/**
* ZSTD_seqStore_resolveOffCodes() reconciles any possible divergences in offset history that may arise
- * due to emission of RLE/raw blocks that disturb the offset history, and replaces any repcodes within
- * the seqStore that may be invalid.
+ * due to emission of RLE/raw blocks that disturb the offset history,
+ * and replaces any repcodes within the seqStore that may be invalid.
*
- * dRepcodes are updated as would be on the decompression side. cRepcodes are updated exactly in
- * accordance with the seqStore.
+ * dRepcodes are updated as would be on the decompression side.
+ * cRepcodes are updated exactly in accordance with the seqStore.
+ *
+ * Note : this function assumes seq->offBase respects the following numbering scheme :
+ * 0 : invalid
+ * 1-3 : repcode 1-3
+ * 4+ : real_offset+3
*/
static void ZSTD_seqStore_resolveOffCodes(repcodes_t* const dRepcodes, repcodes_t* const cRepcodes,
seqStore_t* const seqStore, U32 const nbSeq) {
for (; idx < nbSeq; ++idx) {
seqDef* const seq = seqStore->sequencesStart + idx;
U32 const ll0 = (seq->litLength == 0);
- U32 offCode = seq->offBase - 1;
+ U32 const offCode = seq->offBase - 1;
assert(seq->offBase > 0);
- if (offCode <= ZSTD_REP_MOVE) {
+ if (offCode < ZSTD_REP_NUM) {
U32 const dRawOffset = ZSTD_resolveRepcodeToRawOffset(dRepcodes->rep, offCode, ll0);
U32 const cRawOffset = ZSTD_resolveRepcodeToRawOffset(cRepcodes->rep, offCode, ll0);
/* Adjust simulated decompression repcode history if we come across a mismatch. Replace
size_t posInSrc; /* Number of bytes given by sequences provided so far */
} ZSTD_sequencePosition;
-/* Returns a ZSTD error code if sequence is not valid */
-static size_t ZSTD_validateSequence(U32 offCode, U32 matchLength,
- size_t posInSrc, U32 windowLog, size_t dictSize, U32 minMatch) {
- size_t offsetBound;
- U32 windowSize = 1 << windowLog;
+/* ZSTD_validateSequence() :
+ * @offCode : is presumed to follow format required by ZSTD_storeSeq()
+ * @returns a ZSTD error code if sequence is not valid
+ */
+static size_t
+ZSTD_validateSequence(U32 offCode, U32 matchLength,
+ size_t posInSrc, U32 windowLog, size_t dictSize)
+{
+ U32 const windowSize = 1 << windowLog;
/* posInSrc represents the amount of data the the decoder would decode up to this point.
* As long as the amount of data decoded is less than or equal to window size, offsets may be
* larger than the total length of output decoded in order to reference the dict, even larger than
* window size. After output surpasses windowSize, we're limited to windowSize offsets again.
*/
- offsetBound = posInSrc > windowSize ? (size_t)windowSize : posInSrc + (size_t)dictSize;
- RETURN_ERROR_IF(offCode > offsetBound + ZSTD_REP_MOVE, corruption_detected, "Offset too large!");
- RETURN_ERROR_IF(matchLength < minMatch, corruption_detected, "Matchlength too small");
+ size_t const offsetBound = posInSrc > windowSize ? (size_t)windowSize : posInSrc + (size_t)dictSize;
+ RETURN_ERROR_IF(offCode > STORE_OFFSET(offsetBound), corruption_detected, "Offset too large!");
+ RETURN_ERROR_IF(matchLength < MINMATCH, corruption_detected, "Matchlength too small");
return 0;
}
/* Returns an offset code, given a sequence's raw offset, the ongoing repcode array, and whether litLength == 0 */
-static U32 ZSTD_finalizeOffCode(U32 rawOffset, const U32 rep[ZSTD_REP_NUM], U32 ll0) {
- U32 offCode = rawOffset + ZSTD_REP_MOVE;
- U32 repCode = 0;
+static U32 ZSTD_finalizeOffCode(U32 rawOffset, const U32 rep[ZSTD_REP_NUM], U32 ll0)
+{
+ U32 offCode = STORE_OFFSET(rawOffset);
if (!ll0 && rawOffset == rep[0]) {
- repCode = 1;
+ offCode = STORE_REPCODE_1;
} else if (rawOffset == rep[1]) {
- repCode = 2 - ll0;
+ offCode = STORE_REPCODE(2 - ll0);
} else if (rawOffset == rep[2]) {
- repCode = 3 - ll0;
+ offCode = STORE_REPCODE(3 - ll0);
} else if (ll0 && rawOffset == rep[0] - 1) {
- repCode = 3;
- }
- if (repCode) {
- /* ZSTD_storeSeq expects a number in the range [0, 2] to represent a repcode */
- offCode = repCode - 1;
+ offCode = STORE_REPCODE_3;
}
return offCode;
}
/* Returns 0 on success, and a ZSTD_error otherwise. This function scans through an array of
* ZSTD_Sequence, storing the sequences it finds, until it reaches a block delimiter.
*/
-static size_t ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
- const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
- const void* src, size_t blockSize) {
+static size_t
+ZSTD_copySequencesToSeqStoreExplicitBlockDelim(ZSTD_CCtx* cctx,
+ ZSTD_sequencePosition* seqPos,
+ const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
+ const void* src, size_t blockSize)
+{
U32 idx = seqPos->idx;
BYTE const* ip = (BYTE const*)(src);
const BYTE* const iend = ip + blockSize;
repcodes_t updatedRepcodes;
U32 dictSize;
- U32 litLength;
- U32 matchLength;
- U32 ll0;
- U32 offCode;
if (cctx->cdict) {
dictSize = (U32)cctx->cdict->dictContentSize;
}
ZSTD_memcpy(updatedRepcodes.rep, cctx->blockState.prevCBlock->rep, sizeof(repcodes_t));
for (; (inSeqs[idx].matchLength != 0 || inSeqs[idx].offset != 0) && idx < inSeqsSize; ++idx) {
- litLength = inSeqs[idx].litLength;
- matchLength = inSeqs[idx].matchLength;
- ll0 = litLength == 0;
- offCode = ZSTD_finalizeOffCode(inSeqs[idx].offset, updatedRepcodes.rep, ll0);
+ U32 const litLength = inSeqs[idx].litLength;
+ U32 const ll0 = (litLength == 0);
+ U32 const matchLength = inSeqs[idx].matchLength;
+ U32 const offCode = ZSTD_finalizeOffCode(inSeqs[idx].offset, updatedRepcodes.rep, ll0);
updatedRepcodes = ZSTD_updateRep(updatedRepcodes.rep, offCode, ll0);
DEBUGLOG(6, "Storing sequence: (of: %u, ml: %u, ll: %u)", offCode, matchLength, litLength);
if (cctx->appliedParams.validateSequences) {
seqPos->posInSrc += litLength + matchLength;
FORWARD_IF_ERROR(ZSTD_validateSequence(offCode, matchLength, seqPos->posInSrc,
- cctx->appliedParams.cParams.windowLog, dictSize,
- cctx->appliedParams.cParams.minMatch),
+ cctx->appliedParams.cParams.windowLog, dictSize),
"Sequence validation failed");
}
RETURN_ERROR_IF(idx - seqPos->idx > cctx->seqStore.maxNbSeq, memory_allocation,
* avoid splitting a match, or to avoid splitting a match such that it would produce a match
* smaller than MINMATCH. In this case, we return the number of bytes that we didn't read from this block.
*/
-static size_t ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
- const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
- const void* src, size_t blockSize) {
+static size_t
+ZSTD_copySequencesToSeqStoreNoBlockDelim(ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
+ const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
+ const void* src, size_t blockSize)
+{
U32 idx = seqPos->idx;
U32 startPosInSequence = seqPos->posInSequence;
U32 endPosInSequence = seqPos->posInSequence + (U32)blockSize;
repcodes_t updatedRepcodes;
U32 bytesAdjustment = 0;
U32 finalMatchSplit = 0;
- U32 litLength;
- U32 matchLength;
- U32 rawOffset;
- U32 offCode;
if (cctx->cdict) {
dictSize = cctx->cdict->dictContentSize;
ZSTD_memcpy(updatedRepcodes.rep, cctx->blockState.prevCBlock->rep, sizeof(repcodes_t));
while (endPosInSequence && idx < inSeqsSize && !finalMatchSplit) {
const ZSTD_Sequence currSeq = inSeqs[idx];
- litLength = currSeq.litLength;
- matchLength = currSeq.matchLength;
- rawOffset = currSeq.offset;
+ U32 litLength = currSeq.litLength;
+ U32 matchLength = currSeq.matchLength;
+ U32 const rawOffset = currSeq.offset;
+ U32 offCode;
/* Modify the sequence depending on where endPosInSequence lies */
if (endPosInSequence >= currSeq.litLength + currSeq.matchLength) {
}
}
/* Check if this offset can be represented with a repcode */
- { U32 ll0 = (litLength == 0);
+ { U32 const ll0 = (litLength == 0);
offCode = ZSTD_finalizeOffCode(rawOffset, updatedRepcodes.rep, ll0);
updatedRepcodes = ZSTD_updateRep(updatedRepcodes.rep, offCode, ll0);
}
if (cctx->appliedParams.validateSequences) {
seqPos->posInSrc += litLength + matchLength;
FORWARD_IF_ERROR(ZSTD_validateSequence(offCode, matchLength, seqPos->posInSrc,
- cctx->appliedParams.cParams.windowLog, dictSize,
- cctx->appliedParams.cParams.minMatch),
+ cctx->appliedParams.cParams.windowLog, dictSize),
"Sequence validation failed");
}
DEBUGLOG(6, "Storing sequence: (of: %u, ml: %u, ll: %u)", offCode, matchLength, litLength);
typedef size_t (*ZSTD_sequenceCopier) (ZSTD_CCtx* cctx, ZSTD_sequencePosition* seqPos,
const ZSTD_Sequence* const inSeqs, size_t inSeqsSize,
const void* src, size_t blockSize);
-static ZSTD_sequenceCopier ZSTD_selectSequenceCopier(ZSTD_sequenceFormat_e mode) {
+static ZSTD_sequenceCopier ZSTD_selectSequenceCopier(ZSTD_sequenceFormat_e mode)
+{
ZSTD_sequenceCopier sequenceCopier = NULL;
assert(ZSTD_cParam_withinBounds(ZSTD_c_blockDelimiters, mode));
if (mode == ZSTD_sf_explicitBlockDelimiters) {
/* Compress, block-by-block, all of the sequences given.
*
- * Returns the cumulative size of all compressed blocks (including their headers), otherwise a ZSTD error.
+ * Returns the cumulative size of all compressed blocks (including their headers),
+ * otherwise a ZSTD error.
*/
-static size_t ZSTD_compressSequences_internal(ZSTD_CCtx* cctx,
- void* dst, size_t dstCapacity,
- const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
- const void* src, size_t srcSize) {
+static size_t
+ZSTD_compressSequences_internal(ZSTD_CCtx* cctx,
+ void* dst, size_t dstCapacity,
+ const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
+ const void* src, size_t srcSize)
+{
size_t cSize = 0;
U32 lastBlock;
size_t blockSize;
BYTE const* ip = (BYTE const*)src;
BYTE* op = (BYTE*)dst;
- ZSTD_sequenceCopier sequenceCopier = ZSTD_selectSequenceCopier(cctx->appliedParams.blockDelimiters);
+ ZSTD_sequenceCopier const sequenceCopier = ZSTD_selectSequenceCopier(cctx->appliedParams.blockDelimiters);
DEBUGLOG(4, "ZSTD_compressSequences_internal srcSize: %zu, inSeqsSize: %zu", srcSize, inSeqsSize);
/* Special case: empty frame */
size_t ZSTD_compressSequences(ZSTD_CCtx* const cctx, void* dst, size_t dstCapacity,
const ZSTD_Sequence* inSeqs, size_t inSeqsSize,
- const void* src, size_t srcSize) {
+ const void* src, size_t srcSize)
+{
BYTE* op = (BYTE*)dst;
size_t cSize = 0;
size_t compressedBlocksSize = 0;
U32 rep[3];
} repcodes_t;
-MEM_STATIC repcodes_t ZSTD_updateRep(U32 const rep[3], U32 const offset, U32 const ll0)
+
+/* ZSTD_updateRep() :
+ * @offcode : expects a scale where 0,1,2 represent repcodes 1-3, and 2+ represents real_offset+2
+ */
+MEM_STATIC repcodes_t
+ZSTD_updateRep(U32 const rep[3], U32 const offcode, U32 const ll0)
{
repcodes_t newReps;
- if (offset >= ZSTD_REP_NUM) { /* full offset */
+ if (offcode >= ZSTD_REP_NUM) { /* full offset */
newReps.rep[2] = rep[1];
newReps.rep[1] = rep[0];
- newReps.rep[0] = offset - ZSTD_REP_MOVE;
+ newReps.rep[0] = offcode - ZSTD_REP_MOVE;
} else { /* repcode */
- U32 const repCode = offset + ll0;
+ U32 const repCode = offcode + ll0;
if (repCode > 0) { /* note : if repCode==0, no change */
U32 const currentOffset = (repCode==ZSTD_REP_NUM) ? (rep[0] - 1) : rep[repCode];
newReps.rep[2] = (repCode >= 2) ? rep[1] : rep[2];
while (ip < iend) *op++ = *ip++;
}
+#define STORE_REPCODE_1 STORE_REPCODE(1)
+#define STORE_REPCODE_2 STORE_REPCODE(2)
+#define STORE_REPCODE_3 STORE_REPCODE(3)
+#define STORE_REPCODE(r) (assert((r)>=1), assert((r)<=3), (r)-1)
+#define STORE_OFFSET(o) (assert((o)>0), o + ZSTD_REP_MOVE)
+
/*! ZSTD_storeSeq() :
* Store a sequence (litlen, litPtr, offCode and matchLength) into seqStore_t.
- * `offCode` : distance to match + ZSTD_REP_MOVE (values <= ZSTD_REP_MOVE are repCodes).
+ * @offBase_minus1 : distance to match + ZSTD_REP_MOVE (values <= ZSTD_REP_MOVE are repCodes).
+ * Users should not specify the encoded value directly,
+ * instead use macros STORE_REPCODE_X and STORE_OFFSET().
* @matchLength : must be >= MINMATCH
* Allowed to overread literals up to litLimit.
*/
HINT_INLINE UNUSED_ATTR
-void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offCode, size_t matchLength)
+void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, const BYTE* litLimit, U32 offBase_minus1, size_t matchLength)
{
BYTE const* const litLimit_w = litLimit - WILDCOPY_OVERLENGTH;
BYTE const* const litEnd = literals + litLength;
seqStorePtr->sequences[0].litLength = (U16)litLength;
/* match offset */
- seqStorePtr->sequences[0].offBase = offCode + 1;
+ seqStorePtr->sequences[0].offBase = offBase_minus1 + 1;
/* match Length */
assert(matchLength >= MINMATCH);
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, (size_t)(ip-anchor), anchor, iend, 0, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_REPCODE_1, mLength);
goto _match_stored;
}
hashLong[hl1] = (U32)(ip1 - base);
}
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
_match_stored:
/* match found */
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(seqStore, 0, anchor, iend, 0, rLength);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, rLength);
ip += rLength;
anchor = ip;
continue; /* faster when present ... (?) */
const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
ip++;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_REPCODE_1, mLength);
goto _match_stored;
}
offset_2 = offset_1;
offset_1 = offset;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
_match_stored:
/* match found */
const BYTE* const repEnd2 = repIndex2 < prefixLowestIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixLowest) + 4;
U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
- ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, repLength2);
hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
ip += repLength2;
const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
ip++;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_REPCODE_1, mLength);
} else {
if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;
while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; } /* catch up */
offset_2 = offset_1;
offset_1 = offset;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
} else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {
size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
}
offset_2 = offset_1;
offset_1 = offset;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
} else {
ip += ((ip-anchor) >> kSearchStrength) + 1;
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
U32 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
- ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, repLength2);
hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
ip += repLength2;
mLength = ip0[-1] == match0[-1];
ip0 -= mLength;
match0 -= mLength;
- offcode = 0;
+ offcode = STORE_REPCODE_1;
mLength += 4;
goto _match;
}
match0 = base + idx;
rep_offset2 = rep_offset1;
rep_offset1 = (U32)(ip0-match0);
- offcode = rep_offset1 + ZSTD_REP_MOVE;
+ offcode = STORE_OFFSET(rep_offset1);
mLength = 4;
/* Count the backwards match length. */
{ U32 const tmpOff = rep_offset2; rep_offset2 = rep_offset1; rep_offset1 = tmpOff; } /* swap rep_offset2 <=> rep_offset1 */
hashTable[ZSTD_hashPtr(ip0, hlog, mls)] = (U32)(ip0-base);
ip0 += rLength;
- ZSTD_storeSeq(seqStore, 0 /*litLen*/, anchor, iend, 0 /*offCode*/, rLength);
+ ZSTD_storeSeq(seqStore, 0 /*litLen*/, anchor, iend, STORE_REPCODE_1, rLength);
anchor = ip0;
continue; /* faster when present (confirmed on gcc-8) ... (?) */
} } }
const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
ip++;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_REPCODE_1, mLength);
} else if ( (matchIndex <= prefixStartIndex) ) {
size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
U32 const dictMatchIndex = dictHashTable[dictHash];
} /* catch up */
offset_2 = offset_1;
offset_1 = offset;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
}
} else if (MEM_read32(match) != MEM_read32(ip)) {
/* it's not a match, and we're not going to check the dictionary */
&& (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
offset_2 = offset_1;
offset_1 = offset;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
}
/* match found */
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
- ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, repLength2);
hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
ip += repLength2;
anchor = ip;
const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
size_t const rLength = ZSTD_count_2segments(ip+1 +4, repMatch +4, iend, repMatchEnd, prefixStart) + 4;
ip++;
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, rLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_REPCODE_1, rLength);
ip += rLength;
anchor = ip;
} else {
size_t mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
offset_2 = offset_1; offset_1 = offset; /* update offset history */
- ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength);
+ ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, STORE_OFFSET(offset), mLength);
ip += mLength;
anchor = ip;
} }
const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
{ U32 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; } /* swap offset_2 <=> offset_1 */
- ZSTD_storeSeq(seqStore, 0 /*litlen*/, anchor, iend, 0 /*offcode*/, repLength2);
+ ZSTD_storeSeq(seqStore, 0 /*litlen*/, anchor, iend, STORE_REPCODE_1, repLength2);
hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
ip += repLength2;
anchor = ip;
U32 matchIndex = dictMatchIndex + dictIndexDelta;
if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(curr-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)",
- curr, (U32)bestLength, (U32)matchLength, (U32)*offsetPtr, ZSTD_REP_MOVE + curr - matchIndex, dictMatchIndex, matchIndex);
- bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + curr - matchIndex;
+ curr, (U32)bestLength, (U32)matchLength, (U32)*offsetPtr, STORE_OFFSET(curr - matchIndex), dictMatchIndex, matchIndex);
+ bestLength = matchLength, *offsetPtr = STORE_OFFSET(curr - 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 (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength;
if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit32(curr-matchIndex+1) - ZSTD_highbit32((U32)offsetPtr[0]+1)) )
- bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + curr - matchIndex;
+ bestLength = matchLength, *offsetPtr = STORE_OFFSET(curr - 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
/* save best solution */
if (currentMl > ml) {
ml = currentMl;
- *offsetPtr = curr - (matchIndex + ddsIndexDelta) + ZSTD_REP_MOVE;
+ *offsetPtr = STORE_OFFSET(curr - (matchIndex + ddsIndexDelta));
if (ip+currentMl == iLimit) {
/* best possible, avoids read overflow on next attempt */
return ml;
/* save best solution */
if (currentMl > ml) {
ml = currentMl;
- *offsetPtr = curr - (matchIndex + ddsIndexDelta) + ZSTD_REP_MOVE;
+ *offsetPtr = STORE_OFFSET(curr - (matchIndex + ddsIndexDelta));
if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
}
}
/* save best solution */
if (currentMl > ml) {
ml = currentMl;
- *offsetPtr = curr - matchIndex + ZSTD_REP_MOVE;
+ *offsetPtr = STORE_OFFSET(curr - matchIndex);
if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
}
/* save best solution */
if (currentMl > ml) {
ml = currentMl;
- *offsetPtr = curr - (matchIndex + dmsIndexDelta) + ZSTD_REP_MOVE;
+ assert(curr > matchIndex + dmsIndexDelta);
+ *offsetPtr = STORE_OFFSET(curr - (matchIndex + dmsIndexDelta));
if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
}
/* Save best solution */
if (currentMl > ml) {
ml = currentMl;
- *offsetPtr = curr - matchIndex + ZSTD_REP_MOVE;
+ *offsetPtr = STORE_OFFSET(curr - matchIndex);
if (ip+currentMl == iLimit) break; /* best possible, avoids read overflow on next attempt */
}
}
if (currentMl > ml) {
ml = currentMl;
- *offsetPtr = curr - (matchIndex + dmsIndexDelta) + ZSTD_REP_MOVE;
+ assert(curr > matchIndex + dmsIndexDelta);
+ *offsetPtr = STORE_OFFSET(curr - (matchIndex + dmsIndexDelta));
if (ip+currentMl == iLimit) break;
}
}
const BYTE* const repEnd2 = repIndex < prefixLowestIndex ? dictEnd : iend;
matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd2, prefixLowest) + 4;
offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap offset_2 <=> offset_1 */
- ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, matchLength);
ip += matchLength;
anchor = ip;
continue;
/* store sequence */
matchLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap repcodes */
- ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, matchLength);
ip += matchLength;
anchor = ip;
continue; /* faster when present ... (?) */
const BYTE* const repEnd = repIndex < dictLimit ? dictEnd : iend;
matchLength = ZSTD_count_2segments(ip+4, repMatch+4, iend, repEnd, prefixStart) + 4;
offset = offset_2; offset_2 = offset_1; offset_1 = (U32)offset; /* swap offset history */
- ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength);
+ ZSTD_storeSeq(seqStore, 0, anchor, iend, STORE_REPCODE_1, matchLength);
ip += matchLength;
anchor = ip;
continue; /* faster when present ... (?) */
rep[0] = sequence.offset;
/* Store the sequence */
ZSTD_storeSeq(seqStore, newLitLength, ip - newLitLength, iend,
- sequence.offset + ZSTD_REP_MOVE,
+ STORE_OFFSET(sequence.offset),
sequence.matchLength);
ip += sequence.matchLength;
}
/* ZSTD_getMatchPrice() :
* Provides the cost of the match part (offset + matchLength) of a sequence
* Must be combined with ZSTD_fullLiteralsCost() to get the full cost of a sequence.
- * optLevel: when <2, favors small offset for decompression speed (improved cache efficiency) */
+ * @offcode : expects a scale where 0,1,2 are repcodes 1-3, and 3+ are real_offsets+2
+ * @optLevel: when <2, favors small offset for decompression speed (improved cache efficiency)
+ */
FORCE_INLINE_TEMPLATE U32
-ZSTD_getMatchPrice(U32 const offset,
+ZSTD_getMatchPrice(U32 const offcode,
U32 const matchLength,
const optState_t* const optPtr,
int const optLevel)
{
U32 price;
- U32 const offCode = ZSTD_highbit32(offset+1);
+ U32 const offCode = ZSTD_highbit32(offcode+1);
U32 const mlBase = matchLength - MINMATCH;
assert(matchLength >= MINMATCH);
DEBUGLOG(8, "found repCode %u (ll0:%u, offset:%u) of length %u",
repCode, ll0, repOffset, repLen);
bestLength = repLen;
- matches[mnum].off = repCode - ll0;
+ matches[mnum].off = STORE_REPCODE(repCode - ll0 + 1); /* expect value between 1 and 3 */
matches[mnum].len = (U32)repLen;
mnum++;
if ( (repLen > sufficient_len)
bestLength = mlen;
assert(curr > matchIndex3);
assert(mnum==0); /* no prior solution */
- matches[0].off = (curr - matchIndex3) + ZSTD_REP_MOVE;
+ matches[0].off = STORE_OFFSET(curr - matchIndex3);
matches[0].len = (U32)mlen;
mnum = 1;
if ( (mlen > sufficient_len) |
if (matchLength > bestLength) {
DEBUGLOG(8, "found match of length %u at distance %u (offCode=%u)",
- (U32)matchLength, curr - matchIndex, curr - matchIndex + ZSTD_REP_MOVE);
+ (U32)matchLength, curr - matchIndex, STORE_OFFSET(curr - matchIndex));
assert(matchEndIdx > matchIndex);
if (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength;
bestLength = matchLength;
- matches[mnum].off = (curr - matchIndex) + ZSTD_REP_MOVE;
+ matches[mnum].off = STORE_OFFSET(curr - matchIndex);
matches[mnum].len = (U32)matchLength;
mnum++;
if ( (matchLength > ZSTD_OPT_NUM)
if (matchLength > bestLength) {
matchIndex = dictMatchIndex + dmsIndexDelta;
DEBUGLOG(8, "found dms match of length %u at distance %u (offCode=%u)",
- (U32)matchLength, curr - matchIndex, curr - matchIndex + ZSTD_REP_MOVE);
+ (U32)matchLength, curr - matchIndex, STORE_OFFSET(curr - matchIndex));
if (matchLength > matchEndIdx - matchIndex)
matchEndIdx = matchIndex + (U32)matchLength;
bestLength = matchLength;
- matches[mnum].off = (curr - matchIndex) + ZSTD_REP_MOVE;
+ matches[mnum].off = STORE_OFFSET(curr - matchIndex);
matches[mnum].len = (U32)matchLength;
mnum++;
if ( (matchLength > ZSTD_OPT_NUM)
* and 'matchEndPosInBlock', into 'matches'. Maintains the correct ordering of 'matches'
*/
static void ZSTD_optLdm_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
- ZSTD_optLdm_t* optLdm, U32 currPosInBlock) {
- U32 posDiff = currPosInBlock - optLdm->startPosInBlock;
+ ZSTD_optLdm_t* optLdm, U32 currPosInBlock)
+{
+ U32 const posDiff = currPosInBlock - optLdm->startPosInBlock;
/* Note: ZSTD_match_t actually contains offCode and matchLength (before subtracting MINMATCH) */
- U32 candidateMatchLength = optLdm->endPosInBlock - optLdm->startPosInBlock - posDiff;
- U32 candidateOffCode = optLdm->offset + ZSTD_REP_MOVE;
+ U32 const candidateMatchLength = optLdm->endPosInBlock - optLdm->startPosInBlock - posDiff;
+ U32 const candidateOffCode = STORE_OFFSET(optLdm->offset);
/* Ensure that current block position is not outside of the match */
if (currPosInBlock < optLdm->startPosInBlock
/* large match -> immediate encoding */
{ U32 const maxML = matches[nbMatches-1].len;
- U32 const maxOffset = matches[nbMatches-1].off;
+ U32 const maxOffcode = matches[nbMatches-1].off;
DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new series",
- nbMatches, maxML, maxOffset, (U32)(ip-prefixStart));
+ nbMatches, maxML, maxOffcode, (U32)(ip-prefixStart));
if (maxML > sufficient_len) {
lastSequence.litlen = litlen;
lastSequence.mlen = maxML;
- lastSequence.off = maxOffset;
+ lastSequence.off = maxOffcode;
DEBUGLOG(6, "large match (%u>%u), immediate encoding",
maxML, sufficient_len);
cur = 0;
opt[pos].price = ZSTD_MAX_PRICE; /* mlen, litlen and price will be fixed during forward scanning */
}
for (matchNb = 0; matchNb < nbMatches; matchNb++) {
- U32 const offset = matches[matchNb].off;
+ U32 const offcode = matches[matchNb].off;
U32 const end = matches[matchNb].len;
for ( ; pos <= end ; pos++ ) {
- U32 const matchPrice = ZSTD_getMatchPrice(offset, pos, optStatePtr, optLevel);
+ U32 const matchPrice = ZSTD_getMatchPrice(offcode, pos, optStatePtr, optLevel);
U32 const sequencePrice = literalsPrice + matchPrice;
DEBUGLOG(7, "rPos:%u => set initial price : %.2f",
pos, ZSTD_fCost(sequencePrice));
opt[pos].mlen = pos;
- opt[pos].off = offset;
+ opt[pos].off = offcode;
opt[pos].litlen = litlen;
opt[pos].price = (int)sequencePrice;
} }
}
/* Randomly generate sequence commands */
-static U32 generateSequences(U32* seed, frame_t* frame, seqStore_t* seqStore,
- size_t contentSize, size_t literalsSize, dictInfo info)
+static U32
+generateSequences(U32* seed, frame_t* frame, seqStore_t* seqStore,
+ size_t contentSize, size_t literalsSize, dictInfo info)
{
/* The total length of all the matches */
size_t const remainingMatch = contentSize - literalsSize;
size_t excessMatch = 0;
U32 numSequences = 0;
-
U32 i;
-
const BYTE* literals = LITERAL_BUFFER;
BYTE* srcPtr = frame->src;
lenPastStart = MIN(lenPastStart+MIN_SEQ_LEN, (U32)info.dictContentSize);
offset = (U32)((BYTE*)srcPtr - (BYTE*)frame->srcStart) + lenPastStart;
}
- {
- U32 const matchLenBound = MIN(frame->header.windowSize, lenPastStart);
+ { U32 const matchLenBound = MIN(frame->header.windowSize, lenPastStart);
matchLen = MIN(matchLen, matchLenBound);
}
}
}
- offsetCode = offset + ZSTD_REP_MOVE;
+ offsetCode = STORE_OFFSET(offset);
repIndex = 2;
} else {
/* do a repeat offset */
- offsetCode = RAND(seed) % 3;
+ U32 const randomRepIndex = RAND(seed) % 3;
+ offsetCode = STORE_REPCODE(randomRepIndex + 1); /* expects values between 1 & 3 */
if (literalLen > 0) {
- offset = frame->stats.rep[offsetCode];
- repIndex = offsetCode;
+ offset = frame->stats.rep[randomRepIndex];
+ repIndex = randomRepIndex;
} else {
- /* special case */
- offset = offsetCode == 2 ? frame->stats.rep[0] - 1
- : frame->stats.rep[offsetCode + 1];
- repIndex = MIN(2, offsetCode + 1);
+ /* special case : literalLen == 0 */
+ offset = randomRepIndex == 2 ? frame->stats.rep[0] - 1
+ : frame->stats.rep[randomRepIndex + 1];
+ repIndex = MIN(2, randomRepIndex + 1);
}
}
} while (((!info.useDict) && (offset > (size_t)((BYTE*)srcPtr - (BYTE*)frame->srcStart))) || offset == 0);
- {
+ { BYTE* const dictEnd = info.dictContent + info.dictContentSize;
size_t j;
- BYTE* const dictEnd = info.dictContent + info.dictContentSize;
for (j = 0; j < matchLen; j++) {
if ((U32)((BYTE*)srcPtr - (BYTE*)frame->srcStart) < offset) {
/* copy from dictionary instead of literals */
*srcPtr = *(srcPtr-offset);
}
srcPtr++;
- }
- }
+ } }
{ int r;
for (r = repIndex; r > 0; r--) {
DISPLAYLEVEL(7, " srcPos: %8u seqNb: %3u",
(unsigned)((BYTE*)srcPtr - (BYTE*)frame->srcStart), (unsigned)i);
DISPLAYLEVEL(6, "\n");
- if (offsetCode < 3) {
+ if (offsetCode < ZSTD_REP_NUM) { /* expects @offsetCode to use 0-2 to represents repCodes */
DISPLAYLEVEL(7, " repeat offset: %d\n", (int)repIndex);
}
/* use libzstd sequence handling */
projects / thirdparty / zstd.git / commitdiff
