1 /*- 2 * Copyright (c) 2003-2007 Tim Kientzle 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 15 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 16 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 17 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 18 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 19 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 20 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 21 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 23 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24 */ 25 26 /* 27 * This code borrows heavily from "compress" source code, which is 28 * protected by the following copyright. (Clause 3 dropped by request 29 * of the Regents.) 30 */ 31 32 /*- 33 * Copyright (c) 1985, 1986, 1992, 1993 34 * The Regents of the University of California. All rights reserved. 35 * 36 * This code is derived from software contributed to Berkeley by 37 * Diomidis Spinellis and James A. Woods, derived from original 38 * work by Spencer Thomas and Joseph Orost. 39 * 40 * Redistribution and use in source and binary forms, with or without 41 * modification, are permitted provided that the following conditions 42 * are met: 43 * 1. Redistributions of source code must retain the above copyright 44 * notice, this list of conditions and the following disclaimer. 45 * 2. Redistributions in binary form must reproduce the above copyright 46 * notice, this list of conditions and the following disclaimer in the 47 * documentation and/or other materials provided with the distribution. 48 * 4. Neither the name of the University nor the names of its contributors 49 * may be used to endorse or promote products derived from this software 50 * without specific prior written permission. 51 * 52 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 53 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 54 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 55 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 56 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 57 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 58 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 59 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 60 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 61 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 62 * SUCH DAMAGE. 63 */ 64 65 66 #include "archive_platform.h" 67 __FBSDID("$FreeBSD$"); 68 69 #ifdef HAVE_ERRNO_H 70 #include <errno.h> 71 #endif 72 #ifdef HAVE_STDLIB_H 73 #include <stdlib.h> 74 #endif 75 #ifdef HAVE_STRING_H 76 #include <string.h> 77 #endif 78 #ifdef HAVE_UNISTD_H 79 #include <unistd.h> 80 #endif 81 82 #include "archive.h" 83 #include "archive_private.h" 84 #include "archive_read_private.h" 85 86 /* 87 * Because LZW decompression is pretty simple, I've just implemented 88 * the whole decompressor here (cribbing from "compress" source code, 89 * of course), rather than relying on an external library. I have 90 * made an effort to clarify and simplify the algorithm, so the 91 * names and structure here don't exactly match those used by compress. 92 */ 93 94 struct private_data { 95 /* Input variables. */ 96 const unsigned char *next_in; 97 size_t avail_in; 98 size_t consume_unnotified; 99 int bit_buffer; 100 int bits_avail; 101 size_t bytes_in_section; 102 103 /* Output variables. */ 104 size_t out_block_size; 105 void *out_block; 106 107 /* Decompression status variables. */ 108 int use_reset_code; 109 int end_of_stream; /* EOF status. */ 110 int maxcode; /* Largest code. */ 111 int maxcode_bits; /* Length of largest code. */ 112 int section_end_code; /* When to increase bits. */ 113 int bits; /* Current code length. */ 114 int oldcode; /* Previous code. */ 115 int finbyte; /* Last byte of prev code. */ 116 117 /* Dictionary. */ 118 int free_ent; /* Next dictionary entry. */ 119 unsigned char suffix[65536]; 120 uint16_t prefix[65536]; 121 122 /* 123 * Scratch area for expanding dictionary entries. Note: 124 * "worst" case here comes from compressing /dev/zero: the 125 * last code in the dictionary will code a sequence of 126 * 65536-256 zero bytes. Thus, we need stack space to expand 127 * a 65280-byte dictionary entry. (Of course, 32640:1 128 * compression could also be considered the "best" case. ;-) 129 */ 130 unsigned char *stackp; 131 unsigned char stack[65300]; 132 }; 133 134 static int compress_bidder_bid(struct archive_read_filter_bidder *, struct archive_read_filter *); 135 static int compress_bidder_init(struct archive_read_filter *); 136 137 static ssize_t compress_filter_read(struct archive_read_filter *, const void **); 138 static int compress_filter_close(struct archive_read_filter *); 139 140 static int getbits(struct archive_read_filter *, int n); 141 static int next_code(struct archive_read_filter *); 142 143 #if ARCHIVE_VERSION_NUMBER < 4000000 144 /* Deprecated; remove in libarchive 4.0 */ 145 int 146 archive_read_support_compression_compress(struct archive *a) 147 { 148 return archive_read_support_filter_compress(a); 149 } 150 #endif 151 152 static const struct archive_read_filter_bidder_vtable 153 compress_bidder_vtable = { 154 .bid = compress_bidder_bid, 155 .init = compress_bidder_init, 156 }; 157 158 int 159 archive_read_support_filter_compress(struct archive *_a) 160 { 161 struct archive_read *a = (struct archive_read *)_a; 162 163 return __archive_read_register_bidder(a, NULL, "compress (.Z)", 164 &compress_bidder_vtable); 165 } 166 167 /* 168 * Test whether we can handle this data. 169 * This logic returns zero if any part of the signature fails. 170 */ 171 static int 172 compress_bidder_bid(struct archive_read_filter_bidder *self, 173 struct archive_read_filter *filter) 174 { 175 const unsigned char *buffer; 176 ssize_t avail; 177 int bits_checked; 178 179 (void)self; /* UNUSED */ 180 181 /* Shortest valid compress file is 3 bytes. */ 182 buffer = __archive_read_filter_ahead(filter, 3, &avail); 183 184 if (buffer == NULL) 185 return (0); 186 187 bits_checked = 0; 188 /* First two bytes are the magic value */ 189 if (buffer[0] != 0x1F || buffer[1] != 0x9D) 190 return (0); 191 /* Third byte holds compression parameters. */ 192 if (buffer[2] & 0x20) /* Reserved bit, must be zero. */ 193 return (0); 194 if (buffer[2] & 0x40) /* Reserved bit, must be zero. */ 195 return (0); 196 bits_checked += 18; 197 198 return (bits_checked); 199 } 200 201 static const struct archive_read_filter_vtable 202 compress_reader_vtable = { 203 .read = compress_filter_read, 204 .close = compress_filter_close, 205 }; 206 207 /* 208 * Setup the callbacks. 209 */ 210 static int 211 compress_bidder_init(struct archive_read_filter *self) 212 { 213 struct private_data *state; 214 static const size_t out_block_size = 64 * 1024; 215 void *out_block; 216 int code; 217 218 self->code = ARCHIVE_FILTER_COMPRESS; 219 self->name = "compress (.Z)"; 220 221 state = (struct private_data *)calloc(sizeof(*state), 1); 222 out_block = malloc(out_block_size); 223 if (state == NULL || out_block == NULL) { 224 free(out_block); 225 free(state); 226 archive_set_error(&self->archive->archive, ENOMEM, 227 "Can't allocate data for %s decompression", 228 self->name); 229 return (ARCHIVE_FATAL); 230 } 231 232 self->data = state; 233 state->out_block_size = out_block_size; 234 state->out_block = out_block; 235 self->vtable = &compress_reader_vtable; 236 237 /* XXX MOVE THE FOLLOWING OUT OF INIT() XXX */ 238 239 (void)getbits(self, 8); /* Skip first signature byte. */ 240 (void)getbits(self, 8); /* Skip second signature byte. */ 241 242 /* Get compression parameters. */ 243 code = getbits(self, 8); 244 if ((code & 0x1f) > 16) { 245 archive_set_error(&self->archive->archive, -1, 246 "Invalid compressed data"); 247 return (ARCHIVE_FATAL); 248 } 249 state->maxcode_bits = code & 0x1f; 250 state->maxcode = (1 << state->maxcode_bits); 251 state->use_reset_code = code & 0x80; 252 253 /* Initialize decompressor. */ 254 state->free_ent = 256; 255 state->stackp = state->stack; 256 if (state->use_reset_code) 257 state->free_ent++; 258 state->bits = 9; 259 state->section_end_code = (1<<state->bits) - 1; 260 state->oldcode = -1; 261 for (code = 255; code >= 0; code--) { 262 state->prefix[code] = 0; 263 state->suffix[code] = code; 264 } 265 next_code(self); 266 267 return (ARCHIVE_OK); 268 } 269 270 /* 271 * Return a block of data from the decompression buffer. Decompress more 272 * as necessary. 273 */ 274 static ssize_t 275 compress_filter_read(struct archive_read_filter *self, const void **pblock) 276 { 277 struct private_data *state; 278 unsigned char *p, *start, *end; 279 int ret; 280 281 state = (struct private_data *)self->data; 282 if (state->end_of_stream) { 283 *pblock = NULL; 284 return (0); 285 } 286 p = start = (unsigned char *)state->out_block; 287 end = start + state->out_block_size; 288 289 while (p < end && !state->end_of_stream) { 290 if (state->stackp > state->stack) { 291 *p++ = *--state->stackp; 292 } else { 293 ret = next_code(self); 294 if (ret == -1) 295 state->end_of_stream = ret; 296 else if (ret != ARCHIVE_OK) 297 return (ret); 298 } 299 } 300 301 *pblock = start; 302 return (p - start); 303 } 304 305 /* 306 * Close and release the filter. 307 */ 308 static int 309 compress_filter_close(struct archive_read_filter *self) 310 { 311 struct private_data *state = (struct private_data *)self->data; 312 313 free(state->out_block); 314 free(state); 315 return (ARCHIVE_OK); 316 } 317 318 /* 319 * Process the next code and fill the stack with the expansion 320 * of the code. Returns ARCHIVE_FATAL if there is a fatal I/O or 321 * format error, ARCHIVE_EOF if we hit end of data, ARCHIVE_OK otherwise. 322 */ 323 static int 324 next_code(struct archive_read_filter *self) 325 { 326 struct private_data *state = (struct private_data *)self->data; 327 int code, newcode; 328 329 static int debug_buff[1024]; 330 static unsigned debug_index; 331 332 code = newcode = getbits(self, state->bits); 333 if (code < 0) 334 return (code); 335 336 debug_buff[debug_index++] = code; 337 if (debug_index >= sizeof(debug_buff)/sizeof(debug_buff[0])) 338 debug_index = 0; 339 340 /* If it's a reset code, reset the dictionary. */ 341 if ((code == 256) && state->use_reset_code) { 342 /* 343 * The original 'compress' implementation blocked its 344 * I/O in a manner that resulted in junk bytes being 345 * inserted after every reset. The next section skips 346 * this junk. (Yes, the number of *bytes* to skip is 347 * a function of the current *bit* length.) 348 */ 349 int skip_bytes = state->bits - 350 (state->bytes_in_section % state->bits); 351 skip_bytes %= state->bits; 352 state->bits_avail = 0; /* Discard rest of this byte. */ 353 while (skip_bytes-- > 0) { 354 code = getbits(self, 8); 355 if (code < 0) 356 return (code); 357 } 358 /* Now, actually do the reset. */ 359 state->bytes_in_section = 0; 360 state->bits = 9; 361 state->section_end_code = (1 << state->bits) - 1; 362 state->free_ent = 257; 363 state->oldcode = -1; 364 return (next_code(self)); 365 } 366 367 if (code > state->free_ent 368 || (code == state->free_ent && state->oldcode < 0)) { 369 /* An invalid code is a fatal error. */ 370 archive_set_error(&(self->archive->archive), -1, 371 "Invalid compressed data"); 372 return (ARCHIVE_FATAL); 373 } 374 375 /* Special case for KwKwK string. */ 376 if (code >= state->free_ent) { 377 *state->stackp++ = state->finbyte; 378 code = state->oldcode; 379 } 380 381 /* Generate output characters in reverse order. */ 382 while (code >= 256) { 383 *state->stackp++ = state->suffix[code]; 384 code = state->prefix[code]; 385 } 386 *state->stackp++ = state->finbyte = code; 387 388 /* Generate the new entry. */ 389 code = state->free_ent; 390 if (code < state->maxcode && state->oldcode >= 0) { 391 state->prefix[code] = state->oldcode; 392 state->suffix[code] = state->finbyte; 393 ++state->free_ent; 394 } 395 if (state->free_ent > state->section_end_code) { 396 state->bits++; 397 state->bytes_in_section = 0; 398 if (state->bits == state->maxcode_bits) 399 state->section_end_code = state->maxcode; 400 else 401 state->section_end_code = (1 << state->bits) - 1; 402 } 403 404 /* Remember previous code. */ 405 state->oldcode = newcode; 406 return (ARCHIVE_OK); 407 } 408 409 /* 410 * Return next 'n' bits from stream. 411 * 412 * -1 indicates end of available data. 413 */ 414 static int 415 getbits(struct archive_read_filter *self, int n) 416 { 417 struct private_data *state = (struct private_data *)self->data; 418 int code; 419 ssize_t ret; 420 static const int mask[] = { 421 0x00, 0x01, 0x03, 0x07, 0x0f, 0x1f, 0x3f, 0x7f, 0xff, 422 0x1ff, 0x3ff, 0x7ff, 0xfff, 0x1fff, 0x3fff, 0x7fff, 0xffff 423 }; 424 425 while (state->bits_avail < n) { 426 if (state->avail_in <= 0) { 427 if (state->consume_unnotified) { 428 __archive_read_filter_consume(self->upstream, 429 state->consume_unnotified); 430 state->consume_unnotified = 0; 431 } 432 state->next_in 433 = __archive_read_filter_ahead(self->upstream, 434 1, &ret); 435 if (ret == 0) 436 return (-1); 437 if (ret < 0 || state->next_in == NULL) 438 return (ARCHIVE_FATAL); 439 state->consume_unnotified = state->avail_in = ret; 440 } 441 state->bit_buffer |= *state->next_in++ << state->bits_avail; 442 state->avail_in--; 443 state->bits_avail += 8; 444 state->bytes_in_section++; 445 } 446 447 code = state->bit_buffer; 448 state->bit_buffer >>= n; 449 state->bits_avail -= n; 450 451 return (code & mask[n]); 452 } 453