File: /usr/src/linux/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@hpesjro.fc.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/nls.h>
31 #include <linux/udf_fs.h>
32 #include "udf_sb.h"
33 #else
34 #include <string.h>
35 #endif
36
37 #include "udfdecl.h"
38
39 int udf_ustr_to_dchars(Uint8 *dest, const struct ustr *src, int strlen)
40 {
41 if ( (!dest) || (!src) || (!strlen) || (src->u_len > strlen) )
42 return 0;
43 memcpy(dest+1, src->u_name, src->u_len);
44 dest[0] = src->u_cmpID;
45 return src->u_len + 1;
46 }
47
48 int udf_ustr_to_char(Uint8 *dest, const struct ustr *src, int strlen)
49 {
50 if ( (!dest) || (!src) || (!strlen) || (src->u_len >= strlen) )
51 return 0;
52 memcpy(dest, src->u_name, src->u_len);
53 return src->u_len;
54 }
55
56 int udf_ustr_to_dstring(dstring *dest, const struct ustr *src, int dlength)
57 {
58 if ( udf_ustr_to_dchars(dest, src, dlength-1) )
59 {
60 dest[dlength-1] = src->u_len + 1;
61 return dlength;
62 }
63 else
64 return 0;
65 }
66
67 int udf_dchars_to_ustr(struct ustr *dest, const Uint8 *src, int strlen)
68 {
69 if ( (!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN) )
70 return 0;
71 memset(dest, 0, sizeof(struct ustr));
72 memcpy(dest->u_name, src+1, strlen-1);
73 dest->u_cmpID = src[0];
74 dest->u_len = strlen-1;
75 return strlen-1;
76 }
77
78 int udf_char_to_ustr(struct ustr *dest, const Uint8 *src, int strlen)
79 {
80 if ( (!dest) || (!src) || (!strlen) || (strlen >= UDF_NAME_LEN) )
81 return 0;
82 memset(dest, 0, sizeof(struct ustr));
83 memcpy(dest->u_name, src, strlen);
84 dest->u_cmpID = 0x08;
85 dest->u_len = strlen;
86 return strlen;
87 }
88
89
90 int udf_dstring_to_ustr(struct ustr *dest, const dstring *src, int dlength)
91 {
92 if ( dlength && udf_dchars_to_ustr(dest, src, src[dlength-1]) )
93 return dlength;
94 else
95 return 0;
96 }
97
98 /*
99 * udf_build_ustr
100 */
101 int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
102 {
103 int usesize;
104
105 if ( (!dest) || (!ptr) || (!size) )
106 return -1;
107
108 memset(dest, 0, sizeof(struct ustr));
109 usesize= (size > UDF_NAME_LEN) ? UDF_NAME_LEN : size;
110 dest->u_cmpID=ptr[0];
111 dest->u_len=ptr[size-1];
112 memcpy(dest->u_name, ptr+1, usesize-1);
113 return 0;
114 }
115
116 /*
117 * udf_build_ustr_exact
118 */
119 int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
120 {
121 if ( (!dest) || (!ptr) || (!exactsize) )
122 return -1;
123
124 memset(dest, 0, sizeof(struct ustr));
125 dest->u_cmpID=ptr[0];
126 dest->u_len=exactsize-1;
127 memcpy(dest->u_name, ptr+1, exactsize-1);
128 return 0;
129 }
130
131 /*
132 * udf_ocu_to_utf8
133 *
134 * PURPOSE
135 * Convert OSTA Compressed Unicode to the UTF-8 equivalent.
136 *
137 * DESCRIPTION
138 * This routine is only called by udf_filldir().
139 *
140 * PRE-CONDITIONS
141 * utf Pointer to UTF-8 output buffer.
142 * ocu Pointer to OSTA Compressed Unicode input buffer
143 * of size UDF_NAME_LEN bytes.
144 * both of type "struct ustr *"
145 *
146 * POST-CONDITIONS
147 * <return> Zero on success.
148 *
149 * HISTORY
150 * November 12, 1997 - Andrew E. Mileski
151 * Written, tested, and released.
152 */
153 int udf_CS0toUTF8(struct ustr *utf_o, struct ustr *ocu_i)
154 {
155 Uint8 *ocu;
156 Uint32 c;
157 Uint8 cmp_id, ocu_len;
158 int i;
159
160 ocu = ocu_i->u_name;
161
162 ocu_len = ocu_i->u_len;
163 cmp_id = ocu_i->u_cmpID;
164 utf_o->u_len = 0;
165
166 if (ocu_len == 0)
167 {
168 memset(utf_o, 0, sizeof(struct ustr));
169 utf_o->u_cmpID = 0;
170 utf_o->u_len = 0;
171 return 0;
172 }
173
174 if ((cmp_id != 8) && (cmp_id != 16))
175 {
176 #ifdef __KERNEL__
177 printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name);
178 #endif
179 return 0;
180 }
181
182 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN-3)) ;)
183 {
184
185 /* Expand OSTA compressed Unicode to Unicode */
186 c = ocu[i++];
187 if (cmp_id == 16)
188 c = (c << 8) | ocu[i++];
189
190 /* Compress Unicode to UTF-8 */
191 if (c < 0x80U)
192 utf_o->u_name[utf_o->u_len++] = (Uint8)c;
193 else if (c < 0x800U)
194 {
195 utf_o->u_name[utf_o->u_len++] = (Uint8)(0xc0 | (c >> 6));
196 utf_o->u_name[utf_o->u_len++] = (Uint8)(0x80 | (c & 0x3f));
197 }
198 else
199 {
200 utf_o->u_name[utf_o->u_len++] = (Uint8)(0xe0 | (c >> 12));
201 utf_o->u_name[utf_o->u_len++] = (Uint8)(0x80 | ((c >> 6) & 0x3f));
202 utf_o->u_name[utf_o->u_len++] = (Uint8)(0x80 | (c & 0x3f));
203 }
204 }
205 utf_o->u_cmpID=8;
206 utf_o->u_hash=0L;
207 utf_o->padding=0;
208
209 return utf_o->u_len;
210 }
211
212 /*
213 *
214 * udf_utf8_to_ocu
215 *
216 * PURPOSE
217 * Convert UTF-8 to the OSTA Compressed Unicode equivalent.
218 *
219 * DESCRIPTION
220 * This routine is only called by udf_lookup().
221 *
222 * PRE-CONDITIONS
223 * ocu Pointer to OSTA Compressed Unicode output
224 * buffer of size UDF_NAME_LEN bytes.
225 * utf Pointer to UTF-8 input buffer.
226 * utf_len Length of UTF-8 input buffer in bytes.
227 *
228 * POST-CONDITIONS
229 * <return> Zero on success.
230 *
231 * HISTORY
232 * November 12, 1997 - Andrew E. Mileski
233 * Written, tested, and released.
234 */
235 int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length)
236 {
237 unsigned c, i, max_val, utf_char;
238 int utf_cnt;
239 int u_len = 0;
240
241 memset(ocu, 0, sizeof(dstring) * length);
242 ocu[0] = 8;
243 max_val = 0xffU;
244
245 try_again:
246 utf_char = 0U;
247 utf_cnt = 0U;
248 for (i = 0U; i < utf->u_len; i++)
249 {
250 c = (Uint8)utf->u_name[i];
251
252 /* Complete a multi-byte UTF-8 character */
253 if (utf_cnt)
254 {
255 utf_char = (utf_char << 6) | (c & 0x3fU);
256 if (--utf_cnt)
257 continue;
258 }
259 else
260 {
261 /* Check for a multi-byte UTF-8 character */
262 if (c & 0x80U)
263 {
264 /* Start a multi-byte UTF-8 character */
265 if ((c & 0xe0U) == 0xc0U)
266 {
267 utf_char = c & 0x1fU;
268 utf_cnt = 1;
269 }
270 else if ((c & 0xf0U) == 0xe0U)
271 {
272 utf_char = c & 0x0fU;
273 utf_cnt = 2;
274 }
275 else if ((c & 0xf8U) == 0xf0U)
276 {
277 utf_char = c & 0x07U;
278 utf_cnt = 3;
279 }
280 else if ((c & 0xfcU) == 0xf8U)
281 {
282 utf_char = c & 0x03U;
283 utf_cnt = 4;
284 }
285 else if ((c & 0xfeU) == 0xfcU)
286 {
287 utf_char = c & 0x01U;
288 utf_cnt = 5;
289 }
290 else
291 goto error_out;
292 continue;
293 } else
294 /* Single byte UTF-8 character (most common) */
295 utf_char = c;
296 }
297
298 /* Choose no compression if necessary */
299 if (utf_char > max_val)
300 {
301 if ( 0xffU == max_val )
302 {
303 max_val = 0xffffU;
304 ocu[0] = (Uint8)0x10U;
305 goto try_again;
306 }
307 goto error_out;
308 }
309
310 if (max_val == 0xffffU)
311 {
312 ocu[++u_len] = (Uint8)(utf_char >> 8);
313 }
314 ocu[++u_len] = (Uint8)(utf_char & 0xffU);
315 }
316
317
318 if (utf_cnt)
319 {
320 error_out:
321 #ifdef __KERNEL__
322 printk(KERN_ERR "udf: bad UTF-8 character\n");
323 #endif
324 return 0;
325 }
326
327 ocu[length - 1] = (Uint8)u_len + 1;
328 return u_len + 1;
329 }
330
331 #ifdef __KERNEL__
332 int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o, struct ustr *ocu_i)
333 {
334 Uint8 *ocu;
335 Uint32 c;
336 Uint8 cmp_id, ocu_len;
337 int i;
338
339 ocu = ocu_i->u_name;
340
341 ocu_len = ocu_i->u_len;
342 cmp_id = ocu_i->u_cmpID;
343 utf_o->u_len = 0;
344
345 if (ocu_len == 0)
346 {
347 memset(utf_o, 0, sizeof(struct ustr));
348 utf_o->u_cmpID = 0;
349 utf_o->u_len = 0;
350 return 0;
351 }
352
353 if ((cmp_id != 8) && (cmp_id != 16))
354 {
355 printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name);
356 return 0;
357 }
358
359 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN-3)) ;)
360 {
361 /* Expand OSTA compressed Unicode to Unicode */
362 c = ocu[i++];
363 if (cmp_id == 16)
364 c = (c << 8) | ocu[i++];
365
366 utf_o->u_len += nls->uni2char(c, &utf_o->u_name[utf_o->u_len],
367 UDF_NAME_LEN - utf_o->u_len);
368 }
369 utf_o->u_cmpID=8;
370 utf_o->u_hash=0L;
371 utf_o->padding=0;
372
373 return utf_o->u_len;
374 }
375
376 int udf_NLStoCS0(struct nls_table *nls, dstring *ocu, struct ustr *uni, int length)
377 {
378 unsigned len, i, max_val;
379 Uint16 uni_char;
380 int uni_cnt;
381 int u_len = 0;
382
383 memset(ocu, 0, sizeof(dstring) * length);
384 ocu[0] = 8;
385 max_val = 0xffU;
386
387 try_again:
388 uni_char = 0U;
389 uni_cnt = 0U;
390 for (i = 0U; i < uni->u_len; i++)
391 {
392 len = nls->char2uni(&uni->u_name[i], uni->u_len-i, &uni_char);
393
394 if (len == 2 && max_val == 0xff)
395 {
396 max_val = 0xffffU;
397 ocu[0] = (Uint8)0x10U;
398 goto try_again;
399 }
400
401 if (max_val == 0xffffU)
402 {
403 ocu[++u_len] = (Uint8)(uni_char >> 8);
404 i++;
405 }
406 ocu[++u_len] = (Uint8)(uni_char & 0xffU);
407 }
408
409 ocu[length - 1] = (Uint8)u_len + 1;
410 return u_len + 1;
411 }
412
413 int udf_get_filename(struct super_block *sb, Uint8 *sname, Uint8 *dname, int flen)
414 {
415 struct ustr filename, unifilename;
416 int len;
417
418 if (udf_build_ustr_exact(&unifilename, sname, flen))
419 {
420 return 0;
421 }
422
423 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8))
424 {
425 if (!udf_CS0toUTF8(&filename, &unifilename) )
426 {
427 udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
428 return 0;
429 }
430 }
431 else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP))
432 {
433 if (!udf_CS0toNLS(UDF_SB(sb)->s_nls_map, &filename, &unifilename) )
434 {
435 udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
436 return 0;
437 }
438 }
439 else
440 return 0;
441
442 if ((len = udf_translate_to_linux(dname, filename.u_name, filename.u_len,
443 unifilename.u_name, unifilename.u_len)))
444 {
445 return len;
446 }
447 return 0;
448 }
449 #endif
450
451 #define ILLEGAL_CHAR_MARK '_'
452 #define EXT_MARK '.'
453 #define CRC_MARK '#'
454 #define EXT_SIZE 5
455
456 int udf_translate_to_linux(Uint8 *newName, Uint8 *udfName, int udfLen, Uint8 *fidName, int fidNameLen)
457 {
458 int index, newIndex = 0, needsCRC = 0;
459 int extIndex = 0, newExtIndex = 0, hasExt = 0;
460 unsigned short valueCRC;
461 Uint8 curr;
462 const Uint8 hexChar[] = "0123456789ABCDEF";
463
464 if (udfName[0] == '.' && (udfLen == 1 ||
465 (udfLen == 2 && udfName[1] == '.')))
466 {
467 needsCRC = 1;
468 newIndex = udfLen;
469 memcpy(newName, udfName, udfLen);
470 }
471 else
472 {
473 for (index = 0; index < udfLen; index++)
474 {
475 curr = udfName[index];
476 if (curr == '/' || curr == 0)
477 {
478 needsCRC = 1;
479 curr = ILLEGAL_CHAR_MARK;
480 while (index+1 < udfLen && (udfName[index+1] == '/' ||
481 udfName[index+1] == 0))
482 index++;
483 }
484 if (curr == EXT_MARK && (udfLen - index - 1) <= EXT_SIZE)
485 {
486 if (udfLen == index + 1)
487 hasExt = 0;
488 else
489 {
490 hasExt = 1;
491 extIndex = index;
492 newExtIndex = newIndex;
493 }
494 }
495 if (newIndex < 256)
496 newName[newIndex++] = curr;
497 else
498 needsCRC = 1;
499 }
500 }
501 if (needsCRC)
502 {
503 Uint8 ext[EXT_SIZE];
504 int localExtIndex = 0;
505
506 if (hasExt)
507 {
508 int maxFilenameLen;
509 for(index = 0; index<EXT_SIZE && extIndex + index +1 < udfLen;
510 index++ )
511 {
512 curr = udfName[extIndex + index + 1];
513
514 if (curr == '/' || curr == 0)
515 {
516 needsCRC = 1;
517 curr = ILLEGAL_CHAR_MARK;
518 while(extIndex + index + 2 < udfLen && (index + 1 < EXT_SIZE
519 && (udfName[extIndex + index + 2] == '/' ||
520 udfName[extIndex + index + 2] == 0)))
521 index++;
522 }
523 ext[localExtIndex++] = curr;
524 }
525 maxFilenameLen = 250 - localExtIndex;
526 if (newIndex > maxFilenameLen)
527 newIndex = maxFilenameLen;
528 else
529 newIndex = newExtIndex;
530 }
531 else if (newIndex > 250)
532 newIndex = 250;
533 newName[newIndex++] = CRC_MARK;
534 valueCRC = udf_crc(fidName, fidNameLen, 0);
535 newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
536 newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
537 newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
538 newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
539
540 if (hasExt)
541 {
542 newName[newIndex++] = EXT_MARK;
543 for (index = 0;index < localExtIndex ;index++ )
544 newName[newIndex++] = ext[index];
545 }
546 }
547 return newIndex;
548 }
549