2 * fs/cifs/cifs_unicode.c
4 * Copyright (c) International Business Machines Corp., 2000,2009
7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; either version 2 of the License, or
10 * (at your option) any later version.
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
15 * the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with this program; if not, write to the Free Software
19 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
22 #include <linux/slab.h>
23 #include "cifs_fs_sb.h"
24 #include "cifs_unicode.h"
25 #include "cifs_uniupr.h"
28 #include "cifs_debug.h"
30 int cifs_remap(struct cifs_sb_info *cifs_sb)
34 if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SFM_CHR)
35 map_type = SFM_MAP_UNI_RSVD;
36 else if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MAP_SPECIAL_CHR)
37 map_type = SFU_MAP_UNI_RSVD;
39 map_type = NO_MAP_UNI_RSVD;
44 /* Convert character using the SFU - "Services for Unix" remapping range */
46 convert_sfu_char(const __u16 src_char, char *target)
49 * BB: Cannot handle remapping UNI_SLASH until all the calls to
50 * build_path_from_dentry are modified, as they use slash as
78 /* Convert character using the SFM - "Services for Mac" remapping range */
80 convert_sfm_char(const __u16 src_char, char *target)
82 if (src_char >= 0xF001 && src_char <= 0xF01F) {
83 *target = src_char - 0xF000;
125 * cifs_mapchar - convert a host-endian char to proper char in codepage
126 * @target - where converted character should be copied
127 * @src_char - 2 byte host-endian source character
128 * @cp - codepage to which character should be converted
129 * @map_type - How should the 7 NTFS/SMB reserved characters be mapped to UCS2?
131 * This function handles the conversion of a single character. It is the
132 * responsibility of the caller to ensure that the target buffer is large
133 * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
136 cifs_mapchar(char *target, const __u16 *from, const struct nls_table *cp,
144 if ((maptype == SFM_MAP_UNI_RSVD) && convert_sfm_char(src_char, target))
146 else if ((maptype == SFU_MAP_UNI_RSVD) &&
147 convert_sfu_char(src_char, target))
150 /* if character not one of seven in special remap set */
151 len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
158 /* convert SURROGATE_PAIR and IVS */
159 if (strcmp(cp->charset, "utf8"))
161 len = utf16s_to_utf8s(from, 3, UTF16_LITTLE_ENDIAN, target, 6);
173 * cifs_from_utf16 - convert utf16le string to local charset
174 * @to - destination buffer
175 * @from - source buffer
176 * @tolen - destination buffer size (in bytes)
177 * @fromlen - source buffer size (in bytes)
178 * @codepage - codepage to which characters should be converted
179 * @mapchar - should characters be remapped according to the mapchars option?
181 * Convert a little-endian utf16le string (as sent by the server) to a string
182 * in the provided codepage. The tolen and fromlen parameters are to ensure
183 * that the code doesn't walk off of the end of the buffer (which is always
184 * a danger if the alignment of the source buffer is off). The destination
185 * string is always properly null terminated and fits in the destination
186 * buffer. Returns the length of the destination string in bytes (including
189 * Note that some windows versions actually send multiword UTF-16 characters
190 * instead of straight UTF16-2. The linux nls routines however aren't able to
191 * deal with those characters properly. In the event that we get some of
192 * those characters, they won't be translated properly.
195 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
196 const struct nls_table *codepage, int map_type)
198 int i, charlen, safelen;
200 int nullsize = nls_nullsize(codepage);
201 int fromwords = fromlen / 2;
202 char tmp[NLS_MAX_CHARSET_SIZE];
203 __u16 ftmp[3]; /* ftmp[3] = 3array x 2bytes = 6bytes UTF-16 */
206 * because the chars can be of varying widths, we need to take care
207 * not to overflow the destination buffer when we get close to the
208 * end of it. Until we get to this offset, we don't need to check
209 * for overflow however.
211 safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
213 for (i = 0; i < fromwords; i++) {
214 ftmp[0] = get_unaligned_le16(&from[i]);
217 if (i + 1 < fromwords)
218 ftmp[1] = get_unaligned_le16(&from[i + 1]);
221 if (i + 2 < fromwords)
222 ftmp[2] = get_unaligned_le16(&from[i + 2]);
227 * check to see if converting this character might make the
228 * conversion bleed into the null terminator
230 if (outlen >= safelen) {
231 charlen = cifs_mapchar(tmp, ftmp, codepage, map_type);
232 if ((outlen + charlen) > (tolen - nullsize))
236 /* put converted char into 'to' buffer */
237 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, map_type);
240 /* charlen (=bytes of UTF-8 for 1 character)
241 * 4bytes UTF-8(surrogate pair) is charlen=4
242 * (4bytes UTF-16 code)
243 * 7-8bytes UTF-8(IVS) is charlen=3+4 or 4+4
244 * (2 UTF-8 pairs divided to 2 UTF-16 pairs) */
247 else if (charlen >= 5)
252 /* properly null-terminate string */
253 for (i = 0; i < nullsize; i++)
260 * NAME: cifs_strtoUTF16()
262 * FUNCTION: Convert character string to unicode string
266 cifs_strtoUTF16(__le16 *to, const char *from, int len,
267 const struct nls_table *codepage)
271 wchar_t wchar_to; /* needed to quiet sparse */
273 /* special case for utf8 to handle no plane0 chars */
274 if (!strcmp(codepage->charset, "utf8")) {
276 * convert utf8 -> utf16, we assume we have enough space
277 * as caller should have assumed conversion does not overflow
278 * in destination len is length in wchar_t units (16bits)
280 i = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
281 (wchar_t *) to, len);
283 /* if success terminate and exit */
287 * if fails fall back to UCS encoding as this
288 * function should not return negative values
289 * currently can fail only if source contains
290 * invalid encoded characters
294 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
295 charlen = codepage->char2uni(from, len, &wchar_to);
297 cifs_dbg(VFS, "strtoUTF16: char2uni of 0x%x returned %d\n",
299 /* A question mark */
303 put_unaligned_le16(wchar_to, &to[i]);
307 put_unaligned_le16(0, &to[i]);
312 * cifs_utf16_bytes - how long will a string be after conversion?
313 * @utf16 - pointer to input string
314 * @maxbytes - don't go past this many bytes of input string
315 * @codepage - destination codepage
317 * Walk a utf16le string and return the number of bytes that the string will
318 * be after being converted to the given charset, not including any null
319 * termination required. Don't walk past maxbytes in the source buffer.
322 cifs_utf16_bytes(const __le16 *from, int maxbytes,
323 const struct nls_table *codepage)
326 int charlen, outlen = 0;
327 int maxwords = maxbytes / 2;
328 char tmp[NLS_MAX_CHARSET_SIZE];
331 for (i = 0; i < maxwords; i++) {
332 ftmp[0] = get_unaligned_le16(&from[i]);
335 if (i + 1 < maxwords)
336 ftmp[1] = get_unaligned_le16(&from[i + 1]);
339 if (i + 2 < maxwords)
340 ftmp[2] = get_unaligned_le16(&from[i + 2]);
344 charlen = cifs_mapchar(tmp, ftmp, codepage, NO_MAP_UNI_RSVD);
352 * cifs_strndup_from_utf16 - copy a string from wire format to the local
354 * @src - source string
355 * @maxlen - don't walk past this many bytes in the source string
356 * @is_unicode - is this a unicode string?
357 * @codepage - destination codepage
359 * Take a string given by the server, convert it to the local codepage and
360 * put it in a new buffer. Returns a pointer to the new string or NULL on
364 cifs_strndup_from_utf16(const char *src, const int maxlen,
365 const bool is_unicode, const struct nls_table *codepage)
371 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
372 len += nls_nullsize(codepage);
373 dst = kmalloc(len, GFP_KERNEL);
376 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
379 len = strnlen(src, maxlen);
381 dst = kmalloc(len, GFP_KERNEL);
384 strlcpy(dst, src, len);
390 static __le16 convert_to_sfu_char(char src_char)
396 dest_char = cpu_to_le16(UNI_COLON);
399 dest_char = cpu_to_le16(UNI_ASTERISK);
402 dest_char = cpu_to_le16(UNI_QUESTION);
405 dest_char = cpu_to_le16(UNI_LESSTHAN);
408 dest_char = cpu_to_le16(UNI_GRTRTHAN);
411 dest_char = cpu_to_le16(UNI_PIPE);
420 static __le16 convert_to_sfm_char(char src_char, bool end_of_string)
424 if (src_char >= 0x01 && src_char <= 0x1F) {
425 dest_char = cpu_to_le16(src_char + 0xF000);
430 dest_char = cpu_to_le16(SFM_COLON);
433 dest_char = cpu_to_le16(SFM_DOUBLEQUOTE);
436 dest_char = cpu_to_le16(SFM_ASTERISK);
439 dest_char = cpu_to_le16(SFM_QUESTION);
442 dest_char = cpu_to_le16(SFM_LESSTHAN);
445 dest_char = cpu_to_le16(SFM_GRTRTHAN);
448 dest_char = cpu_to_le16(SFM_PIPE);
452 dest_char = cpu_to_le16(SFM_PERIOD);
458 dest_char = cpu_to_le16(SFM_SPACE);
470 * Convert 16 bit Unicode pathname to wire format from string in current code
471 * page. Conversion may involve remapping up the six characters that are
472 * only legal in POSIX-like OS (if they are present in the string). Path
473 * names are little endian 16 bit Unicode on the wire
476 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
477 const struct nls_table *cp, int map_chars)
484 wchar_t *wchar_to; /* UTF-16 */
488 if (map_chars == NO_MAP_UNI_RSVD)
489 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
491 wchar_to = kzalloc(6, GFP_KERNEL);
493 for (i = 0; i < srclen; j++) {
494 src_char = source[i];
497 /* check if end of string */
501 /* see if we must remap this char */
502 if (map_chars == SFU_MAP_UNI_RSVD)
503 dst_char = convert_to_sfu_char(src_char);
504 else if (map_chars == SFM_MAP_UNI_RSVD) {
508 end_of_string = true;
510 end_of_string = false;
512 dst_char = convert_to_sfm_char(src_char, end_of_string);
516 * FIXME: We can not handle remapping backslash (UNI_SLASH)
517 * until all the calls to build_path_from_dentry are modified,
518 * as they use backslash as separator.
521 charlen = cp->char2uni(source + i, srclen - i, &tmp);
522 dst_char = cpu_to_le16(tmp);
525 * if no match, use question mark, which at least in
526 * some cases serves as wild card
531 /* convert SURROGATE_PAIR */
532 if (strcmp(cp->charset, "utf8") || !wchar_to)
534 if (*(source + i) & 0x80) {
535 charlen = utf8_to_utf32(source + i, 6, &u);
540 ret = utf8s_to_utf16s(source + i, charlen,
547 dst_char = cpu_to_le16(*wchar_to);
549 /* 1-3bytes UTF-8 to 2bytes UTF-16 */
550 put_unaligned(dst_char, &target[j]);
551 else if (charlen == 4) {
552 /* 4bytes UTF-8(surrogate pair) to 4bytes UTF-16
553 * 7-8bytes UTF-8(IVS) divided to 2 UTF-16
554 * (charlen=3+4 or 4+4) */
555 put_unaligned(dst_char, &target[j]);
556 dst_char = cpu_to_le16(*(wchar_to + 1));
558 put_unaligned(dst_char, &target[j]);
559 } else if (charlen >= 5) {
560 /* 5-6bytes UTF-8 to 6bytes UTF-16 */
561 put_unaligned(dst_char, &target[j]);
562 dst_char = cpu_to_le16(*(wchar_to + 1));
564 put_unaligned(dst_char, &target[j]);
565 dst_char = cpu_to_le16(*(wchar_to + 2));
567 put_unaligned(dst_char, &target[j]);
572 dst_char = cpu_to_le16(0x003f);
578 * character may take more than one byte in the source string,
579 * but will take exactly two bytes in the target string
582 put_unaligned(dst_char, &target[j]);
586 put_unaligned(0, &target[j]); /* Null terminate target unicode string */
592 * cifs_local_to_utf16_bytes - how long will a string be after conversion?
593 * @from - pointer to input string
594 * @maxbytes - don't go past this many bytes of input string
595 * @codepage - source codepage
597 * Walk a string and return the number of bytes that the string will
598 * be after being converted to the given charset, not including any null
599 * termination required. Don't walk past maxbytes in the source buffer.
603 cifs_local_to_utf16_bytes(const char *from, int len,
604 const struct nls_table *codepage)
610 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
611 charlen = codepage->char2uni(from, len, &wchar_to);
612 /* Failed conversion defaults to a question mark */
616 return 2 * i; /* UTF16 characters are two bytes */
620 * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
621 * @src - source string
622 * @maxlen - don't walk past this many bytes in the source string
623 * @utf16_len - the length of the allocated string in bytes (including null)
624 * @cp - source codepage
625 * @remap - map special chars
627 * Take a string convert it from the local codepage to UTF16 and
628 * put it in a new buffer. Returns a pointer to the new string or NULL on
632 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
633 const struct nls_table *cp, int remap)
638 len = cifs_local_to_utf16_bytes(src, maxlen, cp);
640 dst = kmalloc(len, GFP_KERNEL);
645 cifsConvertToUTF16(dst, src, strlen(src), cp, remap);