1/* 2 * LZ4 - Fast LZ compression algorithm 3 * Header File 4 * Copyright (C) 2011-2013, Yann Collet. 5 * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions are 9 * met: 10 * 11 * * Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * * Redistributions in binary form must reproduce the above 14 * copyright notice, this list of conditions and the following disclaimer 15 * in the documentation and/or other materials provided with the 16 * distribution. 17 * 18 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 19 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 20 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 21 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 22 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 23 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 24 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 25 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 26 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 27 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 28 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 29 * 30 * You can contact the author at : 31 * - LZ4 homepage : http://fastcompression.blogspot.com/p/lz4.html 32 * - LZ4 source repository : http://code.google.com/p/lz4/ 33 * 34 * $FreeBSD$ 35 */ 36 37static int LZ4_uncompress_unknownOutputSize(const char *source, char *dest, 38 int isize, int maxOutputSize); 39 40/* ARGSUSED */ 41static int 42lz4_decompress(void *s_start, void *d_start, size_t s_len, size_t d_len, int dummy __unused) 43{ 44 const uint8_t *src = s_start; 45 uint32_t bufsiz = htonl(*(uint32_t *)src); 46 47 /* invalid compressed buffer size encoded at start */ 48 if (bufsiz + 4 > s_len) 49 return (1); 50 51 /* 52 * Returns 0 on success (decompression function returned non-negative) 53 * and non-zero on failure (decompression function returned negative). 54 */ 55 return (LZ4_uncompress_unknownOutputSize(s_start + 4, d_start, bufsiz, 56 d_len) < 0); 57} 58 59/* 60 * CPU Feature Detection 61 */ 62 63/* 32 or 64 bits ? */ 64#if (defined(__x86_64__) || defined(__x86_64) || defined(__amd64__) || \ 65 defined(__amd64) || defined(__ppc64__) || defined(_WIN64) || \ 66 defined(__LP64__) || defined(_LP64)) 67#define LZ4_ARCH64 1 68#else 69#define LZ4_ARCH64 0 70#endif 71 72/* 73 * Little Endian or Big Endian? 74 * Note: overwrite the below #define if you know your architecture endianess. 75 */ 76#if BYTE_ORDER == BIG_ENDIAN 77#define LZ4_BIG_ENDIAN 1 78#else 79 /* 80 * Little Endian assumed. PDP Endian and other very rare endian format 81 * are unsupported. 82 */ 83#endif 84 85/* 86 * Compiler Options 87 */ 88#if __STDC_VERSION__ >= 199901L /* C99 */ 89/* "restrict" is a known keyword */ 90#else 91/* Disable restrict */ 92#define restrict 93#endif 94 95#define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) 96 97#define lz4_bswap16(x) ((unsigned short int) ((((x) >> 8) & 0xffu) \ 98 | (((x) & 0xffu) << 8))) 99 100#if (GCC_VERSION >= 302) || (__INTEL_COMPILER >= 800) || defined(__clang__) 101#define expect(expr, value) (__builtin_expect((expr), (value))) 102#else 103#define expect(expr, value) (expr) 104#endif 105 106#define likely(expr) expect((expr) != 0, 1) 107#define unlikely(expr) expect((expr) != 0, 0) 108 109/* Basic types */ 110#define BYTE uint8_t 111#define U16 uint16_t 112#define U32 uint32_t 113#define S32 int32_t 114#define U64 uint64_t 115 116typedef struct _U16_S { 117 U16 v; 118} U16_S; 119typedef struct _U32_S { 120 U32 v; 121} U32_S; 122typedef struct _U64_S { 123 U64 v; 124} U64_S; 125 126#define A64(x) (((U64_S *)(x))->v) 127#define A32(x) (((U32_S *)(x))->v) 128#define A16(x) (((U16_S *)(x))->v) 129 130/* 131 * Constants 132 */ 133#define MINMATCH 4 134 135#define COPYLENGTH 8 136#define LASTLITERALS 5 137 138#define ML_BITS 4 139#define ML_MASK ((1U<<ML_BITS)-1) 140#define RUN_BITS (8-ML_BITS) 141#define RUN_MASK ((1U<<RUN_BITS)-1) 142 143/* 144 * Architecture-specific macros 145 */ 146#if LZ4_ARCH64 147#define STEPSIZE 8 148#define UARCH U64 149#define AARCH A64 150#define LZ4_COPYSTEP(s, d) A64(d) = A64(s); d += 8; s += 8; 151#define LZ4_COPYPACKET(s, d) LZ4_COPYSTEP(s, d) 152#define LZ4_SECURECOPY(s, d, e) if (d < e) LZ4_WILDCOPY(s, d, e) 153#define HTYPE U32 154#define INITBASE(base) const BYTE* const base = ip 155#else 156#define STEPSIZE 4 157#define UARCH U32 158#define AARCH A32 159#define LZ4_COPYSTEP(s, d) A32(d) = A32(s); d += 4; s += 4; 160#define LZ4_COPYPACKET(s, d) LZ4_COPYSTEP(s, d); LZ4_COPYSTEP(s, d); 161#define LZ4_SECURECOPY LZ4_WILDCOPY 162#define HTYPE const BYTE* 163#define INITBASE(base) const int base = 0 164#endif 165 166#if (defined(LZ4_BIG_ENDIAN) && !defined(BIG_ENDIAN_NATIVE_BUT_INCOMPATIBLE)) 167#define LZ4_READ_LITTLEENDIAN_16(d, s, p) \ 168 { U16 v = A16(p); v = lz4_bswap16(v); d = (s) - v; } 169#define LZ4_WRITE_LITTLEENDIAN_16(p, i) \ 170 { U16 v = (U16)(i); v = lz4_bswap16(v); A16(p) = v; p += 2; } 171#else 172#define LZ4_READ_LITTLEENDIAN_16(d, s, p) { d = (s) - A16(p); } 173#define LZ4_WRITE_LITTLEENDIAN_16(p, v) { A16(p) = v; p += 2; } 174#endif 175 176/* Macros */ 177#define LZ4_WILDCOPY(s, d, e) do { LZ4_COPYPACKET(s, d) } while (d < e); 178 179/* Decompression functions */ 180 181static int 182LZ4_uncompress_unknownOutputSize(const char *source, 183 char *dest, int isize, int maxOutputSize) 184{ 185 /* Local Variables */ 186 const BYTE *restrict ip = (const BYTE *) source; 187 const BYTE *const iend = ip + isize; 188 const BYTE *restrict ref; 189 190 BYTE *restrict op = (BYTE *) dest; 191 BYTE *const oend = op + maxOutputSize; 192 BYTE *cpy; 193 194 size_t dec[] = { 0, 3, 2, 3, 0, 0, 0, 0 }; 195 196 /* Main Loop */ 197 while (ip < iend) { 198 BYTE token; 199 int length; 200 201 /* get runlength */ 202 token = *ip++; 203 if ((length = (token >> ML_BITS)) == RUN_MASK) { 204 int s = 255; 205 while ((ip < iend) && (s == 255)) { 206 s = *ip++; 207 length += s; 208 } 209 } 210 /* copy literals */ 211 cpy = op + length; 212 if ((cpy > oend - COPYLENGTH) || 213 (ip + length > iend - COPYLENGTH)) { 214 if (cpy > oend) 215 /* 216 * Error: request to write beyond destination 217 * buffer. 218 */ 219 goto _output_error; 220 if (ip + length > iend) 221 /* 222 * Error : request to read beyond source 223 * buffer. 224 */ 225 goto _output_error; 226 memcpy(op, ip, length); 227 op += length; 228 ip += length; 229 if (ip < iend) 230 /* Error : LZ4 format violation */ 231 goto _output_error; 232 /* Necessarily EOF, due to parsing restrictions. */ 233 break; 234 } 235 LZ4_WILDCOPY(ip, op, cpy); 236 ip -= (op - cpy); 237 op = cpy; 238 239 /* get offset */ 240 LZ4_READ_LITTLEENDIAN_16(ref, cpy, ip); 241 ip += 2; 242 if (ref < (BYTE * const) dest) 243 /* 244 * Error: offset creates reference outside of 245 * destination buffer. 246 */ 247 goto _output_error; 248 249 /* get matchlength */ 250 if ((length = (token & ML_MASK)) == ML_MASK) { 251 while (ip < iend) { 252 int s = *ip++; 253 length += s; 254 if (s == 255) 255 continue; 256 break; 257 } 258 } 259 /* copy repeated sequence */ 260 if unlikely(op - ref < STEPSIZE) { 261#if LZ4_ARCH64 262 size_t dec2table[] = { 0, 0, 0, -1, 0, 1, 2, 3 }; 263 size_t dec2 = dec2table[op - ref]; 264#else 265 const int dec2 = 0; 266#endif 267 *op++ = *ref++; 268 *op++ = *ref++; 269 *op++ = *ref++; 270 *op++ = *ref++; 271 ref -= dec[op - ref]; 272 A32(op) = A32(ref); 273 op += STEPSIZE - 4; 274 ref -= dec2; 275 } else { 276 LZ4_COPYSTEP(ref, op); 277 } 278 cpy = op + length - (STEPSIZE - 4); 279 if (cpy > oend - COPYLENGTH) { 280 if (cpy > oend) 281 /* 282 * Error: request to write outside of 283 * destination buffer. 284 */ 285 goto _output_error; 286 LZ4_SECURECOPY(ref, op, (oend - COPYLENGTH)); 287 while (op < cpy) 288 *op++ = *ref++; 289 op = cpy; 290 if (op == oend) 291 /* 292 * Check EOF (should never happen, since last 293 * 5 bytes are supposed to be literals). 294 */ 295 break; 296 continue; 297 } 298 LZ4_SECURECOPY(ref, op, cpy); 299 op = cpy; /* correction */ 300 } 301 302 /* end of decoding */ 303 return (int)(((char *)op) - dest); 304 305 /* write overflow error detected */ 306 _output_error: 307 return (int)(-(((char *)ip) - source)); 308} 309