/src/zstd/lib/compress/zstd_compress_superblock.c
Line | Count | Source (jump to first uncovered line) |
1 | | /* |
2 | | * Copyright (c) Meta Platforms, Inc. and affiliates. |
3 | | * All rights reserved. |
4 | | * |
5 | | * This source code is licensed under both the BSD-style license (found in the |
6 | | * LICENSE file in the root directory of this source tree) and the GPLv2 (found |
7 | | * in the COPYING file in the root directory of this source tree). |
8 | | * You may select, at your option, one of the above-listed licenses. |
9 | | */ |
10 | | |
11 | | /*-************************************* |
12 | | * Dependencies |
13 | | ***************************************/ |
14 | | #include "zstd_compress_superblock.h" |
15 | | |
16 | | #include "../common/zstd_internal.h" /* ZSTD_getSequenceLength */ |
17 | | #include "hist.h" /* HIST_countFast_wksp */ |
18 | | #include "zstd_compress_internal.h" /* ZSTD_[huf|fse|entropy]CTablesMetadata_t */ |
19 | | #include "zstd_compress_sequences.h" |
20 | | #include "zstd_compress_literals.h" |
21 | | |
22 | | /** ZSTD_compressSubBlock_literal() : |
23 | | * Compresses literals section for a sub-block. |
24 | | * When we have to write the Huffman table we will sometimes choose a header |
25 | | * size larger than necessary. This is because we have to pick the header size |
26 | | * before we know the table size + compressed size, so we have a bound on the |
27 | | * table size. If we guessed incorrectly, we fall back to uncompressed literals. |
28 | | * |
29 | | * We write the header when writeEntropy=1 and set entropyWritten=1 when we succeeded |
30 | | * in writing the header, otherwise it is set to 0. |
31 | | * |
32 | | * hufMetadata->hType has literals block type info. |
33 | | * If it is set_basic, all sub-blocks literals section will be Raw_Literals_Block. |
34 | | * If it is set_rle, all sub-blocks literals section will be RLE_Literals_Block. |
35 | | * If it is set_compressed, first sub-block's literals section will be Compressed_Literals_Block |
36 | | * If it is set_compressed, first sub-block's literals section will be Treeless_Literals_Block |
37 | | * and the following sub-blocks' literals sections will be Treeless_Literals_Block. |
38 | | * @return : compressed size of literals section of a sub-block |
39 | | * Or 0 if unable to compress. |
40 | | * Or error code */ |
41 | | static size_t |
42 | | ZSTD_compressSubBlock_literal(const HUF_CElt* hufTable, |
43 | | const ZSTD_hufCTablesMetadata_t* hufMetadata, |
44 | | const BYTE* literals, size_t litSize, |
45 | | void* dst, size_t dstSize, |
46 | | const int bmi2, int writeEntropy, int* entropyWritten) |
47 | 0 | { |
48 | 0 | size_t const header = writeEntropy ? 200 : 0; |
49 | 0 | size_t const lhSize = 3 + (litSize >= (1 KB - header)) + (litSize >= (16 KB - header)); |
50 | 0 | BYTE* const ostart = (BYTE*)dst; |
51 | 0 | BYTE* const oend = ostart + dstSize; |
52 | 0 | BYTE* op = ostart + lhSize; |
53 | 0 | U32 const singleStream = lhSize == 3; |
54 | 0 | SymbolEncodingType_e hType = writeEntropy ? hufMetadata->hType : set_repeat; |
55 | 0 | size_t cLitSize = 0; |
56 | |
|
57 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_literal (litSize=%zu, lhSize=%zu, writeEntropy=%d)", litSize, lhSize, writeEntropy); |
58 | |
|
59 | 0 | *entropyWritten = 0; |
60 | 0 | if (litSize == 0 || hufMetadata->hType == set_basic) { |
61 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_literal using raw literal"); |
62 | 0 | return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize); |
63 | 0 | } else if (hufMetadata->hType == set_rle) { |
64 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_literal using rle literal"); |
65 | 0 | return ZSTD_compressRleLiteralsBlock(dst, dstSize, literals, litSize); |
66 | 0 | } |
67 | | |
68 | 0 | assert(litSize > 0); |
69 | 0 | assert(hufMetadata->hType == set_compressed || hufMetadata->hType == set_repeat); |
70 | |
|
71 | 0 | if (writeEntropy && hufMetadata->hType == set_compressed) { |
72 | 0 | ZSTD_memcpy(op, hufMetadata->hufDesBuffer, hufMetadata->hufDesSize); |
73 | 0 | op += hufMetadata->hufDesSize; |
74 | 0 | cLitSize += hufMetadata->hufDesSize; |
75 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_literal (hSize=%zu)", hufMetadata->hufDesSize); |
76 | 0 | } |
77 | |
|
78 | 0 | { int const flags = bmi2 ? HUF_flags_bmi2 : 0; |
79 | 0 | const size_t cSize = singleStream ? HUF_compress1X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags) |
80 | 0 | : HUF_compress4X_usingCTable(op, (size_t)(oend-op), literals, litSize, hufTable, flags); |
81 | 0 | op += cSize; |
82 | 0 | cLitSize += cSize; |
83 | 0 | if (cSize == 0 || ERR_isError(cSize)) { |
84 | 0 | DEBUGLOG(5, "Failed to write entropy tables %s", ZSTD_getErrorName(cSize)); |
85 | 0 | return 0; |
86 | 0 | } |
87 | | /* If we expand and we aren't writing a header then emit uncompressed */ |
88 | 0 | if (!writeEntropy && cLitSize >= litSize) { |
89 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_literal using raw literal because uncompressible"); |
90 | 0 | return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize); |
91 | 0 | } |
92 | | /* If we are writing headers then allow expansion that doesn't change our header size. */ |
93 | 0 | if (lhSize < (size_t)(3 + (cLitSize >= 1 KB) + (cLitSize >= 16 KB))) { |
94 | 0 | assert(cLitSize > litSize); |
95 | 0 | DEBUGLOG(5, "Literals expanded beyond allowed header size"); |
96 | 0 | return ZSTD_noCompressLiterals(dst, dstSize, literals, litSize); |
97 | 0 | } |
98 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_literal (cSize=%zu)", cSize); |
99 | 0 | } |
100 | | |
101 | | /* Build header */ |
102 | 0 | switch(lhSize) |
103 | 0 | { |
104 | 0 | case 3: /* 2 - 2 - 10 - 10 */ |
105 | 0 | { U32 const lhc = hType + ((U32)(!singleStream) << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<14); |
106 | 0 | MEM_writeLE24(ostart, lhc); |
107 | 0 | break; |
108 | 0 | } |
109 | 0 | case 4: /* 2 - 2 - 14 - 14 */ |
110 | 0 | { U32 const lhc = hType + (2 << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<18); |
111 | 0 | MEM_writeLE32(ostart, lhc); |
112 | 0 | break; |
113 | 0 | } |
114 | 0 | case 5: /* 2 - 2 - 18 - 18 */ |
115 | 0 | { U32 const lhc = hType + (3 << 2) + ((U32)litSize<<4) + ((U32)cLitSize<<22); |
116 | 0 | MEM_writeLE32(ostart, lhc); |
117 | 0 | ostart[4] = (BYTE)(cLitSize >> 10); |
118 | 0 | break; |
119 | 0 | } |
120 | 0 | default: /* not possible : lhSize is {3,4,5} */ |
121 | 0 | assert(0); |
122 | 0 | } |
123 | 0 | *entropyWritten = 1; |
124 | 0 | DEBUGLOG(5, "Compressed literals: %u -> %u", (U32)litSize, (U32)(op-ostart)); |
125 | 0 | return (size_t)(op-ostart); |
126 | 0 | } |
127 | | |
128 | | static size_t |
129 | | ZSTD_seqDecompressedSize(SeqStore_t const* seqStore, |
130 | | const SeqDef* sequences, size_t nbSeqs, |
131 | | size_t litSize, int lastSubBlock) |
132 | 0 | { |
133 | 0 | size_t matchLengthSum = 0; |
134 | 0 | size_t litLengthSum = 0; |
135 | 0 | size_t n; |
136 | 0 | for (n=0; n<nbSeqs; n++) { |
137 | 0 | const ZSTD_SequenceLength seqLen = ZSTD_getSequenceLength(seqStore, sequences+n); |
138 | 0 | litLengthSum += seqLen.litLength; |
139 | 0 | matchLengthSum += seqLen.matchLength; |
140 | 0 | } |
141 | 0 | DEBUGLOG(5, "ZSTD_seqDecompressedSize: %u sequences from %p: %u literals + %u matchlength", |
142 | 0 | (unsigned)nbSeqs, (const void*)sequences, |
143 | 0 | (unsigned)litLengthSum, (unsigned)matchLengthSum); |
144 | 0 | if (!lastSubBlock) |
145 | 0 | assert(litLengthSum == litSize); |
146 | 0 | else |
147 | 0 | assert(litLengthSum <= litSize); |
148 | 0 | (void)litLengthSum; |
149 | 0 | return matchLengthSum + litSize; |
150 | 0 | } |
151 | | |
152 | | /** ZSTD_compressSubBlock_sequences() : |
153 | | * Compresses sequences section for a sub-block. |
154 | | * fseMetadata->llType, fseMetadata->ofType, and fseMetadata->mlType have |
155 | | * symbol compression modes for the super-block. |
156 | | * The first successfully compressed block will have these in its header. |
157 | | * We set entropyWritten=1 when we succeed in compressing the sequences. |
158 | | * The following sub-blocks will always have repeat mode. |
159 | | * @return : compressed size of sequences section of a sub-block |
160 | | * Or 0 if it is unable to compress |
161 | | * Or error code. */ |
162 | | static size_t |
163 | | ZSTD_compressSubBlock_sequences(const ZSTD_fseCTables_t* fseTables, |
164 | | const ZSTD_fseCTablesMetadata_t* fseMetadata, |
165 | | const SeqDef* sequences, size_t nbSeq, |
166 | | const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode, |
167 | | const ZSTD_CCtx_params* cctxParams, |
168 | | void* dst, size_t dstCapacity, |
169 | | const int bmi2, int writeEntropy, int* entropyWritten) |
170 | 0 | { |
171 | 0 | const int longOffsets = cctxParams->cParams.windowLog > STREAM_ACCUMULATOR_MIN; |
172 | 0 | BYTE* const ostart = (BYTE*)dst; |
173 | 0 | BYTE* const oend = ostart + dstCapacity; |
174 | 0 | BYTE* op = ostart; |
175 | 0 | BYTE* seqHead; |
176 | |
|
177 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (nbSeq=%zu, writeEntropy=%d, longOffsets=%d)", nbSeq, writeEntropy, longOffsets); |
178 | |
|
179 | 0 | *entropyWritten = 0; |
180 | | /* Sequences Header */ |
181 | 0 | RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/, |
182 | 0 | dstSize_tooSmall, ""); |
183 | 0 | if (nbSeq < 128) |
184 | 0 | *op++ = (BYTE)nbSeq; |
185 | 0 | else if (nbSeq < LONGNBSEQ) |
186 | 0 | op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2; |
187 | 0 | else |
188 | 0 | op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3; |
189 | 0 | if (nbSeq==0) { |
190 | 0 | return (size_t)(op - ostart); |
191 | 0 | } |
192 | | |
193 | | /* seqHead : flags for FSE encoding type */ |
194 | 0 | seqHead = op++; |
195 | |
|
196 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (seqHeadSize=%u)", (unsigned)(op-ostart)); |
197 | |
|
198 | 0 | if (writeEntropy) { |
199 | 0 | const U32 LLtype = fseMetadata->llType; |
200 | 0 | const U32 Offtype = fseMetadata->ofType; |
201 | 0 | const U32 MLtype = fseMetadata->mlType; |
202 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (fseTablesSize=%zu)", fseMetadata->fseTablesSize); |
203 | 0 | *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2)); |
204 | 0 | ZSTD_memcpy(op, fseMetadata->fseTablesBuffer, fseMetadata->fseTablesSize); |
205 | 0 | op += fseMetadata->fseTablesSize; |
206 | 0 | } else { |
207 | 0 | const U32 repeat = set_repeat; |
208 | 0 | *seqHead = (BYTE)((repeat<<6) + (repeat<<4) + (repeat<<2)); |
209 | 0 | } |
210 | |
|
211 | 0 | { size_t const bitstreamSize = ZSTD_encodeSequences( |
212 | 0 | op, (size_t)(oend - op), |
213 | 0 | fseTables->matchlengthCTable, mlCode, |
214 | 0 | fseTables->offcodeCTable, ofCode, |
215 | 0 | fseTables->litlengthCTable, llCode, |
216 | 0 | sequences, nbSeq, |
217 | 0 | longOffsets, bmi2); |
218 | 0 | FORWARD_IF_ERROR(bitstreamSize, "ZSTD_encodeSequences failed"); |
219 | 0 | op += bitstreamSize; |
220 | | /* zstd versions <= 1.3.4 mistakenly report corruption when |
221 | | * FSE_readNCount() receives a buffer < 4 bytes. |
222 | | * Fixed by https://github.com/facebook/zstd/pull/1146. |
223 | | * This can happen when the last set_compressed table present is 2 |
224 | | * bytes and the bitstream is only one byte. |
225 | | * In this exceedingly rare case, we will simply emit an uncompressed |
226 | | * block, since it isn't worth optimizing. |
227 | | */ |
228 | | #ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION |
229 | | if (writeEntropy && fseMetadata->lastCountSize && fseMetadata->lastCountSize + bitstreamSize < 4) { |
230 | | /* NCountSize >= 2 && bitstreamSize > 0 ==> lastCountSize == 3 */ |
231 | | assert(fseMetadata->lastCountSize + bitstreamSize == 3); |
232 | | DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.3.4 by " |
233 | | "emitting an uncompressed block."); |
234 | | return 0; |
235 | | } |
236 | | #endif |
237 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_sequences (bitstreamSize=%zu)", bitstreamSize); |
238 | 0 | } |
239 | | |
240 | | /* zstd versions <= 1.4.0 mistakenly report error when |
241 | | * sequences section body size is less than 3 bytes. |
242 | | * Fixed by https://github.com/facebook/zstd/pull/1664. |
243 | | * This can happen when the previous sequences section block is compressed |
244 | | * with rle mode and the current block's sequences section is compressed |
245 | | * with repeat mode where sequences section body size can be 1 byte. |
246 | | */ |
247 | | #ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION |
248 | | if (op-seqHead < 4) { |
249 | | DEBUGLOG(5, "Avoiding bug in zstd decoder in versions <= 1.4.0 by emitting " |
250 | | "an uncompressed block when sequences are < 4 bytes"); |
251 | | return 0; |
252 | | } |
253 | | #endif |
254 | | |
255 | 0 | *entropyWritten = 1; |
256 | 0 | return (size_t)(op - ostart); |
257 | 0 | } |
258 | | |
259 | | /** ZSTD_compressSubBlock() : |
260 | | * Compresses a single sub-block. |
261 | | * @return : compressed size of the sub-block |
262 | | * Or 0 if it failed to compress. */ |
263 | | static size_t ZSTD_compressSubBlock(const ZSTD_entropyCTables_t* entropy, |
264 | | const ZSTD_entropyCTablesMetadata_t* entropyMetadata, |
265 | | const SeqDef* sequences, size_t nbSeq, |
266 | | const BYTE* literals, size_t litSize, |
267 | | const BYTE* llCode, const BYTE* mlCode, const BYTE* ofCode, |
268 | | const ZSTD_CCtx_params* cctxParams, |
269 | | void* dst, size_t dstCapacity, |
270 | | const int bmi2, |
271 | | int writeLitEntropy, int writeSeqEntropy, |
272 | | int* litEntropyWritten, int* seqEntropyWritten, |
273 | | U32 lastBlock) |
274 | 0 | { |
275 | 0 | BYTE* const ostart = (BYTE*)dst; |
276 | 0 | BYTE* const oend = ostart + dstCapacity; |
277 | 0 | BYTE* op = ostart + ZSTD_blockHeaderSize; |
278 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock (litSize=%zu, nbSeq=%zu, writeLitEntropy=%d, writeSeqEntropy=%d, lastBlock=%d)", |
279 | 0 | litSize, nbSeq, writeLitEntropy, writeSeqEntropy, lastBlock); |
280 | 0 | { size_t cLitSize = ZSTD_compressSubBlock_literal((const HUF_CElt*)entropy->huf.CTable, |
281 | 0 | &entropyMetadata->hufMetadata, literals, litSize, |
282 | 0 | op, (size_t)(oend-op), |
283 | 0 | bmi2, writeLitEntropy, litEntropyWritten); |
284 | 0 | FORWARD_IF_ERROR(cLitSize, "ZSTD_compressSubBlock_literal failed"); |
285 | 0 | if (cLitSize == 0) return 0; |
286 | 0 | op += cLitSize; |
287 | 0 | } |
288 | 0 | { size_t cSeqSize = ZSTD_compressSubBlock_sequences(&entropy->fse, |
289 | 0 | &entropyMetadata->fseMetadata, |
290 | 0 | sequences, nbSeq, |
291 | 0 | llCode, mlCode, ofCode, |
292 | 0 | cctxParams, |
293 | 0 | op, (size_t)(oend-op), |
294 | 0 | bmi2, writeSeqEntropy, seqEntropyWritten); |
295 | 0 | FORWARD_IF_ERROR(cSeqSize, "ZSTD_compressSubBlock_sequences failed"); |
296 | 0 | if (cSeqSize == 0) return 0; |
297 | 0 | op += cSeqSize; |
298 | 0 | } |
299 | | /* Write block header */ |
300 | 0 | { size_t cSize = (size_t)(op-ostart) - ZSTD_blockHeaderSize; |
301 | 0 | U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3); |
302 | 0 | MEM_writeLE24(ostart, cBlockHeader24); |
303 | 0 | } |
304 | 0 | return (size_t)(op-ostart); |
305 | 0 | } |
306 | | |
307 | | static size_t ZSTD_estimateSubBlockSize_literal(const BYTE* literals, size_t litSize, |
308 | | const ZSTD_hufCTables_t* huf, |
309 | | const ZSTD_hufCTablesMetadata_t* hufMetadata, |
310 | | void* workspace, size_t wkspSize, |
311 | | int writeEntropy) |
312 | 0 | { |
313 | 0 | unsigned* const countWksp = (unsigned*)workspace; |
314 | 0 | unsigned maxSymbolValue = 255; |
315 | 0 | size_t literalSectionHeaderSize = 3; /* Use hard coded size of 3 bytes */ |
316 | |
|
317 | 0 | if (hufMetadata->hType == set_basic) return litSize; |
318 | 0 | else if (hufMetadata->hType == set_rle) return 1; |
319 | 0 | else if (hufMetadata->hType == set_compressed || hufMetadata->hType == set_repeat) { |
320 | 0 | size_t const largest = HIST_count_wksp (countWksp, &maxSymbolValue, (const BYTE*)literals, litSize, workspace, wkspSize); |
321 | 0 | if (ZSTD_isError(largest)) return litSize; |
322 | 0 | { size_t cLitSizeEstimate = HUF_estimateCompressedSize((const HUF_CElt*)huf->CTable, countWksp, maxSymbolValue); |
323 | 0 | if (writeEntropy) cLitSizeEstimate += hufMetadata->hufDesSize; |
324 | 0 | return cLitSizeEstimate + literalSectionHeaderSize; |
325 | 0 | } } |
326 | 0 | assert(0); /* impossible */ |
327 | 0 | return 0; |
328 | 0 | } |
329 | | |
330 | | static size_t ZSTD_estimateSubBlockSize_symbolType(SymbolEncodingType_e type, |
331 | | const BYTE* codeTable, unsigned maxCode, |
332 | | size_t nbSeq, const FSE_CTable* fseCTable, |
333 | | const U8* additionalBits, |
334 | | short const* defaultNorm, U32 defaultNormLog, U32 defaultMax, |
335 | | void* workspace, size_t wkspSize) |
336 | 0 | { |
337 | 0 | unsigned* const countWksp = (unsigned*)workspace; |
338 | 0 | const BYTE* ctp = codeTable; |
339 | 0 | const BYTE* const ctStart = ctp; |
340 | 0 | const BYTE* const ctEnd = ctStart + nbSeq; |
341 | 0 | size_t cSymbolTypeSizeEstimateInBits = 0; |
342 | 0 | unsigned max = maxCode; |
343 | |
|
344 | 0 | HIST_countFast_wksp(countWksp, &max, codeTable, nbSeq, workspace, wkspSize); /* can't fail */ |
345 | 0 | if (type == set_basic) { |
346 | | /* We selected this encoding type, so it must be valid. */ |
347 | 0 | assert(max <= defaultMax); |
348 | 0 | cSymbolTypeSizeEstimateInBits = max <= defaultMax |
349 | 0 | ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, countWksp, max) |
350 | 0 | : ERROR(GENERIC); |
351 | 0 | } else if (type == set_rle) { |
352 | 0 | cSymbolTypeSizeEstimateInBits = 0; |
353 | 0 | } else if (type == set_compressed || type == set_repeat) { |
354 | 0 | cSymbolTypeSizeEstimateInBits = ZSTD_fseBitCost(fseCTable, countWksp, max); |
355 | 0 | } |
356 | 0 | if (ZSTD_isError(cSymbolTypeSizeEstimateInBits)) return nbSeq * 10; |
357 | 0 | while (ctp < ctEnd) { |
358 | 0 | if (additionalBits) cSymbolTypeSizeEstimateInBits += additionalBits[*ctp]; |
359 | 0 | else cSymbolTypeSizeEstimateInBits += *ctp; /* for offset, offset code is also the number of additional bits */ |
360 | 0 | ctp++; |
361 | 0 | } |
362 | 0 | return cSymbolTypeSizeEstimateInBits / 8; |
363 | 0 | } |
364 | | |
365 | | static size_t ZSTD_estimateSubBlockSize_sequences(const BYTE* ofCodeTable, |
366 | | const BYTE* llCodeTable, |
367 | | const BYTE* mlCodeTable, |
368 | | size_t nbSeq, |
369 | | const ZSTD_fseCTables_t* fseTables, |
370 | | const ZSTD_fseCTablesMetadata_t* fseMetadata, |
371 | | void* workspace, size_t wkspSize, |
372 | | int writeEntropy) |
373 | 0 | { |
374 | 0 | size_t const sequencesSectionHeaderSize = 3; /* Use hard coded size of 3 bytes */ |
375 | 0 | size_t cSeqSizeEstimate = 0; |
376 | 0 | if (nbSeq == 0) return sequencesSectionHeaderSize; |
377 | 0 | cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->ofType, ofCodeTable, MaxOff, |
378 | 0 | nbSeq, fseTables->offcodeCTable, NULL, |
379 | 0 | OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff, |
380 | 0 | workspace, wkspSize); |
381 | 0 | cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->llType, llCodeTable, MaxLL, |
382 | 0 | nbSeq, fseTables->litlengthCTable, LL_bits, |
383 | 0 | LL_defaultNorm, LL_defaultNormLog, MaxLL, |
384 | 0 | workspace, wkspSize); |
385 | 0 | cSeqSizeEstimate += ZSTD_estimateSubBlockSize_symbolType(fseMetadata->mlType, mlCodeTable, MaxML, |
386 | 0 | nbSeq, fseTables->matchlengthCTable, ML_bits, |
387 | 0 | ML_defaultNorm, ML_defaultNormLog, MaxML, |
388 | 0 | workspace, wkspSize); |
389 | 0 | if (writeEntropy) cSeqSizeEstimate += fseMetadata->fseTablesSize; |
390 | 0 | return cSeqSizeEstimate + sequencesSectionHeaderSize; |
391 | 0 | } |
392 | | |
393 | | typedef struct { |
394 | | size_t estLitSize; |
395 | | size_t estBlockSize; |
396 | | } EstimatedBlockSize; |
397 | | static EstimatedBlockSize ZSTD_estimateSubBlockSize(const BYTE* literals, size_t litSize, |
398 | | const BYTE* ofCodeTable, |
399 | | const BYTE* llCodeTable, |
400 | | const BYTE* mlCodeTable, |
401 | | size_t nbSeq, |
402 | | const ZSTD_entropyCTables_t* entropy, |
403 | | const ZSTD_entropyCTablesMetadata_t* entropyMetadata, |
404 | | void* workspace, size_t wkspSize, |
405 | | int writeLitEntropy, int writeSeqEntropy) |
406 | 0 | { |
407 | 0 | EstimatedBlockSize ebs; |
408 | 0 | ebs.estLitSize = ZSTD_estimateSubBlockSize_literal(literals, litSize, |
409 | 0 | &entropy->huf, &entropyMetadata->hufMetadata, |
410 | 0 | workspace, wkspSize, writeLitEntropy); |
411 | 0 | ebs.estBlockSize = ZSTD_estimateSubBlockSize_sequences(ofCodeTable, llCodeTable, mlCodeTable, |
412 | 0 | nbSeq, &entropy->fse, &entropyMetadata->fseMetadata, |
413 | 0 | workspace, wkspSize, writeSeqEntropy); |
414 | 0 | ebs.estBlockSize += ebs.estLitSize + ZSTD_blockHeaderSize; |
415 | 0 | return ebs; |
416 | 0 | } |
417 | | |
418 | | static int ZSTD_needSequenceEntropyTables(ZSTD_fseCTablesMetadata_t const* fseMetadata) |
419 | 0 | { |
420 | 0 | if (fseMetadata->llType == set_compressed || fseMetadata->llType == set_rle) |
421 | 0 | return 1; |
422 | 0 | if (fseMetadata->mlType == set_compressed || fseMetadata->mlType == set_rle) |
423 | 0 | return 1; |
424 | 0 | if (fseMetadata->ofType == set_compressed || fseMetadata->ofType == set_rle) |
425 | 0 | return 1; |
426 | 0 | return 0; |
427 | 0 | } |
428 | | |
429 | | static size_t countLiterals(SeqStore_t const* seqStore, const SeqDef* sp, size_t seqCount) |
430 | 0 | { |
431 | 0 | size_t n, total = 0; |
432 | 0 | assert(sp != NULL); |
433 | 0 | for (n=0; n<seqCount; n++) { |
434 | 0 | total += ZSTD_getSequenceLength(seqStore, sp+n).litLength; |
435 | 0 | } |
436 | 0 | DEBUGLOG(6, "countLiterals for %zu sequences from %p => %zu bytes", seqCount, (const void*)sp, total); |
437 | 0 | return total; |
438 | 0 | } |
439 | | |
440 | 0 | #define BYTESCALE 256 |
441 | | |
442 | | static size_t sizeBlockSequences(const SeqDef* sp, size_t nbSeqs, |
443 | | size_t targetBudget, size_t avgLitCost, size_t avgSeqCost, |
444 | | int firstSubBlock) |
445 | 0 | { |
446 | 0 | size_t n, budget = 0, inSize=0; |
447 | | /* entropy headers */ |
448 | 0 | size_t const headerSize = (size_t)firstSubBlock * 120 * BYTESCALE; /* generous estimate */ |
449 | 0 | assert(firstSubBlock==0 || firstSubBlock==1); |
450 | 0 | budget += headerSize; |
451 | | |
452 | | /* first sequence => at least one sequence*/ |
453 | 0 | budget += sp[0].litLength * avgLitCost + avgSeqCost; |
454 | 0 | if (budget > targetBudget) return 1; |
455 | 0 | inSize = sp[0].litLength + (sp[0].mlBase+MINMATCH); |
456 | | |
457 | | /* loop over sequences */ |
458 | 0 | for (n=1; n<nbSeqs; n++) { |
459 | 0 | size_t currentCost = sp[n].litLength * avgLitCost + avgSeqCost; |
460 | 0 | budget += currentCost; |
461 | 0 | inSize += sp[n].litLength + (sp[n].mlBase+MINMATCH); |
462 | | /* stop when sub-block budget is reached */ |
463 | 0 | if ( (budget > targetBudget) |
464 | | /* though continue to expand until the sub-block is deemed compressible */ |
465 | 0 | && (budget < inSize * BYTESCALE) ) |
466 | 0 | break; |
467 | 0 | } |
468 | |
|
469 | 0 | return n; |
470 | 0 | } |
471 | | |
472 | | /** ZSTD_compressSubBlock_multi() : |
473 | | * Breaks super-block into multiple sub-blocks and compresses them. |
474 | | * Entropy will be written into the first block. |
475 | | * The following blocks use repeat_mode to compress. |
476 | | * Sub-blocks are all compressed, except the last one when beneficial. |
477 | | * @return : compressed size of the super block (which features multiple ZSTD blocks) |
478 | | * or 0 if it failed to compress. */ |
479 | | static size_t ZSTD_compressSubBlock_multi(const SeqStore_t* seqStorePtr, |
480 | | const ZSTD_compressedBlockState_t* prevCBlock, |
481 | | ZSTD_compressedBlockState_t* nextCBlock, |
482 | | const ZSTD_entropyCTablesMetadata_t* entropyMetadata, |
483 | | const ZSTD_CCtx_params* cctxParams, |
484 | | void* dst, size_t dstCapacity, |
485 | | const void* src, size_t srcSize, |
486 | | const int bmi2, U32 lastBlock, |
487 | | void* workspace, size_t wkspSize) |
488 | 0 | { |
489 | 0 | const SeqDef* const sstart = seqStorePtr->sequencesStart; |
490 | 0 | const SeqDef* const send = seqStorePtr->sequences; |
491 | 0 | const SeqDef* sp = sstart; /* tracks progresses within seqStorePtr->sequences */ |
492 | 0 | size_t const nbSeqs = (size_t)(send - sstart); |
493 | 0 | const BYTE* const lstart = seqStorePtr->litStart; |
494 | 0 | const BYTE* const lend = seqStorePtr->lit; |
495 | 0 | const BYTE* lp = lstart; |
496 | 0 | size_t const nbLiterals = (size_t)(lend - lstart); |
497 | 0 | BYTE const* ip = (BYTE const*)src; |
498 | 0 | BYTE const* const iend = ip + srcSize; |
499 | 0 | BYTE* const ostart = (BYTE*)dst; |
500 | 0 | BYTE* const oend = ostart + dstCapacity; |
501 | 0 | BYTE* op = ostart; |
502 | 0 | const BYTE* llCodePtr = seqStorePtr->llCode; |
503 | 0 | const BYTE* mlCodePtr = seqStorePtr->mlCode; |
504 | 0 | const BYTE* ofCodePtr = seqStorePtr->ofCode; |
505 | 0 | size_t const minTarget = ZSTD_TARGETCBLOCKSIZE_MIN; /* enforce minimum size, to reduce undesirable side effects */ |
506 | 0 | size_t const targetCBlockSize = MAX(minTarget, cctxParams->targetCBlockSize); |
507 | 0 | int writeLitEntropy = (entropyMetadata->hufMetadata.hType == set_compressed); |
508 | 0 | int writeSeqEntropy = 1; |
509 | |
|
510 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_multi (srcSize=%u, litSize=%u, nbSeq=%u)", |
511 | 0 | (unsigned)srcSize, (unsigned)(lend-lstart), (unsigned)(send-sstart)); |
512 | | |
513 | | /* let's start by a general estimation for the full block */ |
514 | 0 | if (nbSeqs > 0) { |
515 | 0 | EstimatedBlockSize const ebs = |
516 | 0 | ZSTD_estimateSubBlockSize(lp, nbLiterals, |
517 | 0 | ofCodePtr, llCodePtr, mlCodePtr, nbSeqs, |
518 | 0 | &nextCBlock->entropy, entropyMetadata, |
519 | 0 | workspace, wkspSize, |
520 | 0 | writeLitEntropy, writeSeqEntropy); |
521 | | /* quick estimation */ |
522 | 0 | size_t const avgLitCost = nbLiterals ? (ebs.estLitSize * BYTESCALE) / nbLiterals : BYTESCALE; |
523 | 0 | size_t const avgSeqCost = ((ebs.estBlockSize - ebs.estLitSize) * BYTESCALE) / nbSeqs; |
524 | 0 | const size_t nbSubBlocks = MAX((ebs.estBlockSize + (targetCBlockSize/2)) / targetCBlockSize, 1); |
525 | 0 | size_t n, avgBlockBudget, blockBudgetSupp=0; |
526 | 0 | avgBlockBudget = (ebs.estBlockSize * BYTESCALE) / nbSubBlocks; |
527 | 0 | DEBUGLOG(5, "estimated fullblock size=%u bytes ; avgLitCost=%.2f ; avgSeqCost=%.2f ; targetCBlockSize=%u, nbSubBlocks=%u ; avgBlockBudget=%.0f bytes", |
528 | 0 | (unsigned)ebs.estBlockSize, (double)avgLitCost/BYTESCALE, (double)avgSeqCost/BYTESCALE, |
529 | 0 | (unsigned)targetCBlockSize, (unsigned)nbSubBlocks, (double)avgBlockBudget/BYTESCALE); |
530 | | /* simplification: if estimates states that the full superblock doesn't compress, just bail out immediately |
531 | | * this will result in the production of a single uncompressed block covering @srcSize.*/ |
532 | 0 | if (ebs.estBlockSize > srcSize) return 0; |
533 | | |
534 | | /* compress and write sub-blocks */ |
535 | 0 | assert(nbSubBlocks>0); |
536 | 0 | for (n=0; n < nbSubBlocks-1; n++) { |
537 | | /* determine nb of sequences for current sub-block + nbLiterals from next sequence */ |
538 | 0 | size_t const seqCount = sizeBlockSequences(sp, (size_t)(send-sp), |
539 | 0 | avgBlockBudget + blockBudgetSupp, avgLitCost, avgSeqCost, n==0); |
540 | | /* if reached last sequence : break to last sub-block (simplification) */ |
541 | 0 | assert(seqCount <= (size_t)(send-sp)); |
542 | 0 | if (sp + seqCount == send) break; |
543 | 0 | assert(seqCount > 0); |
544 | | /* compress sub-block */ |
545 | 0 | { int litEntropyWritten = 0; |
546 | 0 | int seqEntropyWritten = 0; |
547 | 0 | size_t litSize = countLiterals(seqStorePtr, sp, seqCount); |
548 | 0 | const size_t decompressedSize = |
549 | 0 | ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 0); |
550 | 0 | size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata, |
551 | 0 | sp, seqCount, |
552 | 0 | lp, litSize, |
553 | 0 | llCodePtr, mlCodePtr, ofCodePtr, |
554 | 0 | cctxParams, |
555 | 0 | op, (size_t)(oend-op), |
556 | 0 | bmi2, writeLitEntropy, writeSeqEntropy, |
557 | 0 | &litEntropyWritten, &seqEntropyWritten, |
558 | 0 | 0); |
559 | 0 | FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed"); |
560 | | |
561 | | /* check compressibility, update state components */ |
562 | 0 | if (cSize > 0 && cSize < decompressedSize) { |
563 | 0 | DEBUGLOG(5, "Committed sub-block compressing %u bytes => %u bytes", |
564 | 0 | (unsigned)decompressedSize, (unsigned)cSize); |
565 | 0 | assert(ip + decompressedSize <= iend); |
566 | 0 | ip += decompressedSize; |
567 | 0 | lp += litSize; |
568 | 0 | op += cSize; |
569 | 0 | llCodePtr += seqCount; |
570 | 0 | mlCodePtr += seqCount; |
571 | 0 | ofCodePtr += seqCount; |
572 | | /* Entropy only needs to be written once */ |
573 | 0 | if (litEntropyWritten) { |
574 | 0 | writeLitEntropy = 0; |
575 | 0 | } |
576 | 0 | if (seqEntropyWritten) { |
577 | 0 | writeSeqEntropy = 0; |
578 | 0 | } |
579 | 0 | sp += seqCount; |
580 | 0 | blockBudgetSupp = 0; |
581 | 0 | } } |
582 | | /* otherwise : do not compress yet, coalesce current sub-block with following one */ |
583 | 0 | } |
584 | 0 | } /* if (nbSeqs > 0) */ |
585 | | |
586 | | /* write last block */ |
587 | 0 | DEBUGLOG(5, "Generate last sub-block: %u sequences remaining", (unsigned)(send - sp)); |
588 | 0 | { int litEntropyWritten = 0; |
589 | 0 | int seqEntropyWritten = 0; |
590 | 0 | size_t litSize = (size_t)(lend - lp); |
591 | 0 | size_t seqCount = (size_t)(send - sp); |
592 | 0 | const size_t decompressedSize = |
593 | 0 | ZSTD_seqDecompressedSize(seqStorePtr, sp, seqCount, litSize, 1); |
594 | 0 | size_t const cSize = ZSTD_compressSubBlock(&nextCBlock->entropy, entropyMetadata, |
595 | 0 | sp, seqCount, |
596 | 0 | lp, litSize, |
597 | 0 | llCodePtr, mlCodePtr, ofCodePtr, |
598 | 0 | cctxParams, |
599 | 0 | op, (size_t)(oend-op), |
600 | 0 | bmi2, writeLitEntropy, writeSeqEntropy, |
601 | 0 | &litEntropyWritten, &seqEntropyWritten, |
602 | 0 | lastBlock); |
603 | 0 | FORWARD_IF_ERROR(cSize, "ZSTD_compressSubBlock failed"); |
604 | | |
605 | | /* update pointers, the nb of literals borrowed from next sequence must be preserved */ |
606 | 0 | if (cSize > 0 && cSize < decompressedSize) { |
607 | 0 | DEBUGLOG(5, "Last sub-block compressed %u bytes => %u bytes", |
608 | 0 | (unsigned)decompressedSize, (unsigned)cSize); |
609 | 0 | assert(ip + decompressedSize <= iend); |
610 | 0 | ip += decompressedSize; |
611 | 0 | lp += litSize; |
612 | 0 | op += cSize; |
613 | 0 | llCodePtr += seqCount; |
614 | 0 | mlCodePtr += seqCount; |
615 | 0 | ofCodePtr += seqCount; |
616 | | /* Entropy only needs to be written once */ |
617 | 0 | if (litEntropyWritten) { |
618 | 0 | writeLitEntropy = 0; |
619 | 0 | } |
620 | 0 | if (seqEntropyWritten) { |
621 | 0 | writeSeqEntropy = 0; |
622 | 0 | } |
623 | 0 | sp += seqCount; |
624 | 0 | } |
625 | 0 | } |
626 | | |
627 | | |
628 | 0 | if (writeLitEntropy) { |
629 | 0 | DEBUGLOG(5, "Literal entropy tables were never written"); |
630 | 0 | ZSTD_memcpy(&nextCBlock->entropy.huf, &prevCBlock->entropy.huf, sizeof(prevCBlock->entropy.huf)); |
631 | 0 | } |
632 | 0 | if (writeSeqEntropy && ZSTD_needSequenceEntropyTables(&entropyMetadata->fseMetadata)) { |
633 | | /* If we haven't written our entropy tables, then we've violated our contract and |
634 | | * must emit an uncompressed block. |
635 | | */ |
636 | 0 | DEBUGLOG(5, "Sequence entropy tables were never written => cancel, emit an uncompressed block"); |
637 | 0 | return 0; |
638 | 0 | } |
639 | | |
640 | 0 | if (ip < iend) { |
641 | | /* some data left : last part of the block sent uncompressed */ |
642 | 0 | size_t const rSize = (size_t)((iend - ip)); |
643 | 0 | size_t const cSize = ZSTD_noCompressBlock(op, (size_t)(oend - op), ip, rSize, lastBlock); |
644 | 0 | DEBUGLOG(5, "Generate last uncompressed sub-block of %u bytes", (unsigned)(rSize)); |
645 | 0 | FORWARD_IF_ERROR(cSize, "ZSTD_noCompressBlock failed"); |
646 | 0 | assert(cSize != 0); |
647 | 0 | op += cSize; |
648 | | /* We have to regenerate the repcodes because we've skipped some sequences */ |
649 | 0 | if (sp < send) { |
650 | 0 | const SeqDef* seq; |
651 | 0 | Repcodes_t rep; |
652 | 0 | ZSTD_memcpy(&rep, prevCBlock->rep, sizeof(rep)); |
653 | 0 | for (seq = sstart; seq < sp; ++seq) { |
654 | 0 | ZSTD_updateRep(rep.rep, seq->offBase, ZSTD_getSequenceLength(seqStorePtr, seq).litLength == 0); |
655 | 0 | } |
656 | 0 | ZSTD_memcpy(nextCBlock->rep, &rep, sizeof(rep)); |
657 | 0 | } |
658 | 0 | } |
659 | | |
660 | 0 | DEBUGLOG(5, "ZSTD_compressSubBlock_multi compressed all subBlocks: total compressed size = %u", |
661 | 0 | (unsigned)(op-ostart)); |
662 | 0 | return (size_t)(op-ostart); |
663 | 0 | } |
664 | | |
665 | | size_t ZSTD_compressSuperBlock(ZSTD_CCtx* zc, |
666 | | void* dst, size_t dstCapacity, |
667 | | const void* src, size_t srcSize, |
668 | | unsigned lastBlock) |
669 | 0 | { |
670 | 0 | ZSTD_entropyCTablesMetadata_t entropyMetadata; |
671 | |
|
672 | 0 | FORWARD_IF_ERROR(ZSTD_buildBlockEntropyStats(&zc->seqStore, |
673 | 0 | &zc->blockState.prevCBlock->entropy, |
674 | 0 | &zc->blockState.nextCBlock->entropy, |
675 | 0 | &zc->appliedParams, |
676 | 0 | &entropyMetadata, |
677 | 0 | zc->tmpWorkspace, zc->tmpWkspSize /* statically allocated in resetCCtx */), ""); |
678 | | |
679 | 0 | return ZSTD_compressSubBlock_multi(&zc->seqStore, |
680 | 0 | zc->blockState.prevCBlock, |
681 | 0 | zc->blockState.nextCBlock, |
682 | 0 | &entropyMetadata, |
683 | 0 | &zc->appliedParams, |
684 | 0 | dst, dstCapacity, |
685 | 0 | src, srcSize, |
686 | 0 | zc->bmi2, lastBlock, |
687 | 0 | zc->tmpWorkspace, zc->tmpWkspSize /* statically allocated in resetCCtx */); |
688 | 0 | } |