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_unicode.h"
  24 #include "cifs_uniupr.h"
  25 #include "cifspdu.h"
  26 #include "cifsglob.h"
  27 #include "cifs_debug.h"
  28
  29 /*
  30  * cifs_utf16_bytes - how long will a string be after conversion?
  31  * @utf16 - pointer to input string
  32  * @maxbytes - don't go past this many bytes of input string
  33  * @codepage - destination codepage
  34  *
  35  * Walk a utf16le string and return the number of bytes that the string will
  36  * be after being converted to the given charset, not including any null
  37  * termination required. Don't walk past maxbytes in the source buffer.
  38  */
  39 int
  40 cifs_utf16_bytes(const __le16 *from, int maxbytes,
  41                 const struct nls_table *codepage)
  42 {
  43         int i;
  44         int charlen, outlen = 0;
  45         int maxwords = maxbytes / 2;
  46         char tmp[NLS_MAX_CHARSET_SIZE];
  47         __u16 ftmp;
  48
  49         for (i = 0; i < maxwords; i++) {
  50                 ftmp = get_unaligned_le16(&from[i]);
  51                 if (ftmp == 0)
  52                         break;
  53
  54                 charlen = codepage->uni2char(ftmp, tmp, NLS_MAX_CHARSET_SIZE);
  55                 if (charlen > 0)
  56                         outlen += charlen;
  57                 else
  58                         outlen++;
  59         }
  60
  61         return outlen;
  62 }
  63
  64 /*
  65  * cifs_mapchar - convert a host-endian char to proper char in codepage
  66  * @target - where converted character should be copied
  67  * @src_char - 2 byte host-endian source character
  68  * @cp - codepage to which character should be converted
  69  * @mapchar - should character be mapped according to mapchars mount option?
  70  *
  71  * This function handles the conversion of a single character. It is the
  72  * responsibility of the caller to ensure that the target buffer is large
  73  * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
  74  */
  75 static int
  76 cifs_mapchar(char *target, const __u16 src_char, const struct nls_table *cp,
  77              bool mapchar)
  78 {
  79         int len = 1;
  80
  81         if (!mapchar)
  82                 goto cp_convert;
  83
  84         /*
  85          * BB: Cannot handle remapping UNI_SLASH until all the calls to
  86          *     build_path_from_dentry are modified, as they use slash as
  87          *     separator.
  88          */
  89         switch (src_char) {
  90         case UNI_COLON:
  91                 *target = ':';
  92                 break;
  93         case UNI_ASTERISK:
  94                 *target = '*';
  95                 break;
  96         case UNI_QUESTION:
  97                 *target = '?';
  98                 break;
  99         case UNI_PIPE:
 100                 *target = '|';
 101                 break;
 102         case UNI_GRTRTHAN:
 103                 *target = '>';
 104                 break;
 105         case UNI_LESSTHAN:
 106                 *target = '<';
 107                 break;
 108         default:
 109                 goto cp_convert;
 110         }
 111
 112 out:
 113         return len;
 114
 115 cp_convert:
 116         len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
 117         if (len <= 0) {
 118                 *target = '?';
 119                 len = 1;
 120         }
 121         goto out;
 122 }
 123
 124 /*
 125  * cifs_from_utf16 - convert utf16le string to local charset
 126  * @to - destination buffer
 127  * @from - source buffer
 128  * @tolen - destination buffer size (in bytes)
 129  * @fromlen - source buffer size (in bytes)
 130  * @codepage - codepage to which characters should be converted
 131  * @mapchar - should characters be remapped according to the mapchars option?
 132  *
 133  * Convert a little-endian utf16le string (as sent by the server) to a string
 134  * in the provided codepage. The tolen and fromlen parameters are to ensure
 135  * that the code doesn't walk off of the end of the buffer (which is always
 136  * a danger if the alignment of the source buffer is off). The destination
 137  * string is always properly null terminated and fits in the destination
 138  * buffer. Returns the length of the destination string in bytes (including
 139  * null terminator).
 140  *
 141  * Note that some windows versions actually send multiword UTF-16 characters
 142  * instead of straight UTF16-2. The linux nls routines however aren't able to
 143  * deal with those characters properly. In the event that we get some of
 144  * those characters, they won't be translated properly.
 145  */
 146 int
 147 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
 148                  const struct nls_table *codepage, bool mapchar)
 149 {
 150         int i, charlen, safelen;
 151         int outlen = 0;
 152         int nullsize = nls_nullsize(codepage);
 153         int fromwords = fromlen / 2;
 154         char tmp[NLS_MAX_CHARSET_SIZE];
 155         __u16 ftmp;
 156
 157         /*
 158          * because the chars can be of varying widths, we need to take care
 159          * not to overflow the destination buffer when we get close to the
 160          * end of it. Until we get to this offset, we don't need to check
 161          * for overflow however.
 162          */
 163         safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
 164
 165         for (i = 0; i < fromwords; i++) {
 166                 ftmp = get_unaligned_le16(&from[i]);
 167                 if (ftmp == 0)
 168                         break;
 169
 170                 /*
 171                  * check to see if converting this character might make the
 172                  * conversion bleed into the null terminator
 173                  */
 174                 if (outlen >= safelen) {
 175                         charlen = cifs_mapchar(tmp, ftmp, codepage, mapchar);
 176                         if ((outlen + charlen) > (tolen - nullsize))
 177                                 break;
 178                 }
 179
 180                 /* put converted char into 'to' buffer */
 181                 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, mapchar);
 182                 outlen += charlen;
 183         }
 184
 185         /* properly null-terminate string */
 186         for (i = 0; i < nullsize; i++)
 187                 to[outlen++] = 0;
 188
 189         return outlen;
 190 }
 191
 192 /*
 193  * NAME:        cifs_strtoUTF16()
 194  *
 195  * FUNCTION:    Convert character string to unicode string
 196  *
 197  */
 198 int
 199 cifs_strtoUTF16(__le16 *to, const char *from, int len,
 200               const struct nls_table *codepage)
 201 {
 202         int charlen;
 203         int i;
 204         wchar_t wchar_to; /* needed to quiet sparse */
 205
 206         /* special case for utf8 to handle no plane0 chars */
 207         if (!strcmp(codepage->charset, "utf8")) {
 208                 /*
 209                  * convert utf8 -> utf16, we assume we have enough space
 210                  * as caller should have assumed conversion does not overflow
 211                  * in destination len is length in wchar_t units (16bits)
 212                  */
 213                 i  = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
 214                                        (wchar_t *) to, len);
 215
 216                 /* if success terminate and exit */
 217                 if (i >= 0)
 218                         goto success;
 219                 /*
 220                  * if fails fall back to UCS encoding as this
 221                  * function should not return negative values
 222                  * currently can fail only if source contains
 223                  * invalid encoded characters
 224                  */
 225         }
 226
 227         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 228                 charlen = codepage->char2uni(from, len, &wchar_to);
 229                 if (charlen < 1) {
 230                         cERROR(1, "strtoUTF16: char2uni of 0x%x returned %d",
 231                                 *from, charlen);
 232                         /* A question mark */
 233                         wchar_to = 0x003f;
 234                         charlen = 1;
 235                 }
 236                 put_unaligned_le16(wchar_to, &to[i]);
 237         }
 238
 239 success:
 240         put_unaligned_le16(0, &to[i]);
 241         return i;
 242 }
 243
 244 /*
 245  * cifs_strndup_from_utf16 - copy a string from wire format to the local
 246  * codepage
 247  * @src - source string
 248  * @maxlen - don't walk past this many bytes in the source string
 249  * @is_unicode - is this a unicode string?
 250  * @codepage - destination codepage
 251  *
 252  * Take a string given by the server, convert it to the local codepage and
 253  * put it in a new buffer. Returns a pointer to the new string or NULL on
 254  * error.
 255  */
 256 char *
 257 cifs_strndup_from_utf16(const char *src, const int maxlen,
 258                         const bool is_unicode, const struct nls_table *codepage)
 259 {
 260         int len;
 261         char *dst;
 262
 263         if (is_unicode) {
 264                 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
 265                 len += nls_nullsize(codepage);
 266                 dst = kmalloc(len, GFP_KERNEL);
 267                 if (!dst)
 268                         return NULL;
 269                 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
 270                                false);
 271         } else {
 272                 len = strnlen(src, maxlen);
 273                 len++;
 274                 dst = kmalloc(len, GFP_KERNEL);
 275                 if (!dst)
 276                         return NULL;
 277                 strlcpy(dst, src, len);
 278         }
 279
 280         return dst;
 281 }
 282
 283 /*
 284  * Convert 16 bit Unicode pathname to wire format from string in current code
 285  * page. Conversion may involve remapping up the six characters that are
 286  * only legal in POSIX-like OS (if they are present in the string). Path
 287  * names are little endian 16 bit Unicode on the wire
 288  */
 289 int
 290 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
 291                  const struct nls_table *cp, int mapChars)
 292 {
 293         int i, j, charlen;
 294         char src_char;
 295         __le16 dst_char;
 296         wchar_t tmp;
 297
 298         if (!mapChars)
 299                 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
 300
 301         for (i = 0, j = 0; i < srclen; j++) {
 302                 src_char = source[i];
 303                 charlen = 1;
 304                 switch (src_char) {
 305                 case 0:
 306                         put_unaligned(0, &target[j]);
 307                         goto ctoUTF16_out;
 308                 case ':':
 309                         dst_char = cpu_to_le16(UNI_COLON);
 310                         break;
 311                 case '*':
 312                         dst_char = cpu_to_le16(UNI_ASTERISK);
 313                         break;
 314                 case '?':
 315                         dst_char = cpu_to_le16(UNI_QUESTION);
 316                         break;
 317                 case '<':
 318                         dst_char = cpu_to_le16(UNI_LESSTHAN);
 319                         break;
 320                 case '>':
 321                         dst_char = cpu_to_le16(UNI_GRTRTHAN);
 322                         break;
 323                 case '|':
 324                         dst_char = cpu_to_le16(UNI_PIPE);
 325                         break;
 326                 /*
 327                  * FIXME: We can not handle remapping backslash (UNI_SLASH)
 328                  * until all the calls to build_path_from_dentry are modified,
 329                  * as they use backslash as separator.
 330                  */
 331                 default:
 332                         charlen = cp->char2uni(source + i, srclen - i, &tmp);
 333                         dst_char = cpu_to_le16(tmp);
 334
 335                         /*
 336                          * if no match, use question mark, which at least in
 337                          * some cases serves as wild card
 338                          */
 339                         if (charlen < 1) {
 340                                 dst_char = cpu_to_le16(0x003f);
 341                                 charlen = 1;
 342                         }
 343                 }
 344                 /*
 345                  * character may take more than one byte in the source string,
 346                  * but will take exactly two bytes in the target string
 347                  */
 348                 i += charlen;
 349                 put_unaligned(dst_char, &target[j]);
 350         }
 351
 352 ctoUTF16_out:
 353         return j;
 354 }
 355
 356 #ifdef CONFIG_CIFS_SMB2
 357 /*
 358  * cifs_local_to_utf16_bytes - how long will a string be after conversion?
 359  * @from - pointer to input string
 360  * @maxbytes - don't go past this many bytes of input string
 361  * @codepage - source codepage
 362  *
 363  * Walk a string and return the number of bytes that the string will
 364  * be after being converted to the given charset, not including any null
 365  * termination required. Don't walk past maxbytes in the source buffer.
 366  */
 367
 368 static int
 369 cifs_local_to_utf16_bytes(const char *from, int len,
 370                           const struct nls_table *codepage)
 371 {
 372         int charlen;
 373         int i;
 374         wchar_t wchar_to;
 375
 376         for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
 377                 charlen = codepage->char2uni(from, len, &wchar_to);
 378                 /* Failed conversion defaults to a question mark */
 379                 if (charlen < 1)
 380                         charlen = 1;
 381         }
 382         return 2 * i; /* UTF16 characters are two bytes */
 383 }
 384
 385 /*
 386  * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
 387  * @src - source string
 388  * @maxlen - don't walk past this many bytes in the source string
 389  * @utf16_len - the length of the allocated string in bytes (including null)
 390  * @cp - source codepage
 391  * @remap - map special chars
 392  *
 393  * Take a string convert it from the local codepage to UTF16 and
 394  * put it in a new buffer. Returns a pointer to the new string or NULL on
 395  * error.
 396  */
 397 __le16 *
 398 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
 399                       const struct nls_table *cp, int remap)
 400 {
 401         int len;
 402         __le16 *dst;
 403
 404         len = cifs_local_to_utf16_bytes(src, maxlen, cp);
 405         len += 2; /* NULL */
 406         dst = kmalloc(len, GFP_KERNEL);
 407         if (!dst) {
 408                 *utf16_len = 0;
 409                 return NULL;
 410         }
 411         cifsConvertToUTF16(dst, src, strlen(src), cp, remap);
 412         *utf16_len = len;
 413         return dst;
 414 }
 415 #endif /* CONFIG_CIFS_SMB2 */