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 * COPYRIGHT 15 * This file is distributed under the terms of the GNU General Public 16 * License (GPL). Copies of the GPL can be obtained from: 17 * ftp://prep.ai.mit.edu/pub/gnu/GPL 18 * Each contributing author retains all rights to their own work. 19 */ 20 21 #include "udfdecl.h" 22 23 #include <linux/kernel.h> 24 #include <linux/string.h> /* for memset */ 25 #include <linux/nls.h> 26 #include <linux/crc-itu-t.h> 27 #include <linux/slab.h> 28 29 #include "udf_sb.h" 30 31 static int udf_translate_to_linux(uint8_t *, int, uint8_t *, int, uint8_t *, 32 int); 33 34 static int udf_char_to_ustr(struct ustr *dest, const uint8_t *src, int strlen) 35 { 36 if ((!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN - 2)) 37 return 0; 38 39 memset(dest, 0, sizeof(struct ustr)); 40 memcpy(dest->u_name, src, strlen); 41 dest->u_cmpID = 0x08; 42 dest->u_len = strlen; 43 44 return strlen; 45 } 46 47 /* 48 * udf_build_ustr 49 */ 50 int udf_build_ustr(struct ustr *dest, dstring *ptr, int size) 51 { 52 int usesize; 53 54 if (!dest || !ptr || !size) 55 return -1; 56 BUG_ON(size < 2); 57 58 usesize = min_t(size_t, ptr[size - 1], sizeof(dest->u_name)); 59 usesize = min(usesize, size - 2); 60 dest->u_cmpID = ptr[0]; 61 dest->u_len = usesize; 62 memcpy(dest->u_name, ptr + 1, usesize); 63 memset(dest->u_name + usesize, 0, sizeof(dest->u_name) - usesize); 64 65 return 0; 66 } 67 68 /* 69 * udf_build_ustr_exact 70 */ 71 static void udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize) 72 { 73 memset(dest, 0, sizeof(struct ustr)); 74 dest->u_cmpID = ptr[0]; 75 dest->u_len = exactsize - 1; 76 memcpy(dest->u_name, ptr + 1, exactsize - 1); 77 } 78 79 /* 80 * udf_CS0toUTF8 81 * 82 * PURPOSE 83 * Convert OSTA Compressed Unicode to the UTF-8 equivalent. 84 * 85 * PRE-CONDITIONS 86 * utf Pointer to UTF-8 output buffer. 87 * ocu Pointer to OSTA Compressed Unicode input buffer 88 * of size UDF_NAME_LEN bytes. 89 * both of type "struct ustr *" 90 * 91 * POST-CONDITIONS 92 * <return> >= 0 on success. 93 * 94 * HISTORY 95 * November 12, 1997 - Andrew E. Mileski 96 * Written, tested, and released. 97 */ 98 int udf_CS0toUTF8(struct ustr *utf_o, const struct ustr *ocu_i) 99 { 100 const uint8_t *ocu; 101 uint8_t cmp_id, ocu_len; 102 int i; 103 104 ocu_len = ocu_i->u_len; 105 if (ocu_len == 0) { 106 memset(utf_o, 0, sizeof(struct ustr)); 107 return 0; 108 } 109 110 cmp_id = ocu_i->u_cmpID; 111 if (cmp_id != 8 && cmp_id != 16) { 112 memset(utf_o, 0, sizeof(struct ustr)); 113 pr_err("unknown compression code (%d) stri=%s\n", 114 cmp_id, ocu_i->u_name); 115 return -EINVAL; 116 } 117 118 ocu = ocu_i->u_name; 119 utf_o->u_len = 0; 120 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) { 121 122 /* Expand OSTA compressed Unicode to Unicode */ 123 uint32_t c = ocu[i++]; 124 if (cmp_id == 16) 125 c = (c << 8) | ocu[i++]; 126 127 /* Compress Unicode to UTF-8 */ 128 if (c < 0x80U) 129 utf_o->u_name[utf_o->u_len++] = (uint8_t)c; 130 else if (c < 0x800U) { 131 if (utf_o->u_len > (UDF_NAME_LEN - 4)) 132 break; 133 utf_o->u_name[utf_o->u_len++] = 134 (uint8_t)(0xc0 | (c >> 6)); 135 utf_o->u_name[utf_o->u_len++] = 136 (uint8_t)(0x80 | (c & 0x3f)); 137 } else { 138 if (utf_o->u_len > (UDF_NAME_LEN - 5)) 139 break; 140 utf_o->u_name[utf_o->u_len++] = 141 (uint8_t)(0xe0 | (c >> 12)); 142 utf_o->u_name[utf_o->u_len++] = 143 (uint8_t)(0x80 | 144 ((c >> 6) & 0x3f)); 145 utf_o->u_name[utf_o->u_len++] = 146 (uint8_t)(0x80 | (c & 0x3f)); 147 } 148 } 149 utf_o->u_cmpID = 8; 150 151 return utf_o->u_len; 152 } 153 154 /* 155 * 156 * udf_UTF8toCS0 157 * 158 * PURPOSE 159 * Convert UTF-8 to the OSTA Compressed Unicode equivalent. 160 * 161 * DESCRIPTION 162 * This routine is only called by udf_lookup(). 163 * 164 * PRE-CONDITIONS 165 * ocu Pointer to OSTA Compressed Unicode output 166 * buffer of size UDF_NAME_LEN bytes. 167 * utf Pointer to UTF-8 input buffer. 168 * utf_len Length of UTF-8 input buffer in bytes. 169 * 170 * POST-CONDITIONS 171 * <return> Zero on success. 172 * 173 * HISTORY 174 * November 12, 1997 - Andrew E. Mileski 175 * Written, tested, and released. 176 */ 177 static int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length) 178 { 179 unsigned c, i, max_val, utf_char; 180 int utf_cnt, u_len; 181 182 memset(ocu, 0, sizeof(dstring) * length); 183 ocu[0] = 8; 184 max_val = 0xffU; 185 186 try_again: 187 u_len = 0U; 188 utf_char = 0U; 189 utf_cnt = 0U; 190 for (i = 0U; i < utf->u_len; i++) { 191 c = (uint8_t)utf->u_name[i]; 192 193 /* Complete a multi-byte UTF-8 character */ 194 if (utf_cnt) { 195 utf_char = (utf_char << 6) | (c & 0x3fU); 196 if (--utf_cnt) 197 continue; 198 } else { 199 /* Check for a multi-byte UTF-8 character */ 200 if (c & 0x80U) { 201 /* Start a multi-byte UTF-8 character */ 202 if ((c & 0xe0U) == 0xc0U) { 203 utf_char = c & 0x1fU; 204 utf_cnt = 1; 205 } else if ((c & 0xf0U) == 0xe0U) { 206 utf_char = c & 0x0fU; 207 utf_cnt = 2; 208 } else if ((c & 0xf8U) == 0xf0U) { 209 utf_char = c & 0x07U; 210 utf_cnt = 3; 211 } else if ((c & 0xfcU) == 0xf8U) { 212 utf_char = c & 0x03U; 213 utf_cnt = 4; 214 } else if ((c & 0xfeU) == 0xfcU) { 215 utf_char = c & 0x01U; 216 utf_cnt = 5; 217 } else { 218 goto error_out; 219 } 220 continue; 221 } else { 222 /* Single byte UTF-8 character (most common) */ 223 utf_char = c; 224 } 225 } 226 227 /* Choose no compression if necessary */ 228 if (utf_char > max_val) { 229 if (max_val == 0xffU) { 230 max_val = 0xffffU; 231 ocu[0] = (uint8_t)0x10U; 232 goto try_again; 233 } 234 goto error_out; 235 } 236 237 if (max_val == 0xffffU) 238 ocu[++u_len] = (uint8_t)(utf_char >> 8); 239 ocu[++u_len] = (uint8_t)(utf_char & 0xffU); 240 } 241 242 if (utf_cnt) { 243 error_out: 244 ocu[++u_len] = '?'; 245 printk(KERN_DEBUG pr_fmt("bad UTF-8 character\n")); 246 } 247 248 ocu[length - 1] = (uint8_t)u_len + 1; 249 250 return u_len + 1; 251 } 252 253 static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o, 254 const struct ustr *ocu_i) 255 { 256 const uint8_t *ocu; 257 uint8_t cmp_id, ocu_len; 258 int i, len; 259 260 261 ocu_len = ocu_i->u_len; 262 if (ocu_len == 0) { 263 memset(utf_o, 0, sizeof(struct ustr)); 264 return 0; 265 } 266 267 cmp_id = ocu_i->u_cmpID; 268 if (cmp_id != 8 && cmp_id != 16) { 269 memset(utf_o, 0, sizeof(struct ustr)); 270 pr_err("unknown compression code (%d) stri=%s\n", 271 cmp_id, ocu_i->u_name); 272 return -EINVAL; 273 } 274 275 ocu = ocu_i->u_name; 276 utf_o->u_len = 0; 277 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) { 278 /* Expand OSTA compressed Unicode to Unicode */ 279 uint32_t c = ocu[i++]; 280 if (cmp_id == 16) 281 c = (c << 8) | ocu[i++]; 282 283 len = nls->uni2char(c, &utf_o->u_name[utf_o->u_len], 284 UDF_NAME_LEN - 2 - utf_o->u_len); 285 /* Valid character? */ 286 if (len >= 0) 287 utf_o->u_len += len; 288 else 289 utf_o->u_name[utf_o->u_len++] = '?'; 290 } 291 utf_o->u_cmpID = 8; 292 293 return utf_o->u_len; 294 } 295 296 static int udf_NLStoCS0(struct nls_table *nls, dstring *ocu, struct ustr *uni, 297 int length) 298 { 299 int len; 300 unsigned i, max_val; 301 uint16_t uni_char; 302 int u_len; 303 304 memset(ocu, 0, sizeof(dstring) * length); 305 ocu[0] = 8; 306 max_val = 0xffU; 307 308 try_again: 309 u_len = 0U; 310 for (i = 0U; i < uni->u_len; i++) { 311 len = nls->char2uni(&uni->u_name[i], uni->u_len - i, &uni_char); 312 if (!len) 313 continue; 314 /* Invalid character, deal with it */ 315 if (len < 0) { 316 len = 1; 317 uni_char = '?'; 318 } 319 320 if (uni_char > max_val) { 321 max_val = 0xffffU; 322 ocu[0] = (uint8_t)0x10U; 323 goto try_again; 324 } 325 326 if (max_val == 0xffffU) 327 ocu[++u_len] = (uint8_t)(uni_char >> 8); 328 ocu[++u_len] = (uint8_t)(uni_char & 0xffU); 329 i += len - 1; 330 } 331 332 ocu[length - 1] = (uint8_t)u_len + 1; 333 return u_len + 1; 334 } 335 336 int udf_get_filename(struct super_block *sb, uint8_t *sname, int slen, 337 uint8_t *dname, int dlen) 338 { 339 struct ustr *filename, *unifilename; 340 int ret; 341 342 if (!slen) 343 return -EIO; 344 345 filename = kmalloc(sizeof(struct ustr), GFP_NOFS); 346 if (!filename) 347 return -ENOMEM; 348 349 unifilename = kmalloc(sizeof(struct ustr), GFP_NOFS); 350 if (!unifilename) { 351 ret = -ENOMEM; 352 goto out1; 353 } 354 355 udf_build_ustr_exact(unifilename, sname, slen); 356 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) { 357 ret = udf_CS0toUTF8(filename, unifilename); 358 if (ret < 0) { 359 udf_debug("Failed in udf_get_filename: sname = %s\n", 360 sname); 361 goto out2; 362 } 363 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) { 364 ret = udf_CS0toNLS(UDF_SB(sb)->s_nls_map, filename, 365 unifilename); 366 if (ret < 0) { 367 udf_debug("Failed in udf_get_filename: sname = %s\n", 368 sname); 369 goto out2; 370 } 371 } else 372 BUG(); 373 374 ret = udf_translate_to_linux(dname, dlen, 375 filename->u_name, filename->u_len, 376 unifilename->u_name, unifilename->u_len); 377 /* Zero length filename isn't valid... */ 378 if (ret == 0) 379 ret = -EINVAL; 380 out2: 381 kfree(unifilename); 382 out1: 383 kfree(filename); 384 return ret; 385 } 386 387 int udf_put_filename(struct super_block *sb, const uint8_t *sname, 388 uint8_t *dname, int flen) 389 { 390 struct ustr unifilename; 391 int namelen; 392 393 if (!udf_char_to_ustr(&unifilename, sname, flen)) 394 return 0; 395 396 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) { 397 namelen = udf_UTF8toCS0(dname, &unifilename, UDF_NAME_LEN); 398 if (!namelen) 399 return 0; 400 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) { 401 namelen = udf_NLStoCS0(UDF_SB(sb)->s_nls_map, dname, 402 &unifilename, UDF_NAME_LEN); 403 if (!namelen) 404 return 0; 405 } else 406 return 0; 407 408 return namelen; 409 } 410 411 #define ILLEGAL_CHAR_MARK '_' 412 #define EXT_MARK '.' 413 #define CRC_MARK '#' 414 #define EXT_SIZE 5 415 /* Number of chars we need to store generated CRC to make filename unique */ 416 #define CRC_LEN 5 417 418 static int udf_translate_to_linux(uint8_t *newName, int newLen, 419 uint8_t *udfName, int udfLen, 420 uint8_t *fidName, int fidNameLen) 421 { 422 int index, newIndex = 0, needsCRC = 0; 423 int extIndex = 0, newExtIndex = 0, hasExt = 0; 424 unsigned short valueCRC; 425 uint8_t curr; 426 427 if (udfName[0] == '.' && 428 (udfLen == 1 || (udfLen == 2 && udfName[1] == '.'))) { 429 needsCRC = 1; 430 newIndex = udfLen; 431 memcpy(newName, udfName, udfLen); 432 } else { 433 for (index = 0; index < udfLen; index++) { 434 curr = udfName[index]; 435 if (curr == '/' || curr == 0) { 436 needsCRC = 1; 437 curr = ILLEGAL_CHAR_MARK; 438 while (index + 1 < udfLen && 439 (udfName[index + 1] == '/' || 440 udfName[index + 1] == 0)) 441 index++; 442 } 443 if (curr == EXT_MARK && 444 (udfLen - index - 1) <= EXT_SIZE) { 445 if (udfLen == index + 1) 446 hasExt = 0; 447 else { 448 hasExt = 1; 449 extIndex = index; 450 newExtIndex = newIndex; 451 } 452 } 453 if (newIndex < newLen) 454 newName[newIndex++] = curr; 455 else 456 needsCRC = 1; 457 } 458 } 459 if (needsCRC) { 460 uint8_t ext[EXT_SIZE]; 461 int localExtIndex = 0; 462 463 if (hasExt) { 464 int maxFilenameLen; 465 for (index = 0; 466 index < EXT_SIZE && extIndex + index + 1 < udfLen; 467 index++) { 468 curr = udfName[extIndex + index + 1]; 469 470 if (curr == '/' || curr == 0) { 471 needsCRC = 1; 472 curr = ILLEGAL_CHAR_MARK; 473 while (extIndex + index + 2 < udfLen && 474 (index + 1 < EXT_SIZE && 475 (udfName[extIndex + index + 2] == '/' || 476 udfName[extIndex + index + 2] == 0))) 477 index++; 478 } 479 ext[localExtIndex++] = curr; 480 } 481 maxFilenameLen = newLen - CRC_LEN - localExtIndex; 482 if (newIndex > maxFilenameLen) 483 newIndex = maxFilenameLen; 484 else 485 newIndex = newExtIndex; 486 } else if (newIndex > newLen - CRC_LEN) 487 newIndex = newLen - CRC_LEN; 488 newName[newIndex++] = CRC_MARK; 489 valueCRC = crc_itu_t(0, fidName, fidNameLen); 490 newName[newIndex++] = hex_asc_upper_hi(valueCRC >> 8); 491 newName[newIndex++] = hex_asc_upper_lo(valueCRC >> 8); 492 newName[newIndex++] = hex_asc_upper_hi(valueCRC); 493 newName[newIndex++] = hex_asc_upper_lo(valueCRC); 494 495 if (hasExt) { 496 newName[newIndex++] = EXT_MARK; 497 for (index = 0; index < localExtIndex; index++) 498 newName[newIndex++] = ext[index]; 499 } 500 } 501 502 return newIndex; 503 } 504