Loading...
1/*
2 * LZ4 - Fast LZ compression algorithm
3 * Copyright (C) 2011-2012, Yann Collet.
4 * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
5
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions are
8 * met:
9 *
10 * * Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * * Redistributions in binary form must reproduce the above
13 * copyright notice, this list of conditions and the following disclaimer
14 * in the documentation and/or other materials provided with the
15 * distribution.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
21 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
22 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
23 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
27 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 *
29 * You can contact the author at :
30 * - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html
31 * - LZ4 source repository : http://code.google.com/p/lz4/
32 *
33 * Changed for kernel use by:
34 * Chanho Min <chanho.min@lge.com>
35 */
36
37#include <linux/module.h>
38#include <linux/kernel.h>
39#include <linux/lz4.h>
40#include <asm/unaligned.h>
41#include "lz4defs.h"
42
43/*
44 * LZ4_compressCtx :
45 * -----------------
46 * Compress 'isize' bytes from 'source' into an output buffer 'dest' of
47 * maximum size 'maxOutputSize'. * If it cannot achieve it, compression
48 * will stop, and result of the function will be zero.
49 * return : the number of bytes written in buffer 'dest', or 0 if the
50 * compression fails
51 */
52static inline int lz4_compressctx(void *ctx,
53 const char *source,
54 char *dest,
55 int isize,
56 int maxoutputsize)
57{
58 HTYPE *hashtable = (HTYPE *)ctx;
59 const u8 *ip = (u8 *)source;
60#if LZ4_ARCH64
61 const BYTE * const base = ip;
62#else
63 const int base = 0;
64#endif
65 const u8 *anchor = ip;
66 const u8 *const iend = ip + isize;
67 const u8 *const mflimit = iend - MFLIMIT;
68 #define MATCHLIMIT (iend - LASTLITERALS)
69
70 u8 *op = (u8 *) dest;
71 u8 *const oend = op + maxoutputsize;
72 int length;
73 const int skipstrength = SKIPSTRENGTH;
74 u32 forwardh;
75 int lastrun;
76
77 /* Init */
78 if (isize < MINLENGTH)
79 goto _last_literals;
80
81 memset((void *)hashtable, 0, LZ4_MEM_COMPRESS);
82
83 /* First Byte */
84 hashtable[LZ4_HASH_VALUE(ip)] = ip - base;
85 ip++;
86 forwardh = LZ4_HASH_VALUE(ip);
87
88 /* Main Loop */
89 for (;;) {
90 int findmatchattempts = (1U << skipstrength) + 3;
91 const u8 *forwardip = ip;
92 const u8 *ref;
93 u8 *token;
94
95 /* Find a match */
96 do {
97 u32 h = forwardh;
98 int step = findmatchattempts++ >> skipstrength;
99 ip = forwardip;
100 forwardip = ip + step;
101
102 if (unlikely(forwardip > mflimit))
103 goto _last_literals;
104
105 forwardh = LZ4_HASH_VALUE(forwardip);
106 ref = base + hashtable[h];
107 hashtable[h] = ip - base;
108 } while ((ref < ip - MAX_DISTANCE) || (A32(ref) != A32(ip)));
109
110 /* Catch up */
111 while ((ip > anchor) && (ref > (u8 *)source) &&
112 unlikely(ip[-1] == ref[-1])) {
113 ip--;
114 ref--;
115 }
116
117 /* Encode Literal length */
118 length = (int)(ip - anchor);
119 token = op++;
120 /* check output limit */
121 if (unlikely(op + length + (2 + 1 + LASTLITERALS) +
122 (length >> 8) > oend))
123 return 0;
124
125 if (length >= (int)RUN_MASK) {
126 int len;
127 *token = (RUN_MASK << ML_BITS);
128 len = length - RUN_MASK;
129 for (; len > 254 ; len -= 255)
130 *op++ = 255;
131 *op++ = (u8)len;
132 } else
133 *token = (length << ML_BITS);
134
135 /* Copy Literals */
136 LZ4_BLINDCOPY(anchor, op, length);
137_next_match:
138 /* Encode Offset */
139 LZ4_WRITE_LITTLEENDIAN_16(op, (u16)(ip - ref));
140
141 /* Start Counting */
142 ip += MINMATCH;
143 /* MinMatch verified */
144 ref += MINMATCH;
145 anchor = ip;
146 while (likely(ip < MATCHLIMIT - (STEPSIZE - 1))) {
147 #if LZ4_ARCH64
148 u64 diff = A64(ref) ^ A64(ip);
149 #else
150 u32 diff = A32(ref) ^ A32(ip);
151 #endif
152 if (!diff) {
153 ip += STEPSIZE;
154 ref += STEPSIZE;
155 continue;
156 }
157 ip += LZ4_NBCOMMONBYTES(diff);
158 goto _endcount;
159 }
160 #if LZ4_ARCH64
161 if ((ip < (MATCHLIMIT - 3)) && (A32(ref) == A32(ip))) {
162 ip += 4;
163 ref += 4;
164 }
165 #endif
166 if ((ip < (MATCHLIMIT - 1)) && (A16(ref) == A16(ip))) {
167 ip += 2;
168 ref += 2;
169 }
170 if ((ip < MATCHLIMIT) && (*ref == *ip))
171 ip++;
172_endcount:
173 /* Encode MatchLength */
174 length = (int)(ip - anchor);
175 /* Check output limit */
176 if (unlikely(op + (1 + LASTLITERALS) + (length >> 8) > oend))
177 return 0;
178 if (length >= (int)ML_MASK) {
179 *token += ML_MASK;
180 length -= ML_MASK;
181 for (; length > 509 ; length -= 510) {
182 *op++ = 255;
183 *op++ = 255;
184 }
185 if (length > 254) {
186 length -= 255;
187 *op++ = 255;
188 }
189 *op++ = (u8)length;
190 } else
191 *token += length;
192
193 /* Test end of chunk */
194 if (ip > mflimit) {
195 anchor = ip;
196 break;
197 }
198
199 /* Fill table */
200 hashtable[LZ4_HASH_VALUE(ip-2)] = ip - 2 - base;
201
202 /* Test next position */
203 ref = base + hashtable[LZ4_HASH_VALUE(ip)];
204 hashtable[LZ4_HASH_VALUE(ip)] = ip - base;
205 if ((ref > ip - (MAX_DISTANCE + 1)) && (A32(ref) == A32(ip))) {
206 token = op++;
207 *token = 0;
208 goto _next_match;
209 }
210
211 /* Prepare next loop */
212 anchor = ip++;
213 forwardh = LZ4_HASH_VALUE(ip);
214 }
215
216_last_literals:
217 /* Encode Last Literals */
218 lastrun = (int)(iend - anchor);
219 if (((char *)op - dest) + lastrun + 1
220 + ((lastrun + 255 - RUN_MASK) / 255) > (u32)maxoutputsize)
221 return 0;
222
223 if (lastrun >= (int)RUN_MASK) {
224 *op++ = (RUN_MASK << ML_BITS);
225 lastrun -= RUN_MASK;
226 for (; lastrun > 254 ; lastrun -= 255)
227 *op++ = 255;
228 *op++ = (u8)lastrun;
229 } else
230 *op++ = (lastrun << ML_BITS);
231 memcpy(op, anchor, iend - anchor);
232 op += iend - anchor;
233
234 /* End */
235 return (int)(((char *)op) - dest);
236}
237
238static inline int lz4_compress64kctx(void *ctx,
239 const char *source,
240 char *dest,
241 int isize,
242 int maxoutputsize)
243{
244 u16 *hashtable = (u16 *)ctx;
245 const u8 *ip = (u8 *) source;
246 const u8 *anchor = ip;
247 const u8 *const base = ip;
248 const u8 *const iend = ip + isize;
249 const u8 *const mflimit = iend - MFLIMIT;
250 #define MATCHLIMIT (iend - LASTLITERALS)
251
252 u8 *op = (u8 *) dest;
253 u8 *const oend = op + maxoutputsize;
254 int len, length;
255 const int skipstrength = SKIPSTRENGTH;
256 u32 forwardh;
257 int lastrun;
258
259 /* Init */
260 if (isize < MINLENGTH)
261 goto _last_literals;
262
263 memset((void *)hashtable, 0, LZ4_MEM_COMPRESS);
264
265 /* First Byte */
266 ip++;
267 forwardh = LZ4_HASH64K_VALUE(ip);
268
269 /* Main Loop */
270 for (;;) {
271 int findmatchattempts = (1U << skipstrength) + 3;
272 const u8 *forwardip = ip;
273 const u8 *ref;
274 u8 *token;
275
276 /* Find a match */
277 do {
278 u32 h = forwardh;
279 int step = findmatchattempts++ >> skipstrength;
280 ip = forwardip;
281 forwardip = ip + step;
282
283 if (forwardip > mflimit)
284 goto _last_literals;
285
286 forwardh = LZ4_HASH64K_VALUE(forwardip);
287 ref = base + hashtable[h];
288 hashtable[h] = (u16)(ip - base);
289 } while (A32(ref) != A32(ip));
290
291 /* Catch up */
292 while ((ip > anchor) && (ref > (u8 *)source)
293 && (ip[-1] == ref[-1])) {
294 ip--;
295 ref--;
296 }
297
298 /* Encode Literal length */
299 length = (int)(ip - anchor);
300 token = op++;
301 /* Check output limit */
302 if (unlikely(op + length + (2 + 1 + LASTLITERALS)
303 + (length >> 8) > oend))
304 return 0;
305 if (length >= (int)RUN_MASK) {
306 *token = (RUN_MASK << ML_BITS);
307 len = length - RUN_MASK;
308 for (; len > 254 ; len -= 255)
309 *op++ = 255;
310 *op++ = (u8)len;
311 } else
312 *token = (length << ML_BITS);
313
314 /* Copy Literals */
315 LZ4_BLINDCOPY(anchor, op, length);
316
317_next_match:
318 /* Encode Offset */
319 LZ4_WRITE_LITTLEENDIAN_16(op, (u16)(ip - ref));
320
321 /* Start Counting */
322 ip += MINMATCH;
323 /* MinMatch verified */
324 ref += MINMATCH;
325 anchor = ip;
326
327 while (ip < MATCHLIMIT - (STEPSIZE - 1)) {
328 #if LZ4_ARCH64
329 u64 diff = A64(ref) ^ A64(ip);
330 #else
331 u32 diff = A32(ref) ^ A32(ip);
332 #endif
333
334 if (!diff) {
335 ip += STEPSIZE;
336 ref += STEPSIZE;
337 continue;
338 }
339 ip += LZ4_NBCOMMONBYTES(diff);
340 goto _endcount;
341 }
342 #if LZ4_ARCH64
343 if ((ip < (MATCHLIMIT - 3)) && (A32(ref) == A32(ip))) {
344 ip += 4;
345 ref += 4;
346 }
347 #endif
348 if ((ip < (MATCHLIMIT - 1)) && (A16(ref) == A16(ip))) {
349 ip += 2;
350 ref += 2;
351 }
352 if ((ip < MATCHLIMIT) && (*ref == *ip))
353 ip++;
354_endcount:
355
356 /* Encode MatchLength */
357 len = (int)(ip - anchor);
358 /* Check output limit */
359 if (unlikely(op + (1 + LASTLITERALS) + (len >> 8) > oend))
360 return 0;
361 if (len >= (int)ML_MASK) {
362 *token += ML_MASK;
363 len -= ML_MASK;
364 for (; len > 509 ; len -= 510) {
365 *op++ = 255;
366 *op++ = 255;
367 }
368 if (len > 254) {
369 len -= 255;
370 *op++ = 255;
371 }
372 *op++ = (u8)len;
373 } else
374 *token += len;
375
376 /* Test end of chunk */
377 if (ip > mflimit) {
378 anchor = ip;
379 break;
380 }
381
382 /* Fill table */
383 hashtable[LZ4_HASH64K_VALUE(ip-2)] = (u16)(ip - 2 - base);
384
385 /* Test next position */
386 ref = base + hashtable[LZ4_HASH64K_VALUE(ip)];
387 hashtable[LZ4_HASH64K_VALUE(ip)] = (u16)(ip - base);
388 if (A32(ref) == A32(ip)) {
389 token = op++;
390 *token = 0;
391 goto _next_match;
392 }
393
394 /* Prepare next loop */
395 anchor = ip++;
396 forwardh = LZ4_HASH64K_VALUE(ip);
397 }
398
399_last_literals:
400 /* Encode Last Literals */
401 lastrun = (int)(iend - anchor);
402 if (op + lastrun + 1 + (lastrun - RUN_MASK + 255) / 255 > oend)
403 return 0;
404 if (lastrun >= (int)RUN_MASK) {
405 *op++ = (RUN_MASK << ML_BITS);
406 lastrun -= RUN_MASK;
407 for (; lastrun > 254 ; lastrun -= 255)
408 *op++ = 255;
409 *op++ = (u8)lastrun;
410 } else
411 *op++ = (lastrun << ML_BITS);
412 memcpy(op, anchor, iend - anchor);
413 op += iend - anchor;
414 /* End */
415 return (int)(((char *)op) - dest);
416}
417
418int lz4_compress(const unsigned char *src, size_t src_len,
419 unsigned char *dst, size_t *dst_len, void *wrkmem)
420{
421 int ret = -1;
422 int out_len = 0;
423
424 if (src_len < LZ4_64KLIMIT)
425 out_len = lz4_compress64kctx(wrkmem, src, dst, src_len,
426 lz4_compressbound(src_len));
427 else
428 out_len = lz4_compressctx(wrkmem, src, dst, src_len,
429 lz4_compressbound(src_len));
430
431 if (out_len < 0)
432 goto exit;
433
434 *dst_len = out_len;
435
436 return 0;
437exit:
438 return ret;
439}
440EXPORT_SYMBOL(lz4_compress);
441
442MODULE_LICENSE("Dual BSD/GPL");
443MODULE_DESCRIPTION("LZ4 compressor");
1/*
2 * LZ4 - Fast LZ compression algorithm
3 * Copyright (C) 2011 - 2016, Yann Collet.
4 * BSD 2 - Clause License (http://www.opensource.org/licenses/bsd - license.php)
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions are
7 * met:
8 * * Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * * Redistributions in binary form must reproduce the above
11 * copyright notice, this list of conditions and the following disclaimer
12 * in the documentation and/or other materials provided with the
13 * distribution.
14 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
15 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
16 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
17 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
18 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
19 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
20 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
21 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
22 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
23 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
24 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
25 * You can contact the author at :
26 * - LZ4 homepage : http://www.lz4.org
27 * - LZ4 source repository : https://github.com/lz4/lz4
28 *
29 * Changed for kernel usage by:
30 * Sven Schmidt <4sschmid@informatik.uni-hamburg.de>
31 */
32
33/*-************************************
34 * Dependencies
35 **************************************/
36#include <linux/lz4.h>
37#include "lz4defs.h"
38#include <linux/module.h>
39#include <linux/kernel.h>
40#include <linux/unaligned.h>
41
42static const int LZ4_minLength = (MFLIMIT + 1);
43static const int LZ4_64Klimit = ((64 * KB) + (MFLIMIT - 1));
44
45/*-******************************
46 * Compression functions
47 ********************************/
48static FORCE_INLINE U32 LZ4_hash4(
49 U32 sequence,
50 tableType_t const tableType)
51{
52 if (tableType == byU16)
53 return ((sequence * 2654435761U)
54 >> ((MINMATCH * 8) - (LZ4_HASHLOG + 1)));
55 else
56 return ((sequence * 2654435761U)
57 >> ((MINMATCH * 8) - LZ4_HASHLOG));
58}
59
60static FORCE_INLINE U32 LZ4_hash5(
61 U64 sequence,
62 tableType_t const tableType)
63{
64 const U32 hashLog = (tableType == byU16)
65 ? LZ4_HASHLOG + 1
66 : LZ4_HASHLOG;
67
68#if LZ4_LITTLE_ENDIAN
69 static const U64 prime5bytes = 889523592379ULL;
70
71 return (U32)(((sequence << 24) * prime5bytes) >> (64 - hashLog));
72#else
73 static const U64 prime8bytes = 11400714785074694791ULL;
74
75 return (U32)(((sequence >> 24) * prime8bytes) >> (64 - hashLog));
76#endif
77}
78
79static FORCE_INLINE U32 LZ4_hashPosition(
80 const void *p,
81 tableType_t const tableType)
82{
83#if LZ4_ARCH64
84 if (tableType == byU32)
85 return LZ4_hash5(LZ4_read_ARCH(p), tableType);
86#endif
87
88 return LZ4_hash4(LZ4_read32(p), tableType);
89}
90
91static void LZ4_putPositionOnHash(
92 const BYTE *p,
93 U32 h,
94 void *tableBase,
95 tableType_t const tableType,
96 const BYTE *srcBase)
97{
98 switch (tableType) {
99 case byPtr:
100 {
101 const BYTE **hashTable = (const BYTE **)tableBase;
102
103 hashTable[h] = p;
104 return;
105 }
106 case byU32:
107 {
108 U32 *hashTable = (U32 *) tableBase;
109
110 hashTable[h] = (U32)(p - srcBase);
111 return;
112 }
113 case byU16:
114 {
115 U16 *hashTable = (U16 *) tableBase;
116
117 hashTable[h] = (U16)(p - srcBase);
118 return;
119 }
120 }
121}
122
123static FORCE_INLINE void LZ4_putPosition(
124 const BYTE *p,
125 void *tableBase,
126 tableType_t tableType,
127 const BYTE *srcBase)
128{
129 U32 const h = LZ4_hashPosition(p, tableType);
130
131 LZ4_putPositionOnHash(p, h, tableBase, tableType, srcBase);
132}
133
134static const BYTE *LZ4_getPositionOnHash(
135 U32 h,
136 void *tableBase,
137 tableType_t tableType,
138 const BYTE *srcBase)
139{
140 if (tableType == byPtr) {
141 const BYTE **hashTable = (const BYTE **) tableBase;
142
143 return hashTable[h];
144 }
145
146 if (tableType == byU32) {
147 const U32 * const hashTable = (U32 *) tableBase;
148
149 return hashTable[h] + srcBase;
150 }
151
152 {
153 /* default, to ensure a return */
154 const U16 * const hashTable = (U16 *) tableBase;
155
156 return hashTable[h] + srcBase;
157 }
158}
159
160static FORCE_INLINE const BYTE *LZ4_getPosition(
161 const BYTE *p,
162 void *tableBase,
163 tableType_t tableType,
164 const BYTE *srcBase)
165{
166 U32 const h = LZ4_hashPosition(p, tableType);
167
168 return LZ4_getPositionOnHash(h, tableBase, tableType, srcBase);
169}
170
171
172/*
173 * LZ4_compress_generic() :
174 * inlined, to ensure branches are decided at compilation time
175 */
176static FORCE_INLINE int LZ4_compress_generic(
177 LZ4_stream_t_internal * const dictPtr,
178 const char * const source,
179 char * const dest,
180 const int inputSize,
181 const int maxOutputSize,
182 const limitedOutput_directive outputLimited,
183 const tableType_t tableType,
184 const dict_directive dict,
185 const dictIssue_directive dictIssue,
186 const U32 acceleration)
187{
188 const BYTE *ip = (const BYTE *) source;
189 const BYTE *base;
190 const BYTE *lowLimit;
191 const BYTE * const lowRefLimit = ip - dictPtr->dictSize;
192 const BYTE * const dictionary = dictPtr->dictionary;
193 const BYTE * const dictEnd = dictionary + dictPtr->dictSize;
194 const size_t dictDelta = dictEnd - (const BYTE *)source;
195 const BYTE *anchor = (const BYTE *) source;
196 const BYTE * const iend = ip + inputSize;
197 const BYTE * const mflimit = iend - MFLIMIT;
198 const BYTE * const matchlimit = iend - LASTLITERALS;
199
200 BYTE *op = (BYTE *) dest;
201 BYTE * const olimit = op + maxOutputSize;
202
203 U32 forwardH;
204 size_t refDelta = 0;
205
206 /* Init conditions */
207 if ((U32)inputSize > (U32)LZ4_MAX_INPUT_SIZE) {
208 /* Unsupported inputSize, too large (or negative) */
209 return 0;
210 }
211
212 switch (dict) {
213 case noDict:
214 default:
215 base = (const BYTE *)source;
216 lowLimit = (const BYTE *)source;
217 break;
218 case withPrefix64k:
219 base = (const BYTE *)source - dictPtr->currentOffset;
220 lowLimit = (const BYTE *)source - dictPtr->dictSize;
221 break;
222 case usingExtDict:
223 base = (const BYTE *)source - dictPtr->currentOffset;
224 lowLimit = (const BYTE *)source;
225 break;
226 }
227
228 if ((tableType == byU16)
229 && (inputSize >= LZ4_64Klimit)) {
230 /* Size too large (not within 64K limit) */
231 return 0;
232 }
233
234 if (inputSize < LZ4_minLength) {
235 /* Input too small, no compression (all literals) */
236 goto _last_literals;
237 }
238
239 /* First Byte */
240 LZ4_putPosition(ip, dictPtr->hashTable, tableType, base);
241 ip++;
242 forwardH = LZ4_hashPosition(ip, tableType);
243
244 /* Main Loop */
245 for ( ; ; ) {
246 const BYTE *match;
247 BYTE *token;
248
249 /* Find a match */
250 {
251 const BYTE *forwardIp = ip;
252 unsigned int step = 1;
253 unsigned int searchMatchNb = acceleration << LZ4_SKIPTRIGGER;
254
255 do {
256 U32 const h = forwardH;
257
258 ip = forwardIp;
259 forwardIp += step;
260 step = (searchMatchNb++ >> LZ4_SKIPTRIGGER);
261
262 if (unlikely(forwardIp > mflimit))
263 goto _last_literals;
264
265 match = LZ4_getPositionOnHash(h,
266 dictPtr->hashTable,
267 tableType, base);
268
269 if (dict == usingExtDict) {
270 if (match < (const BYTE *)source) {
271 refDelta = dictDelta;
272 lowLimit = dictionary;
273 } else {
274 refDelta = 0;
275 lowLimit = (const BYTE *)source;
276 } }
277
278 forwardH = LZ4_hashPosition(forwardIp,
279 tableType);
280
281 LZ4_putPositionOnHash(ip, h, dictPtr->hashTable,
282 tableType, base);
283 } while (((dictIssue == dictSmall)
284 ? (match < lowRefLimit)
285 : 0)
286 || ((tableType == byU16)
287 ? 0
288 : (match + MAX_DISTANCE < ip))
289 || (LZ4_read32(match + refDelta)
290 != LZ4_read32(ip)));
291 }
292
293 /* Catch up */
294 while (((ip > anchor) & (match + refDelta > lowLimit))
295 && (unlikely(ip[-1] == match[refDelta - 1]))) {
296 ip--;
297 match--;
298 }
299
300 /* Encode Literals */
301 {
302 unsigned const int litLength = (unsigned int)(ip - anchor);
303
304 token = op++;
305
306 if ((outputLimited) &&
307 /* Check output buffer overflow */
308 (unlikely(op + litLength +
309 (2 + 1 + LASTLITERALS) +
310 (litLength / 255) > olimit)))
311 return 0;
312
313 if (litLength >= RUN_MASK) {
314 int len = (int)litLength - RUN_MASK;
315
316 *token = (RUN_MASK << ML_BITS);
317
318 for (; len >= 255; len -= 255)
319 *op++ = 255;
320 *op++ = (BYTE)len;
321 } else
322 *token = (BYTE)(litLength << ML_BITS);
323
324 /* Copy Literals */
325 LZ4_wildCopy(op, anchor, op + litLength);
326 op += litLength;
327 }
328
329_next_match:
330 /* Encode Offset */
331 LZ4_writeLE16(op, (U16)(ip - match));
332 op += 2;
333
334 /* Encode MatchLength */
335 {
336 unsigned int matchCode;
337
338 if ((dict == usingExtDict)
339 && (lowLimit == dictionary)) {
340 const BYTE *limit;
341
342 match += refDelta;
343 limit = ip + (dictEnd - match);
344
345 if (limit > matchlimit)
346 limit = matchlimit;
347
348 matchCode = LZ4_count(ip + MINMATCH,
349 match + MINMATCH, limit);
350
351 ip += MINMATCH + matchCode;
352
353 if (ip == limit) {
354 unsigned const int more = LZ4_count(ip,
355 (const BYTE *)source,
356 matchlimit);
357
358 matchCode += more;
359 ip += more;
360 }
361 } else {
362 matchCode = LZ4_count(ip + MINMATCH,
363 match + MINMATCH, matchlimit);
364 ip += MINMATCH + matchCode;
365 }
366
367 if (outputLimited &&
368 /* Check output buffer overflow */
369 (unlikely(op +
370 (1 + LASTLITERALS) +
371 (matchCode >> 8) > olimit)))
372 return 0;
373
374 if (matchCode >= ML_MASK) {
375 *token += ML_MASK;
376 matchCode -= ML_MASK;
377 LZ4_write32(op, 0xFFFFFFFF);
378
379 while (matchCode >= 4 * 255) {
380 op += 4;
381 LZ4_write32(op, 0xFFFFFFFF);
382 matchCode -= 4 * 255;
383 }
384
385 op += matchCode / 255;
386 *op++ = (BYTE)(matchCode % 255);
387 } else
388 *token += (BYTE)(matchCode);
389 }
390
391 anchor = ip;
392
393 /* Test end of chunk */
394 if (ip > mflimit)
395 break;
396
397 /* Fill table */
398 LZ4_putPosition(ip - 2, dictPtr->hashTable, tableType, base);
399
400 /* Test next position */
401 match = LZ4_getPosition(ip, dictPtr->hashTable,
402 tableType, base);
403
404 if (dict == usingExtDict) {
405 if (match < (const BYTE *)source) {
406 refDelta = dictDelta;
407 lowLimit = dictionary;
408 } else {
409 refDelta = 0;
410 lowLimit = (const BYTE *)source;
411 }
412 }
413
414 LZ4_putPosition(ip, dictPtr->hashTable, tableType, base);
415
416 if (((dictIssue == dictSmall) ? (match >= lowRefLimit) : 1)
417 && (match + MAX_DISTANCE >= ip)
418 && (LZ4_read32(match + refDelta) == LZ4_read32(ip))) {
419 token = op++;
420 *token = 0;
421 goto _next_match;
422 }
423
424 /* Prepare next loop */
425 forwardH = LZ4_hashPosition(++ip, tableType);
426 }
427
428_last_literals:
429 /* Encode Last Literals */
430 {
431 size_t const lastRun = (size_t)(iend - anchor);
432
433 if ((outputLimited) &&
434 /* Check output buffer overflow */
435 ((op - (BYTE *)dest) + lastRun + 1 +
436 ((lastRun + 255 - RUN_MASK) / 255) > (U32)maxOutputSize))
437 return 0;
438
439 if (lastRun >= RUN_MASK) {
440 size_t accumulator = lastRun - RUN_MASK;
441 *op++ = RUN_MASK << ML_BITS;
442 for (; accumulator >= 255; accumulator -= 255)
443 *op++ = 255;
444 *op++ = (BYTE) accumulator;
445 } else {
446 *op++ = (BYTE)(lastRun << ML_BITS);
447 }
448
449 LZ4_memcpy(op, anchor, lastRun);
450
451 op += lastRun;
452 }
453
454 /* End */
455 return (int) (((char *)op) - dest);
456}
457
458static int LZ4_compress_fast_extState(
459 void *state,
460 const char *source,
461 char *dest,
462 int inputSize,
463 int maxOutputSize,
464 int acceleration)
465{
466 LZ4_stream_t_internal *ctx = &((LZ4_stream_t *)state)->internal_donotuse;
467#if LZ4_ARCH64
468 const tableType_t tableType = byU32;
469#else
470 const tableType_t tableType = byPtr;
471#endif
472
473 LZ4_resetStream((LZ4_stream_t *)state);
474
475 if (acceleration < 1)
476 acceleration = LZ4_ACCELERATION_DEFAULT;
477
478 if (maxOutputSize >= LZ4_COMPRESSBOUND(inputSize)) {
479 if (inputSize < LZ4_64Klimit)
480 return LZ4_compress_generic(ctx, source,
481 dest, inputSize, 0,
482 noLimit, byU16, noDict,
483 noDictIssue, acceleration);
484 else
485 return LZ4_compress_generic(ctx, source,
486 dest, inputSize, 0,
487 noLimit, tableType, noDict,
488 noDictIssue, acceleration);
489 } else {
490 if (inputSize < LZ4_64Klimit)
491 return LZ4_compress_generic(ctx, source,
492 dest, inputSize,
493 maxOutputSize, limitedOutput, byU16, noDict,
494 noDictIssue, acceleration);
495 else
496 return LZ4_compress_generic(ctx, source,
497 dest, inputSize,
498 maxOutputSize, limitedOutput, tableType, noDict,
499 noDictIssue, acceleration);
500 }
501}
502
503int LZ4_compress_fast(const char *source, char *dest, int inputSize,
504 int maxOutputSize, int acceleration, void *wrkmem)
505{
506 return LZ4_compress_fast_extState(wrkmem, source, dest, inputSize,
507 maxOutputSize, acceleration);
508}
509EXPORT_SYMBOL(LZ4_compress_fast);
510
511int LZ4_compress_default(const char *source, char *dest, int inputSize,
512 int maxOutputSize, void *wrkmem)
513{
514 return LZ4_compress_fast(source, dest, inputSize,
515 maxOutputSize, LZ4_ACCELERATION_DEFAULT, wrkmem);
516}
517EXPORT_SYMBOL(LZ4_compress_default);
518
519/*-******************************
520 * *_destSize() variant
521 ********************************/
522static int LZ4_compress_destSize_generic(
523 LZ4_stream_t_internal * const ctx,
524 const char * const src,
525 char * const dst,
526 int * const srcSizePtr,
527 const int targetDstSize,
528 const tableType_t tableType)
529{
530 const BYTE *ip = (const BYTE *) src;
531 const BYTE *base = (const BYTE *) src;
532 const BYTE *lowLimit = (const BYTE *) src;
533 const BYTE *anchor = ip;
534 const BYTE * const iend = ip + *srcSizePtr;
535 const BYTE * const mflimit = iend - MFLIMIT;
536 const BYTE * const matchlimit = iend - LASTLITERALS;
537
538 BYTE *op = (BYTE *) dst;
539 BYTE * const oend = op + targetDstSize;
540 BYTE * const oMaxLit = op + targetDstSize - 2 /* offset */
541 - 8 /* because 8 + MINMATCH == MFLIMIT */ - 1 /* token */;
542 BYTE * const oMaxMatch = op + targetDstSize
543 - (LASTLITERALS + 1 /* token */);
544 BYTE * const oMaxSeq = oMaxLit - 1 /* token */;
545
546 U32 forwardH;
547
548 /* Init conditions */
549 /* Impossible to store anything */
550 if (targetDstSize < 1)
551 return 0;
552 /* Unsupported input size, too large (or negative) */
553 if ((U32)*srcSizePtr > (U32)LZ4_MAX_INPUT_SIZE)
554 return 0;
555 /* Size too large (not within 64K limit) */
556 if ((tableType == byU16) && (*srcSizePtr >= LZ4_64Klimit))
557 return 0;
558 /* Input too small, no compression (all literals) */
559 if (*srcSizePtr < LZ4_minLength)
560 goto _last_literals;
561
562 /* First Byte */
563 *srcSizePtr = 0;
564 LZ4_putPosition(ip, ctx->hashTable, tableType, base);
565 ip++; forwardH = LZ4_hashPosition(ip, tableType);
566
567 /* Main Loop */
568 for ( ; ; ) {
569 const BYTE *match;
570 BYTE *token;
571
572 /* Find a match */
573 {
574 const BYTE *forwardIp = ip;
575 unsigned int step = 1;
576 unsigned int searchMatchNb = 1 << LZ4_SKIPTRIGGER;
577
578 do {
579 U32 h = forwardH;
580
581 ip = forwardIp;
582 forwardIp += step;
583 step = (searchMatchNb++ >> LZ4_SKIPTRIGGER);
584
585 if (unlikely(forwardIp > mflimit))
586 goto _last_literals;
587
588 match = LZ4_getPositionOnHash(h, ctx->hashTable,
589 tableType, base);
590 forwardH = LZ4_hashPosition(forwardIp,
591 tableType);
592 LZ4_putPositionOnHash(ip, h,
593 ctx->hashTable, tableType,
594 base);
595
596 } while (((tableType == byU16)
597 ? 0
598 : (match + MAX_DISTANCE < ip))
599 || (LZ4_read32(match) != LZ4_read32(ip)));
600 }
601
602 /* Catch up */
603 while ((ip > anchor)
604 && (match > lowLimit)
605 && (unlikely(ip[-1] == match[-1]))) {
606 ip--;
607 match--;
608 }
609
610 /* Encode Literal length */
611 {
612 unsigned int litLength = (unsigned int)(ip - anchor);
613
614 token = op++;
615 if (op + ((litLength + 240) / 255)
616 + litLength > oMaxLit) {
617 /* Not enough space for a last match */
618 op--;
619 goto _last_literals;
620 }
621 if (litLength >= RUN_MASK) {
622 unsigned int len = litLength - RUN_MASK;
623 *token = (RUN_MASK<<ML_BITS);
624 for (; len >= 255; len -= 255)
625 *op++ = 255;
626 *op++ = (BYTE)len;
627 } else
628 *token = (BYTE)(litLength << ML_BITS);
629
630 /* Copy Literals */
631 LZ4_wildCopy(op, anchor, op + litLength);
632 op += litLength;
633 }
634
635_next_match:
636 /* Encode Offset */
637 LZ4_writeLE16(op, (U16)(ip - match)); op += 2;
638
639 /* Encode MatchLength */
640 {
641 size_t matchLength = LZ4_count(ip + MINMATCH,
642 match + MINMATCH, matchlimit);
643
644 if (op + ((matchLength + 240)/255) > oMaxMatch) {
645 /* Match description too long : reduce it */
646 matchLength = (15 - 1) + (oMaxMatch - op) * 255;
647 }
648 ip += MINMATCH + matchLength;
649
650 if (matchLength >= ML_MASK) {
651 *token += ML_MASK;
652 matchLength -= ML_MASK;
653 while (matchLength >= 255) {
654 matchLength -= 255;
655 *op++ = 255;
656 }
657 *op++ = (BYTE)matchLength;
658 } else
659 *token += (BYTE)(matchLength);
660 }
661
662 anchor = ip;
663
664 /* Test end of block */
665 if (ip > mflimit)
666 break;
667 if (op > oMaxSeq)
668 break;
669
670 /* Fill table */
671 LZ4_putPosition(ip - 2, ctx->hashTable, tableType, base);
672
673 /* Test next position */
674 match = LZ4_getPosition(ip, ctx->hashTable, tableType, base);
675 LZ4_putPosition(ip, ctx->hashTable, tableType, base);
676
677 if ((match + MAX_DISTANCE >= ip)
678 && (LZ4_read32(match) == LZ4_read32(ip))) {
679 token = op++; *token = 0;
680 goto _next_match;
681 }
682
683 /* Prepare next loop */
684 forwardH = LZ4_hashPosition(++ip, tableType);
685 }
686
687_last_literals:
688 /* Encode Last Literals */
689 {
690 size_t lastRunSize = (size_t)(iend - anchor);
691
692 if (op + 1 /* token */
693 + ((lastRunSize + 240) / 255) /* litLength */
694 + lastRunSize /* literals */ > oend) {
695 /* adapt lastRunSize to fill 'dst' */
696 lastRunSize = (oend - op) - 1;
697 lastRunSize -= (lastRunSize + 240) / 255;
698 }
699 ip = anchor + lastRunSize;
700
701 if (lastRunSize >= RUN_MASK) {
702 size_t accumulator = lastRunSize - RUN_MASK;
703
704 *op++ = RUN_MASK << ML_BITS;
705 for (; accumulator >= 255; accumulator -= 255)
706 *op++ = 255;
707 *op++ = (BYTE) accumulator;
708 } else {
709 *op++ = (BYTE)(lastRunSize<<ML_BITS);
710 }
711 LZ4_memcpy(op, anchor, lastRunSize);
712 op += lastRunSize;
713 }
714
715 /* End */
716 *srcSizePtr = (int) (((const char *)ip) - src);
717 return (int) (((char *)op) - dst);
718}
719
720static int LZ4_compress_destSize_extState(
721 LZ4_stream_t *state,
722 const char *src,
723 char *dst,
724 int *srcSizePtr,
725 int targetDstSize)
726{
727#if LZ4_ARCH64
728 const tableType_t tableType = byU32;
729#else
730 const tableType_t tableType = byPtr;
731#endif
732
733 LZ4_resetStream(state);
734
735 if (targetDstSize >= LZ4_COMPRESSBOUND(*srcSizePtr)) {
736 /* compression success is guaranteed */
737 return LZ4_compress_fast_extState(
738 state, src, dst, *srcSizePtr,
739 targetDstSize, 1);
740 } else {
741 if (*srcSizePtr < LZ4_64Klimit)
742 return LZ4_compress_destSize_generic(
743 &state->internal_donotuse,
744 src, dst, srcSizePtr,
745 targetDstSize, byU16);
746 else
747 return LZ4_compress_destSize_generic(
748 &state->internal_donotuse,
749 src, dst, srcSizePtr,
750 targetDstSize, tableType);
751 }
752}
753
754
755int LZ4_compress_destSize(
756 const char *src,
757 char *dst,
758 int *srcSizePtr,
759 int targetDstSize,
760 void *wrkmem)
761{
762 return LZ4_compress_destSize_extState(wrkmem, src, dst, srcSizePtr,
763 targetDstSize);
764}
765EXPORT_SYMBOL(LZ4_compress_destSize);
766
767/*-******************************
768 * Streaming functions
769 ********************************/
770void LZ4_resetStream(LZ4_stream_t *LZ4_stream)
771{
772 memset(LZ4_stream, 0, sizeof(LZ4_stream_t));
773}
774
775int LZ4_loadDict(LZ4_stream_t *LZ4_dict,
776 const char *dictionary, int dictSize)
777{
778 LZ4_stream_t_internal *dict = &LZ4_dict->internal_donotuse;
779 const BYTE *p = (const BYTE *)dictionary;
780 const BYTE * const dictEnd = p + dictSize;
781 const BYTE *base;
782
783 if ((dict->initCheck)
784 || (dict->currentOffset > 1 * GB)) {
785 /* Uninitialized structure, or reuse overflow */
786 LZ4_resetStream(LZ4_dict);
787 }
788
789 if (dictSize < (int)HASH_UNIT) {
790 dict->dictionary = NULL;
791 dict->dictSize = 0;
792 return 0;
793 }
794
795 if ((dictEnd - p) > 64 * KB)
796 p = dictEnd - 64 * KB;
797 dict->currentOffset += 64 * KB;
798 base = p - dict->currentOffset;
799 dict->dictionary = p;
800 dict->dictSize = (U32)(dictEnd - p);
801 dict->currentOffset += dict->dictSize;
802
803 while (p <= dictEnd - HASH_UNIT) {
804 LZ4_putPosition(p, dict->hashTable, byU32, base);
805 p += 3;
806 }
807
808 return dict->dictSize;
809}
810EXPORT_SYMBOL(LZ4_loadDict);
811
812static void LZ4_renormDictT(LZ4_stream_t_internal *LZ4_dict,
813 const BYTE *src)
814{
815 if ((LZ4_dict->currentOffset > 0x80000000) ||
816 ((uptrval)LZ4_dict->currentOffset > (uptrval)src)) {
817 /* address space overflow */
818 /* rescale hash table */
819 U32 const delta = LZ4_dict->currentOffset - 64 * KB;
820 const BYTE *dictEnd = LZ4_dict->dictionary + LZ4_dict->dictSize;
821 int i;
822
823 for (i = 0; i < LZ4_HASH_SIZE_U32; i++) {
824 if (LZ4_dict->hashTable[i] < delta)
825 LZ4_dict->hashTable[i] = 0;
826 else
827 LZ4_dict->hashTable[i] -= delta;
828 }
829 LZ4_dict->currentOffset = 64 * KB;
830 if (LZ4_dict->dictSize > 64 * KB)
831 LZ4_dict->dictSize = 64 * KB;
832 LZ4_dict->dictionary = dictEnd - LZ4_dict->dictSize;
833 }
834}
835
836int LZ4_saveDict(LZ4_stream_t *LZ4_dict, char *safeBuffer, int dictSize)
837{
838 LZ4_stream_t_internal * const dict = &LZ4_dict->internal_donotuse;
839 const BYTE * const previousDictEnd = dict->dictionary + dict->dictSize;
840
841 if ((U32)dictSize > 64 * KB) {
842 /* useless to define a dictionary > 64 * KB */
843 dictSize = 64 * KB;
844 }
845 if ((U32)dictSize > dict->dictSize)
846 dictSize = dict->dictSize;
847
848 memmove(safeBuffer, previousDictEnd - dictSize, dictSize);
849
850 dict->dictionary = (const BYTE *)safeBuffer;
851 dict->dictSize = (U32)dictSize;
852
853 return dictSize;
854}
855EXPORT_SYMBOL(LZ4_saveDict);
856
857int LZ4_compress_fast_continue(LZ4_stream_t *LZ4_stream, const char *source,
858 char *dest, int inputSize, int maxOutputSize, int acceleration)
859{
860 LZ4_stream_t_internal *streamPtr = &LZ4_stream->internal_donotuse;
861 const BYTE * const dictEnd = streamPtr->dictionary
862 + streamPtr->dictSize;
863
864 const BYTE *smallest = (const BYTE *) source;
865
866 if (streamPtr->initCheck) {
867 /* Uninitialized structure detected */
868 return 0;
869 }
870
871 if ((streamPtr->dictSize > 0) && (smallest > dictEnd))
872 smallest = dictEnd;
873
874 LZ4_renormDictT(streamPtr, smallest);
875
876 if (acceleration < 1)
877 acceleration = LZ4_ACCELERATION_DEFAULT;
878
879 /* Check overlapping input/dictionary space */
880 {
881 const BYTE *sourceEnd = (const BYTE *) source + inputSize;
882
883 if ((sourceEnd > streamPtr->dictionary)
884 && (sourceEnd < dictEnd)) {
885 streamPtr->dictSize = (U32)(dictEnd - sourceEnd);
886 if (streamPtr->dictSize > 64 * KB)
887 streamPtr->dictSize = 64 * KB;
888 if (streamPtr->dictSize < 4)
889 streamPtr->dictSize = 0;
890 streamPtr->dictionary = dictEnd - streamPtr->dictSize;
891 }
892 }
893
894 /* prefix mode : source data follows dictionary */
895 if (dictEnd == (const BYTE *)source) {
896 int result;
897
898 if ((streamPtr->dictSize < 64 * KB) &&
899 (streamPtr->dictSize < streamPtr->currentOffset)) {
900 result = LZ4_compress_generic(
901 streamPtr, source, dest, inputSize,
902 maxOutputSize, limitedOutput, byU32,
903 withPrefix64k, dictSmall, acceleration);
904 } else {
905 result = LZ4_compress_generic(
906 streamPtr, source, dest, inputSize,
907 maxOutputSize, limitedOutput, byU32,
908 withPrefix64k, noDictIssue, acceleration);
909 }
910 streamPtr->dictSize += (U32)inputSize;
911 streamPtr->currentOffset += (U32)inputSize;
912 return result;
913 }
914
915 /* external dictionary mode */
916 {
917 int result;
918
919 if ((streamPtr->dictSize < 64 * KB) &&
920 (streamPtr->dictSize < streamPtr->currentOffset)) {
921 result = LZ4_compress_generic(
922 streamPtr, source, dest, inputSize,
923 maxOutputSize, limitedOutput, byU32,
924 usingExtDict, dictSmall, acceleration);
925 } else {
926 result = LZ4_compress_generic(
927 streamPtr, source, dest, inputSize,
928 maxOutputSize, limitedOutput, byU32,
929 usingExtDict, noDictIssue, acceleration);
930 }
931 streamPtr->dictionary = (const BYTE *)source;
932 streamPtr->dictSize = (U32)inputSize;
933 streamPtr->currentOffset += (U32)inputSize;
934 return result;
935 }
936}
937EXPORT_SYMBOL(LZ4_compress_fast_continue);
938
939MODULE_LICENSE("Dual BSD/GPL");
940MODULE_DESCRIPTION("LZ4 compressor");