blob: a05b8a47f14ea204183a3fd1d1aa7bd6b6d1f084 [file] [log] [blame]
William Kurkianea869482019-04-09 15:16:11 -04001/*
2 * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
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#include "zstd_compress_internal.h"
12#include "zstd_fast.h"
13
14
15void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
Devmalya Pauleb2e9552019-08-27 19:42:00 -040016 const void* const end,
17 ZSTD_dictTableLoadMethod_e dtlm)
William Kurkianea869482019-04-09 15:16:11 -040018{
Abhilash S.L3b494632019-07-16 15:51:09 +053019 const ZSTD_compressionParameters* const cParams = &ms->cParams;
William Kurkianea869482019-04-09 15:16:11 -040020 U32* const hashTable = ms->hashTable;
21 U32 const hBits = cParams->hashLog;
Abhilash S.L3b494632019-07-16 15:51:09 +053022 U32 const mls = cParams->minMatch;
William Kurkianea869482019-04-09 15:16:11 -040023 const BYTE* const base = ms->window.base;
24 const BYTE* ip = base + ms->nextToUpdate;
25 const BYTE* const iend = ((const BYTE*)end) - HASH_READ_SIZE;
26 const U32 fastHashFillStep = 3;
27
28 /* Always insert every fastHashFillStep position into the hash table.
29 * Insert the other positions if their hash entry is empty.
30 */
Abhilash S.L3b494632019-07-16 15:51:09 +053031 for ( ; ip + fastHashFillStep < iend + 2; ip += fastHashFillStep) {
William Kurkianea869482019-04-09 15:16:11 -040032 U32 const current = (U32)(ip - base);
Abhilash S.L3b494632019-07-16 15:51:09 +053033 size_t const hash0 = ZSTD_hashPtr(ip, hBits, mls);
34 hashTable[hash0] = current;
35 if (dtlm == ZSTD_dtlm_fast) continue;
36 /* Only load extra positions for ZSTD_dtlm_full */
37 { U32 p;
38 for (p = 1; p < fastHashFillStep; ++p) {
39 size_t const hash = ZSTD_hashPtr(ip + p, hBits, mls);
40 if (hashTable[hash] == 0) { /* not yet filled */
41 hashTable[hash] = current + p;
42 } } } }
William Kurkianea869482019-04-09 15:16:11 -040043}
44
Devmalya Pauleb2e9552019-08-27 19:42:00 -040045
William Kurkianea869482019-04-09 15:16:11 -040046FORCE_INLINE_TEMPLATE
47size_t ZSTD_compressBlock_fast_generic(
48 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
49 void const* src, size_t srcSize,
Abhilash S.L3b494632019-07-16 15:51:09 +053050 U32 const mls)
William Kurkianea869482019-04-09 15:16:11 -040051{
Abhilash S.L3b494632019-07-16 15:51:09 +053052 const ZSTD_compressionParameters* const cParams = &ms->cParams;
William Kurkianea869482019-04-09 15:16:11 -040053 U32* const hashTable = ms->hashTable;
Abhilash S.L3b494632019-07-16 15:51:09 +053054 U32 const hlog = cParams->hashLog;
55 /* support stepSize of 0 */
56 size_t const stepSize = cParams->targetLength + !(cParams->targetLength) + 1;
William Kurkianea869482019-04-09 15:16:11 -040057 const BYTE* const base = ms->window.base;
58 const BYTE* const istart = (const BYTE*)src;
Abhilash S.L3b494632019-07-16 15:51:09 +053059 /* We check ip0 (ip + 0) and ip1 (ip + 1) each loop */
60 const BYTE* ip0 = istart;
61 const BYTE* ip1;
William Kurkianea869482019-04-09 15:16:11 -040062 const BYTE* anchor = istart;
Devmalya Pauleb2e9552019-08-27 19:42:00 -040063 const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
64 const U32 maxDistance = 1U << cParams->windowLog;
65 const U32 validStartIndex = ms->window.dictLimit;
66 const U32 prefixStartIndex = (endIndex - validStartIndex > maxDistance) ? endIndex - maxDistance : validStartIndex;
Abhilash S.L3b494632019-07-16 15:51:09 +053067 const BYTE* const prefixStart = base + prefixStartIndex;
William Kurkianea869482019-04-09 15:16:11 -040068 const BYTE* const iend = istart + srcSize;
69 const BYTE* const ilimit = iend - HASH_READ_SIZE;
70 U32 offset_1=rep[0], offset_2=rep[1];
71 U32 offsetSaved = 0;
72
73 /* init */
Abhilash S.L3b494632019-07-16 15:51:09 +053074 ip0 += (ip0 == prefixStart);
75 ip1 = ip0 + 1;
76 {
77 U32 const maxRep = (U32)(ip0 - prefixStart);
William Kurkianea869482019-04-09 15:16:11 -040078 if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
79 if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
80 }
81
82 /* Main Search Loop */
Abhilash S.L3b494632019-07-16 15:51:09 +053083 while (ip1 < ilimit) { /* < instead of <=, because check at ip0+2 */
William Kurkianea869482019-04-09 15:16:11 -040084 size_t mLength;
Abhilash S.L3b494632019-07-16 15:51:09 +053085 BYTE const* ip2 = ip0 + 2;
86 size_t const h0 = ZSTD_hashPtr(ip0, hlog, mls);
87 U32 const val0 = MEM_read32(ip0);
88 size_t const h1 = ZSTD_hashPtr(ip1, hlog, mls);
89 U32 const val1 = MEM_read32(ip1);
90 U32 const current0 = (U32)(ip0-base);
91 U32 const current1 = (U32)(ip1-base);
92 U32 const matchIndex0 = hashTable[h0];
93 U32 const matchIndex1 = hashTable[h1];
94 BYTE const* repMatch = ip2-offset_1;
95 const BYTE* match0 = base + matchIndex0;
96 const BYTE* match1 = base + matchIndex1;
97 U32 offcode;
98 hashTable[h0] = current0; /* update hash table */
99 hashTable[h1] = current1; /* update hash table */
William Kurkianea869482019-04-09 15:16:11 -0400100
Abhilash S.L3b494632019-07-16 15:51:09 +0530101 assert(ip0 + 1 == ip1);
William Kurkianea869482019-04-09 15:16:11 -0400102
Abhilash S.L3b494632019-07-16 15:51:09 +0530103 if ((offset_1 > 0) & (MEM_read32(repMatch) == MEM_read32(ip2))) {
104 mLength = ip2[-1] == repMatch[-1] ? 1 : 0;
105 ip0 = ip2 - mLength;
106 match0 = repMatch - mLength;
107 offcode = 0;
108 goto _match;
109 }
110 if ((matchIndex0 > prefixStartIndex) && MEM_read32(match0) == val0) {
111 /* found a regular match */
112 goto _offset;
113 }
114 if ((matchIndex1 > prefixStartIndex) && MEM_read32(match1) == val1) {
115 /* found a regular match after one literal */
116 ip0 = ip1;
117 match0 = match1;
118 goto _offset;
119 }
120 {
121 size_t const step = ((ip0-anchor) >> (kSearchStrength - 1)) + stepSize;
122 assert(step >= 2);
123 ip0 += step;
124 ip1 += step;
125 continue;
126 }
127_offset: /* Requires: ip0, match0 */
128 /* Compute the offset code */
129 offset_2 = offset_1;
130 offset_1 = (U32)(ip0-match0);
131 offcode = offset_1 + ZSTD_REP_MOVE;
132 mLength = 0;
133 /* Count the backwards match length */
134 while (((ip0>anchor) & (match0>prefixStart))
135 && (ip0[-1] == match0[-1])) { ip0--; match0--; mLength++; } /* catch up */
136
137_match: /* Requires: ip0, match0, offcode */
138 /* Count the forward length */
139 mLength += ZSTD_count(ip0+mLength+4, match0+mLength+4, iend) + 4;
140 ZSTD_storeSeq(seqStore, ip0-anchor, anchor, offcode, mLength-MINMATCH);
William Kurkianea869482019-04-09 15:16:11 -0400141 /* match found */
Abhilash S.L3b494632019-07-16 15:51:09 +0530142 ip0 += mLength;
143 anchor = ip0;
144 ip1 = ip0 + 1;
William Kurkianea869482019-04-09 15:16:11 -0400145
Abhilash S.L3b494632019-07-16 15:51:09 +0530146 if (ip0 <= ilimit) {
William Kurkianea869482019-04-09 15:16:11 -0400147 /* Fill Table */
Abhilash S.L3b494632019-07-16 15:51:09 +0530148 assert(base+current0+2 > istart); /* check base overflow */
149 hashTable[ZSTD_hashPtr(base+current0+2, hlog, mls)] = current0+2; /* here because current+2 could be > iend-8 */
150 hashTable[ZSTD_hashPtr(ip0-2, hlog, mls)] = (U32)(ip0-2-base);
151
152 while ( (ip0 <= ilimit)
William Kurkianea869482019-04-09 15:16:11 -0400153 && ( (offset_2>0)
Abhilash S.L3b494632019-07-16 15:51:09 +0530154 & (MEM_read32(ip0) == MEM_read32(ip0 - offset_2)) )) {
William Kurkianea869482019-04-09 15:16:11 -0400155 /* store sequence */
Abhilash S.L3b494632019-07-16 15:51:09 +0530156 size_t const rLength = ZSTD_count(ip0+4, ip0+4-offset_2, iend) + 4;
157 U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff; /* swap offset_2 <=> offset_1 */
158 hashTable[ZSTD_hashPtr(ip0, hlog, mls)] = (U32)(ip0-base);
159 ip0 += rLength;
160 ip1 = ip0 + 1;
William Kurkianea869482019-04-09 15:16:11 -0400161 ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
Abhilash S.L3b494632019-07-16 15:51:09 +0530162 anchor = ip0;
163 continue; /* faster when present (confirmed on gcc-8) ... (?) */
164 }
165 }
166 }
William Kurkianea869482019-04-09 15:16:11 -0400167
168 /* save reps for next block */
169 rep[0] = offset_1 ? offset_1 : offsetSaved;
170 rep[1] = offset_2 ? offset_2 : offsetSaved;
171
172 /* Return the last literals size */
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400173 return (size_t)(iend - anchor);
William Kurkianea869482019-04-09 15:16:11 -0400174}
175
176
177size_t ZSTD_compressBlock_fast(
178 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
Abhilash S.L3b494632019-07-16 15:51:09 +0530179 void const* src, size_t srcSize)
William Kurkianea869482019-04-09 15:16:11 -0400180{
Abhilash S.L3b494632019-07-16 15:51:09 +0530181 ZSTD_compressionParameters const* cParams = &ms->cParams;
182 U32 const mls = cParams->minMatch;
183 assert(ms->dictMatchState == NULL);
William Kurkianea869482019-04-09 15:16:11 -0400184 switch(mls)
185 {
186 default: /* includes case 3 */
187 case 4 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530188 return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4);
William Kurkianea869482019-04-09 15:16:11 -0400189 case 5 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530190 return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5);
William Kurkianea869482019-04-09 15:16:11 -0400191 case 6 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530192 return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6);
William Kurkianea869482019-04-09 15:16:11 -0400193 case 7 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530194 return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7);
195 }
196}
197
198FORCE_INLINE_TEMPLATE
199size_t ZSTD_compressBlock_fast_dictMatchState_generic(
200 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
201 void const* src, size_t srcSize, U32 const mls)
202{
203 const ZSTD_compressionParameters* const cParams = &ms->cParams;
204 U32* const hashTable = ms->hashTable;
205 U32 const hlog = cParams->hashLog;
206 /* support stepSize of 0 */
207 U32 const stepSize = cParams->targetLength + !(cParams->targetLength);
208 const BYTE* const base = ms->window.base;
209 const BYTE* const istart = (const BYTE*)src;
210 const BYTE* ip = istart;
211 const BYTE* anchor = istart;
212 const U32 prefixStartIndex = ms->window.dictLimit;
213 const BYTE* const prefixStart = base + prefixStartIndex;
214 const BYTE* const iend = istart + srcSize;
215 const BYTE* const ilimit = iend - HASH_READ_SIZE;
216 U32 offset_1=rep[0], offset_2=rep[1];
217 U32 offsetSaved = 0;
218
219 const ZSTD_matchState_t* const dms = ms->dictMatchState;
220 const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;
221 const U32* const dictHashTable = dms->hashTable;
222 const U32 dictStartIndex = dms->window.dictLimit;
223 const BYTE* const dictBase = dms->window.base;
224 const BYTE* const dictStart = dictBase + dictStartIndex;
225 const BYTE* const dictEnd = dms->window.nextSrc;
226 const U32 dictIndexDelta = prefixStartIndex - (U32)(dictEnd - dictBase);
227 const U32 dictAndPrefixLength = (U32)(ip - prefixStart + dictEnd - dictStart);
228 const U32 dictHLog = dictCParams->hashLog;
229
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400230 /* if a dictionary is still attached, it necessarily means that
231 * it is within window size. So we just check it. */
232 const U32 maxDistance = 1U << cParams->windowLog;
233 const U32 endIndex = (U32)((size_t)(ip - base) + srcSize);
234 assert(endIndex - prefixStartIndex <= maxDistance);
235 (void)maxDistance; (void)endIndex; /* these variables are not used when assert() is disabled */
236
237 /* ensure there will be no no underflow
238 * when translating a dict index into a local index */
Abhilash S.L3b494632019-07-16 15:51:09 +0530239 assert(prefixStartIndex >= (U32)(dictEnd - dictBase));
240
241 /* init */
242 ip += (dictAndPrefixLength == 0);
243 /* dictMatchState repCode checks don't currently handle repCode == 0
244 * disabling. */
245 assert(offset_1 <= dictAndPrefixLength);
246 assert(offset_2 <= dictAndPrefixLength);
247
248 /* Main Search Loop */
249 while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
250 size_t mLength;
251 size_t const h = ZSTD_hashPtr(ip, hlog, mls);
252 U32 const current = (U32)(ip-base);
253 U32 const matchIndex = hashTable[h];
254 const BYTE* match = base + matchIndex;
255 const U32 repIndex = current + 1 - offset_1;
256 const BYTE* repMatch = (repIndex < prefixStartIndex) ?
257 dictBase + (repIndex - dictIndexDelta) :
258 base + repIndex;
259 hashTable[h] = current; /* update hash table */
260
261 if ( ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
262 && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
263 const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
264 mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
265 ip++;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400266 ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
Abhilash S.L3b494632019-07-16 15:51:09 +0530267 } else if ( (matchIndex <= prefixStartIndex) ) {
268 size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
269 U32 const dictMatchIndex = dictHashTable[dictHash];
270 const BYTE* dictMatch = dictBase + dictMatchIndex;
271 if (dictMatchIndex <= dictStartIndex ||
272 MEM_read32(dictMatch) != MEM_read32(ip)) {
273 assert(stepSize >= 1);
274 ip += ((ip-anchor) >> kSearchStrength) + stepSize;
275 continue;
276 } else {
277 /* found a dict match */
278 U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);
279 mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
280 while (((ip>anchor) & (dictMatch>dictStart))
281 && (ip[-1] == dictMatch[-1])) {
282 ip--; dictMatch--; mLength++;
283 } /* catch up */
284 offset_2 = offset_1;
285 offset_1 = offset;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400286 ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
Abhilash S.L3b494632019-07-16 15:51:09 +0530287 }
288 } else if (MEM_read32(match) != MEM_read32(ip)) {
289 /* it's not a match, and we're not going to check the dictionary */
290 assert(stepSize >= 1);
291 ip += ((ip-anchor) >> kSearchStrength) + stepSize;
292 continue;
293 } else {
294 /* found a regular match */
295 U32 const offset = (U32)(ip-match);
296 mLength = ZSTD_count(ip+4, match+4, iend) + 4;
297 while (((ip>anchor) & (match>prefixStart))
298 && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
299 offset_2 = offset_1;
300 offset_1 = offset;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400301 ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
Abhilash S.L3b494632019-07-16 15:51:09 +0530302 }
303
304 /* match found */
305 ip += mLength;
306 anchor = ip;
307
308 if (ip <= ilimit) {
309 /* Fill Table */
310 assert(base+current+2 > istart); /* check base overflow */
311 hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2; /* here because current+2 could be > iend-8 */
312 hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
313
314 /* check immediate repcode */
315 while (ip <= ilimit) {
316 U32 const current2 = (U32)(ip-base);
317 U32 const repIndex2 = current2 - offset_2;
318 const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
319 dictBase - dictIndexDelta + repIndex2 :
320 base + repIndex2;
321 if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
322 && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
323 const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
324 size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
325 U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
326 ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
327 hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
328 ip += repLength2;
329 anchor = ip;
330 continue;
331 }
332 break;
333 }
334 }
335 }
336
337 /* save reps for next block */
338 rep[0] = offset_1 ? offset_1 : offsetSaved;
339 rep[1] = offset_2 ? offset_2 : offsetSaved;
340
341 /* Return the last literals size */
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400342 return (size_t)(iend - anchor);
Abhilash S.L3b494632019-07-16 15:51:09 +0530343}
344
345size_t ZSTD_compressBlock_fast_dictMatchState(
346 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
347 void const* src, size_t srcSize)
348{
349 ZSTD_compressionParameters const* cParams = &ms->cParams;
350 U32 const mls = cParams->minMatch;
351 assert(ms->dictMatchState != NULL);
352 switch(mls)
353 {
354 default: /* includes case 3 */
355 case 4 :
356 return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 4);
357 case 5 :
358 return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 5);
359 case 6 :
360 return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 6);
361 case 7 :
362 return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 7);
William Kurkianea869482019-04-09 15:16:11 -0400363 }
364}
365
366
367static size_t ZSTD_compressBlock_fast_extDict_generic(
368 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
Abhilash S.L3b494632019-07-16 15:51:09 +0530369 void const* src, size_t srcSize, U32 const mls)
William Kurkianea869482019-04-09 15:16:11 -0400370{
Abhilash S.L3b494632019-07-16 15:51:09 +0530371 const ZSTD_compressionParameters* const cParams = &ms->cParams;
372 U32* const hashTable = ms->hashTable;
373 U32 const hlog = cParams->hashLog;
374 /* support stepSize of 0 */
375 U32 const stepSize = cParams->targetLength + !(cParams->targetLength);
William Kurkianea869482019-04-09 15:16:11 -0400376 const BYTE* const base = ms->window.base;
377 const BYTE* const dictBase = ms->window.dictBase;
378 const BYTE* const istart = (const BYTE*)src;
379 const BYTE* ip = istart;
380 const BYTE* anchor = istart;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400381 const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
382 const U32 maxDistance = 1U << cParams->windowLog;
383 const U32 validLow = ms->window.lowLimit;
384 const U32 lowLimit = (endIndex - validLow > maxDistance) ? endIndex - maxDistance : validLow;
385 const U32 dictStartIndex = lowLimit;
Abhilash S.L3b494632019-07-16 15:51:09 +0530386 const BYTE* const dictStart = dictBase + dictStartIndex;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400387 const U32 dictLimit = ms->window.dictLimit;
388 const U32 prefixStartIndex = dictLimit < lowLimit ? lowLimit : dictLimit;
Abhilash S.L3b494632019-07-16 15:51:09 +0530389 const BYTE* const prefixStart = base + prefixStartIndex;
390 const BYTE* const dictEnd = dictBase + prefixStartIndex;
William Kurkianea869482019-04-09 15:16:11 -0400391 const BYTE* const iend = istart + srcSize;
392 const BYTE* const ilimit = iend - 8;
393 U32 offset_1=rep[0], offset_2=rep[1];
394
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400395 /* switch to "regular" variant if extDict is invalidated due to maxDistance */
396 if (prefixStartIndex == dictStartIndex)
397 return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, mls);
398
William Kurkianea869482019-04-09 15:16:11 -0400399 /* Search Loop */
400 while (ip < ilimit) { /* < instead of <=, because (ip+1) */
401 const size_t h = ZSTD_hashPtr(ip, hlog, mls);
Abhilash S.L3b494632019-07-16 15:51:09 +0530402 const U32 matchIndex = hashTable[h];
403 const BYTE* const matchBase = matchIndex < prefixStartIndex ? dictBase : base;
404 const BYTE* match = matchBase + matchIndex;
405 const U32 current = (U32)(ip-base);
406 const U32 repIndex = current + 1 - offset_1;
407 const BYTE* const repBase = repIndex < prefixStartIndex ? dictBase : base;
408 const BYTE* const repMatch = repBase + repIndex;
William Kurkianea869482019-04-09 15:16:11 -0400409 size_t mLength;
410 hashTable[h] = current; /* update hash table */
Abhilash S.L3b494632019-07-16 15:51:09 +0530411 assert(offset_1 <= current +1); /* check repIndex */
William Kurkianea869482019-04-09 15:16:11 -0400412
Abhilash S.L3b494632019-07-16 15:51:09 +0530413 if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex))
William Kurkianea869482019-04-09 15:16:11 -0400414 && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
Abhilash S.L3b494632019-07-16 15:51:09 +0530415 const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
416 mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
William Kurkianea869482019-04-09 15:16:11 -0400417 ip++;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400418 ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
William Kurkianea869482019-04-09 15:16:11 -0400419 } else {
Abhilash S.L3b494632019-07-16 15:51:09 +0530420 if ( (matchIndex < dictStartIndex) ||
William Kurkianea869482019-04-09 15:16:11 -0400421 (MEM_read32(match) != MEM_read32(ip)) ) {
422 assert(stepSize >= 1);
423 ip += ((ip-anchor) >> kSearchStrength) + stepSize;
424 continue;
425 }
Abhilash S.L3b494632019-07-16 15:51:09 +0530426 { const BYTE* matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
427 const BYTE* lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
William Kurkianea869482019-04-09 15:16:11 -0400428 U32 offset;
Abhilash S.L3b494632019-07-16 15:51:09 +0530429 mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
William Kurkianea869482019-04-09 15:16:11 -0400430 while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
431 offset = current - matchIndex;
432 offset_2 = offset_1;
433 offset_1 = offset;
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400434 ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
William Kurkianea869482019-04-09 15:16:11 -0400435 } }
436
437 /* found a match : store it */
438 ip += mLength;
439 anchor = ip;
440
441 if (ip <= ilimit) {
442 /* Fill Table */
443 hashTable[ZSTD_hashPtr(base+current+2, hlog, mls)] = current+2;
444 hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
445 /* check immediate repcode */
446 while (ip <= ilimit) {
447 U32 const current2 = (U32)(ip-base);
448 U32 const repIndex2 = current2 - offset_2;
Abhilash S.L3b494632019-07-16 15:51:09 +0530449 const BYTE* repMatch2 = repIndex2 < prefixStartIndex ? dictBase + repIndex2 : base + repIndex2;
450 if ( (((U32)((prefixStartIndex-1) - repIndex2) >= 3) & (repIndex2 > dictStartIndex)) /* intentional overflow */
William Kurkianea869482019-04-09 15:16:11 -0400451 && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
Abhilash S.L3b494632019-07-16 15:51:09 +0530452 const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
453 size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
William Kurkianea869482019-04-09 15:16:11 -0400454 U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
455 ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
456 hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
457 ip += repLength2;
458 anchor = ip;
459 continue;
460 }
461 break;
462 } } }
463
464 /* save reps for next block */
465 rep[0] = offset_1;
466 rep[1] = offset_2;
467
468 /* Return the last literals size */
Devmalya Pauleb2e9552019-08-27 19:42:00 -0400469 return (size_t)(iend - anchor);
William Kurkianea869482019-04-09 15:16:11 -0400470}
471
472
473size_t ZSTD_compressBlock_fast_extDict(
474 ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
Abhilash S.L3b494632019-07-16 15:51:09 +0530475 void const* src, size_t srcSize)
William Kurkianea869482019-04-09 15:16:11 -0400476{
Abhilash S.L3b494632019-07-16 15:51:09 +0530477 ZSTD_compressionParameters const* cParams = &ms->cParams;
478 U32 const mls = cParams->minMatch;
William Kurkianea869482019-04-09 15:16:11 -0400479 switch(mls)
480 {
481 default: /* includes case 3 */
482 case 4 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530483 return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 4);
William Kurkianea869482019-04-09 15:16:11 -0400484 case 5 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530485 return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 5);
William Kurkianea869482019-04-09 15:16:11 -0400486 case 6 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530487 return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 6);
William Kurkianea869482019-04-09 15:16:11 -0400488 case 7 :
Abhilash S.L3b494632019-07-16 15:51:09 +0530489 return ZSTD_compressBlock_fast_extDict_generic(ms, seqStore, rep, src, srcSize, 7);
William Kurkianea869482019-04-09 15:16:11 -0400490 }
491}