zstdlib 0.4.0 → 0.5.0

data/ext/zstdlib/{zstd-1.4.3 → zstd-1.4.4}/lib/compress/zstd_double_fast.c

@@ -148,7 +148,7 @@ size_t ZSTD_compressBlock_doubleFast_generic(
             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, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
             goto _match_stored;
         }
 
@@ -157,7 +157,7 @@ size_t ZSTD_compressBlock_doubleFast_generic(
           && ((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, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
             goto _match_stored;
         }
 
@@ -247,7 +247,7 @@ _match_found:
         offset_2 = offset_1;
         offset_1 = offset;
 
-        ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+        ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
 
 _match_stored:
         /* match found */
@@ -278,7 +278,7 @@ _match_stored:
                         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, 0, repLength2-MINMATCH);
+                        ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2-MINMATCH);
                         hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
                         hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
                         ip += repLength2;
@@ -297,7 +297,7 @@ _match_stored:
                     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, 0, rLength-MINMATCH);
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, rLength-MINMATCH);
                     ip += rLength;
                     anchor = ip;
                     continue;   /* faster when present ... (?) */
@@ -411,7 +411,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
             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, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
         } else {
             if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
                 const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;
@@ -422,7 +422,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
                 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, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
 
             } else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {
                 size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
@@ -447,7 +447,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
                 }
                 offset_2 = offset_1;
                 offset_1 = offset;
-                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
 
             } else {
                 ip += ((ip-anchor) >> kSearchStrength) + 1;
@@ -479,7 +479,7 @@ static size_t ZSTD_compressBlock_doubleFast_extDict_generic(
                     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, 0, repLength2-MINMATCH);
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2-MINMATCH);
                     hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = current2;
                     hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = current2;
                     ip += repLength2;

data/ext/zstdlib/{zstd-1.4.3 → zstd-1.4.4}/lib/compress/zstd_fast.c

@@ -8,7 +8,7 @@
  * You may select, at your option, one of the above-listed licenses.
  */
 
-#include "zstd_compress_internal.h"
+#include "zstd_compress_internal.h"  /* ZSTD_hashPtr, ZSTD_count, ZSTD_storeSeq */
 #include "zstd_fast.h"
 
 
@@ -43,8 +43,8 @@ void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
 }
 
 
-FORCE_INLINE_TEMPLATE
-size_t ZSTD_compressBlock_fast_generic(
+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)
@@ -74,8 +74,7 @@ size_t ZSTD_compressBlock_fast_generic(
     DEBUGLOG(5, "ZSTD_compressBlock_fast_generic");
     ip0 += (ip0 == prefixStart);
     ip1 = ip0 + 1;
-    {
-        U32 const maxRep = (U32)(ip0 - prefixStart);
+    {   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;
     }
@@ -118,8 +117,7 @@ size_t ZSTD_compressBlock_fast_generic(
             match0 = match1;
             goto _offset;
         }
-        {
-            size_t const step = ((ip0-anchor) >> (kSearchStrength - 1)) + stepSize;
+        {   size_t const step = ((size_t)(ip0-anchor) >> (kSearchStrength - 1)) + stepSize;
             assert(step >= 2);
             ip0 += step;
             ip1 += step;
@@ -138,7 +136,7 @@ _offset: /* Requires: ip0, match0 */
 _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);
+        ZSTD_storeSeq(seqStore, (size_t)(ip0-anchor), anchor, iend, offcode, mLength-MINMATCH);
         /* match found */
         ip0 += mLength;
         anchor = ip0;
@@ -150,16 +148,15 @@ _match: /* Requires: ip0, match0, offcode */
             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)) )) {
+            while ( ((ip0 <= ilimit) & (offset_2>0))  /* offset_2==0 means offset_2 is invalidated */
+                 && (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 */
+                { 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);
+                ZSTD_storeSeq(seqStore, 0 /*litLen*/, anchor, iend, 0 /*offCode*/, rLength-MINMATCH);
                 anchor = ip0;
                 continue;   /* faster when present (confirmed on gcc-8) ... (?) */
             }
@@ -179,8 +176,7 @@ 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;
+    U32 const mls = ms->cParams.minMatch;
     assert(ms->dictMatchState == NULL);
     switch(mls)
     {
@@ -265,7 +261,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
             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, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, mLength-MINMATCH);
         } else if ( (matchIndex <= prefixStartIndex) ) {
             size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
             U32 const dictMatchIndex = dictHashTable[dictHash];
@@ -285,7 +281,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
                 } /* catch up */
                 offset_2 = offset_1;
                 offset_1 = offset;
-                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 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 */
@@ -300,7 +296,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
                  && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
             offset_2 = offset_1;
             offset_1 = offset;
-            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
         }
 
         /* match found */
@@ -325,7 +321,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState_generic(
                     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);
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, repLength2-MINMATCH);
                     hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
                     ip += repLength2;
                     anchor = ip;
@@ -348,8 +344,7 @@ size_t ZSTD_compressBlock_fast_dictMatchState(
         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;
+    U32 const mls = ms->cParams.minMatch;
     assert(ms->dictMatchState != NULL);
     switch(mls)
     {
@@ -408,16 +403,17 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
         const U32    repIndex = current + 1 - offset_1;
         const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;
         const BYTE* const repMatch = repBase + repIndex;
-        size_t mLength;
         hashTable[h] = current;   /* update hash table */
         assert(offset_1 <= current +1);   /* check repIndex */
 
         if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex))
            && (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;
+            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, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, 0, rLength-MINMATCH);
+            ip += rLength;
+            anchor = ip;
         } else {
             if ( (matchIndex < dictStartIndex) ||
                  (MEM_read32(match) != MEM_read32(ip)) ) {
@@ -427,19 +423,15 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
             }
             {   const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
                 const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
-                U32 offset;
-                mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
+                U32 const offset = current - matchIndex;
+                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 = current - matchIndex;
-                offset_2 = offset_1;
-                offset_1 = offset;
-                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                offset_2 = offset_1; offset_1 = offset;  /* update offset history */
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, iend, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ip += mLength;
+                anchor = ip;
         }   }
 
-        /* found a match : store it */
-        ip += mLength;
-        anchor = ip;
-
         if (ip <= ilimit) {
             /* Fill Table */
             hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2;
@@ -448,13 +440,13 @@ static size_t ZSTD_compressBlock_fast_extDict_generic(
             while (ip <= ilimit) {
                 U32 const current2 = (U32)(ip-base);
                 U32 const repIndex2 = current2 - offset_2;
-                const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
+                const BYTE* const repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
                 if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (repIndex2 > dictStartIndex))  /* 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 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
-                    ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
+                    { 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-MINMATCH);
                     hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
                     ip += repLength2;
                     anchor = ip;
@@ -476,8 +468,7 @@ size_t ZSTD_compressBlock_fast_extDict(
         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;
+    U32 const mls = ms->cParams.minMatch;
     switch(mls)
     {
     default: /* includes case 3 */

data/ext/zstdlib/{zstd-1.4.3 → zstd-1.4.4}/lib/compress/zstd_lazy.c

@@ -810,7 +810,7 @@ ZSTD_compressBlock_lazy_generic(
         /* store sequence */
 _storeSequence:
         {   size_t const litLength = start - anchor;
-            ZSTD_storeSeq(seqStore, litLength, anchor, (U32)offset, matchLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offset, matchLength-MINMATCH);
             anchor = ip = start + matchLength;
         }
 
@@ -828,7 +828,7 @@ _storeSequence:
                     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, 0, matchLength-MINMATCH);
+                    ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength-MINMATCH);
                     ip += matchLength;
                     anchor = ip;
                     continue;
@@ -843,7 +843,7 @@ _storeSequence:
                 /* 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, 0, matchLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength-MINMATCH);
                 ip += matchLength;
                 anchor = ip;
                 continue;   /* faster when present ... (?) */
@@ -1051,7 +1051,7 @@ size_t ZSTD_compressBlock_lazy_extDict_generic(
         /* store sequence */
 _storeSequence:
         {   size_t const litLength = start - anchor;
-            ZSTD_storeSeq(seqStore, litLength, anchor, (U32)offset, matchLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, litLength, anchor, iend, (U32)offset, matchLength-MINMATCH);
             anchor = ip = start + matchLength;
         }
 
@@ -1066,7 +1066,7 @@ _storeSequence:
                 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, 0, matchLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, 0, anchor, iend, 0, matchLength-MINMATCH);
                 ip += matchLength;
                 anchor = ip;
                 continue;   /* faster when present ... (?) */

data/ext/zstdlib/{zstd-1.4.3 → zstd-1.4.4}/lib/compress/zstd_ldm.c

@@ -49,9 +49,9 @@ size_t ZSTD_ldm_getTableSize(ldmParams_t params)
 {
     size_t const ldmHSize = ((size_t)1) << params.hashLog;
     size_t const ldmBucketSizeLog = MIN(params.bucketSizeLog, params.hashLog);
-    size_t const ldmBucketSize =
-        ((size_t)1) << (params.hashLog - ldmBucketSizeLog);
-    size_t const totalSize = ldmBucketSize + ldmHSize * sizeof(ldmEntry_t);
+    size_t const ldmBucketSize = ((size_t)1) << (params.hashLog - ldmBucketSizeLog);
+    size_t const totalSize = ZSTD_cwksp_alloc_size(ldmBucketSize)
+                           + ZSTD_cwksp_alloc_size(ldmHSize * sizeof(ldmEntry_t));
     return params.enableLdm ? totalSize : 0;
 }
 
@@ -583,7 +583,7 @@ size_t ZSTD_ldm_blockCompress(rawSeqStore_t* rawSeqStore,
                 rep[i] = rep[i-1];
             rep[0] = sequence.offset;
             /* Store the sequence */
-            ZSTD_storeSeq(seqStore, newLitLength, ip - newLitLength,
+            ZSTD_storeSeq(seqStore, newLitLength, ip - newLitLength, iend,
                           sequence.offset + ZSTD_REP_MOVE,
                           sequence.matchLength - MINMATCH);
             ip += sequence.matchLength;

data/ext/zstdlib/{zstd-1.4.3 → zstd-1.4.4}/lib/compress/zstd_opt.c

@@ -1098,7 +1098,7 @@ _shortestPath:   /* cur, last_pos, best_mlen, best_off have to be set */
 
                     assert(anchor + llen <= iend);
                     ZSTD_updateStats(optStatePtr, llen, anchor, offCode, mlen);
-                    ZSTD_storeSeq(seqStore, llen, anchor, offCode, mlen-MINMATCH);
+                    ZSTD_storeSeq(seqStore, llen, anchor, iend, offCode, mlen-MINMATCH);
                     anchor += advance;
                     ip = anchor;
             }   }

data/ext/zstdlib/{zstd-1.4.3 → zstd-1.4.4}/lib/compress/zstdmt_compress.c

@@ -668,7 +668,7 @@ static void ZSTDMT_compressionJob(void* jobDescription)
 
     /* init */
     if (job->cdict) {
-        size_t const initError = ZSTD_compressBegin_advanced_internal(cctx, NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast, job->cdict, jobParams, job->fullFrameSize);
+        size_t const initError = ZSTD_compressBegin_advanced_internal(cctx, NULL, 0, ZSTD_dct_auto, ZSTD_dtlm_fast, job->cdict, &jobParams, job->fullFrameSize);
         assert(job->firstJob);  /* only allowed for first job */
         if (ZSTD_isError(initError)) JOB_ERROR(initError);
     } else {  /* srcStart points at reloaded section */
@@ -680,7 +680,7 @@ static void ZSTDMT_compressionJob(void* jobDescription)
                                         job->prefix.start, job->prefix.size, ZSTD_dct_rawContent, /* load dictionary in "content-only" mode (no header analysis) */
                                         ZSTD_dtlm_fast,
                                         NULL, /*cdict*/
-                                        jobParams, pledgedSrcSize);
+                                        &jobParams, pledgedSrcSize);
             if (ZSTD_isError(initError)) JOB_ERROR(initError);
     }   }
 
@@ -927,12 +927,18 @@ static void ZSTDMT_releaseAllJobResources(ZSTDMT_CCtx* mtctx)
     unsigned jobID;
     DEBUGLOG(3, "ZSTDMT_releaseAllJobResources");
     for (jobID=0; jobID <= mtctx->jobIDMask; jobID++) {
+        /* Copy the mutex/cond out */
+        ZSTD_pthread_mutex_t const mutex = mtctx->jobs[jobID].job_mutex;
+        ZSTD_pthread_cond_t const cond = mtctx->jobs[jobID].job_cond;
+
         DEBUGLOG(4, "job%02u: release dst address %08X", jobID, (U32)(size_t)mtctx->jobs[jobID].dstBuff.start);
         ZSTDMT_releaseBuffer(mtctx->bufPool, mtctx->jobs[jobID].dstBuff);
-        mtctx->jobs[jobID].dstBuff = g_nullBuffer;
-        mtctx->jobs[jobID].cSize = 0;
+
+        /* Clear the job description, but keep the mutex/cond */
+        memset(&mtctx->jobs[jobID], 0, sizeof(mtctx->jobs[jobID]));
+        mtctx->jobs[jobID].job_mutex = mutex;
+        mtctx->jobs[jobID].job_cond = cond;
     }
-    memset(mtctx->jobs, 0, (mtctx->jobIDMask+1)*sizeof(ZSTDMT_jobDescription));
     mtctx->inBuff.buffer = g_nullBuffer;
     mtctx->inBuff.filled = 0;
     mtctx->allJobsCompleted = 1;
@@ -1028,9 +1034,9 @@ size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter,
 
 /* Sets parameters relevant to the compression job,
  * initializing others to default values. */
-static ZSTD_CCtx_params ZSTDMT_initJobCCtxParams(ZSTD_CCtx_params const params)
+static ZSTD_CCtx_params ZSTDMT_initJobCCtxParams(const ZSTD_CCtx_params* params)
 {
-    ZSTD_CCtx_params jobParams = params;
+    ZSTD_CCtx_params jobParams = *params;
     /* Clear parameters related to multithreading */
     jobParams.forceWindow = 0;
     jobParams.nbWorkers = 0;
@@ -1151,16 +1157,16 @@ size_t ZSTDMT_toFlushNow(ZSTDMT_CCtx* mtctx)
 /* =====   Multi-threaded compression   ===== */
 /* ------------------------------------------ */
 
-static unsigned ZSTDMT_computeTargetJobLog(ZSTD_CCtx_params const params)
+static unsigned ZSTDMT_computeTargetJobLog(const ZSTD_CCtx_params* params)
 {
     unsigned jobLog;
-    if (params.ldmParams.enableLdm) {
+    if (params->ldmParams.enableLdm) {
         /* In Long Range Mode, the windowLog is typically oversized.
          * In which case, it's preferable to determine the jobSize
          * based on chainLog instead. */
-        jobLog = MAX(21, params.cParams.chainLog + 4);
+        jobLog = MAX(21, params->cParams.chainLog + 4);
     } else {
-        jobLog = MAX(20, params.cParams.windowLog + 2);
+        jobLog = MAX(20, params->cParams.windowLog + 2);
     }
     return MIN(jobLog, (unsigned)ZSTDMT_JOBLOG_MAX);
 }
@@ -1193,27 +1199,27 @@ static int ZSTDMT_overlapLog(int ovlog, ZSTD_strategy strat)
     return ovlog;
 }
 
-static size_t ZSTDMT_computeOverlapSize(ZSTD_CCtx_params const params)
+static size_t ZSTDMT_computeOverlapSize(const ZSTD_CCtx_params* params)
 {
-    int const overlapRLog = 9 - ZSTDMT_overlapLog(params.overlapLog, params.cParams.strategy);
-    int ovLog = (overlapRLog >= 8) ? 0 : (params.cParams.windowLog - overlapRLog);
+    int const overlapRLog = 9 - ZSTDMT_overlapLog(params->overlapLog, params->cParams.strategy);
+    int ovLog = (overlapRLog >= 8) ? 0 : (params->cParams.windowLog - overlapRLog);
     assert(0 <= overlapRLog && overlapRLog <= 8);
-    if (params.ldmParams.enableLdm) {
+    if (params->ldmParams.enableLdm) {
         /* In Long Range Mode, the windowLog is typically oversized.
          * In which case, it's preferable to determine the jobSize
          * based on chainLog instead.
          * Then, ovLog becomes a fraction of the jobSize, rather than windowSize */
-        ovLog = MIN(params.cParams.windowLog, ZSTDMT_computeTargetJobLog(params) - 2)
+        ovLog = MIN(params->cParams.windowLog, ZSTDMT_computeTargetJobLog(params) - 2)
                 - overlapRLog;
     }
     assert(0 <= ovLog && ovLog <= ZSTD_WINDOWLOG_MAX);
-    DEBUGLOG(4, "overlapLog : %i", params.overlapLog);
+    DEBUGLOG(4, "overlapLog : %i", params->overlapLog);
     DEBUGLOG(4, "overlap size : %i", 1 << ovLog);
     return (ovLog==0) ? 0 : (size_t)1 << ovLog;
 }
 
 static unsigned
-ZSTDMT_computeNbJobs(ZSTD_CCtx_params params, size_t srcSize, unsigned nbWorkers)
+ZSTDMT_computeNbJobs(const ZSTD_CCtx_params* params, size_t srcSize, unsigned nbWorkers)
 {
     assert(nbWorkers>0);
     {   size_t const jobSizeTarget = (size_t)1 << ZSTDMT_computeTargetJobLog(params);
@@ -1236,9 +1242,9 @@ static size_t ZSTDMT_compress_advanced_internal(
           const ZSTD_CDict* cdict,
                 ZSTD_CCtx_params params)
 {
-    ZSTD_CCtx_params const jobParams = ZSTDMT_initJobCCtxParams(params);
-    size_t const overlapSize = ZSTDMT_computeOverlapSize(params);
-    unsigned const nbJobs = ZSTDMT_computeNbJobs(params, srcSize, params.nbWorkers);
+    ZSTD_CCtx_params const jobParams = ZSTDMT_initJobCCtxParams(&params);
+    size_t const overlapSize = ZSTDMT_computeOverlapSize(&params);
+    unsigned const nbJobs = ZSTDMT_computeNbJobs(&params, srcSize, params.nbWorkers);
     size_t const proposedJobSize = (srcSize + (nbJobs-1)) / nbJobs;
     size_t const avgJobSize = (((proposedJobSize-1) & 0x1FFFF) < 0x7FFF) ? proposedJobSize + 0xFFFF : proposedJobSize;   /* avoid too small last block */
     const char* const srcStart = (const char*)src;
@@ -1256,7 +1262,7 @@ static size_t ZSTDMT_compress_advanced_internal(
         ZSTD_CCtx* const cctx = mtctx->cctxPool->cctx[0];
         DEBUGLOG(4, "ZSTDMT_compress_advanced_internal: fallback to single-thread mode");
         if (cdict) return ZSTD_compress_usingCDict_advanced(cctx, dst, dstCapacity, src, srcSize, cdict, jobParams.fParams);
-        return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, NULL, 0, jobParams);
+        return ZSTD_compress_advanced_internal(cctx, dst, dstCapacity, src, srcSize, NULL, 0, &jobParams);
     }
 
     assert(avgJobSize >= 256 KB);  /* condition for ZSTD_compressBound(A) + ZSTD_compressBound(B) <= ZSTD_compressBound(A+B), required to compress directly into Dst (no additional buffer) */
@@ -1404,12 +1410,12 @@ size_t ZSTDMT_initCStream_internal(
 
     mtctx->singleBlockingThread = (pledgedSrcSize <= ZSTDMT_JOBSIZE_MIN);  /* do not trigger multi-threading when srcSize is too small */
     if (mtctx->singleBlockingThread) {
-        ZSTD_CCtx_params const singleThreadParams = ZSTDMT_initJobCCtxParams(params);
+        ZSTD_CCtx_params const singleThreadParams = ZSTDMT_initJobCCtxParams(&params);
         DEBUGLOG(5, "ZSTDMT_initCStream_internal: switch to single blocking thread mode");
         assert(singleThreadParams.nbWorkers == 0);
         return ZSTD_initCStream_internal(mtctx->cctxPool->cctx[0],
                                          dict, dictSize, cdict,
-                                         singleThreadParams, pledgedSrcSize);
+                                         &singleThreadParams, pledgedSrcSize);
     }
 
     DEBUGLOG(4, "ZSTDMT_initCStream_internal: %u workers", params.nbWorkers);
@@ -1435,11 +1441,11 @@ size_t ZSTDMT_initCStream_internal(
         mtctx->cdict = cdict;
     }
 
-    mtctx->targetPrefixSize = ZSTDMT_computeOverlapSize(params);
+    mtctx->targetPrefixSize = ZSTDMT_computeOverlapSize(&params);
     DEBUGLOG(4, "overlapLog=%i => %u KB", params.overlapLog, (U32)(mtctx->targetPrefixSize>>10));
     mtctx->targetSectionSize = params.jobSize;
     if (mtctx->targetSectionSize == 0) {
-        mtctx->targetSectionSize = 1ULL << ZSTDMT_computeTargetJobLog(params);
+        mtctx->targetSectionSize = 1ULL << ZSTDMT_computeTargetJobLog(&params);
     }
     assert(mtctx->targetSectionSize <= (size_t)ZSTDMT_JOBSIZE_MAX);