archive_read_support_format_rar.c revision 311041
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 *)calloc(sizeof(*rar), 1); 651 if (rar == NULL) 652 { 653 archive_set_error(&a->archive, ENOMEM, "Can't allocate rar data"); 654 return (ARCHIVE_FATAL); 655 } 656 657 /* 658 * Until enough data has been read, we cannot tell about 659 * any encrypted entries yet. 660 */ 661 rar->has_encrypted_entries = ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; 662 663 r = __archive_read_register_format(a, 664 rar, 665 "rar", 666 archive_read_format_rar_bid, 667 archive_read_format_rar_options, 668 archive_read_format_rar_read_header, 669 archive_read_format_rar_read_data, 670 archive_read_format_rar_read_data_skip, 671 archive_read_format_rar_seek_data, 672 archive_read_format_rar_cleanup, 673 archive_read_support_format_rar_capabilities, 674 archive_read_format_rar_has_encrypted_entries); 675 676 if (r != ARCHIVE_OK) 677 free(rar); 678 return (r); 679} 680 681static int 682archive_read_support_format_rar_capabilities(struct archive_read * a) 683{ 684 (void)a; /* UNUSED */ 685 return (ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_DATA 686 | ARCHIVE_READ_FORMAT_CAPS_ENCRYPT_METADATA); 687} 688 689static int 690archive_read_format_rar_has_encrypted_entries(struct archive_read *_a) 691{ 692 if (_a && _a->format) { 693 struct rar * rar = (struct rar *)_a->format->data; 694 if (rar) { 695 return rar->has_encrypted_entries; 696 } 697 } 698 return ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW; 699} 700 701 702static int 703archive_read_format_rar_bid(struct archive_read *a, int best_bid) 704{ 705 const char *p; 706 707 /* If there's already a bid > 30, we'll never win. */ 708 if (best_bid > 30) 709 return (-1); 710 711 if ((p = __archive_read_ahead(a, 7, NULL)) == NULL) 712 return (-1); 713 714 if (memcmp(p, RAR_SIGNATURE, 7) == 0) 715 return (30); 716 717 if ((p[0] == 'M' && p[1] == 'Z') || memcmp(p, "\x7F\x45LF", 4) == 0) { 718 /* This is a PE file */ 719 ssize_t offset = 0x10000; 720 ssize_t window = 4096; 721 ssize_t bytes_avail; 722 while (offset + window <= (1024 * 128)) { 723 const char *buff = __archive_read_ahead(a, offset + window, &bytes_avail); 724 if (buff == NULL) { 725 /* Remaining bytes are less than window. */ 726 window >>= 1; 727 if (window < 0x40) 728 return (0); 729 continue; 730 } 731 p = buff + offset; 732 while (p + 7 < buff + bytes_avail) { 733 if (memcmp(p, RAR_SIGNATURE, 7) == 0) 734 return (30); 735 p += 0x10; 736 } 737 offset = p - buff; 738 } 739 } 740 return (0); 741} 742 743static int 744skip_sfx(struct archive_read *a) 745{ 746 const void *h; 747 const char *p, *q; 748 size_t skip, total; 749 ssize_t bytes, window; 750 751 total = 0; 752 window = 4096; 753 while (total + window <= (1024 * 128)) { 754 h = __archive_read_ahead(a, window, &bytes); 755 if (h == NULL) { 756 /* Remaining bytes are less than window. */ 757 window >>= 1; 758 if (window < 0x40) 759 goto fatal; 760 continue; 761 } 762 if (bytes < 0x40) 763 goto fatal; 764 p = h; 765 q = p + bytes; 766 767 /* 768 * Scan ahead until we find something that looks 769 * like the RAR header. 770 */ 771 while (p + 7 < q) { 772 if (memcmp(p, RAR_SIGNATURE, 7) == 0) { 773 skip = p - (const char *)h; 774 __archive_read_consume(a, skip); 775 return (ARCHIVE_OK); 776 } 777 p += 0x10; 778 } 779 skip = p - (const char *)h; 780 __archive_read_consume(a, skip); 781 total += skip; 782 } 783fatal: 784 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 785 "Couldn't find out RAR header"); 786 return (ARCHIVE_FATAL); 787} 788 789static int 790archive_read_format_rar_options(struct archive_read *a, 791 const char *key, const char *val) 792{ 793 struct rar *rar; 794 int ret = ARCHIVE_FAILED; 795 796 rar = (struct rar *)(a->format->data); 797 if (strcmp(key, "hdrcharset") == 0) { 798 if (val == NULL || val[0] == 0) 799 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 800 "rar: hdrcharset option needs a character-set name"); 801 else { 802 rar->opt_sconv = 803 archive_string_conversion_from_charset( 804 &a->archive, val, 0); 805 if (rar->opt_sconv != NULL) 806 ret = ARCHIVE_OK; 807 else 808 ret = ARCHIVE_FATAL; 809 } 810 return (ret); 811 } 812 813 /* Note: The "warn" return is just to inform the options 814 * supervisor that we didn't handle it. It will generate 815 * a suitable error if no one used this option. */ 816 return (ARCHIVE_WARN); 817} 818 819static int 820archive_read_format_rar_read_header(struct archive_read *a, 821 struct archive_entry *entry) 822{ 823 const void *h; 824 const char *p; 825 struct rar *rar; 826 size_t skip; 827 char head_type; 828 int ret; 829 unsigned flags; 830 unsigned long crc32_expected; 831 832 a->archive.archive_format = ARCHIVE_FORMAT_RAR; 833 if (a->archive.archive_format_name == NULL) 834 a->archive.archive_format_name = "RAR"; 835 836 rar = (struct rar *)(a->format->data); 837 838 /* 839 * It should be sufficient to call archive_read_next_header() for 840 * a reader to determine if an entry is encrypted or not. If the 841 * encryption of an entry is only detectable when calling 842 * archive_read_data(), so be it. We'll do the same check there 843 * as well. 844 */ 845 if (rar->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { 846 rar->has_encrypted_entries = 0; 847 } 848 849 /* RAR files can be generated without EOF headers, so return ARCHIVE_EOF if 850 * this fails. 851 */ 852 if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) 853 return (ARCHIVE_EOF); 854 855 p = h; 856 if (rar->found_first_header == 0 && 857 ((p[0] == 'M' && p[1] == 'Z') || memcmp(p, "\x7F\x45LF", 4) == 0)) { 858 /* This is an executable ? Must be self-extracting... */ 859 ret = skip_sfx(a); 860 if (ret < ARCHIVE_WARN) 861 return (ret); 862 } 863 rar->found_first_header = 1; 864 865 while (1) 866 { 867 unsigned long crc32_val; 868 869 if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) 870 return (ARCHIVE_FATAL); 871 p = h; 872 873 head_type = p[2]; 874 switch(head_type) 875 { 876 case MARK_HEAD: 877 if (memcmp(p, RAR_SIGNATURE, 7) != 0) { 878 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 879 "Invalid marker header"); 880 return (ARCHIVE_FATAL); 881 } 882 __archive_read_consume(a, 7); 883 break; 884 885 case MAIN_HEAD: 886 rar->main_flags = archive_le16dec(p + 3); 887 skip = archive_le16dec(p + 5); 888 if (skip < 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)) { 889 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 890 "Invalid header size"); 891 return (ARCHIVE_FATAL); 892 } 893 if ((h = __archive_read_ahead(a, skip, NULL)) == NULL) 894 return (ARCHIVE_FATAL); 895 p = h; 896 memcpy(rar->reserved1, p + 7, sizeof(rar->reserved1)); 897 memcpy(rar->reserved2, p + 7 + sizeof(rar->reserved1), 898 sizeof(rar->reserved2)); 899 if (rar->main_flags & MHD_ENCRYPTVER) { 900 if (skip < 7 + sizeof(rar->reserved1) + sizeof(rar->reserved2)+1) { 901 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 902 "Invalid header size"); 903 return (ARCHIVE_FATAL); 904 } 905 rar->encryptver = *(p + 7 + sizeof(rar->reserved1) + 906 sizeof(rar->reserved2)); 907 } 908 909 /* Main header is password encrytped, so we cannot read any 910 file names or any other info about files from the header. */ 911 if (rar->main_flags & MHD_PASSWORD) 912 { 913 archive_entry_set_is_metadata_encrypted(entry, 1); 914 archive_entry_set_is_data_encrypted(entry, 1); 915 rar->has_encrypted_entries = 1; 916 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 917 "RAR encryption support unavailable."); 918 return (ARCHIVE_FATAL); 919 } 920 921 crc32_val = crc32(0, (const unsigned char *)p + 2, (unsigned)skip - 2); 922 if ((crc32_val & 0xffff) != archive_le16dec(p)) { 923 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 924 "Header CRC error"); 925 return (ARCHIVE_FATAL); 926 } 927 __archive_read_consume(a, skip); 928 break; 929 930 case FILE_HEAD: 931 return read_header(a, entry, head_type); 932 933 case COMM_HEAD: 934 case AV_HEAD: 935 case SUB_HEAD: 936 case PROTECT_HEAD: 937 case SIGN_HEAD: 938 case ENDARC_HEAD: 939 flags = archive_le16dec(p + 3); 940 skip = archive_le16dec(p + 5); 941 if (skip < 7) { 942 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 943 "Invalid header size too small"); 944 return (ARCHIVE_FATAL); 945 } 946 if (flags & HD_ADD_SIZE_PRESENT) 947 { 948 if (skip < 7 + 4) { 949 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 950 "Invalid header size too small"); 951 return (ARCHIVE_FATAL); 952 } 953 if ((h = __archive_read_ahead(a, skip, NULL)) == NULL) 954 return (ARCHIVE_FATAL); 955 p = h; 956 skip += archive_le32dec(p + 7); 957 } 958 959 /* Skip over the 2-byte CRC at the beginning of the header. */ 960 crc32_expected = archive_le16dec(p); 961 __archive_read_consume(a, 2); 962 skip -= 2; 963 964 /* Skim the entire header and compute the CRC. */ 965 crc32_val = 0; 966 while (skip > 0) { 967 size_t to_read = skip; 968 ssize_t did_read; 969 if (to_read > 32 * 1024) { 970 to_read = 32 * 1024; 971 } 972 if ((h = __archive_read_ahead(a, to_read, &did_read)) == NULL) { 973 return (ARCHIVE_FATAL); 974 } 975 p = h; 976 crc32_val = crc32(crc32_val, (const unsigned char *)p, (unsigned)did_read); 977 __archive_read_consume(a, did_read); 978 skip -= did_read; 979 } 980 if ((crc32_val & 0xffff) != crc32_expected) { 981 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 982 "Header CRC error"); 983 return (ARCHIVE_FATAL); 984 } 985 if (head_type == ENDARC_HEAD) 986 return (ARCHIVE_EOF); 987 break; 988 989 case NEWSUB_HEAD: 990 if ((ret = read_header(a, entry, head_type)) < ARCHIVE_WARN) 991 return ret; 992 break; 993 994 default: 995 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 996 "Bad RAR file"); 997 return (ARCHIVE_FATAL); 998 } 999 } 1000} 1001 1002static int 1003archive_read_format_rar_read_data(struct archive_read *a, const void **buff, 1004 size_t *size, int64_t *offset) 1005{ 1006 struct rar *rar = (struct rar *)(a->format->data); 1007 int ret; 1008 1009 if (rar->has_encrypted_entries == ARCHIVE_READ_FORMAT_ENCRYPTION_DONT_KNOW) { 1010 rar->has_encrypted_entries = 0; 1011 } 1012 1013 if (rar->bytes_unconsumed > 0) { 1014 /* Consume as much as the decompressor actually used. */ 1015 __archive_read_consume(a, rar->bytes_unconsumed); 1016 rar->bytes_unconsumed = 0; 1017 } 1018 1019 *buff = NULL; 1020 if (rar->entry_eof || rar->offset_seek >= rar->unp_size) { 1021 *size = 0; 1022 *offset = rar->offset; 1023 if (*offset < rar->unp_size) 1024 *offset = rar->unp_size; 1025 return (ARCHIVE_EOF); 1026 } 1027 1028 switch (rar->compression_method) 1029 { 1030 case COMPRESS_METHOD_STORE: 1031 ret = read_data_stored(a, buff, size, offset); 1032 break; 1033 1034 case COMPRESS_METHOD_FASTEST: 1035 case COMPRESS_METHOD_FAST: 1036 case COMPRESS_METHOD_NORMAL: 1037 case COMPRESS_METHOD_GOOD: 1038 case COMPRESS_METHOD_BEST: 1039 ret = read_data_compressed(a, buff, size, offset); 1040 if (ret != ARCHIVE_OK && ret != ARCHIVE_WARN) 1041 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 1042 break; 1043 1044 default: 1045 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1046 "Unsupported compression method for RAR file."); 1047 ret = ARCHIVE_FATAL; 1048 break; 1049 } 1050 return (ret); 1051} 1052 1053static int 1054archive_read_format_rar_read_data_skip(struct archive_read *a) 1055{ 1056 struct rar *rar; 1057 int64_t bytes_skipped; 1058 int ret; 1059 1060 rar = (struct rar *)(a->format->data); 1061 1062 if (rar->bytes_unconsumed > 0) { 1063 /* Consume as much as the decompressor actually used. */ 1064 __archive_read_consume(a, rar->bytes_unconsumed); 1065 rar->bytes_unconsumed = 0; 1066 } 1067 1068 if (rar->bytes_remaining > 0) { 1069 bytes_skipped = __archive_read_consume(a, rar->bytes_remaining); 1070 if (bytes_skipped < 0) 1071 return (ARCHIVE_FATAL); 1072 } 1073 1074 /* Compressed data to skip must be read from each header in a multivolume 1075 * archive. 1076 */ 1077 if (rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER) 1078 { 1079 ret = archive_read_format_rar_read_header(a, a->entry); 1080 if (ret == (ARCHIVE_EOF)) 1081 ret = archive_read_format_rar_read_header(a, a->entry); 1082 if (ret != (ARCHIVE_OK)) 1083 return ret; 1084 return archive_read_format_rar_read_data_skip(a); 1085 } 1086 1087 return (ARCHIVE_OK); 1088} 1089 1090static int64_t 1091archive_read_format_rar_seek_data(struct archive_read *a, int64_t offset, 1092 int whence) 1093{ 1094 int64_t client_offset, ret; 1095 unsigned int i; 1096 struct rar *rar = (struct rar *)(a->format->data); 1097 1098 if (rar->compression_method == COMPRESS_METHOD_STORE) 1099 { 1100 /* Modify the offset for use with SEEK_SET */ 1101 switch (whence) 1102 { 1103 case SEEK_CUR: 1104 client_offset = rar->offset_seek; 1105 break; 1106 case SEEK_END: 1107 client_offset = rar->unp_size; 1108 break; 1109 case SEEK_SET: 1110 default: 1111 client_offset = 0; 1112 } 1113 client_offset += offset; 1114 if (client_offset < 0) 1115 { 1116 /* Can't seek past beginning of data block */ 1117 return -1; 1118 } 1119 else if (client_offset > rar->unp_size) 1120 { 1121 /* 1122 * Set the returned offset but only seek to the end of 1123 * the data block. 1124 */ 1125 rar->offset_seek = client_offset; 1126 client_offset = rar->unp_size; 1127 } 1128 1129 client_offset += rar->dbo[0].start_offset; 1130 i = 0; 1131 while (i < rar->cursor) 1132 { 1133 i++; 1134 client_offset += rar->dbo[i].start_offset - rar->dbo[i-1].end_offset; 1135 } 1136 if (rar->main_flags & MHD_VOLUME) 1137 { 1138 /* Find the appropriate offset among the multivolume archive */ 1139 while (1) 1140 { 1141 if (client_offset < rar->dbo[rar->cursor].start_offset && 1142 rar->file_flags & FHD_SPLIT_BEFORE) 1143 { 1144 /* Search backwards for the correct data block */ 1145 if (rar->cursor == 0) 1146 { 1147 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1148 "Attempt to seek past beginning of RAR data block"); 1149 return (ARCHIVE_FAILED); 1150 } 1151 rar->cursor--; 1152 client_offset -= rar->dbo[rar->cursor+1].start_offset - 1153 rar->dbo[rar->cursor].end_offset; 1154 if (client_offset < rar->dbo[rar->cursor].start_offset) 1155 continue; 1156 ret = __archive_read_seek(a, rar->dbo[rar->cursor].start_offset - 1157 rar->dbo[rar->cursor].header_size, SEEK_SET); 1158 if (ret < (ARCHIVE_OK)) 1159 return ret; 1160 ret = archive_read_format_rar_read_header(a, a->entry); 1161 if (ret != (ARCHIVE_OK)) 1162 { 1163 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1164 "Error during seek of RAR file"); 1165 return (ARCHIVE_FAILED); 1166 } 1167 rar->cursor--; 1168 break; 1169 } 1170 else if (client_offset > rar->dbo[rar->cursor].end_offset && 1171 rar->file_flags & FHD_SPLIT_AFTER) 1172 { 1173 /* Search forward for the correct data block */ 1174 rar->cursor++; 1175 if (rar->cursor < rar->nodes && 1176 client_offset > rar->dbo[rar->cursor].end_offset) 1177 { 1178 client_offset += rar->dbo[rar->cursor].start_offset - 1179 rar->dbo[rar->cursor-1].end_offset; 1180 continue; 1181 } 1182 rar->cursor--; 1183 ret = __archive_read_seek(a, rar->dbo[rar->cursor].end_offset, 1184 SEEK_SET); 1185 if (ret < (ARCHIVE_OK)) 1186 return ret; 1187 ret = archive_read_format_rar_read_header(a, a->entry); 1188 if (ret == (ARCHIVE_EOF)) 1189 { 1190 rar->has_endarc_header = 1; 1191 ret = archive_read_format_rar_read_header(a, a->entry); 1192 } 1193 if (ret != (ARCHIVE_OK)) 1194 { 1195 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1196 "Error during seek of RAR file"); 1197 return (ARCHIVE_FAILED); 1198 } 1199 client_offset += rar->dbo[rar->cursor].start_offset - 1200 rar->dbo[rar->cursor-1].end_offset; 1201 continue; 1202 } 1203 break; 1204 } 1205 } 1206 1207 ret = __archive_read_seek(a, client_offset, SEEK_SET); 1208 if (ret < (ARCHIVE_OK)) 1209 return ret; 1210 rar->bytes_remaining = rar->dbo[rar->cursor].end_offset - ret; 1211 i = rar->cursor; 1212 while (i > 0) 1213 { 1214 i--; 1215 ret -= rar->dbo[i+1].start_offset - rar->dbo[i].end_offset; 1216 } 1217 ret -= rar->dbo[0].start_offset; 1218 1219 /* Always restart reading the file after a seek */ 1220 __archive_reset_read_data(&a->archive); 1221 1222 rar->bytes_unconsumed = 0; 1223 rar->offset = 0; 1224 1225 /* 1226 * If a seek past the end of file was requested, return the requested 1227 * offset. 1228 */ 1229 if (ret == rar->unp_size && rar->offset_seek > rar->unp_size) 1230 return rar->offset_seek; 1231 1232 /* Return the new offset */ 1233 rar->offset_seek = ret; 1234 return rar->offset_seek; 1235 } 1236 else 1237 { 1238 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1239 "Seeking of compressed RAR files is unsupported"); 1240 } 1241 return (ARCHIVE_FAILED); 1242} 1243 1244static int 1245archive_read_format_rar_cleanup(struct archive_read *a) 1246{ 1247 struct rar *rar; 1248 1249 rar = (struct rar *)(a->format->data); 1250 free_codes(a); 1251 free(rar->filename); 1252 free(rar->filename_save); 1253 free(rar->dbo); 1254 free(rar->unp_buffer); 1255 free(rar->lzss.window); 1256 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 1257 free(rar); 1258 (a->format->data) = NULL; 1259 return (ARCHIVE_OK); 1260} 1261 1262static int 1263read_header(struct archive_read *a, struct archive_entry *entry, 1264 char head_type) 1265{ 1266 const void *h; 1267 const char *p, *endp; 1268 struct rar *rar; 1269 struct rar_header rar_header; 1270 struct rar_file_header file_header; 1271 int64_t header_size; 1272 unsigned filename_size, end; 1273 char *filename; 1274 char *strp; 1275 char packed_size[8]; 1276 char unp_size[8]; 1277 int ttime; 1278 struct archive_string_conv *sconv, *fn_sconv; 1279 unsigned long crc32_val; 1280 int ret = (ARCHIVE_OK), ret2; 1281 1282 rar = (struct rar *)(a->format->data); 1283 1284 /* Setup a string conversion object for non-rar-unicode filenames. */ 1285 sconv = rar->opt_sconv; 1286 if (sconv == NULL) { 1287 if (!rar->init_default_conversion) { 1288 rar->sconv_default = 1289 archive_string_default_conversion_for_read( 1290 &(a->archive)); 1291 rar->init_default_conversion = 1; 1292 } 1293 sconv = rar->sconv_default; 1294 } 1295 1296 1297 if ((h = __archive_read_ahead(a, 7, NULL)) == NULL) 1298 return (ARCHIVE_FATAL); 1299 p = h; 1300 memcpy(&rar_header, p, sizeof(rar_header)); 1301 rar->file_flags = archive_le16dec(rar_header.flags); 1302 header_size = archive_le16dec(rar_header.size); 1303 if (header_size < (int64_t)sizeof(file_header) + 7) { 1304 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1305 "Invalid header size"); 1306 return (ARCHIVE_FATAL); 1307 } 1308 crc32_val = crc32(0, (const unsigned char *)p + 2, 7 - 2); 1309 __archive_read_consume(a, 7); 1310 1311 if (!(rar->file_flags & FHD_SOLID)) 1312 { 1313 rar->compression_method = 0; 1314 rar->packed_size = 0; 1315 rar->unp_size = 0; 1316 rar->mtime = 0; 1317 rar->ctime = 0; 1318 rar->atime = 0; 1319 rar->arctime = 0; 1320 rar->mode = 0; 1321 memset(&rar->salt, 0, sizeof(rar->salt)); 1322 rar->atime = 0; 1323 rar->ansec = 0; 1324 rar->ctime = 0; 1325 rar->cnsec = 0; 1326 rar->mtime = 0; 1327 rar->mnsec = 0; 1328 rar->arctime = 0; 1329 rar->arcnsec = 0; 1330 } 1331 else 1332 { 1333 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1334 "RAR solid archive support unavailable."); 1335 return (ARCHIVE_FATAL); 1336 } 1337 1338 if ((h = __archive_read_ahead(a, (size_t)header_size - 7, NULL)) == NULL) 1339 return (ARCHIVE_FATAL); 1340 1341 /* File Header CRC check. */ 1342 crc32_val = crc32(crc32_val, h, (unsigned)(header_size - 7)); 1343 if ((crc32_val & 0xffff) != archive_le16dec(rar_header.crc)) { 1344 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1345 "Header CRC error"); 1346 return (ARCHIVE_FATAL); 1347 } 1348 /* If no CRC error, Go on parsing File Header. */ 1349 p = h; 1350 endp = p + header_size - 7; 1351 memcpy(&file_header, p, sizeof(file_header)); 1352 p += sizeof(file_header); 1353 1354 rar->compression_method = file_header.method; 1355 1356 ttime = archive_le32dec(file_header.file_time); 1357 rar->mtime = get_time(ttime); 1358 1359 rar->file_crc = archive_le32dec(file_header.file_crc); 1360 1361 if (rar->file_flags & FHD_PASSWORD) 1362 { 1363 archive_entry_set_is_data_encrypted(entry, 1); 1364 rar->has_encrypted_entries = 1; 1365 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1366 "RAR encryption support unavailable."); 1367 /* Since it is only the data part itself that is encrypted we can at least 1368 extract information about the currently processed entry and don't need 1369 to return ARCHIVE_FATAL here. */ 1370 /*return (ARCHIVE_FATAL);*/ 1371 } 1372 1373 if (rar->file_flags & FHD_LARGE) 1374 { 1375 memcpy(packed_size, file_header.pack_size, 4); 1376 memcpy(packed_size + 4, p, 4); /* High pack size */ 1377 p += 4; 1378 memcpy(unp_size, file_header.unp_size, 4); 1379 memcpy(unp_size + 4, p, 4); /* High unpack size */ 1380 p += 4; 1381 rar->packed_size = archive_le64dec(&packed_size); 1382 rar->unp_size = archive_le64dec(&unp_size); 1383 } 1384 else 1385 { 1386 rar->packed_size = archive_le32dec(file_header.pack_size); 1387 rar->unp_size = archive_le32dec(file_header.unp_size); 1388 } 1389 1390 if (rar->packed_size < 0 || rar->unp_size < 0) 1391 { 1392 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1393 "Invalid sizes specified."); 1394 return (ARCHIVE_FATAL); 1395 } 1396 1397 rar->bytes_remaining = rar->packed_size; 1398 1399 /* TODO: RARv3 subblocks contain comments. For now the complete block is 1400 * consumed at the end. 1401 */ 1402 if (head_type == NEWSUB_HEAD) { 1403 size_t distance = p - (const char *)h; 1404 header_size += rar->packed_size; 1405 /* Make sure we have the extended data. */ 1406 if ((h = __archive_read_ahead(a, (size_t)header_size - 7, NULL)) == NULL) 1407 return (ARCHIVE_FATAL); 1408 p = h; 1409 endp = p + header_size - 7; 1410 p += distance; 1411 } 1412 1413 filename_size = archive_le16dec(file_header.name_size); 1414 if (p + filename_size > endp) { 1415 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1416 "Invalid filename size"); 1417 return (ARCHIVE_FATAL); 1418 } 1419 if (rar->filename_allocated < filename_size * 2 + 2) { 1420 char *newptr; 1421 size_t newsize = filename_size * 2 + 2; 1422 newptr = realloc(rar->filename, newsize); 1423 if (newptr == NULL) { 1424 archive_set_error(&a->archive, ENOMEM, 1425 "Couldn't allocate memory."); 1426 return (ARCHIVE_FATAL); 1427 } 1428 rar->filename = newptr; 1429 rar->filename_allocated = newsize; 1430 } 1431 filename = rar->filename; 1432 memcpy(filename, p, filename_size); 1433 filename[filename_size] = '\0'; 1434 if (rar->file_flags & FHD_UNICODE) 1435 { 1436 if (filename_size != strlen(filename)) 1437 { 1438 unsigned char highbyte, flagbits, flagbyte; 1439 unsigned fn_end, offset; 1440 1441 end = filename_size; 1442 fn_end = filename_size * 2; 1443 filename_size = 0; 1444 offset = (unsigned)strlen(filename) + 1; 1445 highbyte = *(p + offset++); 1446 flagbits = 0; 1447 flagbyte = 0; 1448 while (offset < end && filename_size < fn_end) 1449 { 1450 if (!flagbits) 1451 { 1452 flagbyte = *(p + offset++); 1453 flagbits = 8; 1454 } 1455 1456 flagbits -= 2; 1457 switch((flagbyte >> flagbits) & 3) 1458 { 1459 case 0: 1460 filename[filename_size++] = '\0'; 1461 filename[filename_size++] = *(p + offset++); 1462 break; 1463 case 1: 1464 filename[filename_size++] = highbyte; 1465 filename[filename_size++] = *(p + offset++); 1466 break; 1467 case 2: 1468 filename[filename_size++] = *(p + offset + 1); 1469 filename[filename_size++] = *(p + offset); 1470 offset += 2; 1471 break; 1472 case 3: 1473 { 1474 char extra, high; 1475 uint8_t length = *(p + offset++); 1476 1477 if (length & 0x80) { 1478 extra = *(p + offset++); 1479 high = (char)highbyte; 1480 } else 1481 extra = high = 0; 1482 length = (length & 0x7f) + 2; 1483 while (length && filename_size < fn_end) { 1484 unsigned cp = filename_size >> 1; 1485 filename[filename_size++] = high; 1486 filename[filename_size++] = p[cp] + extra; 1487 length--; 1488 } 1489 } 1490 break; 1491 } 1492 } 1493 if (filename_size > fn_end) { 1494 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1495 "Invalid filename"); 1496 return (ARCHIVE_FATAL); 1497 } 1498 filename[filename_size++] = '\0'; 1499 filename[filename_size++] = '\0'; 1500 1501 /* Decoded unicode form is UTF-16BE, so we have to update a string 1502 * conversion object for it. */ 1503 if (rar->sconv_utf16be == NULL) { 1504 rar->sconv_utf16be = archive_string_conversion_from_charset( 1505 &a->archive, "UTF-16BE", 1); 1506 if (rar->sconv_utf16be == NULL) 1507 return (ARCHIVE_FATAL); 1508 } 1509 fn_sconv = rar->sconv_utf16be; 1510 1511 strp = filename; 1512 while (memcmp(strp, "\x00\x00", 2)) 1513 { 1514 if (!memcmp(strp, "\x00\\", 2)) 1515 *(strp + 1) = '/'; 1516 strp += 2; 1517 } 1518 p += offset; 1519 } else { 1520 /* 1521 * If FHD_UNICODE is set but no unicode data, this file name form 1522 * is UTF-8, so we have to update a string conversion object for 1523 * it accordingly. 1524 */ 1525 if (rar->sconv_utf8 == NULL) { 1526 rar->sconv_utf8 = archive_string_conversion_from_charset( 1527 &a->archive, "UTF-8", 1); 1528 if (rar->sconv_utf8 == NULL) 1529 return (ARCHIVE_FATAL); 1530 } 1531 fn_sconv = rar->sconv_utf8; 1532 while ((strp = strchr(filename, '\\')) != NULL) 1533 *strp = '/'; 1534 p += filename_size; 1535 } 1536 } 1537 else 1538 { 1539 fn_sconv = sconv; 1540 while ((strp = strchr(filename, '\\')) != NULL) 1541 *strp = '/'; 1542 p += filename_size; 1543 } 1544 1545 /* Split file in multivolume RAR. No more need to process header. */ 1546 if (rar->filename_save && 1547 filename_size == rar->filename_save_size && 1548 !memcmp(rar->filename, rar->filename_save, filename_size + 1)) 1549 { 1550 __archive_read_consume(a, header_size - 7); 1551 rar->cursor++; 1552 if (rar->cursor >= rar->nodes) 1553 { 1554 rar->nodes++; 1555 if ((rar->dbo = 1556 realloc(rar->dbo, sizeof(*rar->dbo) * rar->nodes)) == NULL) 1557 { 1558 archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); 1559 return (ARCHIVE_FATAL); 1560 } 1561 rar->dbo[rar->cursor].header_size = header_size; 1562 rar->dbo[rar->cursor].start_offset = -1; 1563 rar->dbo[rar->cursor].end_offset = -1; 1564 } 1565 if (rar->dbo[rar->cursor].start_offset < 0) 1566 { 1567 rar->dbo[rar->cursor].start_offset = a->filter->position; 1568 rar->dbo[rar->cursor].end_offset = rar->dbo[rar->cursor].start_offset + 1569 rar->packed_size; 1570 } 1571 return ret; 1572 } 1573 1574 rar->filename_save = (char*)realloc(rar->filename_save, 1575 filename_size + 1); 1576 memcpy(rar->filename_save, rar->filename, filename_size + 1); 1577 rar->filename_save_size = filename_size; 1578 1579 /* Set info for seeking */ 1580 free(rar->dbo); 1581 if ((rar->dbo = calloc(1, sizeof(*rar->dbo))) == NULL) 1582 { 1583 archive_set_error(&a->archive, ENOMEM, "Couldn't allocate memory."); 1584 return (ARCHIVE_FATAL); 1585 } 1586 rar->dbo[0].header_size = header_size; 1587 rar->dbo[0].start_offset = -1; 1588 rar->dbo[0].end_offset = -1; 1589 rar->cursor = 0; 1590 rar->nodes = 1; 1591 1592 if (rar->file_flags & FHD_SALT) 1593 { 1594 if (p + 8 > endp) { 1595 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1596 "Invalid header size"); 1597 return (ARCHIVE_FATAL); 1598 } 1599 memcpy(rar->salt, p, 8); 1600 p += 8; 1601 } 1602 1603 if (rar->file_flags & FHD_EXTTIME) { 1604 if (read_exttime(p, rar, endp) < 0) { 1605 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1606 "Invalid header size"); 1607 return (ARCHIVE_FATAL); 1608 } 1609 } 1610 1611 __archive_read_consume(a, header_size - 7); 1612 rar->dbo[0].start_offset = a->filter->position; 1613 rar->dbo[0].end_offset = rar->dbo[0].start_offset + rar->packed_size; 1614 1615 switch(file_header.host_os) 1616 { 1617 case OS_MSDOS: 1618 case OS_OS2: 1619 case OS_WIN32: 1620 rar->mode = archive_le32dec(file_header.file_attr); 1621 if (rar->mode & FILE_ATTRIBUTE_DIRECTORY) 1622 rar->mode = AE_IFDIR | S_IXUSR | S_IXGRP | S_IXOTH; 1623 else 1624 rar->mode = AE_IFREG; 1625 rar->mode |= S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH; 1626 break; 1627 1628 case OS_UNIX: 1629 case OS_MAC_OS: 1630 case OS_BEOS: 1631 rar->mode = archive_le32dec(file_header.file_attr); 1632 break; 1633 1634 default: 1635 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1636 "Unknown file attributes from RAR file's host OS"); 1637 return (ARCHIVE_FATAL); 1638 } 1639 1640 rar->bytes_uncopied = rar->bytes_unconsumed = 0; 1641 rar->lzss.position = rar->offset = 0; 1642 rar->offset_seek = 0; 1643 rar->dictionary_size = 0; 1644 rar->offset_outgoing = 0; 1645 rar->br.cache_avail = 0; 1646 rar->br.avail_in = 0; 1647 rar->crc_calculated = 0; 1648 rar->entry_eof = 0; 1649 rar->valid = 1; 1650 rar->is_ppmd_block = 0; 1651 rar->start_new_table = 1; 1652 free(rar->unp_buffer); 1653 rar->unp_buffer = NULL; 1654 rar->unp_offset = 0; 1655 rar->unp_buffer_size = UNP_BUFFER_SIZE; 1656 memset(rar->lengthtable, 0, sizeof(rar->lengthtable)); 1657 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 1658 rar->ppmd_valid = rar->ppmd_eod = 0; 1659 1660 /* Don't set any archive entries for non-file header types */ 1661 if (head_type == NEWSUB_HEAD) 1662 return ret; 1663 1664 archive_entry_set_mtime(entry, rar->mtime, rar->mnsec); 1665 archive_entry_set_ctime(entry, rar->ctime, rar->cnsec); 1666 archive_entry_set_atime(entry, rar->atime, rar->ansec); 1667 archive_entry_set_size(entry, rar->unp_size); 1668 archive_entry_set_mode(entry, rar->mode); 1669 1670 if (archive_entry_copy_pathname_l(entry, filename, filename_size, fn_sconv)) 1671 { 1672 if (errno == ENOMEM) 1673 { 1674 archive_set_error(&a->archive, ENOMEM, 1675 "Can't allocate memory for Pathname"); 1676 return (ARCHIVE_FATAL); 1677 } 1678 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1679 "Pathname cannot be converted from %s to current locale.", 1680 archive_string_conversion_charset_name(fn_sconv)); 1681 ret = (ARCHIVE_WARN); 1682 } 1683 1684 if (((rar->mode) & AE_IFMT) == AE_IFLNK) 1685 { 1686 /* Make sure a symbolic-link file does not have its body. */ 1687 rar->bytes_remaining = 0; 1688 archive_entry_set_size(entry, 0); 1689 1690 /* Read a symbolic-link name. */ 1691 if ((ret2 = read_symlink_stored(a, entry, sconv)) < (ARCHIVE_WARN)) 1692 return ret2; 1693 if (ret > ret2) 1694 ret = ret2; 1695 } 1696 1697 if (rar->bytes_remaining == 0) 1698 rar->entry_eof = 1; 1699 1700 return ret; 1701} 1702 1703static time_t 1704get_time(int ttime) 1705{ 1706 struct tm tm; 1707 tm.tm_sec = 2 * (ttime & 0x1f); 1708 tm.tm_min = (ttime >> 5) & 0x3f; 1709 tm.tm_hour = (ttime >> 11) & 0x1f; 1710 tm.tm_mday = (ttime >> 16) & 0x1f; 1711 tm.tm_mon = ((ttime >> 21) & 0x0f) - 1; 1712 tm.tm_year = ((ttime >> 25) & 0x7f) + 80; 1713 tm.tm_isdst = -1; 1714 return mktime(&tm); 1715} 1716 1717static int 1718read_exttime(const char *p, struct rar *rar, const char *endp) 1719{ 1720 unsigned rmode, flags, rem, j, count; 1721 int ttime, i; 1722 struct tm *tm; 1723 time_t t; 1724 long nsec; 1725 1726 if (p + 2 > endp) 1727 return (-1); 1728 flags = archive_le16dec(p); 1729 p += 2; 1730 1731 for (i = 3; i >= 0; i--) 1732 { 1733 t = 0; 1734 if (i == 3) 1735 t = rar->mtime; 1736 rmode = flags >> i * 4; 1737 if (rmode & 8) 1738 { 1739 if (!t) 1740 { 1741 if (p + 4 > endp) 1742 return (-1); 1743 ttime = archive_le32dec(p); 1744 t = get_time(ttime); 1745 p += 4; 1746 } 1747 rem = 0; 1748 count = rmode & 3; 1749 if (p + count > endp) 1750 return (-1); 1751 for (j = 0; j < count; j++) 1752 { 1753 rem = ((*p) << 16) | (rem >> 8); 1754 p++; 1755 } 1756 tm = localtime(&t); 1757 nsec = tm->tm_sec + rem / NS_UNIT; 1758 if (rmode & 4) 1759 { 1760 tm->tm_sec++; 1761 t = mktime(tm); 1762 } 1763 if (i == 3) 1764 { 1765 rar->mtime = t; 1766 rar->mnsec = nsec; 1767 } 1768 else if (i == 2) 1769 { 1770 rar->ctime = t; 1771 rar->cnsec = nsec; 1772 } 1773 else if (i == 1) 1774 { 1775 rar->atime = t; 1776 rar->ansec = nsec; 1777 } 1778 else 1779 { 1780 rar->arctime = t; 1781 rar->arcnsec = nsec; 1782 } 1783 } 1784 } 1785 return (0); 1786} 1787 1788static int 1789read_symlink_stored(struct archive_read *a, struct archive_entry *entry, 1790 struct archive_string_conv *sconv) 1791{ 1792 const void *h; 1793 const char *p; 1794 struct rar *rar; 1795 int ret = (ARCHIVE_OK); 1796 1797 rar = (struct rar *)(a->format->data); 1798 if ((h = rar_read_ahead(a, (size_t)rar->packed_size, NULL)) == NULL) 1799 return (ARCHIVE_FATAL); 1800 p = h; 1801 1802 if (archive_entry_copy_symlink_l(entry, 1803 p, (size_t)rar->packed_size, sconv)) 1804 { 1805 if (errno == ENOMEM) 1806 { 1807 archive_set_error(&a->archive, ENOMEM, 1808 "Can't allocate memory for link"); 1809 return (ARCHIVE_FATAL); 1810 } 1811 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1812 "link cannot be converted from %s to current locale.", 1813 archive_string_conversion_charset_name(sconv)); 1814 ret = (ARCHIVE_WARN); 1815 } 1816 __archive_read_consume(a, rar->packed_size); 1817 return ret; 1818} 1819 1820static int 1821read_data_stored(struct archive_read *a, const void **buff, size_t *size, 1822 int64_t *offset) 1823{ 1824 struct rar *rar; 1825 ssize_t bytes_avail; 1826 1827 rar = (struct rar *)(a->format->data); 1828 if (rar->bytes_remaining == 0 && 1829 !(rar->main_flags & MHD_VOLUME && rar->file_flags & FHD_SPLIT_AFTER)) 1830 { 1831 *buff = NULL; 1832 *size = 0; 1833 *offset = rar->offset; 1834 if (rar->file_crc != rar->crc_calculated) { 1835 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1836 "File CRC error"); 1837 return (ARCHIVE_FATAL); 1838 } 1839 rar->entry_eof = 1; 1840 return (ARCHIVE_EOF); 1841 } 1842 1843 *buff = rar_read_ahead(a, 1, &bytes_avail); 1844 if (bytes_avail <= 0) 1845 { 1846 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1847 "Truncated RAR file data"); 1848 return (ARCHIVE_FATAL); 1849 } 1850 1851 *size = bytes_avail; 1852 *offset = rar->offset; 1853 rar->offset += bytes_avail; 1854 rar->offset_seek += bytes_avail; 1855 rar->bytes_remaining -= bytes_avail; 1856 rar->bytes_unconsumed = bytes_avail; 1857 /* Calculate File CRC. */ 1858 rar->crc_calculated = crc32(rar->crc_calculated, *buff, 1859 (unsigned)bytes_avail); 1860 return (ARCHIVE_OK); 1861} 1862 1863static int 1864read_data_compressed(struct archive_read *a, const void **buff, size_t *size, 1865 int64_t *offset) 1866{ 1867 struct rar *rar; 1868 int64_t start, end, actualend; 1869 size_t bs; 1870 int ret = (ARCHIVE_OK), sym, code, lzss_offset, length, i; 1871 1872 rar = (struct rar *)(a->format->data); 1873 1874 do { 1875 if (!rar->valid) 1876 return (ARCHIVE_FATAL); 1877 if (rar->ppmd_eod || 1878 (rar->dictionary_size && rar->offset >= rar->unp_size)) 1879 { 1880 if (rar->unp_offset > 0) { 1881 /* 1882 * We have unprocessed extracted data. write it out. 1883 */ 1884 *buff = rar->unp_buffer; 1885 *size = rar->unp_offset; 1886 *offset = rar->offset_outgoing; 1887 rar->offset_outgoing += *size; 1888 /* Calculate File CRC. */ 1889 rar->crc_calculated = crc32(rar->crc_calculated, *buff, 1890 (unsigned)*size); 1891 rar->unp_offset = 0; 1892 return (ARCHIVE_OK); 1893 } 1894 *buff = NULL; 1895 *size = 0; 1896 *offset = rar->offset; 1897 if (rar->file_crc != rar->crc_calculated) { 1898 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1899 "File CRC error"); 1900 return (ARCHIVE_FATAL); 1901 } 1902 rar->entry_eof = 1; 1903 return (ARCHIVE_EOF); 1904 } 1905 1906 if (!rar->is_ppmd_block && rar->dictionary_size && rar->bytes_uncopied > 0) 1907 { 1908 if (rar->bytes_uncopied > (rar->unp_buffer_size - rar->unp_offset)) 1909 bs = rar->unp_buffer_size - rar->unp_offset; 1910 else 1911 bs = (size_t)rar->bytes_uncopied; 1912 ret = copy_from_lzss_window(a, buff, rar->offset, (int)bs); 1913 if (ret != ARCHIVE_OK) 1914 return (ret); 1915 rar->offset += bs; 1916 rar->bytes_uncopied -= bs; 1917 if (*buff != NULL) { 1918 rar->unp_offset = 0; 1919 *size = rar->unp_buffer_size; 1920 *offset = rar->offset_outgoing; 1921 rar->offset_outgoing += *size; 1922 /* Calculate File CRC. */ 1923 rar->crc_calculated = crc32(rar->crc_calculated, *buff, 1924 (unsigned)*size); 1925 return (ret); 1926 } 1927 continue; 1928 } 1929 1930 if (!rar->br.next_in && 1931 (ret = rar_br_preparation(a, &(rar->br))) < ARCHIVE_WARN) 1932 return (ret); 1933 if (rar->start_new_table && ((ret = parse_codes(a)) < (ARCHIVE_WARN))) 1934 return (ret); 1935 1936 if (rar->is_ppmd_block) 1937 { 1938 if ((sym = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1939 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1940 { 1941 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1942 "Invalid symbol"); 1943 return (ARCHIVE_FATAL); 1944 } 1945 if(sym != rar->ppmd_escape) 1946 { 1947 lzss_emit_literal(rar, sym); 1948 rar->bytes_uncopied++; 1949 } 1950 else 1951 { 1952 if ((code = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1953 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1954 { 1955 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1956 "Invalid symbol"); 1957 return (ARCHIVE_FATAL); 1958 } 1959 1960 switch(code) 1961 { 1962 case 0: 1963 rar->start_new_table = 1; 1964 return read_data_compressed(a, buff, size, offset); 1965 1966 case 2: 1967 rar->ppmd_eod = 1;/* End Of ppmd Data. */ 1968 continue; 1969 1970 case 3: 1971 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1972 "Parsing filters is unsupported."); 1973 return (ARCHIVE_FAILED); 1974 1975 case 4: 1976 lzss_offset = 0; 1977 for (i = 2; i >= 0; i--) 1978 { 1979 if ((code = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1980 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1981 { 1982 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1983 "Invalid symbol"); 1984 return (ARCHIVE_FATAL); 1985 } 1986 lzss_offset |= code << (i * 8); 1987 } 1988 if ((length = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 1989 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 1990 { 1991 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1992 "Invalid symbol"); 1993 return (ARCHIVE_FATAL); 1994 } 1995 lzss_emit_match(rar, lzss_offset + 2, length + 32); 1996 rar->bytes_uncopied += length + 32; 1997 break; 1998 1999 case 5: 2000 if ((length = __archive_ppmd7_functions.Ppmd7_DecodeSymbol( 2001 &rar->ppmd7_context, &rar->range_dec.p)) < 0) 2002 { 2003 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2004 "Invalid symbol"); 2005 return (ARCHIVE_FATAL); 2006 } 2007 lzss_emit_match(rar, 1, length + 4); 2008 rar->bytes_uncopied += length + 4; 2009 break; 2010 2011 default: 2012 lzss_emit_literal(rar, sym); 2013 rar->bytes_uncopied++; 2014 } 2015 } 2016 } 2017 else 2018 { 2019 start = rar->offset; 2020 end = start + rar->dictionary_size; 2021 rar->filterstart = INT64_MAX; 2022 2023 if ((actualend = expand(a, end)) < 0) 2024 return ((int)actualend); 2025 2026 rar->bytes_uncopied = actualend - start; 2027 if (rar->bytes_uncopied == 0) { 2028 /* Broken RAR files cause this case. 2029 * NOTE: If this case were possible on a normal RAR file 2030 * we would find out where it was actually bad and 2031 * what we would do to solve it. */ 2032 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2033 "Internal error extracting RAR file"); 2034 return (ARCHIVE_FATAL); 2035 } 2036 } 2037 if (rar->bytes_uncopied > (rar->unp_buffer_size - rar->unp_offset)) 2038 bs = rar->unp_buffer_size - rar->unp_offset; 2039 else 2040 bs = (size_t)rar->bytes_uncopied; 2041 ret = copy_from_lzss_window(a, buff, rar->offset, (int)bs); 2042 if (ret != ARCHIVE_OK) 2043 return (ret); 2044 rar->offset += bs; 2045 rar->bytes_uncopied -= bs; 2046 /* 2047 * If *buff is NULL, it means unp_buffer is not full. 2048 * So we have to continue extracting a RAR file. 2049 */ 2050 } while (*buff == NULL); 2051 2052 rar->unp_offset = 0; 2053 *size = rar->unp_buffer_size; 2054 *offset = rar->offset_outgoing; 2055 rar->offset_outgoing += *size; 2056 /* Calculate File CRC. */ 2057 rar->crc_calculated = crc32(rar->crc_calculated, *buff, (unsigned)*size); 2058 return ret; 2059} 2060 2061static int 2062parse_codes(struct archive_read *a) 2063{ 2064 int i, j, val, n, r; 2065 unsigned char bitlengths[MAX_SYMBOLS], zerocount, ppmd_flags; 2066 unsigned int maxorder; 2067 struct huffman_code precode; 2068 struct rar *rar = (struct rar *)(a->format->data); 2069 struct rar_br *br = &(rar->br); 2070 2071 free_codes(a); 2072 2073 /* Skip to the next byte */ 2074 rar_br_consume_unalined_bits(br); 2075 2076 /* PPMd block flag */ 2077 if (!rar_br_read_ahead(a, br, 1)) 2078 goto truncated_data; 2079 if ((rar->is_ppmd_block = rar_br_bits(br, 1)) != 0) 2080 { 2081 rar_br_consume(br, 1); 2082 if (!rar_br_read_ahead(a, br, 7)) 2083 goto truncated_data; 2084 ppmd_flags = rar_br_bits(br, 7); 2085 rar_br_consume(br, 7); 2086 2087 /* Memory is allocated in MB */ 2088 if (ppmd_flags & 0x20) 2089 { 2090 if (!rar_br_read_ahead(a, br, 8)) 2091 goto truncated_data; 2092 rar->dictionary_size = (rar_br_bits(br, 8) + 1) << 20; 2093 rar_br_consume(br, 8); 2094 } 2095 2096 if (ppmd_flags & 0x40) 2097 { 2098 if (!rar_br_read_ahead(a, br, 8)) 2099 goto truncated_data; 2100 rar->ppmd_escape = rar->ppmd7_context.InitEsc = rar_br_bits(br, 8); 2101 rar_br_consume(br, 8); 2102 } 2103 else 2104 rar->ppmd_escape = 2; 2105 2106 if (ppmd_flags & 0x20) 2107 { 2108 maxorder = (ppmd_flags & 0x1F) + 1; 2109 if(maxorder > 16) 2110 maxorder = 16 + (maxorder - 16) * 3; 2111 2112 if (maxorder == 1) 2113 { 2114 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2115 "Truncated RAR file data"); 2116 return (ARCHIVE_FATAL); 2117 } 2118 2119 /* Make sure ppmd7_contest is freed before Ppmd7_Construct 2120 * because reading a broken file cause this abnormal sequence. */ 2121 __archive_ppmd7_functions.Ppmd7_Free(&rar->ppmd7_context, &g_szalloc); 2122 2123 rar->bytein.a = a; 2124 rar->bytein.Read = &ppmd_read; 2125 __archive_ppmd7_functions.PpmdRAR_RangeDec_CreateVTable(&rar->range_dec); 2126 rar->range_dec.Stream = &rar->bytein; 2127 __archive_ppmd7_functions.Ppmd7_Construct(&rar->ppmd7_context); 2128 2129 if (rar->dictionary_size == 0) { 2130 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2131 "Invalid zero dictionary size"); 2132 return (ARCHIVE_FATAL); 2133 } 2134 2135 if (!__archive_ppmd7_functions.Ppmd7_Alloc(&rar->ppmd7_context, 2136 rar->dictionary_size, &g_szalloc)) 2137 { 2138 archive_set_error(&a->archive, ENOMEM, 2139 "Out of memory"); 2140 return (ARCHIVE_FATAL); 2141 } 2142 if (!__archive_ppmd7_functions.PpmdRAR_RangeDec_Init(&rar->range_dec)) 2143 { 2144 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2145 "Unable to initialize PPMd range decoder"); 2146 return (ARCHIVE_FATAL); 2147 } 2148 __archive_ppmd7_functions.Ppmd7_Init(&rar->ppmd7_context, maxorder); 2149 rar->ppmd_valid = 1; 2150 } 2151 else 2152 { 2153 if (!rar->ppmd_valid) { 2154 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2155 "Invalid PPMd sequence"); 2156 return (ARCHIVE_FATAL); 2157 } 2158 if (!__archive_ppmd7_functions.PpmdRAR_RangeDec_Init(&rar->range_dec)) 2159 { 2160 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2161 "Unable to initialize PPMd range decoder"); 2162 return (ARCHIVE_FATAL); 2163 } 2164 } 2165 } 2166 else 2167 { 2168 rar_br_consume(br, 1); 2169 2170 /* Keep existing table flag */ 2171 if (!rar_br_read_ahead(a, br, 1)) 2172 goto truncated_data; 2173 if (!rar_br_bits(br, 1)) 2174 memset(rar->lengthtable, 0, sizeof(rar->lengthtable)); 2175 rar_br_consume(br, 1); 2176 2177 memset(&bitlengths, 0, sizeof(bitlengths)); 2178 for (i = 0; i < MAX_SYMBOLS;) 2179 { 2180 if (!rar_br_read_ahead(a, br, 4)) 2181 goto truncated_data; 2182 bitlengths[i++] = rar_br_bits(br, 4); 2183 rar_br_consume(br, 4); 2184 if (bitlengths[i-1] == 0xF) 2185 { 2186 if (!rar_br_read_ahead(a, br, 4)) 2187 goto truncated_data; 2188 zerocount = rar_br_bits(br, 4); 2189 rar_br_consume(br, 4); 2190 if (zerocount) 2191 { 2192 i--; 2193 for (j = 0; j < zerocount + 2 && i < MAX_SYMBOLS; j++) 2194 bitlengths[i++] = 0; 2195 } 2196 } 2197 } 2198 2199 memset(&precode, 0, sizeof(precode)); 2200 r = create_code(a, &precode, bitlengths, MAX_SYMBOLS, MAX_SYMBOL_LENGTH); 2201 if (r != ARCHIVE_OK) { 2202 free(precode.tree); 2203 free(precode.table); 2204 return (r); 2205 } 2206 2207 for (i = 0; i < HUFFMAN_TABLE_SIZE;) 2208 { 2209 if ((val = read_next_symbol(a, &precode)) < 0) { 2210 free(precode.tree); 2211 free(precode.table); 2212 return (ARCHIVE_FATAL); 2213 } 2214 if (val < 16) 2215 { 2216 rar->lengthtable[i] = (rar->lengthtable[i] + val) & 0xF; 2217 i++; 2218 } 2219 else if (val < 18) 2220 { 2221 if (i == 0) 2222 { 2223 free(precode.tree); 2224 free(precode.table); 2225 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2226 "Internal error extracting RAR file."); 2227 return (ARCHIVE_FATAL); 2228 } 2229 2230 if(val == 16) { 2231 if (!rar_br_read_ahead(a, br, 3)) { 2232 free(precode.tree); 2233 free(precode.table); 2234 goto truncated_data; 2235 } 2236 n = rar_br_bits(br, 3) + 3; 2237 rar_br_consume(br, 3); 2238 } else { 2239 if (!rar_br_read_ahead(a, br, 7)) { 2240 free(precode.tree); 2241 free(precode.table); 2242 goto truncated_data; 2243 } 2244 n = rar_br_bits(br, 7) + 11; 2245 rar_br_consume(br, 7); 2246 } 2247 2248 for (j = 0; j < n && i < HUFFMAN_TABLE_SIZE; j++) 2249 { 2250 rar->lengthtable[i] = rar->lengthtable[i-1]; 2251 i++; 2252 } 2253 } 2254 else 2255 { 2256 if(val == 18) { 2257 if (!rar_br_read_ahead(a, br, 3)) { 2258 free(precode.tree); 2259 free(precode.table); 2260 goto truncated_data; 2261 } 2262 n = rar_br_bits(br, 3) + 3; 2263 rar_br_consume(br, 3); 2264 } else { 2265 if (!rar_br_read_ahead(a, br, 7)) { 2266 free(precode.tree); 2267 free(precode.table); 2268 goto truncated_data; 2269 } 2270 n = rar_br_bits(br, 7) + 11; 2271 rar_br_consume(br, 7); 2272 } 2273 2274 for(j = 0; j < n && i < HUFFMAN_TABLE_SIZE; j++) 2275 rar->lengthtable[i++] = 0; 2276 } 2277 } 2278 free(precode.tree); 2279 free(precode.table); 2280 2281 r = create_code(a, &rar->maincode, &rar->lengthtable[0], MAINCODE_SIZE, 2282 MAX_SYMBOL_LENGTH); 2283 if (r != ARCHIVE_OK) 2284 return (r); 2285 r = create_code(a, &rar->offsetcode, &rar->lengthtable[MAINCODE_SIZE], 2286 OFFSETCODE_SIZE, MAX_SYMBOL_LENGTH); 2287 if (r != ARCHIVE_OK) 2288 return (r); 2289 r = create_code(a, &rar->lowoffsetcode, 2290 &rar->lengthtable[MAINCODE_SIZE + OFFSETCODE_SIZE], 2291 LOWOFFSETCODE_SIZE, MAX_SYMBOL_LENGTH); 2292 if (r != ARCHIVE_OK) 2293 return (r); 2294 r = create_code(a, &rar->lengthcode, 2295 &rar->lengthtable[MAINCODE_SIZE + OFFSETCODE_SIZE + 2296 LOWOFFSETCODE_SIZE], LENGTHCODE_SIZE, MAX_SYMBOL_LENGTH); 2297 if (r != ARCHIVE_OK) 2298 return (r); 2299 } 2300 2301 if (!rar->dictionary_size || !rar->lzss.window) 2302 { 2303 /* Seems as though dictionary sizes are not used. Even so, minimize 2304 * memory usage as much as possible. 2305 */ 2306 void *new_window; 2307 unsigned int new_size; 2308 2309 if (rar->unp_size >= DICTIONARY_MAX_SIZE) 2310 new_size = DICTIONARY_MAX_SIZE; 2311 else 2312 new_size = rar_fls((unsigned int)rar->unp_size) << 1; 2313 new_window = realloc(rar->lzss.window, new_size); 2314 if (new_window == NULL) { 2315 archive_set_error(&a->archive, ENOMEM, 2316 "Unable to allocate memory for uncompressed data."); 2317 return (ARCHIVE_FATAL); 2318 } 2319 rar->lzss.window = (unsigned char *)new_window; 2320 rar->dictionary_size = new_size; 2321 memset(rar->lzss.window, 0, rar->dictionary_size); 2322 rar->lzss.mask = rar->dictionary_size - 1; 2323 } 2324 2325 rar->start_new_table = 0; 2326 return (ARCHIVE_OK); 2327truncated_data: 2328 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2329 "Truncated RAR file data"); 2330 rar->valid = 0; 2331 return (ARCHIVE_FATAL); 2332} 2333 2334static void 2335free_codes(struct archive_read *a) 2336{ 2337 struct rar *rar = (struct rar *)(a->format->data); 2338 free(rar->maincode.tree); 2339 free(rar->offsetcode.tree); 2340 free(rar->lowoffsetcode.tree); 2341 free(rar->lengthcode.tree); 2342 free(rar->maincode.table); 2343 free(rar->offsetcode.table); 2344 free(rar->lowoffsetcode.table); 2345 free(rar->lengthcode.table); 2346 memset(&rar->maincode, 0, sizeof(rar->maincode)); 2347 memset(&rar->offsetcode, 0, sizeof(rar->offsetcode)); 2348 memset(&rar->lowoffsetcode, 0, sizeof(rar->lowoffsetcode)); 2349 memset(&rar->lengthcode, 0, sizeof(rar->lengthcode)); 2350} 2351 2352 2353static int 2354read_next_symbol(struct archive_read *a, struct huffman_code *code) 2355{ 2356 unsigned char bit; 2357 unsigned int bits; 2358 int length, value, node; 2359 struct rar *rar; 2360 struct rar_br *br; 2361 2362 if (!code->table) 2363 { 2364 if (make_table(a, code) != (ARCHIVE_OK)) 2365 return -1; 2366 } 2367 2368 rar = (struct rar *)(a->format->data); 2369 br = &(rar->br); 2370 2371 /* Look ahead (peek) at bits */ 2372 if (!rar_br_read_ahead(a, br, code->tablesize)) { 2373 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2374 "Truncated RAR file data"); 2375 rar->valid = 0; 2376 return -1; 2377 } 2378 bits = rar_br_bits(br, code->tablesize); 2379 2380 length = code->table[bits].length; 2381 value = code->table[bits].value; 2382 2383 if (length < 0) 2384 { 2385 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2386 "Invalid prefix code in bitstream"); 2387 return -1; 2388 } 2389 2390 if (length <= code->tablesize) 2391 { 2392 /* Skip length bits */ 2393 rar_br_consume(br, length); 2394 return value; 2395 } 2396 2397 /* Skip tablesize bits */ 2398 rar_br_consume(br, code->tablesize); 2399 2400 node = value; 2401 while (!(code->tree[node].branches[0] == 2402 code->tree[node].branches[1])) 2403 { 2404 if (!rar_br_read_ahead(a, br, 1)) { 2405 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2406 "Truncated RAR file data"); 2407 rar->valid = 0; 2408 return -1; 2409 } 2410 bit = rar_br_bits(br, 1); 2411 rar_br_consume(br, 1); 2412 2413 if (code->tree[node].branches[bit] < 0) 2414 { 2415 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2416 "Invalid prefix code in bitstream"); 2417 return -1; 2418 } 2419 node = code->tree[node].branches[bit]; 2420 } 2421 2422 return code->tree[node].branches[0]; 2423} 2424 2425static int 2426create_code(struct archive_read *a, struct huffman_code *code, 2427 unsigned char *lengths, int numsymbols, char maxlength) 2428{ 2429 int i, j, codebits = 0, symbolsleft = numsymbols; 2430 2431 code->numentries = 0; 2432 code->numallocatedentries = 0; 2433 if (new_node(code) < 0) { 2434 archive_set_error(&a->archive, ENOMEM, 2435 "Unable to allocate memory for node data."); 2436 return (ARCHIVE_FATAL); 2437 } 2438 code->numentries = 1; 2439 code->minlength = INT_MAX; 2440 code->maxlength = INT_MIN; 2441 codebits = 0; 2442 for(i = 1; i <= maxlength; i++) 2443 { 2444 for(j = 0; j < numsymbols; j++) 2445 { 2446 if (lengths[j] != i) continue; 2447 if (add_value(a, code, j, codebits, i) != ARCHIVE_OK) 2448 return (ARCHIVE_FATAL); 2449 codebits++; 2450 if (--symbolsleft <= 0) { break; break; } 2451 } 2452 codebits <<= 1; 2453 } 2454 return (ARCHIVE_OK); 2455} 2456 2457static int 2458add_value(struct archive_read *a, struct huffman_code *code, int value, 2459 int codebits, int length) 2460{ 2461 int repeatpos, lastnode, bitpos, bit, repeatnode, nextnode; 2462 2463 free(code->table); 2464 code->table = NULL; 2465 2466 if(length > code->maxlength) 2467 code->maxlength = length; 2468 if(length < code->minlength) 2469 code->minlength = length; 2470 2471 repeatpos = -1; 2472 if (repeatpos == 0 || (repeatpos >= 0 2473 && (((codebits >> (repeatpos - 1)) & 3) == 0 2474 || ((codebits >> (repeatpos - 1)) & 3) == 3))) 2475 { 2476 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2477 "Invalid repeat position"); 2478 return (ARCHIVE_FATAL); 2479 } 2480 2481 lastnode = 0; 2482 for (bitpos = length - 1; bitpos >= 0; bitpos--) 2483 { 2484 bit = (codebits >> bitpos) & 1; 2485 2486 /* Leaf node check */ 2487 if (code->tree[lastnode].branches[0] == 2488 code->tree[lastnode].branches[1]) 2489 { 2490 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2491 "Prefix found"); 2492 return (ARCHIVE_FATAL); 2493 } 2494 2495 if (bitpos == repeatpos) 2496 { 2497 /* Open branch check */ 2498 if (!(code->tree[lastnode].branches[bit] < 0)) 2499 { 2500 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2501 "Invalid repeating code"); 2502 return (ARCHIVE_FATAL); 2503 } 2504 2505 if ((repeatnode = new_node(code)) < 0) { 2506 archive_set_error(&a->archive, ENOMEM, 2507 "Unable to allocate memory for node data."); 2508 return (ARCHIVE_FATAL); 2509 } 2510 if ((nextnode = new_node(code)) < 0) { 2511 archive_set_error(&a->archive, ENOMEM, 2512 "Unable to allocate memory for node data."); 2513 return (ARCHIVE_FATAL); 2514 } 2515 2516 /* Set branches */ 2517 code->tree[lastnode].branches[bit] = repeatnode; 2518 code->tree[repeatnode].branches[bit] = repeatnode; 2519 code->tree[repeatnode].branches[bit^1] = nextnode; 2520 lastnode = nextnode; 2521 2522 bitpos++; /* terminating bit already handled, skip it */ 2523 } 2524 else 2525 { 2526 /* Open branch check */ 2527 if (code->tree[lastnode].branches[bit] < 0) 2528 { 2529 if (new_node(code) < 0) { 2530 archive_set_error(&a->archive, ENOMEM, 2531 "Unable to allocate memory for node data."); 2532 return (ARCHIVE_FATAL); 2533 } 2534 code->tree[lastnode].branches[bit] = code->numentries++; 2535 } 2536 2537 /* set to branch */ 2538 lastnode = code->tree[lastnode].branches[bit]; 2539 } 2540 } 2541 2542 if (!(code->tree[lastnode].branches[0] == -1 2543 && code->tree[lastnode].branches[1] == -2)) 2544 { 2545 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2546 "Prefix found"); 2547 return (ARCHIVE_FATAL); 2548 } 2549 2550 /* Set leaf value */ 2551 code->tree[lastnode].branches[0] = value; 2552 code->tree[lastnode].branches[1] = value; 2553 2554 return (ARCHIVE_OK); 2555} 2556 2557static int 2558new_node(struct huffman_code *code) 2559{ 2560 void *new_tree; 2561 if (code->numallocatedentries == code->numentries) { 2562 int new_num_entries = 256; 2563 if (code->numentries > 0) { 2564 new_num_entries = code->numentries * 2; 2565 } 2566 new_tree = realloc(code->tree, new_num_entries * sizeof(*code->tree)); 2567 if (new_tree == NULL) 2568 return (-1); 2569 code->tree = (struct huffman_tree_node *)new_tree; 2570 code->numallocatedentries = new_num_entries; 2571 } 2572 code->tree[code->numentries].branches[0] = -1; 2573 code->tree[code->numentries].branches[1] = -2; 2574 return 1; 2575} 2576 2577static int 2578make_table(struct archive_read *a, struct huffman_code *code) 2579{ 2580 if (code->maxlength < code->minlength || code->maxlength > 10) 2581 code->tablesize = 10; 2582 else 2583 code->tablesize = code->maxlength; 2584 2585 code->table = 2586 (struct huffman_table_entry *)calloc(1, sizeof(*code->table) 2587 * ((size_t)1 << code->tablesize)); 2588 2589 return make_table_recurse(a, code, 0, code->table, 0, code->tablesize); 2590} 2591 2592static int 2593make_table_recurse(struct archive_read *a, struct huffman_code *code, int node, 2594 struct huffman_table_entry *table, int depth, 2595 int maxdepth) 2596{ 2597 int currtablesize, i, ret = (ARCHIVE_OK); 2598 2599 if (!code->tree) 2600 { 2601 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2602 "Huffman tree was not created."); 2603 return (ARCHIVE_FATAL); 2604 } 2605 if (node < 0 || node >= code->numentries) 2606 { 2607 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2608 "Invalid location to Huffman tree specified."); 2609 return (ARCHIVE_FATAL); 2610 } 2611 2612 currtablesize = 1 << (maxdepth - depth); 2613 2614 if (code->tree[node].branches[0] == 2615 code->tree[node].branches[1]) 2616 { 2617 for(i = 0; i < currtablesize; i++) 2618 { 2619 table[i].length = depth; 2620 table[i].value = code->tree[node].branches[0]; 2621 } 2622 } 2623 else if (node < 0) 2624 { 2625 for(i = 0; i < currtablesize; i++) 2626 table[i].length = -1; 2627 } 2628 else 2629 { 2630 if(depth == maxdepth) 2631 { 2632 table[0].length = maxdepth + 1; 2633 table[0].value = node; 2634 } 2635 else 2636 { 2637 ret |= make_table_recurse(a, code, code->tree[node].branches[0], table, 2638 depth + 1, maxdepth); 2639 ret |= make_table_recurse(a, code, code->tree[node].branches[1], 2640 table + currtablesize / 2, depth + 1, maxdepth); 2641 } 2642 } 2643 return ret; 2644} 2645 2646static int64_t 2647expand(struct archive_read *a, int64_t end) 2648{ 2649 static const unsigned char lengthbases[] = 2650 { 0, 1, 2, 3, 4, 5, 6, 2651 7, 8, 10, 12, 14, 16, 20, 2652 24, 28, 32, 40, 48, 56, 64, 2653 80, 96, 112, 128, 160, 192, 224 }; 2654 static const unsigned char lengthbits[] = 2655 { 0, 0, 0, 0, 0, 0, 0, 2656 0, 1, 1, 1, 1, 2, 2, 2657 2, 2, 3, 3, 3, 3, 4, 2658 4, 4, 4, 5, 5, 5, 5 }; 2659 static const unsigned int offsetbases[] = 2660 { 0, 1, 2, 3, 4, 6, 2661 8, 12, 16, 24, 32, 48, 2662 64, 96, 128, 192, 256, 384, 2663 512, 768, 1024, 1536, 2048, 3072, 2664 4096, 6144, 8192, 12288, 16384, 24576, 2665 32768, 49152, 65536, 98304, 131072, 196608, 2666 262144, 327680, 393216, 458752, 524288, 589824, 2667 655360, 720896, 786432, 851968, 917504, 983040, 2668 1048576, 1310720, 1572864, 1835008, 2097152, 2359296, 2669 2621440, 2883584, 3145728, 3407872, 3670016, 3932160 }; 2670 static const unsigned char offsetbits[] = 2671 { 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 2672 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 2673 11, 11, 12, 12, 13, 13, 14, 14, 15, 15, 16, 16, 2674 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 2675 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18 }; 2676 static const unsigned char shortbases[] = 2677 { 0, 4, 8, 16, 32, 64, 128, 192 }; 2678 static const unsigned char shortbits[] = 2679 { 2, 2, 3, 4, 5, 6, 6, 6 }; 2680 2681 int symbol, offs, len, offsindex, lensymbol, i, offssymbol, lowoffsetsymbol; 2682 unsigned char newfile; 2683 struct rar *rar = (struct rar *)(a->format->data); 2684 struct rar_br *br = &(rar->br); 2685 2686 if (rar->filterstart < end) 2687 end = rar->filterstart; 2688 2689 while (1) 2690 { 2691 if (rar->output_last_match && 2692 lzss_position(&rar->lzss) + rar->lastlength <= end) 2693 { 2694 lzss_emit_match(rar, rar->lastoffset, rar->lastlength); 2695 rar->output_last_match = 0; 2696 } 2697 2698 if(rar->is_ppmd_block || rar->output_last_match || 2699 lzss_position(&rar->lzss) >= end) 2700 return lzss_position(&rar->lzss); 2701 2702 if ((symbol = read_next_symbol(a, &rar->maincode)) < 0) 2703 return (ARCHIVE_FATAL); 2704 rar->output_last_match = 0; 2705 2706 if (symbol < 256) 2707 { 2708 lzss_emit_literal(rar, symbol); 2709 continue; 2710 } 2711 else if (symbol == 256) 2712 { 2713 if (!rar_br_read_ahead(a, br, 1)) 2714 goto truncated_data; 2715 newfile = !rar_br_bits(br, 1); 2716 rar_br_consume(br, 1); 2717 2718 if(newfile) 2719 { 2720 rar->start_new_block = 1; 2721 if (!rar_br_read_ahead(a, br, 1)) 2722 goto truncated_data; 2723 rar->start_new_table = rar_br_bits(br, 1); 2724 rar_br_consume(br, 1); 2725 return lzss_position(&rar->lzss); 2726 } 2727 else 2728 { 2729 if (parse_codes(a) != ARCHIVE_OK) 2730 return (ARCHIVE_FATAL); 2731 continue; 2732 } 2733 } 2734 else if(symbol==257) 2735 { 2736 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 2737 "Parsing filters is unsupported."); 2738 return (ARCHIVE_FAILED); 2739 } 2740 else if(symbol==258) 2741 { 2742 if(rar->lastlength == 0) 2743 continue; 2744 2745 offs = rar->lastoffset; 2746 len = rar->lastlength; 2747 } 2748 else if (symbol <= 262) 2749 { 2750 offsindex = symbol - 259; 2751 offs = rar->oldoffset[offsindex]; 2752 2753 if ((lensymbol = read_next_symbol(a, &rar->lengthcode)) < 0) 2754 goto bad_data; 2755 if (lensymbol > (int)(sizeof(lengthbases)/sizeof(lengthbases[0]))) 2756 goto bad_data; 2757 if (lensymbol > (int)(sizeof(lengthbits)/sizeof(lengthbits[0]))) 2758 goto bad_data; 2759 len = lengthbases[lensymbol] + 2; 2760 if (lengthbits[lensymbol] > 0) { 2761 if (!rar_br_read_ahead(a, br, lengthbits[lensymbol])) 2762 goto truncated_data; 2763 len += rar_br_bits(br, lengthbits[lensymbol]); 2764 rar_br_consume(br, lengthbits[lensymbol]); 2765 } 2766 2767 for (i = offsindex; i > 0; i--) 2768 rar->oldoffset[i] = rar->oldoffset[i-1]; 2769 rar->oldoffset[0] = offs; 2770 } 2771 else if(symbol<=270) 2772 { 2773 offs = shortbases[symbol-263] + 1; 2774 if(shortbits[symbol-263] > 0) { 2775 if (!rar_br_read_ahead(a, br, shortbits[symbol-263])) 2776 goto truncated_data; 2777 offs += rar_br_bits(br, shortbits[symbol-263]); 2778 rar_br_consume(br, shortbits[symbol-263]); 2779 } 2780 2781 len = 2; 2782 2783 for(i = 3; i > 0; i--) 2784 rar->oldoffset[i] = rar->oldoffset[i-1]; 2785 rar->oldoffset[0] = offs; 2786 } 2787 else 2788 { 2789 if (symbol-271 > (int)(sizeof(lengthbases)/sizeof(lengthbases[0]))) 2790 goto bad_data; 2791 if (symbol-271 > (int)(sizeof(lengthbits)/sizeof(lengthbits[0]))) 2792 goto bad_data; 2793 len = lengthbases[symbol-271]+3; 2794 if(lengthbits[symbol-271] > 0) { 2795 if (!rar_br_read_ahead(a, br, lengthbits[symbol-271])) 2796 goto truncated_data; 2797 len += rar_br_bits(br, lengthbits[symbol-271]); 2798 rar_br_consume(br, lengthbits[symbol-271]); 2799 } 2800 2801 if ((offssymbol = read_next_symbol(a, &rar->offsetcode)) < 0) 2802 goto bad_data; 2803 if (offssymbol > (int)(sizeof(offsetbases)/sizeof(offsetbases[0]))) 2804 goto bad_data; 2805 if (offssymbol > (int)(sizeof(offsetbits)/sizeof(offsetbits[0]))) 2806 goto bad_data; 2807 offs = offsetbases[offssymbol]+1; 2808 if(offsetbits[offssymbol] > 0) 2809 { 2810 if(offssymbol > 9) 2811 { 2812 if(offsetbits[offssymbol] > 4) { 2813 if (!rar_br_read_ahead(a, br, offsetbits[offssymbol] - 4)) 2814 goto truncated_data; 2815 offs += rar_br_bits(br, offsetbits[offssymbol] - 4) << 4; 2816 rar_br_consume(br, offsetbits[offssymbol] - 4); 2817 } 2818 2819 if(rar->numlowoffsetrepeats > 0) 2820 { 2821 rar->numlowoffsetrepeats--; 2822 offs += rar->lastlowoffset; 2823 } 2824 else 2825 { 2826 if ((lowoffsetsymbol = 2827 read_next_symbol(a, &rar->lowoffsetcode)) < 0) 2828 return (ARCHIVE_FATAL); 2829 if(lowoffsetsymbol == 16) 2830 { 2831 rar->numlowoffsetrepeats = 15; 2832 offs += rar->lastlowoffset; 2833 } 2834 else 2835 { 2836 offs += lowoffsetsymbol; 2837 rar->lastlowoffset = lowoffsetsymbol; 2838 } 2839 } 2840 } 2841 else { 2842 if (!rar_br_read_ahead(a, br, offsetbits[offssymbol])) 2843 goto truncated_data; 2844 offs += rar_br_bits(br, offsetbits[offssymbol]); 2845 rar_br_consume(br, offsetbits[offssymbol]); 2846 } 2847 } 2848 2849 if (offs >= 0x40000) 2850 len++; 2851 if (offs >= 0x2000) 2852 len++; 2853 2854 for(i = 3; i > 0; i--) 2855 rar->oldoffset[i] = rar->oldoffset[i-1]; 2856 rar->oldoffset[0] = offs; 2857 } 2858 2859 rar->lastoffset = offs; 2860 rar->lastlength = len; 2861 rar->output_last_match = 1; 2862 } 2863truncated_data: 2864 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2865 "Truncated RAR file data"); 2866 rar->valid = 0; 2867 return (ARCHIVE_FATAL); 2868bad_data: 2869 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2870 "Bad RAR file data"); 2871 return (ARCHIVE_FATAL); 2872} 2873 2874static int 2875copy_from_lzss_window(struct archive_read *a, const void **buffer, 2876 int64_t startpos, int length) 2877{ 2878 int windowoffs, firstpart; 2879 struct rar *rar = (struct rar *)(a->format->data); 2880 2881 if (!rar->unp_buffer) 2882 { 2883 if ((rar->unp_buffer = malloc(rar->unp_buffer_size)) == NULL) 2884 { 2885 archive_set_error(&a->archive, ENOMEM, 2886 "Unable to allocate memory for uncompressed data."); 2887 return (ARCHIVE_FATAL); 2888 } 2889 } 2890 2891 windowoffs = lzss_offset_for_position(&rar->lzss, startpos); 2892 if(windowoffs + length <= lzss_size(&rar->lzss)) { 2893 memcpy(&rar->unp_buffer[rar->unp_offset], &rar->lzss.window[windowoffs], 2894 length); 2895 } else if (length <= lzss_size(&rar->lzss)) { 2896 firstpart = lzss_size(&rar->lzss) - windowoffs; 2897 if (firstpart < 0) { 2898 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2899 "Bad RAR file data"); 2900 return (ARCHIVE_FATAL); 2901 } 2902 if (firstpart < length) { 2903 memcpy(&rar->unp_buffer[rar->unp_offset], 2904 &rar->lzss.window[windowoffs], firstpart); 2905 memcpy(&rar->unp_buffer[rar->unp_offset + firstpart], 2906 &rar->lzss.window[0], length - firstpart); 2907 } else { 2908 memcpy(&rar->unp_buffer[rar->unp_offset], 2909 &rar->lzss.window[windowoffs], length); 2910 } 2911 } else { 2912 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 2913 "Bad RAR file data"); 2914 return (ARCHIVE_FATAL); 2915 } 2916 rar->unp_offset += length; 2917 if (rar->unp_offset >= rar->unp_buffer_size) 2918 *buffer = rar->unp_buffer; 2919 else 2920 *buffer = NULL; 2921 return (ARCHIVE_OK); 2922} 2923 2924static const void * 2925rar_read_ahead(struct archive_read *a, size_t min, ssize_t *avail) 2926{ 2927 struct rar *rar = (struct rar *)(a->format->data); 2928 const void *h = __archive_read_ahead(a, min, avail); 2929 int ret; 2930 if (avail) 2931 { 2932 if (a->archive.read_data_is_posix_read && *avail > (ssize_t)a->archive.read_data_requested) 2933 *avail = a->archive.read_data_requested; 2934 if (*avail > rar->bytes_remaining) 2935 *avail = (ssize_t)rar->bytes_remaining; 2936 if (*avail < 0) 2937 return NULL; 2938 else if (*avail == 0 && rar->main_flags & MHD_VOLUME && 2939 rar->file_flags & FHD_SPLIT_AFTER) 2940 { 2941 ret = archive_read_format_rar_read_header(a, a->entry); 2942 if (ret == (ARCHIVE_EOF)) 2943 { 2944 rar->has_endarc_header = 1; 2945 ret = archive_read_format_rar_read_header(a, a->entry); 2946 } 2947 if (ret != (ARCHIVE_OK)) 2948 return NULL; 2949 return rar_read_ahead(a, min, avail); 2950 } 2951 } 2952 return h; 2953} 2954