fs/udf/unicode.c

   1 /*
   2  * unicode.c
   3  *
   4  * PURPOSE
   5  *      Routines for converting between UTF-8 and OSTA Compressed Unicode.
   6  *      Also handles filename mangling
   7  *
   8  * DESCRIPTION
   9  *      OSTA Compressed Unicode is explained in the OSTA UDF specification.
  10  *              http://www.osta.org/
  11  *      UTF-8 is explained in the IETF RFC XXXX.
  12  *              ftp://ftp.internic.net/rfc/rfcxxxx.txt
  13  *
  14  * CONTACTS
  15  *      E-mail regarding any portion of the Linux UDF file system should be
  16  *      directed to the development team's mailing list (run by majordomo):
  17  *              linux_udf@hootie.lvld.hp.com
  18  *
  19  * COPYRIGHT
  20  *      This file is distributed under the terms of the GNU General Public
  21  *      License (GPL). Copies of the GPL can be obtained from:
  22  *              ftp://prep.ai.mit.edu/pub/gnu/GPL
  23  *      Each contributing author retains all rights to their own work.
  24  */
  25
  26
  27 #ifdef __KERNEL__
  28 #include <linux/kernel.h>
  29 #include <linux/string.h>       /* for memset */
  30 #include <linux/udf_fs.h>
  31 #else
  32 #include <string.h>
  33 #endif
  34
  35 #include "udfdecl.h"
  36
  37 int udf_ustr_to_dchars(Uint8 *dest, const struct ustr *src, int strlen)
  38 {
  39         if ( (!dest) || (!src) || (!strlen) || (src->u_len > strlen) )
  40                 return 0;
  41         memcpy(dest+1, src->u_name, src->u_len);
  42         dest[0] = src->u_cmpID;
  43         return src->u_len + 1;
  44 }
  45
  46 int udf_ustr_to_char(Uint8 *dest, const struct ustr *src, int strlen)
  47 {
  48         if ( (!dest) || (!src) || (!strlen) || (src->u_len >= strlen) )
  49                 return 0;
  50         memcpy(dest, src->u_name, src->u_len);
  51         return src->u_len;
  52 }
  53
  54 int udf_ustr_to_dstring(dstring *dest, const struct ustr *src, int dlength)
  55 {
  56         if ( udf_ustr_to_dchars(dest, src, dlength-1) )
  57         {
  58                 dest[dlength-1] = src->u_len + 1;
  59                 return dlength;
  60         }
  61         else
  62                 return 0;
  63 }
  64
  65 int udf_dchars_to_ustr(struct ustr *dest, const Uint8 *src, int strlen)
  66 {
  67         if ( (!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN) )
  68                 return 0;
  69         memset(dest, 0, sizeof(struct ustr));
  70         memcpy(dest->u_name, src+1, strlen-1);
  71         dest->u_cmpID = src[0];
  72         dest->u_len = strlen-1;
  73         return strlen-1;
  74 }
  75
  76 int udf_char_to_ustr(struct ustr *dest, const Uint8 *src, int strlen)
  77 {
  78         if ( (!dest) || (!src) || (!strlen) || (strlen >= UDF_NAME_LEN) )
  79                 return 0;
  80         memset(dest, 0, sizeof(struct ustr));
  81         memcpy(dest->u_name, src, strlen);
  82         dest->u_cmpID = 0x08;
  83         dest->u_len = strlen;
  84         return strlen;
  85 }
  86
  87
  88 int udf_dstring_to_ustr(struct ustr *dest, const dstring *src, int dlength)
  89 {
  90         if ( dlength && udf_dchars_to_ustr(dest, src, src[dlength-1]) )
  91                 return dlength;
  92         else
  93                 return 0;
  94 }
  95
  96 /*
  97  * udf_build_ustr
  98  */
  99 int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
 100 {
 101         int usesize;
 102
 103         if ( (!dest) || (!ptr) || (!size) )
 104                 return -1;
 105
 106         memset(dest, 0, sizeof(struct ustr));
 107         usesize= (size > UDF_NAME_LEN) ? UDF_NAME_LEN : size;
 108         dest->u_cmpID=ptr[0];
 109         dest->u_len=ptr[size-1];
 110         memcpy(dest->u_name, ptr+1, usesize-1);
 111         return 0;
 112 }
 113
 114 /*
 115  * udf_build_ustr_exact
 116  */
 117 int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
 118 {
 119         if ( (!dest) || (!ptr) || (!exactsize) )
 120                 return -1;
 121
 122         memset(dest, 0, sizeof(struct ustr));
 123         dest->u_cmpID=ptr[0];
 124         dest->u_len=exactsize-1;
 125         memcpy(dest->u_name, ptr+1, exactsize-1);
 126         return 0;
 127 }
 128
 129 /*
 130  * udf_ocu_to_udf8
 131  *
 132  * PURPOSE
 133  *      Convert OSTA Compressed Unicode to the UTF-8 equivalent.
 134  *
 135  * DESCRIPTION
 136  *      This routine is only called by udf_filldir().
 137  *
 138  * PRE-CONDITIONS
 139  *      utf                     Pointer to UTF-8 output buffer.
 140  *      ocu                     Pointer to OSTA Compressed Unicode input buffer
 141  *                              of size UDF_NAME_LEN bytes.
 142  *                              both of type "struct ustr *"
 143  *
 144  * POST-CONDITIONS
 145  *      <return>                Zero on success.
 146  *
 147  * HISTORY
 148  *      November 12, 1997 - Andrew E. Mileski
 149  *      Written, tested, and released.
 150  */
 151 int udf_CS0toUTF8(struct ustr *utf_o, struct ustr *ocu_i)
 152 {
 153         Uint8 *ocu;
 154         Uint32 c;
 155         Uint8 cmp_id, ocu_len;
 156         int i;
 157
 158         ocu = ocu_i->u_name;
 159
 160         ocu_len = ocu_i->u_len;
 161         cmp_id = ocu_i->u_cmpID;
 162         utf_o->u_len = 0;
 163
 164         if (ocu_len == 0)
 165         {
 166                 memset(utf_o, 0, sizeof(struct ustr));
 167                 utf_o->u_cmpID = 0;
 168                 utf_o->u_len = 0;
 169                 return 0;
 170         }
 171
 172         if ((cmp_id != 8) && (cmp_id != 16))
 173         {
 174 #ifdef __KERNEL__
 175                 printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name);
 176 #endif
 177                 return 0;
 178         }
 179
 180         for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN-3)) ;)
 181         {
 182
 183                 /* Expand OSTA compressed Unicode to Unicode */
 184                 c = ocu[i++];
 185                 if (cmp_id == 16)
 186                         c = (c << 8) | ocu[i++];
 187
 188                 /* Compress Unicode to UTF-8 */
 189                 if (c < 0x80U)
 190                         utf_o->u_name[utf_o->u_len++] = (Uint8)c;
 191                 else if (c < 0x800U)
 192                 {
 193                         utf_o->u_name[utf_o->u_len++] = (Uint8)(0xc0 | (c >> 6));
 194                         utf_o->u_name[utf_o->u_len++] = (Uint8)(0x80 | (c & 0x3f));
 195                 }
 196                 else
 197                 {
 198                         utf_o->u_name[utf_o->u_len++] = (Uint8)(0xe0 | (c >> 12));
 199                         utf_o->u_name[utf_o->u_len++] = (Uint8)(0x80 | ((c >> 6) & 0x3f));
 200                         utf_o->u_name[utf_o->u_len++] = (Uint8)(0x80 | (c & 0x3f));
 201                 }
 202         }
 203         utf_o->u_cmpID=8;
 204         utf_o->u_hash=0L;
 205         utf_o->padding=0;
 206
 207         return utf_o->u_len;
 208 }
 209
 210 /*
 211  *
 212  * udf_utf8_to_ocu
 213  *
 214  * PURPOSE
 215  *      Convert UTF-8 to the OSTA Compressed Unicode equivalent.
 216  *
 217  * DESCRIPTION
 218  *      This routine is only called by udf_lookup().
 219  *
 220  * PRE-CONDITIONS
 221  *      ocu                     Pointer to OSTA Compressed Unicode output
 222  *                              buffer of size UDF_NAME_LEN bytes.
 223  *      utf                     Pointer to UTF-8 input buffer.
 224  *      utf_len                 Length of UTF-8 input buffer in bytes.
 225  *
 226  * POST-CONDITIONS
 227  *      <return>                Zero on success.
 228  *
 229  * HISTORY
 230  *      November 12, 1997 - Andrew E. Mileski
 231  *      Written, tested, and released.
 232  */
 233 int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length)
 234 {
 235         unsigned c, i, max_val, utf_char;
 236         int utf_cnt;
 237         int u_len = 0;
 238
 239         memset(ocu, 0, sizeof(dstring) * length);
 240         ocu[0] = 8;
 241         max_val = 0xffU;
 242
 243 try_again:
 244         utf_char = 0U;
 245         utf_cnt = 0U;
 246         for (i = 0U; i < utf->u_len; i++)
 247         {
 248                 c = (Uint8)utf->u_name[i];
 249
 250                 /* Complete a multi-byte UTF-8 character */
 251                 if (utf_cnt)
 252                 {
 253                         utf_char = (utf_char << 6) | (c & 0x3fU);
 254                         if (--utf_cnt)
 255                                 continue;
 256                 }
 257                 else
 258                 {
 259                         /* Check for a multi-byte UTF-8 character */
 260                         if (c & 0x80U)
 261                         {
 262                                 /* Start a multi-byte UTF-8 character */
 263                                 if ((c & 0xe0U) == 0xc0U)
 264                                 {
 265                                         utf_char = c & 0x1fU;
 266                                         utf_cnt = 1;
 267                                 }
 268                                 else if ((c & 0xf0U) == 0xe0U)
 269                                 {
 270                                         utf_char = c & 0x0fU;
 271                                         utf_cnt = 2;
 272                                 }
 273                                 else if ((c & 0xf8U) == 0xf0U)
 274                                 {
 275                                         utf_char = c & 0x07U;
 276                                         utf_cnt = 3;
 277                                 }
 278                                 else if ((c & 0xfcU) == 0xf8U)
 279                                 {
 280                                         utf_char = c & 0x03U;
 281                                         utf_cnt = 4;
 282                                 }
 283                                 else if ((c & 0xfeU) == 0xfcU)
 284                                 {
 285                                         utf_char = c & 0x01U;
 286                                         utf_cnt = 5;
 287                                 }
 288                                 else
 289                                         goto error_out;
 290                                 continue;
 291                         } else
 292                                 /* Single byte UTF-8 character (most common) */
 293                                 utf_char = c;
 294                 }
 295
 296                 /* Choose no compression if necessary */
 297                 if (utf_char > max_val)
 298                 {
 299                         if ( 0xffU == max_val )
 300                         {
 301                                 max_val = 0xffffU;
 302                                 ocu[0] = (Uint8)0x10U;
 303                                 goto try_again;
 304                         }
 305                         goto error_out;
 306                 }
 307
 308                 if (max_val == 0xffffU)
 309                 {
 310                         ocu[++u_len] = (Uint8)(utf_char >> 8);
 311                 }
 312                 ocu[++u_len] = (Uint8)(utf_char & 0xffU);
 313         }
 314
 315
 316         if (utf_cnt)
 317         {
 318 error_out:
 319 #ifdef __KERNEL__
 320                 printk(KERN_ERR "udf: bad UTF-8 character\n");
 321 #endif
 322                 return 0;
 323         }
 324
 325         ocu[length - 1] = (Uint8)u_len + 1;
 326         return u_len + 1;
 327 }
 328
 329 #ifdef __KERNEL__
 330 int udf_get_filename(Uint8 *sname, Uint8 *dname, int flen)
 331 {
 332         struct ustr filename, unifilename;
 333         int len;
 334
 335         if (udf_build_ustr_exact(&unifilename, sname, flen))
 336         {
 337                 return 0;
 338         }
 339
 340         if (!udf_CS0toUTF8(&filename, &unifilename) )
 341         {
 342                 udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
 343                 return 0;
 344         }
 345
 346         if ((len = udf_translate_to_linux(dname, filename.u_name, filename.u_len,
 347                 unifilename.u_name, unifilename.u_len)))
 348         {
 349                 return len;
 350         }
 351         return 0;
 352 }
 353 #endif
 354
 355 #define ILLEGAL_CHAR_MARK       '_'
 356 #define EXT_MARK                        '.'
 357 #define CRC_MARK                        '#'
 358 #define EXT_SIZE                        5
 359
 360 int udf_translate_to_linux(Uint8 *newName, Uint8 *udfName, int udfLen, Uint8 *fidName, int fidNameLen)
 361 {
 362         int index, newIndex = 0, needsCRC = 0;
 363         int extIndex = 0, newExtIndex = 0, hasExt = 0;
 364         unsigned short valueCRC;
 365         Uint8 curr;
 366         const Uint8 hexChar[] = "0123456789ABCDEF";
 367
 368         if (udfName[0] == '.' && (udfLen == 1 ||
 369                 (udfLen == 2 && udfName[1] == '.')))
 370         {
 371                 needsCRC = 1;
 372                 newIndex = udfLen;
 373                 memcpy(newName, udfName, udfLen);
 374         }
 375         else
 376         {
 377                 for (index = 0; index < udfLen; index++)
 378                 {
 379                         curr = udfName[index];
 380                         if (curr == '/' || curr == 0)
 381                         {
 382                                 needsCRC = 1;
 383                                 curr = ILLEGAL_CHAR_MARK;
 384                                 while (index+1 < udfLen && (udfName[index+1] == '/' ||
 385                                         udfName[index+1] == 0))
 386                                         index++;
 387                         }
 388                         if (curr == EXT_MARK && (udfLen - index - 1) <= EXT_SIZE)
 389                         {
 390                                 if (udfLen == index + 1)
 391                                         hasExt = 0;
 392                                 else
 393                                 {
 394                                         hasExt = 1;
 395                                         extIndex = index;
 396                                         newExtIndex = newIndex;
 397                                 }
 398                         }
 399                         if (newIndex < 256)
 400                                 newName[newIndex++] = curr;
 401                         else
 402                                 needsCRC = 1;
 403                 }
 404         }
 405         if (needsCRC)
 406         {
 407                 Uint8 ext[EXT_SIZE];
 408                 int localExtIndex = 0;
 409
 410                 if (hasExt)
 411                 {
 412                         int maxFilenameLen;
 413                         for(index = 0; index<EXT_SIZE && extIndex + index +1 < udfLen;
 414                                 index++ )
 415                         {
 416                                 curr = udfName[extIndex + index + 1];
 417
 418                                 if (curr == '/' || curr == 0)
 419                                 {
 420                                         needsCRC = 1;
 421                                         curr = ILLEGAL_CHAR_MARK;
 422                                         while(extIndex + index + 2 < udfLen && (index + 1 < EXT_SIZE
 423                                                 && (udfName[extIndex + index + 2] == '/' ||
 424                                                         udfName[extIndex + index + 2] == 0)))
 425                                                 index++;
 426                                 }
 427                                 ext[localExtIndex++] = curr;
 428                         }
 429                         maxFilenameLen = 250 - localExtIndex;
 430                         if (newIndex > maxFilenameLen)
 431                                 newIndex = maxFilenameLen;
 432                         else
 433                                 newIndex = newExtIndex;
 434                 }
 435                 else if (newIndex > 250)
 436                         newIndex = 250;
 437                 newName[newIndex++] = CRC_MARK;
 438                 valueCRC = udf_crc(fidName, fidNameLen, 0);
 439                 newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
 440                 newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
 441                 newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
 442                 newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
 443
 444                 if (hasExt)
 445                 {
 446                         newName[newIndex++] = EXT_MARK;
 447                         for (index = 0;index < localExtIndex ;index++ )
 448                                 newName[newIndex++] = ext[index];
 449                 }
 450         }
 451         return newIndex;
 452 }