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 #include "udfdecl.h"
27 
28 #include <linux/kernel.h>
29 #include <linux/string.h>	/* for memset */
30 #include <linux/nls.h>
31 #include <linux/udf_fs.h>
32 
33 #include "udf_sb.h"
34 
udf_ustr_to_dchars(uint8_t * dest,const struct ustr * src,int strlen)35 int udf_ustr_to_dchars(uint8_t *dest, const struct ustr *src, int strlen)
36 {
37 	if ( (!dest) || (!src) || (!strlen) || (src->u_len > strlen) )
38 		return 0;
39 	memcpy(dest+1, src->u_name, src->u_len);
40 	dest[0] = src->u_cmpID;
41 	return src->u_len + 1;
42 }
43 
udf_ustr_to_char(uint8_t * dest,const struct ustr * src,int strlen)44 int udf_ustr_to_char(uint8_t *dest, const struct ustr *src, int strlen)
45 {
46 	if ( (!dest) || (!src) || (!strlen) || (src->u_len >= strlen) )
47 		return 0;
48 	memcpy(dest, src->u_name, src->u_len);
49 	return src->u_len;
50 }
51 
udf_ustr_to_dstring(dstring * dest,const struct ustr * src,int dlength)52 int udf_ustr_to_dstring(dstring *dest, const struct ustr *src, int dlength)
53 {
54 	if ( udf_ustr_to_dchars(dest, src, dlength-1) )
55 	{
56 		dest[dlength-1] = src->u_len + 1;
57 		return dlength;
58 	}
59 	else
60 		return 0;
61 }
62 
udf_dchars_to_ustr(struct ustr * dest,const uint8_t * src,int strlen)63 int udf_dchars_to_ustr(struct ustr *dest, const uint8_t *src, int strlen)
64 {
65 	if ( (!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN) )
66 		return 0;
67 	memset(dest, 0, sizeof(struct ustr));
68 	memcpy(dest->u_name, src+1, strlen-1);
69 	dest->u_cmpID = src[0];
70 	dest->u_len = strlen-1;
71 	return strlen-1;
72 }
73 
udf_char_to_ustr(struct ustr * dest,const uint8_t * src,int strlen)74 int udf_char_to_ustr(struct ustr *dest, const uint8_t *src, int strlen)
75 {
76 	if ( (!dest) || (!src) || (!strlen) || (strlen >= UDF_NAME_LEN) )
77 		return 0;
78 	memset(dest, 0, sizeof(struct ustr));
79 	memcpy(dest->u_name, src, strlen);
80 	dest->u_cmpID = 0x08;
81 	dest->u_len = strlen;
82 	return strlen;
83 }
84 
85 
udf_dstring_to_ustr(struct ustr * dest,const dstring * src,int dlength)86 int udf_dstring_to_ustr(struct ustr *dest, const dstring *src, int dlength)
87 {
88 	if ( dlength && udf_dchars_to_ustr(dest, src, src[dlength-1]) )
89 		return dlength;
90 	else
91 		return 0;
92 }
93 
94 /*
95  * udf_build_ustr
96  */
udf_build_ustr(struct ustr * dest,dstring * ptr,int size)97 int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
98 {
99 	int usesize;
100 
101 	if ( (!dest) || (!ptr) || (!size) )
102 		return -1;
103 
104 	memset(dest, 0, sizeof(struct ustr));
105 	usesize= (size > UDF_NAME_LEN) ? UDF_NAME_LEN : size;
106 	dest->u_cmpID=ptr[0];
107 	dest->u_len=ptr[size-1];
108 	memcpy(dest->u_name, ptr+1, usesize-1);
109 	return 0;
110 }
111 
112 /*
113  * udf_build_ustr_exact
114  */
udf_build_ustr_exact(struct ustr * dest,dstring * ptr,int exactsize)115 int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
116 {
117 	if ( (!dest) || (!ptr) || (!exactsize) )
118 		return -1;
119 
120 	memset(dest, 0, sizeof(struct ustr));
121 	dest->u_cmpID=ptr[0];
122 	dest->u_len=exactsize-1;
123 	memcpy(dest->u_name, ptr+1, exactsize-1);
124 	return 0;
125 }
126 
127 /*
128  * udf_ocu_to_utf8
129  *
130  * PURPOSE
131  *	Convert OSTA Compressed Unicode to the UTF-8 equivalent.
132  *
133  * DESCRIPTION
134  *	This routine is only called by udf_filldir().
135  *
136  * PRE-CONDITIONS
137  *	utf			Pointer to UTF-8 output buffer.
138  *	ocu			Pointer to OSTA Compressed Unicode input buffer
139  *				of size UDF_NAME_LEN bytes.
140  * 				both of type "struct ustr *"
141  *
142  * POST-CONDITIONS
143  *	<return>		Zero on success.
144  *
145  * HISTORY
146  *	November 12, 1997 - Andrew E. Mileski
147  *	Written, tested, and released.
148  */
udf_CS0toUTF8(struct ustr * utf_o,struct ustr * ocu_i)149 int udf_CS0toUTF8(struct ustr *utf_o, struct ustr *ocu_i)
150 {
151 	uint8_t *ocu;
152 	uint32_t c;
153 	uint8_t cmp_id, ocu_len;
154 	int i;
155 
156 	ocu = ocu_i->u_name;
157 
158 	ocu_len = ocu_i->u_len;
159 	cmp_id = ocu_i->u_cmpID;
160 	utf_o->u_len = 0;
161 
162 	if (ocu_len == 0)
163 	{
164 		memset(utf_o, 0, sizeof(struct ustr));
165 		utf_o->u_cmpID = 0;
166 		utf_o->u_len = 0;
167 		return 0;
168 	}
169 
170 	if ((cmp_id != 8) && (cmp_id != 16))
171 	{
172 		printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name);
173 		return 0;
174 	}
175 
176 	for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN-3)) ;)
177 	{
178 
179 		/* Expand OSTA compressed Unicode to Unicode */
180 		c = ocu[i++];
181 		if (cmp_id == 16)
182 			c = (c << 8) | ocu[i++];
183 
184 		/* Compress Unicode to UTF-8 */
185 		if (c < 0x80U)
186 			utf_o->u_name[utf_o->u_len++] = (uint8_t)c;
187 		else if (c < 0x800U)
188 		{
189 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0xc0 | (c >> 6));
190 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0x80 | (c & 0x3f));
191 		}
192 		else
193 		{
194 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0xe0 | (c >> 12));
195 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0x80 | ((c >> 6) & 0x3f));
196 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0x80 | (c & 0x3f));
197 		}
198 	}
199 	utf_o->u_cmpID=8;
200 
201 	return utf_o->u_len;
202 }
203 
204 /*
205  *
206  * udf_utf8_to_ocu
207  *
208  * PURPOSE
209  *	Convert UTF-8 to the OSTA Compressed Unicode equivalent.
210  *
211  * DESCRIPTION
212  *	This routine is only called by udf_lookup().
213  *
214  * PRE-CONDITIONS
215  *	ocu			Pointer to OSTA Compressed Unicode output
216  *				buffer of size UDF_NAME_LEN bytes.
217  *	utf			Pointer to UTF-8 input buffer.
218  *	utf_len			Length of UTF-8 input buffer in bytes.
219  *
220  * POST-CONDITIONS
221  *	<return>		Zero on success.
222  *
223  * HISTORY
224  *	November 12, 1997 - Andrew E. Mileski
225  *	Written, tested, and released.
226  */
udf_UTF8toCS0(dstring * ocu,struct ustr * utf,int length)227 int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length)
228 {
229 	unsigned c, i, max_val, utf_char;
230 	int utf_cnt;
231 	int u_len = 0;
232 
233 	memset(ocu, 0, sizeof(dstring) * length);
234 	ocu[0] = 8;
235 	max_val = 0xffU;
236 
237 try_again:
238 	utf_char = 0U;
239 	utf_cnt = 0U;
240 	for (i = 0U; i < utf->u_len; i++)
241 	{
242 		c = (uint8_t)utf->u_name[i];
243 
244 		/* Complete a multi-byte UTF-8 character */
245 		if (utf_cnt)
246 		{
247 			utf_char = (utf_char << 6) | (c & 0x3fU);
248 			if (--utf_cnt)
249 				continue;
250 		}
251 		else
252 		{
253 			/* Check for a multi-byte UTF-8 character */
254 			if (c & 0x80U)
255 			{
256 				/* Start a multi-byte UTF-8 character */
257 				if ((c & 0xe0U) == 0xc0U)
258 				{
259 					utf_char = c & 0x1fU;
260 					utf_cnt = 1;
261 				}
262 				else if ((c & 0xf0U) == 0xe0U)
263 				{
264 					utf_char = c & 0x0fU;
265 					utf_cnt = 2;
266 				}
267 				else if ((c & 0xf8U) == 0xf0U)
268 				{
269 					utf_char = c & 0x07U;
270 					utf_cnt = 3;
271 				}
272 				else if ((c & 0xfcU) == 0xf8U)
273 				{
274 					utf_char = c & 0x03U;
275 					utf_cnt = 4;
276 				}
277 				else if ((c & 0xfeU) == 0xfcU)
278 				{
279 					utf_char = c & 0x01U;
280 					utf_cnt = 5;
281 				}
282 				else
283 					goto error_out;
284 				continue;
285 			} else
286 				/* Single byte UTF-8 character (most common) */
287 				utf_char = c;
288 		}
289 
290 		/* Choose no compression if necessary */
291 		if (utf_char > max_val)
292 		{
293 			if ( 0xffU == max_val )
294 			{
295 				max_val = 0xffffU;
296 				ocu[0] = (uint8_t)0x10U;
297 				goto try_again;
298 			}
299 			goto error_out;
300 		}
301 
302 		if (max_val == 0xffffU)
303 		{
304 			ocu[++u_len] = (uint8_t)(utf_char >> 8);
305 		}
306 		ocu[++u_len] = (uint8_t)(utf_char & 0xffU);
307 	}
308 
309 
310 	if (utf_cnt)
311 	{
312 error_out:
313 		printk(KERN_ERR "udf: bad UTF-8 character\n");
314 		return 0;
315 	}
316 
317 	ocu[length - 1] = (uint8_t)u_len + 1;
318 	return u_len + 1;
319 }
320 
udf_CS0toNLS(struct nls_table * nls,struct ustr * utf_o,struct ustr * ocu_i)321 int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o, struct ustr *ocu_i)
322 {
323 	uint8_t *ocu;
324 	uint32_t c;
325 	uint8_t cmp_id, ocu_len;
326 	int i;
327 
328 	ocu = ocu_i->u_name;
329 
330 	ocu_len = ocu_i->u_len;
331 	cmp_id = ocu_i->u_cmpID;
332 	utf_o->u_len = 0;
333 
334 	if (ocu_len == 0)
335 	{
336 		memset(utf_o, 0, sizeof(struct ustr));
337 		utf_o->u_cmpID = 0;
338 		utf_o->u_len = 0;
339 		return 0;
340 	}
341 
342 	if ((cmp_id != 8) && (cmp_id != 16))
343 	{
344 		printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name);
345 		return 0;
346 	}
347 
348 	for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN-3)) ;)
349 	{
350 		/* Expand OSTA compressed Unicode to Unicode */
351 		c = ocu[i++];
352 		if (cmp_id == 16)
353 			c = (c << 8) | ocu[i++];
354 
355 		utf_o->u_len += nls->uni2char(c, &utf_o->u_name[utf_o->u_len],
356 			UDF_NAME_LEN - utf_o->u_len);
357 	}
358 	utf_o->u_cmpID=8;
359 
360 	return utf_o->u_len;
361 }
362 
udf_NLStoCS0(struct nls_table * nls,dstring * ocu,struct ustr * uni,int length)363 int udf_NLStoCS0(struct nls_table *nls, dstring *ocu, struct ustr *uni, int length)
364 {
365 	unsigned len, i, max_val;
366 	uint16_t uni_char;
367 	int uni_cnt;
368 	int u_len = 0;
369 
370 	memset(ocu, 0, sizeof(dstring) * length);
371 	ocu[0] = 8;
372 	max_val = 0xffU;
373 
374 try_again:
375 	uni_char = 0U;
376 	uni_cnt = 0U;
377 	for (i = 0U; i < uni->u_len; i++)
378 	{
379 		len = nls->char2uni(&uni->u_name[i], uni->u_len-i, &uni_char);
380 
381 		if (len == 2 && max_val == 0xff)
382 		{
383 			max_val = 0xffffU;
384 			ocu[0] = (uint8_t)0x10U;
385 			goto try_again;
386 		}
387 
388 		if (max_val == 0xffffU)
389 		{
390 			ocu[++u_len] = (uint8_t)(uni_char >> 8);
391 			i++;
392 		}
393 		ocu[++u_len] = (uint8_t)(uni_char & 0xffU);
394 	}
395 
396 	ocu[length - 1] = (uint8_t)u_len + 1;
397 	return u_len + 1;
398 }
399 
udf_get_filename(struct super_block * sb,uint8_t * sname,uint8_t * dname,int flen)400 int udf_get_filename(struct super_block *sb, uint8_t *sname, uint8_t *dname, int flen)
401 {
402 	struct ustr filename, unifilename;
403 	int len;
404 
405 	if (udf_build_ustr_exact(&unifilename, sname, flen))
406 	{
407 		return 0;
408 	}
409 
410 	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8))
411 	{
412 		if (!udf_CS0toUTF8(&filename, &unifilename) )
413 		{
414 			udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
415 			return 0;
416 		}
417 	}
418 	else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP))
419 	{
420 		if (!udf_CS0toNLS(UDF_SB(sb)->s_nls_map, &filename, &unifilename) )
421 		{
422 			udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
423 			return 0;
424 		}
425 	}
426 	else
427 		return 0;
428 
429 	if ((len = udf_translate_to_linux(dname, filename.u_name, filename.u_len,
430 		unifilename.u_name, unifilename.u_len)))
431 	{
432 		return len;
433 	}
434 	return 0;
435 }
436 
437 #define ILLEGAL_CHAR_MARK	'_'
438 #define EXT_MARK			'.'
439 #define CRC_MARK			'#'
440 #define EXT_SIZE			5
441 
udf_translate_to_linux(uint8_t * newName,uint8_t * udfName,int udfLen,uint8_t * fidName,int fidNameLen)442 int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName, int udfLen, uint8_t *fidName, int fidNameLen)
443 {
444 	int index, newIndex = 0, needsCRC = 0;
445 	int extIndex = 0, newExtIndex = 0, hasExt = 0;
446 	unsigned short valueCRC;
447 	uint8_t curr;
448 	const uint8_t hexChar[] = "0123456789ABCDEF";
449 
450 	if (udfName[0] == '.' && (udfLen == 1 ||
451 		(udfLen == 2 && udfName[1] == '.')))
452 	{
453 		needsCRC = 1;
454 		newIndex = udfLen;
455 		memcpy(newName, udfName, udfLen);
456 	}
457 	else
458 	{
459 		for (index = 0; index < udfLen; index++)
460 		{
461 			curr = udfName[index];
462 			if (curr == '/' || curr == 0)
463 			{
464 				needsCRC = 1;
465 				curr = ILLEGAL_CHAR_MARK;
466 				while (index+1 < udfLen && (udfName[index+1] == '/' ||
467 					udfName[index+1] == 0))
468 					index++;
469 			}
470 			if (curr == EXT_MARK && (udfLen - index - 1) <= EXT_SIZE)
471 			{
472 				if (udfLen == index + 1)
473 					hasExt = 0;
474 				else
475 				{
476 					hasExt = 1;
477 					extIndex = index;
478 					newExtIndex = newIndex;
479 				}
480 			}
481 			if (newIndex < 256)
482 				newName[newIndex++] = curr;
483 			else
484 				needsCRC = 1;
485 		}
486 	}
487 	if (needsCRC)
488 	{
489 		uint8_t ext[EXT_SIZE];
490 		int localExtIndex = 0;
491 
492 		if (hasExt)
493 		{
494 			int maxFilenameLen;
495 			for(index = 0; index<EXT_SIZE && extIndex + index +1 < udfLen;
496 				index++ )
497 			{
498 				curr = udfName[extIndex + index + 1];
499 
500 				if (curr == '/' || curr == 0)
501 				{
502 					needsCRC = 1;
503 					curr = ILLEGAL_CHAR_MARK;
504 					while(extIndex + index + 2 < udfLen && (index + 1 < EXT_SIZE
505 						&& (udfName[extIndex + index + 2] == '/' ||
506 							udfName[extIndex + index + 2] == 0)))
507 						index++;
508 				}
509 				ext[localExtIndex++] = curr;
510 			}
511 			maxFilenameLen = 250 - localExtIndex;
512 			if (newIndex > maxFilenameLen)
513 				newIndex = maxFilenameLen;
514 			else
515 				newIndex = newExtIndex;
516 		}
517 		else if (newIndex > 250)
518 			newIndex = 250;
519 		newName[newIndex++] = CRC_MARK;
520 		valueCRC = udf_crc(fidName, fidNameLen, 0);
521 		newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
522 		newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
523 		newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
524 		newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
525 
526 		if (hasExt)
527 		{
528 			newName[newIndex++] = EXT_MARK;
529 			for (index = 0;index < localExtIndex ;index++ )
530 				newName[newIndex++] = ext[index];
531 		}
532 	}
533 	return newIndex;
534 }
535