ms->window.dictLimit);
assert(ip + 8 <= iend); /* condition for ZSTD_hashPtr */
(void)iend;
assert(idx >= ms->window.dictLimit); /* condition for valid base+idx */
for ( ; idx < target ; idx++) {
size_t const h = ZSTD_hashPtr(base + idx, hashLog, mls); /* assumption : ip + 8 <= iend */
U32 const matchIndex = hashTable[h];
U32* const nextCandidatePtr = bt + 2*(idx&btMask);
U32* const sortMarkPtr = nextCandidatePtr + 1;
DEBUGLOG(8, "ZSTD_updateDUBT: insert %u", idx);
hashTable[h] = idx; /* Update Hash Table */
*nextCandidatePtr = matchIndex; /* update BT like a chain */
*sortMarkPtr = ZSTD_DUBT_UNSORTED_MARK;
}
ms->nextToUpdate = target;
}
/* inlining is important to hardwire a hot branch (template emulation) */
FORCE_INLINE_TEMPLATE
size_t ZSTD_HcFindBestMatch_generic (
ZSTD_matchState_t* ms,
const BYTE* const ip, const BYTE* const iLimit,
size_t* offsetPtr,
const U32 mls, const ZSTD_dictMode_e dictMode)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const chainTable = ms->chainTable;
const U32 chainSize = (1 << cParams->chainLog);
const U32 chainMask = chainSize-1;
const BYTE* const base = ms->window.base;
const BYTE* const dictBase = ms->window.dictBase;
const U32 dictLimit = ms->window.dictLimit;
const BYTE* const prefixStart = base + dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit;
const U32 lowLimit = ms->window.lowLimit;
const U32 current = (U32)(ip-base);
const U32 minChain = current > chainSize ? current - chainSize : 0;
U32 nbAttempts = 1U << cParams->searchLog;
size_t ml=4-1;
/* HC4 match finder */
U32 matchIndex = ZSTD_insertAndFindFirstIndex_internal(ms, cParams, ip, mls);
for ( ; (matchIndex>lowLimit) & (nbAttempts>0) ; nbAttempts--) {
size_t currentMl=0;
if ((dictMode != ZSTD_extDict) || matchIndex >= dictLimit) {
const BYTE* const match = base + matchIndex;
assert(matchIndex >= dictLimit); /* ensures this is true if dictMode != ZSTD_extDict */
if (match[ml] == ip[ml]) /* potentially better */
currentMl = ZSTD_count(ip, match, iLimit);
} else {
const BYTE* const match = dictBase + matchIndex;
assert(match+4 <= dictEnd);
if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */
currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dictEnd, prefixStart) + 4;
}
/* save best solution */
if (currentMl > ml) {
ml = currentMl;
*offsetPtr = current - matchIndex + ZSTD_REP_MOVE;
if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
}
if (matchIndex <= minChain) break;
matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
}
if (dictMode == ZSTD_dictMatchState) {
const ZSTD_matchState_t* const dms = ms->dictMatchState;
const U32* const dmsChainTable = dms->chainTable;
const U32 dmsChainSize = (1 << dms->cParams.chainLog);
const U32 dmsChainMask = dmsChainSize - 1;
const U32 dmsLowestIndex = dms->window.dictLimit;
const BYTE* const dmsBase = dms->window.base;
const BYTE* const dmsEnd = dms->window.nextSrc;
const U32 dmsSize = (U32)(dmsEnd - dmsBase);
const U32 dmsIndexDelta = dictLimit - dmsSize;
const U32 dmsMinChain = dmsSize > dmsChainSize ? dmsSize - dmsChainSize : 0;
matchIndex = dms->hashTable[ZSTD_hashPtr(ip, dms->cParams.hashLog, mls)];
for ( ; (matchIndex>dmsLowestIndex) & (nbAttempts>0) ; nbAttempts--) {
size_t currentMl=0;
const BYTE* const match = dmsBase + matchIndex;
assert(match+4 <= dmsEnd);
if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */
currentMl = ZSTD_count_2segments(ip+4, match+4, iLimit, dmsEnd, prefixStart) + 4;
/* save best solution */
if (currentMl > ml) {
ml = currentMl;
*offsetPtr = current - (matchIndex + dmsIndexDelta) + ZSTD_REP_MOVE;
if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
}
if (matchIndex <= dmsMinChain) break;
matchIndex = dmsChainTable[matchIndex & dmsChainMask];
}
}
return ml;
}
static size_t
ZSTD_DUBT_findBestMatch(ZSTD_matchState_t* ms,
const BYTE* const ip, const BYTE* const iend,
size_t* offsetPtr,
U32 const mls,
const ZSTD_dictMode_e dictMode)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable;
U32 const hashLog = cParams->hashLog;
size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
U32 matchIndex = hashTable[h];
const BYTE* const base = ms->window.base;
U32 const current = (U32)(ip-base);
U32 const windowLow = ms->window.lowLimit;
U32* const bt = ms->chainTable;
U32 const btLog = cParams->chainLog - 1;
U32 const btMask = (1 << btLog) - 1;
U32 const btLow = (btMask >= current) ? 0 : current - btMask;
U32 const unsortLimit = MAX(btLow, windowLow);
U32* nextCandidate = bt + 2*(matchIndex&btMask);
U32* unsortedMark = bt + 2*(matchIndex&btMask) + 1;
U32 nbCompares = 1U << cParams->searchLog;
U32 nbCandidates = nbCompares;
U32 previousCandidate = 0;
DEBUGLOG(7, "ZSTD_DUBT_findBestMatch (%u) ", current);
assert(ip <= iend-8); /* required for h calculation */
/* reach end of unsorted candidates list */
while ( (matchIndex > unsortLimit)
&& (*unsortedMark == ZSTD_DUBT_UNSORTED_MARK)
&& (nbCandidates > 1) ) {
DEBUGLOG(8, "ZSTD_DUBT_findBestMatch: candidate %u is unsorted",
matchIndex);
*unsortedMark = previousCandidate; /* the unsortedMark becomes a reversed chain, to move up back to original position */
previousCandidate = matchIndex;
matchIndex = *nextCandidate;
nextCandidate = bt + 2*(matchIndex&btMask);
unsortedMark = bt + 2*(matchIndex&btMask) + 1;
nbCandidates --;
}
/* nullify last candidate if it's still unsorted
* simplification, detrimental to compression ratio, beneficial for speed */
if ( (matchIndex > unsortLimit)
&& (*unsortedMark==ZSTD_DUBT_UNSORTED_MARK) ) {
DEBUGLOG(7, "ZSTD_DUBT_findBestMatch: nullify last unsorted candidate %u",
matchIndex);
*nextCandidate = *unsortedMark = 0;
}
/* batch sort stacked candidates */
matchIndex = previousCandidate;
while (matchIndex) { /* will end on matchIndex == 0 */
U32* const nextCandidateIdxPtr = bt + 2*(matchIndex&btMask) + 1;
U32 const nextCandidateIdx = *nextCandidateIdxPtr;
ZSTD_insertDUBT1(ms, matchIndex, iend,
nbCandidates, unsortLimit, dictMode);
matchIndex = nextCandidateIdx;
nbCandidates++;
}
/* find longest match */
{ size_t commonLengthSmaller = 0, commonLengthLarger = 0;
const BYTE* const dictBase = ms->window.dictBase;
const U32 dictLimit = ms->window.dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const prefixStart = base + dictLimit;
U32* smallerPtr = bt + 2*(current&btMask);
U32* largerPtr = bt + 2*(current&btMask) + 1;
U32 matchEndIdx = current + 8 + 1;
U32 dummy32; /* to be nullified at the end */
size_t bestLength = 0;
matchIndex = hashTable[h];
hashTable[h] = current; /* Update Hash Table */
while (nbCompares-- && (matchIndex > windowLow)) {
U32* const nextPtr = bt + 2*(matchIndex & btMask);
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
const BYTE* match;
if ((dictMode != ZSTD_extDict) || (matchIndex+matchLength >= dictLimit)) {
match = base + matchIndex;
matchLength += ZSTD_count(ip+matchLength, match+matchLength, iend);
} else {
match = dictBase + matchIndex;
matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
if (matchIndex+matchLength >= dictLimit)
match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
}
if (matchLength > bestLength) {
if (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength;
if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )
bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;
if (ip+matchLength == iend) { /* equal : no way to know if inf or sup */
if (dictMode == ZSTD_dictMatchState) {
nbCompares = 0; /* in addition to avoiding checking any
* further in this loop, make sure we
* skip checking in the dictionary. */
}
break; /* drop, to guarantee consistency (miss a little bit of compression) */
}
}
if (match[matchLength] < ip[matchLength]) {
/* match is smaller than current */
*smallerPtr = matchIndex; /* update smaller idx */
commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */
smallerPtr = nextPtr+1; /* new "smaller" => larger of match */
matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */
} else {
/* match is larger than current */
*largerPtr = matchIndex;
commonLengthLarger = matchLength;
if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */
largerPtr = nextPtr;
matchIndex = nextPtr[0];
} }
*smallerPtr = *largerPtr = 0;
if (dictMode == ZSTD_dictMatchState && nbCompares) {
bestLength = ZSTD_DUBT_findBetterDictMatch(
ms, ip, iend,
offsetPtr, bestLength, nbCompares,
mls, dictMode);
}
assert(matchEndIdx > current+8); /* ensure nextToUpdate is increased */
ms->nextToUpdate = matchEndIdx - 8; /* skip repetitive patterns */
if (bestLength >= MINMATCH) {
U32 const mIndex = current - ((U32)*offsetPtr - ZSTD_REP_MOVE); (void)mIndex;
DEBUGLOG(8, "ZSTD_DUBT_findBestMatch(%u) : found match of length %u and offsetCode %u (pos %u)",
current, (U32)bestLength, (U32)*offsetPtr, mIndex);
}
return bestLength;
}
}
static size_t
ZSTD_DUBT_findBetterDictMatch (
ZSTD_matchState_t* ms,
const BYTE* const ip, const BYTE* const iend,
size_t* offsetPtr,
size_t bestLength,
U32 nbCompares,
U32 const mls,
const ZSTD_dictMode_e dictMode)
{
const ZSTD_matchState_t * const dms = ms->dictMatchState;
const ZSTD_compressionParameters* const dmsCParams = &dms->cParams;
const U32 * const dictHashTable = dms->hashTable;
U32 const hashLog = dmsCParams->hashLog;
size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
U32 dictMatchIndex = dictHashTable[h];
const BYTE* const base = ms->window.base;
const BYTE* const prefixStart = base + ms->window.dictLimit;
U32 const current = (U32)(ip-base);
const BYTE* const dictBase = dms->window.base;
const BYTE* const dictEnd = dms->window.nextSrc;
U32 const dictHighLimit = (U32)(dms->window.nextSrc - dms->window.base);
U32 const dictLowLimit = dms->window.lowLimit;
U32 const dictIndexDelta = ms->window.lowLimit - dictHighLimit;
U32* const dictBt = dms->chainTable;
U32 const btLog = dmsCParams->chainLog - 1;
U32 const btMask = (1 << btLog) - 1;
U32 const btLow = (btMask >= dictHighLimit - dictLowLimit) ? dictLowLimit : dictHighLimit - btMask;
size_t commonLengthSmaller=0, commonLengthLarger=0;
(void)dictMode;
assert(dictMode == ZSTD_dictMatchState);
while (nbCompares-- && (dictMatchIndex > dictLowLimit)) {
U32* const nextPtr = dictBt + 2*(dictMatchIndex & btMask);
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
const BYTE* match = dictBase + dictMatchIndex;
matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
if (dictMatchIndex+matchLength >= dictHighLimit)
match = base + dictMatchIndex + dictIndexDelta; /* to prepare for next usage of match[matchLength] */
if (matchLength > bestLength) {
U32 matchIndex = dictMatchIndex + dictIndexDelta;
if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) ) {
DEBUGLOG(9, "ZSTD_DUBT_findBetterDictMatch(%u) : found better match length %u -> %u and offsetCode %u -> %u (dictMatchIndex %u, matchIndex %u)",
current, (U32)bestLength, (U32)matchLength, (U32)*offsetPtr, ZSTD_REP_MOVE + current - matchIndex, dictMatchIndex, matchIndex);
bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;
}
if (ip+matchLength == iend) { /* reached end of input : ip[matchLength] is not valid, no way to know if it's larger or smaller than match */
break; /* drop, to guarantee consistency (miss a little bit of compression) */
}
}
if (match[matchLength] < ip[matchLength]) {
if (dictMatchIndex <= btLow) { break; } /* beyond tree size, stop the search */
commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
dictMatchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */
} else {
/* match is larger than current */
if (dictMatchIndex <= btLow) { break; } /* beyond tree size, stop the search */
commonLengthLarger = matchLength;
dictMatchIndex = nextPtr[0];
}
}
if (bestLength >= MINMATCH) {
U32 const mIndex = current - ((U32)*offsetPtr - ZSTD_REP_MOVE); (void)mIndex;
DEBUGLOG(8, "ZSTD_DUBT_findBetterDictMatch(%u) : found match of length %u and offsetCode %u (pos %u)",
current, (U32)bestLength, (U32)*offsetPtr, mIndex);
}
return bestLength;
}
/** ZSTD_insertBt1() : add one or multiple positions to tree.
* ip : assumed <= iend-8 .
* @return : nb of positions added */
static U32 ZSTD_insertBt1(ZSTD_CCtx* zc, const BYTE* const ip, const U32 mls, const BYTE* const iend, U32 nbCompares,
U32 extDict)
{
U32* const hashTable = zc->hashTable;
U32 const hashLog = zc->appliedParams.cParams.hashLog;
size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
U32* const bt = zc->chainTable;
U32 const btLog = zc->appliedParams.cParams.chainLog - 1;
U32 const btMask = (1 << btLog) - 1;
U32 matchIndex = hashTable[h];
size_t commonLengthSmaller=0, commonLengthLarger=0;
const BYTE* const base = zc->base;
const BYTE* const dictBase = zc->dictBase;
const U32 dictLimit = zc->dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const prefixStart = base + dictLimit;
const BYTE* match;
const U32 current = (U32)(ip-base);
const U32 btLow = btMask >= current ? 0 : current - btMask;
U32* smallerPtr = bt + 2*(current&btMask);
U32* largerPtr = smallerPtr + 1;
U32 dummy32; /* to be nullified at the end */
U32 const windowLow = zc->lowLimit;
U32 matchEndIdx = current+8;
size_t bestLength = 8;
#ifdef ZSTD_C_PREDICT
U32 predictedSmall = *(bt + 2*((current-1)&btMask) + 0);
U32 predictedLarge = *(bt + 2*((current-1)&btMask) + 1);
predictedSmall += (predictedSmall>0);
predictedLarge += (predictedLarge>0);
#endif /* ZSTD_C_PREDICT */
assert(ip <= iend-8); /* required for h calculation */
hashTable[h] = current; /* Update Hash Table */
while (nbCompares-- && (matchIndex > windowLow)) {
U32* const nextPtr = bt + 2*(matchIndex & btMask);
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
#ifdef ZSTD_C_PREDICT /* note : can create issues when hlog small <= 11 */
const U32* predictPtr = bt + 2*((matchIndex-1) & btMask); /* written this way, as bt is a roll buffer */
if (matchIndex == predictedSmall) {
/* no need to check length, result known */
*smallerPtr = matchIndex;
if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */
smallerPtr = nextPtr+1; /* new "smaller" => larger of match */
matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */
predictedSmall = predictPtr[1] + (predictPtr[1]>0);
continue;
}
if (matchIndex == predictedLarge) {
*largerPtr = matchIndex;
if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */
largerPtr = nextPtr;
matchIndex = nextPtr[0];
predictedLarge = predictPtr[0] + (predictPtr[0]>0);
continue;
}
#endif
if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
match = base + matchIndex;
if (match[matchLength] == ip[matchLength])
matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
} else {
match = dictBase + matchIndex;
matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
if (matchIndex+matchLength >= dictLimit)
match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
}
if (matchLength > bestLength) {
bestLength = matchLength;
if (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength;
}
if (ip+matchLength == iend) /* equal : no way to know if inf or sup */
break; /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt tree */
if (match[matchLength] < ip[matchLength]) { /* necessarily within buffer */
/* match+1 is smaller than current */
*smallerPtr = matchIndex; /* update smaller idx */
commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop searching */
smallerPtr = nextPtr+1; /* new "smaller" => larger of match */
matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */
} else {
/* match is larger than current */
*largerPtr = matchIndex;
commonLengthLarger = matchLength;
if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop searching */
largerPtr = nextPtr;
matchIndex = nextPtr[0];
} }
*smallerPtr = *largerPtr = 0;
if (bestLength > 384) return MIN(192, (U32)(bestLength - 384)); /* speed optimization */
if (matchEndIdx > current + 8) return matchEndIdx - (current + 8);
return 1;
}
static size_t ZSTD_insertBtAndFindBestMatch (
ZSTD_CCtx* zc,
const BYTE* const ip, const BYTE* const iend,
size_t* offsetPtr,
U32 nbCompares, const U32 mls,
U32 extDict)
{
U32* const hashTable = zc->hashTable;
U32 const hashLog = zc->appliedParams.cParams.hashLog;
size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
U32* const bt = zc->chainTable;
U32 const btLog = zc->appliedParams.cParams.chainLog - 1;
U32 const btMask = (1 << btLog) - 1;
U32 matchIndex = hashTable[h];
size_t commonLengthSmaller=0, commonLengthLarger=0;
const BYTE* const base = zc->base;
const BYTE* const dictBase = zc->dictBase;
const U32 dictLimit = zc->dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const prefixStart = base + dictLimit;
const U32 current = (U32)(ip-base);
const U32 btLow = btMask >= current ? 0 : current - btMask;
const U32 windowLow = zc->lowLimit;
U32* smallerPtr = bt + 2*(current&btMask);
U32* largerPtr = bt + 2*(current&btMask) + 1;
U32 matchEndIdx = current+8;
U32 dummy32; /* to be nullified at the end */
size_t bestLength = 0;
assert(ip <= iend-8); /* required for h calculation */
hashTable[h] = current; /* Update Hash Table */
while (nbCompares-- && (matchIndex > windowLow)) {
U32* const nextPtr = bt + 2*(matchIndex & btMask);
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
const BYTE* match;
if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
match = base + matchIndex;
if (match[matchLength] == ip[matchLength])
matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
} else {
match = dictBase + matchIndex;
matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
if (matchIndex+matchLength >= dictLimit)
match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
}
if (matchLength > bestLength) {
if (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength;
if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(current-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )
bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + current - matchIndex;
if (ip+matchLength == iend) /* equal : no way to know if inf or sup */
break; /* drop, to guarantee consistency (miss a little bit of compression) */
}
if (match[matchLength] < ip[matchLength]) {
/* match is smaller than current */
*smallerPtr = matchIndex; /* update smaller idx */
commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */
smallerPtr = nextPtr+1; /* new "smaller" => larger of match */
matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */
} else {
/* match is larger than current */
*largerPtr = matchIndex;
commonLengthLarger = matchLength;
if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */
largerPtr = nextPtr;
matchIndex = nextPtr[0];
} }
*smallerPtr = *largerPtr = 0;
zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;
return bestLength;
}
FORCE_INLINE_TEMPLATE
size_t ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx* cctx,
const void* src, size_t srcSize,
const U32 mls)
{
U32* const hashLong = cctx->hashTable;
const U32 hBitsL = cctx->appliedParams.cParams.hashLog;
U32* const hashSmall = cctx->chainTable;
const U32 hBitsS = cctx->appliedParams.cParams.chainLog;
seqStore_t* seqStorePtr = &(cctx->seqStore);
const BYTE* const base = cctx->base;
const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart;
const BYTE* anchor = istart;
const U32 lowestIndex = cctx->dictLimit;
const BYTE* const lowest = base + lowestIndex;
const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - HASH_READ_SIZE;
U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
U32 offsetSaved = 0;
/* init */
ip += (ip==lowest);
{ U32 const maxRep = (U32)(ip-lowest);
if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
}
/* Main Search Loop */
while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
size_t mLength;
size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);
size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);
U32 const current = (U32)(ip-base);
U32 const matchIndexL = hashLong[h2];
U32 const matchIndexS = hashSmall[h];
const BYTE* matchLong = base + matchIndexL;
const BYTE* match = base + matchIndexS;
hashLong[h2] = hashSmall[h] = current; /* update hash tables */
assert(offset_1 <= current); /* supposed guaranteed by construction */
if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {
/* favor repcode */
mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
ip++;
ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
} else {
U32 offset;
if ( (matchIndexL > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip)) ) {
mLength = ZSTD_count(ip+8, matchLong+8, iend) + 8;
offset = (U32)(ip-matchLong);
while (((ip>anchor) & (matchLong>lowest)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
} else if ( (matchIndexS > lowestIndex) && (MEM_read32(match) == MEM_read32(ip)) ) {
size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
U32 const matchIndexL3 = hashLong[hl3];
const BYTE* matchL3 = base + matchIndexL3;
hashLong[hl3] = current + 1;
if ( (matchIndexL3 > lowestIndex) && (MEM_read64(matchL3) == MEM_read64(ip+1)) ) {
mLength = ZSTD_count(ip+9, matchL3+8, iend) + 8;
ip++;
offset = (U32)(ip-matchL3);
while (((ip>anchor) & (matchL3>lowest)) && (ip[-1] == matchL3[-1])) { ip--; matchL3--; mLength++; } /* catch up */
} else {
mLength = ZSTD_count(ip+4, match+4, iend) + 4;
offset = (U32)(ip-match);
while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
}
} else {
ip += ((ip-anchor) >> g_searchStrength) + 1;
continue;
}
offset_2 = offset_1;
offset_1 = offset;
ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
}
/* match found */
ip += mLength;
anchor = ip;
if (ip <= ilimit) {
/* Fill Table */
hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] =
hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2; /* here because current+2 could be > iend-8 */
hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] =
hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
/* check immediate repcode */
while ( (ip <= ilimit)
&& ( (offset_2>0)
& (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
/* store sequence */
size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
{ U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */
hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip-base);
hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip-base);
ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength-MINMATCH);
ip += rLength;
anchor = ip;
continue; /* faster when present ... (?) */
} } }
/* save reps for next block */
seqStorePtr->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
seqStorePtr->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
/* Return the last literals size */
return iend - anchor;
}
static size_t ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx* ctx,
const void* src, size_t srcSize,
const U32 mls)
{
U32* const hashLong = ctx->hashTable;
U32 const hBitsL = ctx->appliedParams.cParams.hashLog;
U32* const hashSmall = ctx->chainTable;
U32 const hBitsS = ctx->appliedParams.cParams.chainLog;
seqStore_t* seqStorePtr = &(ctx->seqStore);
const BYTE* const base = ctx->base;
const BYTE* const dictBase = ctx->dictBase;
const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart;
const BYTE* anchor = istart;
const U32 lowestIndex = ctx->lowLimit;
const BYTE* const dictStart = dictBase + lowestIndex;
const U32 dictLimit = ctx->dictLimit;
const BYTE* const lowPrefixPtr = base + dictLimit;
const BYTE* const dictEnd = dictBase + dictLimit;
const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - 8;
U32 offset_1=seqStorePtr->rep[0], offset_2=seqStorePtr->rep[1];
/* Search Loop */
while (ip < ilimit) { /* < instead of <=, because (ip+1) */
const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
const U32 matchIndex = hashSmall[hSmall];
const BYTE* matchBase = matchIndex < dictLimit ? dictBase : base;
const BYTE* match = matchBase + matchIndex;
const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
const U32 matchLongIndex = hashLong[hLong];
const BYTE* matchLongBase = matchLongIndex < dictLimit ? dictBase : base;
const BYTE* matchLong = matchLongBase + matchLongIndex;
const U32 current = (U32)(ip-base);
const U32 repIndex = current + 1 - offset_1; /* offset_1 expected <= current +1 */
const BYTE* repBase = repIndex < dictLimit ? dictBase : base;
const BYTE* repMatch = repBase + repIndex;
size_t mLength;
hashSmall[hSmall] = hashLong[hLong] = current; /* update hash table */
if ( (((U32)((dictLimit-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex))
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
const BYTE* repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, lowPrefixPtr) + 4;
ip++;
ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, 0, mLength-MINMATCH);
} else {
if ((matchLongIndex > lowestIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
const BYTE* matchEnd = matchLongIndex < dictLimit ? dictEnd : iend;
const BYTE* lowMatchPtr = matchLongIndex < dictLimit ? dictStart : lowPrefixPtr;
U32 offset;
mLength = ZSTD_count_2segments(ip+8, matchLong+8, iend, matchEnd, lowPrefixPtr) + 8;
offset = current - matchLongIndex;
while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
offset_2 = offset_1;
offset_1 = offset;
ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} else if ((matchIndex > lowestIndex) && (MEM_read32(match) == MEM_read32(ip))) {
size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
U32 const matchIndex3 = hashLong[h3];
const BYTE* const match3Base = matchIndex3 < dictLimit ? dictBase : base;
const BYTE* match3 = match3Base + matchIndex3;
U32 offset;
hashLong[h3] = current + 1;
if ( (matchIndex3 > lowestIndex) && (MEM_read64(match3) == MEM_read64(ip+1)) ) {
const BYTE* matchEnd = matchIndex3 < dictLimit ? dictEnd : iend;
const BYTE* lowMatchPtr = matchIndex3 < dictLimit ? dictStart : lowPrefixPtr;
mLength = ZSTD_count_2segments(ip+9, match3+8, iend, matchEnd, lowPrefixPtr) + 8;
ip++;
offset = current+1 - matchIndex3;
while (((ip>anchor) & (match3>lowMatchPtr)) && (ip[-1] == match3[-1])) { ip--; match3--; mLength++; } /* catch up */
} else {
const BYTE* matchEnd = matchIndex < dictLimit ? dictEnd : iend;
const BYTE* lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, lowPrefixPtr) + 4;
offset = current - matchIndex;
while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
}
offset_2 = offset_1;
offset_1 = offset;
ZSTD_storeSeq(seqStorePtr, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} else {
ip += ((ip-anchor) >> g_searchStrength) + 1;
continue;
} }
/* found a match : store it */
ip += mLength;
anchor = ip;
if (ip <= ilimit) {
/* Fill Table */
hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;
hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2;
hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
/* check immediate repcode */
while (ip <= ilimit) {
U32 const current2 = (U32)(ip-base);
U32 const repIndex2 = current2 - offset_2;
const BYTE* repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
if ( (((U32)((dictLimit-1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */
&& (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
const BYTE* const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, lowPrefixPtr) + 4;
U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2-MINMATCH);
hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
ip += repLength2;
anchor = ip;
continue;
}
break;
} } }
/* save reps for next block */
seqStorePtr->repToConfirm[0] = offset_1; seqStorePtr->repToConfirm[1] = offset_2;
/* Return the last literals size */
return iend - anchor;
}
FORCE_INLINE_TEMPLATE
size_t ZSTD_compressBlock_fast_generic(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize,
U32 const hlog, U32 const stepSize, U32 const mls)
{
U32* const hashTable = ms->hashTable;
const BYTE* const base = ms->window.base;
const BYTE* const istart = (const BYTE*)src;
const BYTE* ip = istart;
const BYTE* anchor = istart;
const U32 lowestIndex = ms->window.dictLimit;
const BYTE* const lowest = base + lowestIndex;
const BYTE* const iend = istart + srcSize;
const BYTE* const ilimit = iend - HASH_READ_SIZE;
U32 offset_1=rep[0], offset_2=rep[1];
U32 offsetSaved = 0;
/* init */
ip += (ip==lowest);
{ U32 const maxRep = (U32)(ip-lowest);
if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
}
/* Main Search Loop */
while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
size_t mLength;
size_t const h = ZSTD_hashPtr(ip, hlog, mls);
U32 const current = (U32)(ip-base);
U32 const matchIndex = hashTable[h];
const BYTE* match = base + matchIndex;
hashTable[h] = current; /* update hash table */
if ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1))) {
mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
ip++;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
} else {
if ( (matchIndex <= lowestIndex)
|| (MEM_read32(match) != MEM_read32(ip)) ) {
assert(stepSize >= 1);
ip += ((ip-anchor) >> kSearchStrength) + stepSize;
continue;
}
mLength = ZSTD_count(ip+4, match+4, iend) + 4;
{ U32 const offset = (U32)(ip-match);
while (((ip>anchor) & (match>lowest)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
offset_2 = offset_1;
offset_1 = offset;
ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
} }
/* match found */
ip += mLength;
anchor = ip;
if (ip <= ilimit) {
/* Fill Table */
hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2; /* here because current+2 could be > iend-8 */
hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
/* check immediate repcode */
while ( (ip <= ilimit)
&& ( (offset_2>0)
& (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
/* store sequence */
size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
{ U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; } /* swap offset_2 <=> offset_1 */
hashTable[ZSTD_hashPtr(ip, hlog, mls)] = (U32)(ip-base);
ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
ip += rLength;
anchor = ip;
continue; /* faster when present ... (?) */
} } }
