xref: /openbmc/linux/fs/udf/unicode.c (revision c21b37f6)
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/udf_fs.h>
27 
28 #include "udf_sb.h"
29 
30 static int udf_translate_to_linux(uint8_t *, uint8_t *, int, uint8_t *, int);
31 
32 static int udf_char_to_ustr(struct ustr *dest, const uint8_t *src, int strlen)
33 {
34 	if ((!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN - 2))
35 		return 0;
36 
37 	memset(dest, 0, sizeof(struct ustr));
38 	memcpy(dest->u_name, src, strlen);
39 	dest->u_cmpID = 0x08;
40 	dest->u_len = strlen;
41 
42 	return strlen;
43 }
44 
45 /*
46  * udf_build_ustr
47  */
48 int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
49 {
50 	int usesize;
51 
52 	if ((!dest) || (!ptr) || (!size))
53 		return -1;
54 
55 	memset(dest, 0, sizeof(struct ustr));
56 	usesize = (size > UDF_NAME_LEN) ? UDF_NAME_LEN : size;
57 	dest->u_cmpID = ptr[0];
58 	dest->u_len = ptr[size - 1];
59 	memcpy(dest->u_name, ptr + 1, usesize - 1);
60 
61 	return 0;
62 }
63 
64 /*
65  * udf_build_ustr_exact
66  */
67 static int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
68 {
69 	if ((!dest) || (!ptr) || (!exactsize))
70 		return -1;
71 
72 	memset(dest, 0, sizeof(struct ustr));
73 	dest->u_cmpID = ptr[0];
74 	dest->u_len = exactsize - 1;
75 	memcpy(dest->u_name, ptr + 1, exactsize - 1);
76 
77 	return 0;
78 }
79 
80 /*
81  * udf_ocu_to_utf8
82  *
83  * PURPOSE
84  *	Convert OSTA Compressed Unicode to the UTF-8 equivalent.
85  *
86  * DESCRIPTION
87  *	This routine is only called by udf_filldir().
88  *
89  * PRE-CONDITIONS
90  *	utf			Pointer to UTF-8 output buffer.
91  *	ocu			Pointer to OSTA Compressed Unicode input buffer
92  *				of size UDF_NAME_LEN bytes.
93  * 				both of type "struct ustr *"
94  *
95  * POST-CONDITIONS
96  *	<return>		Zero on success.
97  *
98  * HISTORY
99  *	November 12, 1997 - Andrew E. Mileski
100  *	Written, tested, and released.
101  */
102 int udf_CS0toUTF8(struct ustr *utf_o, struct ustr *ocu_i)
103 {
104 	uint8_t *ocu;
105 	uint32_t c;
106 	uint8_t cmp_id, ocu_len;
107 	int i;
108 
109 	ocu = ocu_i->u_name;
110 
111 	ocu_len = ocu_i->u_len;
112 	cmp_id = ocu_i->u_cmpID;
113 	utf_o->u_len = 0;
114 
115 	if (ocu_len == 0) {
116 		memset(utf_o, 0, sizeof(struct ustr));
117 		utf_o->u_cmpID = 0;
118 		utf_o->u_len = 0;
119 		return 0;
120 	}
121 
122 	if ((cmp_id != 8) && (cmp_id != 16)) {
123 		printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n",
124 		       cmp_id, ocu_i->u_name);
125 		return 0;
126 	}
127 
128 	for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
129 
130 		/* Expand OSTA compressed Unicode to Unicode */
131 		c = ocu[i++];
132 		if (cmp_id == 16)
133 			c = (c << 8) | ocu[i++];
134 
135 		/* Compress Unicode to UTF-8 */
136 		if (c < 0x80U) {
137 			utf_o->u_name[utf_o->u_len++] = (uint8_t)c;
138 		} else if (c < 0x800U) {
139 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0xc0 | (c >> 6));
140 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0x80 | (c & 0x3f));
141 		} else {
142 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0xe0 | (c >> 12));
143 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0x80 | ((c >> 6) & 0x3f));
144 			utf_o->u_name[utf_o->u_len++] = (uint8_t)(0x80 | (c & 0x3f));
145 		}
146 	}
147 	utf_o->u_cmpID = 8;
148 
149 	return utf_o->u_len;
150 }
151 
152 /*
153  *
154  * udf_utf8_to_ocu
155  *
156  * PURPOSE
157  *	Convert UTF-8 to the OSTA Compressed Unicode equivalent.
158  *
159  * DESCRIPTION
160  *	This routine is only called by udf_lookup().
161  *
162  * PRE-CONDITIONS
163  *	ocu			Pointer to OSTA Compressed Unicode output
164  *				buffer of size UDF_NAME_LEN bytes.
165  *	utf			Pointer to UTF-8 input buffer.
166  *	utf_len			Length of UTF-8 input buffer in bytes.
167  *
168  * POST-CONDITIONS
169  *	<return>		Zero on success.
170  *
171  * HISTORY
172  *	November 12, 1997 - Andrew E. Mileski
173  *	Written, tested, and released.
174  */
175 static int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length)
176 {
177 	unsigned c, i, max_val, utf_char;
178 	int utf_cnt, u_len;
179 
180 	memset(ocu, 0, sizeof(dstring) * length);
181 	ocu[0] = 8;
182 	max_val = 0xffU;
183 
184 try_again:
185 	u_len = 0U;
186 	utf_char = 0U;
187 	utf_cnt = 0U;
188 	for (i = 0U; i < utf->u_len; i++) {
189 		c = (uint8_t)utf->u_name[i];
190 
191 		/* Complete a multi-byte UTF-8 character */
192 		if (utf_cnt) {
193 			utf_char = (utf_char << 6) | (c & 0x3fU);
194 			if (--utf_cnt)
195 				continue;
196 		} else {
197 			/* Check for a multi-byte UTF-8 character */
198 			if (c & 0x80U) {
199 				/* Start a multi-byte UTF-8 character */
200 				if ((c & 0xe0U) == 0xc0U) {
201 					utf_char = c & 0x1fU;
202 					utf_cnt = 1;
203 				} else if ((c & 0xf0U) == 0xe0U) {
204 					utf_char = c & 0x0fU;
205 					utf_cnt = 2;
206 				} else if ((c & 0xf8U) == 0xf0U) {
207 					utf_char = c & 0x07U;
208 					utf_cnt = 3;
209 				} else if ((c & 0xfcU) == 0xf8U) {
210 					utf_char = c & 0x03U;
211 					utf_cnt = 4;
212 				} else if ((c & 0xfeU) == 0xfcU) {
213 					utf_char = c & 0x01U;
214 					utf_cnt = 5;
215 				} else {
216 					goto error_out;
217 				}
218 				continue;
219 			} else {
220 				/* Single byte UTF-8 character (most common) */
221 				utf_char = c;
222 			}
223 		}
224 
225 		/* Choose no compression if necessary */
226 		if (utf_char > max_val) {
227 			if (max_val == 0xffU) {
228 				max_val = 0xffffU;
229 				ocu[0] = (uint8_t)0x10U;
230 				goto try_again;
231 			}
232 			goto error_out;
233 		}
234 
235 		if (max_val == 0xffffU) {
236 			ocu[++u_len] = (uint8_t)(utf_char >> 8);
237 		}
238 		ocu[++u_len] = (uint8_t)(utf_char & 0xffU);
239 	}
240 
241 	if (utf_cnt) {
242 error_out:
243 		ocu[++u_len] = '?';
244 		printk(KERN_DEBUG "udf: bad UTF-8 character\n");
245 	}
246 
247 	ocu[length - 1] = (uint8_t)u_len + 1;
248 
249 	return u_len + 1;
250 }
251 
252 static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o,
253 			struct ustr *ocu_i)
254 {
255 	uint8_t *ocu;
256 	uint32_t c;
257 	uint8_t cmp_id, ocu_len;
258 	int i;
259 
260 	ocu = ocu_i->u_name;
261 
262 	ocu_len = ocu_i->u_len;
263 	cmp_id = ocu_i->u_cmpID;
264 	utf_o->u_len = 0;
265 
266 	if (ocu_len == 0) {
267 		memset(utf_o, 0, sizeof(struct ustr));
268 		utf_o->u_cmpID = 0;
269 		utf_o->u_len = 0;
270 		return 0;
271 	}
272 
273 	if ((cmp_id != 8) && (cmp_id != 16)) {
274 		printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n",
275 		       cmp_id, ocu_i->u_name);
276 		return 0;
277 	}
278 
279 	for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
280 		/* Expand OSTA compressed Unicode to Unicode */
281 		c = ocu[i++];
282 		if (cmp_id == 16)
283 			c = (c << 8) | ocu[i++];
284 
285 		utf_o->u_len += nls->uni2char(c, &utf_o->u_name[utf_o->u_len],
286 					      UDF_NAME_LEN - utf_o->u_len);
287 	}
288 	utf_o->u_cmpID = 8;
289 
290 	return utf_o->u_len;
291 }
292 
293 static int udf_NLStoCS0(struct nls_table *nls, dstring *ocu, struct ustr *uni,
294 			int length)
295 {
296 	unsigned len, i, max_val;
297 	uint16_t uni_char;
298 	int u_len;
299 
300 	memset(ocu, 0, sizeof(dstring) * length);
301 	ocu[0] = 8;
302 	max_val = 0xffU;
303 
304 try_again:
305 	u_len = 0U;
306 	for (i = 0U; i < uni->u_len; i++) {
307 		len = nls->char2uni(&uni->u_name[i], uni->u_len - i, &uni_char);
308 		if (len <= 0)
309 			continue;
310 
311 		if (uni_char > max_val) {
312 			max_val = 0xffffU;
313 			ocu[0] = (uint8_t)0x10U;
314 			goto try_again;
315 		}
316 
317 		if (max_val == 0xffffU)
318 			ocu[++u_len] = (uint8_t)(uni_char >> 8);
319 		ocu[++u_len] = (uint8_t)(uni_char & 0xffU);
320 		i += len - 1;
321 	}
322 
323 	ocu[length - 1] = (uint8_t)u_len + 1;
324 	return u_len + 1;
325 }
326 
327 int udf_get_filename(struct super_block *sb, uint8_t *sname, uint8_t *dname,
328 		     int flen)
329 {
330 	struct ustr filename, unifilename;
331 	int len;
332 
333 	if (udf_build_ustr_exact(&unifilename, sname, flen)) {
334 		return 0;
335 	}
336 
337 	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
338 		if (!udf_CS0toUTF8(&filename, &unifilename)) {
339 			udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
340 			return 0;
341 		}
342 	} else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
343 		if (!udf_CS0toNLS(UDF_SB(sb)->s_nls_map, &filename, &unifilename)) {
344 			udf_debug("Failed in udf_get_filename: sname = %s\n", sname);
345 			return 0;
346 		}
347 	} else {
348 		return 0;
349 	}
350 
351 	len = udf_translate_to_linux(dname, filename.u_name, filename.u_len,
352 				     unifilename.u_name, unifilename.u_len);
353 	if (len) {
354 		return len;
355 	}
356 
357 	return 0;
358 }
359 
360 int udf_put_filename(struct super_block *sb, const uint8_t *sname,
361 		     uint8_t *dname, int flen)
362 {
363 	struct ustr unifilename;
364 	int namelen;
365 
366 	if (!(udf_char_to_ustr(&unifilename, sname, flen))) {
367 		return 0;
368 	}
369 
370 	if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
371 		namelen = udf_UTF8toCS0(dname, &unifilename, UDF_NAME_LEN);
372 		if (!namelen) {
373 			return 0;
374 		}
375 	} else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
376 		namelen = udf_NLStoCS0(UDF_SB(sb)->s_nls_map, dname, &unifilename, UDF_NAME_LEN);
377 		if (!namelen) {
378 			return 0;
379 		}
380 	} else {
381 		return 0;
382 	}
383 
384 	return namelen;
385 }
386 
387 #define ILLEGAL_CHAR_MARK	'_'
388 #define EXT_MARK		'.'
389 #define CRC_MARK		'#'
390 #define EXT_SIZE 		5
391 
392 static int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName, int udfLen,
393 				  uint8_t *fidName, int fidNameLen)
394 {
395 	int index, newIndex = 0, needsCRC = 0;
396 	int extIndex = 0, newExtIndex = 0, hasExt = 0;
397 	unsigned short valueCRC;
398 	uint8_t curr;
399 	const uint8_t hexChar[] = "0123456789ABCDEF";
400 
401 	if (udfName[0] == '.' &&
402 	    (udfLen == 1 || (udfLen == 2 && udfName[1] == '.'))) {
403 		needsCRC = 1;
404 		newIndex = udfLen;
405 		memcpy(newName, udfName, udfLen);
406 	} else {
407 		for (index = 0; index < udfLen; index++) {
408 			curr = udfName[index];
409 			if (curr == '/' || curr == 0) {
410 				needsCRC = 1;
411 				curr = ILLEGAL_CHAR_MARK;
412 				while (index + 1 < udfLen && (udfName[index + 1] == '/' ||
413 							      udfName[index + 1] == 0))
414 					index++;
415 			} if (curr == EXT_MARK && (udfLen - index - 1) <= EXT_SIZE) {
416 				if (udfLen == index + 1) {
417 					hasExt = 0;
418 				} else {
419 					hasExt = 1;
420 					extIndex = index;
421 					newExtIndex = newIndex;
422 				}
423 			}
424 			if (newIndex < 256)
425 				newName[newIndex++] = curr;
426 			else
427 				needsCRC = 1;
428 		}
429 	}
430 	if (needsCRC) {
431 		uint8_t ext[EXT_SIZE];
432 		int localExtIndex = 0;
433 
434 		if (hasExt) {
435 			int maxFilenameLen;
436 			for(index = 0; index < EXT_SIZE && extIndex + index + 1 < udfLen; index++) {
437 				curr = udfName[extIndex + index + 1];
438 
439 				if (curr == '/' || curr == 0) {
440 					needsCRC = 1;
441 					curr = ILLEGAL_CHAR_MARK;
442 					while(extIndex + index + 2 < udfLen &&
443 					      (index + 1 < EXT_SIZE
444 					       && (udfName[extIndex + index + 2] == '/' ||
445 						   udfName[extIndex + index + 2] == 0)))
446 						index++;
447 				}
448 				ext[localExtIndex++] = curr;
449 			}
450 			maxFilenameLen = 250 - localExtIndex;
451 			if (newIndex > maxFilenameLen)
452 				newIndex = maxFilenameLen;
453 			else
454 				newIndex = newExtIndex;
455 		} else if (newIndex > 250) {
456 			newIndex = 250;
457 		}
458 		newName[newIndex++] = CRC_MARK;
459 		valueCRC = udf_crc(fidName, fidNameLen, 0);
460 		newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
461 		newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
462 		newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
463 		newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
464 
465 		if (hasExt) {
466 			newName[newIndex++] = EXT_MARK;
467 			for (index = 0; index < localExtIndex; index++)
468 				newName[newIndex++] = ext[index];
469 		}
470 	}
471 
472 	return newIndex;
473 }
474