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_SLASH:
 109                 *target = '\\';
 110                 break;
 111         case SFM_SPACE:
 112                 *target = ' ';
 113                 break;
 114         case SFM_PERIOD:
 115                 *target = '.';
 116                 break;
 117         default:
 118                 return false;
 119         }
 120         return true;
 121 }
 122
 123
 124 /*
 125  * cifs_mapchar - convert a host-endian char to proper char in codepage
 126  * @target - where converted character should be copied
 127  * @src_char - 2 byte host-endian source character
 128  * @cp - codepage to which character should be converted
 129  * @map_type - How should the 7 NTFS/SMB reserved characters be mapped to UCS2?
 130  *
 131  * This function handles the conversion of a single character. It is the
 132  * responsibility of the caller to ensure that the target buffer is large
 133  * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
 134  */
 135 static int
 136 cifs_mapchar(char *target, const __u16 *from, const struct nls_table *cp,
 137              int maptype)
 138 {
 139         int len = 1;
 140         __u16 src_char;
 141
 142         src_char = *from;
 143
 144         if ((maptype == SFM_MAP_UNI_RSVD) && convert_sfm_char(src_char, target))
 145                 return len;
 146         else if ((maptype == SFU_MAP_UNI_RSVD) &&
 147                   convert_sfu_char(src_char, target))
 148                 return len;
 149
 150         /* if character not one of seven in special remap set */
 151         len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
 152         if (len <= 0)
 153                 goto surrogate_pair;
 154
 155         return len;
 156
 157 surrogate_pair:
 158         /* convert SURROGATE_PAIR and IVS */
 159         if (strcmp(cp->charset, "utf8"))
 160                 goto unknown;
 161         len = utf16s_to_utf8s(from, 3, UTF16_LITTLE_ENDIAN, target, 6);
 162         if (len <= 0)
 163                 goto unknown;
 164         return len;
 165
 166 unknown:
 167         *target = '?';
 168         len = 1;
 169         return len;
 170 }
 171
 172 /*
 173  * cifs_from_utf16 - convert utf16le string to local charset
 174  * @to - destination buffer
 175  * @from - source buffer
 176  * @tolen - destination buffer size (in bytes)
 177  * @fromlen - source buffer size (in bytes)
 178  * @codepage - codepage to which characters should be converted
 179  * @mapchar - should characters be remapped according to the mapchars option?
 180  *
 181  * Convert a little-endian utf16le string (as sent by the server) to a string
 182  * in the provided codepage. The tolen and fromlen parameters are to ensure
 183  * that the code doesn't walk off of the end of the buffer (which is always
 184  * a danger if the alignment of the source buffer is off). The destination
 185  * string is always properly null terminated and fits in the destination
 186  * buffer. Returns the length of the destination string in bytes (including
 187  * null terminator).
 188  *
 189  * Note that some windows versions actually send multiword UTF-16 characters
 190  * instead of straight UTF16-2. The linux nls routines however aren't able to
 191  * deal with those characters properly. In the event that we get some of
 192  * those characters, they won't be translated properly.
 193  */
 194 int
 195 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
 196                 const struct nls_table *codepage, int map_type)
 197 {
 198         int i, charlen, safelen;
 199         int outlen = 0;
 200         int nullsize = nls_nullsize(codepage);
 201         int fromwords = fromlen / 2;
 202         char tmp[NLS_MAX_CHARSET_SIZE];
 203         __u16 ftmp[3];          /* ftmp[3] = 3array x 2bytes = 6bytes UTF-16 */
 204
 205         /*
 206          * because the chars can be of varying widths, we need to take care
 207          * not to overflow the destination buffer when we get close to the
 208          * end of it. Until we get to this offset, we don't need to check
 209          * for overflow however.
 210          */
 211         safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
 212
 213         for (i = 0; i < fromwords; i++) {
 214                 ftmp[0] = get_unaligned_le16(&from[i]);
 215                 if (ftmp[0] == 0)
 216                         break;
 217                 if (i + 1 < fromwords)
 218                         ftmp[1] = get_unaligned_le16(&from[i + 1]);
 219                 else
 220                         ftmp[1] = 0;
 221                 if (i + 2 < fromwords)
 222                         ftmp[2] = get_unaligned_le16(&from[i + 2]);
 223                 else
 224                         ftmp[2] = 0;
 225
 226                 /*
 227                  * check to see if converting this character might make the
 228                  * conversion bleed into the null terminator
 229                  */
 230                 if (outlen >= safelen) {
 231                         charlen = cifs_mapchar(tmp, ftmp, codepage, map_type);
 232                         if ((outlen + charlen) > (tolen - nullsize))
 233                                 break;
 234                 }
 235
 236                 /* put converted char into 'to' buffer */
 237                 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, map_type);
 238                 outlen += charlen;
 239
 240                 /* charlen (=bytes of UTF-8 for 1 character)
 241                  * 4bytes UTF-8(surrogate pair) is charlen=4
 242                  *   (4bytes UTF-16 code)
 243                  * 7-8bytes UTF-8(IVS) is charlen=3+4 or 4+4
 244                  *   (2 UTF-8 pairs divided to 2 UTF-16 pairs) */
 245                 if (charlen == 4)
 246                         i++;
 247                 else if (charlen >= 5)
 248                         /* 5-6bytes UTF-8 */
 249                         i += 2;
 250         }
 251
 252         /* properly null-terminate string */
 253         for (i = 0; i < nullsize; i++)
 254                 to[outlen++] = 0;
 255
 256         return outlen;
 257 }
 258
 259 /*
 260  * NAME:        cifs_strtoUTF16()
 261  *
 262  * FUNCTION:    Convert character string to unicode string
 263  *
 264  */
 265 int
 266 cifs_strtoUTF16(__le16 *to, const char *from, int len,
 267               const struct nls_table *codepage)
 268 {
 269         int charlen;
 270         int i;
 271         wchar_t wchar_to; /* needed to quiet sparse */
 272
 273         /* special case for utf8 to handle no plane0 chars */
 274         if (!strcmp(codepage->charset, "utf8")) {
 275                 /*
 276                  * convert utf8 -> utf16, we assume we have enough space
 277                  * as caller should have assumed conversion does not overflow
 278                  * in destination len is length in wchar_t units (16bits)
 279                  */
 280                 i  = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
 281                                        (wchar_t *) to, len);
 282
 283                 /* if success terminate and exit */
 284                 if (i >= 0)
 285                         goto success;
 286                 /*
 287                  * if fails fall back to UCS encoding as this
 288                  * function should not return negative values
 289                  * currently can fail only if source contains
 290                  * invalid encoded characters
 291                  */
 292         }
 293
 294         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 295                 charlen = codepage->char2uni(from, len, &wchar_to);
 296                 if (charlen < 1) {
 297                         cifs_dbg(VFS, "strtoUTF16: char2uni of 0x%x returned %d\n",
 298                                  *from, charlen);
 299                         /* A question mark */
 300                         wchar_to = 0x003f;
 301                         charlen = 1;
 302                 }
 303                 put_unaligned_le16(wchar_to, &to[i]);
 304         }
 305
 306 success:
 307         put_unaligned_le16(0, &to[i]);
 308         return i;
 309 }
 310
 311 /*
 312  * cifs_utf16_bytes - how long will a string be after conversion?
 313  * @utf16 - pointer to input string
 314  * @maxbytes - don't go past this many bytes of input string
 315  * @codepage - destination codepage
 316  *
 317  * Walk a utf16le string and return the number of bytes that the string will
 318  * be after being converted to the given charset, not including any null
 319  * termination required. Don't walk past maxbytes in the source buffer.
 320  */
 321 int
 322 cifs_utf16_bytes(const __le16 *from, int maxbytes,
 323                 const struct nls_table *codepage)
 324 {
 325         int i;
 326         int charlen, outlen = 0;
 327         int maxwords = maxbytes / 2;
 328         char tmp[NLS_MAX_CHARSET_SIZE];
 329         __u16 ftmp[3];
 330
 331         for (i = 0; i < maxwords; i++) {
 332                 ftmp[0] = get_unaligned_le16(&from[i]);
 333                 if (ftmp[0] == 0)
 334                         break;
 335                 if (i + 1 < maxwords)
 336                         ftmp[1] = get_unaligned_le16(&from[i + 1]);
 337                 else
 338                         ftmp[1] = 0;
 339                 if (i + 2 < maxwords)
 340                         ftmp[2] = get_unaligned_le16(&from[i + 2]);
 341                 else
 342                         ftmp[2] = 0;
 343
 344                 charlen = cifs_mapchar(tmp, ftmp, codepage, NO_MAP_UNI_RSVD);
 345                 outlen += charlen;
 346         }
 347
 348         return outlen;
 349 }
 350
 351 /*
 352  * cifs_strndup_from_utf16 - copy a string from wire format to the local
 353  * codepage
 354  * @src - source string
 355  * @maxlen - don't walk past this many bytes in the source string
 356  * @is_unicode - is this a unicode string?
 357  * @codepage - destination codepage
 358  *
 359  * Take a string given by the server, convert it to the local codepage and
 360  * put it in a new buffer. Returns a pointer to the new string or NULL on
 361  * error.
 362  */
 363 char *
 364 cifs_strndup_from_utf16(const char *src, const int maxlen,
 365                         const bool is_unicode, const struct nls_table *codepage)
 366 {
 367         int len;
 368         char *dst;
 369
 370         if (is_unicode) {
 371                 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
 372                 len += nls_nullsize(codepage);
 373                 dst = kmalloc(len, GFP_KERNEL);
 374                 if (!dst)
 375                         return NULL;
 376                 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
 377                                NO_MAP_UNI_RSVD);
 378         } else {
 379                 len = strnlen(src, maxlen);
 380                 len++;
 381                 dst = kmalloc(len, GFP_KERNEL);
 382                 if (!dst)
 383                         return NULL;
 384                 strlcpy(dst, src, len);
 385         }
 386
 387         return dst;
 388 }
 389
 390 static __le16 convert_to_sfu_char(char src_char)
 391 {
 392         __le16 dest_char;
 393
 394         switch (src_char) {
 395         case ':':
 396                 dest_char = cpu_to_le16(UNI_COLON);
 397                 break;
 398         case '*':
 399                 dest_char = cpu_to_le16(UNI_ASTERISK);
 400                 break;
 401         case '?':
 402                 dest_char = cpu_to_le16(UNI_QUESTION);
 403                 break;
 404         case '<':
 405                 dest_char = cpu_to_le16(UNI_LESSTHAN);
 406                 break;
 407         case '>':
 408                 dest_char = cpu_to_le16(UNI_GRTRTHAN);
 409                 break;
 410         case '|':
 411                 dest_char = cpu_to_le16(UNI_PIPE);
 412                 break;
 413         default:
 414                 dest_char = 0;
 415         }
 416
 417         return dest_char;
 418 }
 419
 420 static __le16 convert_to_sfm_char(char src_char, bool end_of_string)
 421 {
 422         __le16 dest_char;
 423
 424         if (src_char >= 0x01 && src_char <= 0x1F) {
 425                 dest_char = cpu_to_le16(src_char + 0xF000);
 426                 return dest_char;
 427         }
 428         switch (src_char) {
 429         case ':':
 430                 dest_char = cpu_to_le16(SFM_COLON);
 431                 break;
 432         case '"':
 433                 dest_char = cpu_to_le16(SFM_DOUBLEQUOTE);
 434                 break;
 435         case '*':
 436                 dest_char = cpu_to_le16(SFM_ASTERISK);
 437                 break;
 438         case '?':
 439                 dest_char = cpu_to_le16(SFM_QUESTION);
 440                 break;
 441         case '<':
 442                 dest_char = cpu_to_le16(SFM_LESSTHAN);
 443                 break;
 444         case '>':
 445                 dest_char = cpu_to_le16(SFM_GRTRTHAN);
 446                 break;
 447         case '|':
 448                 dest_char = cpu_to_le16(SFM_PIPE);
 449                 break;
 450         case '.':
 451                 if (end_of_string)
 452                         dest_char = cpu_to_le16(SFM_PERIOD);
 453                 else
 454                         dest_char = 0;
 455                 break;
 456         case ' ':
 457                 if (end_of_string)
 458                         dest_char = cpu_to_le16(SFM_SPACE);
 459                 else
 460                         dest_char = 0;
 461                 break;
 462         default:
 463                 dest_char = 0;
 464         }
 465
 466         return dest_char;
 467 }
 468
 469 /*
 470  * Convert 16 bit Unicode pathname to wire format from string in current code
 471  * page. Conversion may involve remapping up the six characters that are
 472  * only legal in POSIX-like OS (if they are present in the string). Path
 473  * names are little endian 16 bit Unicode on the wire
 474  */
 475 int
 476 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
 477                  const struct nls_table *cp, int map_chars)
 478 {
 479         int i, charlen;
 480         int j = 0;
 481         char src_char;
 482         __le16 dst_char;
 483         wchar_t tmp;
 484         wchar_t *wchar_to;      /* UTF-16 */
 485         int ret;
 486         unicode_t u;
 487
 488         if (map_chars == NO_MAP_UNI_RSVD)
 489                 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
 490
 491         wchar_to = kzalloc(6, GFP_KERNEL);
 492
 493         for (i = 0; i < srclen; j++) {
 494                 src_char = source[i];
 495                 charlen = 1;
 496
 497                 /* check if end of string */
 498                 if (src_char == 0)
 499                         goto ctoUTF16_out;
 500
 501                 /* see if we must remap this char */
 502                 if (map_chars == SFU_MAP_UNI_RSVD)
 503                         dst_char = convert_to_sfu_char(src_char);
 504                 else if (map_chars == SFM_MAP_UNI_RSVD) {
 505                         bool end_of_string;
 506
 507                         if (i == srclen - 1)
 508                                 end_of_string = true;
 509                         else
 510                                 end_of_string = false;
 511
 512                         dst_char = convert_to_sfm_char(src_char, end_of_string);
 513                 } else
 514                         dst_char = 0;
 515                 /*
 516                  * FIXME: We can not handle remapping backslash (UNI_SLASH)
 517                  * until all the calls to build_path_from_dentry are modified,
 518                  * as they use backslash as separator.
 519                  */
 520                 if (dst_char == 0) {
 521                         charlen = cp->char2uni(source + i, srclen - i, &tmp);
 522                         dst_char = cpu_to_le16(tmp);
 523
 524                         /*
 525                          * if no match, use question mark, which at least in
 526                          * some cases serves as wild card
 527                          */
 528                         if (charlen > 0)
 529                                 goto ctoUTF16;
 530
 531                         /* convert SURROGATE_PAIR */
 532                         if (strcmp(cp->charset, "utf8") || !wchar_to)
 533                                 goto unknown;
 534                         if (*(source + i) & 0x80) {
 535                                 charlen = utf8_to_utf32(source + i, 6, &u);
 536                                 if (charlen < 0)
 537                                         goto unknown;
 538                         } else
 539                                 goto unknown;
 540                         ret  = utf8s_to_utf16s(source + i, charlen,
 541                                                UTF16_LITTLE_ENDIAN,
 542                                                wchar_to, 6);
 543                         if (ret < 0)
 544                                 goto unknown;
 545
 546                         i += charlen;
 547                         dst_char = cpu_to_le16(*wchar_to);
 548                         if (charlen <= 3)
 549                                 /* 1-3bytes UTF-8 to 2bytes UTF-16 */
 550                                 put_unaligned(dst_char, &target[j]);
 551                         else if (charlen == 4) {
 552                                 /* 4bytes UTF-8(surrogate pair) to 4bytes UTF-16
 553                                  * 7-8bytes UTF-8(IVS) divided to 2 UTF-16
 554                                  *   (charlen=3+4 or 4+4) */
 555                                 put_unaligned(dst_char, &target[j]);
 556                                 dst_char = cpu_to_le16(*(wchar_to + 1));
 557                                 j++;
 558                                 put_unaligned(dst_char, &target[j]);
 559                         } else if (charlen >= 5) {
 560                                 /* 5-6bytes UTF-8 to 6bytes UTF-16 */
 561                                 put_unaligned(dst_char, &target[j]);
 562                                 dst_char = cpu_to_le16(*(wchar_to + 1));
 563                                 j++;
 564                                 put_unaligned(dst_char, &target[j]);
 565                                 dst_char = cpu_to_le16(*(wchar_to + 2));
 566                                 j++;
 567                                 put_unaligned(dst_char, &target[j]);
 568                         }
 569                         continue;
 570
 571 unknown:
 572                         dst_char = cpu_to_le16(0x003f);
 573                         charlen = 1;
 574                 }
 575
 576 ctoUTF16:
 577                 /*
 578                  * character may take more than one byte in the source string,
 579                  * but will take exactly two bytes in the target string
 580                  */
 581                 i += charlen;
 582                 put_unaligned(dst_char, &target[j]);
 583         }
 584
 585 ctoUTF16_out:
 586         put_unaligned(0, &target[j]); /* Null terminate target unicode string */
 587         kfree(wchar_to);
 588         return j;
 589 }
 590
 591 /*
 592  * cifs_local_to_utf16_bytes - how long will a string be after conversion?
 593  * @from - pointer to input string
 594  * @maxbytes - don't go past this many bytes of input string
 595  * @codepage - source codepage
 596  *
 597  * Walk a string and return the number of bytes that the string will
 598  * be after being converted to the given charset, not including any null
 599  * termination required. Don't walk past maxbytes in the source buffer.
 600  */
 601
 602 static int
 603 cifs_local_to_utf16_bytes(const char *from, int len,
 604                           const struct nls_table *codepage)
 605 {
 606         int charlen;
 607         int i;
 608         wchar_t wchar_to;
 609
 610         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 611                 charlen = codepage->char2uni(from, len, &wchar_to);
 612                 /* Failed conversion defaults to a question mark */
 613                 if (charlen < 1)
 614                         charlen = 1;
 615         }
 616         return 2 * i; /* UTF16 characters are two bytes */
 617 }
 618
 619 /*
 620  * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
 621  * @src - source string
 622  * @maxlen - don't walk past this many bytes in the source string
 623  * @utf16_len - the length of the allocated string in bytes (including null)
 624  * @cp - source codepage
 625  * @remap - map special chars
 626  *
 627  * Take a string convert it from the local codepage to UTF16 and
 628  * put it in a new buffer. Returns a pointer to the new string or NULL on
 629  * error.
 630  */
 631 __le16 *
 632 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
 633                       const struct nls_table *cp, int remap)
 634 {
 635         int len;
 636         __le16 *dst;
 637
 638         len = cifs_local_to_utf16_bytes(src, maxlen, cp);
 639         len += 2; /* NULL */
 640         dst = kmalloc(len, GFP_KERNEL);
 641         if (!dst) {
 642                 *utf16_len = 0;
 643                 return NULL;
 644         }
 645         cifsConvertToUTF16(dst, src, strlen(src), cp, remap);
 646         *utf16_len = len;
 647         return dst;
 648 }