summaryrefslogtreecommitdiff
path: root/thirdparty/zstd/compress/zstd_fast.c
diff options
context:
space:
mode:
Diffstat (limited to 'thirdparty/zstd/compress/zstd_fast.c')
-rw-r--r--thirdparty/zstd/compress/zstd_fast.c327
1 files changed, 203 insertions, 124 deletions
diff --git a/thirdparty/zstd/compress/zstd_fast.c b/thirdparty/zstd/compress/zstd_fast.c
index 40ba0f73e6..ed997b441c 100644
--- a/thirdparty/zstd/compress/zstd_fast.c
+++ b/thirdparty/zstd/compress/zstd_fast.c
@@ -45,7 +45,155 @@ 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 mls, ZSTD_dictMode_e const dictMode)
+ U32 const mls)
+{
+ const ZSTD_compressionParameters* const cParams = &ms->cParams;
+ U32* const hashTable = ms->hashTable;
+ U32 const hlog = cParams->hashLog;
+ /* support stepSize of 0 */
+ size_t const stepSize = cParams->targetLength + !(cParams->targetLength) + 1;
+ const BYTE* const base = ms->window.base;
+ const BYTE* const istart = (const BYTE*)src;
+ /* We check ip0 (ip + 0) and ip1 (ip + 1) each loop */
+ const BYTE* ip0 = istart;
+ const BYTE* ip1;
+ const BYTE* anchor = istart;
+ const U32 prefixStartIndex = ms->window.dictLimit;
+ const BYTE* const prefixStart = base + prefixStartIndex;
+ 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 */
+ ip0 += (ip0 == prefixStart);
+ ip1 = ip0 + 1;
+ {
+ U32 const maxRep = (U32)(ip0 - prefixStart);
+ if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
+ if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
+ }
+
+ /* Main Search Loop */
+ while (ip1 < ilimit) { /* < instead of <=, because check at ip0+2 */
+ size_t mLength;
+ BYTE const* ip2 = ip0 + 2;
+ size_t const h0 = ZSTD_hashPtr(ip0, hlog, mls);
+ U32 const val0 = MEM_read32(ip0);
+ size_t const h1 = ZSTD_hashPtr(ip1, hlog, mls);
+ U32 const val1 = MEM_read32(ip1);
+ U32 const current0 = (U32)(ip0-base);
+ U32 const current1 = (U32)(ip1-base);
+ U32 const matchIndex0 = hashTable[h0];
+ U32 const matchIndex1 = hashTable[h1];
+ BYTE const* repMatch = ip2-offset_1;
+ const BYTE* match0 = base + matchIndex0;
+ const BYTE* match1 = base + matchIndex1;
+ U32 offcode;
+ hashTable[h0] = current0; /* update hash table */
+ hashTable[h1] = current1; /* update hash table */
+
+ assert(ip0 + 1 == ip1);
+
+ if ((offset_1 > 0) & (MEM_read32(repMatch) == MEM_read32(ip2))) {
+ mLength = ip2[-1] == repMatch[-1] ? 1 : 0;
+ ip0 = ip2 - mLength;
+ match0 = repMatch - mLength;
+ offcode = 0;
+ goto _match;
+ }
+ if ((matchIndex0 > prefixStartIndex) && MEM_read32(match0) == val0) {
+ /* found a regular match */
+ goto _offset;
+ }
+ if ((matchIndex1 > prefixStartIndex) && MEM_read32(match1) == val1) {
+ /* found a regular match after one literal */
+ ip0 = ip1;
+ match0 = match1;
+ goto _offset;
+ }
+ {
+ size_t const step = ((ip0-anchor) >> (kSearchStrength - 1)) + stepSize;
+ assert(step >= 2);
+ ip0 += step;
+ ip1 += step;
+ continue;
+ }
+_offset: /* Requires: ip0, match0 */
+ /* Compute the offset code */
+ offset_2 = offset_1;
+ offset_1 = (U32)(ip0-match0);
+ offcode = offset_1 + ZSTD_REP_MOVE;
+ mLength = 0;
+ /* Count the backwards match length */
+ while (((ip0>anchor) & (match0>prefixStart))
+ && (ip0[-1] == match0[-1])) { ip0--; match0--; mLength++; } /* catch up */
+
+_match: /* Requires: ip0, match0, offcode */
+ /* Count the forward length */
+ mLength += ZSTD_count(ip0+mLength+4, match0+mLength+4, iend) + 4;
+ ZSTD_storeSeq(seqStore, ip0-anchor, anchor, offcode, mLength-MINMATCH);
+ /* match found */
+ ip0 += mLength;
+ anchor = ip0;
+ ip1 = ip0 + 1;
+
+ if (ip0 <= ilimit) {
+ /* Fill Table */
+ assert(base+current0+2 > istart); /* check base overflow */
+ hashTable[ZSTD_hashPtr(base+current0+2, hlog, mls)] = current0+2; /* here because current+2 could be > iend-8 */
+ hashTable[ZSTD_hashPtr(ip0-2, hlog, mls)] = (U32)(ip0-2-base);
+
+ while ( (ip0 <= ilimit)
+ && ( (offset_2>0)
+ & (MEM_read32(ip0) == MEM_read32(ip0 - offset_2)) )) {
+ /* store sequence */
+ size_t const rLength = ZSTD_count(ip0+4, ip0+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(ip0, hlog, mls)] = (U32)(ip0-base);
+ ip0 += rLength;
+ ip1 = ip0 + 1;
+ ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
+ anchor = ip0;
+ continue; /* faster when present (confirmed on gcc-8) ... (?) */
+ }
+ }
+ }
+
+ /* save reps for next block */
+ rep[0] = offset_1 ? offset_1 : offsetSaved;
+ rep[1] = offset_2 ? offset_2 : offsetSaved;
+
+ /* Return the last literals size */
+ return iend - anchor;
+}
+
+
+size_t ZSTD_compressBlock_fast(
+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+ void const* src, size_t srcSize)
+{
+ ZSTD_compressionParameters const* cParams = &ms->cParams;
+ U32 const mls = cParams->minMatch;
+ assert(ms->dictMatchState == NULL);
+ switch(mls)
+ {
+ default: /* includes case 3 */
+ case 4 :
+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4);
+ case 5 :
+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5);
+ case 6 :
+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6);
+ case 7 :
+ return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7);
+ }
+}
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_fast_dictMatchState_generic(
+ ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+ void const* src, size_t srcSize, U32 const mls)
{
const ZSTD_compressionParameters* const cParams = &ms->cParams;
U32* const hashTable = ms->hashTable;
@@ -64,46 +212,26 @@ size_t ZSTD_compressBlock_fast_generic(
U32 offsetSaved = 0;
const ZSTD_matchState_t* const dms = ms->dictMatchState;
- const ZSTD_compressionParameters* const dictCParams =
- dictMode == ZSTD_dictMatchState ?
- &dms->cParams : NULL;
- const U32* const dictHashTable = dictMode == ZSTD_dictMatchState ?
- dms->hashTable : NULL;
- const U32 dictStartIndex = dictMode == ZSTD_dictMatchState ?
- dms->window.dictLimit : 0;
- const BYTE* const dictBase = dictMode == ZSTD_dictMatchState ?
- dms->window.base : NULL;
- const BYTE* const dictStart = dictMode == ZSTD_dictMatchState ?
- dictBase + dictStartIndex : NULL;
- const BYTE* const dictEnd = dictMode == ZSTD_dictMatchState ?
- dms->window.nextSrc : NULL;
- const U32 dictIndexDelta = dictMode == ZSTD_dictMatchState ?
- prefixStartIndex - (U32)(dictEnd - dictBase) :
- 0;
+ const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;
+ const U32* const dictHashTable = dms->hashTable;
+ const U32 dictStartIndex = dms->window.dictLimit;
+ const BYTE* const dictBase = dms->window.base;
+ const BYTE* const dictStart = dictBase + dictStartIndex;
+ const BYTE* const dictEnd = dms->window.nextSrc;
+ const U32 dictIndexDelta = prefixStartIndex - (U32)(dictEnd - dictBase);
const U32 dictAndPrefixLength = (U32)(ip - prefixStart + dictEnd - dictStart);
- const U32 dictHLog = dictMode == ZSTD_dictMatchState ?
- dictCParams->hashLog : hlog;
-
- assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);
+ const U32 dictHLog = dictCParams->hashLog;
/* otherwise, we would get index underflow when translating a dict index
* into a local index */
- assert(dictMode != ZSTD_dictMatchState
- || prefixStartIndex >= (U32)(dictEnd - dictBase));
+ assert(prefixStartIndex >= (U32)(dictEnd - dictBase));
/* init */
ip += (dictAndPrefixLength == 0);
- if (dictMode == ZSTD_noDict) {
- U32 const maxRep = (U32)(ip - prefixStart);
- if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
- if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
- }
- if (dictMode == ZSTD_dictMatchState) {
- /* dictMatchState repCode checks don't currently handle repCode == 0
- * disabling. */
- assert(offset_1 <= dictAndPrefixLength);
- assert(offset_2 <= dictAndPrefixLength);
- }
+ /* dictMatchState repCode checks don't currently handle repCode == 0
+ * disabling. */
+ assert(offset_1 <= dictAndPrefixLength);
+ assert(offset_2 <= dictAndPrefixLength);
/* Main Search Loop */
while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
@@ -113,50 +241,37 @@ size_t ZSTD_compressBlock_fast_generic(
U32 const matchIndex = hashTable[h];
const BYTE* match = base + matchIndex;
const U32 repIndex = current + 1 - offset_1;
- const BYTE* repMatch = (dictMode == ZSTD_dictMatchState
- && repIndex < prefixStartIndex) ?
+ const BYTE* repMatch = (repIndex < prefixStartIndex) ?
dictBase + (repIndex - dictIndexDelta) :
base + repIndex;
hashTable[h] = current; /* update hash table */
- if ( (dictMode == ZSTD_dictMatchState)
- && ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
+ if ( ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
&& (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
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, ip-anchor, anchor, 0, mLength-MINMATCH);
- } else if ( dictMode == ZSTD_noDict
- && ((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 <= prefixStartIndex) ) {
- if (dictMode == ZSTD_dictMatchState) {
- size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
- U32 const dictMatchIndex = dictHashTable[dictHash];
- const BYTE* dictMatch = dictBase + dictMatchIndex;
- if (dictMatchIndex <= dictStartIndex ||
- MEM_read32(dictMatch) != MEM_read32(ip)) {
- assert(stepSize >= 1);
- ip += ((ip-anchor) >> kSearchStrength) + stepSize;
- continue;
- } else {
- /* found a dict match */
- U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);
- mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
- while (((ip>anchor) & (dictMatch>dictStart))
- && (ip[-1] == dictMatch[-1])) {
- ip--; dictMatch--; mLength++;
- } /* catch up */
- offset_2 = offset_1;
- offset_1 = offset;
- ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
- }
- } else {
+ size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
+ U32 const dictMatchIndex = dictHashTable[dictHash];
+ const BYTE* dictMatch = dictBase + dictMatchIndex;
+ if (dictMatchIndex <= dictStartIndex ||
+ MEM_read32(dictMatch) != MEM_read32(ip)) {
assert(stepSize >= 1);
ip += ((ip-anchor) >> kSearchStrength) + stepSize;
continue;
+ } else {
+ /* found a dict match */
+ U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);
+ mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
+ while (((ip>anchor) & (dictMatch>dictStart))
+ && (ip[-1] == dictMatch[-1])) {
+ ip--; dictMatch--; mLength++;
+ } /* catch up */
+ offset_2 = offset_1;
+ offset_1 = offset;
+ ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
}
} else if (MEM_read32(match) != MEM_read32(ip)) {
/* it's not a match, and we're not going to check the dictionary */
@@ -185,41 +300,27 @@ size_t ZSTD_compressBlock_fast_generic(
hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
/* check immediate repcode */
- if (dictMode == ZSTD_dictMatchState) {
- while (ip <= ilimit) {
- U32 const current2 = (U32)(ip-base);
- U32 const repIndex2 = current2 - offset_2;
- const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
- dictBase - dictIndexDelta + repIndex2 :
- base + repIndex2;
- if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
- && (MEM_read32(repMatch2) == MEM_read32(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 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
- ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
- hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
- ip += repLength2;
- anchor = ip;
- continue;
- }
- break;
+ while (ip <= ilimit) {
+ U32 const current2 = (U32)(ip-base);
+ U32 const repIndex2 = current2 - offset_2;
+ const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
+ dictBase - dictIndexDelta + repIndex2 :
+ base + repIndex2;
+ if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
+ && (MEM_read32(repMatch2) == MEM_read32(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 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
+ ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
+ hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
+ ip += repLength2;
+ anchor = ip;
+ continue;
}
+ break;
}
-
- if (dictMode == ZSTD_noDict) {
- 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 ... (?) */
- } } } }
+ }
+ }
/* save reps for next block */
rep[0] = offset_1 ? offset_1 : offsetSaved;
@@ -229,28 +330,6 @@ size_t ZSTD_compressBlock_fast_generic(
return iend - anchor;
}
-
-size_t ZSTD_compressBlock_fast(
- ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
- void const* src, size_t srcSize)
-{
- ZSTD_compressionParameters const* cParams = &ms->cParams;
- U32 const mls = cParams->minMatch;
- assert(ms->dictMatchState == NULL);
- switch(mls)
- {
- default: /* includes case 3 */
- case 4 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_noDict);
- case 5 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_noDict);
- case 6 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_noDict);
- case 7 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_noDict);
- }
-}
-
size_t ZSTD_compressBlock_fast_dictMatchState(
ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
void const* src, size_t srcSize)
@@ -262,13 +341,13 @@ size_t ZSTD_compressBlock_fast_dictMatchState(
{
default: /* includes case 3 */
case 4 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_dictMatchState);
+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 4);
case 5 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_dictMatchState);
+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 5);
case 6 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_dictMatchState);
+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 6);
case 7 :
- return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_dictMatchState);
+ return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 7);
}
}
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-25 04:54:48 +0000