archive_read_support_format_lha.c revision 358090
1/*- 2 * Copyright (c) 2008-2014 Michihiro NAKAJIMA 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR 15 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 16 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 17 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, 18 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 19 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 20 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 21 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 22 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 23 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 24 */ 25 26#include "archive_platform.h" 27 28#ifdef HAVE_ERRNO_H 29#include <errno.h> 30#endif 31#ifdef HAVE_LIMITS_H 32#include <limits.h> 33#endif 34#ifdef HAVE_STDLIB_H 35#include <stdlib.h> 36#endif 37#ifdef HAVE_STRING_H 38#include <string.h> 39#endif 40 41#include "archive.h" 42#include "archive_entry.h" 43#include "archive_entry_locale.h" 44#include "archive_private.h" 45#include "archive_read_private.h" 46#include "archive_endian.h" 47 48 49#define MAXMATCH 256 /* Maximum match length. */ 50#define MINMATCH 3 /* Minimum match length. */ 51/* 52 * Literal table format: 53 * +0 +256 +510 54 * +---------------+-------------------------+ 55 * | literal code | match length | 56 * | 0 ... 255 | MINMATCH ... MAXMATCH | 57 * +---------------+-------------------------+ 58 * <--- LT_BITLEN_SIZE ---> 59 */ 60/* Literal table size. */ 61#define LT_BITLEN_SIZE (UCHAR_MAX + 1 + MAXMATCH - MINMATCH + 1) 62/* Position table size. 63 * Note: this used for both position table and pre literal table.*/ 64#define PT_BITLEN_SIZE (3 + 16) 65 66struct lzh_dec { 67 /* Decoding status. */ 68 int state; 69 70 /* 71 * Window to see last 8Ki(lh5),32Ki(lh6),64Ki(lh7) bytes of decoded 72 * data. 73 */ 74 int w_size; 75 int w_mask; 76 /* Window buffer, which is a loop buffer. */ 77 unsigned char *w_buff; 78 /* The insert position to the window. */ 79 int w_pos; 80 /* The position where we can copy decoded code from the window. */ 81 int copy_pos; 82 /* The length how many bytes we can copy decoded code from 83 * the window. */ 84 int copy_len; 85 86 /* 87 * Bit stream reader. 88 */ 89 struct lzh_br { 90#define CACHE_TYPE uint64_t 91#define CACHE_BITS (8 * sizeof(CACHE_TYPE)) 92 /* Cache buffer. */ 93 CACHE_TYPE cache_buffer; 94 /* Indicates how many bits avail in cache_buffer. */ 95 int cache_avail; 96 } br; 97 98 /* 99 * Huffman coding. 100 */ 101 struct huffman { 102 int len_size; 103 int len_avail; 104 int len_bits; 105 int freq[17]; 106 unsigned char *bitlen; 107 108 /* 109 * Use a index table. It's faster than searching a huffman 110 * coding tree, which is a binary tree. But a use of a large 111 * index table causes L1 cache read miss many times. 112 */ 113#define HTBL_BITS 10 114 int max_bits; 115 int shift_bits; 116 int tbl_bits; 117 int tree_used; 118 int tree_avail; 119 /* Direct access table. */ 120 uint16_t *tbl; 121 /* Binary tree table for extra bits over the direct access. */ 122 struct htree_t { 123 uint16_t left; 124 uint16_t right; 125 } *tree; 126 } lt, pt; 127 128 int blocks_avail; 129 int pos_pt_len_size; 130 int pos_pt_len_bits; 131 int literal_pt_len_size; 132 int literal_pt_len_bits; 133 int reading_position; 134 int loop; 135 int error; 136}; 137 138struct lzh_stream { 139 const unsigned char *next_in; 140 int avail_in; 141 int64_t total_in; 142 const unsigned char *ref_ptr; 143 int avail_out; 144 int64_t total_out; 145 struct lzh_dec *ds; 146}; 147 148struct lha { 149 /* entry_bytes_remaining is the number of bytes we expect. */ 150 int64_t entry_offset; 151 int64_t entry_bytes_remaining; 152 int64_t entry_unconsumed; 153 uint16_t entry_crc_calculated; 154 155 size_t header_size; /* header size */ 156 unsigned char level; /* header level */ 157 char method[3]; /* compress type */ 158 int64_t compsize; /* compressed data size */ 159 int64_t origsize; /* original file size */ 160 int setflag; 161#define BIRTHTIME_IS_SET 1 162#define ATIME_IS_SET 2 163#define UNIX_MODE_IS_SET 4 164#define CRC_IS_SET 8 165 time_t birthtime; 166 long birthtime_tv_nsec; 167 time_t mtime; 168 long mtime_tv_nsec; 169 time_t atime; 170 long atime_tv_nsec; 171 mode_t mode; 172 int64_t uid; 173 int64_t gid; 174 struct archive_string uname; 175 struct archive_string gname; 176 uint16_t header_crc; 177 uint16_t crc; 178 /* dirname and filename could be in different codepages */ 179 struct archive_string_conv *sconv_dir; 180 struct archive_string_conv *sconv_fname; 181 struct archive_string_conv *opt_sconv; 182 183 struct archive_string dirname; 184 struct archive_string filename; 185 struct archive_wstring ws; 186 187 unsigned char dos_attr; 188 189 /* Flag to mark progress that an archive was read their first header.*/ 190 char found_first_header; 191 /* Flag to mark that indicates an empty directory. */ 192 char directory; 193 194 /* Flags to mark progress of decompression. */ 195 char decompress_init; 196 char end_of_entry; 197 char end_of_entry_cleanup; 198 char entry_is_compressed; 199 200 char format_name[64]; 201 202 struct lzh_stream strm; 203}; 204 205/* 206 * LHA header common member offset. 207 */ 208#define H_METHOD_OFFSET 2 /* Compress type. */ 209#define H_ATTR_OFFSET 19 /* DOS attribute. */ 210#define H_LEVEL_OFFSET 20 /* Header Level. */ 211#define H_SIZE 22 /* Minimum header size. */ 212 213static int archive_read_format_lha_bid(struct archive_read *, int); 214static int archive_read_format_lha_options(struct archive_read *, 215 const char *, const char *); 216static int archive_read_format_lha_read_header(struct archive_read *, 217 struct archive_entry *); 218static int archive_read_format_lha_read_data(struct archive_read *, 219 const void **, size_t *, int64_t *); 220static int archive_read_format_lha_read_data_skip(struct archive_read *); 221static int archive_read_format_lha_cleanup(struct archive_read *); 222 223static void lha_replace_path_separator(struct lha *, 224 struct archive_entry *); 225static int lha_read_file_header_0(struct archive_read *, struct lha *); 226static int lha_read_file_header_1(struct archive_read *, struct lha *); 227static int lha_read_file_header_2(struct archive_read *, struct lha *); 228static int lha_read_file_header_3(struct archive_read *, struct lha *); 229static int lha_read_file_extended_header(struct archive_read *, 230 struct lha *, uint16_t *, int, size_t, size_t *); 231static size_t lha_check_header_format(const void *); 232static int lha_skip_sfx(struct archive_read *); 233static time_t lha_dos_time(const unsigned char *); 234static time_t lha_win_time(uint64_t, long *); 235static unsigned char lha_calcsum(unsigned char, const void *, 236 int, size_t); 237static int lha_parse_linkname(struct archive_wstring *, 238 struct archive_wstring *); 239static int lha_read_data_none(struct archive_read *, const void **, 240 size_t *, int64_t *); 241static int lha_read_data_lzh(struct archive_read *, const void **, 242 size_t *, int64_t *); 243static void lha_crc16_init(void); 244static uint16_t lha_crc16(uint16_t, const void *, size_t); 245static int lzh_decode_init(struct lzh_stream *, const char *); 246static void lzh_decode_free(struct lzh_stream *); 247static int lzh_decode(struct lzh_stream *, int); 248static int lzh_br_fillup(struct lzh_stream *, struct lzh_br *); 249static int lzh_huffman_init(struct huffman *, size_t, int); 250static void lzh_huffman_free(struct huffman *); 251static int lzh_read_pt_bitlen(struct lzh_stream *, int start, int end); 252static int lzh_make_fake_table(struct huffman *, uint16_t); 253static int lzh_make_huffman_table(struct huffman *); 254static inline int lzh_decode_huffman(struct huffman *, unsigned); 255static int lzh_decode_huffman_tree(struct huffman *, unsigned, int); 256 257 258int 259archive_read_support_format_lha(struct archive *_a) 260{ 261 struct archive_read *a = (struct archive_read *)_a; 262 struct lha *lha; 263 int r; 264 265 archive_check_magic(_a, ARCHIVE_READ_MAGIC, 266 ARCHIVE_STATE_NEW, "archive_read_support_format_lha"); 267 268 lha = (struct lha *)calloc(1, sizeof(*lha)); 269 if (lha == NULL) { 270 archive_set_error(&a->archive, ENOMEM, 271 "Can't allocate lha data"); 272 return (ARCHIVE_FATAL); 273 } 274 archive_string_init(&lha->ws); 275 276 r = __archive_read_register_format(a, 277 lha, 278 "lha", 279 archive_read_format_lha_bid, 280 archive_read_format_lha_options, 281 archive_read_format_lha_read_header, 282 archive_read_format_lha_read_data, 283 archive_read_format_lha_read_data_skip, 284 NULL, 285 archive_read_format_lha_cleanup, 286 NULL, 287 NULL); 288 289 if (r != ARCHIVE_OK) 290 free(lha); 291 return (ARCHIVE_OK); 292} 293 294static size_t 295lha_check_header_format(const void *h) 296{ 297 const unsigned char *p = h; 298 size_t next_skip_bytes; 299 300 switch (p[H_METHOD_OFFSET+3]) { 301 /* 302 * "-lh0-" ... "-lh7-" "-lhd-" 303 * "-lzs-" "-lz5-" 304 */ 305 case '0': case '1': case '2': case '3': 306 case '4': case '5': case '6': case '7': 307 case 'd': 308 case 's': 309 next_skip_bytes = 4; 310 311 /* b0 == 0 means the end of an LHa archive file. */ 312 if (p[0] == 0) 313 break; 314 if (p[H_METHOD_OFFSET] != '-' || p[H_METHOD_OFFSET+1] != 'l' 315 || p[H_METHOD_OFFSET+4] != '-') 316 break; 317 318 if (p[H_METHOD_OFFSET+2] == 'h') { 319 /* "-lh?-" */ 320 if (p[H_METHOD_OFFSET+3] == 's') 321 break; 322 if (p[H_LEVEL_OFFSET] == 0) 323 return (0); 324 if (p[H_LEVEL_OFFSET] <= 3 && p[H_ATTR_OFFSET] == 0x20) 325 return (0); 326 } 327 if (p[H_METHOD_OFFSET+2] == 'z') { 328 /* LArc extensions: -lzs-,-lz4- and -lz5- */ 329 if (p[H_LEVEL_OFFSET] != 0) 330 break; 331 if (p[H_METHOD_OFFSET+3] == 's' 332 || p[H_METHOD_OFFSET+3] == '4' 333 || p[H_METHOD_OFFSET+3] == '5') 334 return (0); 335 } 336 break; 337 case 'h': next_skip_bytes = 1; break; 338 case 'z': next_skip_bytes = 1; break; 339 case 'l': next_skip_bytes = 2; break; 340 case '-': next_skip_bytes = 3; break; 341 default : next_skip_bytes = 4; break; 342 } 343 344 return (next_skip_bytes); 345} 346 347static int 348archive_read_format_lha_bid(struct archive_read *a, int best_bid) 349{ 350 const char *p; 351 const void *buff; 352 ssize_t bytes_avail, offset, window; 353 size_t next; 354 355 /* If there's already a better bid than we can ever 356 make, don't bother testing. */ 357 if (best_bid > 30) 358 return (-1); 359 360 if ((p = __archive_read_ahead(a, H_SIZE, NULL)) == NULL) 361 return (-1); 362 363 if (lha_check_header_format(p) == 0) 364 return (30); 365 366 if (p[0] == 'M' && p[1] == 'Z') { 367 /* PE file */ 368 offset = 0; 369 window = 4096; 370 while (offset < (1024 * 20)) { 371 buff = __archive_read_ahead(a, offset + window, 372 &bytes_avail); 373 if (buff == NULL) { 374 /* Remaining bytes are less than window. */ 375 window >>= 1; 376 if (window < (H_SIZE + 3)) 377 return (0); 378 continue; 379 } 380 p = (const char *)buff + offset; 381 while (p + H_SIZE < (const char *)buff + bytes_avail) { 382 if ((next = lha_check_header_format(p)) == 0) 383 return (30); 384 p += next; 385 } 386 offset = p - (const char *)buff; 387 } 388 } 389 return (0); 390} 391 392static int 393archive_read_format_lha_options(struct archive_read *a, 394 const char *key, const char *val) 395{ 396 struct lha *lha; 397 int ret = ARCHIVE_FAILED; 398 399 lha = (struct lha *)(a->format->data); 400 if (strcmp(key, "hdrcharset") == 0) { 401 if (val == NULL || val[0] == 0) 402 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 403 "lha: hdrcharset option needs a character-set name"); 404 else { 405 lha->opt_sconv = 406 archive_string_conversion_from_charset( 407 &a->archive, val, 0); 408 if (lha->opt_sconv != NULL) 409 ret = ARCHIVE_OK; 410 else 411 ret = ARCHIVE_FATAL; 412 } 413 return (ret); 414 } 415 416 /* Note: The "warn" return is just to inform the options 417 * supervisor that we didn't handle it. It will generate 418 * a suitable error if no one used this option. */ 419 return (ARCHIVE_WARN); 420} 421 422static int 423lha_skip_sfx(struct archive_read *a) 424{ 425 const void *h; 426 const char *p, *q; 427 size_t next, skip; 428 ssize_t bytes, window; 429 430 window = 4096; 431 for (;;) { 432 h = __archive_read_ahead(a, window, &bytes); 433 if (h == NULL) { 434 /* Remaining bytes are less than window. */ 435 window >>= 1; 436 if (window < (H_SIZE + 3)) 437 goto fatal; 438 continue; 439 } 440 if (bytes < H_SIZE) 441 goto fatal; 442 p = h; 443 q = p + bytes; 444 445 /* 446 * Scan ahead until we find something that looks 447 * like the lha header. 448 */ 449 while (p + H_SIZE < q) { 450 if ((next = lha_check_header_format(p)) == 0) { 451 skip = p - (const char *)h; 452 __archive_read_consume(a, skip); 453 return (ARCHIVE_OK); 454 } 455 p += next; 456 } 457 skip = p - (const char *)h; 458 __archive_read_consume(a, skip); 459 } 460fatal: 461 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 462 "Couldn't find out LHa header"); 463 return (ARCHIVE_FATAL); 464} 465 466static int 467truncated_error(struct archive_read *a) 468{ 469 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 470 "Truncated LHa header"); 471 return (ARCHIVE_FATAL); 472} 473 474static int 475archive_read_format_lha_read_header(struct archive_read *a, 476 struct archive_entry *entry) 477{ 478 struct archive_wstring linkname; 479 struct archive_wstring pathname; 480 struct lha *lha; 481 const unsigned char *p; 482 const char *signature; 483 int err; 484 struct archive_mstring conv_buffer; 485 const wchar_t *conv_buffer_p; 486 487 lha_crc16_init(); 488 489 a->archive.archive_format = ARCHIVE_FORMAT_LHA; 490 if (a->archive.archive_format_name == NULL) 491 a->archive.archive_format_name = "lha"; 492 493 lha = (struct lha *)(a->format->data); 494 lha->decompress_init = 0; 495 lha->end_of_entry = 0; 496 lha->end_of_entry_cleanup = 0; 497 lha->entry_unconsumed = 0; 498 499 if ((p = __archive_read_ahead(a, H_SIZE, NULL)) == NULL) { 500 /* 501 * LHa archiver added 0 to the tail of its archive file as 502 * the mark of the end of the archive. 503 */ 504 signature = __archive_read_ahead(a, sizeof(signature[0]), NULL); 505 if (signature == NULL || signature[0] == 0) 506 return (ARCHIVE_EOF); 507 return (truncated_error(a)); 508 } 509 510 signature = (const char *)p; 511 if (lha->found_first_header == 0 && 512 signature[0] == 'M' && signature[1] == 'Z') { 513 /* This is an executable? Must be self-extracting... */ 514 err = lha_skip_sfx(a); 515 if (err < ARCHIVE_WARN) 516 return (err); 517 518 if ((p = __archive_read_ahead(a, sizeof(*p), NULL)) == NULL) 519 return (truncated_error(a)); 520 signature = (const char *)p; 521 } 522 /* signature[0] == 0 means the end of an LHa archive file. */ 523 if (signature[0] == 0) 524 return (ARCHIVE_EOF); 525 526 /* 527 * Check the header format and method type. 528 */ 529 if (lha_check_header_format(p) != 0) { 530 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 531 "Bad LHa file"); 532 return (ARCHIVE_FATAL); 533 } 534 535 /* We've found the first header. */ 536 lha->found_first_header = 1; 537 /* Set a default value and common data */ 538 lha->header_size = 0; 539 lha->level = p[H_LEVEL_OFFSET]; 540 lha->method[0] = p[H_METHOD_OFFSET+1]; 541 lha->method[1] = p[H_METHOD_OFFSET+2]; 542 lha->method[2] = p[H_METHOD_OFFSET+3]; 543 if (memcmp(lha->method, "lhd", 3) == 0) 544 lha->directory = 1; 545 else 546 lha->directory = 0; 547 if (memcmp(lha->method, "lh0", 3) == 0 || 548 memcmp(lha->method, "lz4", 3) == 0) 549 lha->entry_is_compressed = 0; 550 else 551 lha->entry_is_compressed = 1; 552 553 lha->compsize = 0; 554 lha->origsize = 0; 555 lha->setflag = 0; 556 lha->birthtime = 0; 557 lha->birthtime_tv_nsec = 0; 558 lha->mtime = 0; 559 lha->mtime_tv_nsec = 0; 560 lha->atime = 0; 561 lha->atime_tv_nsec = 0; 562 lha->mode = (lha->directory)? 0777 : 0666; 563 lha->uid = 0; 564 lha->gid = 0; 565 archive_string_empty(&lha->dirname); 566 archive_string_empty(&lha->filename); 567 lha->dos_attr = 0; 568 if (lha->opt_sconv != NULL) { 569 lha->sconv_dir = lha->opt_sconv; 570 lha->sconv_fname = lha->opt_sconv; 571 } else { 572 lha->sconv_dir = NULL; 573 lha->sconv_fname = NULL; 574 } 575 576 switch (p[H_LEVEL_OFFSET]) { 577 case 0: 578 err = lha_read_file_header_0(a, lha); 579 break; 580 case 1: 581 err = lha_read_file_header_1(a, lha); 582 break; 583 case 2: 584 err = lha_read_file_header_2(a, lha); 585 break; 586 case 3: 587 err = lha_read_file_header_3(a, lha); 588 break; 589 default: 590 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 591 "Unsupported LHa header level %d", p[H_LEVEL_OFFSET]); 592 err = ARCHIVE_FATAL; 593 break; 594 } 595 if (err < ARCHIVE_WARN) 596 return (err); 597 598 599 if (!lha->directory && archive_strlen(&lha->filename) == 0) 600 /* The filename has not been set */ 601 return (truncated_error(a)); 602 603 /* 604 * Make a pathname from a dirname and a filename, after converting to Unicode. 605 * This is because codepages might differ between dirname and filename. 606 */ 607 archive_string_init(&pathname); 608 archive_string_init(&linkname); 609 archive_string_init(&conv_buffer.aes_mbs); 610 archive_string_init(&conv_buffer.aes_mbs_in_locale); 611 archive_string_init(&conv_buffer.aes_utf8); 612 archive_string_init(&conv_buffer.aes_wcs); 613 if (0 != archive_mstring_copy_mbs_len_l(&conv_buffer, lha->dirname.s, lha->dirname.length, lha->sconv_dir)) { 614 archive_set_error(&a->archive, 615 ARCHIVE_ERRNO_FILE_FORMAT, 616 "Pathname cannot be converted " 617 "from %s to Unicode.", 618 archive_string_conversion_charset_name(lha->sconv_dir)); 619 err = ARCHIVE_FATAL; 620 } else if (0 != archive_mstring_get_wcs(&a->archive, &conv_buffer, &conv_buffer_p)) 621 err = ARCHIVE_FATAL; 622 if (err == ARCHIVE_FATAL) { 623 archive_mstring_clean(&conv_buffer); 624 archive_wstring_free(&pathname); 625 archive_wstring_free(&linkname); 626 return (err); 627 } 628 archive_wstring_copy(&pathname, &conv_buffer.aes_wcs); 629 630 archive_string_empty(&conv_buffer.aes_mbs); 631 archive_string_empty(&conv_buffer.aes_mbs_in_locale); 632 archive_string_empty(&conv_buffer.aes_utf8); 633 archive_wstring_empty(&conv_buffer.aes_wcs); 634 if (0 != archive_mstring_copy_mbs_len_l(&conv_buffer, lha->filename.s, lha->filename.length, lha->sconv_fname)) { 635 archive_set_error(&a->archive, 636 ARCHIVE_ERRNO_FILE_FORMAT, 637 "Pathname cannot be converted " 638 "from %s to Unicode.", 639 archive_string_conversion_charset_name(lha->sconv_fname)); 640 err = ARCHIVE_FATAL; 641 } 642 else if (0 != archive_mstring_get_wcs(&a->archive, &conv_buffer, &conv_buffer_p)) 643 err = ARCHIVE_FATAL; 644 if (err == ARCHIVE_FATAL) { 645 archive_mstring_clean(&conv_buffer); 646 archive_wstring_free(&pathname); 647 archive_wstring_free(&linkname); 648 return (err); 649 } 650 archive_wstring_concat(&pathname, &conv_buffer.aes_wcs); 651 archive_mstring_clean(&conv_buffer); 652 653 if ((lha->mode & AE_IFMT) == AE_IFLNK) { 654 /* 655 * Extract the symlink-name if it's included in the pathname. 656 */ 657 if (!lha_parse_linkname(&linkname, &pathname)) { 658 /* We couldn't get the symlink-name. */ 659 archive_set_error(&a->archive, 660 ARCHIVE_ERRNO_FILE_FORMAT, 661 "Unknown symlink-name"); 662 archive_wstring_free(&pathname); 663 archive_wstring_free(&linkname); 664 return (ARCHIVE_FAILED); 665 } 666 } else { 667 /* 668 * Make sure a file-type is set. 669 * The mode has been overridden if it is in the extended data. 670 */ 671 lha->mode = (lha->mode & ~AE_IFMT) | 672 ((lha->directory)? AE_IFDIR: AE_IFREG); 673 } 674 if ((lha->setflag & UNIX_MODE_IS_SET) == 0 && 675 (lha->dos_attr & 1) != 0) 676 lha->mode &= ~(0222);/* read only. */ 677 678 /* 679 * Set basic file parameters. 680 */ 681 archive_entry_copy_pathname_w(entry, pathname.s); 682 archive_wstring_free(&pathname); 683 if (archive_strlen(&linkname) > 0) { 684 archive_entry_copy_symlink_w(entry, linkname.s); 685 } else 686 archive_entry_set_symlink(entry, NULL); 687 archive_wstring_free(&linkname); 688 /* 689 * When a header level is 0, there is a possibility that 690 * a pathname and a symlink has '\' character, a directory 691 * separator in DOS/Windows. So we should convert it to '/'. 692 */ 693 if (p[H_LEVEL_OFFSET] == 0) 694 lha_replace_path_separator(lha, entry); 695 696 archive_entry_set_mode(entry, lha->mode); 697 archive_entry_set_uid(entry, lha->uid); 698 archive_entry_set_gid(entry, lha->gid); 699 if (archive_strlen(&lha->uname) > 0) 700 archive_entry_set_uname(entry, lha->uname.s); 701 if (archive_strlen(&lha->gname) > 0) 702 archive_entry_set_gname(entry, lha->gname.s); 703 if (lha->setflag & BIRTHTIME_IS_SET) { 704 archive_entry_set_birthtime(entry, lha->birthtime, 705 lha->birthtime_tv_nsec); 706 archive_entry_set_ctime(entry, lha->birthtime, 707 lha->birthtime_tv_nsec); 708 } else { 709 archive_entry_unset_birthtime(entry); 710 archive_entry_unset_ctime(entry); 711 } 712 archive_entry_set_mtime(entry, lha->mtime, lha->mtime_tv_nsec); 713 if (lha->setflag & ATIME_IS_SET) 714 archive_entry_set_atime(entry, lha->atime, 715 lha->atime_tv_nsec); 716 else 717 archive_entry_unset_atime(entry); 718 if (lha->directory || archive_entry_symlink(entry) != NULL) 719 archive_entry_unset_size(entry); 720 else 721 archive_entry_set_size(entry, lha->origsize); 722 723 /* 724 * Prepare variables used to read a file content. 725 */ 726 lha->entry_bytes_remaining = lha->compsize; 727 if (lha->entry_bytes_remaining < 0) { 728 archive_set_error(&a->archive, 729 ARCHIVE_ERRNO_FILE_FORMAT, 730 "Invalid LHa entry size"); 731 return (ARCHIVE_FATAL); 732 } 733 lha->entry_offset = 0; 734 lha->entry_crc_calculated = 0; 735 736 /* 737 * This file does not have a content. 738 */ 739 if (lha->directory || lha->compsize == 0) 740 lha->end_of_entry = 1; 741 742 sprintf(lha->format_name, "lha -%c%c%c-", 743 lha->method[0], lha->method[1], lha->method[2]); 744 a->archive.archive_format_name = lha->format_name; 745 746 return (err); 747} 748 749/* 750 * Replace a DOS path separator '\' by a character '/'. 751 * Some multi-byte character set have a character '\' in its second byte. 752 */ 753static void 754lha_replace_path_separator(struct lha *lha, struct archive_entry *entry) 755{ 756 const wchar_t *wp; 757 size_t i; 758 759 if ((wp = archive_entry_pathname_w(entry)) != NULL) { 760 archive_wstrcpy(&(lha->ws), wp); 761 for (i = 0; i < archive_strlen(&(lha->ws)); i++) { 762 if (lha->ws.s[i] == L'\\') 763 lha->ws.s[i] = L'/'; 764 } 765 archive_entry_copy_pathname_w(entry, lha->ws.s); 766 } 767 768 if ((wp = archive_entry_symlink_w(entry)) != NULL) { 769 archive_wstrcpy(&(lha->ws), wp); 770 for (i = 0; i < archive_strlen(&(lha->ws)); i++) { 771 if (lha->ws.s[i] == L'\\') 772 lha->ws.s[i] = L'/'; 773 } 774 archive_entry_copy_symlink_w(entry, lha->ws.s); 775 } 776} 777 778/* 779 * Header 0 format 780 * 781 * +0 +1 +2 +7 +11 782 * +---------------+----------+----------------+-------------------+ 783 * |header size(*1)|header sum|compression type|compressed size(*2)| 784 * +---------------+----------+----------------+-------------------+ 785 * <---------------------(*1)----------* 786 * 787 * +11 +15 +17 +19 +20 +21 788 * +-----------------+---------+---------+--------------+----------------+ 789 * |uncompressed size|time(DOS)|date(DOS)|attribute(DOS)|header level(=0)| 790 * +-----------------+---------+---------+--------------+----------------+ 791 * *--------------------------------(*1)---------------------------------* 792 * 793 * +21 +22 +22+(*3) +22+(*3)+2 +22+(*3)+2+(*4) 794 * +---------------+---------+----------+----------------+------------------+ 795 * |name length(*3)|file name|file CRC16|extra header(*4)| compressed data | 796 * +---------------+---------+----------+----------------+------------------+ 797 * <--(*3)-> <------(*2)------> 798 * *----------------------(*1)--------------------------> 799 * 800 */ 801#define H0_HEADER_SIZE_OFFSET 0 802#define H0_HEADER_SUM_OFFSET 1 803#define H0_COMP_SIZE_OFFSET 7 804#define H0_ORIG_SIZE_OFFSET 11 805#define H0_DOS_TIME_OFFSET 15 806#define H0_NAME_LEN_OFFSET 21 807#define H0_FILE_NAME_OFFSET 22 808#define H0_FIXED_SIZE 24 809static int 810lha_read_file_header_0(struct archive_read *a, struct lha *lha) 811{ 812 const unsigned char *p; 813 int extdsize, namelen; 814 unsigned char headersum, sum_calculated; 815 816 if ((p = __archive_read_ahead(a, H0_FIXED_SIZE, NULL)) == NULL) 817 return (truncated_error(a)); 818 lha->header_size = p[H0_HEADER_SIZE_OFFSET] + 2; 819 headersum = p[H0_HEADER_SUM_OFFSET]; 820 lha->compsize = archive_le32dec(p + H0_COMP_SIZE_OFFSET); 821 lha->origsize = archive_le32dec(p + H0_ORIG_SIZE_OFFSET); 822 lha->mtime = lha_dos_time(p + H0_DOS_TIME_OFFSET); 823 namelen = p[H0_NAME_LEN_OFFSET]; 824 extdsize = (int)lha->header_size - H0_FIXED_SIZE - namelen; 825 if ((namelen > 221 || extdsize < 0) && extdsize != -2) { 826 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 827 "Invalid LHa header"); 828 return (ARCHIVE_FATAL); 829 } 830 if ((p = __archive_read_ahead(a, lha->header_size, NULL)) == NULL) 831 return (truncated_error(a)); 832 833 archive_strncpy(&lha->filename, p + H0_FILE_NAME_OFFSET, namelen); 834 /* When extdsize == -2, A CRC16 value is not present in the header. */ 835 if (extdsize >= 0) { 836 lha->crc = archive_le16dec(p + H0_FILE_NAME_OFFSET + namelen); 837 lha->setflag |= CRC_IS_SET; 838 } 839 sum_calculated = lha_calcsum(0, p, 2, lha->header_size - 2); 840 841 /* Read an extended header */ 842 if (extdsize > 0) { 843 /* This extended data is set by 'LHa for UNIX' only. 844 * Maybe fixed size. 845 */ 846 p += H0_FILE_NAME_OFFSET + namelen + 2; 847 if (p[0] == 'U' && extdsize == 12) { 848 /* p[1] is a minor version. */ 849 lha->mtime = archive_le32dec(&p[2]); 850 lha->mode = archive_le16dec(&p[6]); 851 lha->uid = archive_le16dec(&p[8]); 852 lha->gid = archive_le16dec(&p[10]); 853 lha->setflag |= UNIX_MODE_IS_SET; 854 } 855 } 856 __archive_read_consume(a, lha->header_size); 857 858 if (sum_calculated != headersum) { 859 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 860 "LHa header sum error"); 861 return (ARCHIVE_FATAL); 862 } 863 864 return (ARCHIVE_OK); 865} 866 867/* 868 * Header 1 format 869 * 870 * +0 +1 +2 +7 +11 871 * +---------------+----------+----------------+-------------+ 872 * |header size(*1)|header sum|compression type|skip size(*2)| 873 * +---------------+----------+----------------+-------------+ 874 * <---------------(*1)----------* 875 * 876 * +11 +15 +17 +19 +20 +21 877 * +-----------------+---------+---------+--------------+----------------+ 878 * |uncompressed size|time(DOS)|date(DOS)|attribute(DOS)|header level(=1)| 879 * +-----------------+---------+---------+--------------+----------------+ 880 * *-------------------------------(*1)----------------------------------* 881 * 882 * +21 +22 +22+(*3) +22+(*3)+2 +22+(*3)+3 +22+(*3)+3+(*4) 883 * +---------------+---------+----------+-----------+-----------+ 884 * |name length(*3)|file name|file CRC16| creator |padding(*4)| 885 * +---------------+---------+----------+-----------+-----------+ 886 * <--(*3)-> 887 * *----------------------------(*1)----------------------------* 888 * 889 * +22+(*3)+3+(*4) +22+(*3)+3+(*4)+2 +22+(*3)+3+(*4)+2+(*5) 890 * +----------------+---------------------+------------------------+ 891 * |next header size| extended header(*5) | compressed data | 892 * +----------------+---------------------+------------------------+ 893 * *------(*1)-----> <--------------------(*2)--------------------> 894 */ 895#define H1_HEADER_SIZE_OFFSET 0 896#define H1_HEADER_SUM_OFFSET 1 897#define H1_COMP_SIZE_OFFSET 7 898#define H1_ORIG_SIZE_OFFSET 11 899#define H1_DOS_TIME_OFFSET 15 900#define H1_NAME_LEN_OFFSET 21 901#define H1_FILE_NAME_OFFSET 22 902#define H1_FIXED_SIZE 27 903static int 904lha_read_file_header_1(struct archive_read *a, struct lha *lha) 905{ 906 const unsigned char *p; 907 size_t extdsize; 908 int i, err, err2; 909 int namelen, padding; 910 unsigned char headersum, sum_calculated; 911 912 err = ARCHIVE_OK; 913 914 if ((p = __archive_read_ahead(a, H1_FIXED_SIZE, NULL)) == NULL) 915 return (truncated_error(a)); 916 917 lha->header_size = p[H1_HEADER_SIZE_OFFSET] + 2; 918 headersum = p[H1_HEADER_SUM_OFFSET]; 919 /* Note: An extended header size is included in a compsize. */ 920 lha->compsize = archive_le32dec(p + H1_COMP_SIZE_OFFSET); 921 lha->origsize = archive_le32dec(p + H1_ORIG_SIZE_OFFSET); 922 lha->mtime = lha_dos_time(p + H1_DOS_TIME_OFFSET); 923 namelen = p[H1_NAME_LEN_OFFSET]; 924 /* Calculate a padding size. The result will be normally 0 only(?) */ 925 padding = ((int)lha->header_size) - H1_FIXED_SIZE - namelen; 926 927 if (namelen > 230 || padding < 0) 928 goto invalid; 929 930 if ((p = __archive_read_ahead(a, lha->header_size, NULL)) == NULL) 931 return (truncated_error(a)); 932 933 for (i = 0; i < namelen; i++) { 934 if (p[i + H1_FILE_NAME_OFFSET] == 0xff) 935 goto invalid;/* Invalid filename. */ 936 } 937 archive_strncpy(&lha->filename, p + H1_FILE_NAME_OFFSET, namelen); 938 lha->crc = archive_le16dec(p + H1_FILE_NAME_OFFSET + namelen); 939 lha->setflag |= CRC_IS_SET; 940 941 sum_calculated = lha_calcsum(0, p, 2, lha->header_size - 2); 942 /* Consume used bytes but not include `next header size' data 943 * since it will be consumed in lha_read_file_extended_header(). */ 944 __archive_read_consume(a, lha->header_size - 2); 945 946 /* Read extended headers */ 947 err2 = lha_read_file_extended_header(a, lha, NULL, 2, 948 (size_t)(lha->compsize + 2), &extdsize); 949 if (err2 < ARCHIVE_WARN) 950 return (err2); 951 if (err2 < err) 952 err = err2; 953 /* Get a real compressed file size. */ 954 lha->compsize -= extdsize - 2; 955 956 if (lha->compsize < 0) 957 goto invalid; /* Invalid compressed file size */ 958 959 if (sum_calculated != headersum) { 960 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 961 "LHa header sum error"); 962 return (ARCHIVE_FATAL); 963 } 964 return (err); 965invalid: 966 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 967 "Invalid LHa header"); 968 return (ARCHIVE_FATAL); 969} 970 971/* 972 * Header 2 format 973 * 974 * +0 +2 +7 +11 +15 975 * +---------------+----------------+-------------------+-----------------+ 976 * |header size(*1)|compression type|compressed size(*2)|uncompressed size| 977 * +---------------+----------------+-------------------+-----------------+ 978 * <--------------------------------(*1)---------------------------------* 979 * 980 * +15 +19 +20 +21 +23 +24 981 * +-----------------+------------+----------------+----------+-----------+ 982 * |data/time(time_t)| 0x20 fixed |header level(=2)|file CRC16| creator | 983 * +-----------------+------------+----------------+----------+-----------+ 984 * *---------------------------------(*1)---------------------------------* 985 * 986 * +24 +26 +26+(*3) +26+(*3)+(*4) 987 * +----------------+-------------------+-------------+-------------------+ 988 * |next header size|extended header(*3)| padding(*4) | compressed data | 989 * +----------------+-------------------+-------------+-------------------+ 990 * *--------------------------(*1)-------------------> <------(*2)-------> 991 * 992 */ 993#define H2_HEADER_SIZE_OFFSET 0 994#define H2_COMP_SIZE_OFFSET 7 995#define H2_ORIG_SIZE_OFFSET 11 996#define H2_TIME_OFFSET 15 997#define H2_CRC_OFFSET 21 998#define H2_FIXED_SIZE 24 999static int 1000lha_read_file_header_2(struct archive_read *a, struct lha *lha) 1001{ 1002 const unsigned char *p; 1003 size_t extdsize; 1004 int err, padding; 1005 uint16_t header_crc; 1006 1007 if ((p = __archive_read_ahead(a, H2_FIXED_SIZE, NULL)) == NULL) 1008 return (truncated_error(a)); 1009 1010 lha->header_size =archive_le16dec(p + H2_HEADER_SIZE_OFFSET); 1011 lha->compsize = archive_le32dec(p + H2_COMP_SIZE_OFFSET); 1012 lha->origsize = archive_le32dec(p + H2_ORIG_SIZE_OFFSET); 1013 lha->mtime = archive_le32dec(p + H2_TIME_OFFSET); 1014 lha->crc = archive_le16dec(p + H2_CRC_OFFSET); 1015 lha->setflag |= CRC_IS_SET; 1016 1017 if (lha->header_size < H2_FIXED_SIZE) { 1018 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1019 "Invalid LHa header size"); 1020 return (ARCHIVE_FATAL); 1021 } 1022 1023 header_crc = lha_crc16(0, p, H2_FIXED_SIZE); 1024 __archive_read_consume(a, H2_FIXED_SIZE); 1025 1026 /* Read extended headers */ 1027 err = lha_read_file_extended_header(a, lha, &header_crc, 2, 1028 lha->header_size - H2_FIXED_SIZE, &extdsize); 1029 if (err < ARCHIVE_WARN) 1030 return (err); 1031 1032 /* Calculate a padding size. The result will be normally 0 or 1. */ 1033 padding = (int)lha->header_size - (int)(H2_FIXED_SIZE + extdsize); 1034 if (padding > 0) { 1035 if ((p = __archive_read_ahead(a, padding, NULL)) == NULL) 1036 return (truncated_error(a)); 1037 header_crc = lha_crc16(header_crc, p, padding); 1038 __archive_read_consume(a, padding); 1039 } 1040 1041 if (header_crc != lha->header_crc) { 1042 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1043 "LHa header CRC error"); 1044 return (ARCHIVE_FATAL); 1045 } 1046 return (err); 1047} 1048 1049/* 1050 * Header 3 format 1051 * 1052 * +0 +2 +7 +11 +15 1053 * +------------+----------------+-------------------+-----------------+ 1054 * | 0x04 fixed |compression type|compressed size(*2)|uncompressed size| 1055 * +------------+----------------+-------------------+-----------------+ 1056 * <-------------------------------(*1)-------------------------------* 1057 * 1058 * +15 +19 +20 +21 +23 +24 1059 * +-----------------+------------+----------------+----------+-----------+ 1060 * |date/time(time_t)| 0x20 fixed |header level(=3)|file CRC16| creator | 1061 * +-----------------+------------+----------------+----------+-----------+ 1062 * *--------------------------------(*1)----------------------------------* 1063 * 1064 * +24 +28 +32 +32+(*3) 1065 * +---------------+----------------+-------------------+-----------------+ 1066 * |header size(*1)|next header size|extended header(*3)| compressed data | 1067 * +---------------+----------------+-------------------+-----------------+ 1068 * *------------------------(*1)-----------------------> <------(*2)-----> 1069 * 1070 */ 1071#define H3_FIELD_LEN_OFFSET 0 1072#define H3_COMP_SIZE_OFFSET 7 1073#define H3_ORIG_SIZE_OFFSET 11 1074#define H3_TIME_OFFSET 15 1075#define H3_CRC_OFFSET 21 1076#define H3_HEADER_SIZE_OFFSET 24 1077#define H3_FIXED_SIZE 28 1078static int 1079lha_read_file_header_3(struct archive_read *a, struct lha *lha) 1080{ 1081 const unsigned char *p; 1082 size_t extdsize; 1083 int err; 1084 uint16_t header_crc; 1085 1086 if ((p = __archive_read_ahead(a, H3_FIXED_SIZE, NULL)) == NULL) 1087 return (truncated_error(a)); 1088 1089 if (archive_le16dec(p + H3_FIELD_LEN_OFFSET) != 4) 1090 goto invalid; 1091 lha->header_size =archive_le32dec(p + H3_HEADER_SIZE_OFFSET); 1092 lha->compsize = archive_le32dec(p + H3_COMP_SIZE_OFFSET); 1093 lha->origsize = archive_le32dec(p + H3_ORIG_SIZE_OFFSET); 1094 lha->mtime = archive_le32dec(p + H3_TIME_OFFSET); 1095 lha->crc = archive_le16dec(p + H3_CRC_OFFSET); 1096 lha->setflag |= CRC_IS_SET; 1097 1098 if (lha->header_size < H3_FIXED_SIZE + 4) 1099 goto invalid; 1100 header_crc = lha_crc16(0, p, H3_FIXED_SIZE); 1101 __archive_read_consume(a, H3_FIXED_SIZE); 1102 1103 /* Read extended headers */ 1104 err = lha_read_file_extended_header(a, lha, &header_crc, 4, 1105 lha->header_size - H3_FIXED_SIZE, &extdsize); 1106 if (err < ARCHIVE_WARN) 1107 return (err); 1108 1109 if (header_crc != lha->header_crc) { 1110 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1111 "LHa header CRC error"); 1112 return (ARCHIVE_FATAL); 1113 } 1114 return (err); 1115invalid: 1116 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1117 "Invalid LHa header"); 1118 return (ARCHIVE_FATAL); 1119} 1120 1121/* 1122 * Extended header format 1123 * 1124 * +0 +2 +3 -- used in header 1 and 2 1125 * +0 +4 +5 -- used in header 3 1126 * +--------------+---------+-------------------+--------------+-- 1127 * |ex-header size|header id| data |ex-header size| ....... 1128 * +--------------+---------+-------------------+--------------+-- 1129 * <-------------( ex-header size)------------> <-- next extended header --* 1130 * 1131 * If the ex-header size is zero, it is the make of the end of extended 1132 * headers. 1133 * 1134 */ 1135static int 1136lha_read_file_extended_header(struct archive_read *a, struct lha *lha, 1137 uint16_t *crc, int sizefield_length, size_t limitsize, size_t *total_size) 1138{ 1139 const void *h; 1140 const unsigned char *extdheader; 1141 size_t extdsize; 1142 size_t datasize; 1143 unsigned int i; 1144 unsigned char extdtype; 1145 1146#define EXT_HEADER_CRC 0x00 /* Header CRC and information*/ 1147#define EXT_FILENAME 0x01 /* Filename */ 1148#define EXT_DIRECTORY 0x02 /* Directory name */ 1149#define EXT_DOS_ATTR 0x40 /* MS-DOS attribute */ 1150#define EXT_TIMESTAMP 0x41 /* Windows time stamp */ 1151#define EXT_FILESIZE 0x42 /* Large file size */ 1152#define EXT_TIMEZONE 0x43 /* Time zone */ 1153#define EXT_UTF16_FILENAME 0x44 /* UTF-16 filename */ 1154#define EXT_UTF16_DIRECTORY 0x45 /* UTF-16 directory name */ 1155#define EXT_CODEPAGE 0x46 /* Codepage */ 1156#define EXT_UNIX_MODE 0x50 /* File permission */ 1157#define EXT_UNIX_GID_UID 0x51 /* gid,uid */ 1158#define EXT_UNIX_GNAME 0x52 /* Group name */ 1159#define EXT_UNIX_UNAME 0x53 /* User name */ 1160#define EXT_UNIX_MTIME 0x54 /* Modified time */ 1161#define EXT_OS2_NEW_ATTR 0x7f /* new attribute(OS/2 only) */ 1162#define EXT_NEW_ATTR 0xff /* new attribute */ 1163 1164 *total_size = sizefield_length; 1165 1166 for (;;) { 1167 /* Read an extended header size. */ 1168 if ((h = 1169 __archive_read_ahead(a, sizefield_length, NULL)) == NULL) 1170 return (truncated_error(a)); 1171 /* Check if the size is the zero indicates the end of the 1172 * extended header. */ 1173 if (sizefield_length == sizeof(uint16_t)) 1174 extdsize = archive_le16dec(h); 1175 else 1176 extdsize = archive_le32dec(h); 1177 if (extdsize == 0) { 1178 /* End of extended header */ 1179 if (crc != NULL) 1180 *crc = lha_crc16(*crc, h, sizefield_length); 1181 __archive_read_consume(a, sizefield_length); 1182 return (ARCHIVE_OK); 1183 } 1184 1185 /* Sanity check to the extended header size. */ 1186 if (((uint64_t)*total_size + extdsize) > 1187 (uint64_t)limitsize || 1188 extdsize <= (size_t)sizefield_length) 1189 goto invalid; 1190 1191 /* Read the extended header. */ 1192 if ((h = __archive_read_ahead(a, extdsize, NULL)) == NULL) 1193 return (truncated_error(a)); 1194 *total_size += extdsize; 1195 1196 extdheader = (const unsigned char *)h; 1197 /* Get the extended header type. */ 1198 extdtype = extdheader[sizefield_length]; 1199 /* Calculate an extended data size. */ 1200 datasize = extdsize - (1 + sizefield_length); 1201 /* Skip an extended header size field and type field. */ 1202 extdheader += sizefield_length + 1; 1203 1204 if (crc != NULL && extdtype != EXT_HEADER_CRC) 1205 *crc = lha_crc16(*crc, h, extdsize); 1206 switch (extdtype) { 1207 case EXT_HEADER_CRC: 1208 /* We only use a header CRC. Following data will not 1209 * be used. */ 1210 if (datasize >= 2) { 1211 lha->header_crc = archive_le16dec(extdheader); 1212 if (crc != NULL) { 1213 static const char zeros[2] = {0, 0}; 1214 *crc = lha_crc16(*crc, h, 1215 extdsize - datasize); 1216 /* CRC value itself as zero */ 1217 *crc = lha_crc16(*crc, zeros, 2); 1218 *crc = lha_crc16(*crc, 1219 extdheader+2, datasize - 2); 1220 } 1221 } 1222 break; 1223 case EXT_FILENAME: 1224 if (datasize == 0) { 1225 /* maybe directory header */ 1226 archive_string_empty(&lha->filename); 1227 break; 1228 } 1229 if (extdheader[0] == '\0') 1230 goto invalid; 1231 archive_strncpy(&lha->filename, 1232 (const char *)extdheader, datasize); 1233 break; 1234 case EXT_UTF16_FILENAME: 1235 if (datasize == 0) { 1236 /* maybe directory header */ 1237 archive_string_empty(&lha->filename); 1238 break; 1239 } else if (datasize & 1) { 1240 /* UTF-16 characters take always 2 or 4 bytes */ 1241 goto invalid; 1242 } 1243 if (extdheader[0] == '\0') 1244 goto invalid; 1245 archive_string_empty(&lha->filename); 1246 archive_array_append(&lha->filename, 1247 (const char *)extdheader, datasize); 1248 /* Setup a string conversion for a filename. */ 1249 lha->sconv_fname = 1250 archive_string_conversion_from_charset(&a->archive, 1251 "UTF-16LE", 1); 1252 if (lha->sconv_fname == NULL) 1253 return (ARCHIVE_FATAL); 1254 break; 1255 case EXT_DIRECTORY: 1256 if (datasize == 0 || extdheader[0] == '\0') 1257 /* no directory name data. exit this case. */ 1258 goto invalid; 1259 1260 archive_strncpy(&lha->dirname, 1261 (const char *)extdheader, datasize); 1262 /* 1263 * Convert directory delimiter from 0xFF 1264 * to '/' for local system. 1265 */ 1266 for (i = 0; i < lha->dirname.length; i++) { 1267 if ((unsigned char)lha->dirname.s[i] == 0xFF) 1268 lha->dirname.s[i] = '/'; 1269 } 1270 /* Is last character directory separator? */ 1271 if (lha->dirname.s[lha->dirname.length-1] != '/') 1272 /* invalid directory data */ 1273 goto invalid; 1274 break; 1275 case EXT_UTF16_DIRECTORY: 1276 /* UTF-16 characters take always 2 or 4 bytes */ 1277 if (datasize == 0 || (datasize & 1) || 1278 extdheader[0] == '\0') { 1279 /* no directory name data. exit this case. */ 1280 goto invalid; 1281 } 1282 1283 archive_string_empty(&lha->dirname); 1284 archive_array_append(&lha->dirname, 1285 (const char *)extdheader, datasize); 1286 lha->sconv_dir = 1287 archive_string_conversion_from_charset(&a->archive, 1288 "UTF-16LE", 1); 1289 if (lha->sconv_dir == NULL) 1290 return (ARCHIVE_FATAL); 1291 else { 1292 /* 1293 * Convert directory delimiter from 0xFFFF 1294 * to '/' for local system. 1295 */ 1296 uint16_t dirSep; 1297 uint16_t d = 1; 1298 if (archive_be16dec(&d) == 1) 1299 dirSep = 0x2F00; 1300 else 1301 dirSep = 0x002F; 1302 1303 /* UTF-16LE character */ 1304 uint16_t *utf16name = 1305 (uint16_t *)lha->dirname.s; 1306 for (i = 0; i < lha->dirname.length / 2; i++) { 1307 if (utf16name[i] == 0xFFFF) { 1308 utf16name[i] = dirSep; 1309 } 1310 } 1311 /* Is last character directory separator? */ 1312 if (utf16name[lha->dirname.length / 2 - 1] != 1313 dirSep) { 1314 /* invalid directory data */ 1315 goto invalid; 1316 } 1317 } 1318 break; 1319 case EXT_DOS_ATTR: 1320 if (datasize == 2) 1321 lha->dos_attr = (unsigned char) 1322 (archive_le16dec(extdheader) & 0xff); 1323 break; 1324 case EXT_TIMESTAMP: 1325 if (datasize == (sizeof(uint64_t) * 3)) { 1326 lha->birthtime = lha_win_time( 1327 archive_le64dec(extdheader), 1328 &lha->birthtime_tv_nsec); 1329 extdheader += sizeof(uint64_t); 1330 lha->mtime = lha_win_time( 1331 archive_le64dec(extdheader), 1332 &lha->mtime_tv_nsec); 1333 extdheader += sizeof(uint64_t); 1334 lha->atime = lha_win_time( 1335 archive_le64dec(extdheader), 1336 &lha->atime_tv_nsec); 1337 lha->setflag |= BIRTHTIME_IS_SET | 1338 ATIME_IS_SET; 1339 } 1340 break; 1341 case EXT_FILESIZE: 1342 if (datasize == sizeof(uint64_t) * 2) { 1343 lha->compsize = archive_le64dec(extdheader); 1344 extdheader += sizeof(uint64_t); 1345 lha->origsize = archive_le64dec(extdheader); 1346 } 1347 break; 1348 case EXT_CODEPAGE: 1349 /* Get an archived filename charset from codepage. 1350 * This overwrites the charset specified by 1351 * hdrcharset option. */ 1352 if (datasize == sizeof(uint32_t)) { 1353 struct archive_string cp; 1354 const char *charset; 1355 1356 archive_string_init(&cp); 1357 switch (archive_le32dec(extdheader)) { 1358 case 65001: /* UTF-8 */ 1359 charset = "UTF-8"; 1360 break; 1361 default: 1362 archive_string_sprintf(&cp, "CP%d", 1363 (int)archive_le32dec(extdheader)); 1364 charset = cp.s; 1365 break; 1366 } 1367 lha->sconv_dir = 1368 archive_string_conversion_from_charset( 1369 &(a->archive), charset, 1); 1370 lha->sconv_fname = 1371 archive_string_conversion_from_charset( 1372 &(a->archive), charset, 1); 1373 archive_string_free(&cp); 1374 if (lha->sconv_dir == NULL) 1375 return (ARCHIVE_FATAL); 1376 if (lha->sconv_fname == NULL) 1377 return (ARCHIVE_FATAL); 1378 } 1379 break; 1380 case EXT_UNIX_MODE: 1381 if (datasize == sizeof(uint16_t)) { 1382 lha->mode = archive_le16dec(extdheader); 1383 lha->setflag |= UNIX_MODE_IS_SET; 1384 } 1385 break; 1386 case EXT_UNIX_GID_UID: 1387 if (datasize == (sizeof(uint16_t) * 2)) { 1388 lha->gid = archive_le16dec(extdheader); 1389 lha->uid = archive_le16dec(extdheader+2); 1390 } 1391 break; 1392 case EXT_UNIX_GNAME: 1393 if (datasize > 0) 1394 archive_strncpy(&lha->gname, 1395 (const char *)extdheader, datasize); 1396 break; 1397 case EXT_UNIX_UNAME: 1398 if (datasize > 0) 1399 archive_strncpy(&lha->uname, 1400 (const char *)extdheader, datasize); 1401 break; 1402 case EXT_UNIX_MTIME: 1403 if (datasize == sizeof(uint32_t)) 1404 lha->mtime = archive_le32dec(extdheader); 1405 break; 1406 case EXT_OS2_NEW_ATTR: 1407 /* This extended header is OS/2 depend. */ 1408 if (datasize == 16) { 1409 lha->dos_attr = (unsigned char) 1410 (archive_le16dec(extdheader) & 0xff); 1411 lha->mode = archive_le16dec(extdheader+2); 1412 lha->gid = archive_le16dec(extdheader+4); 1413 lha->uid = archive_le16dec(extdheader+6); 1414 lha->birthtime = archive_le32dec(extdheader+8); 1415 lha->atime = archive_le32dec(extdheader+12); 1416 lha->setflag |= UNIX_MODE_IS_SET 1417 | BIRTHTIME_IS_SET | ATIME_IS_SET; 1418 } 1419 break; 1420 case EXT_NEW_ATTR: 1421 if (datasize == 20) { 1422 lha->mode = (mode_t)archive_le32dec(extdheader); 1423 lha->gid = archive_le32dec(extdheader+4); 1424 lha->uid = archive_le32dec(extdheader+8); 1425 lha->birthtime = archive_le32dec(extdheader+12); 1426 lha->atime = archive_le32dec(extdheader+16); 1427 lha->setflag |= UNIX_MODE_IS_SET 1428 | BIRTHTIME_IS_SET | ATIME_IS_SET; 1429 } 1430 break; 1431 case EXT_TIMEZONE: /* Not supported */ 1432 break; 1433 default: 1434 break; 1435 } 1436 1437 __archive_read_consume(a, extdsize); 1438 } 1439invalid: 1440 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1441 "Invalid extended LHa header"); 1442 return (ARCHIVE_FATAL); 1443} 1444 1445static int 1446lha_end_of_entry(struct archive_read *a) 1447{ 1448 struct lha *lha = (struct lha *)(a->format->data); 1449 int r = ARCHIVE_EOF; 1450 1451 if (!lha->end_of_entry_cleanup) { 1452 if ((lha->setflag & CRC_IS_SET) && 1453 lha->crc != lha->entry_crc_calculated) { 1454 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1455 "LHa data CRC error"); 1456 r = ARCHIVE_WARN; 1457 } 1458 1459 /* End-of-entry cleanup done. */ 1460 lha->end_of_entry_cleanup = 1; 1461 } 1462 return (r); 1463} 1464 1465static int 1466archive_read_format_lha_read_data(struct archive_read *a, 1467 const void **buff, size_t *size, int64_t *offset) 1468{ 1469 struct lha *lha = (struct lha *)(a->format->data); 1470 int r; 1471 1472 if (lha->entry_unconsumed) { 1473 /* Consume as much as the decompressor actually used. */ 1474 __archive_read_consume(a, lha->entry_unconsumed); 1475 lha->entry_unconsumed = 0; 1476 } 1477 if (lha->end_of_entry) { 1478 *offset = lha->entry_offset; 1479 *size = 0; 1480 *buff = NULL; 1481 return (lha_end_of_entry(a)); 1482 } 1483 1484 if (lha->entry_is_compressed) 1485 r = lha_read_data_lzh(a, buff, size, offset); 1486 else 1487 /* No compression. */ 1488 r = lha_read_data_none(a, buff, size, offset); 1489 return (r); 1490} 1491 1492/* 1493 * Read a file content in no compression. 1494 * 1495 * Returns ARCHIVE_OK if successful, ARCHIVE_FATAL otherwise, sets 1496 * lha->end_of_entry if it consumes all of the data. 1497 */ 1498static int 1499lha_read_data_none(struct archive_read *a, const void **buff, 1500 size_t *size, int64_t *offset) 1501{ 1502 struct lha *lha = (struct lha *)(a->format->data); 1503 ssize_t bytes_avail; 1504 1505 if (lha->entry_bytes_remaining == 0) { 1506 *buff = NULL; 1507 *size = 0; 1508 *offset = lha->entry_offset; 1509 lha->end_of_entry = 1; 1510 return (ARCHIVE_OK); 1511 } 1512 /* 1513 * Note: '1' here is a performance optimization. 1514 * Recall that the decompression layer returns a count of 1515 * available bytes; asking for more than that forces the 1516 * decompressor to combine reads by copying data. 1517 */ 1518 *buff = __archive_read_ahead(a, 1, &bytes_avail); 1519 if (bytes_avail <= 0) { 1520 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1521 "Truncated LHa file data"); 1522 return (ARCHIVE_FATAL); 1523 } 1524 if (bytes_avail > lha->entry_bytes_remaining) 1525 bytes_avail = (ssize_t)lha->entry_bytes_remaining; 1526 lha->entry_crc_calculated = 1527 lha_crc16(lha->entry_crc_calculated, *buff, bytes_avail); 1528 *size = bytes_avail; 1529 *offset = lha->entry_offset; 1530 lha->entry_offset += bytes_avail; 1531 lha->entry_bytes_remaining -= bytes_avail; 1532 if (lha->entry_bytes_remaining == 0) 1533 lha->end_of_entry = 1; 1534 lha->entry_unconsumed = bytes_avail; 1535 return (ARCHIVE_OK); 1536} 1537 1538/* 1539 * Read a file content in LZHUFF encoding. 1540 * 1541 * Returns ARCHIVE_OK if successful, returns ARCHIVE_WARN if compression is 1542 * unsupported, ARCHIVE_FATAL otherwise, sets lha->end_of_entry if it consumes 1543 * all of the data. 1544 */ 1545static int 1546lha_read_data_lzh(struct archive_read *a, const void **buff, 1547 size_t *size, int64_t *offset) 1548{ 1549 struct lha *lha = (struct lha *)(a->format->data); 1550 ssize_t bytes_avail; 1551 int r; 1552 1553 /* If we haven't yet read any data, initialize the decompressor. */ 1554 if (!lha->decompress_init) { 1555 r = lzh_decode_init(&(lha->strm), lha->method); 1556 switch (r) { 1557 case ARCHIVE_OK: 1558 break; 1559 case ARCHIVE_FAILED: 1560 /* Unsupported compression. */ 1561 *buff = NULL; 1562 *size = 0; 1563 *offset = 0; 1564 archive_set_error(&a->archive, 1565 ARCHIVE_ERRNO_FILE_FORMAT, 1566 "Unsupported lzh compression method -%c%c%c-", 1567 lha->method[0], lha->method[1], lha->method[2]); 1568 /* We know compressed size; just skip it. */ 1569 archive_read_format_lha_read_data_skip(a); 1570 return (ARCHIVE_WARN); 1571 default: 1572 archive_set_error(&a->archive, ENOMEM, 1573 "Couldn't allocate memory " 1574 "for lzh decompression"); 1575 return (ARCHIVE_FATAL); 1576 } 1577 /* We've initialized decompression for this stream. */ 1578 lha->decompress_init = 1; 1579 lha->strm.avail_out = 0; 1580 lha->strm.total_out = 0; 1581 } 1582 1583 /* 1584 * Note: '1' here is a performance optimization. 1585 * Recall that the decompression layer returns a count of 1586 * available bytes; asking for more than that forces the 1587 * decompressor to combine reads by copying data. 1588 */ 1589 lha->strm.next_in = __archive_read_ahead(a, 1, &bytes_avail); 1590 if (bytes_avail <= 0) { 1591 archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, 1592 "Truncated LHa file body"); 1593 return (ARCHIVE_FATAL); 1594 } 1595 if (bytes_avail > lha->entry_bytes_remaining) 1596 bytes_avail = (ssize_t)lha->entry_bytes_remaining; 1597 1598 lha->strm.avail_in = (int)bytes_avail; 1599 lha->strm.total_in = 0; 1600 lha->strm.avail_out = 0; 1601 1602 r = lzh_decode(&(lha->strm), bytes_avail == lha->entry_bytes_remaining); 1603 switch (r) { 1604 case ARCHIVE_OK: 1605 break; 1606 case ARCHIVE_EOF: 1607 lha->end_of_entry = 1; 1608 break; 1609 default: 1610 archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, 1611 "Bad lzh data"); 1612 return (ARCHIVE_FAILED); 1613 } 1614 lha->entry_unconsumed = lha->strm.total_in; 1615 lha->entry_bytes_remaining -= lha->strm.total_in; 1616 1617 if (lha->strm.avail_out) { 1618 *offset = lha->entry_offset; 1619 *size = lha->strm.avail_out; 1620 *buff = lha->strm.ref_ptr; 1621 lha->entry_crc_calculated = 1622 lha_crc16(lha->entry_crc_calculated, *buff, *size); 1623 lha->entry_offset += *size; 1624 } else { 1625 *offset = lha->entry_offset; 1626 *size = 0; 1627 *buff = NULL; 1628 if (lha->end_of_entry) 1629 return (lha_end_of_entry(a)); 1630 } 1631 return (ARCHIVE_OK); 1632} 1633 1634/* 1635 * Skip a file content. 1636 */ 1637static int 1638archive_read_format_lha_read_data_skip(struct archive_read *a) 1639{ 1640 struct lha *lha; 1641 int64_t bytes_skipped; 1642 1643 lha = (struct lha *)(a->format->data); 1644 1645 if (lha->entry_unconsumed) { 1646 /* Consume as much as the decompressor actually used. */ 1647 __archive_read_consume(a, lha->entry_unconsumed); 1648 lha->entry_unconsumed = 0; 1649 } 1650 1651 /* if we've already read to end of data, we're done. */ 1652 if (lha->end_of_entry_cleanup) 1653 return (ARCHIVE_OK); 1654 1655 /* 1656 * If the length is at the beginning, we can skip the 1657 * compressed data much more quickly. 1658 */ 1659 bytes_skipped = __archive_read_consume(a, lha->entry_bytes_remaining); 1660 if (bytes_skipped < 0) 1661 return (ARCHIVE_FATAL); 1662 1663 /* This entry is finished and done. */ 1664 lha->end_of_entry_cleanup = lha->end_of_entry = 1; 1665 return (ARCHIVE_OK); 1666} 1667 1668static int 1669archive_read_format_lha_cleanup(struct archive_read *a) 1670{ 1671 struct lha *lha = (struct lha *)(a->format->data); 1672 1673 lzh_decode_free(&(lha->strm)); 1674 archive_string_free(&(lha->dirname)); 1675 archive_string_free(&(lha->filename)); 1676 archive_string_free(&(lha->uname)); 1677 archive_string_free(&(lha->gname)); 1678 archive_wstring_free(&(lha->ws)); 1679 free(lha); 1680 (a->format->data) = NULL; 1681 return (ARCHIVE_OK); 1682} 1683 1684/* 1685 * 'LHa for UNIX' utility has archived a symbolic-link name after 1686 * a pathname with '|' character. 1687 * This function extracts the symbolic-link name from the pathname. 1688 * 1689 * example. 1690 * 1. a symbolic-name is 'aaa/bb/cc' 1691 * 2. a filename is 'xxx/bbb' 1692 * then a archived pathname is 'xxx/bbb|aaa/bb/cc' 1693 */ 1694static int 1695lha_parse_linkname(struct archive_wstring *linkname, 1696 struct archive_wstring *pathname) 1697{ 1698 wchar_t * linkptr; 1699 size_t symlen; 1700 1701 linkptr = wcschr(pathname->s, L'|'); 1702 if (linkptr != NULL) { 1703 symlen = wcslen(linkptr + 1); 1704 archive_wstrncpy(linkname, linkptr+1, symlen); 1705 1706 *linkptr = 0; 1707 pathname->length = wcslen(pathname->s); 1708 1709 return (1); 1710 } 1711 return (0); 1712} 1713 1714/* Convert an MSDOS-style date/time into Unix-style time. */ 1715static time_t 1716lha_dos_time(const unsigned char *p) 1717{ 1718 int msTime, msDate; 1719 struct tm ts; 1720 1721 msTime = archive_le16dec(p); 1722 msDate = archive_le16dec(p+2); 1723 1724 memset(&ts, 0, sizeof(ts)); 1725 ts.tm_year = ((msDate >> 9) & 0x7f) + 80; /* Years since 1900. */ 1726 ts.tm_mon = ((msDate >> 5) & 0x0f) - 1; /* Month number. */ 1727 ts.tm_mday = msDate & 0x1f; /* Day of month. */ 1728 ts.tm_hour = (msTime >> 11) & 0x1f; 1729 ts.tm_min = (msTime >> 5) & 0x3f; 1730 ts.tm_sec = (msTime << 1) & 0x3e; 1731 ts.tm_isdst = -1; 1732 return (mktime(&ts)); 1733} 1734 1735/* Convert an MS-Windows-style date/time into Unix-style time. */ 1736static time_t 1737lha_win_time(uint64_t wintime, long *ns) 1738{ 1739#define EPOC_TIME ARCHIVE_LITERAL_ULL(116444736000000000) 1740 1741 if (wintime >= EPOC_TIME) { 1742 wintime -= EPOC_TIME; /* 1970-01-01 00:00:00 (UTC) */ 1743 if (ns != NULL) 1744 *ns = (long)(wintime % 10000000) * 100; 1745 return (wintime / 10000000); 1746 } else { 1747 if (ns != NULL) 1748 *ns = 0; 1749 return (0); 1750 } 1751} 1752 1753static unsigned char 1754lha_calcsum(unsigned char sum, const void *pp, int offset, size_t size) 1755{ 1756 unsigned char const *p = (unsigned char const *)pp; 1757 1758 p += offset; 1759 for (;size > 0; --size) 1760 sum += *p++; 1761 return (sum); 1762} 1763 1764static uint16_t crc16tbl[2][256]; 1765static void 1766lha_crc16_init(void) 1767{ 1768 unsigned int i; 1769 static int crc16init = 0; 1770 1771 if (crc16init) 1772 return; 1773 crc16init = 1; 1774 1775 for (i = 0; i < 256; i++) { 1776 unsigned int j; 1777 uint16_t crc = (uint16_t)i; 1778 for (j = 8; j; j--) 1779 crc = (crc >> 1) ^ ((crc & 1) * 0xA001); 1780 crc16tbl[0][i] = crc; 1781 } 1782 1783 for (i = 0; i < 256; i++) { 1784 crc16tbl[1][i] = (crc16tbl[0][i] >> 8) 1785 ^ crc16tbl[0][crc16tbl[0][i] & 0xff]; 1786 } 1787} 1788 1789static uint16_t 1790lha_crc16(uint16_t crc, const void *pp, size_t len) 1791{ 1792 const unsigned char *p = (const unsigned char *)pp; 1793 const uint16_t *buff; 1794 const union { 1795 uint32_t i; 1796 char c[4]; 1797 } u = { 0x01020304 }; 1798 1799 if (len == 0) 1800 return crc; 1801 1802 /* Process unaligned address. */ 1803 if (((uintptr_t)p) & (uintptr_t)0x1) { 1804 crc = (crc >> 8) ^ crc16tbl[0][(crc ^ *p++) & 0xff]; 1805 len--; 1806 } 1807 buff = (const uint16_t *)p; 1808 /* 1809 * Modern C compiler such as GCC does not unroll automatically yet 1810 * without unrolling pragma, and Clang is so. So we should 1811 * unroll this loop for its performance. 1812 */ 1813 for (;len >= 8; len -= 8) { 1814 /* This if statement expects compiler optimization will 1815 * remove the statement which will not be executed. */ 1816#undef bswap16 1817#if defined(_MSC_VER) && _MSC_VER >= 1400 /* Visual Studio */ 1818# define bswap16(x) _byteswap_ushort(x) 1819#elif defined(__GNUC__) && ((__GNUC__ == 4 && __GNUC_MINOR__ >= 8) || __GNUC__ > 4) 1820/* GCC 4.8 and later has __builtin_bswap16() */ 1821# define bswap16(x) __builtin_bswap16(x) 1822#elif defined(__clang__) 1823/* All clang versions have __builtin_bswap16() */ 1824# define bswap16(x) __builtin_bswap16(x) 1825#else 1826# define bswap16(x) ((((x) >> 8) & 0xff) | ((x) << 8)) 1827#endif 1828#define CRC16W do { \ 1829 if(u.c[0] == 1) { /* Big endian */ \ 1830 crc ^= bswap16(*buff); buff++; \ 1831 } else \ 1832 crc ^= *buff++; \ 1833 crc = crc16tbl[1][crc & 0xff] ^ crc16tbl[0][crc >> 8];\ 1834} while (0) 1835 CRC16W; 1836 CRC16W; 1837 CRC16W; 1838 CRC16W; 1839#undef CRC16W 1840#undef bswap16 1841 } 1842 1843 p = (const unsigned char *)buff; 1844 for (;len; len--) { 1845 crc = (crc >> 8) ^ crc16tbl[0][(crc ^ *p++) & 0xff]; 1846 } 1847 return crc; 1848} 1849 1850/* 1851 * Initialize LZHUF decoder. 1852 * 1853 * Returns ARCHIVE_OK if initialization was successful. 1854 * Returns ARCHIVE_FAILED if method is unsupported. 1855 * Returns ARCHIVE_FATAL if initialization failed; memory allocation 1856 * error occurred. 1857 */ 1858static int 1859lzh_decode_init(struct lzh_stream *strm, const char *method) 1860{ 1861 struct lzh_dec *ds; 1862 int w_bits, w_size; 1863 1864 if (strm->ds == NULL) { 1865 strm->ds = calloc(1, sizeof(*strm->ds)); 1866 if (strm->ds == NULL) 1867 return (ARCHIVE_FATAL); 1868 } 1869 ds = strm->ds; 1870 ds->error = ARCHIVE_FAILED; 1871 if (method == NULL || method[0] != 'l' || method[1] != 'h') 1872 return (ARCHIVE_FAILED); 1873 switch (method[2]) { 1874 case '5': 1875 w_bits = 13;/* 8KiB for window */ 1876 break; 1877 case '6': 1878 w_bits = 15;/* 32KiB for window */ 1879 break; 1880 case '7': 1881 w_bits = 16;/* 64KiB for window */ 1882 break; 1883 default: 1884 return (ARCHIVE_FAILED);/* Not supported. */ 1885 } 1886 ds->error = ARCHIVE_FATAL; 1887 /* Expand a window size up to 128 KiB for decompressing process 1888 * performance whatever its original window size is. */ 1889 ds->w_size = 1U << 17; 1890 ds->w_mask = ds->w_size -1; 1891 if (ds->w_buff == NULL) { 1892 ds->w_buff = malloc(ds->w_size); 1893 if (ds->w_buff == NULL) 1894 return (ARCHIVE_FATAL); 1895 } 1896 w_size = 1U << w_bits; 1897 memset(ds->w_buff + ds->w_size - w_size, 0x20, w_size); 1898 ds->w_pos = 0; 1899 ds->state = 0; 1900 ds->pos_pt_len_size = w_bits + 1; 1901 ds->pos_pt_len_bits = (w_bits == 15 || w_bits == 16)? 5: 4; 1902 ds->literal_pt_len_size = PT_BITLEN_SIZE; 1903 ds->literal_pt_len_bits = 5; 1904 ds->br.cache_buffer = 0; 1905 ds->br.cache_avail = 0; 1906 1907 if (lzh_huffman_init(&(ds->lt), LT_BITLEN_SIZE, 16) 1908 != ARCHIVE_OK) 1909 return (ARCHIVE_FATAL); 1910 ds->lt.len_bits = 9; 1911 if (lzh_huffman_init(&(ds->pt), PT_BITLEN_SIZE, 16) 1912 != ARCHIVE_OK) 1913 return (ARCHIVE_FATAL); 1914 ds->error = 0; 1915 1916 return (ARCHIVE_OK); 1917} 1918 1919/* 1920 * Release LZHUF decoder. 1921 */ 1922static void 1923lzh_decode_free(struct lzh_stream *strm) 1924{ 1925 1926 if (strm->ds == NULL) 1927 return; 1928 free(strm->ds->w_buff); 1929 lzh_huffman_free(&(strm->ds->lt)); 1930 lzh_huffman_free(&(strm->ds->pt)); 1931 free(strm->ds); 1932 strm->ds = NULL; 1933} 1934 1935/* 1936 * Bit stream reader. 1937 */ 1938/* Check that the cache buffer has enough bits. */ 1939#define lzh_br_has(br, n) ((br)->cache_avail >= n) 1940/* Get compressed data by bit. */ 1941#define lzh_br_bits(br, n) \ 1942 (((uint16_t)((br)->cache_buffer >> \ 1943 ((br)->cache_avail - (n)))) & cache_masks[n]) 1944#define lzh_br_bits_forced(br, n) \ 1945 (((uint16_t)((br)->cache_buffer << \ 1946 ((n) - (br)->cache_avail))) & cache_masks[n]) 1947/* Read ahead to make sure the cache buffer has enough compressed data we 1948 * will use. 1949 * True : completed, there is enough data in the cache buffer. 1950 * False : we met that strm->next_in is empty, we have to get following 1951 * bytes. */ 1952#define lzh_br_read_ahead_0(strm, br, n) \ 1953 (lzh_br_has(br, (n)) || lzh_br_fillup(strm, br)) 1954/* True : the cache buffer has some bits as much as we need. 1955 * False : there are no enough bits in the cache buffer to be used, 1956 * we have to get following bytes if we could. */ 1957#define lzh_br_read_ahead(strm, br, n) \ 1958 (lzh_br_read_ahead_0((strm), (br), (n)) || lzh_br_has((br), (n))) 1959 1960/* Notify how many bits we consumed. */ 1961#define lzh_br_consume(br, n) ((br)->cache_avail -= (n)) 1962#define lzh_br_unconsume(br, n) ((br)->cache_avail += (n)) 1963 1964static const uint16_t cache_masks[] = { 1965 0x0000, 0x0001, 0x0003, 0x0007, 1966 0x000F, 0x001F, 0x003F, 0x007F, 1967 0x00FF, 0x01FF, 0x03FF, 0x07FF, 1968 0x0FFF, 0x1FFF, 0x3FFF, 0x7FFF, 1969 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF 1970}; 1971 1972/* 1973 * Shift away used bits in the cache data and fill it up with following bits. 1974 * Call this when cache buffer does not have enough bits you need. 1975 * 1976 * Returns 1 if the cache buffer is full. 1977 * Returns 0 if the cache buffer is not full; input buffer is empty. 1978 */ 1979static int 1980lzh_br_fillup(struct lzh_stream *strm, struct lzh_br *br) 1981{ 1982 int n = CACHE_BITS - br->cache_avail; 1983 1984 for (;;) { 1985 const int x = n >> 3; 1986 if (strm->avail_in >= x) { 1987 switch (x) { 1988 case 8: 1989 br->cache_buffer = 1990 ((uint64_t)strm->next_in[0]) << 56 | 1991 ((uint64_t)strm->next_in[1]) << 48 | 1992 ((uint64_t)strm->next_in[2]) << 40 | 1993 ((uint64_t)strm->next_in[3]) << 32 | 1994 ((uint32_t)strm->next_in[4]) << 24 | 1995 ((uint32_t)strm->next_in[5]) << 16 | 1996 ((uint32_t)strm->next_in[6]) << 8 | 1997 (uint32_t)strm->next_in[7]; 1998 strm->next_in += 8; 1999 strm->avail_in -= 8; 2000 br->cache_avail += 8 * 8; 2001 return (1); 2002 case 7: 2003 br->cache_buffer = 2004 (br->cache_buffer << 56) | 2005 ((uint64_t)strm->next_in[0]) << 48 | 2006 ((uint64_t)strm->next_in[1]) << 40 | 2007 ((uint64_t)strm->next_in[2]) << 32 | 2008 ((uint32_t)strm->next_in[3]) << 24 | 2009 ((uint32_t)strm->next_in[4]) << 16 | 2010 ((uint32_t)strm->next_in[5]) << 8 | 2011 (uint32_t)strm->next_in[6]; 2012 strm->next_in += 7; 2013 strm->avail_in -= 7; 2014 br->cache_avail += 7 * 8; 2015 return (1); 2016 case 6: 2017 br->cache_buffer = 2018 (br->cache_buffer << 48) | 2019 ((uint64_t)strm->next_in[0]) << 40 | 2020 ((uint64_t)strm->next_in[1]) << 32 | 2021 ((uint32_t)strm->next_in[2]) << 24 | 2022 ((uint32_t)strm->next_in[3]) << 16 | 2023 ((uint32_t)strm->next_in[4]) << 8 | 2024 (uint32_t)strm->next_in[5]; 2025 strm->next_in += 6; 2026 strm->avail_in -= 6; 2027 br->cache_avail += 6 * 8; 2028 return (1); 2029 case 0: 2030 /* We have enough compressed data in 2031 * the cache buffer.*/ 2032 return (1); 2033 default: 2034 break; 2035 } 2036 } 2037 if (strm->avail_in == 0) { 2038 /* There is not enough compressed data to fill up the 2039 * cache buffer. */ 2040 return (0); 2041 } 2042 br->cache_buffer = 2043 (br->cache_buffer << 8) | *strm->next_in++; 2044 strm->avail_in--; 2045 br->cache_avail += 8; 2046 n -= 8; 2047 } 2048} 2049 2050/* 2051 * Decode LZHUF. 2052 * 2053 * 1. Returns ARCHIVE_OK if output buffer or input buffer are empty. 2054 * Please set available buffer and call this function again. 2055 * 2. Returns ARCHIVE_EOF if decompression has been completed. 2056 * 3. Returns ARCHIVE_FAILED if an error occurred; compressed data 2057 * is broken or you do not set 'last' flag properly. 2058 * 4. 'last' flag is very important, you must set 1 to the flag if there 2059 * is no input data. The lha compressed data format does not provide how 2060 * to know the compressed data is really finished. 2061 * Note: lha command utility check if the total size of output bytes is 2062 * reached the uncompressed size recorded in its header. it does not mind 2063 * that the decoding process is properly finished. 2064 * GNU ZIP can decompress another compressed file made by SCO LZH compress. 2065 * it handles EOF as null to fill read buffer with zero until the decoding 2066 * process meet 2 bytes of zeros at reading a size of a next chunk, so the 2067 * zeros are treated as the mark of the end of the data although the zeros 2068 * is dummy, not the file data. 2069 */ 2070static int lzh_read_blocks(struct lzh_stream *, int); 2071static int lzh_decode_blocks(struct lzh_stream *, int); 2072#define ST_RD_BLOCK 0 2073#define ST_RD_PT_1 1 2074#define ST_RD_PT_2 2 2075#define ST_RD_PT_3 3 2076#define ST_RD_PT_4 4 2077#define ST_RD_LITERAL_1 5 2078#define ST_RD_LITERAL_2 6 2079#define ST_RD_LITERAL_3 7 2080#define ST_RD_POS_DATA_1 8 2081#define ST_GET_LITERAL 9 2082#define ST_GET_POS_1 10 2083#define ST_GET_POS_2 11 2084#define ST_COPY_DATA 12 2085 2086static int 2087lzh_decode(struct lzh_stream *strm, int last) 2088{ 2089 struct lzh_dec *ds = strm->ds; 2090 int avail_in; 2091 int r; 2092 2093 if (ds->error) 2094 return (ds->error); 2095 2096 avail_in = strm->avail_in; 2097 do { 2098 if (ds->state < ST_GET_LITERAL) 2099 r = lzh_read_blocks(strm, last); 2100 else 2101 r = lzh_decode_blocks(strm, last); 2102 } while (r == 100); 2103 strm->total_in += avail_in - strm->avail_in; 2104 return (r); 2105} 2106 2107static void 2108lzh_emit_window(struct lzh_stream *strm, size_t s) 2109{ 2110 strm->ref_ptr = strm->ds->w_buff; 2111 strm->avail_out = (int)s; 2112 strm->total_out += s; 2113} 2114 2115static int 2116lzh_read_blocks(struct lzh_stream *strm, int last) 2117{ 2118 struct lzh_dec *ds = strm->ds; 2119 struct lzh_br *br = &(ds->br); 2120 int c = 0, i; 2121 unsigned rbits; 2122 2123 for (;;) { 2124 switch (ds->state) { 2125 case ST_RD_BLOCK: 2126 /* 2127 * Read a block number indicates how many blocks 2128 * we will handle. The block is composed of a 2129 * literal and a match, sometimes a literal only 2130 * in particular, there are no reference data at 2131 * the beginning of the decompression. 2132 */ 2133 if (!lzh_br_read_ahead_0(strm, br, 16)) { 2134 if (!last) 2135 /* We need following data. */ 2136 return (ARCHIVE_OK); 2137 if (lzh_br_has(br, 8)) { 2138 /* 2139 * It seems there are extra bits. 2140 * 1. Compressed data is broken. 2141 * 2. `last' flag does not properly 2142 * set. 2143 */ 2144 goto failed; 2145 } 2146 if (ds->w_pos > 0) { 2147 lzh_emit_window(strm, ds->w_pos); 2148 ds->w_pos = 0; 2149 return (ARCHIVE_OK); 2150 } 2151 /* End of compressed data; we have completely 2152 * handled all compressed data. */ 2153 return (ARCHIVE_EOF); 2154 } 2155 ds->blocks_avail = lzh_br_bits(br, 16); 2156 if (ds->blocks_avail == 0) 2157 goto failed; 2158 lzh_br_consume(br, 16); 2159 /* 2160 * Read a literal table compressed in huffman 2161 * coding. 2162 */ 2163 ds->pt.len_size = ds->literal_pt_len_size; 2164 ds->pt.len_bits = ds->literal_pt_len_bits; 2165 ds->reading_position = 0; 2166 /* FALL THROUGH */ 2167 case ST_RD_PT_1: 2168 /* Note: ST_RD_PT_1, ST_RD_PT_2 and ST_RD_PT_4 are 2169 * used in reading both a literal table and a 2170 * position table. */ 2171 if (!lzh_br_read_ahead(strm, br, ds->pt.len_bits)) { 2172 if (last) 2173 goto failed;/* Truncated data. */ 2174 ds->state = ST_RD_PT_1; 2175 return (ARCHIVE_OK); 2176 } 2177 ds->pt.len_avail = lzh_br_bits(br, ds->pt.len_bits); 2178 lzh_br_consume(br, ds->pt.len_bits); 2179 /* FALL THROUGH */ 2180 case ST_RD_PT_2: 2181 if (ds->pt.len_avail == 0) { 2182 /* There is no bitlen. */ 2183 if (!lzh_br_read_ahead(strm, br, 2184 ds->pt.len_bits)) { 2185 if (last) 2186 goto failed;/* Truncated data.*/ 2187 ds->state = ST_RD_PT_2; 2188 return (ARCHIVE_OK); 2189 } 2190 if (!lzh_make_fake_table(&(ds->pt), 2191 lzh_br_bits(br, ds->pt.len_bits))) 2192 goto failed;/* Invalid data. */ 2193 lzh_br_consume(br, ds->pt.len_bits); 2194 if (ds->reading_position) 2195 ds->state = ST_GET_LITERAL; 2196 else 2197 ds->state = ST_RD_LITERAL_1; 2198 break; 2199 } else if (ds->pt.len_avail > ds->pt.len_size) 2200 goto failed;/* Invalid data. */ 2201 ds->loop = 0; 2202 memset(ds->pt.freq, 0, sizeof(ds->pt.freq)); 2203 if (ds->pt.len_avail < 3 || 2204 ds->pt.len_size == ds->pos_pt_len_size) { 2205 ds->state = ST_RD_PT_4; 2206 break; 2207 } 2208 /* FALL THROUGH */ 2209 case ST_RD_PT_3: 2210 ds->loop = lzh_read_pt_bitlen(strm, ds->loop, 3); 2211 if (ds->loop < 3) { 2212 if (ds->loop < 0 || last) 2213 goto failed;/* Invalid data. */ 2214 /* Not completed, get following data. */ 2215 ds->state = ST_RD_PT_3; 2216 return (ARCHIVE_OK); 2217 } 2218 /* There are some null in bitlen of the literal. */ 2219 if (!lzh_br_read_ahead(strm, br, 2)) { 2220 if (last) 2221 goto failed;/* Truncated data. */ 2222 ds->state = ST_RD_PT_3; 2223 return (ARCHIVE_OK); 2224 } 2225 c = lzh_br_bits(br, 2); 2226 lzh_br_consume(br, 2); 2227 if (c > ds->pt.len_avail - 3) 2228 goto failed;/* Invalid data. */ 2229 for (i = 3; c-- > 0 ;) 2230 ds->pt.bitlen[i++] = 0; 2231 ds->loop = i; 2232 /* FALL THROUGH */ 2233 case ST_RD_PT_4: 2234 ds->loop = lzh_read_pt_bitlen(strm, ds->loop, 2235 ds->pt.len_avail); 2236 if (ds->loop < ds->pt.len_avail) { 2237 if (ds->loop < 0 || last) 2238 goto failed;/* Invalid data. */ 2239 /* Not completed, get following data. */ 2240 ds->state = ST_RD_PT_4; 2241 return (ARCHIVE_OK); 2242 } 2243 if (!lzh_make_huffman_table(&(ds->pt))) 2244 goto failed;/* Invalid data */ 2245 if (ds->reading_position) { 2246 ds->state = ST_GET_LITERAL; 2247 break; 2248 } 2249 /* FALL THROUGH */ 2250 case ST_RD_LITERAL_1: 2251 if (!lzh_br_read_ahead(strm, br, ds->lt.len_bits)) { 2252 if (last) 2253 goto failed;/* Truncated data. */ 2254 ds->state = ST_RD_LITERAL_1; 2255 return (ARCHIVE_OK); 2256 } 2257 ds->lt.len_avail = lzh_br_bits(br, ds->lt.len_bits); 2258 lzh_br_consume(br, ds->lt.len_bits); 2259 /* FALL THROUGH */ 2260 case ST_RD_LITERAL_2: 2261 if (ds->lt.len_avail == 0) { 2262 /* There is no bitlen. */ 2263 if (!lzh_br_read_ahead(strm, br, 2264 ds->lt.len_bits)) { 2265 if (last) 2266 goto failed;/* Truncated data.*/ 2267 ds->state = ST_RD_LITERAL_2; 2268 return (ARCHIVE_OK); 2269 } 2270 if (!lzh_make_fake_table(&(ds->lt), 2271 lzh_br_bits(br, ds->lt.len_bits))) 2272 goto failed;/* Invalid data */ 2273 lzh_br_consume(br, ds->lt.len_bits); 2274 ds->state = ST_RD_POS_DATA_1; 2275 break; 2276 } else if (ds->lt.len_avail > ds->lt.len_size) 2277 goto failed;/* Invalid data */ 2278 ds->loop = 0; 2279 memset(ds->lt.freq, 0, sizeof(ds->lt.freq)); 2280 /* FALL THROUGH */ 2281 case ST_RD_LITERAL_3: 2282 i = ds->loop; 2283 while (i < ds->lt.len_avail) { 2284 if (!lzh_br_read_ahead(strm, br, 2285 ds->pt.max_bits)) { 2286 if (last) 2287 goto failed;/* Truncated data.*/ 2288 ds->loop = i; 2289 ds->state = ST_RD_LITERAL_3; 2290 return (ARCHIVE_OK); 2291 } 2292 rbits = lzh_br_bits(br, ds->pt.max_bits); 2293 c = lzh_decode_huffman(&(ds->pt), rbits); 2294 if (c > 2) { 2295 /* Note: 'c' will never be more than 2296 * eighteen since it's limited by 2297 * PT_BITLEN_SIZE, which is being set 2298 * to ds->pt.len_size through 2299 * ds->literal_pt_len_size. */ 2300 lzh_br_consume(br, ds->pt.bitlen[c]); 2301 c -= 2; 2302 ds->lt.freq[c]++; 2303 ds->lt.bitlen[i++] = c; 2304 } else if (c == 0) { 2305 lzh_br_consume(br, ds->pt.bitlen[c]); 2306 ds->lt.bitlen[i++] = 0; 2307 } else { 2308 /* c == 1 or c == 2 */ 2309 int n = (c == 1)?4:9; 2310 if (!lzh_br_read_ahead(strm, br, 2311 ds->pt.bitlen[c] + n)) { 2312 if (last) /* Truncated data. */ 2313 goto failed; 2314 ds->loop = i; 2315 ds->state = ST_RD_LITERAL_3; 2316 return (ARCHIVE_OK); 2317 } 2318 lzh_br_consume(br, ds->pt.bitlen[c]); 2319 c = lzh_br_bits(br, n); 2320 lzh_br_consume(br, n); 2321 c += (n == 4)?3:20; 2322 if (i + c > ds->lt.len_avail) 2323 goto failed;/* Invalid data */ 2324 memset(&(ds->lt.bitlen[i]), 0, c); 2325 i += c; 2326 } 2327 } 2328 if (i > ds->lt.len_avail || 2329 !lzh_make_huffman_table(&(ds->lt))) 2330 goto failed;/* Invalid data */ 2331 /* FALL THROUGH */ 2332 case ST_RD_POS_DATA_1: 2333 /* 2334 * Read a position table compressed in huffman 2335 * coding. 2336 */ 2337 ds->pt.len_size = ds->pos_pt_len_size; 2338 ds->pt.len_bits = ds->pos_pt_len_bits; 2339 ds->reading_position = 1; 2340 ds->state = ST_RD_PT_1; 2341 break; 2342 case ST_GET_LITERAL: 2343 return (100); 2344 } 2345 } 2346failed: 2347 return (ds->error = ARCHIVE_FAILED); 2348} 2349 2350static int 2351lzh_decode_blocks(struct lzh_stream *strm, int last) 2352{ 2353 struct lzh_dec *ds = strm->ds; 2354 struct lzh_br bre = ds->br; 2355 struct huffman *lt = &(ds->lt); 2356 struct huffman *pt = &(ds->pt); 2357 unsigned char *w_buff = ds->w_buff; 2358 unsigned char *lt_bitlen = lt->bitlen; 2359 unsigned char *pt_bitlen = pt->bitlen; 2360 int blocks_avail = ds->blocks_avail, c = 0; 2361 int copy_len = ds->copy_len, copy_pos = ds->copy_pos; 2362 int w_pos = ds->w_pos, w_mask = ds->w_mask, w_size = ds->w_size; 2363 int lt_max_bits = lt->max_bits, pt_max_bits = pt->max_bits; 2364 int state = ds->state; 2365 2366 for (;;) { 2367 switch (state) { 2368 case ST_GET_LITERAL: 2369 for (;;) { 2370 if (blocks_avail == 0) { 2371 /* We have decoded all blocks. 2372 * Let's handle next blocks. */ 2373 ds->state = ST_RD_BLOCK; 2374 ds->br = bre; 2375 ds->blocks_avail = 0; 2376 ds->w_pos = w_pos; 2377 ds->copy_pos = 0; 2378 return (100); 2379 } 2380 2381 /* lzh_br_read_ahead() always try to fill the 2382 * cache buffer up. In specific situation we 2383 * are close to the end of the data, the cache 2384 * buffer will not be full and thus we have to 2385 * determine if the cache buffer has some bits 2386 * as much as we need after lzh_br_read_ahead() 2387 * failed. */ 2388 if (!lzh_br_read_ahead(strm, &bre, 2389 lt_max_bits)) { 2390 if (!last) 2391 goto next_data; 2392 /* Remaining bits are less than 2393 * maximum bits(lt.max_bits) but maybe 2394 * it still remains as much as we need, 2395 * so we should try to use it with 2396 * dummy bits. */ 2397 c = lzh_decode_huffman(lt, 2398 lzh_br_bits_forced(&bre, 2399 lt_max_bits)); 2400 lzh_br_consume(&bre, lt_bitlen[c]); 2401 if (!lzh_br_has(&bre, 0)) 2402 goto failed;/* Over read. */ 2403 } else { 2404 c = lzh_decode_huffman(lt, 2405 lzh_br_bits(&bre, lt_max_bits)); 2406 lzh_br_consume(&bre, lt_bitlen[c]); 2407 } 2408 blocks_avail--; 2409 if (c > UCHAR_MAX) 2410 /* Current block is a match data. */ 2411 break; 2412 /* 2413 * 'c' is exactly a literal code. 2414 */ 2415 /* Save a decoded code to reference it 2416 * afterward. */ 2417 w_buff[w_pos] = c; 2418 if (++w_pos >= w_size) { 2419 w_pos = 0; 2420 lzh_emit_window(strm, w_size); 2421 goto next_data; 2422 } 2423 } 2424 /* 'c' is the length of a match pattern we have 2425 * already extracted, which has be stored in 2426 * window(ds->w_buff). */ 2427 copy_len = c - (UCHAR_MAX + 1) + MINMATCH; 2428 /* FALL THROUGH */ 2429 case ST_GET_POS_1: 2430 /* 2431 * Get a reference position. 2432 */ 2433 if (!lzh_br_read_ahead(strm, &bre, pt_max_bits)) { 2434 if (!last) { 2435 state = ST_GET_POS_1; 2436 ds->copy_len = copy_len; 2437 goto next_data; 2438 } 2439 copy_pos = lzh_decode_huffman(pt, 2440 lzh_br_bits_forced(&bre, pt_max_bits)); 2441 lzh_br_consume(&bre, pt_bitlen[copy_pos]); 2442 if (!lzh_br_has(&bre, 0)) 2443 goto failed;/* Over read. */ 2444 } else { 2445 copy_pos = lzh_decode_huffman(pt, 2446 lzh_br_bits(&bre, pt_max_bits)); 2447 lzh_br_consume(&bre, pt_bitlen[copy_pos]); 2448 } 2449 /* FALL THROUGH */ 2450 case ST_GET_POS_2: 2451 if (copy_pos > 1) { 2452 /* We need an additional adjustment number to 2453 * the position. */ 2454 int p = copy_pos - 1; 2455 if (!lzh_br_read_ahead(strm, &bre, p)) { 2456 if (last) 2457 goto failed;/* Truncated data.*/ 2458 state = ST_GET_POS_2; 2459 ds->copy_len = copy_len; 2460 ds->copy_pos = copy_pos; 2461 goto next_data; 2462 } 2463 copy_pos = (1 << p) + lzh_br_bits(&bre, p); 2464 lzh_br_consume(&bre, p); 2465 } 2466 /* The position is actually a distance from the last 2467 * code we had extracted and thus we have to convert 2468 * it to a position of the window. */ 2469 copy_pos = (w_pos - copy_pos - 1) & w_mask; 2470 /* FALL THROUGH */ 2471 case ST_COPY_DATA: 2472 /* 2473 * Copy `copy_len' bytes as extracted data from 2474 * the window into the output buffer. 2475 */ 2476 for (;;) { 2477 int l; 2478 2479 l = copy_len; 2480 if (copy_pos > w_pos) { 2481 if (l > w_size - copy_pos) 2482 l = w_size - copy_pos; 2483 } else { 2484 if (l > w_size - w_pos) 2485 l = w_size - w_pos; 2486 } 2487 if ((copy_pos + l < w_pos) 2488 || (w_pos + l < copy_pos)) { 2489 /* No overlap. */ 2490 memcpy(w_buff + w_pos, 2491 w_buff + copy_pos, l); 2492 } else { 2493 const unsigned char *s; 2494 unsigned char *d; 2495 int li; 2496 2497 d = w_buff + w_pos; 2498 s = w_buff + copy_pos; 2499 for (li = 0; li < l-1;) { 2500 d[li] = s[li];li++; 2501 d[li] = s[li];li++; 2502 } 2503 if (li < l) 2504 d[li] = s[li]; 2505 } 2506 w_pos += l; 2507 if (w_pos == w_size) { 2508 w_pos = 0; 2509 lzh_emit_window(strm, w_size); 2510 if (copy_len <= l) 2511 state = ST_GET_LITERAL; 2512 else { 2513 state = ST_COPY_DATA; 2514 ds->copy_len = copy_len - l; 2515 ds->copy_pos = 2516 (copy_pos + l) & w_mask; 2517 } 2518 goto next_data; 2519 } 2520 if (copy_len <= l) 2521 /* A copy of current pattern ended. */ 2522 break; 2523 copy_len -= l; 2524 copy_pos = (copy_pos + l) & w_mask; 2525 } 2526 state = ST_GET_LITERAL; 2527 break; 2528 } 2529 } 2530failed: 2531 return (ds->error = ARCHIVE_FAILED); 2532next_data: 2533 ds->br = bre; 2534 ds->blocks_avail = blocks_avail; 2535 ds->state = state; 2536 ds->w_pos = w_pos; 2537 return (ARCHIVE_OK); 2538} 2539 2540static int 2541lzh_huffman_init(struct huffman *hf, size_t len_size, int tbl_bits) 2542{ 2543 int bits; 2544 2545 if (hf->bitlen == NULL) { 2546 hf->bitlen = malloc(len_size * sizeof(hf->bitlen[0])); 2547 if (hf->bitlen == NULL) 2548 return (ARCHIVE_FATAL); 2549 } 2550 if (hf->tbl == NULL) { 2551 if (tbl_bits < HTBL_BITS) 2552 bits = tbl_bits; 2553 else 2554 bits = HTBL_BITS; 2555 hf->tbl = malloc(((size_t)1 << bits) * sizeof(hf->tbl[0])); 2556 if (hf->tbl == NULL) 2557 return (ARCHIVE_FATAL); 2558 } 2559 if (hf->tree == NULL && tbl_bits > HTBL_BITS) { 2560 hf->tree_avail = 1 << (tbl_bits - HTBL_BITS + 4); 2561 hf->tree = malloc(hf->tree_avail * sizeof(hf->tree[0])); 2562 if (hf->tree == NULL) 2563 return (ARCHIVE_FATAL); 2564 } 2565 hf->len_size = (int)len_size; 2566 hf->tbl_bits = tbl_bits; 2567 return (ARCHIVE_OK); 2568} 2569 2570static void 2571lzh_huffman_free(struct huffman *hf) 2572{ 2573 free(hf->bitlen); 2574 free(hf->tbl); 2575 free(hf->tree); 2576} 2577 2578static const char bitlen_tbl[0x400] = { 2579 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2580 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2581 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2582 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2583 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2584 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2585 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2586 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2587 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2588 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2589 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2590 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2591 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2592 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2593 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2594 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2595 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2596 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2597 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2598 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2599 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2600 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2601 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2602 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2603 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2604 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2605 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2606 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2607 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2608 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2609 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2610 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 2611 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2612 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2613 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2614 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2615 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2616 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2617 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2618 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2619 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2620 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2621 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2622 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2623 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2624 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2625 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2626 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 2627 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2628 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2629 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2630 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2631 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2632 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2633 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2634 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 2635 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 2636 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 2637 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 2638 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 2639 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 2640 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 2641 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 2642 13, 13, 13, 13, 13, 13, 13, 13, 14, 14, 14, 14, 15, 15, 16, 0 2643}; 2644static int 2645lzh_read_pt_bitlen(struct lzh_stream *strm, int start, int end) 2646{ 2647 struct lzh_dec *ds = strm->ds; 2648 struct lzh_br *br = &(ds->br); 2649 int c, i; 2650 2651 for (i = start; i < end; ) { 2652 /* 2653 * bit pattern the number we need 2654 * 000 -> 0 2655 * 001 -> 1 2656 * 010 -> 2 2657 * ... 2658 * 110 -> 6 2659 * 1110 -> 7 2660 * 11110 -> 8 2661 * ... 2662 * 1111111111110 -> 16 2663 */ 2664 if (!lzh_br_read_ahead(strm, br, 3)) 2665 return (i); 2666 if ((c = lzh_br_bits(br, 3)) == 7) { 2667 if (!lzh_br_read_ahead(strm, br, 13)) 2668 return (i); 2669 c = bitlen_tbl[lzh_br_bits(br, 13) & 0x3FF]; 2670 if (c) 2671 lzh_br_consume(br, c - 3); 2672 else 2673 return (-1);/* Invalid data. */ 2674 } else 2675 lzh_br_consume(br, 3); 2676 ds->pt.bitlen[i++] = c; 2677 ds->pt.freq[c]++; 2678 } 2679 return (i); 2680} 2681 2682static int 2683lzh_make_fake_table(struct huffman *hf, uint16_t c) 2684{ 2685 if (c >= hf->len_size) 2686 return (0); 2687 hf->tbl[0] = c; 2688 hf->max_bits = 0; 2689 hf->shift_bits = 0; 2690 hf->bitlen[hf->tbl[0]] = 0; 2691 return (1); 2692} 2693 2694/* 2695 * Make a huffman coding table. 2696 */ 2697static int 2698lzh_make_huffman_table(struct huffman *hf) 2699{ 2700 uint16_t *tbl; 2701 const unsigned char *bitlen; 2702 int bitptn[17], weight[17]; 2703 int i, maxbits = 0, ptn, tbl_size, w; 2704 int diffbits, len_avail; 2705 2706 /* 2707 * Initialize bit patterns. 2708 */ 2709 ptn = 0; 2710 for (i = 1, w = 1 << 15; i <= 16; i++, w >>= 1) { 2711 bitptn[i] = ptn; 2712 weight[i] = w; 2713 if (hf->freq[i]) { 2714 ptn += hf->freq[i] * w; 2715 maxbits = i; 2716 } 2717 } 2718 if (ptn != 0x10000 || maxbits > hf->tbl_bits) 2719 return (0);/* Invalid */ 2720 2721 hf->max_bits = maxbits; 2722 2723 /* 2724 * Cut out extra bits which we won't house in the table. 2725 * This preparation reduces the same calculation in the for-loop 2726 * making the table. 2727 */ 2728 if (maxbits < 16) { 2729 int ebits = 16 - maxbits; 2730 for (i = 1; i <= maxbits; i++) { 2731 bitptn[i] >>= ebits; 2732 weight[i] >>= ebits; 2733 } 2734 } 2735 if (maxbits > HTBL_BITS) { 2736 unsigned htbl_max; 2737 uint16_t *p; 2738 2739 diffbits = maxbits - HTBL_BITS; 2740 for (i = 1; i <= HTBL_BITS; i++) { 2741 bitptn[i] >>= diffbits; 2742 weight[i] >>= diffbits; 2743 } 2744 htbl_max = bitptn[HTBL_BITS] + 2745 weight[HTBL_BITS] * hf->freq[HTBL_BITS]; 2746 p = &(hf->tbl[htbl_max]); 2747 while (p < &hf->tbl[1U<<HTBL_BITS]) 2748 *p++ = 0; 2749 } else 2750 diffbits = 0; 2751 hf->shift_bits = diffbits; 2752 2753 /* 2754 * Make the table. 2755 */ 2756 tbl_size = 1 << HTBL_BITS; 2757 tbl = hf->tbl; 2758 bitlen = hf->bitlen; 2759 len_avail = hf->len_avail; 2760 hf->tree_used = 0; 2761 for (i = 0; i < len_avail; i++) { 2762 uint16_t *p; 2763 int len, cnt; 2764 uint16_t bit; 2765 int extlen; 2766 struct htree_t *ht; 2767 2768 if (bitlen[i] == 0) 2769 continue; 2770 /* Get a bit pattern */ 2771 len = bitlen[i]; 2772 ptn = bitptn[len]; 2773 cnt = weight[len]; 2774 if (len <= HTBL_BITS) { 2775 /* Calculate next bit pattern */ 2776 if ((bitptn[len] = ptn + cnt) > tbl_size) 2777 return (0);/* Invalid */ 2778 /* Update the table */ 2779 p = &(tbl[ptn]); 2780 if (cnt > 7) { 2781 uint16_t *pc; 2782 2783 cnt -= 8; 2784 pc = &p[cnt]; 2785 pc[0] = (uint16_t)i; 2786 pc[1] = (uint16_t)i; 2787 pc[2] = (uint16_t)i; 2788 pc[3] = (uint16_t)i; 2789 pc[4] = (uint16_t)i; 2790 pc[5] = (uint16_t)i; 2791 pc[6] = (uint16_t)i; 2792 pc[7] = (uint16_t)i; 2793 if (cnt > 7) { 2794 cnt -= 8; 2795 memcpy(&p[cnt], pc, 2796 8 * sizeof(uint16_t)); 2797 pc = &p[cnt]; 2798 while (cnt > 15) { 2799 cnt -= 16; 2800 memcpy(&p[cnt], pc, 2801 16 * sizeof(uint16_t)); 2802 } 2803 } 2804 if (cnt) 2805 memcpy(p, pc, cnt * sizeof(uint16_t)); 2806 } else { 2807 while (cnt > 1) { 2808 p[--cnt] = (uint16_t)i; 2809 p[--cnt] = (uint16_t)i; 2810 } 2811 if (cnt) 2812 p[--cnt] = (uint16_t)i; 2813 } 2814 continue; 2815 } 2816 2817 /* 2818 * A bit length is too big to be housed to a direct table, 2819 * so we use a tree model for its extra bits. 2820 */ 2821 bitptn[len] = ptn + cnt; 2822 bit = 1U << (diffbits -1); 2823 extlen = len - HTBL_BITS; 2824 2825 p = &(tbl[ptn >> diffbits]); 2826 if (*p == 0) { 2827 *p = len_avail + hf->tree_used; 2828 ht = &(hf->tree[hf->tree_used++]); 2829 if (hf->tree_used > hf->tree_avail) 2830 return (0);/* Invalid */ 2831 ht->left = 0; 2832 ht->right = 0; 2833 } else { 2834 if (*p < len_avail || 2835 *p >= (len_avail + hf->tree_used)) 2836 return (0);/* Invalid */ 2837 ht = &(hf->tree[*p - len_avail]); 2838 } 2839 while (--extlen > 0) { 2840 if (ptn & bit) { 2841 if (ht->left < len_avail) { 2842 ht->left = len_avail + hf->tree_used; 2843 ht = &(hf->tree[hf->tree_used++]); 2844 if (hf->tree_used > hf->tree_avail) 2845 return (0);/* Invalid */ 2846 ht->left = 0; 2847 ht->right = 0; 2848 } else { 2849 ht = &(hf->tree[ht->left - len_avail]); 2850 } 2851 } else { 2852 if (ht->right < len_avail) { 2853 ht->right = len_avail + hf->tree_used; 2854 ht = &(hf->tree[hf->tree_used++]); 2855 if (hf->tree_used > hf->tree_avail) 2856 return (0);/* Invalid */ 2857 ht->left = 0; 2858 ht->right = 0; 2859 } else { 2860 ht = &(hf->tree[ht->right - len_avail]); 2861 } 2862 } 2863 bit >>= 1; 2864 } 2865 if (ptn & bit) { 2866 if (ht->left != 0) 2867 return (0);/* Invalid */ 2868 ht->left = (uint16_t)i; 2869 } else { 2870 if (ht->right != 0) 2871 return (0);/* Invalid */ 2872 ht->right = (uint16_t)i; 2873 } 2874 } 2875 return (1); 2876} 2877 2878static int 2879lzh_decode_huffman_tree(struct huffman *hf, unsigned rbits, int c) 2880{ 2881 struct htree_t *ht; 2882 int extlen; 2883 2884 ht = hf->tree; 2885 extlen = hf->shift_bits; 2886 while (c >= hf->len_avail) { 2887 c -= hf->len_avail; 2888 if (extlen-- <= 0 || c >= hf->tree_used) 2889 return (0); 2890 if (rbits & (1U << extlen)) 2891 c = ht[c].left; 2892 else 2893 c = ht[c].right; 2894 } 2895 return (c); 2896} 2897 2898static inline int 2899lzh_decode_huffman(struct huffman *hf, unsigned rbits) 2900{ 2901 int c; 2902 /* 2903 * At first search an index table for a bit pattern. 2904 * If it fails, search a huffman tree for. 2905 */ 2906 c = hf->tbl[rbits >> hf->shift_bits]; 2907 if (c < hf->len_avail || hf->len_avail == 0) 2908 return (c); 2909 /* This bit pattern needs to be found out at a huffman tree. */ 2910 return (lzh_decode_huffman_tree(hf, rbits, c)); 2911} 2912 2913