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/udf_fs.h>
30 static int udf_translate_to_linux(uint8_t *, uint8_t *, int, uint8_t *, int);
32 static int udf_char_to_ustr(struct ustr *dest, const uint8_t * src, int strlen)
34 if ((!dest) || (!src) || (!strlen) || (strlen > UDF_NAME_LEN - 2))
36 memset(dest, 0, sizeof(struct ustr));
37 memcpy(dest->u_name, src, strlen);
46 int udf_build_ustr(struct ustr *dest, dstring * ptr, int size)
50 if ((!dest) || (!ptr) || (!size))
53 memset(dest, 0, sizeof(struct ustr));
54 usesize = (size > UDF_NAME_LEN) ? UDF_NAME_LEN : size;
55 dest->u_cmpID = ptr[0];
56 dest->u_len = ptr[size - 1];
57 memcpy(dest->u_name, ptr + 1, usesize - 1);
62 * udf_build_ustr_exact
64 static int udf_build_ustr_exact(struct ustr *dest, dstring * ptr, int exactsize)
66 if ((!dest) || (!ptr) || (!exactsize))
69 memset(dest, 0, sizeof(struct ustr));
70 dest->u_cmpID = ptr[0];
71 dest->u_len = exactsize - 1;
72 memcpy(dest->u_name, ptr + 1, exactsize - 1);
80 * Convert OSTA Compressed Unicode to the UTF-8 equivalent.
83 * This routine is only called by udf_filldir().
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> Zero on success.
95 * November 12, 1997 - Andrew E. Mileski
96 * Written, tested, and released.
98 int udf_CS0toUTF8(struct ustr *utf_o, struct ustr *ocu_i)
102 uint8_t cmp_id, ocu_len;
107 ocu_len = ocu_i->u_len;
108 cmp_id = ocu_i->u_cmpID;
112 memset(utf_o, 0, sizeof(struct ustr));
118 if ((cmp_id != 8) && (cmp_id != 16)) {
119 printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n",
120 cmp_id, ocu_i->u_name);
124 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
126 /* Expand OSTA compressed Unicode to Unicode */
129 c = (c << 8) | ocu[i++];
131 /* Compress Unicode to UTF-8 */
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));
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 | ((c >> 6) & 0x3f));
144 utf_o->u_name[utf_o->u_len++] =
145 (uint8_t) (0x80 | (c & 0x3f));
158 * Convert UTF-8 to the OSTA Compressed Unicode equivalent.
161 * This routine is only called by udf_lookup().
164 * ocu Pointer to OSTA Compressed Unicode output
165 * buffer of size UDF_NAME_LEN bytes.
166 * utf Pointer to UTF-8 input buffer.
167 * utf_len Length of UTF-8 input buffer in bytes.
170 * <return> Zero on success.
173 * November 12, 1997 - Andrew E. Mileski
174 * Written, tested, and released.
176 static int udf_UTF8toCS0(dstring * ocu, struct ustr *utf, int length)
178 unsigned c, i, max_val, utf_char;
181 memset(ocu, 0, sizeof(dstring) * length);
189 for (i = 0U; i < utf->u_len; i++) {
190 c = (uint8_t) utf->u_name[i];
192 /* Complete a multi-byte UTF-8 character */
194 utf_char = (utf_char << 6) | (c & 0x3fU);
198 /* Check for a multi-byte UTF-8 character */
200 /* Start a multi-byte UTF-8 character */
201 if ((c & 0xe0U) == 0xc0U) {
202 utf_char = c & 0x1fU;
204 } else if ((c & 0xf0U) == 0xe0U) {
205 utf_char = c & 0x0fU;
207 } else if ((c & 0xf8U) == 0xf0U) {
208 utf_char = c & 0x07U;
210 } else if ((c & 0xfcU) == 0xf8U) {
211 utf_char = c & 0x03U;
213 } else if ((c & 0xfeU) == 0xfcU) {
214 utf_char = c & 0x01U;
220 /* Single byte UTF-8 character (most common) */
224 /* Choose no compression if necessary */
225 if (utf_char > max_val) {
226 if (0xffU == max_val) {
228 ocu[0] = (uint8_t) 0x10U;
234 if (max_val == 0xffffU) {
235 ocu[++u_len] = (uint8_t) (utf_char >> 8);
237 ocu[++u_len] = (uint8_t) (utf_char & 0xffU);
243 printk(KERN_DEBUG "udf: bad UTF-8 character\n");
246 ocu[length - 1] = (uint8_t) u_len + 1;
250 static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o,
255 uint8_t cmp_id, ocu_len;
260 ocu_len = ocu_i->u_len;
261 cmp_id = ocu_i->u_cmpID;
265 memset(utf_o, 0, sizeof(struct ustr));
271 if ((cmp_id != 8) && (cmp_id != 16)) {
272 printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n",
273 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 */
281 c = (c << 8) | ocu[i++];
283 utf_o->u_len += nls->uni2char(c, &utf_o->u_name[utf_o->u_len],
284 UDF_NAME_LEN - utf_o->u_len);
291 static int udf_NLStoCS0(struct nls_table *nls, dstring * ocu, struct ustr *uni,
294 unsigned len, i, max_val;
298 memset(ocu, 0, sizeof(dstring) * length);
304 for (i = 0U; i < uni->u_len; i++) {
305 len = nls->char2uni(&uni->u_name[i], uni->u_len - i, &uni_char);
309 if (uni_char > max_val) {
311 ocu[0] = (uint8_t) 0x10U;
315 if (max_val == 0xffffU)
316 ocu[++u_len] = (uint8_t) (uni_char >> 8);
317 ocu[++u_len] = (uint8_t) (uni_char & 0xffU);
321 ocu[length - 1] = (uint8_t) u_len + 1;
325 int udf_get_filename(struct super_block *sb, uint8_t * sname, uint8_t * dname,
328 struct ustr filename, unifilename;
331 if (udf_build_ustr_exact(&unifilename, sname, flen)) {
335 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
336 if (!udf_CS0toUTF8(&filename, &unifilename)) {
337 udf_debug("Failed in udf_get_filename: sname = %s\n",
341 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
343 (UDF_SB(sb)->s_nls_map, &filename, &unifilename)) {
344 udf_debug("Failed in udf_get_filename: sname = %s\n",
352 udf_translate_to_linux(dname, filename.u_name, filename.u_len,
353 unifilename.u_name, unifilename.u_len))) {
359 int udf_put_filename(struct super_block *sb, const uint8_t * sname,
360 uint8_t * dname, int flen)
362 struct ustr unifilename;
365 if (!(udf_char_to_ustr(&unifilename, sname, flen))) {
369 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
372 udf_UTF8toCS0(dname, &unifilename, UDF_NAME_LEN))) {
375 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
378 udf_NLStoCS0(UDF_SB(sb)->s_nls_map, dname, &unifilename,
388 #define ILLEGAL_CHAR_MARK '_'
393 static int udf_translate_to_linux(uint8_t * newName, uint8_t * udfName,
394 int udfLen, uint8_t * fidName, int fidNameLen)
396 int index, newIndex = 0, needsCRC = 0;
397 int extIndex = 0, newExtIndex = 0, hasExt = 0;
398 unsigned short valueCRC;
400 const uint8_t hexChar[] = "0123456789ABCDEF";
402 if (udfName[0] == '.' && (udfLen == 1 ||
403 (udfLen == 2 && udfName[1] == '.'))) {
406 memcpy(newName, udfName, udfLen);
408 for (index = 0; index < udfLen; index++) {
409 curr = udfName[index];
410 if (curr == '/' || curr == 0) {
412 curr = ILLEGAL_CHAR_MARK;
413 while (index + 1 < udfLen
414 && (udfName[index + 1] == '/'
415 || udfName[index + 1] == 0))
419 && (udfLen - index - 1) <= EXT_SIZE) {
420 if (udfLen == index + 1)
425 newExtIndex = newIndex;
429 newName[newIndex++] = curr;
435 uint8_t ext[EXT_SIZE];
436 int localExtIndex = 0;
441 index < EXT_SIZE && extIndex + index + 1 < udfLen;
443 curr = udfName[extIndex + index + 1];
445 if (curr == '/' || curr == 0) {
447 curr = ILLEGAL_CHAR_MARK;
448 while (extIndex + index + 2 < udfLen
449 && (index + 1 < EXT_SIZE
452 [extIndex + index + 2] ==
454 || udfName[extIndex +
459 ext[localExtIndex++] = curr;
461 maxFilenameLen = 250 - localExtIndex;
462 if (newIndex > maxFilenameLen)
463 newIndex = maxFilenameLen;
465 newIndex = newExtIndex;
466 } else if (newIndex > 250)
468 newName[newIndex++] = CRC_MARK;
469 valueCRC = udf_crc(fidName, fidNameLen, 0);
470 newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
471 newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
472 newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
473 newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
476 newName[newIndex++] = EXT_MARK;
477 for (index = 0; index < localExtIndex; index++)
478 newName[newIndex++] = ext[index];