1 /** 2 * unistr.c - Unicode string handling. Originated from the Linux-NTFS project. 3 * 4 * Copyright (c) 2000-2004 Anton Altaparmakov 5 * Copyright (c) 2002-2006 Szabolcs Szakacsits 6 * 7 * This program/include file is free software; you can redistribute it and/or 8 * modify it under the terms of the GNU General Public License as published 9 * by the Free Software Foundation; either version 2 of the License, or 10 * (at your option) any later version. 11 * 12 * This program/include file is distributed in the hope that it will be 13 * useful, but WITHOUT ANY WARRANTY; without even the implied warranty 14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 15 * GNU General Public License for more details. 16 * 17 * You should have received a copy of the GNU General Public License 18 * along with this program (in the main directory of the NTFS-3G 19 * distribution in the file COPYING); if not, write to the Free Software 20 * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 21 */ 22 23 #ifdef HAVE_CONFIG_H 24 #include "config.h" 25 #endif 26 27 #ifdef HAVE_STDIO_H 28 #include <stdio.h> 29 #endif 30 #ifdef HAVE_STDLIB_H 31 #include <stdlib.h> 32 #endif 33 #ifdef HAVE_WCHAR_H 34 #include <wchar.h> 35 #endif 36 #ifdef HAVE_STRING_H 37 #include <string.h> 38 #endif 39 #ifdef HAVE_ERRNO_H 40 #include <errno.h> 41 #endif 42 43 #include "attrib.h" 44 #include "types.h" 45 #include "unistr.h" 46 #include "debug.h" 47 #include "logging.h" 48 #include "misc.h" 49 50 /* 51 * IMPORTANT 52 * ========= 53 * 54 * All these routines assume that the Unicode characters are in little endian 55 * encoding inside the strings!!! 56 */ 57 58 /* 59 * This is used by the name collation functions to quickly determine what 60 * characters are (in)valid. 61 */ 62 #if 0 63 static const u8 legal_ansi_char_array[0x40] = { 64 0x00, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 65 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 66 67 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 68 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 69 70 0x17, 0x07, 0x18, 0x17, 0x17, 0x17, 0x17, 0x17, 71 0x17, 0x17, 0x18, 0x16, 0x16, 0x17, 0x07, 0x00, 72 73 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 74 0x17, 0x17, 0x04, 0x16, 0x18, 0x16, 0x18, 0x18, 75 }; 76 #endif 77 78 /** 79 * ntfs_names_are_equal - compare two Unicode names for equality 80 * @s1: name to compare to @s2 81 * @s1_len: length in Unicode characters of @s1 82 * @s2: name to compare to @s1 83 * @s2_len: length in Unicode characters of @s2 84 * @ic: ignore case bool 85 * @upcase: upcase table (only if @ic == IGNORE_CASE) 86 * @upcase_size: length in Unicode characters of @upcase (if present) 87 * 88 * Compare the names @s1 and @s2 and return TRUE (1) if the names are 89 * identical, or FALSE (0) if they are not identical. If @ic is IGNORE_CASE, 90 * the @upcase table is used to perform a case insensitive comparison. 91 */ 92 BOOL ntfs_names_are_equal(const ntfschar *s1, size_t s1_len, 93 const ntfschar *s2, size_t s2_len, 94 const IGNORE_CASE_BOOL ic, 95 const ntfschar *upcase, const u32 upcase_size) 96 { 97 if (s1_len != s2_len) 98 return FALSE; 99 if (!s1_len) 100 return TRUE; 101 if (ic == CASE_SENSITIVE) 102 return ntfs_ucsncmp(s1, s2, s1_len) ? FALSE: TRUE; 103 return ntfs_ucsncasecmp(s1, s2, s1_len, upcase, upcase_size) ? FALSE: 104 TRUE; 105 } 106 107 /** 108 * ntfs_names_collate - collate two Unicode names 109 * @name1: first Unicode name to compare 110 * @name1_len: length of first Unicode name to compare 111 * @name2: second Unicode name to compare 112 * @name2_len: length of second Unicode name to compare 113 * @err_val: if @name1 contains an invalid character return this value 114 * @ic: either CASE_SENSITIVE or IGNORE_CASE 115 * @upcase: upcase table (ignored if @ic is CASE_SENSITIVE) 116 * @upcase_len: upcase table size (ignored if @ic is CASE_SENSITIVE) 117 * 118 * ntfs_names_collate() collates two Unicode names and returns: 119 * 120 * -1 if the first name collates before the second one, 121 * 0 if the names match, 122 * 1 if the second name collates before the first one, or 123 * @err_val if an invalid character is found in @name1 during the comparison. 124 * 125 * The following characters are considered invalid: '"', '*', '<', '>' and '?'. 126 */ 127 int ntfs_names_collate(const ntfschar *name1, const u32 name1_len, 128 const ntfschar *name2, const u32 name2_len, 129 const int err_val __attribute__((unused)), 130 const IGNORE_CASE_BOOL ic, const ntfschar *upcase, 131 const u32 upcase_len) 132 { 133 u32 cnt; 134 ntfschar c1, c2; 135 136 #ifdef DEBUG 137 if (!name1 || !name2 || (ic && (!upcase || !upcase_len))) { 138 ntfs_log_debug("ntfs_names_collate received NULL pointer!\n"); 139 exit(1); 140 } 141 #endif 142 for (cnt = 0; cnt < min(name1_len, name2_len); ++cnt) { 143 c1 = le16_to_cpu(*name1); 144 name1++; 145 c2 = le16_to_cpu(*name2); 146 name2++; 147 if (ic) { 148 if (c1 < upcase_len) 149 c1 = le16_to_cpu(upcase[c1]); 150 if (c2 < upcase_len) 151 c2 = le16_to_cpu(upcase[c2]); 152 } 153 #if 0 154 if (c1 < 64 && legal_ansi_char_array[c1] & 8) 155 return err_val; 156 #endif 157 if (c1 < c2) 158 return -1; 159 if (c1 > c2) 160 return 1; 161 } 162 if (name1_len < name2_len) 163 return -1; 164 if (name1_len == name2_len) 165 return 0; 166 /* name1_len > name2_len */ 167 #if 0 168 c1 = le16_to_cpu(*name1); 169 if (c1 < 64 && legal_ansi_char_array[c1] & 8) 170 return err_val; 171 #endif 172 return 1; 173 } 174 175 /** 176 * ntfs_ucsncmp - compare two little endian Unicode strings 177 * @s1: first string 178 * @s2: second string 179 * @n: maximum unicode characters to compare 180 * 181 * Compare the first @n characters of the Unicode strings @s1 and @s2, 182 * The strings in little endian format and appropriate le16_to_cpu() 183 * conversion is performed on non-little endian machines. 184 * 185 * The function returns an integer less than, equal to, or greater than zero 186 * if @s1 (or the first @n Unicode characters thereof) is found, respectively, 187 * to be less than, to match, or be greater than @s2. 188 */ 189 int ntfs_ucsncmp(const ntfschar *s1, const ntfschar *s2, size_t n) 190 { 191 ntfschar c1, c2; 192 size_t i; 193 194 #ifdef DEBUG 195 if (!s1 || !s2) { 196 ntfs_log_debug("ntfs_wcsncmp() received NULL pointer!\n"); 197 exit(1); 198 } 199 #endif 200 for (i = 0; i < n; ++i) { 201 c1 = le16_to_cpu(s1[i]); 202 c2 = le16_to_cpu(s2[i]); 203 if (c1 < c2) 204 return -1; 205 if (c1 > c2) 206 return 1; 207 if (!c1) 208 break; 209 } 210 return 0; 211 } 212 213 /** 214 * ntfs_ucsncasecmp - compare two little endian Unicode strings, ignoring case 215 * @s1: first string 216 * @s2: second string 217 * @n: maximum unicode characters to compare 218 * @upcase: upcase table 219 * @upcase_size: upcase table size in Unicode characters 220 * 221 * Compare the first @n characters of the Unicode strings @s1 and @s2, 222 * ignoring case. The strings in little endian format and appropriate 223 * le16_to_cpu() conversion is performed on non-little endian machines. 224 * 225 * Each character is uppercased using the @upcase table before the comparison. 226 * 227 * The function returns an integer less than, equal to, or greater than zero 228 * if @s1 (or the first @n Unicode characters thereof) is found, respectively, 229 * to be less than, to match, or be greater than @s2. 230 */ 231 int ntfs_ucsncasecmp(const ntfschar *s1, const ntfschar *s2, size_t n, 232 const ntfschar *upcase, const u32 upcase_size) 233 { 234 ntfschar c1, c2; 235 size_t i; 236 237 #ifdef DEBUG 238 if (!s1 || !s2 || !upcase) { 239 ntfs_log_debug("ntfs_wcsncasecmp() received NULL pointer!\n"); 240 exit(1); 241 } 242 #endif 243 for (i = 0; i < n; ++i) { 244 if ((c1 = le16_to_cpu(s1[i])) < upcase_size) 245 c1 = le16_to_cpu(upcase[c1]); 246 if ((c2 = le16_to_cpu(s2[i])) < upcase_size) 247 c2 = le16_to_cpu(upcase[c2]); 248 if (c1 < c2) 249 return -1; 250 if (c1 > c2) 251 return 1; 252 if (!c1) 253 break; 254 } 255 return 0; 256 } 257 258 /** 259 * ntfs_ucsnlen - determine the length of a little endian Unicode string 260 * @s: pointer to Unicode string 261 * @maxlen: maximum length of string @s 262 * 263 * Return the number of Unicode characters in the little endian Unicode 264 * string @s up to a maximum of maxlen Unicode characters, not including 265 * the terminating (ntfschar)'\0'. If there is no (ntfschar)'\0' between @s 266 * and @s + @maxlen, @maxlen is returned. 267 * 268 * This function never looks beyond @s + @maxlen. 269 */ 270 u32 ntfs_ucsnlen(const ntfschar *s, u32 maxlen) 271 { 272 u32 i; 273 274 for (i = 0; i < maxlen; i++) { 275 if (!le16_to_cpu(s[i])) 276 break; 277 } 278 return i; 279 } 280 281 /** 282 * ntfs_ucsndup - duplicate little endian Unicode string 283 * @s: pointer to Unicode string 284 * @maxlen: maximum length of string @s 285 * 286 * Return a pointer to a new little endian Unicode string which is a duplicate 287 * of the string s. Memory for the new string is obtained with ntfs_malloc(3), 288 * and can be freed with free(3). 289 * 290 * A maximum of @maxlen Unicode characters are copied and a terminating 291 * (ntfschar)'\0' little endian Unicode character is added. 292 * 293 * This function never looks beyond @s + @maxlen. 294 * 295 * Return a pointer to the new little endian Unicode string on success and NULL 296 * on failure with errno set to the error code. 297 */ 298 ntfschar *ntfs_ucsndup(const ntfschar *s, u32 maxlen) 299 { 300 ntfschar *dst; 301 u32 len; 302 303 len = ntfs_ucsnlen(s, maxlen); 304 dst = ntfs_malloc((len + 1) * sizeof(ntfschar)); 305 if (dst) { 306 memcpy(dst, s, len * sizeof(ntfschar)); 307 dst[len] = cpu_to_le16(L'\0'); 308 } 309 return dst; 310 } 311 312 /** 313 * ntfs_name_upcase - Map an Unicode name to its uppercase equivalent 314 * @name: 315 * @name_len: 316 * @upcase: 317 * @upcase_len: 318 * 319 * Description... 320 * 321 * Returns: 322 */ 323 void ntfs_name_upcase(ntfschar *name, u32 name_len, const ntfschar *upcase, 324 const u32 upcase_len) 325 { 326 u32 i; 327 ntfschar u; 328 329 for (i = 0; i < name_len; i++) 330 if ((u = le16_to_cpu(name[i])) < upcase_len) 331 name[i] = upcase[u]; 332 } 333 334 /** 335 * ntfs_file_value_upcase - Convert a filename to upper case 336 * @file_name_attr: 337 * @upcase: 338 * @upcase_len: 339 * 340 * Description... 341 * 342 * Returns: 343 */ 344 void ntfs_file_value_upcase(FILE_NAME_ATTR *file_name_attr, 345 const ntfschar *upcase, const u32 upcase_len) 346 { 347 ntfs_name_upcase((ntfschar*)&file_name_attr->file_name, 348 file_name_attr->file_name_length, upcase, upcase_len); 349 } 350 351 /** 352 * ntfs_file_values_compare - Which of two filenames should be listed first 353 * @file_name_attr1: 354 * @file_name_attr2: 355 * @err_val: 356 * @ic: 357 * @upcase: 358 * @upcase_len: 359 * 360 * Description... 361 * 362 * Returns: 363 */ 364 int ntfs_file_values_compare(const FILE_NAME_ATTR *file_name_attr1, 365 const FILE_NAME_ATTR *file_name_attr2, 366 const int err_val, const IGNORE_CASE_BOOL ic, 367 const ntfschar *upcase, const u32 upcase_len) 368 { 369 return ntfs_names_collate((ntfschar*)&file_name_attr1->file_name, 370 file_name_attr1->file_name_length, 371 (ntfschar*)&file_name_attr2->file_name, 372 file_name_attr2->file_name_length, 373 err_val, ic, upcase, upcase_len); 374 } 375 376 #ifdef _BEOS_ 377 378 379 static int 380 to_utf8(wchar_t c,unsigned char* buf) 381 { 382 if(c==0) 383 return 0; /* No support for embedded 0 runes */ 384 if(c<0x80){ 385 if(buf)buf[0]=(unsigned char)c; 386 return 1; 387 } 388 if(c<0x800){ 389 if(buf){ 390 buf[0] = 0xc0 | (c>>6); 391 buf[1] = 0x80 | (c & 0x3f); 392 } 393 return 2; 394 } 395 if(c<0x10000){ 396 if(buf){ 397 buf[0] = 0xe0 | (c>>12); 398 buf[1] = 0x80 | ((c>>6) & 0x3f); 399 buf[2] = 0x80 | (c & 0x3f); 400 } 401 return 3; 402 } 403 /* We don't support characters above 0xFFFF in NTFS */ 404 return 0; 405 } 406 407 408 /* Decodes a sequence of utf8 bytes into a single wide character. 409 * The character is returned in host byte order. 410 * Returns the number of bytes consumed, or 0 on error. 411 */ 412 static int 413 from_utf8(const unsigned char* str,wchar_t *c) 414 { 415 int l=0,i; 416 417 if(*str<0x80){ 418 *c = *str; 419 return 1; 420 } 421 if(*str<0xc0) /* lead byte must not be 10xxxxxx */ 422 return 0; /* is c0 a possible lead byte? */ 423 if(*str<0xe0){ /* 110xxxxx */ 424 *c = *str & 0x1f; 425 l=2; 426 }else if(*str<0xf0){ /* 1110xxxx */ 427 *c = *str & 0xf; 428 l=3; 429 }else if(*str<0xf8){ /* 11110xxx */ 430 *c = *str & 7; 431 l=4; 432 }else /* We don't support characters above 0xFFFF in NTFS */ 433 return 0; 434 435 436 for(i=1;i<l;i++){ 437 /* all other bytes must be 10xxxxxx */ 438 if((str[i] & 0xc0) != 0x80) 439 return 0; 440 *c <<= 6; 441 *c |= str[i] & 0x3f; 442 } 443 return l; 444 } 445 446 447 /* Converts wide string to UTF-8. Expects two in- and two out-parameters. 448 * Returns 0 on success, or error code. 449 * The caller has to free the result string. 450 * There is no support for UTF-16, yet 451 */ 452 static int ntfs_dupuni2utf8(wchar_t* in, int in_len,char **out,int *out_len) 453 { 454 int i,tmp; 455 int len8; 456 unsigned char *result; 457 458 /* count the length of the resulting UTF-8 */ 459 for(i=len8=0;i<in_len;i++){ 460 tmp=to_utf8(le16_to_cpu( *(in+i) ),0); 461 if(!tmp) 462 /* invalid character */ 463 return EILSEQ; 464 len8+=tmp; 465 } 466 *out=result=ntfs_malloc(len8+1); /* allow for zero-termination */ 467 468 if(!result) 469 return ENOMEM; 470 result[len8]='\0'; 471 *out_len=len8; 472 for(i=len8=0;i<in_len;i++) 473 len8+=to_utf8(le16_to_cpu( *(in+i) ),result+len8); 474 return 0; 475 } 476 477 /* Converts an UTF-8 sequence to a wide string. Same conventions as the 478 * previous function 479 */ 480 static int ntfs_duputf82uni(unsigned char* in, int in_len,wchar_t** out,int *out_len) 481 { 482 int i,tmp; 483 int len16; 484 485 wchar_t* result; 486 wchar_t wtmp; 487 for(i=len16=0;i<in_len;i+=tmp,len16++){ 488 tmp=from_utf8(in+i,&wtmp); 489 if(!tmp) 490 return EILSEQ; 491 } 492 *out=result=ntfs_malloc(2*(len16+1)); 493 if(!result) 494 return ENOMEM; 495 result[len16]=0; 496 *out_len=len16; 497 for(i=len16=0;i<in_len;i+=tmp,len16++) 498 { 499 tmp=from_utf8(in+i, &wtmp); 500 *(result+len16) = cpu_to_le16(wtmp); 501 } 502 return 0; 503 } 504 505 506 /** 507 * ntfs_ucstombs - convert a little endian Unicode string to a multibyte string 508 * @ins: input Unicode string buffer 509 * @ins_len: length of input string in Unicode characters 510 * @outs: on return contains the (allocated) output multibyte string 511 * @outs_len: length of output buffer in bytes 512 * 513 * Convert the input little endian, 2-byte Unicode string @ins, of length 514 * @ins_len into the multibyte string format dictated by the current locale. 515 * 516 * If *@outs is NULL, the function allocates the string and the caller is 517 * responsible for calling free(*@outs); when finished with it. 518 * 519 * On success the function returns the number of bytes written to the output 520 * string *@outs (>= 0), not counting the terminating NULL byte. If the output 521 * string buffer was allocated, *@outs is set to it. 522 * 523 * On error, -1 is returned, and errno is set to the error code. The following 524 * error codes can be expected: 525 * EINVAL Invalid arguments (e.g. @ins or @outs is NULL). 526 * EILSEQ The input string cannot be represented as a multibyte 527 * sequence according to the current locale. 528 * ENAMETOOLONG Destination buffer is too small for input string. 529 * ENOMEM Not enough memory to allocate destination buffer. 530 */ 531 int ntfs_ucstombs(const ntfschar *ins, const int ins_len, char **outs, int outs_len) 532 { 533 int out_len = outs_len; 534 if(ntfs_dupuni2utf8(ins,ins_len,outs,&out_len)==0) 535 return out_len; 536 else 537 return EINVAL; 538 } 539 540 /** 541 * ntfs_mbstoucs - convert a multibyte string to a little endian Unicode string 542 * @ins: input multibyte string buffer 543 * @outs: on return contains the (allocated) output Unicode string 544 * @outs_len: length of output buffer in Unicode characters 545 * 546 * Convert the input multibyte string @ins, from the current locale into the 547 * corresponding little endian, 2-byte Unicode string. 548 * 549 * If *@outs is NULL, the function allocates the string and the caller is 550 * responsible for calling free(*@outs); when finished with it. 551 * 552 * On success the function returns the number of Unicode characters written to 553 * the output string *@outs (>= 0), not counting the terminating Unicode NULL 554 * character. If the output string buffer was allocated, *@outs is set to it. 555 * 556 * On error, -1 is returned, and errno is set to the error code. The following 557 * error codes can be expected: 558 * EINVAL Invalid arguments (e.g. @ins or @outs is NULL). 559 * EILSEQ The input string cannot be represented as a Unicode 560 * string according to the current locale. 561 * ENAMETOOLONG Destination buffer is too small for input string. 562 * ENOMEM Not enough memory to allocate destination buffer. 563 */ 564 int ntfs_mbstoucs(const char *ins, ntfschar **outs, int outs_len) 565 { 566 int in_len = strlen(ins); 567 int out_len = outs_len; 568 if(ntfs_duputf82uni(ins,in_len,outs,&out_len)==0) 569 return out_len; 570 else 571 return EILSEQ; 572 } 573 574 #else 575 576 /** 577 * ntfs_ucstombs - convert a little endian Unicode string to a multibyte string 578 * @ins: input Unicode string buffer 579 * @ins_len: length of input string in Unicode characters 580 * @outs: on return contains the (allocated) output multibyte string 581 * @outs_len: length of output buffer in bytes 582 * 583 * Convert the input little endian, 2-byte Unicode string @ins, of length 584 * @ins_len into the multibyte string format dictated by the current locale. 585 * 586 * If *@outs is NULL, the function allocates the string and the caller is 587 * responsible for calling free(*@outs); when finished with it. 588 * 589 * On success the function returns the number of bytes written to the output 590 * string *@outs (>= 0), not counting the terminating NULL byte. If the output 591 * string buffer was allocated, *@outs is set to it. 592 * 593 * On error, -1 is returned, and errno is set to the error code. The following 594 * error codes can be expected: 595 * EINVAL Invalid arguments (e.g. @ins or @outs is NULL). 596 * EILSEQ The input string cannot be represented as a multibyte 597 * sequence according to the current locale. 598 * ENAMETOOLONG Destination buffer is too small for input string. 599 * ENOMEM Not enough memory to allocate destination buffer. 600 */ 601 int ntfs_ucstombs(const ntfschar *ins, const int ins_len, char **outs, 602 int outs_len) 603 { 604 char *mbs; 605 wchar_t wc; 606 int i, o, mbs_len; 607 int cnt = 0; 608 #ifdef HAVE_MBSINIT 609 mbstate_t mbstate; 610 #endif 611 612 if (!ins || !outs) { 613 errno = EINVAL; 614 return -1; 615 } 616 mbs = *outs; 617 mbs_len = outs_len; 618 if (mbs && !mbs_len) { 619 errno = ENAMETOOLONG; 620 return -1; 621 } 622 if (!mbs) { 623 mbs_len = (ins_len + 1) * MB_CUR_MAX; 624 mbs = ntfs_malloc(mbs_len); 625 if (!mbs) 626 return -1; 627 } 628 #ifdef HAVE_MBSINIT 629 memset(&mbstate, 0, sizeof(mbstate)); 630 #else 631 wctomb(NULL, 0); 632 #endif 633 for (i = o = 0; i < ins_len; i++) { 634 /* Reallocate memory if necessary or abort. */ 635 if ((int)(o + MB_CUR_MAX) > mbs_len) { 636 char *tc; 637 if (mbs == *outs) { 638 errno = ENAMETOOLONG; 639 return -1; 640 } 641 tc = ntfs_malloc((mbs_len + 64) & ~63); 642 if (!tc) 643 goto err_out; 644 memcpy(tc, mbs, mbs_len); 645 mbs_len = (mbs_len + 64) & ~63; 646 free(mbs); 647 mbs = tc; 648 } 649 /* Convert the LE Unicode character to a CPU wide character. */ 650 wc = (wchar_t)le16_to_cpu(ins[i]); 651 if (!wc) 652 break; 653 /* Convert the CPU endian wide character to multibyte. */ 654 #ifdef HAVE_MBSINIT 655 cnt = wcrtomb(mbs + o, wc, &mbstate); 656 #else 657 cnt = wctomb(mbs + o, wc); 658 #endif 659 if (cnt == -1) 660 goto err_out; 661 if (cnt <= 0) { 662 ntfs_log_debug("Eeek. cnt <= 0, cnt = %i\n", cnt); 663 errno = EINVAL; 664 goto err_out; 665 } 666 o += cnt; 667 } 668 #ifdef HAVE_MBSINIT 669 /* Make sure we are back in the initial state. */ 670 if (!mbsinit(&mbstate)) { 671 ntfs_log_debug("Eeek. mbstate not in initial state!\n"); 672 errno = EILSEQ; 673 goto err_out; 674 } 675 #endif 676 /* Now write the NULL character. */ 677 mbs[o] = '\0'; 678 if (*outs != mbs) 679 *outs = mbs; 680 return o; 681 err_out: 682 if (mbs != *outs) { 683 int eo = errno; 684 free(mbs); 685 errno = eo; 686 } 687 return -1; 688 } 689 690 /** 691 * ntfs_mbstoucs - convert a multibyte string to a little endian Unicode string 692 * @ins: input multibyte string buffer 693 * @outs: on return contains the (allocated) output Unicode string 694 * @outs_len: length of output buffer in Unicode characters 695 * 696 * Convert the input multibyte string @ins, from the current locale into the 697 * corresponding little endian, 2-byte Unicode string. 698 * 699 * If *@outs is NULL, the function allocates the string and the caller is 700 * responsible for calling free(*@outs); when finished with it. 701 * 702 * On success the function returns the number of Unicode characters written to 703 * the output string *@outs (>= 0), not counting the terminating Unicode NULL 704 * character. If the output string buffer was allocated, *@outs is set to it. 705 * 706 * On error, -1 is returned, and errno is set to the error code. The following 707 * error codes can be expected: 708 * EINVAL Invalid arguments (e.g. @ins or @outs is NULL). 709 * EILSEQ The input string cannot be represented as a Unicode 710 * string according to the current locale. 711 * ENAMETOOLONG Destination buffer is too small for input string. 712 * ENOMEM Not enough memory to allocate destination buffer. 713 */ 714 int ntfs_mbstoucs(const char *ins, ntfschar **outs, int outs_len) 715 { 716 ntfschar *ucs; 717 const char *s; 718 wchar_t wc; 719 int i, o, cnt, ins_len, ucs_len, ins_size; 720 #ifdef HAVE_MBSINIT 721 mbstate_t mbstate; 722 #endif 723 724 if (!ins || !outs) { 725 errno = EINVAL; 726 return -1; 727 } 728 ucs = *outs; 729 ucs_len = outs_len; 730 if (ucs && !ucs_len) { 731 errno = ENAMETOOLONG; 732 return -1; 733 } 734 /* Determine the size of the multi-byte string in bytes. */ 735 ins_size = strlen(ins); 736 /* Determine the length of the multi-byte string. */ 737 s = ins; 738 #if defined(HAVE_MBSINIT) 739 memset(&mbstate, 0, sizeof(mbstate)); 740 ins_len = mbsrtowcs(NULL, (const char **)&s, 0, &mbstate); 741 #ifdef __CYGWIN32__ 742 if (!ins_len && *ins) { 743 /* Older Cygwin had broken mbsrtowcs() implementation. */ 744 ins_len = strlen(ins); 745 } 746 #endif 747 #elif !defined(DJGPP) 748 ins_len = mbstowcs(NULL, s, 0); 749 #else 750 /* Eeek!!! DJGPP has broken mbstowcs() implementation!!! */ 751 ins_len = strlen(ins); 752 #endif 753 if (ins_len == -1) 754 return ins_len; 755 #ifdef HAVE_MBSINIT 756 if ((s != ins) || !mbsinit(&mbstate)) { 757 #else 758 if (s != ins) { 759 #endif 760 errno = EILSEQ; 761 return -1; 762 } 763 /* Add the NULL terminator. */ 764 ins_len++; 765 if (!ucs) { 766 ucs_len = ins_len; 767 ucs = ntfs_malloc(ucs_len * sizeof(ntfschar)); 768 if (!ucs) 769 return -1; 770 } 771 #ifdef HAVE_MBSINIT 772 memset(&mbstate, 0, sizeof(mbstate)); 773 #else 774 mbtowc(NULL, NULL, 0); 775 #endif 776 for (i = o = cnt = 0; i < ins_size; i += cnt, o++) { 777 /* Reallocate memory if necessary or abort. */ 778 if (o >= ucs_len) { 779 ntfschar *tc; 780 if (ucs == *outs) { 781 errno = ENAMETOOLONG; 782 return -1; 783 } 784 /* 785 * We will never get here but hey, it's only a bit of 786 * extra code... 787 */ 788 ucs_len = (ucs_len * sizeof(ntfschar) + 64) & ~63; 789 tc = (ntfschar*)realloc(ucs, ucs_len); 790 if (!tc) 791 goto err_out; 792 ucs = tc; 793 ucs_len /= sizeof(ntfschar); 794 } 795 /* Convert the multibyte character to a wide character. */ 796 #ifdef HAVE_MBSINIT 797 cnt = mbrtowc(&wc, ins + i, ins_size - i, &mbstate); 798 #else 799 cnt = mbtowc(&wc, ins + i, ins_size - i); 800 #endif 801 if (!cnt) 802 break; 803 if (cnt == -1) 804 goto err_out; 805 if (cnt < -1) { 806 ntfs_log_trace("Eeek. cnt = %i\n", cnt); 807 errno = EINVAL; 808 goto err_out; 809 } 810 /* Make sure we are not overflowing the NTFS Unicode set. */ 811 if ((unsigned long)wc >= (unsigned long)(1 << 812 (8 * sizeof(ntfschar)))) { 813 errno = EILSEQ; 814 goto err_out; 815 } 816 /* Convert the CPU wide character to a LE Unicode character. */ 817 ucs[o] = cpu_to_le16(wc); 818 } 819 #ifdef HAVE_MBSINIT 820 /* Make sure we are back in the initial state. */ 821 if (!mbsinit(&mbstate)) { 822 ntfs_log_trace("Eeek. mbstate not in initial state!\n"); 823 errno = EILSEQ; 824 goto err_out; 825 } 826 #endif 827 /* Now write the NULL character. */ 828 ucs[o] = cpu_to_le16(L'\0'); 829 if (*outs != ucs) 830 *outs = ucs; 831 return o; 832 err_out: 833 if (ucs != *outs) { 834 int eo = errno; 835 free(ucs); 836 errno = eo; 837 } 838 return -1; 839 } 840 841 #endif 842 843 /** 844 * ntfs_upcase_table_build - build the default upcase table for NTFS 845 * @uc: destination buffer where to store the built table 846 * @uc_len: size of destination buffer in bytes 847 * 848 * ntfs_upcase_table_build() builds the default upcase table for NTFS and 849 * stores it in the caller supplied buffer @uc of size @uc_len. 850 * 851 * Note, @uc_len must be at least 128kiB in size or bad things will happen! 852 */ 853 void ntfs_upcase_table_build(ntfschar *uc, u32 uc_len) 854 { 855 static int uc_run_table[][3] = { /* Start, End, Add */ 856 {0x0061, 0x007B, -32}, {0x0451, 0x045D, -80}, {0x1F70, 0x1F72, 74}, 857 {0x00E0, 0x00F7, -32}, {0x045E, 0x0460, -80}, {0x1F72, 0x1F76, 86}, 858 {0x00F8, 0x00FF, -32}, {0x0561, 0x0587, -48}, {0x1F76, 0x1F78, 100}, 859 {0x0256, 0x0258, -205}, {0x1F00, 0x1F08, 8}, {0x1F78, 0x1F7A, 128}, 860 {0x028A, 0x028C, -217}, {0x1F10, 0x1F16, 8}, {0x1F7A, 0x1F7C, 112}, 861 {0x03AC, 0x03AD, -38}, {0x1F20, 0x1F28, 8}, {0x1F7C, 0x1F7E, 126}, 862 {0x03AD, 0x03B0, -37}, {0x1F30, 0x1F38, 8}, {0x1FB0, 0x1FB2, 8}, 863 {0x03B1, 0x03C2, -32}, {0x1F40, 0x1F46, 8}, {0x1FD0, 0x1FD2, 8}, 864 {0x03C2, 0x03C3, -31}, {0x1F51, 0x1F52, 8}, {0x1FE0, 0x1FE2, 8}, 865 {0x03C3, 0x03CC, -32}, {0x1F53, 0x1F54, 8}, {0x1FE5, 0x1FE6, 7}, 866 {0x03CC, 0x03CD, -64}, {0x1F55, 0x1F56, 8}, {0x2170, 0x2180, -16}, 867 {0x03CD, 0x03CF, -63}, {0x1F57, 0x1F58, 8}, {0x24D0, 0x24EA, -26}, 868 {0x0430, 0x0450, -32}, {0x1F60, 0x1F68, 8}, {0xFF41, 0xFF5B, -32}, 869 {0} 870 }; 871 static int uc_dup_table[][2] = { /* Start, End */ 872 {0x0100, 0x012F}, {0x01A0, 0x01A6}, {0x03E2, 0x03EF}, {0x04CB, 0x04CC}, 873 {0x0132, 0x0137}, {0x01B3, 0x01B7}, {0x0460, 0x0481}, {0x04D0, 0x04EB}, 874 {0x0139, 0x0149}, {0x01CD, 0x01DD}, {0x0490, 0x04BF}, {0x04EE, 0x04F5}, 875 {0x014A, 0x0178}, {0x01DE, 0x01EF}, {0x04BF, 0x04BF}, {0x04F8, 0x04F9}, 876 {0x0179, 0x017E}, {0x01F4, 0x01F5}, {0x04C1, 0x04C4}, {0x1E00, 0x1E95}, 877 {0x018B, 0x018B}, {0x01FA, 0x0218}, {0x04C7, 0x04C8}, {0x1EA0, 0x1EF9}, 878 {0} 879 }; 880 static int uc_byte_table[][2] = { /* Offset, Value */ 881 {0x00FF, 0x0178}, {0x01AD, 0x01AC}, {0x01F3, 0x01F1}, {0x0269, 0x0196}, 882 {0x0183, 0x0182}, {0x01B0, 0x01AF}, {0x0253, 0x0181}, {0x026F, 0x019C}, 883 {0x0185, 0x0184}, {0x01B9, 0x01B8}, {0x0254, 0x0186}, {0x0272, 0x019D}, 884 {0x0188, 0x0187}, {0x01BD, 0x01BC}, {0x0259, 0x018F}, {0x0275, 0x019F}, 885 {0x018C, 0x018B}, {0x01C6, 0x01C4}, {0x025B, 0x0190}, {0x0283, 0x01A9}, 886 {0x0192, 0x0191}, {0x01C9, 0x01C7}, {0x0260, 0x0193}, {0x0288, 0x01AE}, 887 {0x0199, 0x0198}, {0x01CC, 0x01CA}, {0x0263, 0x0194}, {0x0292, 0x01B7}, 888 {0x01A8, 0x01A7}, {0x01DD, 0x018E}, {0x0268, 0x0197}, 889 {0} 890 }; 891 int i, r; 892 893 memset((char*)uc, 0, uc_len); 894 uc_len >>= 1; 895 if (uc_len > 65536) 896 uc_len = 65536; 897 for (i = 0; (u32)i < uc_len; i++) 898 uc[i] = i; 899 for (r = 0; uc_run_table[r][0]; r++) 900 for (i = uc_run_table[r][0]; i < uc_run_table[r][1]; i++) 901 uc[i] += uc_run_table[r][2]; 902 for (r = 0; uc_dup_table[r][0]; r++) 903 for (i = uc_dup_table[r][0]; i < uc_dup_table[r][1]; i += 2) 904 uc[i + 1]--; 905 for (r = 0; uc_byte_table[r][0]; r++) 906 uc[uc_byte_table[r][0]] = uc_byte_table[r][1]; 907 } 908 909 /** 910 * ntfs_str2ucs - convert a string to a valid NTFS file name 911 * @s: input string 912 * @len: length of output buffer in Unicode characters 913 * 914 * Convert the input @s string into the corresponding little endian, 915 * 2-byte Unicode string. The length of the converted string is less 916 * or equal to the maximum length allowed by the NTFS format (255). 917 * 918 * If @s is NULL then return AT_UNNAMED. 919 * 920 * On success the function returns the Unicode string in an allocated 921 * buffer and the caller is responsible to free it when it's not needed 922 * anymore. 923 * 924 * On error NULL is returned and errno is set to the error code. 925 */ 926 ntfschar *ntfs_str2ucs(const char *s, int *len) 927 { 928 ntfschar *ucs = NULL; 929 930 if (s && ((*len = ntfs_mbstoucs(s, &ucs, 0)) == -1)) { 931 ntfs_log_perror("Couldn't convert '%s' to Unicode", s); 932 return NULL; 933 } 934 if (*len > NTFS_MAX_NAME_LEN) { 935 free(ucs); 936 errno = ENAMETOOLONG; 937 return NULL; 938 } 939 if (!ucs || !*len) { 940 ucs = AT_UNNAMED; 941 *len = 0; 942 } 943 return ucs; 944 } 945 946 /** 947 * ntfs_ucsfree - free memory allocated by ntfs_str2ucs() 948 * @ucs input string to be freed 949 * 950 * Free memory at @ucs and which was allocated by ntfs_str2ucs. 951 * 952 * Return value: none. 953 */ 954 void ntfs_ucsfree(ntfschar *ucs) 955 { 956 if (ucs && (ucs != AT_UNNAMED)) 957 free(ucs); 958 } 959 960