archive_read_support_format_rar.c revision 302295
1/*- 2* Copyright (c) 2003-2007 Tim Kientzle 3* Copyright (c) 2011 Andres Mejia 4* All rights reserved. 5* 6* Redistribution and use in source and binary forms, with or without 7* modification, are permitted provided that the following conditions 8* are met: 9* 1. Redistributions of source code must retain the above copyright 10* notice, this list of conditions and the following disclaimer. 11* 2. Redistributions in binary form must reproduce the above copyright 12* notice, this list of conditions and the following disclaimer in the 13* documentation and/or other materials provided with the distribution. 14* 15* THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 16* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 17* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 18* IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 19* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 20* NOT 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 OF 24* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 25*/ 26 27#include "archive_platform.h" 28 29#ifdef HAVE_ERRNO_H 30#include <errno.h> 31#endif 32#include <time.h> 33#include <limits.h> 34#ifdef HAVE_ZLIB_H 35#include <zlib.h> /* crc32 */ 36#endif 37 38#include "archive.h" 39#ifndef HAVE_ZLIB_H 40#include "archive_crc32.h" 41#endif 42#include "archive_endian.h" 43#include "archive_entry.h" 44#include "archive_entry_locale.h" 45#include "archive_ppmd7_private.h" 46#include "archive_private.h" 47#include "archive_read_private.h" 48 49/* RAR signature, also known as the mark header */ 50#define RAR_SIGNATURE "\x52\x61\x72\x21\x1A\x07\x00" 51 52/* Header types */ 53#define MARK_HEAD 0x72 54#define MAIN_HEAD 0x73 55#define FILE_HEAD 0x74 56#define COMM_HEAD 0x75 57#define AV_HEAD 0x76 58#define SUB_HEAD 0x77 59#define PROTECT_HEAD 0x78 60#define SIGN_HEAD 0x79 61#define NEWSUB_HEAD 0x7a 62#define ENDARC_HEAD 0x7b 63 64/* Main Header Flags */ 65#define MHD_VOLUME 0x0001 66#define MHD_COMMENT 0x0002 67#define MHD_LOCK 0x0004 68#define MHD_SOLID 0x0008 69#define MHD_NEWNUMBERING 0x0010 70#define MHD_AV 0x0020 71#define MHD_PROTECT 0x0040 72#define MHD_PASSWORD 0x0080 73#define MHD_FIRSTVOLUME 0x0100 74#define MHD_ENCRYPTVER 0x0200 75 76/* Flags common to all headers */ 77#define HD_MARKDELETION 0x4000 78#define HD_ADD_SIZE_PRESENT 0x8000 79 80/* File Header Flags */ 81#define FHD_SPLIT_BEFORE 0x0001 82#define FHD_SPLIT_AFTER 0x0002 83#define FHD_PASSWORD 0x0004 84#define FHD_COMMENT 0x0008 85#define FHD_SOLID 0x0010 86#define FHD_LARGE 0x0100 87#define FHD_UNICODE 0x0200 88#define FHD_SALT 0x0400 89#define FHD_VERSION 0x0800 90#define FHD_EXTTIME 0x1000 91#define FHD_EXTFLAGS 0x2000 92 93/* File dictionary sizes */ 94#define DICTIONARY_SIZE_64 0x00 95#define DICTIONARY_SIZE_128 0x20 96#define DICTIONARY_SIZE_256 0x40 97#define DICTIONARY_SIZE_512 0x60 98#define DICTIONARY_SIZE_1024 0x80 99#define DICTIONARY_SIZE_2048 0xA0 100#define DICTIONARY_SIZE_4096 0xC0 101#define FILE_IS_DIRECTORY 0xE0 102#define DICTIONARY_MASK FILE_IS_DIRECTORY 103 104/* OS Flags */ 105#define OS_MSDOS 0 106#define OS_OS2 1 107#define OS_WIN32 2 108#define OS_UNIX 3 109#define OS_MAC_OS 4 110#define OS_BEOS 5 111 112/* Compression Methods */ 113#define COMPRESS_METHOD_STORE 0x30 114/* LZSS */ 115#define COMPRESS_METHOD_FASTEST 0x31 116#define COMPRESS_METHOD_FAST 0x32 117#define COMPRESS_METHOD_NORMAL 0x33 118/* PPMd Variant H */ 119#define COMPRESS_METHOD_GOOD 0x34 120#define COMPRESS_METHOD_BEST 0x35 121 122#define CRC_POLYNOMIAL 0xEDB88320 123 124#define NS_UNIT 10000000 125 126#define DICTIONARY_MAX_SIZE 0x400000 127 128#define MAINCODE_SIZE 299 129#define OFFSETCODE_SIZE 60 130#define LOWOFFSETCODE_SIZE 17 131#define LENGTHCODE_SIZE 28 132#define HUFFMAN_TABLE_SIZE \ 133 MAINCODE_SIZE + OFFSETCODE_SIZE + LOWOFFSETCODE_SIZE + LENGTHCODE_SIZE 134 135#define MAX_SYMBOL_LENGTH 0xF 136#define MAX_SYMBOLS 20 137 138/* 139 * Considering L1,L2 cache miss and a calling of write system-call, 140 * the best size of the output buffer(uncompressed buffer) is 128K. 141 * If the structure of extracting process is changed, this value 142 * might be researched again. 143 */ 144#define UNP_BUFFER_SIZE (128 * 1024) 145 146/* Define this here for non-Windows platforms */ 147#if !((defined(__WIN32__) || defined(_WIN32) || defined(__WIN32)) && !defined(__CYGWIN__)) 148#define FILE_ATTRIBUTE_DIRECTORY 0x10 149#endif 150 151/* Fields common to all headers */ 152struct rar_header 153{ 154 char crc[2]; 155 char type; 156 char flags[2]; 157 char size[2]; 158}; 159 160/* Fields common to all file headers */ 161struct rar_file_header 162{ 163 char pack_size[4]; 164 char unp_size[4]; 165 char host_os; 166 char file_crc[4]; 167 char file_time[4]; 168 char unp_ver; 169 char method; 170 char name_size[2]; 171 char file_attr[4]; 172}; 173 174struct huffman_tree_node 175{ 176 int branches[2]; 177}; 178 179struct huffman_table_entry 180{ 181 unsigned int length; 182 int value; 183}; 184 185struct huffman_code 186{ 187 struct huffman_tree_node *tree; 188 int numentries; 189 int numallocatedentries; 190 int minlength; 191 int maxlength; 192 int tablesize; 193 struct huffman_table_entry *table; 194}; 195 196struct lzss 197{ 198 unsigned char *window; 199 int mask; 200 int64_t position; 201}; 202 203struct data_block_offsets 204{ 205 int64_t header_size; 206 int64_t start_offset; 207 int64_t end_offset; 208}; 209 210struct rar 211{ 212 /* Entries from main RAR header */ 213 unsigned main_flags; 214 unsigned long file_crc; 215 char reserved1[2]; 216 char reserved2[4]; 217 char encryptver; 218 219 /* File header entries */ 220 char compression_method; 221 unsigned file_flags; 222 int64_t packed_size; 223 int64_t unp_size; 224 time_t mtime; 225 long mnsec; 226 mode_t mode; 227 char *filename; 228 char *filename_save; 229 size_t filename_save_size; 230 size_t filename_allocated; 231 232 /* File header optional entries */ 233 char salt[8]; 234 time_t atime; 235 long ansec; 236 time_t ctime; 237 long cnsec; 238 time_t arctime; 239 long arcnsec; 240 241 /* Fields to help with tracking decompression of files. */ 242 int64_t bytes_unconsumed; 243 int64_t bytes_remaining; 244 int64_t bytes_uncopied; 245 int64_t offset; 246 int64_t offset_outgoing; 247 int64_t offset_seek; 248 char valid; 249 unsigned int unp_offset; 250 unsigned int unp_buffer_size; 251 unsigned char *unp_buffer; 252 unsigned int dictionary_size; 253 char start_new_block; 254 char entry_eof; 255 unsigned long crc_calculated; 256 int found_first_header; 257 char has_endarc_header; 258 struct data_block_offsets *dbo; 259 unsigned int cursor; 260 unsigned int nodes; 261 262 /* LZSS members */ 263 struct huffman_code maincode; 264 struct huffman_code offsetcode; 265 struct huffman_code lowoffsetcode; 266 struct huffman_code lengthcode; 267 unsigned char lengthtable[HUFFMAN_TABLE_SIZE]; 268 struct lzss lzss; 269 char output_last_match; 270 unsigned int lastlength; 271 unsigned int lastoffset; 272 unsigned int oldoffset[4]; 273 unsigned int lastlowoffset; 274 unsigned int numlowoffsetrepeats; 275 int64_t filterstart; 276 char start_new_table; 277 278 /* PPMd Variant H members */ 279 char ppmd_valid; 280 char ppmd_eod; 281 char is_ppmd_block; 282 int ppmd_escape; 283 CPpmd7 ppmd7_context; 284 CPpmd7z_RangeDec range_dec; 285 IByteIn bytein; 286 287 /* 288 * String conversion object. 289 */ 290 int init_default_conversion; 291 struct archive_string_conv *sconv_default; 292 struct archive_string_conv *opt_sconv; 293 struct archive_string_conv *sconv_utf8; 294 struct archive_string_conv *sconv_utf16be; 295 296 /* 297 * Bit stream reader. 298 */ 299 struct rar_br { 300#define CACHE_TYPE uint64_t 301#define CACHE_BITS (8 * sizeof(CACHE_TYPE)) 302 /* Cache buffer. */ 303 CACHE_TYPE cache_buffer; 304 /* Indicates how many bits avail in cache_buffer. */ 305 int cache_avail; 306 ssize_t avail_in; 307 const unsigned char *next_in; 308 } br; 309 310 /* 311 * Custom field to denote that this archive contains encrypted entries 312 */ 313 int has_encrypted_entries; 314}; 315 316static int archive_read_support_format_rar_capabilities(struct archive_read *); 317static int archive_read_format_rar_has_encrypted_entries(struct archive_read *); 318static int archive_read_format_rar_bid(struct archive_read *, int); 319static int archive_read_format_rar_options(struct archive_read *, 320 const char *, const char *); 321static int archive_read_format_rar_read_header(struct archive_read *, 322 struct archive_entry *); 323static int archive_read_format_rar_read_data(struct archive_read *, 324 const void **, size_t *, int64_t *); 325static int archive_read_format_rar_read_data_skip(struct archive_read *a); 326static int64_t archive_read_format_rar_seek_data(struct archive_read *, int64_t, 327 int); 328static int archive_read_format_rar_cleanup(struct archive_read *); 329 330/* Support functions */ 331static int read_header(struct archive_read *, struct archive_entry *, char); 332static time_t get_time(int); 333static int read_exttime(const char *, struct rar *, const char *); 334static int read_symlink_stored(struct archive_read *, struct archive_entry *, 335 struct archive_string_conv *); 336static int read_data_stored(struct archive_read *, const void **, size_t *, 337 int64_t *); 338static int read_data_compressed(struct archive_read *, const void **, size_t *, 339 int64_t *); 340static int rar_br_preparation(struct archive_read *, struct rar_br *); 341static int parse_codes(struct archive_read *); 342static void free_codes(struct archive_read *); 343static int read_next_symbol(struct archive_read *, struct huffman_code *); 344static int create_code(struct archive_read *, struct huffman_code *, 345 unsigned char *, int, char); 346static int add_value(struct archive_read *, struct huffman_code *, int, int, 347 int); 348static int new_node(struct huffman_code *); 349static int make_table(struct archive_read *, struct huffman_code *); 350static int make_table_recurse(struct archive_read *, struct huffman_code *, int, 351 struct huffman_table_entry *, int, int); 352static int64_t expand(struct archive_read *, int64_t); 353static int copy_from_lzss_window(struct archive_read *, const void **, 354 int64_t, int); 355static const void *rar_read_ahead(struct archive_read *, size_t, ssize_t *); 356 357/* 358 * Bit stream reader. 359 */ 360/* Check that the cache buffer has enough bits. */ 361#define rar_br_has(br, n) ((br)->cache_avail >= n) 362/* Get compressed data by bit. */ 363#define rar_br_bits(br, n) \ 364 (((uint32_t)((br)->cache_buffer >> \ 365 ((br)->cache_avail - (n)))) & cache_masks[n]) 366#define rar_br_bits_forced(br, n) \ 367 (((uint32_t)((br)->cache_buffer << \ 368 ((n) - (br)->cache_avail))) & cache_masks[n]) 369/* Read ahead to make sure the cache buffer has enough compressed data we 370 * will use. 371 * True : completed, there is enough data in the cache buffer. 372 * False : there is no data in the stream. */ 373#define rar_br_read_ahead(a, br, n) \ 374 ((rar_br_has(br, (n)) || rar_br_fillup(a, br)) || rar_br_has(br, (n))) 375/* Notify how many bits we consumed. */ 376#define rar_br_consume(br, n) ((br)->cache_avail -= (n)) 377#define rar_br_consume_unalined_bits(br) ((br)->cache_avail &= ~7) 378 379static const uint32_t cache_masks[] = { 380 0x00000000, 0x00000001, 0x00000003, 0x00000007, 381 0x0000000F, 0x0000001F, 0x0000003F, 0x0000007F, 382 0x000000FF, 0x000001FF, 0x000003FF, 0x000007FF, 383 0x00000FFF, 0x00001FFF, 0x00003FFF, 0x00007FFF, 384 0x0000FFFF, 0x0001FFFF, 0x0003FFFF, 0x0007FFFF, 385 0x000FFFFF, 0x001FFFFF, 0x003FFFFF, 0x007FFFFF, 386 0x00FFFFFF, 0x01FFFFFF, 0x03FFFFFF, 0x07FFFFFF, 387 0x0FFFFFFF, 0x1FFFFFFF, 0x3FFFFFFF, 0x7FFFFFFF, 388 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF 389}; 390 391/* 392 * Shift away used bits in the cache data and fill it up with following bits. 393 * Call this when cache buffer does not have enough bits you need. 394 * 395 * Returns 1 if the cache buffer is full. 396 * Returns 0 if the cache buffer is not full; input buffer is empty. 397 */ 398static int 399rar_br_fillup(struct archive_read *a, struct rar_br *br) 400{ 401 struct rar *rar = (struct rar *)(a->format->data); 402 int n = CACHE_BITS - br->cache_avail; 403 404 for (;;) { 405 switch (n >> 3) { 406 case 8: 407 if (br->avail_in >= 8) { 408 br->cache_buffer = 409 ((uint64_t)br->next_in[0]) << 56 | 410 ((uint64_t)br->next_in[1]) << 48 | 411 ((uint64_t)br->next_in[2]) << 40 | 412 ((uint64_t)br->next_in[3]) << 32 | 413 ((uint32_t)br->next_in[4]) << 24 | 414 ((uint32_t)br->next_in[5]) << 16 | 415 ((uint32_t)br->next_in[6]) << 8 | 416 (uint32_t)br->next_in[7]; 417 br->next_in += 8; 418 br->avail_in -= 8; 419 br->cache_avail += 8 * 8; 420 rar->bytes_unconsumed += 8; 421 rar->bytes_remaining -= 8; 422 return (1); 423 } 424 break; 425 case 7: 426 if (br->avail_in >= 7) { 427 br->cache_buffer = 428 (br->cache_buffer << 56) | 429 ((uint64_t)br->next_in[0]) << 48 | 430 ((uint64_t)br->next_in[1]) << 40 | 431 ((uint64_t)br->next_in[2]) << 32 | 432 ((uint32_t)br->next_in[3]) << 24 | 433 ((uint32_t)br->next_in[4]) << 16 | 434 ((uint32_t)br->next_in[5]) << 8 | 435 (uint32_t)br->next_in[6]; 436 br->next_in += 7; 437 br->avail_in -= 7; 438 br->cache_avail += 7 * 8; 439 rar->bytes_unconsumed += 7; 440 rar->bytes_remaining -= 7; 441 return (1); 442 } 443 break; 444 case 6: 445 if (br->avail_in >= 6) { 446 br->cache_buffer = 447 (br->cache_buffer << 48) | 448 ((uint64_t)br->next_in[0]) << 40 | 449 ((uint64_t)br->next_in[1]) << 32 | 450 ((uint32_t)br->next_in[2]) << 24 | 451 ((uint32_t)br->next_in[3]) << 16 | 452 ((uint32_t)br->next_in[4]) << 8 | 453 (uint32_t)br->next_in[5]; 454 br->next_in += 6; 455 br->avail_in -= 6; 456 br->cache_avail += 6 * 8; 457 rar->bytes_unconsumed += 6; 458 rar->bytes_remaining -= 6; 459 return (1); 460 } 461 break; 462 case 0: 463 /* We have enough compressed data in 464 * the cache buffer.*/ 465 return (1); 466 default: 467 break; 468 } 469 if (br->avail_in <= 0) { 470 471 if (rar->bytes_unconsumed > 0) { 472 /* Consume as much as the decompressor 473 * actually used. */ 474 __archive_read_consume(a, rar->bytes_unconsumed); 475 rar->bytes_unconsumed = 0; 476 } 477 br->next_in = rar_read_ahead(a, 1, &(br->avail_in)); 478 if (br->next_in == NULL) 479 return (0); 480 if (br->avail_in == 0) 481 return (0); 482 } 483 br->cache_buffer = 484 (br->cache_buffer << 8) | *br->next_in++; 485 br->avail_in--; 486 br->cache_avail += 8; 487 n -= 8; 488 rar->bytes_unconsumed++; 489 rar->bytes_remaining--; 490 } 491} 492 493static int 494rar_br_preparation(struct archive_read *a, struct rar_br *br) 495{ 496 struct rar *rar = (struct rar *)(a->format->data); 497 498 if (rar->bytes_remaining > 0) { 499 br->next_in = rar_read_ahead(a, 1, &(br->avail_in)); 500 if (br->next_in == NULL) { 501 archive_set_error(&a->archive, 502 ARCHIVE_ERRNO_FILE_FORMAT, 503 "Truncated RAR file data"); 504 return (ARCHIVE_FATAL); 505 } 506 if (br->cache_avail == 0) 507 (void)rar_br_fillup(a, br); 508 } 509 return (ARCHIVE_OK); 510} 511 512/* Find last bit set */ 513static inline int 514rar_fls(unsigned int word) 515{ 516 word |= (word >> 1); 517 word |= (word >> 2); 518 word |= (word >> 4); 519 word |= (word >> 8); 520 word |= (word >> 16); 521 return word - (word >> 1); 522} 523 524/* LZSS functions */ 525static inline int64_t 526lzss_position(struct lzss *lzss) 527{ 528 return lzss->position; 529} 530 531static inline int 532lzss_mask(struct lzss *lzss) 533{ 534 return lzss->mask; 535} 536 537static inline int 538lzss_size(struct lzss *lzss) 539{ 540 return lzss->mask + 1; 541} 542 543static inline int 544lzss_offset_for_position(struct lzss *lzss, int64_t pos) 545{ 546 return (int)(pos & lzss->mask); 547} 548 549static inline unsigned char * 550lzss_pointer_for_position(struct lzss *lzss, int64_t pos) 551{ 552 return &lzss->window[lzss_offset_for_position(lzss, pos)]; 553} 554 555static inline int 556lzss_current_offset(struct lzss *lzss) 557{ 558 return lzss_offset_for_position(lzss, lzss->position); 559} 560 561static inline uint8_t * 562lzss_current_pointer(struct lzss *lzss) 563{ 564 return lzss_pointer_for_position(lzss, lzss->position); 565} 566 567static inline void 568lzss_emit_literal(struct rar *rar, uint8_t literal) 569{ 570 *lzss_current_pointer(&rar->lzss) = literal; 571 rar->lzss.position++; 572} 573 574static inline void 575lzss_emit_match(struct rar *rar, int offset, int length) 576{ 577 int dstoffs = lzss_current_offset(&rar->lzss); 578 int srcoffs = (dstoffs - offset) & lzss_mask(&rar->lzss); 579 int l, li, remaining; 580 unsigned char *d, *s; 581 582 remaining = length; 583 while (remaining > 0) { 584 l = remaining; 585 if (dstoffs > srcoffs) { 586 if (l > lzss_size(&rar->lzss) - dstoffs) 587 l = lzss_size(&rar->lzss) - dstoffs; 588 } else { 589 if (l > lzss_size(&rar->lzss) - srcoffs) 590 l = lzss_size(&rar->lzss) - srcoffs; 591 } 592 d = &(rar->lzss.window[dstoffs]); 593 s = &(rar->lzss.window[srcoffs]); 594 if ((dstoffs + l < srcoffs) || (srcoffs + l < dstoffs)) 595 memcpy(d, s, l); 596 else { 597 for (li = 0; li < l; li++) 598 d[li] = s[li]; 599 } 600 remaining -= l; 601 dstoffs = (dstoffs + l) & lzss_mask(&(rar->lzss)); 602 srcoffs = (srcoffs + l) & lzss_mask(&(rar->lzss)); 603 } 604 rar->lzss.position += length; 605} 606 607static void * 608ppmd_alloc(void *p, size_t size) 609{ 610 (void)p; 611 return malloc(size); 612} 613static void 614ppmd_free(void *p, void *address) 615{ 616 (void)p; 617 free(address); 618} 619static ISzAlloc g_szalloc = { ppmd_alloc, ppmd_free }; 620 621static Byte 622ppmd_read(void *p) 623{ 624 struct archive_read *a = ((IByteIn*)p)->a; 625 struct rar *rar = (struct rar *)(a->format->data); 626 struct rar_br *br = &(rar->br); 627 Byte b; 628 if (!rar_br_read_ahead(a, br, 8)) 629 { 630 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 631 "Truncated RAR file data"); 632 rar->valid = 0; 633 return 0; 634 } 635 b = rar_br_bits(br, 8); 636 rar_br_consume(br, 8); 637 return b; 638} 639 640int 641archive_read_support_format_rar(struct archive *_a) 642{ 643 struct archive_read *a = (struct archive_read *)_a; 644 struct rar *rar; 645 int r; 646 647 archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW, 648 "archive_read_support_format_rar"); 649 650 rar = (struct rar *)malloc(sizeof(*rar)); 651 if (rar == NULL) 652 { 653 archive_set_error(&a->archive, ENOMEM, "Can't allocate rar data"); 654 return (ARCHIVE_FATAL); 655 } 656 memset(rar, 0, sizeof(*rar)); 657 658 /* 659 * Until enough data has been read, we cannot tell about 660 * any encrypted entries yet. 661 */ 662 rar->has_encrypted_entries = ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; 663 664 r = __archive_read_register_format(a, 665 rar, 666 "rar", 667 archive_read_format_rar_bid, 668 archive_read_format_rar_options, 669 archive_read_format_rar_read_header, 670 archive_read_format_rar_read_data, 671 archive_read_format_rar_read_data_skip, 672 archive_read_format_rar_seek_data, 673 archive_read_format_rar_cleanup, 674 archive_read_support_format_rar_capabilities, 675 archive_read_format_rar_has_encrypted_entries); 676 677 if (r != ARCHIVE_OK) 678 free(rar); 679 return (r); 680} 681 682static int 683archive_read_support_format_rar_capabilities(struct archive_read * a) 684{ 685 (void)a; /* UNUSED */ 686 return (ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_DATA 687 | ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_METADATA); 688} 689 690static int 691archive_read_format_rar_has_encrypted_entries(struct archive_read *_a) 692{ 693 if (_a && _a->format) { 694 struct rar * rar = (struct rar *)_a->format->data; 695 if (rar) { 696 return rar->has_encrypted_entries; 697 } 698 } 699 return ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; 700} 701 702 703static int 704archive_read_format_rar_bid(struct archive_read *a, int best_bid) 705{ 706 const char *p; 707 708 /* If there's already a bid > 30, we'll never win. */ 709 if (best_bid > 30) 710 return (-1); 711 712 if ((p = __archive_read_ahead(a, 7, NULL)) == NULL) 713 return (-1); 714 715 if (memcmp(p, RAR_SIGNATURE, 7) == 0) 716 return (30); 717 718 if ((p[0] == 'M' && p[1] == 'Z') || memcmp(p, "\x7F\x45LF", 4) == 0) { 719 /* This is a PE file */ 720 ssize_t offset = 0x10000; 721 ssize_t window = 4096; 722 ssize_t bytes_avail; 723 while (offset + window <= (1024 * 128)) { 724 const char *buff = __archive_read_ahead(a, offset + window, &bytes_avail); 725 if (buff == NULL) { 726 /* Remaining bytes are less than window. */ 727 window >>= 1; 728 if (window < 0x40) 729 return (0); 730 continue; 731 } 732 p = buff + offset; 733 while (p + 7 < buff + bytes_avail) { 734 if (memcmp(p, RAR_SIGNATURE, 7) == 0) 735 return (30); 736 p += 0x10; 737 } 738 offset = p - buff; 739 } 740 } 741 return (0); 742} 743 744static int 745skip_sfx(struct archive_read *a) 746{ 747 const void *h; 748 const char *p, *q; 749 size_t skip, total; 750 ssize_t bytes, window; 751 752 total = 0; 753 window = 4096; 754 while (total + window <= (1024 * 128)) { 755 h = __archive_read_ahead(a, window, &bytes); 756 if (h == NULL) { 757 /* Remaining bytes are less than window. */ 758 window >>= 1; 759 if (window < 0x40) 760 goto fatal; 761 continue; 762 } 763 if (bytes < 0x40) 764 goto fatal; 765 p = h; 766 q = p + bytes; 767 768 /* 769 * Scan ahead until we find something that looks 770 * like the RAR header. 771 */ 772 while (p + 7 < q) { 773 if (memcmp(p, RAR_SIGNATURE, 7) == 0) { 774 skip = p - (const char *)h; 775 __archive_read_consume(a, skip); 776 return (ARCHIVE_OK); 777 } 778 p += 0x10; 779 } 780 skip = p - (const char *)h; 781 __archive_read_consume(a, skip); 782 total += skip; 783 } 784fatal: 785 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 786 "Couldn't find out RAR header"); 787 return (ARCHIVE_FATAL); 788} 789 790static int 791archive_read_format_rar_options(struct archive_read *a, 792 const char *key, const char *val) 793{ 794 struct rar *rar; 795 int ret = ARCHIVE_FAILED; 796 797 rar = (struct rar *)(a->format->data); 798 if (strcmp(key, "hdrcharset") == 0) { 799 if (val == NULL || val[0] == 0) 800 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 801 "rar: hdrcharset option needs a character-set name"); 802 else { 803 rar->opt_sconv = 804 archive_string_conversion_from_charset( 805 &a->archive, val, 0); 806 if (rar->opt_sconv != NULL) 807 ret = ARCHIVE_OK; 808 else 809 ret = ARCHIVE_FATAL; 810 } 811 return (ret); 812 } 813 814 /* Note: The "warn" return is just to inform the options 815 * supervisor that we didn't handle it. It will generate 816 * a suitable error if no one used this option. */ 817 return (ARCHIVE_WARN); 818} 819 820static int 821archive_read_format_rar_read_header(struct archive_read *a, 822 struct archive_entry *entry) 823{ 824 const void *h; 825 const char *p; 826 struct rar *rar; 827 size_t skip; 828 char head_type; 829 int ret; 830 unsigned flags; 831 unsigned long crc32_expected; 832 833 a->archive.archive_format = ARCHIVE_FORMAT_RAR; 834 if (a->archive.archive_format_name == NULL) 835 a->archive.archive_format_name = "RAR"; 836 837 rar = (struct rar *)(a->format->data); 838 839 /* 840 * It should be sufficient to call archive_read_next_header() for 841 * a reader to determine if an entry is encrypted or not. If the 842 * encryption of an entry is only detectable when calling 843 * archive_read_data(), so be it. We'll do the same check there 844 * as well. 845 */ 846 if (rar->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { 847 rar->has_encrypted_entries = 0; 848 } 849 850 /* RAR files can be generated without EOF headers, so return ARCHIVE_EOF if 851 * this fails. 852 */ 853 if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) 854 return (ARCHIVE_EOF); 855 856 p = h; 857 if (rar->found_first_header == 0 && 858 ((p[0] == 'M' && p[1] == 'Z') || memcmp(p, "\x7F\x45LF", 4) == 0)) { 859 /* This is an executable ? Must be self-extracting... */ 860 ret = skip_sfx(a); 861 if (ret < ARCHIVE_WARN) 862 return (ret); 863 } 864 rar->found_first_header = 1; 865 866 while (1) 867 { 868 unsigned long crc32_val; 869 870 if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) 871 return (ARCHIVE_FATAL); 872 p = h; 873 874 head_type = p[2]; 875 switch(head_type) 876 { 877 case MARK_HEAD: 878 if (memcmp(p, RAR_SIGNATURE, 7) != 0) { 879 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 880 "Invalid marker header"); 881 return (ARCHIVE_FATAL); 882 } 883 __archive_read_consume(a, 7); 884 break; 885 886 case MAIN_HEAD: 887 rar->main_flags = archive_le16dec(p + 3); 888 skip = archive_le16dec(p + 5); 889 if (skip < 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)) { 890 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 891 "Invalid header size"); 892 return (ARCHIVE_FATAL); 893 } 894 if ((h = __archive_read_ahead(a, skip, NULL)) == NULL) 895 return (ARCHIVE_FATAL); 896 p = h; 897 memcpy(rar->reserved1, p + 7, sizeof(rar->reserved1)); 898 memcpy(rar->reserved2, p + 7 + sizeof(rar->reserved1), 899 sizeof(rar->reserved2)); 900 if (rar->main_flags & MHD_ENCRYPTVER) { 901 if (skip < 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)+1) { 902 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 903 "Invalid header size"); 904 return (ARCHIVE_FATAL); 905 } 906 rar->encryptver = *(p + 7 + sizeof(rar->reserved1) + 907 sizeof(rar->reserved2)); 908 } 909 910 /* Main header is password encrytped, so we cannot read any 911 file names or any other info about files from the header. */ 912 if (rar->main_flags & MHD_PASSWORD) 913 { 914 archive_entry_set_is_metadata_encrypted(entry, 1); 915 archive_entry_set_is_data_encrypted(entry, 1); 916 rar->has_encrypted_entries = 1; 917 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 918 "RAR encryption support unavailable."); 919 return (ARCHIVE_FATAL); 920 } 921 922 crc32_val = crc32(0, (const unsigned char *)p + 2, (unsigned)skip - 2); 923 if ((crc32_val & 0xffff) != archive_le16dec(p)) { 924 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 925 "Header CRC error"); 926 return (ARCHIVE_FATAL); 927 } 928 __archive_read_consume(a, skip); 929 break; 930 931 case FILE_HEAD: 932 return read_header(a, entry, head_type); 933 934 case COMM_HEAD: 935 case AV_HEAD: 936 case SUB_HEAD: 937 case PROTECT_HEAD: 938 case SIGN_HEAD: 939 case ENDARC_HEAD: 940 flags = archive_le16dec(p + 3); 941 skip = archive_le16dec(p + 5); 942 if (skip < 7) { 943 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 944 "Invalid header size too small"); 945 return (ARCHIVE_FATAL); 946 } 947 if (flags & HD_ADD_SIZE_PRESENT) 948 { 949 if (skip < 7 + 4) { 950 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 951 "Invalid header size too small"); 952 return (ARCHIVE_FATAL); 953 } 954 if ((h = __archive_read_ahead(a, skip, NULL)) == NULL) 955 return (ARCHIVE_FATAL); 956 p = h; 957 skip += archive_le32dec(p + 7); 958 } 959 960 /* Skip over the 2-byte CRC at the beginning of the header. */ 961 crc32_expected = archive_le16dec(p); 962 __archive_read_consume(a, 2); 963 skip -= 2; 964 965 /* Skim the entire header and compute the CRC. */ 966 crc32_val = 0; 967 while (skip > 0) { 968 size_t to_read = skip; 969 ssize_t did_read; 970 if (to_read > 32 * 1024) { 971 to_read = 32 * 1024; 972 } 973 if ((h = __archive_read_ahead(a, to_read, &did_read)) == NULL) { 974 return (ARCHIVE_FATAL); 975 } 976 p = h; 977 crc32_val = crc32(crc32_val, (const unsigned char *)p, (unsigned)did_read); 978 __archive_read_consume(a, did_read); 979 skip -= did_read; 980 } 981 if ((crc32_val & 0xffff) != crc32_expected) { 982 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 983 "Header CRC error"); 984 return (ARCHIVE_FATAL); 985 } 986 if (head_type == ENDARC_HEAD) 987 return (ARCHIVE_EOF); 988 break; 989 990 case NEWSUB_HEAD: 991 if ((ret = read_header(a, entry, head_type)) < ARCHIVE_WARN) 992 return ret; 993 break; 994 995 default: 996 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 997 "Bad RAR file"); 998 return (ARCHIVE_FATAL); 999 } 1000 } 1001} 1002 1003static int 1004archive_read_format_rar_read_data(struct archive_read *a, const void **buff, 1005 size_t *size, int64_t *offset) 1006{ 1007 struct rar *rar = (struct rar *)(a->format->data); 1008 int ret; 1009 1010 if (rar->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { 1011 rar->has_encrypted_entries = 0; 1012 } 1013 1014 if (rar->bytes_unconsumed > 0) { 1015 /* Consume as much as the decompressor actually used. */ 1016 __archive_read_consume(a, rar->bytes_unconsumed); 1017 rar->bytes_unconsumed = 0; 1018 } 1019 1020 *buff = NULL; 1021 if (rar->entry_eof || rar->offset_seek >= rar->unp_size) { 1022 *size = 0; 1023 *offset = rar->offset; 1024 if (*offset < rar->unp_size) 1025 *offset = rar->unp_size; 1026 return (ARCHIVE_EOF); 1027 } 1028 1029 switch (rar->compression_method) 1030 { 1031 case COMPRESS_METHOD_STORE: 1032 ret = read_data_stored(a, buff, size, offset); 1033 break; 1034 1035 case COMPRESS_METHOD_FASTEST: 1036 case COMPRESS_METHOD_FAST: 1037 case COMPRESS_METHOD_NORMAL: 1038 case COMPRESS_METHOD_GOOD: 1039 case COMPRESS_METHOD_BEST: 1040 ret = read_data_compressed(a, buff, size, offset); 1041 if (ret != ARCHIVE_OK && ret != ARCHIVE_WARN) 1042 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 1043 break; 1044 1045 default: 1046 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1047 "Unsupported compression method for RAR file."); 1048 ret = ARCHIVE_FATAL; 1049 break; 1050 } 1051 return (ret); 1052} 1053 1054static int 1055archive_read_format_rar_read_data_skip(struct archive_read *a) 1056{ 1057 struct rar *rar; 1058 int64_t bytes_skipped; 1059 int ret; 1060 1061 rar = (struct rar *)(a->format->data); 1062 1063 if (rar->bytes_unconsumed > 0) { 1064 /* Consume as much as the decompressor actually used. */ 1065 __archive_read_consume(a, rar->bytes_unconsumed); 1066 rar->bytes_unconsumed = 0; 1067 } 1068 1069 if (rar->bytes_remaining > 0) { 1070 bytes_skipped = __archive_read_consume(a, rar->bytes_remaining); 1071 if (bytes_skipped < 0) 1072 return (ARCHIVE_FATAL); 1073 } 1074 1075 /* Compressed data to skip must be read from each header in a multivolume 1076 * archive. 1077 */ 1078 if (rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER) 1079 { 1080 ret = archive_read_format_rar_read_header(a, a->entry); 1081 if (ret == (ARCHIVE_EOF)) 1082 ret = archive_read_format_rar_read_header(a, a->entry); 1083 if (ret != (ARCHIVE_OK)) 1084 return ret; 1085 return archive_read_format_rar_read_data_skip(a); 1086 } 1087 1088 return (ARCHIVE_OK); 1089} 1090 1091static int64_t 1092archive_read_format_rar_seek_data(struct archive_read *a, int64_t offset, 1093 int whence) 1094{ 1095 int64_t client_offset, ret; 1096 unsigned int i; 1097 struct rar *rar = (struct rar *)(a->format->data); 1098 1099 if (rar->compression_method == COMPRESS_METHOD_STORE) 1100 { 1101 /* Modify the offset for use with SEEK_SET */ 1102 switch (whence) 1103 { 1104 case SEEK_CUR: 1105 client_offset = rar->offset_seek; 1106 break; 1107 case SEEK_END: 1108 client_offset = rar->unp_size; 1109 break; 1110 case SEEK_SET: 1111 default: 1112 client_offset = 0; 1113 } 1114 client_offset += offset; 1115 if (client_offset < 0) 1116 { 1117 /* Can't seek past beginning of data block */ 1118 return -1; 1119 } 1120 else if (client_offset > rar->unp_size) 1121 { 1122 /* 1123 * Set the returned offset but only seek to the end of 1124 * the data block. 1125 */ 1126 rar->offset_seek = client_offset; 1127 client_offset = rar->unp_size; 1128 } 1129 1130 client_offset += rar->dbo[0].start_offset; 1131 i = 0; 1132 while (i < rar->cursor) 1133 { 1134 i++; 1135 client_offset += rar->dbo[i].start_offset - rar->dbo[i-1].end_offset; 1136 } 1137 if (rar->main_flags & MHD_VOLUME) 1138 { 1139 /* Find the appropriate offset among the multivolume archive */ 1140 while (1) 1141 { 1142 if (client_offset < rar->dbo[rar->cursor].start_offset && 1143 rar->file_flags & FHD_SPLIT_BEFORE) 1144 { 1145 /* Search backwards for the correct data block */ 1146 if (rar->cursor == 0) 1147 { 1148 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1149 "Attempt to seek past beginning of RAR data block"); 1150 return (ARCHIVE_FAILED); 1151 } 1152 rar->cursor--; 1153 client_offset -= rar->dbo[rar->cursor+1].start_offset - 1154 rar->dbo[rar->cursor].end_offset; 1155 if (client_offset < rar->dbo[rar->cursor].start_offset) 1156 continue; 1157 ret = __archive_read_seek(a, rar->dbo[rar->cursor].start_offset - 1158 rar->dbo[rar->cursor].header_size, SEEK_SET); 1159 if (ret < (ARCHIVE_OK)) 1160 return ret; 1161 ret = archive_read_format_rar_read_header(a, a->entry); 1162 if (ret != (ARCHIVE_OK)) 1163 { 1164 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1165 "Error during seek of RAR file"); 1166 return (ARCHIVE_FAILED); 1167 } 1168 rar->cursor--; 1169 break; 1170 } 1171 else if (client_offset > rar->dbo[rar->cursor].end_offset && 1172 rar->file_flags & FHD_SPLIT_AFTER) 1173 { 1174 /* Search forward for the correct data block */ 1175 rar->cursor++; 1176 if (rar->cursor < rar->nodes && 1177 client_offset > rar->dbo[rar->cursor].end_offset) 1178 { 1179 client_offset += rar->dbo[rar->cursor].start_offset - 1180 rar->dbo[rar->cursor-1].end_offset; 1181 continue; 1182 } 1183 rar->cursor--; 1184 ret = __archive_read_seek(a, rar->dbo[rar->cursor].end_offset, 1185 SEEK_SET); 1186 if (ret < (ARCHIVE_OK)) 1187 return ret; 1188 ret = archive_read_format_rar_read_header(a, a->entry); 1189 if (ret == (ARCHIVE_EOF)) 1190 { 1191 rar->has_endarc_header = 1; 1192 ret = archive_read_format_rar_read_header(a, a->entry); 1193 } 1194 if (ret != (ARCHIVE_OK)) 1195 { 1196 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1197 "Error during seek of RAR file"); 1198 return (ARCHIVE_FAILED); 1199 } 1200 client_offset += rar->dbo[rar->cursor].start_offset - 1201 rar->dbo[rar->cursor-1].end_offset; 1202 continue; 1203 } 1204 break; 1205 } 1206 } 1207 1208 ret = __archive_read_seek(a, client_offset, SEEK_SET); 1209 if (ret < (ARCHIVE_OK)) 1210 return ret; 1211 rar->bytes_remaining = rar->dbo[rar->cursor].end_offset - ret; 1212 i = rar->cursor; 1213 while (i > 0) 1214 { 1215 i--; 1216 ret -= rar->dbo[i+1].start_offset - rar->dbo[i].end_offset; 1217 } 1218 ret -= rar->dbo[0].start_offset; 1219 1220 /* Always restart reading the file after a seek */ 1221 __archive_reset_read_data(&a->archive); 1222 1223 rar->bytes_unconsumed = 0; 1224 rar->offset = 0; 1225 1226 /* 1227 * If a seek past the end of file was requested, return the requested 1228 * offset. 1229 */ 1230 if (ret == rar->unp_size && rar->offset_seek > rar->unp_size) 1231 return rar->offset_seek; 1232 1233 /* Return the new offset */ 1234 rar->offset_seek = ret; 1235 return rar->offset_seek; 1236 } 1237 else 1238 { 1239 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1240 "Seeking of compressed RAR files is unsupported"); 1241 } 1242 return (ARCHIVE_FAILED); 1243} 1244 1245static int 1246archive_read_format_rar_cleanup(struct archive_read *a) 1247{ 1248 struct rar *rar; 1249 1250 rar = (struct rar *)(a->format->data); 1251 free_codes(a); 1252 free(rar->filename); 1253 free(rar->filename_save); 1254 free(rar->dbo); 1255 free(rar->unp_buffer); 1256 free(rar->lzss.window); 1257 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 1258 free(rar); 1259 (a->format->data) = NULL; 1260 return (ARCHIVE_OK); 1261} 1262 1263static int 1264read_header(struct archive_read *a, struct archive_entry *entry, 1265 char head_type) 1266{ 1267 const void *h; 1268 const char *p, *endp; 1269 struct rar *rar; 1270 struct rar_header rar_header; 1271 struct rar_file_header file_header; 1272 int64_t header_size; 1273 unsigned filename_size, end; 1274 char *filename; 1275 char *strp; 1276 char packed_size[8]; 1277 char unp_size[8]; 1278 int ttime; 1279 struct archive_string_conv *sconv, *fn_sconv; 1280 unsigned long crc32_val; 1281 int ret = (ARCHIVE_OK), ret2; 1282 1283 rar = (struct rar *)(a->format->data); 1284 1285 /* Setup a string conversion object for non-rar-unicode filenames. */ 1286 sconv = rar->opt_sconv; 1287 if (sconv == NULL) { 1288 if (!rar->init_default_conversion) { 1289 rar->sconv_default = 1290 archive_string_default_conversion_for_read( 1291 &(a->archive)); 1292 rar->init_default_conversion = 1; 1293 } 1294 sconv = rar->sconv_default; 1295 } 1296 1297 1298 if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) 1299 return (ARCHIVE_FATAL); 1300 p = h; 1301 memcpy(&rar_header, p, sizeof(rar_header)); 1302 rar->file_flags = archive_le16dec(rar_header.flags); 1303 header_size = archive_le16dec(rar_header.size); 1304 if (header_size < (int64_t)sizeof(file_header) + 7) { 1305 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1306 "Invalid header size"); 1307 return (ARCHIVE_FATAL); 1308 } 1309 crc32_val = crc32(0, (const unsigned char *)p + 2, 7 - 2); 1310 __archive_read_consume(a, 7); 1311 1312 if (!(rar->file_flags & FHD_SOLID)) 1313 { 1314 rar->compression_method = 0; 1315 rar->packed_size = 0; 1316 rar->unp_size = 0; 1317 rar->mtime = 0; 1318 rar->ctime = 0; 1319 rar->atime = 0; 1320 rar->arctime = 0; 1321 rar->mode = 0; 1322 memset(&rar->salt, 0, sizeof(rar->salt)); 1323 rar->atime = 0; 1324 rar->ansec = 0; 1325 rar->ctime = 0; 1326 rar->cnsec = 0; 1327 rar->mtime = 0; 1328 rar->mnsec = 0; 1329 rar->arctime = 0; 1330 rar->arcnsec = 0; 1331 } 1332 else 1333 { 1334 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1335 "RAR solid archive support unavailable."); 1336 return (ARCHIVE_FATAL); 1337 } 1338 1339 if ((h = __archive_read_ahead(a, (size_t)header_size - 7, NULL)) == NULL) 1340 return (ARCHIVE_FATAL); 1341 1342 /* File Header CRC check. */ 1343 crc32_val = crc32(crc32_val, h, (unsigned)(header_size - 7)); 1344 if ((crc32_val & 0xffff) != archive_le16dec(rar_header.crc)) { 1345 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1346 "Header CRC error"); 1347 return (ARCHIVE_FATAL); 1348 } 1349 /* If no CRC error, Go on parsing File Header. */ 1350 p = h; 1351 endp = p + header_size - 7; 1352 memcpy(&file_header, p, sizeof(file_header)); 1353 p += sizeof(file_header); 1354 1355 rar->compression_method = file_header.method; 1356 1357 ttime = archive_le32dec(file_header.file_time); 1358 rar->mtime = get_time(ttime); 1359 1360 rar->file_crc = archive_le32dec(file_header.file_crc); 1361 1362 if (rar->file_flags & FHD_PASSWORD) 1363 { 1364 archive_entry_set_is_data_encrypted(entry, 1); 1365 rar->has_encrypted_entries = 1; 1366 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1367 "RAR encryption support unavailable."); 1368 /* Since it is only the data part itself that is encrypted we can at least 1369 extract information about the currently processed entry and don't need 1370 to return ARCHIVE_FATAL here. */ 1371 /*return (ARCHIVE_FATAL);*/ 1372 } 1373 1374 if (rar->file_flags & FHD_LARGE) 1375 { 1376 memcpy(packed_size, file_header.pack_size, 4); 1377 memcpy(packed_size + 4, p, 4); /* High pack size */ 1378 p += 4; 1379 memcpy(unp_size, file_header.unp_size, 4); 1380 memcpy(unp_size + 4, p, 4); /* High unpack size */ 1381 p += 4; 1382 rar->packed_size = archive_le64dec(&packed_size); 1383 rar->unp_size = archive_le64dec(&unp_size); 1384 } 1385 else 1386 { 1387 rar->packed_size = archive_le32dec(file_header.pack_size); 1388 rar->unp_size = archive_le32dec(file_header.unp_size); 1389 } 1390 1391 if (rar->packed_size < 0 || rar->unp_size < 0) 1392 { 1393 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1394 "Invalid sizes specified."); 1395 return (ARCHIVE_FATAL); 1396 } 1397 1398 rar->bytes_remaining = rar->packed_size; 1399 1400 /* TODO: RARv3 subblocks contain comments. For now the complete block is 1401 * consumed at the end. 1402 */ 1403 if (head_type == NEWSUB_HEAD) { 1404 size_t distance = p - (const char *)h; 1405 header_size += rar->packed_size; 1406 /* Make sure we have the extended data. */ 1407 if ((h = __archive_read_ahead(a, (size_t)header_size - 7, NULL)) == NULL) 1408 return (ARCHIVE_FATAL); 1409 p = h; 1410 endp = p + header_size - 7; 1411 p += distance; 1412 } 1413 1414 filename_size = archive_le16dec(file_header.name_size); 1415 if (p + filename_size > endp) { 1416 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1417 "Invalid filename size"); 1418 return (ARCHIVE_FATAL); 1419 } 1420 if (rar->filename_allocated < filename_size * 2 + 2) { 1421 char *newptr; 1422 size_t newsize = filename_size * 2 + 2; 1423 newptr = realloc(rar->filename, newsize); 1424 if (newptr == NULL) { 1425 archive_set_error(&a->archive, ENOMEM, 1426 "Couldn't allocate memory."); 1427 return (ARCHIVE_FATAL); 1428 } 1429 rar->filename = newptr; 1430 rar->filename_allocated = newsize; 1431 } 1432 filename = rar->filename; 1433 memcpy(filename, p, filename_size); 1434 filename[filename_size] = '\0'; 1435 if (rar->file_flags & FHD_UNICODE) 1436 { 1437 if (filename_size != strlen(filename)) 1438 { 1439 unsigned char highbyte, flagbits, flagbyte; 1440 unsigned fn_end, offset; 1441 1442 end = filename_size; 1443 fn_end = filename_size * 2; 1444 filename_size = 0; 1445 offset = (unsigned)strlen(filename) + 1; 1446 highbyte = *(p + offset++); 1447 flagbits = 0; 1448 flagbyte = 0; 1449 while (offset < end && filename_size < fn_end) 1450 { 1451 if (!flagbits) 1452 { 1453 flagbyte = *(p + offset++); 1454 flagbits = 8; 1455 } 1456 1457 flagbits -= 2; 1458 switch((flagbyte >> flagbits) & 3) 1459 { 1460 case 0: 1461 filename[filename_size++] = '\0'; 1462 filename[filename_size++] = *(p + offset++); 1463 break; 1464 case 1: 1465 filename[filename_size++] = highbyte; 1466 filename[filename_size++] = *(p + offset++); 1467 break; 1468 case 2: 1469 filename[filename_size++] = *(p + offset + 1); 1470 filename[filename_size++] = *(p + offset); 1471 offset += 2; 1472 break; 1473 case 3: 1474 { 1475 char extra, high; 1476 uint8_t length = *(p + offset++); 1477 1478 if (length & 0x80) { 1479 extra = *(p + offset++); 1480 high = (char)highbyte; 1481 } else 1482 extra = high = 0; 1483 length = (length & 0x7f) + 2; 1484 while (length && filename_size < fn_end) { 1485 unsigned cp = filename_size >> 1; 1486 filename[filename_size++] = high; 1487 filename[filename_size++] = p[cp] + extra; 1488 length--; 1489 } 1490 } 1491 break; 1492 } 1493 } 1494 if (filename_size > fn_end) { 1495 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1496 "Invalid filename"); 1497 return (ARCHIVE_FATAL); 1498 } 1499 filename[filename_size++] = '\0'; 1500 filename[filename_size++] = '\0'; 1501 1502 /* Decoded unicode form is UTF-16BE, so we have to update a string 1503 * conversion object for it. */ 1504 if (rar->sconv_utf16be == NULL) { 1505 rar->sconv_utf16be = archive_string_conversion_from_charset( 1506 &a->archive, "UTF-16BE", 1); 1507 if (rar->sconv_utf16be == NULL) 1508 return (ARCHIVE_FATAL); 1509 } 1510 fn_sconv = rar->sconv_utf16be; 1511 1512 strp = filename; 1513 while (memcmp(strp, "\x00\x00", 2)) 1514 { 1515 if (!memcmp(strp, "\x00\\", 2)) 1516 *(strp + 1) = '/'; 1517 strp += 2; 1518 } 1519 p += offset; 1520 } else { 1521 /* 1522 * If FHD_UNICODE is set but no unicode data, this file name form 1523 * is UTF-8, so we have to update a string conversion object for 1524 * it accordingly. 1525 */ 1526 if (rar->sconv_utf8 == NULL) { 1527 rar->sconv_utf8 = archive_string_conversion_from_charset( 1528 &a->archive, "UTF-8", 1); 1529 if (rar->sconv_utf8 == NULL) 1530 return (ARCHIVE_FATAL); 1531 } 1532 fn_sconv = rar->sconv_utf8; 1533 while ((strp = strchr(filename, '\\')) != NULL) 1534 *strp = '/'; 1535 p += filename_size; 1536 } 1537 } 1538 else 1539 { 1540 fn_sconv = sconv; 1541 while ((strp = strchr(filename, '\\')) != NULL) 1542 *strp = '/'; 1543 p += filename_size; 1544 } 1545 1546 /* Split file in multivolume RAR. No more need to process header. */ 1547 if (rar->filename_save && 1548 filename_size == rar->filename_save_size && 1549 !memcmp(rar->filename, rar->filename_save, filename_size + 1)) 1550 { 1551 __archive_read_consume(a, header_size - 7); 1552 rar->cursor++; 1553 if (rar->cursor >= rar->nodes) 1554 { 1555 rar->nodes++; 1556 if ((rar->dbo = 1557 realloc(rar->dbo, sizeof(*rar->dbo) * rar->nodes)) == NULL) 1558 { 1559 archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); 1560 return (ARCHIVE_FATAL); 1561 } 1562 rar->dbo[rar->cursor].header_size = header_size; 1563 rar->dbo[rar->cursor].start_offset = -1; 1564 rar->dbo[rar->cursor].end_offset = -1; 1565 } 1566 if (rar->dbo[rar->cursor].start_offset < 0) 1567 { 1568 rar->dbo[rar->cursor].start_offset = a->filter->position; 1569 rar->dbo[rar->cursor].end_offset = rar->dbo[rar->cursor].start_offset + 1570 rar->packed_size; 1571 } 1572 return ret; 1573 } 1574 1575 rar->filename_save = (char*)realloc(rar->filename_save, 1576 filename_size + 1); 1577 memcpy(rar->filename_save, rar->filename, filename_size + 1); 1578 rar->filename_save_size = filename_size; 1579 1580 /* Set info for seeking */ 1581 free(rar->dbo); 1582 if ((rar->dbo = calloc(1, sizeof(*rar->dbo))) == NULL) 1583 { 1584 archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); 1585 return (ARCHIVE_FATAL); 1586 } 1587 rar->dbo[0].header_size = header_size; 1588 rar->dbo[0].start_offset = -1; 1589 rar->dbo[0].end_offset = -1; 1590 rar->cursor = 0; 1591 rar->nodes = 1; 1592 1593 if (rar->file_flags & FHD_SALT) 1594 { 1595 if (p + 8 > endp) { 1596 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1597 "Invalid header size"); 1598 return (ARCHIVE_FATAL); 1599 } 1600 memcpy(rar->salt, p, 8); 1601 p += 8; 1602 } 1603 1604 if (rar->file_flags & FHD_EXTTIME) { 1605 if (read_exttime(p, rar, endp) < 0) { 1606 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1607 "Invalid header size"); 1608 return (ARCHIVE_FATAL); 1609 } 1610 } 1611 1612 __archive_read_consume(a, header_size - 7); 1613 rar->dbo[0].start_offset = a->filter->position; 1614 rar->dbo[0].end_offset = rar->dbo[0].start_offset + rar->packed_size; 1615 1616 switch(file_header.host_os) 1617 { 1618 case OS_MSDOS: 1619 case OS_OS2: 1620 case OS_WIN32: 1621 rar->mode = archive_le32dec(file_header.file_attr); 1622 if (rar->mode & FILE_ATTRIBUTE_DIRECTORY) 1623 rar->mode = AE_IFDIR | S_IXUSR | S_IXGRP | S_IXOTH; 1624 else 1625 rar->mode = AE_IFREG; 1626 rar->mode |= S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH; 1627 break; 1628 1629 case OS_UNIX: 1630 case OS_MAC_OS: 1631 case OS_BEOS: 1632 rar->mode = archive_le32dec(file_header.file_attr); 1633 break; 1634 1635 default: 1636 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1637 "Unknown file attributes from RAR file's host OS"); 1638 return (ARCHIVE_FATAL); 1639 } 1640 1641 rar->bytes_uncopied = rar->bytes_unconsumed = 0; 1642 rar->lzss.position = rar->offset = 0; 1643 rar->offset_seek = 0; 1644 rar->dictionary_size = 0; 1645 rar->offset_outgoing = 0; 1646 rar->br.cache_avail = 0; 1647 rar->br.avail_in = 0; 1648 rar->crc_calculated = 0; 1649 rar->entry_eof = 0; 1650 rar->valid = 1; 1651 rar->is_ppmd_block = 0; 1652 rar->start_new_table = 1; 1653 free(rar->unp_buffer); 1654 rar->unp_buffer = NULL; 1655 rar->unp_offset = 0; 1656 rar->unp_buffer_size = UNP_BUFFER_SIZE; 1657 memset(rar->lengthtable, 0, sizeof(rar->lengthtable)); 1658 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 1659 rar->ppmd_valid = rar->ppmd_eod = 0; 1660 1661 /* Don't set any archive entries for non-file header types */ 1662 if (head_type == NEWSUB_HEAD) 1663 return ret; 1664 1665 archive_entry_set_mtime(entry, rar->mtime, rar->mnsec); 1666 archive_entry_set_ctime(entry, rar->ctime, rar->cnsec); 1667 archive_entry_set_atime(entry, rar->atime, rar->ansec); 1668 archive_entry_set_size(entry, rar->unp_size); 1669 archive_entry_set_mode(entry, rar->mode); 1670 1671 if (archive_entry_copy_pathname_l(entry, filename, filename_size, fn_sconv)) 1672 { 1673 if (errno == ENOMEM) 1674 { 1675 archive_set_error(&a->archive, ENOMEM, 1676 "Can't allocate memory for Pathname"); 1677 return (ARCHIVE_FATAL); 1678 } 1679 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1680 "Pathname cannot be converted from %s to current locale.", 1681 archive_string_conversion_charset_name(fn_sconv)); 1682 ret = (ARCHIVE_WARN); 1683 } 1684 1685 if (((rar->mode) & AE_IFMT) == AE_IFLNK) 1686 { 1687 /* Make sure a symbolic-link file does not have its body. */ 1688 rar->bytes_remaining = 0; 1689 archive_entry_set_size(entry, 0); 1690 1691 /* Read a symbolic-link name. */ 1692 if ((ret2 = read_symlink_stored(a, entry, sconv)) < (ARCHIVE_WARN)) 1693 return ret2; 1694 if (ret > ret2) 1695 ret = ret2; 1696 } 1697 1698 if (rar->bytes_remaining == 0) 1699 rar->entry_eof = 1; 1700 1701 return ret; 1702} 1703 1704static time_t 1705get_time(int ttime) 1706{ 1707 struct tm tm; 1708 tm.tm_sec = 2 * (ttime & 0x1f); 1709 tm.tm_min = (ttime >> 5) & 0x3f; 1710 tm.tm_hour = (ttime >> 11) & 0x1f; 1711 tm.tm_mday = (ttime >> 16) & 0x1f; 1712 tm.tm_mon = ((ttime >> 21) & 0x0f) - 1; 1713 tm.tm_year = ((ttime >> 25) & 0x7f) + 80; 1714 tm.tm_isdst = -1; 1715 return mktime(&tm); 1716} 1717 1718static int 1719read_exttime(const char *p, struct rar *rar, const char *endp) 1720{ 1721 unsigned rmode, flags, rem, j, count; 1722 int ttime, i; 1723 struct tm *tm; 1724 time_t t; 1725 long nsec; 1726 1727 if (p + 2 > endp) 1728 return (-1); 1729 flags = archive_le16dec(p); 1730 p += 2; 1731 1732 for (i = 3; i >= 0; i--) 1733 { 1734 t = 0; 1735 if (i == 3) 1736 t = rar->mtime; 1737 rmode = flags >> i * 4; 1738 if (rmode & 8) 1739 { 1740 if (!t) 1741 { 1742 if (p + 4 > endp) 1743 return (-1); 1744 ttime = archive_le32dec(p); 1745 t = get_time(ttime); 1746 p += 4; 1747 } 1748 rem = 0; 1749 count = rmode & 3; 1750 if (p + count > endp) 1751 return (-1); 1752 for (j = 0; j < count; j++) 1753 { 1754 rem = ((*p) << 16) | (rem >> 8); 1755 p++; 1756 } 1757 tm = localtime(&t); 1758 nsec = tm->tm_sec + rem / NS_UNIT; 1759 if (rmode & 4) 1760 { 1761 tm->tm_sec++; 1762 t = mktime(tm); 1763 } 1764 if (i == 3) 1765 { 1766 rar->mtime = t; 1767 rar->mnsec = nsec; 1768 } 1769 else if (i == 2) 1770 { 1771 rar->ctime = t; 1772 rar->cnsec = nsec; 1773 } 1774 else if (i == 1) 1775 { 1776 rar->atime = t; 1777 rar->ansec = nsec; 1778 } 1779 else 1780 { 1781 rar->arctime = t; 1782 rar->arcnsec = nsec; 1783 } 1784 } 1785 } 1786 return (0); 1787} 1788 1789static int 1790read_symlink_stored(struct archive_read *a, struct archive_entry *entry, 1791 struct archive_string_conv *sconv) 1792{ 1793 const void *h; 1794 const char *p; 1795 struct rar *rar; 1796 int ret = (ARCHIVE_OK); 1797 1798 rar = (struct rar *)(a->format->data); 1799 if ((h = rar_read_ahead(a, (size_t)rar->packed_size, NULL)) == NULL) 1800 return (ARCHIVE_FATAL); 1801 p = h; 1802 1803 if (archive_entry_copy_symlink_l(entry, 1804 p, (size_t)rar->packed_size, sconv)) 1805 { 1806 if (errno == ENOMEM) 1807 { 1808 archive_set_error(&a->archive, ENOMEM, 1809 "Can't allocate memory for link"); 1810 return (ARCHIVE_FATAL); 1811 } 1812 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1813 "link cannot be converted from %s to current locale.", 1814 archive_string_conversion_charset_name(sconv)); 1815 ret = (ARCHIVE_WARN); 1816 } 1817 __archive_read_consume(a, rar->packed_size); 1818 return ret; 1819} 1820 1821static int 1822read_data_stored(struct archive_read *a, const void **buff, size_t *size, 1823 int64_t *offset) 1824{ 1825 struct rar *rar; 1826 ssize_t bytes_avail; 1827 1828 rar = (struct rar *)(a->format->data); 1829 if (rar->bytes_remaining == 0 && 1830 !(rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER)) 1831 { 1832 *buff = NULL; 1833 *size = 0; 1834 *offset = rar->offset; 1835 if (rar->file_crc != rar->crc_calculated) { 1836 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1837 "File CRC error"); 1838 return (ARCHIVE_FATAL); 1839 } 1840 rar->entry_eof = 1; 1841 return (ARCHIVE_EOF); 1842 } 1843 1844 *buff = rar_read_ahead(a, 1, &bytes_avail); 1845 if (bytes_avail <= 0) 1846 { 1847 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1848 "Truncated RAR file data"); 1849 return (ARCHIVE_FATAL); 1850 } 1851 1852 *size = bytes_avail; 1853 *offset = rar->offset; 1854 rar->offset += bytes_avail; 1855 rar->offset_seek += bytes_avail; 1856 rar->bytes_remaining -= bytes_avail; 1857 rar->bytes_unconsumed = bytes_avail; 1858 /* Calculate File CRC. */ 1859 rar->crc_calculated = crc32(rar->crc_calculated, *buff, 1860 (unsigned)bytes_avail); 1861 return (ARCHIVE_OK); 1862} 1863 1864static int 1865read_data_compressed(struct archive_read *a, const void **buff, size_t *size, 1866 int64_t *offset) 1867{ 1868 struct rar *rar; 1869 int64_t start, end, actualend; 1870 size_t bs; 1871 int ret = (ARCHIVE_OK), sym, code, lzss_offset, length, i; 1872 1873 rar = (struct rar *)(a->format->data); 1874 1875 do { 1876 if (!rar->valid) 1877 return (ARCHIVE_FATAL); 1878 if (rar->ppmd_eod || 1879 (rar->dictionary_size && rar->offset >= rar->unp_size)) 1880 { 1881 if (rar->unp_offset > 0) { 1882 /* 1883 * We have unprocessed extracted data. write it out. 1884 */ 1885 *buff = rar->unp_buffer; 1886 *size = rar->unp_offset; 1887 *offset = rar->offset_outgoing; 1888 rar->offset_outgoing += *size; 1889 /* Calculate File CRC. */ 1890 rar->crc_calculated = crc32(rar->crc_calculated, *buff, 1891 (unsigned)*size); 1892 rar->unp_offset = 0; 1893 return (ARCHIVE_OK); 1894 } 1895 *buff = NULL; 1896 *size = 0; 1897 *offset = rar->offset; 1898 if (rar->file_crc != rar->crc_calculated) { 1899 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1900 "File CRC error"); 1901 return (ARCHIVE_FATAL); 1902 } 1903 rar->entry_eof = 1; 1904 return (ARCHIVE_EOF); 1905 } 1906 1907 if (!rar->is_ppmd_block && rar->dictionary_size && rar->bytes_uncopied > 0) 1908 { 1909 if (rar->bytes_uncopied > (rar->unp_buffer_size - rar->unp_offset)) 1910 bs = rar->unp_buffer_size - rar->unp_offset; 1911 else 1912 bs = (size_t)rar->bytes_uncopied; 1913 ret = copy_from_lzss_window(a, buff, rar->offset, (int)bs); 1914 if (ret != ARCHIVE_OK) 1915 return (ret); 1916 rar->offset += bs; 1917 rar->bytes_uncopied -= bs; 1918 if (*buff != NULL) { 1919 rar->unp_offset = 0; 1920 *size = rar->unp_buffer_size; 1921 *offset = rar->offset_outgoing; 1922 rar->offset_outgoing += *size; 1923 /* Calculate File CRC. */ 1924 rar->crc_calculated = crc32(rar->crc_calculated, *buff, 1925 (unsigned)*size); 1926 return (ret); 1927 } 1928 continue; 1929 } 1930 1931 if (!rar->br.next_in && 1932 (ret = rar_br_preparation(a, &(rar->br))) < ARCHIVE_WARN) 1933 return (ret); 1934 if (rar->start_new_table && ((ret = parse_codes(a)) < (ARCHIVE_WARN))) 1935 return (ret); 1936 1937 if (rar->is_ppmd_block) 1938 { 1939 if ((sym = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1940 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1941 { 1942 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1943 "Invalid symbol"); 1944 return (ARCHIVE_FATAL); 1945 } 1946 if(sym != rar->ppmd_escape) 1947 { 1948 lzss_emit_literal(rar, sym); 1949 rar->bytes_uncopied++; 1950 } 1951 else 1952 { 1953 if ((code = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1954 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1955 { 1956 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1957 "Invalid symbol"); 1958 return (ARCHIVE_FATAL); 1959 } 1960 1961 switch(code) 1962 { 1963 case 0: 1964 rar->start_new_table = 1; 1965 return read_data_compressed(a, buff, size, offset); 1966 1967 case 2: 1968 rar->ppmd_eod = 1;/* End Of ppmd Data. */ 1969 continue; 1970 1971 case 3: 1972 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1973 "Parsing filters is unsupported."); 1974 return (ARCHIVE_FAILED); 1975 1976 case 4: 1977 lzss_offset = 0; 1978 for (i = 2; i >= 0; i--) 1979 { 1980 if ((code = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1981 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1982 { 1983 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1984 "Invalid symbol"); 1985 return (ARCHIVE_FATAL); 1986 } 1987 lzss_offset |= code << (i * 8); 1988 } 1989 if ((length = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1990 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1991 { 1992 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1993 "Invalid symbol"); 1994 return (ARCHIVE_FATAL); 1995 } 1996 lzss_emit_match(rar, lzss_offset + 2, length + 32); 1997 rar->bytes_uncopied += length + 32; 1998 break; 1999 2000 case 5: 2001 if ((length = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 2002 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 2003 { 2004 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2005 "Invalid symbol"); 2006 return (ARCHIVE_FATAL); 2007 } 2008 lzss_emit_match(rar, 1, length + 4); 2009 rar->bytes_uncopied += length + 4; 2010 break; 2011 2012 default: 2013 lzss_emit_literal(rar, sym); 2014 rar->bytes_uncopied++; 2015 } 2016 } 2017 } 2018 else 2019 { 2020 start = rar->offset; 2021 end = start + rar->dictionary_size; 2022 rar->filterstart = INT64_MAX; 2023 2024 if ((actualend = expand(a, end)) < 0) 2025 return ((int)actualend); 2026 2027 rar->bytes_uncopied = actualend - start; 2028 if (rar->bytes_uncopied == 0) { 2029 /* Broken RAR files cause this case. 2030 * NOTE: If this case were possible on a normal RAR file 2031 * we would find out where it was actually bad and 2032 * what we would do to solve it. */ 2033 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2034 "Internal error extracting RAR file"); 2035 return (ARCHIVE_FATAL); 2036 } 2037 } 2038 if (rar->bytes_uncopied > (rar->unp_buffer_size - rar->unp_offset)) 2039 bs = rar->unp_buffer_size - rar->unp_offset; 2040 else 2041 bs = (size_t)rar->bytes_uncopied; 2042 ret = copy_from_lzss_window(a, buff, rar->offset, (int)bs); 2043 if (ret != ARCHIVE_OK) 2044 return (ret); 2045 rar->offset += bs; 2046 rar->bytes_uncopied -= bs; 2047 /* 2048 * If *buff is NULL, it means unp_buffer is not full. 2049 * So we have to continue extracting a RAR file. 2050 */ 2051 } while (*buff == NULL); 2052 2053 rar->unp_offset = 0; 2054 *size = rar->unp_buffer_size; 2055 *offset = rar->offset_outgoing; 2056 rar->offset_outgoing += *size; 2057 /* Calculate File CRC. */ 2058 rar->crc_calculated = crc32(rar->crc_calculated, *buff, (unsigned)*size); 2059 return ret; 2060} 2061 2062static int 2063parse_codes(struct archive_read *a) 2064{ 2065 int i, j, val, n, r; 2066 unsigned char bitlengths[MAX_SYMBOLS], zerocount, ppmd_flags; 2067 unsigned int maxorder; 2068 struct huffman_code precode; 2069 struct rar *rar = (struct rar *)(a->format->data); 2070 struct rar_br *br = &(rar->br); 2071 2072 free_codes(a); 2073 2074 /* Skip to the next byte */ 2075 rar_br_consume_unalined_bits(br); 2076 2077 /* PPMd block flag */ 2078 if (!rar_br_read_ahead(a, br, 1)) 2079 goto truncated_data; 2080 if ((rar->is_ppmd_block = rar_br_bits(br, 1)) != 0) 2081 { 2082 rar_br_consume(br, 1); 2083 if (!rar_br_read_ahead(a, br, 7)) 2084 goto truncated_data; 2085 ppmd_flags = rar_br_bits(br, 7); 2086 rar_br_consume(br, 7); 2087 2088 /* Memory is allocated in MB */ 2089 if (ppmd_flags & 0x20) 2090 { 2091 if (!rar_br_read_ahead(a, br, 8)) 2092 goto truncated_data; 2093 rar->dictionary_size = (rar_br_bits(br, 8) + 1) << 20; 2094 rar_br_consume(br, 8); 2095 } 2096 2097 if (ppmd_flags & 0x40) 2098 { 2099 if (!rar_br_read_ahead(a, br, 8)) 2100 goto truncated_data; 2101 rar->ppmd_escape = rar->ppmd7_context.InitEsc = rar_br_bits(br, 8); 2102 rar_br_consume(br, 8); 2103 } 2104 else 2105 rar->ppmd_escape = 2; 2106 2107 if (ppmd_flags & 0x20) 2108 { 2109 maxorder = (ppmd_flags & 0x1F) + 1; 2110 if(maxorder > 16) 2111 maxorder = 16 + (maxorder - 16) * 3; 2112 2113 if (maxorder == 1) 2114 { 2115 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2116 "Truncated RAR file data"); 2117 return (ARCHIVE_FATAL); 2118 } 2119 2120 /* Make sure ppmd7_contest is freed before Ppmd7_Construct 2121 * because reading a broken file cause this abnormal sequence. */ 2122 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 2123 2124 rar->bytein.a = a; 2125 rar->bytein.Read = &ppmd_read; 2126 __archive_ppmd7_functions.PpmdRAR_RangeDec_CreateVTable(&rar->range_dec); 2127 rar->range_dec.Stream = &rar->bytein; 2128 __archive_ppmd7_functions.Ppmd7_Construct(&rar->ppmd7_context); 2129 2130 if (rar->dictionary_size == 0) { 2131 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2132 "Invalid zero dictionary size"); 2133 return (ARCHIVE_FATAL); 2134 } 2135 2136 if (!__archive_ppmd7_functions.Ppmd7_Alloc(&rar->ppmd7_context, 2137 rar->dictionary_size, &g_szalloc)) 2138 { 2139 archive_set_error(&a->archive, ENOMEM, 2140 "Out of memory"); 2141 return (ARCHIVE_FATAL); 2142 } 2143 if (!__archive_ppmd7_functions.PpmdRAR_RangeDec_Init(&rar->range_dec)) 2144 { 2145 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2146 "Unable to initialize PPMd range decoder"); 2147 return (ARCHIVE_FATAL); 2148 } 2149 __archive_ppmd7_functions.Ppmd7_Init(&rar->ppmd7_context, maxorder); 2150 rar->ppmd_valid = 1; 2151 } 2152 else 2153 { 2154 if (!rar->ppmd_valid) { 2155 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2156 "Invalid PPMd sequence"); 2157 return (ARCHIVE_FATAL); 2158 } 2159 if (!__archive_ppmd7_functions.PpmdRAR_RangeDec_Init(&rar->range_dec)) 2160 { 2161 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2162 "Unable to initialize PPMd range decoder"); 2163 return (ARCHIVE_FATAL); 2164 } 2165 } 2166 } 2167 else 2168 { 2169 rar_br_consume(br, 1); 2170 2171 /* Keep existing table flag */ 2172 if (!rar_br_read_ahead(a, br, 1)) 2173 goto truncated_data; 2174 if (!rar_br_bits(br, 1)) 2175 memset(rar->lengthtable, 0, sizeof(rar->lengthtable)); 2176 rar_br_consume(br, 1); 2177 2178 memset(&bitlengths, 0, sizeof(bitlengths)); 2179 for (i = 0; i < MAX_SYMBOLS;) 2180 { 2181 if (!rar_br_read_ahead(a, br, 4)) 2182 goto truncated_data; 2183 bitlengths[i++] = rar_br_bits(br, 4); 2184 rar_br_consume(br, 4); 2185 if (bitlengths[i-1] == 0xF) 2186 { 2187 if (!rar_br_read_ahead(a, br, 4)) 2188 goto truncated_data; 2189 zerocount = rar_br_bits(br, 4); 2190 rar_br_consume(br, 4); 2191 if (zerocount) 2192 { 2193 i--; 2194 for (j = 0; j < zerocount + 2 && i < MAX_SYMBOLS; j++) 2195 bitlengths[i++] = 0; 2196 } 2197 } 2198 } 2199 2200 memset(&precode, 0, sizeof(precode)); 2201 r = create_code(a, &precode, bitlengths, MAX_SYMBOLS, MAX_SYMBOL_LENGTH); 2202 if (r != ARCHIVE_OK) { 2203 free(precode.tree); 2204 free(precode.table); 2205 return (r); 2206 } 2207 2208 for (i = 0; i < HUFFMAN_TABLE_SIZE;) 2209 { 2210 if ((val = read_next_symbol(a, &precode)) < 0) { 2211 free(precode.tree); 2212 free(precode.table); 2213 return (ARCHIVE_FATAL); 2214 } 2215 if (val < 16) 2216 { 2217 rar->lengthtable[i] = (rar->lengthtable[i] + val) & 0xF; 2218 i++; 2219 } 2220 else if (val < 18) 2221 { 2222 if (i == 0) 2223 { 2224 free(precode.tree); 2225 free(precode.table); 2226 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2227 "Internal error extracting RAR file."); 2228 return (ARCHIVE_FATAL); 2229 } 2230 2231 if(val == 16) { 2232 if (!rar_br_read_ahead(a, br, 3)) { 2233 free(precode.tree); 2234 free(precode.table); 2235 goto truncated_data; 2236 } 2237 n = rar_br_bits(br, 3) + 3; 2238 rar_br_consume(br, 3); 2239 } else { 2240 if (!rar_br_read_ahead(a, br, 7)) { 2241 free(precode.tree); 2242 free(precode.table); 2243 goto truncated_data; 2244 } 2245 n = rar_br_bits(br, 7) + 11; 2246 rar_br_consume(br, 7); 2247 } 2248 2249 for (j = 0; j < n && i < HUFFMAN_TABLE_SIZE; j++) 2250 { 2251 rar->lengthtable[i] = rar->lengthtable[i-1]; 2252 i++; 2253 } 2254 } 2255 else 2256 { 2257 if(val == 18) { 2258 if (!rar_br_read_ahead(a, br, 3)) { 2259 free(precode.tree); 2260 free(precode.table); 2261 goto truncated_data; 2262 } 2263 n = rar_br_bits(br, 3) + 3; 2264 rar_br_consume(br, 3); 2265 } else { 2266 if (!rar_br_read_ahead(a, br, 7)) { 2267 free(precode.tree); 2268 free(precode.table); 2269 goto truncated_data; 2270 } 2271 n = rar_br_bits(br, 7) + 11; 2272 rar_br_consume(br, 7); 2273 } 2274 2275 for(j = 0; j < n && i < HUFFMAN_TABLE_SIZE; j++) 2276 rar->lengthtable[i++] = 0; 2277 } 2278 } 2279 free(precode.tree); 2280 free(precode.table); 2281 2282 r = create_code(a, &rar->maincode, &rar->lengthtable[0], MAINCODE_SIZE, 2283 MAX_SYMBOL_LENGTH); 2284 if (r != ARCHIVE_OK) 2285 return (r); 2286 r = create_code(a, &rar->offsetcode, &rar->lengthtable[MAINCODE_SIZE], 2287 OFFSETCODE_SIZE, MAX_SYMBOL_LENGTH); 2288 if (r != ARCHIVE_OK) 2289 return (r); 2290 r = create_code(a, &rar->lowoffsetcode, 2291 &rar->lengthtable[MAINCODE_SIZE + OFFSETCODE_SIZE], 2292 LOWOFFSETCODE_SIZE, MAX_SYMBOL_LENGTH); 2293 if (r != ARCHIVE_OK) 2294 return (r); 2295 r = create_code(a, &rar->lengthcode, 2296 &rar->lengthtable[MAINCODE_SIZE + OFFSETCODE_SIZE + 2297 LOWOFFSETCODE_SIZE], LENGTHCODE_SIZE, MAX_SYMBOL_LENGTH); 2298 if (r != ARCHIVE_OK) 2299 return (r); 2300 } 2301 2302 if (!rar->dictionary_size || !rar->lzss.window) 2303 { 2304 /* Seems as though dictionary sizes are not used. Even so, minimize 2305 * memory usage as much as possible. 2306 */ 2307 void *new_window; 2308 unsigned int new_size; 2309 2310 if (rar->unp_size >= DICTIONARY_MAX_SIZE) 2311 new_size = DICTIONARY_MAX_SIZE; 2312 else 2313 new_size = rar_fls((unsigned int)rar->unp_size) << 1; 2314 new_window = realloc(rar->lzss.window, new_size); 2315 if (new_window == NULL) { 2316 archive_set_error(&a->archive, ENOMEM, 2317 "Unable to allocate memory for uncompressed data."); 2318 return (ARCHIVE_FATAL); 2319 } 2320 rar->lzss.window = (unsigned char *)new_window; 2321 rar->dictionary_size = new_size; 2322 memset(rar->lzss.window, 0, rar->dictionary_size); 2323 rar->lzss.mask = rar->dictionary_size - 1; 2324 } 2325 2326 rar->start_new_table = 0; 2327 return (ARCHIVE_OK); 2328truncated_data: 2329 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2330 "Truncated RAR file data"); 2331 rar->valid = 0; 2332 return (ARCHIVE_FATAL); 2333} 2334 2335static void 2336free_codes(struct archive_read *a) 2337{ 2338 struct rar *rar = (struct rar *)(a->format->data); 2339 free(rar->maincode.tree); 2340 free(rar->offsetcode.tree); 2341 free(rar->lowoffsetcode.tree); 2342 free(rar->lengthcode.tree); 2343 free(rar->maincode.table); 2344 free(rar->offsetcode.table); 2345 free(rar->lowoffsetcode.table); 2346 free(rar->lengthcode.table); 2347 memset(&rar->maincode, 0, sizeof(rar->maincode)); 2348 memset(&rar->offsetcode, 0, sizeof(rar->offsetcode)); 2349 memset(&rar->lowoffsetcode, 0, sizeof(rar->lowoffsetcode)); 2350 memset(&rar->lengthcode, 0, sizeof(rar->lengthcode)); 2351} 2352 2353 2354static int 2355read_next_symbol(struct archive_read *a, struct huffman_code *code) 2356{ 2357 unsigned char bit; 2358 unsigned int bits; 2359 int length, value, node; 2360 struct rar *rar; 2361 struct rar_br *br; 2362 2363 if (!code->table) 2364 { 2365 if (make_table(a, code) != (ARCHIVE_OK)) 2366 return -1; 2367 } 2368 2369 rar = (struct rar *)(a->format->data); 2370 br = &(rar->br); 2371 2372 /* Look ahead (peek) at bits */ 2373 if (!rar_br_read_ahead(a, br, code->tablesize)) { 2374 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2375 "Truncated RAR file data"); 2376 rar->valid = 0; 2377 return -1; 2378 } 2379 bits = rar_br_bits(br, code->tablesize); 2380 2381 length = code->table[bits].length; 2382 value = code->table[bits].value; 2383 2384 if (length < 0) 2385 { 2386 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2387 "Invalid prefix code in bitstream"); 2388 return -1; 2389 } 2390 2391 if (length <= code->tablesize) 2392 { 2393 /* Skip length bits */ 2394 rar_br_consume(br, length); 2395 return value; 2396 } 2397 2398 /* Skip tablesize bits */ 2399 rar_br_consume(br, code->tablesize); 2400 2401 node = value; 2402 while (!(code->tree[node].branches[0] == 2403 code->tree[node].branches[1])) 2404 { 2405 if (!rar_br_read_ahead(a, br, 1)) { 2406 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2407 "Truncated RAR file data"); 2408 rar->valid = 0; 2409 return -1; 2410 } 2411 bit = rar_br_bits(br, 1); 2412 rar_br_consume(br, 1); 2413 2414 if (code->tree[node].branches[bit] < 0) 2415 { 2416 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2417 "Invalid prefix code in bitstream"); 2418 return -1; 2419 } 2420 node = code->tree[node].branches[bit]; 2421 } 2422 2423 return code->tree[node].branches[0]; 2424} 2425 2426static int 2427create_code(struct archive_read *a, struct huffman_code *code, 2428 unsigned char *lengths, int numsymbols, char maxlength) 2429{ 2430 int i, j, codebits = 0, symbolsleft = numsymbols; 2431 2432 code->numentries = 0; 2433 code->numallocatedentries = 0; 2434 if (new_node(code) < 0) { 2435 archive_set_error(&a->archive, ENOMEM, 2436 "Unable to allocate memory for node data."); 2437 return (ARCHIVE_FATAL); 2438 } 2439 code->numentries = 1; 2440 code->minlength = INT_MAX; 2441 code->maxlength = INT_MIN; 2442 codebits = 0; 2443 for(i = 1; i <= maxlength; i++) 2444 { 2445 for(j = 0; j < numsymbols; j++) 2446 { 2447 if (lengths[j] != i) continue; 2448 if (add_value(a, code, j, codebits, i) != ARCHIVE_OK) 2449 return (ARCHIVE_FATAL); 2450 codebits++; 2451 if (--symbolsleft <= 0) { break; break; } 2452 } 2453 codebits <<= 1; 2454 } 2455 return (ARCHIVE_OK); 2456} 2457 2458static int 2459add_value(struct archive_read *a, struct huffman_code *code, int value, 2460 int codebits, int length) 2461{ 2462 int repeatpos, lastnode, bitpos, bit, repeatnode, nextnode; 2463 2464 free(code->table); 2465 code->table = NULL; 2466 2467 if(length > code->maxlength) 2468 code->maxlength = length; 2469 if(length < code->minlength) 2470 code->minlength = length; 2471 2472 repeatpos = -1; 2473 if (repeatpos == 0 || (repeatpos >= 0 2474 && (((codebits >> (repeatpos - 1)) & 3) == 0 2475 || ((codebits >> (repeatpos - 1)) & 3) == 3))) 2476 { 2477 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2478 "Invalid repeat position"); 2479 return (ARCHIVE_FATAL); 2480 } 2481 2482 lastnode = 0; 2483 for (bitpos = length - 1; bitpos >= 0; bitpos--) 2484 { 2485 bit = (codebits >> bitpos) & 1; 2486 2487 /* Leaf node check */ 2488 if (code->tree[lastnode].branches[0] == 2489 code->tree[lastnode].branches[1]) 2490 { 2491 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2492 "Prefix found"); 2493 return (ARCHIVE_FATAL); 2494 } 2495 2496 if (bitpos == repeatpos) 2497 { 2498 /* Open branch check */ 2499 if (!(code->tree[lastnode].branches[bit] < 0)) 2500 { 2501 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2502 "Invalid repeating code"); 2503 return (ARCHIVE_FATAL); 2504 } 2505 2506 if ((repeatnode = new_node(code)) < 0) { 2507 archive_set_error(&a->archive, ENOMEM, 2508 "Unable to allocate memory for node data."); 2509 return (ARCHIVE_FATAL); 2510 } 2511 if ((nextnode = new_node(code)) < 0) { 2512 archive_set_error(&a->archive, ENOMEM, 2513 "Unable to allocate memory for node data."); 2514 return (ARCHIVE_FATAL); 2515 } 2516 2517 /* Set branches */ 2518 code->tree[lastnode].branches[bit] = repeatnode; 2519 code->tree[repeatnode].branches[bit] = repeatnode; 2520 code->tree[repeatnode].branches[bit^1] = nextnode; 2521 lastnode = nextnode; 2522 2523 bitpos++; /* terminating bit already handled, skip it */ 2524 } 2525 else 2526 { 2527 /* Open branch check */ 2528 if (code->tree[lastnode].branches[bit] < 0) 2529 { 2530 if (new_node(code) < 0) { 2531 archive_set_error(&a->archive, ENOMEM, 2532 "Unable to allocate memory for node data."); 2533 return (ARCHIVE_FATAL); 2534 } 2535 code->tree[lastnode].branches[bit] = code->numentries++; 2536 } 2537 2538 /* set to branch */ 2539 lastnode = code->tree[lastnode].branches[bit]; 2540 } 2541 } 2542 2543 if (!(code->tree[lastnode].branches[0] == -1 2544 && code->tree[lastnode].branches[1] == -2)) 2545 { 2546 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2547 "Prefix found"); 2548 return (ARCHIVE_FATAL); 2549 } 2550 2551 /* Set leaf value */ 2552 code->tree[lastnode].branches[0] = value; 2553 code->tree[lastnode].branches[1] = value; 2554 2555 return (ARCHIVE_OK); 2556} 2557 2558static int 2559new_node(struct huffman_code *code) 2560{ 2561 void *new_tree; 2562 if (code->numallocatedentries == code->numentries) { 2563 int new_num_entries = 256; 2564 if (code->numentries > 0) { 2565 new_num_entries = code->numentries * 2; 2566 } 2567 new_tree = realloc(code->tree, new_num_entries * sizeof(*code->tree)); 2568 if (new_tree == NULL) 2569 return (-1); 2570 code->tree = (struct huffman_tree_node *)new_tree; 2571 code->numallocatedentries = new_num_entries; 2572 } 2573 code->tree[code->numentries].branches[0] = -1; 2574 code->tree[code->numentries].branches[1] = -2; 2575 return 1; 2576} 2577 2578static int 2579make_table(struct archive_read *a, struct huffman_code *code) 2580{ 2581 if (code->maxlength < code->minlength || code->maxlength > 10) 2582 code->tablesize = 10; 2583 else 2584 code->tablesize = code->maxlength; 2585 2586 code->table = 2587 (struct huffman_table_entry *)calloc(1, sizeof(*code->table) 2588 * ((size_t)1 << code->tablesize)); 2589 2590 return make_table_recurse(a, code, 0, code->table, 0, code->tablesize); 2591} 2592 2593static int 2594make_table_recurse(struct archive_read *a, struct huffman_code *code, int node, 2595 struct huffman_table_entry *table, int depth, 2596 int maxdepth) 2597{ 2598 int currtablesize, i, ret = (ARCHIVE_OK); 2599 2600 if (!code->tree) 2601 { 2602 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2603 "Huffman tree was not created."); 2604 return (ARCHIVE_FATAL); 2605 } 2606 if (node < 0 || node >= code->numentries) 2607 { 2608 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2609 "Invalid location to Huffman tree specified."); 2610 return (ARCHIVE_FATAL); 2611 } 2612 2613 currtablesize = 1 << (maxdepth - depth); 2614 2615 if (code->tree[node].branches[0] == 2616 code->tree[node].branches[1]) 2617 { 2618 for(i = 0; i < currtablesize; i++) 2619 { 2620 table[i].length = depth; 2621 table[i].value = code->tree[node].branches[0]; 2622 } 2623 } 2624 else if (node < 0) 2625 { 2626 for(i = 0; i < currtablesize; i++) 2627 table[i].length = -1; 2628 } 2629 else 2630 { 2631 if(depth == maxdepth) 2632 { 2633 table[0].length = maxdepth + 1; 2634 table[0].value = node; 2635 } 2636 else 2637 { 2638 ret |= make_table_recurse(a, code, code->tree[node].branches[0], table, 2639 depth + 1, maxdepth); 2640 ret |= make_table_recurse(a, code, code->tree[node].branches[1], 2641 table + currtablesize / 2, depth + 1, maxdepth); 2642 } 2643 } 2644 return ret; 2645} 2646 2647static int64_t 2648expand(struct archive_read *a, int64_t end) 2649{ 2650 static const unsigned char lengthbases[] = 2651 { 0, 1, 2, 3, 4, 5, 6, 2652 7, 8, 10, 12, 14, 16, 20, 2653 24, 28, 32, 40, 48, 56, 64, 2654 80, 96, 112, 128, 160, 192, 224 }; 2655 static const unsigned char lengthbits[] = 2656 { 0, 0, 0, 0, 0, 0, 0, 2657 0, 1, 1, 1, 1, 2, 2, 2658 2, 2, 3, 3, 3, 3, 4, 2659 4, 4, 4, 5, 5, 5, 5 }; 2660 static const unsigned int offsetbases[] = 2661 { 0, 1, 2, 3, 4, 6, 2662 8, 12, 16, 24, 32, 48, 2663 64, 96, 128, 192, 256, 384, 2664 512, 768, 1024, 1536, 2048, 3072, 2665 4096, 6144, 8192, 12288, 16384, 24576, 2666 32768, 49152, 65536, 98304, 131072, 196608, 2667 262144, 327680, 393216, 458752, 524288, 589824, 2668 655360, 720896, 786432, 851968, 917504, 983040, 2669 1048576, 1310720, 1572864, 1835008, 2097152, 2359296, 2670 2621440, 2883584, 3145728, 3407872, 3670016, 3932160 }; 2671 static const unsigned char offsetbits[] = 2672 { 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 2673 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 2674 11, 11, 12, 12, 13, 13, 14, 14, 15, 15, 16, 16, 2675 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 2676 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18 }; 2677 static const unsigned char shortbases[] = 2678 { 0, 4, 8, 16, 32, 64, 128, 192 }; 2679 static const unsigned char shortbits[] = 2680 { 2, 2, 3, 4, 5, 6, 6, 6 }; 2681 2682 int symbol, offs, len, offsindex, lensymbol, i, offssymbol, lowoffsetsymbol; 2683 unsigned char newfile; 2684 struct rar *rar = (struct rar *)(a->format->data); 2685 struct rar_br *br = &(rar->br); 2686 2687 if (rar->filterstart < end) 2688 end = rar->filterstart; 2689 2690 while (1) 2691 { 2692 if (rar->output_last_match && 2693 lzss_position(&rar->lzss) + rar->lastlength <= end) 2694 { 2695 lzss_emit_match(rar, rar->lastoffset, rar->lastlength); 2696 rar->output_last_match = 0; 2697 } 2698 2699 if(rar->is_ppmd_block || rar->output_last_match || 2700 lzss_position(&rar->lzss) >= end) 2701 return lzss_position(&rar->lzss); 2702 2703 if ((symbol = read_next_symbol(a, &rar->maincode)) < 0) 2704 return (ARCHIVE_FATAL); 2705 rar->output_last_match = 0; 2706 2707 if (symbol < 256) 2708 { 2709 lzss_emit_literal(rar, symbol); 2710 continue; 2711 } 2712 else if (symbol == 256) 2713 { 2714 if (!rar_br_read_ahead(a, br, 1)) 2715 goto truncated_data; 2716 newfile = !rar_br_bits(br, 1); 2717 rar_br_consume(br, 1); 2718 2719 if(newfile) 2720 { 2721 rar->start_new_block = 1; 2722 if (!rar_br_read_ahead(a, br, 1)) 2723 goto truncated_data; 2724 rar->start_new_table = rar_br_bits(br, 1); 2725 rar_br_consume(br, 1); 2726 return lzss_position(&rar->lzss); 2727 } 2728 else 2729 { 2730 if (parse_codes(a) != ARCHIVE_OK) 2731 return (ARCHIVE_FATAL); 2732 continue; 2733 } 2734 } 2735 else if(symbol==257) 2736 { 2737 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 2738 "Parsing filters is unsupported."); 2739 return (ARCHIVE_FAILED); 2740 } 2741 else if(symbol==258) 2742 { 2743 if(rar->lastlength == 0) 2744 continue; 2745 2746 offs = rar->lastoffset; 2747 len = rar->lastlength; 2748 } 2749 else if (symbol <= 262) 2750 { 2751 offsindex = symbol - 259; 2752 offs = rar->oldoffset[offsindex]; 2753 2754 if ((lensymbol = read_next_symbol(a, &rar->lengthcode)) < 0) 2755 goto bad_data; 2756 if (lensymbol > (int)(sizeof(lengthbases)/sizeof(lengthbases[0]))) 2757 goto bad_data; 2758 if (lensymbol > (int)(sizeof(lengthbits)/sizeof(lengthbits[0]))) 2759 goto bad_data; 2760 len = lengthbases[lensymbol] + 2; 2761 if (lengthbits[lensymbol] > 0) { 2762 if (!rar_br_read_ahead(a, br, lengthbits[lensymbol])) 2763 goto truncated_data; 2764 len += rar_br_bits(br, lengthbits[lensymbol]); 2765 rar_br_consume(br, lengthbits[lensymbol]); 2766 } 2767 2768 for (i = offsindex; i > 0; i--) 2769 rar->oldoffset[i] = rar->oldoffset[i-1]; 2770 rar->oldoffset[0] = offs; 2771 } 2772 else if(symbol<=270) 2773 { 2774 offs = shortbases[symbol-263] + 1; 2775 if(shortbits[symbol-263] > 0) { 2776 if (!rar_br_read_ahead(a, br, shortbits[symbol-263])) 2777 goto truncated_data; 2778 offs += rar_br_bits(br, shortbits[symbol-263]); 2779 rar_br_consume(br, shortbits[symbol-263]); 2780 } 2781 2782 len = 2; 2783 2784 for(i = 3; i > 0; i--) 2785 rar->oldoffset[i] = rar->oldoffset[i-1]; 2786 rar->oldoffset[0] = offs; 2787 } 2788 else 2789 { 2790 if (symbol-271 > (int)(sizeof(lengthbases)/sizeof(lengthbases[0]))) 2791 goto bad_data; 2792 if (symbol-271 > (int)(sizeof(lengthbits)/sizeof(lengthbits[0]))) 2793 goto bad_data; 2794 len = lengthbases[symbol-271]+3; 2795 if(lengthbits[symbol-271] > 0) { 2796 if (!rar_br_read_ahead(a, br, lengthbits[symbol-271])) 2797 goto truncated_data; 2798 len += rar_br_bits(br, lengthbits[symbol-271]); 2799 rar_br_consume(br, lengthbits[symbol-271]); 2800 } 2801 2802 if ((offssymbol = read_next_symbol(a, &rar->offsetcode)) < 0) 2803 goto bad_data; 2804 if (offssymbol > (int)(sizeof(offsetbases)/sizeof(offsetbases[0]))) 2805 goto bad_data; 2806 if (offssymbol > (int)(sizeof(offsetbits)/sizeof(offsetbits[0]))) 2807 goto bad_data; 2808 offs = offsetbases[offssymbol]+1; 2809 if(offsetbits[offssymbol] > 0) 2810 { 2811 if(offssymbol > 9) 2812 { 2813 if(offsetbits[offssymbol] > 4) { 2814 if (!rar_br_read_ahead(a, br, offsetbits[offssymbol] - 4)) 2815 goto truncated_data; 2816 offs += rar_br_bits(br, offsetbits[offssymbol] - 4) << 4; 2817 rar_br_consume(br, offsetbits[offssymbol] - 4); 2818 } 2819 2820 if(rar->numlowoffsetrepeats > 0) 2821 { 2822 rar->numlowoffsetrepeats--; 2823 offs += rar->lastlowoffset; 2824 } 2825 else 2826 { 2827 if ((lowoffsetsymbol = 2828 read_next_symbol(a, &rar->lowoffsetcode)) < 0) 2829 return (ARCHIVE_FATAL); 2830 if(lowoffsetsymbol == 16) 2831 { 2832 rar->numlowoffsetrepeats = 15; 2833 offs += rar->lastlowoffset; 2834 } 2835 else 2836 { 2837 offs += lowoffsetsymbol; 2838 rar->lastlowoffset = lowoffsetsymbol; 2839 } 2840 } 2841 } 2842 else { 2843 if (!rar_br_read_ahead(a, br, offsetbits[offssymbol])) 2844 goto truncated_data; 2845 offs += rar_br_bits(br, offsetbits[offssymbol]); 2846 rar_br_consume(br, offsetbits[offssymbol]); 2847 } 2848 } 2849 2850 if (offs >= 0x40000) 2851 len++; 2852 if (offs >= 0x2000) 2853 len++; 2854 2855 for(i = 3; i > 0; i--) 2856 rar->oldoffset[i] = rar->oldoffset[i-1]; 2857 rar->oldoffset[0] = offs; 2858 } 2859 2860 rar->lastoffset = offs; 2861 rar->lastlength = len; 2862 rar->output_last_match = 1; 2863 } 2864truncated_data: 2865 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2866 "Truncated RAR file data"); 2867 rar->valid = 0; 2868 return (ARCHIVE_FATAL); 2869bad_data: 2870 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2871 "Bad RAR file data"); 2872 return (ARCHIVE_FATAL); 2873} 2874 2875static int 2876copy_from_lzss_window(struct archive_read *a, const void **buffer, 2877 int64_t startpos, int length) 2878{ 2879 int windowoffs, firstpart; 2880 struct rar *rar = (struct rar *)(a->format->data); 2881 2882 if (!rar->unp_buffer) 2883 { 2884 if ((rar->unp_buffer = malloc(rar->unp_buffer_size)) == NULL) 2885 { 2886 archive_set_error(&a->archive, ENOMEM, 2887 "Unable to allocate memory for uncompressed data."); 2888 return (ARCHIVE_FATAL); 2889 } 2890 } 2891 2892 windowoffs = lzss_offset_for_position(&rar->lzss, startpos); 2893 if(windowoffs + length <= lzss_size(&rar->lzss)) { 2894 memcpy(&rar->unp_buffer[rar->unp_offset], &rar->lzss.window[windowoffs], 2895 length); 2896 } else if (length <= lzss_size(&rar->lzss)) { 2897 firstpart = lzss_size(&rar->lzss) - windowoffs; 2898 if (firstpart < 0) { 2899 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2900 "Bad RAR file data"); 2901 return (ARCHIVE_FATAL); 2902 } 2903 if (firstpart < length) { 2904 memcpy(&rar->unp_buffer[rar->unp_offset], 2905 &rar->lzss.window[windowoffs], firstpart); 2906 memcpy(&rar->unp_buffer[rar->unp_offset + firstpart], 2907 &rar->lzss.window[0], length - firstpart); 2908 } else { 2909 memcpy(&rar->unp_buffer[rar->unp_offset], 2910 &rar->lzss.window[windowoffs], length); 2911 } 2912 } else { 2913 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2914 "Bad RAR file data"); 2915 return (ARCHIVE_FATAL); 2916 } 2917 rar->unp_offset += length; 2918 if (rar->unp_offset >= rar->unp_buffer_size) 2919 *buffer = rar->unp_buffer; 2920 else 2921 *buffer = NULL; 2922 return (ARCHIVE_OK); 2923} 2924 2925static const void * 2926rar_read_ahead(struct archive_read *a, size_t min, ssize_t *avail) 2927{ 2928 struct rar *rar = (struct rar *)(a->format->data); 2929 const void *h = __archive_read_ahead(a, min, avail); 2930 int ret; 2931 if (avail) 2932 { 2933 if (a->archive.read_data_is_posix_read && *avail > (ssize_t)a->archive.read_data_requested) 2934 *avail = a->archive.read_data_requested; 2935 if (*avail > rar->bytes_remaining) 2936 *avail = (ssize_t)rar->bytes_remaining; 2937 if (*avail < 0) 2938 return NULL; 2939 else if (*avail == 0 && rar->main_flags & MHD_VOLUME && 2940 rar->file_flags & FHD_SPLIT_AFTER) 2941 { 2942 ret = archive_read_format_rar_read_header(a, a->entry); 2943 if (ret == (ARCHIVE_EOF)) 2944 { 2945 rar->has_endarc_header = 1; 2946 ret = archive_read_format_rar_read_header(a, a->entry); 2947 } 2948 if (ret != (ARCHIVE_OK)) 2949 return NULL; 2950 return rar_read_ahead(a, min, avail); 2951 } 2952 } 2953 return h; 2954} 2955