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 int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize)
73 if ((!dest) || (!ptr) || (!exactsize))
76 memset(dest, 0, sizeof(struct ustr));
77 dest->u_cmpID = ptr[0];
78 dest->u_len = exactsize - 1;
79 memcpy(dest->u_name, ptr + 1, exactsize - 1);
88 * Convert OSTA Compressed Unicode to the UTF-8 equivalent.
91 * utf Pointer to UTF-8 output buffer.
92 * ocu Pointer to OSTA Compressed Unicode input buffer
93 * of size UDF_NAME_LEN bytes.
94 * both of type "struct ustr *"
97 * <return> Zero on success.
100 * November 12, 1997 - Andrew E. Mileski
101 * Written, tested, and released.
103 int udf_CS0toUTF8(struct ustr *utf_o, const struct ustr *ocu_i)
106 uint8_t cmp_id, ocu_len;
109 ocu_len = ocu_i->u_len;
111 memset(utf_o, 0, sizeof(struct ustr));
115 cmp_id = ocu_i->u_cmpID;
116 if (cmp_id != 8 && cmp_id != 16) {
117 memset(utf_o, 0, sizeof(struct ustr));
118 pr_err("unknown compression code (%d) stri=%s\n",
119 cmp_id, ocu_i->u_name);
125 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
127 /* Expand OSTA compressed Unicode to Unicode */
128 uint32_t c = ocu[i++];
130 c = (c << 8) | ocu[i++];
132 /* Compress Unicode to UTF-8 */
134 utf_o->u_name[utf_o->u_len++] = (uint8_t)c;
135 else if (c < 0x800U) {
136 if (utf_o->u_len > (UDF_NAME_LEN - 4))
138 utf_o->u_name[utf_o->u_len++] =
139 (uint8_t)(0xc0 | (c >> 6));
140 utf_o->u_name[utf_o->u_len++] =
141 (uint8_t)(0x80 | (c & 0x3f));
143 if (utf_o->u_len > (UDF_NAME_LEN - 5))
145 utf_o->u_name[utf_o->u_len++] =
146 (uint8_t)(0xe0 | (c >> 12));
147 utf_o->u_name[utf_o->u_len++] =
150 utf_o->u_name[utf_o->u_len++] =
151 (uint8_t)(0x80 | (c & 0x3f));
164 * Convert UTF-8 to the OSTA Compressed Unicode equivalent.
167 * This routine is only called by udf_lookup().
170 * ocu Pointer to OSTA Compressed Unicode output
171 * buffer of size UDF_NAME_LEN bytes.
172 * utf Pointer to UTF-8 input buffer.
173 * utf_len Length of UTF-8 input buffer in bytes.
176 * <return> Zero on success.
179 * November 12, 1997 - Andrew E. Mileski
180 * Written, tested, and released.
182 static int udf_UTF8toCS0(dstring *ocu, struct ustr *utf, int length)
184 unsigned c, i, max_val, utf_char;
185 int utf_cnt, u_len, u_ch;
187 memset(ocu, 0, sizeof(dstring) * length);
196 for (i = 0U; i < utf->u_len; i++) {
197 /* Name didn't fit? */
198 if (u_len + 1 + u_ch >= length)
201 c = (uint8_t)utf->u_name[i];
203 /* Complete a multi-byte UTF-8 character */
205 utf_char = (utf_char << 6) | (c & 0x3fU);
209 /* Check for a multi-byte UTF-8 character */
211 /* Start a multi-byte UTF-8 character */
212 if ((c & 0xe0U) == 0xc0U) {
213 utf_char = c & 0x1fU;
215 } else if ((c & 0xf0U) == 0xe0U) {
216 utf_char = c & 0x0fU;
218 } else if ((c & 0xf8U) == 0xf0U) {
219 utf_char = c & 0x07U;
221 } else if ((c & 0xfcU) == 0xf8U) {
222 utf_char = c & 0x03U;
224 } else if ((c & 0xfeU) == 0xfcU) {
225 utf_char = c & 0x01U;
232 /* Single byte UTF-8 character (most common) */
237 /* Choose no compression if necessary */
238 if (utf_char > max_val) {
239 if (max_val == 0xffU) {
241 ocu[0] = (uint8_t)0x10U;
248 if (max_val == 0xffffU)
249 ocu[++u_len] = (uint8_t)(utf_char >> 8);
250 ocu[++u_len] = (uint8_t)(utf_char & 0xffU);
256 printk(KERN_DEBUG pr_fmt("bad UTF-8 character\n"));
259 ocu[length - 1] = (uint8_t)u_len + 1;
264 static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o,
265 const struct ustr *ocu_i)
268 uint8_t cmp_id, ocu_len;
272 ocu_len = ocu_i->u_len;
274 memset(utf_o, 0, sizeof(struct ustr));
278 cmp_id = ocu_i->u_cmpID;
279 if (cmp_id != 8 && cmp_id != 16) {
280 memset(utf_o, 0, sizeof(struct ustr));
281 pr_err("unknown compression code (%d) stri=%s\n",
282 cmp_id, ocu_i->u_name);
288 for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) {
289 /* Expand OSTA compressed Unicode to Unicode */
290 uint32_t c = ocu[i++];
292 c = (c << 8) | ocu[i++];
294 len = nls->uni2char(c, &utf_o->u_name[utf_o->u_len],
295 UDF_NAME_LEN - 2 - utf_o->u_len);
296 /* Valid character? */
300 utf_o->u_name[utf_o->u_len++] = '?';
307 static int udf_NLStoCS0(struct nls_table *nls, dstring *ocu, struct ustr *uni,
315 memset(ocu, 0, sizeof(dstring) * length);
322 for (i = 0U; i < uni->u_len; i++) {
323 /* Name didn't fit? */
324 if (u_len + 1 + u_ch >= length)
326 len = nls->char2uni(&uni->u_name[i], uni->u_len - i, &uni_char);
329 /* Invalid character, deal with it */
335 if (uni_char > max_val) {
337 ocu[0] = (uint8_t)0x10U;
342 if (max_val == 0xffffU)
343 ocu[++u_len] = (uint8_t)(uni_char >> 8);
344 ocu[++u_len] = (uint8_t)(uni_char & 0xffU);
348 ocu[length - 1] = (uint8_t)u_len + 1;
352 int udf_get_filename(struct super_block *sb, uint8_t *sname, int slen,
353 uint8_t *dname, int dlen)
355 struct ustr *filename, *unifilename;
358 filename = kmalloc(sizeof(struct ustr), GFP_NOFS);
362 unifilename = kmalloc(sizeof(struct ustr), GFP_NOFS);
366 if (udf_build_ustr_exact(unifilename, sname, slen))
369 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
370 if (!udf_CS0toUTF8(filename, unifilename)) {
371 udf_debug("Failed in udf_get_filename: sname = %s\n",
375 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
376 if (!udf_CS0toNLS(UDF_SB(sb)->s_nls_map, filename,
378 udf_debug("Failed in udf_get_filename: sname = %s\n",
385 len = udf_translate_to_linux(dname, dlen,
386 filename->u_name, filename->u_len,
387 unifilename->u_name, unifilename->u_len);
395 int udf_put_filename(struct super_block *sb, const uint8_t *sname,
396 uint8_t *dname, int flen)
398 struct ustr unifilename;
401 if (!udf_char_to_ustr(&unifilename, sname, flen))
404 if (UDF_QUERY_FLAG(sb, UDF_FLAG_UTF8)) {
405 namelen = udf_UTF8toCS0(dname, &unifilename, UDF_NAME_LEN);
408 } else if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) {
409 namelen = udf_NLStoCS0(UDF_SB(sb)->s_nls_map, dname,
410 &unifilename, UDF_NAME_LEN);
419 #define ILLEGAL_CHAR_MARK '_'
423 /* Number of chars we need to store generated CRC to make filename unique */
426 static int udf_translate_to_linux(uint8_t *newName, int newLen,
427 uint8_t *udfName, int udfLen,
428 uint8_t *fidName, int fidNameLen)
430 int index, newIndex = 0, needsCRC = 0;
431 int extIndex = 0, newExtIndex = 0, hasExt = 0;
432 unsigned short valueCRC;
434 const uint8_t hexChar[] = "0123456789ABCDEF";
436 if (udfName[0] == '.' &&
437 (udfLen == 1 || (udfLen == 2 && udfName[1] == '.'))) {
440 memcpy(newName, udfName, udfLen);
442 for (index = 0; index < udfLen; index++) {
443 curr = udfName[index];
444 if (curr == '/' || curr == 0) {
446 curr = ILLEGAL_CHAR_MARK;
447 while (index + 1 < udfLen &&
448 (udfName[index + 1] == '/' ||
449 udfName[index + 1] == 0))
452 if (curr == EXT_MARK &&
453 (udfLen - index - 1) <= EXT_SIZE) {
454 if (udfLen == index + 1)
459 newExtIndex = newIndex;
462 if (newIndex < newLen)
463 newName[newIndex++] = curr;
469 uint8_t ext[EXT_SIZE];
470 int localExtIndex = 0;
475 index < EXT_SIZE && extIndex + index + 1 < udfLen;
477 curr = udfName[extIndex + index + 1];
479 if (curr == '/' || curr == 0) {
481 curr = ILLEGAL_CHAR_MARK;
482 while (extIndex + index + 2 < udfLen &&
483 (index + 1 < EXT_SIZE &&
484 (udfName[extIndex + index + 2] == '/' ||
485 udfName[extIndex + index + 2] == 0)))
488 ext[localExtIndex++] = curr;
490 maxFilenameLen = newLen - CRC_LEN - localExtIndex;
491 if (newIndex > maxFilenameLen)
492 newIndex = maxFilenameLen;
494 newIndex = newExtIndex;
495 } else if (newIndex > newLen - CRC_LEN)
496 newIndex = newLen - CRC_LEN;
497 newName[newIndex++] = CRC_MARK;
498 valueCRC = crc_itu_t(0, fidName, fidNameLen);
499 newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12];
500 newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8];
501 newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4];
502 newName[newIndex++] = hexChar[(valueCRC & 0x000f)];
505 newName[newIndex++] = EXT_MARK;
506 for (index = 0; index < localExtIndex; index++)
507 newName[newIndex++] = ext[index];