facebook
/
zstd
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Refactor usage of ms->ldmSeqStore so that it is not modified during compressBlock(), and simplify skipRawSeqStoreBytes

dev
senhuang42 2020-10-06 13:43:13 -04:00
parent a5500cf2af
commit 0fac8e07e1
2 changed files with 64 additions and 62 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
lib/compress/zstd_ldm.c
View File

@ -555,6 +555,26 @@ static rawSeq maybeSplitSequence(rawSeqStore_t* rawSeqStore,
return sequence; return sequence;
} }
/* ZSTD_ldm_skipRawSeqStoreBytes():
* Moves forward in rawSeqStore by nbBytes, updating fields 'pos' and 'posInSequence'.
*/
static void ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes) {
U32 currPos = rawSeqStore->posInSequence + nbBytes;
while (currPos && rawSeqStore->pos < rawSeqStore->size) {
rawSeq currSeq = rawSeqStore->seq[rawSeqStore->pos];
if (currPos >= currSeq.litLength + currSeq.matchLength) {
currPos -= currSeq.litLength + currSeq.matchLength;
rawSeqStore->pos++;
if (currPos == 0) {
rawSeqStore->posInSequence = 0;
}
} else {
rawSeqStore->posInSequence = currPos;
break;
}
}
}
size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore, size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM], ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize) void const* src, size_t srcSize)
@ -574,7 +594,11 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
size_t lastLLSize; size_t lastLLSize;
ms->ldmSeqStore = *rawSeqStore; ms->ldmSeqStore = *rawSeqStore;
lastLLSize = blockCompressor(ms, seqStore, rep, src, srcSize); lastLLSize = blockCompressor(ms, seqStore, rep, src, srcSize);
*rawSeqStore = ms->ldmSeqStore; ZSTD_ldm_skipRawSeqStoreBytes(rawSeqStore, srcSize);
if (rawSeqStore->pos >= rawSeqStore->size) {
/* If we're done with rawSeqStore, invalidate the one in matchState as well */
ms->ldmSeqStore.pos = rawSeqStore->size;
}
return lastLLSize; return lastLLSize;
} }

100
lib/compress/zstd_opt.c
View File

@ -768,39 +768,30 @@ FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllMatches (
* LDM helper functions * * LDM helper functions *
*************************/ *************************/
/* Struct containing info needed to make decision about ldm inclusion */
typedef struct { typedef struct {
rawSeqStore_t* seqStore; /* Reference to struct containing external candidates */ rawSeqStore_t seqStore; /* External match candidates store for this block */
U32 startPosInBlock; /* Start position of the current match candidate */ U32 startPosInBlock; /* Start position of the current match candidate */
U32 endPosInBlock; /* End position of the current match candidate */ U32 endPosInBlock; /* End position of the current match candidate */
U32 offset; /* Offset of the match candidate */ U32 offset; /* Offset of the match candidate */
} ZSTD_optLdm_t; } ZSTD_optLdm_t;
/* ZSTD_opt_skipRawSeqStoreBytes(): /* ZSTD_optLdm_skipRawSeqStoreBytes():
* Moves forward in rawSeqStore by nbBytes, which will update the fields 'pos' and 'posInSequence'. * Moves forward in rawSeqStore by nbBytes, which will update the fields 'pos' and 'posInSequence'.
*/ */
static void ZSTD_opt_skipRawSeqStoreBytes(rawSeqStore_t* ldmSeqStore, size_t nbBytes) { static void ZSTD_optLdm_skipRawSeqStoreBytes(rawSeqStore_t* rawSeqStore, size_t nbBytes) {
while (nbBytes && ldmSeqStore->pos < ldmSeqStore->size) { U32 currPos = rawSeqStore->posInSequence + nbBytes;
rawSeq currSeq = ldmSeqStore->seq[ldmSeqStore->pos]; while (currPos && rawSeqStore->pos < rawSeqStore->size) {
/* posInSequence necessarily must never represent a value beyond the sequence */ rawSeq currSeq = rawSeqStore->seq[rawSeqStore->pos];
assert(ldmSeqStore->posInSequence <= currSeq.matchLength + currSeq.litLength); if (currPos >= currSeq.litLength + currSeq.matchLength) {
currPos -= currSeq.litLength + currSeq.matchLength;
if (nbBytes <= currSeq.litLength) { rawSeqStore->pos++;
ldmSeqStore->posInSequence += nbBytes; if (currPos == 0) {
return; rawSeqStore->posInSequence = 0;
}
} else { } else {
ldmSeqStore->posInSequence += currSeq.litLength; rawSeqStore->posInSequence = currPos;
nbBytes -= currSeq.litLength; break;
}
if (nbBytes < currSeq.matchLength) {
ldmSeqStore->posInSequence += nbBytes;
return;
} else {
nbBytes -= currSeq.matchLength;
/* We have moved through this entire sequence - move the read pos
forward to the next sequence, and reset posInSequence */
ldmSeqStore->pos++;
ldmSeqStore->posInSequence = 0;
} }
} }
} }
@ -817,28 +808,28 @@ static void ZSTD_opt_getNextMatchAndUpdateSeqStore(ZSTD_optLdm_t* optLdm, U32 cu
U32 matchBytesRemaining; U32 matchBytesRemaining;
/* Setting match end position to MAX to ensure we never use an LDM during this block */ /* Setting match end position to MAX to ensure we never use an LDM during this block */
if (optLdm->seqStore->pos >= optLdm->seqStore->size) { if (optLdm->seqStore.pos >= optLdm->seqStore.size || optLdm->seqStore.size == 0) {
optLdm->startPosInBlock = UINT_MAX; optLdm->startPosInBlock = UINT_MAX;
optLdm->endPosInBlock = UINT_MAX; optLdm->endPosInBlock = UINT_MAX;
return; return;
} }
/* Calculate appropriate bytes left in matchLength and litLength after adjusting /* Calculate appropriate bytes left in matchLength and litLength after adjusting
based on ldmSeqStore->posInSequence */ based on ldmSeqStore->posInSequence */
currSeq = optLdm->seqStore->seq[optLdm->seqStore->pos]; currSeq = optLdm->seqStore.seq[optLdm->seqStore.pos];
assert(optLdm->seqStore->posInSequence <= currSeq.litLength + currSeq.matchLength); assert(optLdm->seqStore.posInSequence <= currSeq.litLength + currSeq.matchLength);
currBlockEndPos = currPosInBlock + blockBytesRemaining; currBlockEndPos = currPosInBlock + blockBytesRemaining;
literalsBytesRemaining = (optLdm->seqStore->posInSequence < currSeq.litLength) ? literalsBytesRemaining = (optLdm->seqStore.posInSequence < currSeq.litLength) ?
currSeq.litLength - (U32)optLdm->seqStore->posInSequence : currSeq.litLength - (U32)optLdm->seqStore.posInSequence :
0; 0;
matchBytesRemaining = (literalsBytesRemaining == 0) ? matchBytesRemaining = (literalsBytesRemaining == 0) ?
currSeq.matchLength - ((U32)optLdm->seqStore->posInSequence - currSeq.litLength) : currSeq.matchLength - ((U32)optLdm->seqStore.posInSequence - currSeq.litLength) :
currSeq.matchLength; currSeq.matchLength;
/* If there are more literal bytes than bytes remaining in block, no ldm is possible */ /* If there are more literal bytes than bytes remaining in block, no ldm is possible */
if (literalsBytesRemaining >= blockBytesRemaining) { if (literalsBytesRemaining >= blockBytesRemaining) {
optLdm->startPosInBlock = UINT_MAX; optLdm->startPosInBlock = UINT_MAX;
optLdm->endPosInBlock = UINT_MAX; optLdm->endPosInBlock = UINT_MAX;
ZSTD_opt_skipRawSeqStoreBytes(optLdm->seqStore, blockBytesRemaining); ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, blockBytesRemaining);
return; return;
} }
@ -851,21 +842,18 @@ static void ZSTD_opt_getNextMatchAndUpdateSeqStore(ZSTD_optLdm_t* optLdm, U32 cu
if (optLdm->endPosInBlock > currBlockEndPos) { if (optLdm->endPosInBlock > currBlockEndPos) {
/* Match ends after the block ends, we can't use the whole match */ /* Match ends after the block ends, we can't use the whole match */
optLdm->endPosInBlock = currBlockEndPos; optLdm->endPosInBlock = currBlockEndPos;
ZSTD_opt_skipRawSeqStoreBytes(optLdm->seqStore, currBlockEndPos - currPosInBlock); ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, currBlockEndPos - currPosInBlock);
} else { } else {
/* If we can use the whole match point the ldmSeqStore at the next match */ /* Consume nb of bytes equal to size of sequence left */
optLdm->seqStore->posInSequence = 0; ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, literalsBytesRemaining + matchBytesRemaining);
optLdm->seqStore->pos++;
} }
DEBUGLOG(6, "ZSTD_opt_getNextMatchAndUpdateSeqStore(): got an ldm that beginning at pos: %u, end at pos: %u, with offset: %u",
optLdm->startPosInBlock, optLdm->endPosInBlock, optLdm->offset);
} }
/* ZSTD_opt_maybeAddMatch(): /* ZSTD_optLdm_maybeAddMatch():
* Adds a match if it's long enough, based on it's 'matchStartPosInBlock' * Adds a match if it's long enough, based on it's 'matchStartPosInBlock'
* and 'matchEndPosInBlock', into 'matches'. Maintains the correct ordering of 'matches' * and 'matchEndPosInBlock', into 'matches'. Maintains the correct ordering of 'matches'
*/ */
static void ZSTD_opt_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches, static void ZSTD_optLdm_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
ZSTD_optLdm_t* optLdm, U32 currPosInBlock) { ZSTD_optLdm_t* optLdm, U32 currPosInBlock) {
U32 posDiff = currPosInBlock - optLdm->startPosInBlock; U32 posDiff = currPosInBlock - optLdm->startPosInBlock;
/* Note: ZSTD_match_t actually contains offCode and matchLength (before subtracting MINMATCH) */ /* Note: ZSTD_match_t actually contains offCode and matchLength (before subtracting MINMATCH) */
@ -879,7 +867,7 @@ static void ZSTD_opt_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
return; return;
} }
DEBUGLOG(6, "ZSTD_opt_maybeAddMatch(): Adding ldm candidate match (offCode: %u matchLength %u) at block position=%u", DEBUGLOG(6, "ZSTD_optLdm_maybeAddMatch(): Adding ldm candidate match (offCode: %u matchLength %u) at block position=%u",
candidateOffCode, candidateMatchLength, currPosInBlock); candidateOffCode, candidateMatchLength, currPosInBlock);
if (*nbMatches == 0) { if (*nbMatches == 0) {
matches[*nbMatches].len = candidateMatchLength; matches[*nbMatches].len = candidateMatchLength;
@ -896,27 +884,27 @@ static void ZSTD_opt_maybeAddMatch(ZSTD_match_t* matches, U32* nbMatches,
} }
} }
/* ZSTD_opt_processMatchCandidate(): /* ZSTD_optLdm_processMatchCandidate():
* Wrapper function to update ldm seq store and call ldm functions as necessary. * Wrapper function to update ldm seq store and call ldm functions as necessary.
*/ */
static void ZSTD_opt_processMatchCandidate(ZSTD_optLdm_t* optLdm, ZSTD_match_t* matches, U32* nbMatches, static void ZSTD_optLdm_processMatchCandidate(ZSTD_optLdm_t* optLdm, ZSTD_match_t* matches, U32* nbMatches,
U32 currPosInBlock, U32 remainingBytes) { U32 currPosInBlock, U32 remainingBytes) {
if (optLdm->seqStore->size == 0 && optLdm->seqStore->pos >= optLdm->seqStore->size) { if (optLdm->seqStore.size == 0 && optLdm->seqStore.pos >= optLdm->seqStore.size) {
return; return;
} }
if (currPosInBlock >= optLdm->endPosInBlock) { if (currPosInBlock >= optLdm->endPosInBlock) {
if (currPosInBlock > optLdm->endPosInBlock) { if (currPosInBlock > optLdm->endPosInBlock) {
/* The position at which ZSTD_opt_processMatchCandidate() is called is not necessarily /* The position at which ZSTD_optLdm_processMatchCandidate() is called is not necessarily
* at the end of a match from the ldm seq store, and will often be some bytes * at the end of a match from the ldm seq store, and will often be some bytes
* over beyond matchEndPosInBlock. As such, we need to correct for these "overshoots" * over beyond matchEndPosInBlock. As such, we need to correct for these "overshoots"
*/ */
U32 posOvershoot = currPosInBlock - optLdm->endPosInBlock; U32 posOvershoot = currPosInBlock - optLdm->endPosInBlock;
ZSTD_opt_skipRawSeqStoreBytes(optLdm->seqStore, posOvershoot); ZSTD_optLdm_skipRawSeqStoreBytes(&optLdm->seqStore, posOvershoot);
} }
ZSTD_opt_getNextMatchAndUpdateSeqStore(optLdm, currPosInBlock, remainingBytes); ZSTD_opt_getNextMatchAndUpdateSeqStore(optLdm, currPosInBlock, remainingBytes);
} }
ZSTD_opt_maybeAddMatch(matches, nbMatches, optLdm, currPosInBlock); ZSTD_optLdm_maybeAddMatch(matches, nbMatches, optLdm, currPosInBlock);
} }
/*-******************************* /*-*******************************
@ -971,14 +959,10 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
ZSTD_optimal_t* const opt = optStatePtr->priceTable; ZSTD_optimal_t* const opt = optStatePtr->priceTable;
ZSTD_match_t* const matches = optStatePtr->matchTable; ZSTD_match_t* const matches = optStatePtr->matchTable;
ZSTD_optimal_t lastSequence; ZSTD_optimal_t lastSequence;
ZSTD_optLdm_t optLdm = {ms->ldmSeqStore, 0, 0, 0}; /* ms->ldmSeqStore itself is immutable in this function */
/* Make a copy so that ms->ldmSeqStore stays immutable during compressBlock() */
ZSTD_optLdm_t optLdm = {&ms->ldmSeqStore, 0, 0, 0};
/* Get first match from ldm seq store if long mode is enabled */ /* Get first match from ldm seq store if long mode is enabled */
if (ms->ldmSeqStore.size > 0 && ms->ldmSeqStore.pos < ms->ldmSeqStore.size) { ZSTD_opt_getNextMatchAndUpdateSeqStore(&optLdm, (U32)(ip-istart), (U32)(iend-ip));
ZSTD_opt_getNextMatchAndUpdateSeqStore(&optLdm, (U32)(ip-istart), (U32)(iend-ip));
}
/* init */ /* init */
DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u", DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u",
@ -995,7 +979,7 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
{ U32 const litlen = (U32)(ip - anchor); { U32 const litlen = (U32)(ip - anchor);
U32 const ll0 = !litlen; U32 const ll0 = !litlen;
U32 nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, ip, iend, dictMode, rep, ll0, minMatch); U32 nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, ip, iend, dictMode, rep, ll0, minMatch);
ZSTD_opt_processMatchCandidate(&optLdm, matches, &nbMatches, ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
(U32)(ip-istart), (U32)(iend - ip)); (U32)(ip-istart), (U32)(iend - ip));
if (!nbMatches) { ip++; continue; } if (!nbMatches) { ip++; continue; }
@ -1112,7 +1096,7 @@ ZSTD_compressBlock_opt_generic(ZSTD_matchState_t* ms,
U32 nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, inr, iend, dictMode, opt[cur].rep, ll0, minMatch); U32 nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, inr, iend, dictMode, opt[cur].rep, ll0, minMatch);
U32 matchNb; U32 matchNb;
ZSTD_opt_processMatchCandidate(&optLdm, matches, &nbMatches, ZSTD_optLdm_processMatchCandidate(&optLdm, matches, &nbMatches,
(U32)(inr-istart), (U32)(iend-inr)); (U32)(inr-istart), (U32)(iend-inr));
if (!nbMatches) { if (!nbMatches) {
@ -1228,13 +1212,7 @@ _shortestPath: /* cur, last_pos, best_mlen, best_off have to be set */
ZSTD_setBasePrices(optStatePtr, optLevel); ZSTD_setBasePrices(optStatePtr, optLevel);
} }
} /* while (ip < ilimit) */ } /* while (ip < ilimit) */
if (optLdm.endPosInBlock < srcSize) {
/* This can occur if after adding the final match in an ldm seq store within this block,
ip reaches end of the block without calling ZSTD_opt_getNextLdmAndUpdateSeqStore() */
ZSTD_opt_skipRawSeqStoreBytes(optLdm.seqStore, srcSize - optLdm.endPosInBlock);
}
/* Return the last literals size */ /* Return the last literals size */
return (size_t)(iend - anchor); return (size_t)(iend - anchor);
} }