if (!tableLog) tableLog = FSE_DEFAULT_TABLELOG;
/* Scan input and build symbol stats */
- { CHECK_V_F(maxCount, HIST_count_wksp(count, &maxSymbolValue, src, srcSize, (unsigned*)scratchBuffer) );
+ { CHECK_V_F(maxCount, HIST_count_wksp(count, &maxSymbolValue, src, srcSize, scratchBuffer, scratchBufferSize) );
if (maxCount == srcSize) return 1; /* only a single symbol in src : rle */
if (maxCount == 1) return 0; /* each symbol present maximum once => not compressible */
if (maxCount < (srcSize >> 7)) return 0; /* Heuristic : not compressible enough */
return largestCount;
}
+typedef enum { trustInput, checkMaxSymbolValue } HIST_checkInput_e;
/* HIST_count_parallel_wksp() :
* store histogram into 4 intermediate tables, recombined at the end.
static size_t HIST_count_parallel_wksp(
unsigned* count, unsigned* maxSymbolValuePtr,
const void* source, size_t sourceSize,
- unsigned checkMax,
- unsigned* const workSpace)
+ HIST_checkInput_e check,
+ U32* const workSpace)
{
const BYTE* ip = (const BYTE*)source;
const BYTE* const iend = ip+sourceSize;
/* finish last symbols */
while (ip<iend) Counting1[*ip++]++;
- if (checkMax) { /* verify stats will fit into destination table */
+ if (check) { /* verify stats will fit into destination table */
U32 s; for (s=255; s>maxSymbolValue; s--) {
Counting1[s] += Counting2[s] + Counting3[s] + Counting4[s];
if (Counting1[s]) return ERROR(maxSymbolValue_tooSmall);
/* HIST_countFast_wksp() :
* Same as HIST_countFast(), but using an externally provided scratch buffer.
- * `workSpace` size must be table of >= HIST_WKSP_SIZE_U32 unsigned */
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
+ */
size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
const void* source, size_t sourceSize,
- unsigned* workSpace)
+ void* workSpace, size_t workSpaceSize)
{
if (sourceSize < 1500) /* heuristic threshold */
return HIST_count_simple(count, maxSymbolValuePtr, source, sourceSize);
- return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, 0, workSpace);
+ if ((size_t)workSpace & 3) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */
+ if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);
+ return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, trustInput, (U32*)workSpace);
}
/* fast variant (unsafe : won't check if src contains values beyond count[] limit) */
const void* source, size_t sourceSize)
{
unsigned tmpCounters[HIST_WKSP_SIZE_U32];
- return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, tmpCounters);
+ return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, tmpCounters, sizeof(tmpCounters));
}
/* HIST_count_wksp() :
* Same as HIST_count(), but using an externally provided scratch buffer.
* `workSpace` size must be table of >= HIST_WKSP_SIZE_U32 unsigned */
size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
- const void* source, size_t sourceSize, unsigned* workSpace)
+ const void* source, size_t sourceSize,
+ void* workSpace, size_t workSpaceSize)
{
+ if ((size_t)workSpace & 3) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */
+ if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);
if (*maxSymbolValuePtr < 255)
- return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, 1, workSpace);
+ return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, checkMaxSymbolValue, (U32*)workSpace);
*maxSymbolValuePtr = 255;
- return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace);
+ return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace, workSpaceSize);
}
size_t HIST_count(unsigned* count, unsigned* maxSymbolValuePtr,
const void* src, size_t srcSize)
{
unsigned tmpCounters[HIST_WKSP_SIZE_U32];
- return HIST_count_wksp(count, maxSymbolValuePtr, src, srcSize, tmpCounters);
+ return HIST_count_wksp(count, maxSymbolValuePtr, src, srcSize, tmpCounters, sizeof(tmpCounters));
}
/*! HIST_count():
* Provides the precise count of each byte within a table 'count'.
- * 'count' is a table of unsigned int, of minimum size (*maxSymbolValuePtr+1).
+ * 'count' is a table of unsigned int, of minimum size (*maxSymbolValuePtr+1).
* Updates *maxSymbolValuePtr with actual largest symbol value detected.
- * @return : count of the most frequent symbol (which isn't identified).
- * or an error code, which can be tested using HIST_isError().
- * note : if return == srcSize, there is only one symbol.
+ * @return : count of the most frequent symbol (which isn't identified).
+ * or an error code, which can be tested using HIST_isError().
+ * note : if return == srcSize, there is only one symbol.
*/
size_t HIST_count(unsigned* count, unsigned* maxSymbolValuePtr,
const void* src, size_t srcSize);
/* --- advanced histogram functions --- */
#define HIST_WKSP_SIZE_U32 1024
+#define HIST_WKSP_SIZE (HIST_WKSP_SIZE_U32 * sizeof(unsigned))
/** HIST_count_wksp() :
* Same as HIST_count(), but using an externally provided scratch buffer.
* Benefit is this function will use very little stack space.
- * `workSpace` must be a table of unsigned of size >= HIST_WKSP_SIZE_U32
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
*/
size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
const void* src, size_t srcSize,
- unsigned* workSpace);
+ void* workSpace, size_t workSpaceSize);
/** HIST_countFast() :
* same as HIST_count(), but blindly trusts that all byte values within src are <= *maxSymbolValuePtr.
/** HIST_countFast_wksp() :
* Same as HIST_countFast(), but using an externally provided scratch buffer.
- * `workSpace` must be a table of unsigned of size >= HIST_WKSP_SIZE_U32
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
*/
size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
const void* src, size_t srcSize,
- unsigned* workSpace);
+ void* workSpace, size_t workSpaceSize);
/*! HIST_count_simple() :
* Same as HIST_countFast(), this function is unsafe,
return HUF_compress4X_usingCTable_internal(dst, dstSize, src, srcSize, CTable, /* bmi2 */ 0);
}
+typedef enum { HUF_singleStream, HUF_fourStreams } HUF_nbStreams_e;
static size_t HUF_compressCTable_internal(
BYTE* const ostart, BYTE* op, BYTE* const oend,
const void* src, size_t srcSize,
- unsigned singleStream, const HUF_CElt* CTable, const int bmi2)
+ HUF_nbStreams_e nbStreams, const HUF_CElt* CTable, const int bmi2)
{
- size_t const cSize = singleStream ?
+ size_t const cSize = (nbStreams==HUF_singleStream) ?
HUF_compress1X_usingCTable_internal(op, oend - op, src, srcSize, CTable, bmi2) :
HUF_compress4X_usingCTable_internal(op, oend - op, src, srcSize, CTable, bmi2);
if (HUF_isError(cSize)) { return cSize; }
/* HUF_compress_internal() :
* `workSpace` must a table of at least HUF_WORKSPACE_SIZE_U32 unsigned */
-static size_t HUF_compress_internal (
- void* dst, size_t dstSize,
- const void* src, size_t srcSize,
- unsigned maxSymbolValue, unsigned huffLog,
- unsigned singleStream,
- void* workSpace, size_t wkspSize,
- HUF_CElt* oldHufTable, HUF_repeat* repeat, int preferRepeat,
- const int bmi2)
+static size_t
+HUF_compress_internal (void* dst, size_t dstSize,
+ const void* src, size_t srcSize,
+ unsigned maxSymbolValue, unsigned huffLog,
+ HUF_nbStreams_e nbStreams,
+ void* workSpace, size_t wkspSize,
+ HUF_CElt* oldHufTable, HUF_repeat* repeat, int preferRepeat,
+ const int bmi2)
{
HUF_compress_tables_t* const table = (HUF_compress_tables_t*)workSpace;
BYTE* const ostart = (BYTE*)dst;
/* checks & inits */
if (((size_t)workSpace & 3) != 0) return ERROR(GENERIC); /* must be aligned on 4-bytes boundaries */
- if (wkspSize < sizeof(*table)) return ERROR(workSpace_tooSmall);
+ if (wkspSize < HUF_WORKSPACE_SIZE) return ERROR(workSpace_tooSmall);
if (!srcSize) return 0; /* Uncompressed */
if (!dstSize) return 0; /* cannot fit anything within dst budget */
if (srcSize > HUF_BLOCKSIZE_MAX) return ERROR(srcSize_wrong); /* current block size limit */
if (preferRepeat && repeat && *repeat == HUF_repeat_valid) {
return HUF_compressCTable_internal(ostart, op, oend,
src, srcSize,
- singleStream, oldHufTable, bmi2);
+ nbStreams, oldHufTable, bmi2);
}
/* Scan input and build symbol stats */
- { CHECK_V_F(largest, HIST_count_wksp (table->count, &maxSymbolValue, (const BYTE*)src, srcSize, table->count) );
+ { CHECK_V_F(largest, HIST_count_wksp (table->count, &maxSymbolValue, (const BYTE*)src, srcSize, workSpace, wkspSize) );
if (largest == srcSize) { *ostart = ((const BYTE*)src)[0]; return 1; } /* single symbol, rle */
if (largest <= (srcSize >> 7)+4) return 0; /* heuristic : probably not compressible enough */
}
if (preferRepeat && repeat && *repeat != HUF_repeat_none) {
return HUF_compressCTable_internal(ostart, op, oend,
src, srcSize,
- singleStream, oldHufTable, bmi2);
+ nbStreams, oldHufTable, bmi2);
}
/* Build Huffman Tree */
if (oldSize <= hSize + newSize || hSize + 12 >= srcSize) {
return HUF_compressCTable_internal(ostart, op, oend,
src, srcSize,
- singleStream, oldHufTable, bmi2);
+ nbStreams, oldHufTable, bmi2);
} }
/* Use the new huffman table */
}
return HUF_compressCTable_internal(ostart, op, oend,
src, srcSize,
- singleStream, table->CTable, bmi2);
+ nbStreams, table->CTable, bmi2);
}
void* workSpace, size_t wkspSize)
{
return HUF_compress_internal(dst, dstSize, src, srcSize,
- maxSymbolValue, huffLog, 1 /*single stream*/,
+ maxSymbolValue, huffLog, HUF_singleStream,
workSpace, wkspSize,
NULL, NULL, 0, 0 /*bmi2*/);
}
HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2)
{
return HUF_compress_internal(dst, dstSize, src, srcSize,
- maxSymbolValue, huffLog, 1 /*single stream*/,
+ maxSymbolValue, huffLog, HUF_singleStream,
workSpace, wkspSize, hufTable,
repeat, preferRepeat, bmi2);
}
void* workSpace, size_t wkspSize)
{
return HUF_compress_internal(dst, dstSize, src, srcSize,
- maxSymbolValue, huffLog, 0 /*4 streams*/,
+ maxSymbolValue, huffLog, HUF_fourStreams,
workSpace, wkspSize,
NULL, NULL, 0, 0 /*bmi2*/);
}
HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2)
{
return HUF_compress_internal(dst, dstSize, src, srcSize,
- maxSymbolValue, huffLog, 0 /* 4 streams */,
+ maxSymbolValue, huffLog, HUF_fourStreams,
workSpace, wkspSize,
hufTable, repeat, preferRepeat, bmi2);
}
ZSTD_strategy strategy, int disableLiteralCompression,
void* dst, size_t dstCapacity,
const void* src, size_t srcSize,
- U32* workspace, const int bmi2)
+ void* workspace, size_t wkspSize,
+ const int bmi2)
{
size_t const minGain = ZSTD_minGain(srcSize, strategy);
size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
int const preferRepeat = strategy < ZSTD_lazy ? srcSize <= 1024 : 0;
if (repeat == HUF_repeat_valid && lhSize == 3) singleStream = 1;
cLitSize = singleStream ? HUF_compress1X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11,
- workspace, HUF_WORKSPACE_SIZE, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2)
+ workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2)
: HUF_compress4X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11,
- workspace, HUF_WORKSPACE_SIZE, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2);
+ workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2);
if (repeat != HUF_repeat_none) {
/* reused the existing table */
hType = set_repeat;
sequences, nbSeq, longOffsets);
}
-MEM_STATIC size_t ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,
+MEM_STATIC size_t
+ZSTD_compressSequences_internal(seqStore_t* seqStorePtr,
ZSTD_entropyCTables_t const* prevEntropy,
ZSTD_entropyCTables_t* nextEntropy,
ZSTD_CCtx_params const* cctxParams,
- void* dst, size_t dstCapacity, U32* workspace,
+ void* dst, size_t dstCapacity,
+ void* workspace, size_t wkspSize,
const int bmi2)
{
const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN;
cctxParams->cParams.strategy, disableLiteralCompression,
op, dstCapacity,
literals, litSize,
- workspace, bmi2);
+ workspace, wkspSize,
+ bmi2);
if (ZSTD_isError(cSize))
return cSize;
assert(cSize <= dstCapacity);
ZSTD_seqToCodes(seqStorePtr);
/* build CTable for Literal Lengths */
{ U32 max = MaxLL;
- size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace); /* can't fail */
+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace, wkspSize); /* can't fail */
DEBUGLOG(5, "Building LL table");
nextEntropy->fse.litlength_repeatMode = prevEntropy->fse.litlength_repeatMode;
LLtype = ZSTD_selectEncodingType(&nextEntropy->fse.litlength_repeatMode, count, max, mostFrequent, nbSeq, LLFSELog, prevEntropy->fse.litlengthCTable, LL_defaultNorm, LL_defaultNormLog, ZSTD_defaultAllowed, strategy);
{ size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL,
prevEntropy->fse.litlengthCTable, sizeof(prevEntropy->fse.litlengthCTable),
- workspace, HUF_WORKSPACE_SIZE);
+ workspace, wkspSize);
if (ZSTD_isError(countSize)) return countSize;
if (LLtype == set_compressed)
lastNCount = op;
} }
/* build CTable for Offsets */
{ U32 max = MaxOff;
- size_t const mostFrequent = HIST_countFast_wksp(count, &max, ofCodeTable, nbSeq, workspace); /* can't fail */
+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, ofCodeTable, nbSeq, workspace, wkspSize); /* can't fail */
/* We can only use the basic table if max <= DefaultMaxOff, otherwise the offsets are too large */
ZSTD_defaultPolicy_e const defaultPolicy = (max <= DefaultMaxOff) ? ZSTD_defaultAllowed : ZSTD_defaultDisallowed;
DEBUGLOG(5, "Building OF table");
{ size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
prevEntropy->fse.offcodeCTable, sizeof(prevEntropy->fse.offcodeCTable),
- workspace, HUF_WORKSPACE_SIZE);
+ workspace, wkspSize);
if (ZSTD_isError(countSize)) return countSize;
if (Offtype == set_compressed)
lastNCount = op;
} }
/* build CTable for MatchLengths */
{ U32 max = MaxML;
- size_t const mostFrequent = HIST_countFast_wksp(count, &max, mlCodeTable, nbSeq, workspace); /* can't fail */
+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, mlCodeTable, nbSeq, workspace, wkspSize); /* can't fail */
DEBUGLOG(5, "Building ML table");
nextEntropy->fse.matchlength_repeatMode = prevEntropy->fse.matchlength_repeatMode;
MLtype = ZSTD_selectEncodingType(&nextEntropy->fse.matchlength_repeatMode, count, max, mostFrequent, nbSeq, MLFSELog, prevEntropy->fse.matchlengthCTable, ML_defaultNorm, ML_defaultNormLog, ZSTD_defaultAllowed, strategy);
{ size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML,
prevEntropy->fse.matchlengthCTable, sizeof(prevEntropy->fse.matchlengthCTable),
- workspace, HUF_WORKSPACE_SIZE);
+ workspace, wkspSize);
if (ZSTD_isError(countSize)) return countSize;
if (MLtype == set_compressed)
lastNCount = op;
return op - ostart;
}
-MEM_STATIC size_t ZSTD_compressSequences(seqStore_t* seqStorePtr,
- const ZSTD_entropyCTables_t* prevEntropy,
- ZSTD_entropyCTables_t* nextEntropy,
- const ZSTD_CCtx_params* cctxParams,
- void* dst, size_t dstCapacity,
- size_t srcSize, U32* workspace, int bmi2)
+MEM_STATIC size_t
+ZSTD_compressSequences(seqStore_t* seqStorePtr,
+ const ZSTD_entropyCTables_t* prevEntropy,
+ ZSTD_entropyCTables_t* nextEntropy,
+ const ZSTD_CCtx_params* cctxParams,
+ void* dst, size_t dstCapacity,
+ size_t srcSize,
+ void* workspace, size_t wkspSize,
+ int bmi2)
{
size_t const cSize = ZSTD_compressSequences_internal(
- seqStorePtr, prevEntropy, nextEntropy, cctxParams, dst, dstCapacity,
- workspace, bmi2);
+ seqStorePtr, prevEntropy, nextEntropy, cctxParams,
+ dst, dstCapacity,
+ workspace, wkspSize, bmi2);
if (cSize == 0) return 0;
/* When srcSize <= dstCapacity, there is enough space to write a raw uncompressed block.
* Since we ran out of space, block must be not compressible, so fall back to raw uncompressed block.
&zc->blockState.prevCBlock->entropy, &zc->blockState.nextCBlock->entropy,
&zc->appliedParams,
dst, dstCapacity,
- srcSize, zc->entropyWorkspace, zc->bmi2);
+ srcSize,
+ zc->entropyWorkspace, HUF_WORKSPACE_SIZE /* statically allocated in resetCCtx */,
+ zc->bmi2);
out:
if (!ZSTD_isError(cSize) && cSize != 0) {
U32 count[HUF_SYMBOLVALUE_MAX+1];
/* Scan input and build symbol stats */
- { size_t const largest = HIST_count_wksp (count, &maxSymbolValue, (const BYTE*)src, srcSize, WKSP);
+ { size_t const largest = HIST_count_wksp (count, &maxSymbolValue, (const BYTE*)src, srcSize, WKSP, sizeof(WKSP));
assert(!HIST_isError(largest));
if (largest == srcSize) { *ostart = ((const BYTE*)src)[0]; return 0; } /* single symbol, rle */
if (largest <= (srcSize >> 7)+1) return 0; /* Fast heuristic : not compressible enough */
/* CTable for Literal Lengths */
{ U32 max = MaxLL;
- size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, WKSP); /* cannot fail */
+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, llCodeTable, nbSeq, WKSP, sizeof(WKSP)); /* cannot fail */
assert(!HIST_isError(mostFrequent));
if (mostFrequent == nbSeq) {
/* do RLE if we have the chance */
/* CTable for Offsets */
/* see Literal Lengths for descriptions of mode choices */
{ U32 max = MaxOff;
- size_t const mostFrequent = HIST_countFast_wksp(count, &max, ofCodeTable, nbSeq, WKSP); /* cannot fail */
+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, ofCodeTable, nbSeq, WKSP, sizeof(WKSP)); /* cannot fail */
assert(!HIST_isError(mostFrequent));
if (mostFrequent == nbSeq) {
*op++ = ofCodeTable[0];
/* CTable for MatchLengths */
/* see Literal Lengths for descriptions of mode choices */
{ U32 max = MaxML;
- size_t const mostFrequent = HIST_countFast_wksp(count, &max, mlCodeTable, nbSeq, WKSP); /* cannot fail */
+ size_t const mostFrequent = HIST_countFast_wksp(count, &max, mlCodeTable, nbSeq, WKSP, sizeof(WKSP)); /* cannot fail */
assert(!HIST_isError(mostFrequent));
if (mostFrequent == nbSeq) {
*op++ = *mlCodeTable;
projects / thirdparty / zstd.git / commitdiff
