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         switch (src_char) {
  83         case SFM_COLON:
  84                 *target = ':';
  85                 break;
  86         case SFM_DOUBLEQUOTE:
  87                 *target = '"';
  88                 break;
  89         case SFM_ASTERISK:
  90                 *target = '*';
  91                 break;
  92         case SFM_QUESTION:
  93                 *target = '?';
  94                 break;
  95         case SFM_PIPE:
  96                 *target = '|';
  97                 break;
  98         case SFM_GRTRTHAN:
  99                 *target = '>';
 100                 break;
 101         case SFM_LESSTHAN:
 102                 *target = '<';
 103                 break;
 104         case SFM_SLASH:
 105                 *target = '\\';
 106                 break;
 107         case SFM_SPACE:
 108                 *target = ' ';
 109                 break;
 110         case SFM_PERIOD:
 111                 *target = '.';
 112                 break;
 113         default:
 114                 return false;
 115         }
 116         return true;
 117 }
 118
 119
 120 /*
 121  * cifs_mapchar - convert a host-endian char to proper char in codepage
 122  * @target - where converted character should be copied
 123  * @src_char - 2 byte host-endian source character
 124  * @cp - codepage to which character should be converted
 125  * @map_type - How should the 7 NTFS/SMB reserved characters be mapped to UCS2?
 126  *
 127  * This function handles the conversion of a single character. It is the
 128  * responsibility of the caller to ensure that the target buffer is large
 129  * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
 130  */
 131 static int
 132 cifs_mapchar(char *target, const __u16 *from, const struct nls_table *cp,
 133              int maptype)
 134 {
 135         int len = 1;
 136         __u16 src_char;
 137
 138         src_char = *from;
 139
 140         if ((maptype == SFM_MAP_UNI_RSVD) && convert_sfm_char(src_char, target))
 141                 return len;
 142         else if ((maptype == SFU_MAP_UNI_RSVD) &&
 143                   convert_sfu_char(src_char, target))
 144                 return len;
 145
 146         /* if character not one of seven in special remap set */
 147         len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
 148         if (len <= 0)
 149                 goto surrogate_pair;
 150
 151         return len;
 152
 153 surrogate_pair:
 154         /* convert SURROGATE_PAIR and IVS */
 155         if (strcmp(cp->charset, "utf8"))
 156                 goto unknown;
 157         len = utf16s_to_utf8s(from, 3, UTF16_LITTLE_ENDIAN, target, 6);
 158         if (len <= 0)
 159                 goto unknown;
 160         return len;
 161
 162 unknown:
 163         *target = '?';
 164         len = 1;
 165         return len;
 166 }
 167
 168 /*
 169  * cifs_from_utf16 - convert utf16le string to local charset
 170  * @to - destination buffer
 171  * @from - source buffer
 172  * @tolen - destination buffer size (in bytes)
 173  * @fromlen - source buffer size (in bytes)
 174  * @codepage - codepage to which characters should be converted
 175  * @mapchar - should characters be remapped according to the mapchars option?
 176  *
 177  * Convert a little-endian utf16le string (as sent by the server) to a string
 178  * in the provided codepage. The tolen and fromlen parameters are to ensure
 179  * that the code doesn't walk off of the end of the buffer (which is always
 180  * a danger if the alignment of the source buffer is off). The destination
 181  * string is always properly null terminated and fits in the destination
 182  * buffer. Returns the length of the destination string in bytes (including
 183  * null terminator).
 184  *
 185  * Note that some windows versions actually send multiword UTF-16 characters
 186  * instead of straight UTF16-2. The linux nls routines however aren't able to
 187  * deal with those characters properly. In the event that we get some of
 188  * those characters, they won't be translated properly.
 189  */
 190 int
 191 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
 192                 const struct nls_table *codepage, int map_type)
 193 {
 194         int i, charlen, safelen;
 195         int outlen = 0;
 196         int nullsize = nls_nullsize(codepage);
 197         int fromwords = fromlen / 2;
 198         char tmp[NLS_MAX_CHARSET_SIZE];
 199         __u16 ftmp[3];          /* ftmp[3] = 3array x 2bytes = 6bytes UTF-16 */
 200
 201         /*
 202          * because the chars can be of varying widths, we need to take care
 203          * not to overflow the destination buffer when we get close to the
 204          * end of it. Until we get to this offset, we don't need to check
 205          * for overflow however.
 206          */
 207         safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
 208
 209         for (i = 0; i < fromwords; i++) {
 210                 ftmp[0] = get_unaligned_le16(&from[i]);
 211                 if (ftmp[0] == 0)
 212                         break;
 213                 if (i + 1 < fromwords)
 214                         ftmp[1] = get_unaligned_le16(&from[i + 1]);
 215                 else
 216                         ftmp[1] = 0;
 217                 if (i + 2 < fromwords)
 218                         ftmp[2] = get_unaligned_le16(&from[i + 2]);
 219                 else
 220                         ftmp[2] = 0;
 221
 222                 /*
 223                  * check to see if converting this character might make the
 224                  * conversion bleed into the null terminator
 225                  */
 226                 if (outlen >= safelen) {
 227                         charlen = cifs_mapchar(tmp, ftmp, codepage, map_type);
 228                         if ((outlen + charlen) > (tolen - nullsize))
 229                                 break;
 230                 }
 231
 232                 /* put converted char into 'to' buffer */
 233                 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, map_type);
 234                 outlen += charlen;
 235
 236                 /* charlen (=bytes of UTF-8 for 1 character)
 237                  * 4bytes UTF-8(surrogate pair) is charlen=4
 238                  *   (4bytes UTF-16 code)
 239                  * 7-8bytes UTF-8(IVS) is charlen=3+4 or 4+4
 240                  *   (2 UTF-8 pairs divided to 2 UTF-16 pairs) */
 241                 if (charlen == 4)
 242                         i++;
 243                 else if (charlen >= 5)
 244                         /* 5-6bytes UTF-8 */
 245                         i += 2;
 246         }
 247
 248         /* properly null-terminate string */
 249         for (i = 0; i < nullsize; i++)
 250                 to[outlen++] = 0;
 251
 252         return outlen;
 253 }
 254
 255 /*
 256  * NAME:        cifs_strtoUTF16()
 257  *
 258  * FUNCTION:    Convert character string to unicode string
 259  *
 260  */
 261 int
 262 cifs_strtoUTF16(__le16 *to, const char *from, int len,
 263               const struct nls_table *codepage)
 264 {
 265         int charlen;
 266         int i;
 267         wchar_t wchar_to; /* needed to quiet sparse */
 268
 269         /* special case for utf8 to handle no plane0 chars */
 270         if (!strcmp(codepage->charset, "utf8")) {
 271                 /*
 272                  * convert utf8 -> utf16, we assume we have enough space
 273                  * as caller should have assumed conversion does not overflow
 274                  * in destination len is length in wchar_t units (16bits)
 275                  */
 276                 i  = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
 277                                        (wchar_t *) to, len);
 278
 279                 /* if success terminate and exit */
 280                 if (i >= 0)
 281                         goto success;
 282                 /*
 283                  * if fails fall back to UCS encoding as this
 284                  * function should not return negative values
 285                  * currently can fail only if source contains
 286                  * invalid encoded characters
 287                  */
 288         }
 289
 290         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 291                 charlen = codepage->char2uni(from, len, &wchar_to);
 292                 if (charlen < 1) {
 293                         cifs_dbg(VFS, "strtoUTF16: char2uni of 0x%x returned %d\n",
 294                                  *from, charlen);
 295                         /* A question mark */
 296                         wchar_to = 0x003f;
 297                         charlen = 1;
 298                 }
 299                 put_unaligned_le16(wchar_to, &to[i]);
 300         }
 301
 302 success:
 303         put_unaligned_le16(0, &to[i]);
 304         return i;
 305 }
 306
 307 /*
 308  * cifs_utf16_bytes - how long will a string be after conversion?
 309  * @utf16 - pointer to input string
 310  * @maxbytes - don't go past this many bytes of input string
 311  * @codepage - destination codepage
 312  *
 313  * Walk a utf16le string and return the number of bytes that the string will
 314  * be after being converted to the given charset, not including any null
 315  * termination required. Don't walk past maxbytes in the source buffer.
 316  */
 317 int
 318 cifs_utf16_bytes(const __le16 *from, int maxbytes,
 319                 const struct nls_table *codepage)
 320 {
 321         int i;
 322         int charlen, outlen = 0;
 323         int maxwords = maxbytes / 2;
 324         char tmp[NLS_MAX_CHARSET_SIZE];
 325         __u16 ftmp[3];
 326
 327         for (i = 0; i < maxwords; i++) {
 328                 ftmp[0] = get_unaligned_le16(&from[i]);
 329                 if (ftmp[0] == 0)
 330                         break;
 331                 if (i + 1 < maxwords)
 332                         ftmp[1] = get_unaligned_le16(&from[i + 1]);
 333                 else
 334                         ftmp[1] = 0;
 335                 if (i + 2 < maxwords)
 336                         ftmp[2] = get_unaligned_le16(&from[i + 2]);
 337                 else
 338                         ftmp[2] = 0;
 339
 340                 charlen = cifs_mapchar(tmp, ftmp, codepage, NO_MAP_UNI_RSVD);
 341                 outlen += charlen;
 342         }
 343
 344         return outlen;
 345 }
 346
 347 /*
 348  * cifs_strndup_from_utf16 - copy a string from wire format to the local
 349  * codepage
 350  * @src - source string
 351  * @maxlen - don't walk past this many bytes in the source string
 352  * @is_unicode - is this a unicode string?
 353  * @codepage - destination codepage
 354  *
 355  * Take a string given by the server, convert it to the local codepage and
 356  * put it in a new buffer. Returns a pointer to the new string or NULL on
 357  * error.
 358  */
 359 char *
 360 cifs_strndup_from_utf16(const char *src, const int maxlen,
 361                         const bool is_unicode, const struct nls_table *codepage)
 362 {
 363         int len;
 364         char *dst;
 365
 366         if (is_unicode) {
 367                 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
 368                 len += nls_nullsize(codepage);
 369                 dst = kmalloc(len, GFP_KERNEL);
 370                 if (!dst)
 371                         return NULL;
 372                 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
 373                                NO_MAP_UNI_RSVD);
 374         } else {
 375                 len = strnlen(src, maxlen);
 376                 len++;
 377                 dst = kmalloc(len, GFP_KERNEL);
 378                 if (!dst)
 379                         return NULL;
 380                 strlcpy(dst, src, len);
 381         }
 382
 383         return dst;
 384 }
 385
 386 static __le16 convert_to_sfu_char(char src_char)
 387 {
 388         __le16 dest_char;
 389
 390         switch (src_char) {
 391         case ':':
 392                 dest_char = cpu_to_le16(UNI_COLON);
 393                 break;
 394         case '*':
 395                 dest_char = cpu_to_le16(UNI_ASTERISK);
 396                 break;
 397         case '?':
 398                 dest_char = cpu_to_le16(UNI_QUESTION);
 399                 break;
 400         case '<':
 401                 dest_char = cpu_to_le16(UNI_LESSTHAN);
 402                 break;
 403         case '>':
 404                 dest_char = cpu_to_le16(UNI_GRTRTHAN);
 405                 break;
 406         case '|':
 407                 dest_char = cpu_to_le16(UNI_PIPE);
 408                 break;
 409         default:
 410                 dest_char = 0;
 411         }
 412
 413         return dest_char;
 414 }
 415
 416 static __le16 convert_to_sfm_char(char src_char, bool end_of_string)
 417 {
 418         __le16 dest_char;
 419
 420         switch (src_char) {
 421         case ':':
 422                 dest_char = cpu_to_le16(SFM_COLON);
 423                 break;
 424         case '"':
 425                 dest_char = cpu_to_le16(SFM_DOUBLEQUOTE);
 426                 break;
 427         case '*':
 428                 dest_char = cpu_to_le16(SFM_ASTERISK);
 429                 break;
 430         case '?':
 431                 dest_char = cpu_to_le16(SFM_QUESTION);
 432                 break;
 433         case '<':
 434                 dest_char = cpu_to_le16(SFM_LESSTHAN);
 435                 break;
 436         case '>':
 437                 dest_char = cpu_to_le16(SFM_GRTRTHAN);
 438                 break;
 439         case '|':
 440                 dest_char = cpu_to_le16(SFM_PIPE);
 441                 break;
 442         case '.':
 443                 if (end_of_string)
 444                         dest_char = cpu_to_le16(SFM_PERIOD);
 445                 else
 446                         dest_char = 0;
 447                 break;
 448         case ' ':
 449                 if (end_of_string)
 450                         dest_char = cpu_to_le16(SFM_SPACE);
 451                 else
 452                         dest_char = 0;
 453                 break;
 454         default:
 455                 dest_char = 0;
 456         }
 457
 458         return dest_char;
 459 }
 460
 461 /*
 462  * Convert 16 bit Unicode pathname to wire format from string in current code
 463  * page. Conversion may involve remapping up the six characters that are
 464  * only legal in POSIX-like OS (if they are present in the string). Path
 465  * names are little endian 16 bit Unicode on the wire
 466  */
 467 int
 468 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
 469                  const struct nls_table *cp, int map_chars)
 470 {
 471         int i, charlen;
 472         int j = 0;
 473         char src_char;
 474         __le16 dst_char;
 475         wchar_t tmp;
 476         wchar_t *wchar_to;      /* UTF-16 */
 477         int ret;
 478         unicode_t u;
 479
 480         if (map_chars == NO_MAP_UNI_RSVD)
 481                 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
 482
 483         wchar_to = kzalloc(6, GFP_KERNEL);
 484
 485         for (i = 0; i < srclen; j++) {
 486                 src_char = source[i];
 487                 charlen = 1;
 488
 489                 /* check if end of string */
 490                 if (src_char == 0)
 491                         goto ctoUTF16_out;
 492
 493                 /* see if we must remap this char */
 494                 if (map_chars == SFU_MAP_UNI_RSVD)
 495                         dst_char = convert_to_sfu_char(src_char);
 496                 else if (map_chars == SFM_MAP_UNI_RSVD) {
 497                         bool end_of_string;
 498
 499                         if (i == srclen - 1)
 500                                 end_of_string = true;
 501                         else
 502                                 end_of_string = false;
 503
 504                         dst_char = convert_to_sfm_char(src_char, end_of_string);
 505                 } else
 506                         dst_char = 0;
 507                 /*
 508                  * FIXME: We can not handle remapping backslash (UNI_SLASH)
 509                  * until all the calls to build_path_from_dentry are modified,
 510                  * as they use backslash as separator.
 511                  */
 512                 if (dst_char == 0) {
 513                         charlen = cp->char2uni(source + i, srclen - i, &tmp);
 514                         dst_char = cpu_to_le16(tmp);
 515
 516                         /*
 517                          * if no match, use question mark, which at least in
 518                          * some cases serves as wild card
 519                          */
 520                         if (charlen > 0)
 521                                 goto ctoUTF16;
 522
 523                         /* convert SURROGATE_PAIR */
 524                         if (strcmp(cp->charset, "utf8") || !wchar_to)
 525                                 goto unknown;
 526                         if (*(source + i) & 0x80) {
 527                                 charlen = utf8_to_utf32(source + i, 6, &u);
 528                                 if (charlen < 0)
 529                                         goto unknown;
 530                         } else
 531                                 goto unknown;
 532                         ret  = utf8s_to_utf16s(source + i, charlen,
 533                                                UTF16_LITTLE_ENDIAN,
 534                                                wchar_to, 6);
 535                         if (ret < 0)
 536                                 goto unknown;
 537
 538                         i += charlen;
 539                         dst_char = cpu_to_le16(*wchar_to);
 540                         if (charlen <= 3)
 541                                 /* 1-3bytes UTF-8 to 2bytes UTF-16 */
 542                                 put_unaligned(dst_char, &target[j]);
 543                         else if (charlen == 4) {
 544                                 /* 4bytes UTF-8(surrogate pair) to 4bytes UTF-16
 545                                  * 7-8bytes UTF-8(IVS) divided to 2 UTF-16
 546                                  *   (charlen=3+4 or 4+4) */
 547                                 put_unaligned(dst_char, &target[j]);
 548                                 dst_char = cpu_to_le16(*(wchar_to + 1));
 549                                 j++;
 550                                 put_unaligned(dst_char, &target[j]);
 551                         } else if (charlen >= 5) {
 552                                 /* 5-6bytes UTF-8 to 6bytes UTF-16 */
 553                                 put_unaligned(dst_char, &target[j]);
 554                                 dst_char = cpu_to_le16(*(wchar_to + 1));
 555                                 j++;
 556                                 put_unaligned(dst_char, &target[j]);
 557                                 dst_char = cpu_to_le16(*(wchar_to + 2));
 558                                 j++;
 559                                 put_unaligned(dst_char, &target[j]);
 560                         }
 561                         continue;
 562
 563 unknown:
 564                         dst_char = cpu_to_le16(0x003f);
 565                         charlen = 1;
 566                 }
 567
 568 ctoUTF16:
 569                 /*
 570                  * character may take more than one byte in the source string,
 571                  * but will take exactly two bytes in the target string
 572                  */
 573                 i += charlen;
 574                 put_unaligned(dst_char, &target[j]);
 575         }
 576
 577 ctoUTF16_out:
 578         put_unaligned(0, &target[j]); /* Null terminate target unicode string */
 579         kfree(wchar_to);
 580         return j;
 581 }
 582
 583 #ifdef CONFIG_CIFS_SMB2
 584 /*
 585  * cifs_local_to_utf16_bytes - how long will a string be after conversion?
 586  * @from - pointer to input string
 587  * @maxbytes - don't go past this many bytes of input string
 588  * @codepage - source codepage
 589  *
 590  * Walk a string and return the number of bytes that the string will
 591  * be after being converted to the given charset, not including any null
 592  * termination required. Don't walk past maxbytes in the source buffer.
 593  */
 594
 595 static int
 596 cifs_local_to_utf16_bytes(const char *from, int len,
 597                           const struct nls_table *codepage)
 598 {
 599         int charlen;
 600         int i;
 601         wchar_t wchar_to;
 602
 603         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 604                 charlen = codepage->char2uni(from, len, &wchar_to);
 605                 /* Failed conversion defaults to a question mark */
 606                 if (charlen < 1)
 607                         charlen = 1;
 608         }
 609         return 2 * i; /* UTF16 characters are two bytes */
 610 }
 611
 612 /*
 613  * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
 614  * @src - source string
 615  * @maxlen - don't walk past this many bytes in the source string
 616  * @utf16_len - the length of the allocated string in bytes (including null)
 617  * @cp - source codepage
 618  * @remap - map special chars
 619  *
 620  * Take a string convert it from the local codepage to UTF16 and
 621  * put it in a new buffer. Returns a pointer to the new string or NULL on
 622  * error.
 623  */
 624 __le16 *
 625 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
 626                       const struct nls_table *cp, int remap)
 627 {
 628         int len;
 629         __le16 *dst;
 630
 631         len = cifs_local_to_utf16_bytes(src, maxlen, cp);
 632         len += 2; /* NULL */
 633         dst = kmalloc(len, GFP_KERNEL);
 634         if (!dst) {
 635                 *utf16_len = 0;
 636                 return NULL;
 637         }
 638         cifsConvertToUTF16(dst, src, strlen(src), cp, remap);
 639         *utf16_len = len;
 640         return dst;
 641 }
 642 #endif /* CONFIG_CIFS_SMB2 */