Loading...
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
31static int udf_translate_to_linux(uint8_t *, uint8_t *, int, uint8_t *, int);
32
33static int udf_char_to_ustr(struct ustr *dest, const uint8_t *src, int strlen)
34{
35 if ((!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN - 2))
36 return 0;
37
38 memset(dest, 0, sizeof(struct ustr));
39 memcpy(dest->u_name, src, strlen);
40 dest->u_cmpID = 0x08;
41 dest->u_len = strlen;
42
43 return strlen;
44}
45
46/*
47 * udf_build_ustr
48 */
49int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
50{
51 int usesize;
52
53 if (!dest || !ptr || !size)
54 return -1;
55 BUG_ON(size < 2);
56
57 usesize = min_t(size_t, ptr[size - 1], sizeof(dest->u_name));
58 usesize = min(usesize, size - 2);
59 dest->u_cmpID = ptr[0];
60 dest->u_len = usesize;
61 memcpy(dest->u_name, ptr + 1, usesize);
62 memset(dest->u_name + usesize, 0, sizeof(dest->u_name) - usesize);
63
64 return 0;
65}
66
67/*
68 * udf_build_ustr_exact
69 */
70static int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
71{
72 if ((!dest) || (!ptr) || (!exactsize))
73 return -1;
74
75 memset(dest, 0, sizeof(struct ustr));
76 dest->u_cmpID = ptr[0];
77 dest->u_len = exactsize - 1;
78 memcpy(dest->u_name, ptr + 1, exactsize - 1);
79
80 return 0;
81}
82
83/*
84 * udf_ocu_to_utf8
85 *
86 * PURPOSE
87 * Convert OSTA Compressed Unicode to the UTF-8 equivalent.
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 */
102int udf_CS0toUTF8(struct ustr *utf_o, const struct ustr *ocu_i)
103{
104 const uint8_t *ocu;
105 uint8_t cmp_id, ocu_len;
106 int i;
107
108 ocu_len = ocu_i->u_len;
109 if (ocu_len == 0) {
110 memset(utf_o, 0, sizeof(struct ustr));
111 return 0;
112 }
113
114 cmp_id = ocu_i->u_cmpID;
115 if (cmp_id != 8 && cmp_id != 16) {
116 memset(utf_o, 0, sizeof(struct ustr));
117 pr_err("unknown compression code (%d) stri=%s\n",
118 cmp_id, ocu_i->u_name);
119 return 0;
120 }
121
122 ocu = ocu_i->u_name;
123 utf_o->u_len = 0;
124 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
125
126 /* Expand OSTA compressed Unicode to Unicode */
127 uint32_t c = ocu[i++];
128 if (cmp_id == 16)
129 c = (c << 8) | ocu[i++];
130
131 /* Compress Unicode to UTF-8 */
132 if (c < 0x80U)
133 utf_o->u_name[utf_o->u_len++] = (uint8_t)c;
134 else if (c < 0x800U) {
135 utf_o->u_name[utf_o->u_len++] =
136 (uint8_t)(0xc0 | (c >> 6));
137 utf_o->u_name[utf_o->u_len++] =
138 (uint8_t)(0x80 | (c & 0x3f));
139 } else {
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_utf8_to_ocu
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 */
177static 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
186try_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) {
243error_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
253static 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 0;
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 - 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
296static 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
308try_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
336int udf_get_filename(struct super_block *sb, uint8_t *sname, uint8_t *dname,
337 int flen)
338{
339 struct ustr *filename, *unifilename;
340 int len = 0;
341
342 filename = kmalloc(sizeof(struct ustr), GFP_NOFS);
343 if (!filename)
344 return 0;
345
346 unifilename = kmalloc(sizeof(struct ustr), GFP_NOFS);
347 if (!unifilename)
348 goto out1;
349
350 if (udf_build_ustr_exact(unifilename, sname, flen))
351 goto out2;
352
353 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
354 if (!udf_CS0toUTF8(filename, unifilename)) {
355 udf_debug("Failed in udf_get_filename: sname = %s\n",
356 sname);
357 goto out2;
358 }
359 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
360 if (!udf_CS0toNLS(UDF_SB(sb)->s_nls_map, filename,
361 unifilename)) {
362 udf_debug("Failed in udf_get_filename: sname = %s\n",
363 sname);
364 goto out2;
365 }
366 } else
367 goto out2;
368
369 len = udf_translate_to_linux(dname, filename->u_name, filename->u_len,
370 unifilename->u_name, unifilename->u_len);
371out2:
372 kfree(unifilename);
373out1:
374 kfree(filename);
375 return len;
376}
377
378int udf_put_filename(struct super_block *sb, const uint8_t *sname,
379 uint8_t *dname, int flen)
380{
381 struct ustr unifilename;
382 int namelen;
383
384 if (!udf_char_to_ustr(&unifilename, sname, flen))
385 return 0;
386
387 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
388 namelen = udf_UTF8toCS0(dname, &unifilename, UDF_NAME_LEN);
389 if (!namelen)
390 return 0;
391 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
392 namelen = udf_NLStoCS0(UDF_SB(sb)->s_nls_map, dname,
393 &unifilename, UDF_NAME_LEN);
394 if (!namelen)
395 return 0;
396 } else
397 return 0;
398
399 return namelen;
400}
401
402#define ILLEGAL_CHAR_MARK '_'
403#define EXT_MARK '.'
404#define CRC_MARK '#'
405#define EXT_SIZE 5
406
407static int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName,
408 int udfLen, uint8_t *fidName,
409 int fidNameLen)
410{
411 int index, newIndex = 0, needsCRC = 0;
412 int extIndex = 0, newExtIndex = 0, hasExt = 0;
413 unsigned short valueCRC;
414 uint8_t curr;
415 const uint8_t hexChar[] = "0123456789ABCDEF";
416
417 if (udfName[0] == '.' &&
418 (udfLen == 1 || (udfLen == 2 && udfName[1] == '.'))) {
419 needsCRC = 1;
420 newIndex = udfLen;
421 memcpy(newName, udfName, udfLen);
422 } else {
423 for (index = 0; index < udfLen; index++) {
424 curr = udfName[index];
425 if (curr == '/' || curr == 0) {
426 needsCRC = 1;
427 curr = ILLEGAL_CHAR_MARK;
428 while (index + 1 < udfLen &&
429 (udfName[index + 1] == '/' ||
430 udfName[index + 1] == 0))
431 index++;
432 }
433 if (curr == EXT_MARK &&
434 (udfLen - index - 1) <= EXT_SIZE) {
435 if (udfLen == index + 1)
436 hasExt = 0;
437 else {
438 hasExt = 1;
439 extIndex = index;
440 newExtIndex = newIndex;
441 }
442 }
443 if (newIndex < 256)
444 newName[newIndex++] = curr;
445 else
446 needsCRC = 1;
447 }
448 }
449 if (needsCRC) {
450 uint8_t ext[EXT_SIZE];
451 int localExtIndex = 0;
452
453 if (hasExt) {
454 int maxFilenameLen;
455 for (index = 0;
456 index < EXT_SIZE && extIndex + index + 1 < udfLen;
457 index++) {
458 curr = udfName[extIndex + index + 1];
459
460 if (curr == '/' || curr == 0) {
461 needsCRC = 1;
462 curr = ILLEGAL_CHAR_MARK;
463 while (extIndex + index + 2 < udfLen &&
464 (index + 1 < EXT_SIZE &&
465 (udfName[extIndex + index + 2] == '/' ||
466 udfName[extIndex + index + 2] == 0)))
467 index++;
468 }
469 ext[localExtIndex++] = curr;
470 }
471 maxFilenameLen = 250 - localExtIndex;
472 if (newIndex > maxFilenameLen)
473 newIndex = maxFilenameLen;
474 else
475 newIndex = newExtIndex;
476 } else if (newIndex > 250)
477 newIndex = 250;
478 newName[newIndex++] = CRC_MARK;
479 valueCRC = crc_itu_t(0, fidName, fidNameLen);
480 newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
481 newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
482 newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
483 newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
484
485 if (hasExt) {
486 newName[newIndex++] = EXT_MARK;
487 for (index = 0; index < localExtIndex; index++)
488 newName[newIndex++] = ext[index];
489 }
490 }
491
492 return newIndex;
493}
1// SPDX-License-Identifier: GPL-2.0-only
2/*
3 * unicode.c
4 *
5 * PURPOSE
6 * Routines for converting between UTF-8 and OSTA Compressed Unicode.
7 * Also handles filename mangling
8 *
9 * DESCRIPTION
10 * OSTA Compressed Unicode is explained in the OSTA UDF specification.
11 * http://www.osta.org/
12 * UTF-8 is explained in the IETF RFC XXXX.
13 * ftp://ftp.internic.net/rfc/rfcxxxx.txt
14 *
15 */
16
17#include "udfdecl.h"
18
19#include <linux/kernel.h>
20#include <linux/string.h> /* for memset */
21#include <linux/nls.h>
22#include <linux/crc-itu-t.h>
23#include <linux/slab.h>
24
25#include "udf_sb.h"
26
27#define PLANE_SIZE 0x10000
28#define UNICODE_MAX 0x10ffff
29#define SURROGATE_MASK 0xfffff800
30#define SURROGATE_PAIR 0x0000d800
31#define SURROGATE_LOW 0x00000400
32#define SURROGATE_CHAR_BITS 10
33#define SURROGATE_CHAR_MASK ((1 << SURROGATE_CHAR_BITS) - 1)
34
35#define ILLEGAL_CHAR_MARK '_'
36#define EXT_MARK '.'
37#define CRC_MARK '#'
38#define EXT_SIZE 5
39/* Number of chars we need to store generated CRC to make filename unique */
40#define CRC_LEN 5
41
42static unicode_t get_utf16_char(const uint8_t *str_i, int str_i_max_len,
43 int str_i_idx, int u_ch, unicode_t *ret)
44{
45 unicode_t c;
46 int start_idx = str_i_idx;
47
48 /* Expand OSTA compressed Unicode to Unicode */
49 c = str_i[str_i_idx++];
50 if (u_ch > 1)
51 c = (c << 8) | str_i[str_i_idx++];
52 if ((c & SURROGATE_MASK) == SURROGATE_PAIR) {
53 unicode_t next;
54
55 /* Trailing surrogate char */
56 if (str_i_idx >= str_i_max_len) {
57 c = UNICODE_MAX + 1;
58 goto out;
59 }
60
61 /* Low surrogate must follow the high one... */
62 if (c & SURROGATE_LOW) {
63 c = UNICODE_MAX + 1;
64 goto out;
65 }
66
67 WARN_ON_ONCE(u_ch != 2);
68 next = str_i[str_i_idx++] << 8;
69 next |= str_i[str_i_idx++];
70 if ((next & SURROGATE_MASK) != SURROGATE_PAIR ||
71 !(next & SURROGATE_LOW)) {
72 c = UNICODE_MAX + 1;
73 goto out;
74 }
75
76 c = PLANE_SIZE +
77 ((c & SURROGATE_CHAR_MASK) << SURROGATE_CHAR_BITS) +
78 (next & SURROGATE_CHAR_MASK);
79 }
80out:
81 *ret = c;
82 return str_i_idx - start_idx;
83}
84
85
86static int udf_name_conv_char(uint8_t *str_o, int str_o_max_len,
87 int *str_o_idx,
88 const uint8_t *str_i, int str_i_max_len,
89 int *str_i_idx,
90 int u_ch, int *needsCRC,
91 int (*conv_f)(wchar_t, unsigned char *, int),
92 int translate)
93{
94 unicode_t c;
95 int illChar = 0;
96 int len, gotch = 0;
97
98 while (!gotch && *str_i_idx < str_i_max_len) {
99 if (*str_o_idx >= str_o_max_len) {
100 *needsCRC = 1;
101 return gotch;
102 }
103
104 len = get_utf16_char(str_i, str_i_max_len, *str_i_idx, u_ch,
105 &c);
106 /* These chars cannot be converted. Replace them. */
107 if (c == 0 || c > UNICODE_MAX || (conv_f && c > MAX_WCHAR_T) ||
108 (translate && c == '/')) {
109 illChar = 1;
110 if (!translate)
111 gotch = 1;
112 } else if (illChar)
113 break;
114 else
115 gotch = 1;
116 *str_i_idx += len;
117 }
118 if (illChar) {
119 *needsCRC = 1;
120 c = ILLEGAL_CHAR_MARK;
121 gotch = 1;
122 }
123 if (gotch) {
124 if (conv_f) {
125 len = conv_f(c, &str_o[*str_o_idx],
126 str_o_max_len - *str_o_idx);
127 } else {
128 len = utf32_to_utf8(c, &str_o[*str_o_idx],
129 str_o_max_len - *str_o_idx);
130 if (len < 0)
131 len = -ENAMETOOLONG;
132 }
133 /* Valid character? */
134 if (len >= 0)
135 *str_o_idx += len;
136 else if (len == -ENAMETOOLONG) {
137 *needsCRC = 1;
138 gotch = 0;
139 } else {
140 str_o[(*str_o_idx)++] = ILLEGAL_CHAR_MARK;
141 *needsCRC = 1;
142 }
143 }
144 return gotch;
145}
146
147static int udf_name_from_CS0(struct super_block *sb,
148 uint8_t *str_o, int str_max_len,
149 const uint8_t *ocu, int ocu_len,
150 int translate)
151{
152 uint32_t c;
153 uint8_t cmp_id;
154 int idx, len;
155 int u_ch;
156 int needsCRC = 0;
157 int ext_i_len, ext_max_len;
158 int str_o_len = 0; /* Length of resulting output */
159 int ext_o_len = 0; /* Extension output length */
160 int ext_crc_len = 0; /* Extension output length if used with CRC */
161 int i_ext = -1; /* Extension position in input buffer */
162 int o_crc = 0; /* Rightmost possible output pos for CRC+ext */
163 unsigned short valueCRC;
164 uint8_t ext[EXT_SIZE * NLS_MAX_CHARSET_SIZE + 1];
165 uint8_t crc[CRC_LEN];
166 int (*conv_f)(wchar_t, unsigned char *, int);
167
168 if (str_max_len <= 0)
169 return 0;
170
171 if (ocu_len == 0) {
172 memset(str_o, 0, str_max_len);
173 return 0;
174 }
175
176 if (UDF_SB(sb)->s_nls_map)
177 conv_f = UDF_SB(sb)->s_nls_map->uni2char;
178 else
179 conv_f = NULL;
180
181 cmp_id = ocu[0];
182 if (cmp_id != 8 && cmp_id != 16) {
183 memset(str_o, 0, str_max_len);
184 pr_err("unknown compression code (%u)\n", cmp_id);
185 return -EINVAL;
186 }
187 u_ch = cmp_id >> 3;
188
189 ocu++;
190 ocu_len--;
191
192 if (ocu_len % u_ch) {
193 pr_err("incorrect filename length (%d)\n", ocu_len + 1);
194 return -EINVAL;
195 }
196
197 if (translate) {
198 /* Look for extension */
199 for (idx = ocu_len - u_ch, ext_i_len = 0;
200 (idx >= 0) && (ext_i_len < EXT_SIZE);
201 idx -= u_ch, ext_i_len++) {
202 c = ocu[idx];
203 if (u_ch > 1)
204 c = (c << 8) | ocu[idx + 1];
205
206 if (c == EXT_MARK) {
207 if (ext_i_len)
208 i_ext = idx;
209 break;
210 }
211 }
212 if (i_ext >= 0) {
213 /* Convert extension */
214 ext_max_len = min_t(int, sizeof(ext), str_max_len);
215 ext[ext_o_len++] = EXT_MARK;
216 idx = i_ext + u_ch;
217 while (udf_name_conv_char(ext, ext_max_len, &ext_o_len,
218 ocu, ocu_len, &idx,
219 u_ch, &needsCRC,
220 conv_f, translate)) {
221 if ((ext_o_len + CRC_LEN) < str_max_len)
222 ext_crc_len = ext_o_len;
223 }
224 }
225 }
226
227 idx = 0;
228 while (1) {
229 if (translate && (idx == i_ext)) {
230 if (str_o_len > (str_max_len - ext_o_len))
231 needsCRC = 1;
232 break;
233 }
234
235 if (!udf_name_conv_char(str_o, str_max_len, &str_o_len,
236 ocu, ocu_len, &idx,
237 u_ch, &needsCRC, conv_f, translate))
238 break;
239
240 if (translate &&
241 (str_o_len <= (str_max_len - ext_o_len - CRC_LEN)))
242 o_crc = str_o_len;
243 }
244
245 if (translate) {
246 if (str_o_len > 0 && str_o_len <= 2 && str_o[0] == '.' &&
247 (str_o_len == 1 || str_o[1] == '.'))
248 needsCRC = 1;
249 if (needsCRC) {
250 str_o_len = o_crc;
251 valueCRC = crc_itu_t(0, ocu, ocu_len);
252 crc[0] = CRC_MARK;
253 crc[1] = hex_asc_upper_hi(valueCRC >> 8);
254 crc[2] = hex_asc_upper_lo(valueCRC >> 8);
255 crc[3] = hex_asc_upper_hi(valueCRC);
256 crc[4] = hex_asc_upper_lo(valueCRC);
257 len = min_t(int, CRC_LEN, str_max_len - str_o_len);
258 memcpy(&str_o[str_o_len], crc, len);
259 str_o_len += len;
260 ext_o_len = ext_crc_len;
261 }
262 if (ext_o_len > 0) {
263 memcpy(&str_o[str_o_len], ext, ext_o_len);
264 str_o_len += ext_o_len;
265 }
266 }
267
268 return str_o_len;
269}
270
271static int udf_name_to_CS0(struct super_block *sb,
272 uint8_t *ocu, int ocu_max_len,
273 const uint8_t *str_i, int str_len)
274{
275 int i, len;
276 unsigned int max_val;
277 int u_len, u_ch;
278 unicode_t uni_char;
279 int (*conv_f)(const unsigned char *, int, wchar_t *);
280
281 if (ocu_max_len <= 0)
282 return 0;
283
284 if (UDF_SB(sb)->s_nls_map)
285 conv_f = UDF_SB(sb)->s_nls_map->char2uni;
286 else
287 conv_f = NULL;
288
289 memset(ocu, 0, ocu_max_len);
290 ocu[0] = 8;
291 max_val = 0xff;
292 u_ch = 1;
293
294try_again:
295 u_len = 1;
296 for (i = 0; i < str_len; i += len) {
297 /* Name didn't fit? */
298 if (u_len + u_ch > ocu_max_len)
299 return 0;
300 if (conv_f) {
301 wchar_t wchar;
302
303 len = conv_f(&str_i[i], str_len - i, &wchar);
304 if (len > 0)
305 uni_char = wchar;
306 } else {
307 len = utf8_to_utf32(&str_i[i], str_len - i,
308 &uni_char);
309 }
310 /* Invalid character, deal with it */
311 if (len <= 0 || uni_char > UNICODE_MAX) {
312 len = 1;
313 uni_char = '?';
314 }
315
316 if (uni_char > max_val) {
317 unicode_t c;
318
319 if (max_val == 0xff) {
320 max_val = 0xffff;
321 ocu[0] = 0x10;
322 u_ch = 2;
323 goto try_again;
324 }
325 /*
326 * Use UTF-16 encoding for chars outside we
327 * cannot encode directly.
328 */
329 if (u_len + 2 * u_ch > ocu_max_len)
330 return 0;
331
332 uni_char -= PLANE_SIZE;
333 c = SURROGATE_PAIR |
334 ((uni_char >> SURROGATE_CHAR_BITS) &
335 SURROGATE_CHAR_MASK);
336 ocu[u_len++] = (uint8_t)(c >> 8);
337 ocu[u_len++] = (uint8_t)(c & 0xff);
338 uni_char = SURROGATE_PAIR | SURROGATE_LOW |
339 (uni_char & SURROGATE_CHAR_MASK);
340 }
341
342 if (max_val == 0xffff)
343 ocu[u_len++] = (uint8_t)(uni_char >> 8);
344 ocu[u_len++] = (uint8_t)(uni_char & 0xff);
345 }
346
347 return u_len;
348}
349
350/*
351 * Convert CS0 dstring to output charset. Warning: This function may truncate
352 * input string if it is too long as it is used for informational strings only
353 * and it is better to truncate the string than to refuse mounting a media.
354 */
355int udf_dstrCS0toChar(struct super_block *sb, uint8_t *utf_o, int o_len,
356 const uint8_t *ocu_i, int i_len)
357{
358 int s_len = 0;
359
360 if (i_len > 0) {
361 s_len = ocu_i[i_len - 1];
362 if (s_len >= i_len) {
363 pr_warn("incorrect dstring lengths (%d/%d),"
364 " truncating\n", s_len, i_len);
365 s_len = i_len - 1;
366 /* 2-byte encoding? Need to round properly... */
367 if (ocu_i[0] == 16)
368 s_len -= (s_len - 1) & 2;
369 }
370 }
371
372 return udf_name_from_CS0(sb, utf_o, o_len, ocu_i, s_len, 0);
373}
374
375int udf_get_filename(struct super_block *sb, const uint8_t *sname, int slen,
376 uint8_t *dname, int dlen)
377{
378 int ret;
379
380 if (!slen)
381 return -EIO;
382
383 if (dlen <= 0)
384 return 0;
385
386 ret = udf_name_from_CS0(sb, dname, dlen, sname, slen, 1);
387 /* Zero length filename isn't valid... */
388 if (ret == 0)
389 ret = -EINVAL;
390 return ret;
391}
392
393int udf_put_filename(struct super_block *sb, const uint8_t *sname, int slen,
394 uint8_t *dname, int dlen)
395{
396 return udf_name_to_CS0(sb, dname, dlen, sname, slen);
397}
398