5 * Routines for converting between UTF-8 and OSTA Compressed Unicode.
6 * Also handles filename mangling
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
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.
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>
31 static int udf_translate_to_linux(uint8_t *, int, uint8_t *, int, uint8_t *,
34 static int udf_char_to_ustr(struct ustr *dest, const uint8_t *src, int strlen)
36 if ((!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN - 2))
39 memset(dest, 0, sizeof(struct ustr));
40 memcpy(dest->u_name, src, strlen);
50 int udf_build_ustr(struct ustr *dest, dstring *ptr, int size)
54 if (!dest || !ptr || !size)
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);
69 * udf_build_ustr_exact
71 static void udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
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);
83 * Convert OSTA Compressed Unicode to the UTF-8 equivalent.
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 *"
92 * <return> >= 0 on success.
95 * November 12, 1997 - Andrew E. Mileski
96 * Written, tested, and released.
98 int udf_CS0toUTF8(struct ustr *utf_o, const struct ustr *ocu_i)
101 uint8_t cmp_id, ocu_len;
104 ocu_len = ocu_i->u_len;
106 memset(utf_o, 0, sizeof(struct ustr));
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);
120 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
122 /* Expand OSTA compressed Unicode to Unicode */
123 uint32_t c = ocu[i++];
125 c = (c << 8) | ocu[i++];
127 /* Compress Unicode to UTF-8 */
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))
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));
138 if (utf_o->u_len > (UDF_NAME_LEN - 5))
140 utf_o->u_name[utf_o->u_len++] =
141 (uint8_t)(0xe0 | (c >> 12));
142 utf_o->u_name[utf_o->u_len++] =
145 utf_o->u_name[utf_o->u_len++] =
146 (uint8_t)(0x80 | (c & 0x3f));
159 * Convert UTF-8 to the OSTA Compressed Unicode equivalent.
162 * This routine is only called by udf_lookup().
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.
171 * <return> Zero on success.
174 * November 12, 1997 - Andrew E. Mileski
175 * Written, tested, and released.
177 static int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length)
179 unsigned c, i, max_val, utf_char;
182 memset(ocu, 0, sizeof(dstring) * length);
190 for (i = 0U; i < utf->u_len; i++) {
191 c = (uint8_t)utf->u_name[i];
193 /* Complete a multi-byte UTF-8 character */
195 utf_char = (utf_char << 6) | (c & 0x3fU);
199 /* Check for a multi-byte UTF-8 character */
201 /* Start a multi-byte UTF-8 character */
202 if ((c & 0xe0U) == 0xc0U) {
203 utf_char = c & 0x1fU;
205 } else if ((c & 0xf0U) == 0xe0U) {
206 utf_char = c & 0x0fU;
208 } else if ((c & 0xf8U) == 0xf0U) {
209 utf_char = c & 0x07U;
211 } else if ((c & 0xfcU) == 0xf8U) {
212 utf_char = c & 0x03U;
214 } else if ((c & 0xfeU) == 0xfcU) {
215 utf_char = c & 0x01U;
222 /* Single byte UTF-8 character (most common) */
227 /* Choose no compression if necessary */
228 if (utf_char > max_val) {
229 if (max_val == 0xffU) {
231 ocu[0] = (uint8_t)0x10U;
237 if (max_val == 0xffffU)
238 ocu[++u_len] = (uint8_t)(utf_char >> 8);
239 ocu[++u_len] = (uint8_t)(utf_char & 0xffU);
245 printk(KERN_DEBUG pr_fmt("bad UTF-8 character\n"));
248 ocu[length - 1] = (uint8_t)u_len + 1;
253 static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o,
254 const struct ustr *ocu_i)
257 uint8_t cmp_id, ocu_len;
261 ocu_len = ocu_i->u_len;
263 memset(utf_o, 0, sizeof(struct ustr));
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);
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++];
281 c = (c << 8) | ocu[i++];
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? */
289 utf_o->u_name[utf_o->u_len++] = '?';
296 static int udf_NLStoCS0(struct nls_table *nls, dstring *ocu, struct ustr *uni,
304 memset(ocu, 0, sizeof(dstring) * length);
310 for (i = 0U; i < uni->u_len; i++) {
311 len = nls->char2uni(&uni->u_name[i], uni->u_len - i, &uni_char);
314 /* Invalid character, deal with it */
320 if (uni_char > max_val) {
322 ocu[0] = (uint8_t)0x10U;
326 if (max_val == 0xffffU)
327 ocu[++u_len] = (uint8_t)(uni_char >> 8);
328 ocu[++u_len] = (uint8_t)(uni_char & 0xffU);
332 ocu[length - 1] = (uint8_t)u_len + 1;
336 int udf_get_filename(struct super_block *sb, uint8_t *sname, int slen,
337 uint8_t *dname, int dlen)
339 struct ustr *filename, *unifilename;
345 filename = kmalloc(sizeof(struct ustr), GFP_NOFS);
349 unifilename = kmalloc(sizeof(struct ustr), GFP_NOFS);
355 udf_build_ustr_exact(unifilename, sname, slen);
356 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
357 ret = udf_CS0toUTF8(filename, unifilename);
359 udf_debug("Failed in udf_get_filename: sname = %s\n",
363 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
364 ret = udf_CS0toNLS(UDF_SB(sb)->s_nls_map, filename,
367 udf_debug("Failed in udf_get_filename: sname = %s\n",
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... */
387 int udf_put_filename(struct super_block *sb, const uint8_t *sname,
388 uint8_t *dname, int flen)
390 struct ustr unifilename;
393 if (!udf_char_to_ustr(&unifilename, sname, flen))
396 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
397 namelen = udf_UTF8toCS0(dname, &unifilename, UDF_NAME_LEN);
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);
411 #define ILLEGAL_CHAR_MARK '_'
415 /* Number of chars we need to store generated CRC to make filename unique */
418 static int udf_translate_to_linux(uint8_t *newName, int newLen,
419 uint8_t *udfName, int udfLen,
420 uint8_t *fidName, int fidNameLen)
422 int index, newIndex = 0, needsCRC = 0;
423 int extIndex = 0, newExtIndex = 0, hasExt = 0;
424 unsigned short valueCRC;
427 if (udfName[0] == '.' &&
428 (udfLen == 1 || (udfLen == 2 && udfName[1] == '.'))) {
431 memcpy(newName, udfName, udfLen);
433 for (index = 0; index < udfLen; index++) {
434 curr = udfName[index];
435 if (curr == '/' || curr == 0) {
437 curr = ILLEGAL_CHAR_MARK;
438 while (index + 1 < udfLen &&
439 (udfName[index + 1] == '/' ||
440 udfName[index + 1] == 0))
443 if (curr == EXT_MARK &&
444 (udfLen - index - 1) <= EXT_SIZE) {
445 if (udfLen == index + 1)
450 newExtIndex = newIndex;
453 if (newIndex < newLen)
454 newName[newIndex++] = curr;
460 uint8_t ext[EXT_SIZE];
461 int localExtIndex = 0;
466 index < EXT_SIZE && extIndex + index + 1 < udfLen;
468 curr = udfName[extIndex + index + 1];
470 if (curr == '/' || curr == 0) {
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)))
479 ext[localExtIndex++] = curr;
481 maxFilenameLen = newLen - CRC_LEN - localExtIndex;
482 if (newIndex > maxFilenameLen)
483 newIndex = maxFilenameLen;
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);
496 newName[newIndex++] = EXT_MARK;
497 for (index = 0; index < localExtIndex; index++)
498 newName[newIndex++] = ext[index];