fs/cifs/cifs_unicode.c

   1 /*
   2  *   fs/cifs/cifs_unicode.c
   3  *
   4  *   Copyright (c) International Business Machines  Corp., 2000,2009
   5  *   Modified by Steve French (sfrench@us.ibm.com)
   6  *
   7  *   This program is free software;  you can redistribute it and/or modify
   8  *   it under the terms of the GNU General Public License as published by
   9  *   the Free Software Foundation; either version 2 of the License, or
  10  *   (at your option) any later version.
  11  *
  12  *   This program is distributed in the hope that it will be useful,
  13  *   but WITHOUT ANY WARRANTY;  without even the implied warranty of
  14  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See
  15  *   the 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;  if not, write to the Free Software
  19  *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
  20  */
  21 #include <linux/fs.h>
  22 #include <linux/slab.h>
  23 #include "cifs_fs_sb.h"
  24 #include "cifs_unicode.h"
  25 #include "cifs_uniupr.h"
  26 #include "cifspdu.h"
  27 #include "cifsglob.h"
  28 #include "cifs_debug.h"
  29
  30 int cifs_remap(struct cifs_sb_info *cifs_sb)
  31 {
  32         int map_type;
  33
  34         if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SFM_CHR)
  35                 map_type = SFM_MAP_UNI_RSVD;
  36         else if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR)
  37                 map_type = SFU_MAP_UNI_RSVD;
  38         else
  39                 map_type = NO_MAP_UNI_RSVD;
  40
  41         return map_type;
  42 }
  43
  44 /* Convert character using the SFU - "Services for Unix" remapping range */
  45 static bool
  46 convert_sfu_char(const __u16 src_char, char *target)
  47 {
  48         /*
  49          * BB: Cannot handle remapping UNI_SLASH until all the calls to
  50          *     build_path_from_dentry are modified, as they use slash as
  51          *     separator.
  52          */
  53         switch (src_char) {
  54         case UNI_COLON:
  55                 *target = ':';
  56                 break;
  57         case UNI_ASTERISK:
  58                 *target = '*';
  59                 break;
  60         case UNI_QUESTION:
  61                 *target = '?';
  62                 break;
  63         case UNI_PIPE:
  64                 *target = '|';
  65                 break;
  66         case UNI_GRTRTHAN:
  67                 *target = '>';
  68                 break;
  69         case UNI_LESSTHAN:
  70                 *target = '<';
  71                 break;
  72         default:
  73                 return false;
  74         }
  75         return true;
  76 }
  77
  78 /* Convert character using the SFM - "Services for Mac" remapping range */
  79 static bool
  80 convert_sfm_char(const __u16 src_char, char *target)
  81 {
  82         if (src_char >= 0xF001 && src_char <= 0xF01F) {
  83                 *target = src_char - 0xF000;
  84                 return true;
  85         }
  86         switch (src_char) {
  87         case SFM_COLON:
  88                 *target = ':';
  89                 break;
  90         case SFM_DOUBLEQUOTE:
  91                 *target = '"';
  92                 break;
  93         case SFM_ASTERISK:
  94                 *target = '*';
  95                 break;
  96         case SFM_QUESTION:
  97                 *target = '?';
  98                 break;
  99         case SFM_PIPE:
 100                 *target = '|';
 101                 break;
 102         case SFM_GRTRTHAN:
 103                 *target = '>';
 104                 break;
 105         case SFM_LESSTHAN:
 106                 *target = '<';
 107                 break;
 108         case SFM_SPACE:
 109                 *target = ' ';
 110                 break;
 111         case SFM_PERIOD:
 112                 *target = '.';
 113                 break;
 114         default:
 115                 return false;
 116         }
 117         return true;
 118 }
 119
 120
 121 /*
 122  * cifs_mapchar - convert a host-endian char to proper char in codepage
 123  * @target - where converted character should be copied
 124  * @src_char - 2 byte host-endian source character
 125  * @cp - codepage to which character should be converted
 126  * @map_type - How should the 7 NTFS/SMB reserved characters be mapped to UCS2?
 127  *
 128  * This function handles the conversion of a single character. It is the
 129  * responsibility of the caller to ensure that the target buffer is large
 130  * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
 131  */
 132 static int
 133 cifs_mapchar(char *target, const __u16 *from, const struct nls_table *cp,
 134              int maptype)
 135 {
 136         int len = 1;
 137         __u16 src_char;
 138
 139         src_char = *from;
 140
 141         if ((maptype == SFM_MAP_UNI_RSVD) && convert_sfm_char(src_char, target))
 142                 return len;
 143         else if ((maptype == SFU_MAP_UNI_RSVD) &&
 144                   convert_sfu_char(src_char, target))
 145                 return len;
 146
 147         /* if character not one of seven in special remap set */
 148         len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
 149         if (len <= 0)
 150                 goto surrogate_pair;
 151
 152         return len;
 153
 154 surrogate_pair:
 155         /* convert SURROGATE_PAIR and IVS */
 156         if (strcmp(cp->charset, "utf8"))
 157                 goto unknown;
 158         len = utf16s_to_utf8s(from, 3, UTF16_LITTLE_ENDIAN, target, 6);
 159         if (len <= 0)
 160                 goto unknown;
 161         return len;
 162
 163 unknown:
 164         *target = '?';
 165         len = 1;
 166         return len;
 167 }
 168
 169 /*
 170  * cifs_from_utf16 - convert utf16le string to local charset
 171  * @to - destination buffer
 172  * @from - source buffer
 173  * @tolen - destination buffer size (in bytes)
 174  * @fromlen - source buffer size (in bytes)
 175  * @codepage - codepage to which characters should be converted
 176  * @mapchar - should characters be remapped according to the mapchars option?
 177  *
 178  * Convert a little-endian utf16le string (as sent by the server) to a string
 179  * in the provided codepage. The tolen and fromlen parameters are to ensure
 180  * that the code doesn't walk off of the end of the buffer (which is always
 181  * a danger if the alignment of the source buffer is off). The destination
 182  * string is always properly null terminated and fits in the destination
 183  * buffer. Returns the length of the destination string in bytes (including
 184  * null terminator).
 185  *
 186  * Note that some windows versions actually send multiword UTF-16 characters
 187  * instead of straight UTF16-2. The linux nls routines however aren't able to
 188  * deal with those characters properly. In the event that we get some of
 189  * those characters, they won't be translated properly.
 190  */
 191 int
 192 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
 193                 const struct nls_table *codepage, int map_type)
 194 {
 195         int i, charlen, safelen;
 196         int outlen = 0;
 197         int nullsize = nls_nullsize(codepage);
 198         int fromwords = fromlen / 2;
 199         char tmp[NLS_MAX_CHARSET_SIZE];
 200         __u16 ftmp[3];          /* ftmp[3] = 3array x 2bytes = 6bytes UTF-16 */
 201
 202         /*
 203          * because the chars can be of varying widths, we need to take care
 204          * not to overflow the destination buffer when we get close to the
 205          * end of it. Until we get to this offset, we don't need to check
 206          * for overflow however.
 207          */
 208         safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
 209
 210         for (i = 0; i < fromwords; i++) {
 211                 ftmp[0] = get_unaligned_le16(&from[i]);
 212                 if (ftmp[0] == 0)
 213                         break;
 214                 if (i + 1 < fromwords)
 215                         ftmp[1] = get_unaligned_le16(&from[i + 1]);
 216                 else
 217                         ftmp[1] = 0;
 218                 if (i + 2 < fromwords)
 219                         ftmp[2] = get_unaligned_le16(&from[i + 2]);
 220                 else
 221                         ftmp[2] = 0;
 222
 223                 /*
 224                  * check to see if converting this character might make the
 225                  * conversion bleed into the null terminator
 226                  */
 227                 if (outlen >= safelen) {
 228                         charlen = cifs_mapchar(tmp, ftmp, codepage, map_type);
 229                         if ((outlen + charlen) > (tolen - nullsize))
 230                                 break;
 231                 }
 232
 233                 /* put converted char into 'to' buffer */
 234                 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, map_type);
 235                 outlen += charlen;
 236
 237                 /* charlen (=bytes of UTF-8 for 1 character)
 238                  * 4bytes UTF-8(surrogate pair) is charlen=4
 239                  *   (4bytes UTF-16 code)
 240                  * 7-8bytes UTF-8(IVS) is charlen=3+4 or 4+4
 241                  *   (2 UTF-8 pairs divided to 2 UTF-16 pairs) */
 242                 if (charlen == 4)
 243                         i++;
 244                 else if (charlen >= 5)
 245                         /* 5-6bytes UTF-8 */
 246                         i += 2;
 247         }
 248
 249         /* properly null-terminate string */
 250         for (i = 0; i < nullsize; i++)
 251                 to[outlen++] = 0;
 252
 253         return outlen;
 254 }
 255
 256 /*
 257  * NAME:        cifs_strtoUTF16()
 258  *
 259  * FUNCTION:    Convert character string to unicode string
 260  *
 261  */
 262 int
 263 cifs_strtoUTF16(__le16 *to, const char *from, int len,
 264               const struct nls_table *codepage)
 265 {
 266         int charlen;
 267         int i;
 268         wchar_t wchar_to; /* needed to quiet sparse */
 269
 270         /* special case for utf8 to handle no plane0 chars */
 271         if (!strcmp(codepage->charset, "utf8")) {
 272                 /*
 273                  * convert utf8 -> utf16, we assume we have enough space
 274                  * as caller should have assumed conversion does not overflow
 275                  * in destination len is length in wchar_t units (16bits)
 276                  */
 277                 i  = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
 278                                        (wchar_t *) to, len);
 279
 280                 /* if success terminate and exit */
 281                 if (i >= 0)
 282                         goto success;
 283                 /*
 284                  * if fails fall back to UCS encoding as this
 285                  * function should not return negative values
 286                  * currently can fail only if source contains
 287                  * invalid encoded characters
 288                  */
 289         }
 290
 291         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 292                 charlen = codepage->char2uni(from, len, &wchar_to);
 293                 if (charlen < 1) {
 294                         cifs_dbg(VFS, "strtoUTF16: char2uni of 0x%x returned %d\n",
 295                                  *from, charlen);
 296                         /* A question mark */
 297                         wchar_to = 0x003f;
 298                         charlen = 1;
 299                 }
 300                 put_unaligned_le16(wchar_to, &to[i]);
 301         }
 302
 303 success:
 304         put_unaligned_le16(0, &to[i]);
 305         return i;
 306 }
 307
 308 /*
 309  * cifs_utf16_bytes - how long will a string be after conversion?
 310  * @utf16 - pointer to input string
 311  * @maxbytes - don't go past this many bytes of input string
 312  * @codepage - destination codepage
 313  *
 314  * Walk a utf16le string and return the number of bytes that the string will
 315  * be after being converted to the given charset, not including any null
 316  * termination required. Don't walk past maxbytes in the source buffer.
 317  */
 318 int
 319 cifs_utf16_bytes(const __le16 *from, int maxbytes,
 320                 const struct nls_table *codepage)
 321 {
 322         int i;
 323         int charlen, outlen = 0;
 324         int maxwords = maxbytes / 2;
 325         char tmp[NLS_MAX_CHARSET_SIZE];
 326         __u16 ftmp[3];
 327
 328         for (i = 0; i < maxwords; i++) {
 329                 ftmp[0] = get_unaligned_le16(&from[i]);
 330                 if (ftmp[0] == 0)
 331                         break;
 332                 if (i + 1 < maxwords)
 333                         ftmp[1] = get_unaligned_le16(&from[i + 1]);
 334                 else
 335                         ftmp[1] = 0;
 336                 if (i + 2 < maxwords)
 337                         ftmp[2] = get_unaligned_le16(&from[i + 2]);
 338                 else
 339                         ftmp[2] = 0;
 340
 341                 charlen = cifs_mapchar(tmp, ftmp, codepage, NO_MAP_UNI_RSVD);
 342                 outlen += charlen;
 343         }
 344
 345         return outlen;
 346 }
 347
 348 /*
 349  * cifs_strndup_from_utf16 - copy a string from wire format to the local
 350  * codepage
 351  * @src - source string
 352  * @maxlen - don't walk past this many bytes in the source string
 353  * @is_unicode - is this a unicode string?
 354  * @codepage - destination codepage
 355  *
 356  * Take a string given by the server, convert it to the local codepage and
 357  * put it in a new buffer. Returns a pointer to the new string or NULL on
 358  * error.
 359  */
 360 char *
 361 cifs_strndup_from_utf16(const char *src, const int maxlen,
 362                         const bool is_unicode, const struct nls_table *codepage)
 363 {
 364         int len;
 365         char *dst;
 366
 367         if (is_unicode) {
 368                 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
 369                 len += nls_nullsize(codepage);
 370                 dst = kmalloc(len, GFP_KERNEL);
 371                 if (!dst)
 372                         return NULL;
 373                 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
 374                                NO_MAP_UNI_RSVD);
 375         } else {
 376                 len = strnlen(src, maxlen);
 377                 len++;
 378                 dst = kmalloc(len, GFP_KERNEL);
 379                 if (!dst)
 380                         return NULL;
 381                 strlcpy(dst, src, len);
 382         }
 383
 384         return dst;
 385 }
 386
 387 static __le16 convert_to_sfu_char(char src_char)
 388 {
 389         __le16 dest_char;
 390
 391         switch (src_char) {
 392         case ':':
 393                 dest_char = cpu_to_le16(UNI_COLON);
 394                 break;
 395         case '*':
 396                 dest_char = cpu_to_le16(UNI_ASTERISK);
 397                 break;
 398         case '?':
 399                 dest_char = cpu_to_le16(UNI_QUESTION);
 400                 break;
 401         case '<':
 402                 dest_char = cpu_to_le16(UNI_LESSTHAN);
 403                 break;
 404         case '>':
 405                 dest_char = cpu_to_le16(UNI_GRTRTHAN);
 406                 break;
 407         case '|':
 408                 dest_char = cpu_to_le16(UNI_PIPE);
 409                 break;
 410         default:
 411                 dest_char = 0;
 412         }
 413
 414         return dest_char;
 415 }
 416
 417 static __le16 convert_to_sfm_char(char src_char, bool end_of_string)
 418 {
 419         __le16 dest_char;
 420
 421         if (src_char >= 0x01 && src_char <= 0x1F) {
 422                 dest_char = cpu_to_le16(src_char + 0xF000);
 423                 return dest_char;
 424         }
 425         switch (src_char) {
 426         case ':':
 427                 dest_char = cpu_to_le16(SFM_COLON);
 428                 break;
 429         case '"':
 430                 dest_char = cpu_to_le16(SFM_DOUBLEQUOTE);
 431                 break;
 432         case '*':
 433                 dest_char = cpu_to_le16(SFM_ASTERISK);
 434                 break;
 435         case '?':
 436                 dest_char = cpu_to_le16(SFM_QUESTION);
 437                 break;
 438         case '<':
 439                 dest_char = cpu_to_le16(SFM_LESSTHAN);
 440                 break;
 441         case '>':
 442                 dest_char = cpu_to_le16(SFM_GRTRTHAN);
 443                 break;
 444         case '|':
 445                 dest_char = cpu_to_le16(SFM_PIPE);
 446                 break;
 447         case '.':
 448                 if (end_of_string)
 449                         dest_char = cpu_to_le16(SFM_PERIOD);
 450                 else
 451                         dest_char = 0;
 452                 break;
 453         case ' ':
 454                 if (end_of_string)
 455                         dest_char = cpu_to_le16(SFM_SPACE);
 456                 else
 457                         dest_char = 0;
 458                 break;
 459         default:
 460                 dest_char = 0;
 461         }
 462
 463         return dest_char;
 464 }
 465
 466 /*
 467  * Convert 16 bit Unicode pathname to wire format from string in current code
 468  * page. Conversion may involve remapping up the six characters that are
 469  * only legal in POSIX-like OS (if they are present in the string). Path
 470  * names are little endian 16 bit Unicode on the wire
 471  */
 472 int
 473 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
 474                  const struct nls_table *cp, int map_chars)
 475 {
 476         int i, charlen;
 477         int j = 0;
 478         char src_char;
 479         __le16 dst_char;
 480         wchar_t tmp;
 481         wchar_t *wchar_to;      /* UTF-16 */
 482         int ret;
 483         unicode_t u;
 484
 485         if (map_chars == NO_MAP_UNI_RSVD)
 486                 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
 487
 488         wchar_to = kzalloc(6, GFP_KERNEL);
 489
 490         for (i = 0; i < srclen; j++) {
 491                 src_char = source[i];
 492                 charlen = 1;
 493
 494                 /* check if end of string */
 495                 if (src_char == 0)
 496                         goto ctoUTF16_out;
 497
 498                 /* see if we must remap this char */
 499                 if (map_chars == SFU_MAP_UNI_RSVD)
 500                         dst_char = convert_to_sfu_char(src_char);
 501                 else if (map_chars == SFM_MAP_UNI_RSVD) {
 502                         bool end_of_string;
 503
 504                         if (i == srclen - 1)
 505                                 end_of_string = true;
 506                         else
 507                                 end_of_string = false;
 508
 509                         dst_char = convert_to_sfm_char(src_char, end_of_string);
 510                 } else
 511                         dst_char = 0;
 512                 /*
 513                  * FIXME: We can not handle remapping backslash (UNI_SLASH)
 514                  * until all the calls to build_path_from_dentry are modified,
 515                  * as they use backslash as separator.
 516                  */
 517                 if (dst_char == 0) {
 518                         charlen = cp->char2uni(source + i, srclen - i, &tmp);
 519                         dst_char = cpu_to_le16(tmp);
 520
 521                         /*
 522                          * if no match, use question mark, which at least in
 523                          * some cases serves as wild card
 524                          */
 525                         if (charlen > 0)
 526                                 goto ctoUTF16;
 527
 528                         /* convert SURROGATE_PAIR */
 529                         if (strcmp(cp->charset, "utf8") || !wchar_to)
 530                                 goto unknown;
 531                         if (*(source + i) & 0x80) {
 532                                 charlen = utf8_to_utf32(source + i, 6, &u);
 533                                 if (charlen < 0)
 534                                         goto unknown;
 535                         } else
 536                                 goto unknown;
 537                         ret  = utf8s_to_utf16s(source + i, charlen,
 538                                                UTF16_LITTLE_ENDIAN,
 539                                                wchar_to, 6);
 540                         if (ret < 0)
 541                                 goto unknown;
 542
 543                         i += charlen;
 544                         dst_char = cpu_to_le16(*wchar_to);
 545                         if (charlen <= 3)
 546                                 /* 1-3bytes UTF-8 to 2bytes UTF-16 */
 547                                 put_unaligned(dst_char, &target[j]);
 548                         else if (charlen == 4) {
 549                                 /* 4bytes UTF-8(surrogate pair) to 4bytes UTF-16
 550                                  * 7-8bytes UTF-8(IVS) divided to 2 UTF-16
 551                                  *   (charlen=3+4 or 4+4) */
 552                                 put_unaligned(dst_char, &target[j]);
 553                                 dst_char = cpu_to_le16(*(wchar_to + 1));
 554                                 j++;
 555                                 put_unaligned(dst_char, &target[j]);
 556                         } else if (charlen >= 5) {
 557                                 /* 5-6bytes UTF-8 to 6bytes UTF-16 */
 558                                 put_unaligned(dst_char, &target[j]);
 559                                 dst_char = cpu_to_le16(*(wchar_to + 1));
 560                                 j++;
 561                                 put_unaligned(dst_char, &target[j]);
 562                                 dst_char = cpu_to_le16(*(wchar_to + 2));
 563                                 j++;
 564                                 put_unaligned(dst_char, &target[j]);
 565                         }
 566                         continue;
 567
 568 unknown:
 569                         dst_char = cpu_to_le16(0x003f);
 570                         charlen = 1;
 571                 }
 572
 573 ctoUTF16:
 574                 /*
 575                  * character may take more than one byte in the source string,
 576                  * but will take exactly two bytes in the target string
 577                  */
 578                 i += charlen;
 579                 put_unaligned(dst_char, &target[j]);
 580         }
 581
 582 ctoUTF16_out:
 583         put_unaligned(0, &target[j]); /* Null terminate target unicode string */
 584         kfree(wchar_to);
 585         return j;
 586 }
 587
 588 /*
 589  * cifs_local_to_utf16_bytes - how long will a string be after conversion?
 590  * @from - pointer to input string
 591  * @maxbytes - don't go past this many bytes of input string
 592  * @codepage - source codepage
 593  *
 594  * Walk a string and return the number of bytes that the string will
 595  * be after being converted to the given charset, not including any null
 596  * termination required. Don't walk past maxbytes in the source buffer.
 597  */
 598
 599 static int
 600 cifs_local_to_utf16_bytes(const char *from, int len,
 601                           const struct nls_table *codepage)
 602 {
 603         int charlen;
 604         int i;
 605         wchar_t wchar_to;
 606
 607         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 608                 charlen = codepage->char2uni(from, len, &wchar_to);
 609                 /* Failed conversion defaults to a question mark */
 610                 if (charlen < 1)
 611                         charlen = 1;
 612         }
 613         return 2 * i; /* UTF16 characters are two bytes */
 614 }
 615
 616 /*
 617  * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
 618  * @src - source string
 619  * @maxlen - don't walk past this many bytes in the source string
 620  * @utf16_len - the length of the allocated string in bytes (including null)
 621  * @cp - source codepage
 622  * @remap - map special chars
 623  *
 624  * Take a string convert it from the local codepage to UTF16 and
 625  * put it in a new buffer. Returns a pointer to the new string or NULL on
 626  * error.
 627  */
 628 __le16 *
 629 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
 630                       const struct nls_table *cp, int remap)
 631 {
 632         int len;
 633         __le16 *dst;
 634
 635         len = cifs_local_to_utf16_bytes(src, maxlen, cp);
 636         len += 2; /* NULL */
 637         dst = kmalloc(len, GFP_KERNEL);
 638         if (!dst) {
 639                 *utf16_len = 0;
 640                 return NULL;
 641         }
 642         cifsConvertToUTF16(dst, src, strlen(src), cp, remap);
 643         *utf16_len = len;
 644         return dst;
 645 }