2 * fs/cifs/cifs_unicode.c
4 * Copyright (c) International Business Machines Corp., 2000,2009
5 * Modified by Steve French (sfrench@us.ibm.com)
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;
122 * cifs_mapchar - convert a host-endian char to proper char in codepage
123 * @target - where converted character should be copied
124 * @src_char - 2 byte host-endian source character
125 * @cp - codepage to which character should be converted
126 * @map_type - How should the 7 NTFS/SMB reserved characters be mapped to UCS2?
128 * This function handles the conversion of a single character. It is the
129 * responsibility of the caller to ensure that the target buffer is large
130 * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
133 cifs_mapchar(char *target, const __u16 *from, const struct nls_table *cp,
141 if ((maptype == SFM_MAP_UNI_RSVD) && convert_sfm_char(src_char, target))
143 else if ((maptype == SFU_MAP_UNI_RSVD) &&
144 convert_sfu_char(src_char, target))
147 /* if character not one of seven in special remap set */
148 len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
155 /* convert SURROGATE_PAIR and IVS */
156 if (strcmp(cp->charset, "utf8"))
158 len = utf16s_to_utf8s(from, 3, UTF16_LITTLE_ENDIAN, target, 6);
170 * cifs_from_utf16 - convert utf16le string to local charset
171 * @to - destination buffer
172 * @from - source buffer
173 * @tolen - destination buffer size (in bytes)
174 * @fromlen - source buffer size (in bytes)
175 * @codepage - codepage to which characters should be converted
176 * @mapchar - should characters be remapped according to the mapchars option?
178 * Convert a little-endian utf16le string (as sent by the server) to a string
179 * in the provided codepage. The tolen and fromlen parameters are to ensure
180 * that the code doesn't walk off of the end of the buffer (which is always
181 * a danger if the alignment of the source buffer is off). The destination
182 * string is always properly null terminated and fits in the destination
183 * buffer. Returns the length of the destination string in bytes (including
186 * Note that some windows versions actually send multiword UTF-16 characters
187 * instead of straight UTF16-2. The linux nls routines however aren't able to
188 * deal with those characters properly. In the event that we get some of
189 * those characters, they won't be translated properly.
192 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
193 const struct nls_table *codepage, int map_type)
195 int i, charlen, safelen;
197 int nullsize = nls_nullsize(codepage);
198 int fromwords = fromlen / 2;
199 char tmp[NLS_MAX_CHARSET_SIZE];
200 __u16 ftmp[3]; /* ftmp[3] = 3array x 2bytes = 6bytes UTF-16 */
203 * because the chars can be of varying widths, we need to take care
204 * not to overflow the destination buffer when we get close to the
205 * end of it. Until we get to this offset, we don't need to check
206 * for overflow however.
208 safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
210 for (i = 0; i < fromwords; i++) {
211 ftmp[0] = get_unaligned_le16(&from[i]);
214 if (i + 1 < fromwords)
215 ftmp[1] = get_unaligned_le16(&from[i + 1]);
218 if (i + 2 < fromwords)
219 ftmp[2] = get_unaligned_le16(&from[i + 2]);
224 * check to see if converting this character might make the
225 * conversion bleed into the null terminator
227 if (outlen >= safelen) {
228 charlen = cifs_mapchar(tmp, ftmp, codepage, map_type);
229 if ((outlen + charlen) > (tolen - nullsize))
233 /* put converted char into 'to' buffer */
234 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, map_type);
237 /* charlen (=bytes of UTF-8 for 1 character)
238 * 4bytes UTF-8(surrogate pair) is charlen=4
239 * (4bytes UTF-16 code)
240 * 7-8bytes UTF-8(IVS) is charlen=3+4 or 4+4
241 * (2 UTF-8 pairs divided to 2 UTF-16 pairs) */
244 else if (charlen >= 5)
249 /* properly null-terminate string */
250 for (i = 0; i < nullsize; i++)
257 * NAME: cifs_strtoUTF16()
259 * FUNCTION: Convert character string to unicode string
263 cifs_strtoUTF16(__le16 *to, const char *from, int len,
264 const struct nls_table *codepage)
268 wchar_t wchar_to; /* needed to quiet sparse */
270 /* special case for utf8 to handle no plane0 chars */
271 if (!strcmp(codepage->charset, "utf8")) {
273 * convert utf8 -> utf16, we assume we have enough space
274 * as caller should have assumed conversion does not overflow
275 * in destination len is length in wchar_t units (16bits)
277 i = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
278 (wchar_t *) to, len);
280 /* if success terminate and exit */
284 * if fails fall back to UCS encoding as this
285 * function should not return negative values
286 * currently can fail only if source contains
287 * invalid encoded characters
291 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
292 charlen = codepage->char2uni(from, len, &wchar_to);
294 cifs_dbg(VFS, "strtoUTF16: char2uni of 0x%x returned %d\n",
296 /* A question mark */
300 put_unaligned_le16(wchar_to, &to[i]);
304 put_unaligned_le16(0, &to[i]);
309 * cifs_utf16_bytes - how long will a string be after conversion?
310 * @utf16 - pointer to input string
311 * @maxbytes - don't go past this many bytes of input string
312 * @codepage - destination codepage
314 * Walk a utf16le string and return the number of bytes that the string will
315 * be after being converted to the given charset, not including any null
316 * termination required. Don't walk past maxbytes in the source buffer.
319 cifs_utf16_bytes(const __le16 *from, int maxbytes,
320 const struct nls_table *codepage)
323 int charlen, outlen = 0;
324 int maxwords = maxbytes / 2;
325 char tmp[NLS_MAX_CHARSET_SIZE];
328 for (i = 0; i < maxwords; i++) {
329 ftmp[0] = get_unaligned_le16(&from[i]);
332 if (i + 1 < maxwords)
333 ftmp[1] = get_unaligned_le16(&from[i + 1]);
336 if (i + 2 < maxwords)
337 ftmp[2] = get_unaligned_le16(&from[i + 2]);
341 charlen = cifs_mapchar(tmp, ftmp, codepage, NO_MAP_UNI_RSVD);
349 * cifs_strndup_from_utf16 - copy a string from wire format to the local
351 * @src - source string
352 * @maxlen - don't walk past this many bytes in the source string
353 * @is_unicode - is this a unicode string?
354 * @codepage - destination codepage
356 * Take a string given by the server, convert it to the local codepage and
357 * put it in a new buffer. Returns a pointer to the new string or NULL on
361 cifs_strndup_from_utf16(const char *src, const int maxlen,
362 const bool is_unicode, const struct nls_table *codepage)
368 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
369 len += nls_nullsize(codepage);
370 dst = kmalloc(len, GFP_KERNEL);
373 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
376 dst = kstrndup(src, maxlen, GFP_KERNEL);
382 static __le16 convert_to_sfu_char(char src_char)
388 dest_char = cpu_to_le16(UNI_COLON);
391 dest_char = cpu_to_le16(UNI_ASTERISK);
394 dest_char = cpu_to_le16(UNI_QUESTION);
397 dest_char = cpu_to_le16(UNI_LESSTHAN);
400 dest_char = cpu_to_le16(UNI_GRTRTHAN);
403 dest_char = cpu_to_le16(UNI_PIPE);
412 static __le16 convert_to_sfm_char(char src_char, bool end_of_string)
416 if (src_char >= 0x01 && src_char <= 0x1F) {
417 dest_char = cpu_to_le16(src_char + 0xF000);
422 dest_char = cpu_to_le16(SFM_COLON);
425 dest_char = cpu_to_le16(SFM_DOUBLEQUOTE);
428 dest_char = cpu_to_le16(SFM_ASTERISK);
431 dest_char = cpu_to_le16(SFM_QUESTION);
434 dest_char = cpu_to_le16(SFM_LESSTHAN);
437 dest_char = cpu_to_le16(SFM_GRTRTHAN);
440 dest_char = cpu_to_le16(SFM_PIPE);
444 dest_char = cpu_to_le16(SFM_PERIOD);
450 dest_char = cpu_to_le16(SFM_SPACE);
462 * Convert 16 bit Unicode pathname to wire format from string in current code
463 * page. Conversion may involve remapping up the six characters that are
464 * only legal in POSIX-like OS (if they are present in the string). Path
465 * names are little endian 16 bit Unicode on the wire
468 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
469 const struct nls_table *cp, int map_chars)
476 wchar_t *wchar_to; /* UTF-16 */
480 if (map_chars == NO_MAP_UNI_RSVD)
481 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
483 wchar_to = kzalloc(6, GFP_KERNEL);
485 for (i = 0; i < srclen; j++) {
486 src_char = source[i];
489 /* check if end of string */
493 /* see if we must remap this char */
494 if (map_chars == SFU_MAP_UNI_RSVD)
495 dst_char = convert_to_sfu_char(src_char);
496 else if (map_chars == SFM_MAP_UNI_RSVD) {
500 * Remap spaces and periods found at the end of every
501 * component of the path. The special cases of '.' and
502 * '..' do not need to be dealt with explicitly because
503 * they are addressed in namei.c:link_path_walk().
505 if ((i == srclen - 1) || (source[i+1] == '\\'))
506 end_of_string = true;
508 end_of_string = false;
510 dst_char = convert_to_sfm_char(src_char, end_of_string);
514 * FIXME: We can not handle remapping backslash (UNI_SLASH)
515 * until all the calls to build_path_from_dentry are modified,
516 * as they use backslash as separator.
519 charlen = cp->char2uni(source + i, srclen - i, &tmp);
520 dst_char = cpu_to_le16(tmp);
523 * if no match, use question mark, which at least in
524 * some cases serves as wild card
529 /* convert SURROGATE_PAIR */
530 if (strcmp(cp->charset, "utf8") || !wchar_to)
532 if (*(source + i) & 0x80) {
533 charlen = utf8_to_utf32(source + i, 6, &u);
538 ret = utf8s_to_utf16s(source + i, charlen,
545 dst_char = cpu_to_le16(*wchar_to);
547 /* 1-3bytes UTF-8 to 2bytes UTF-16 */
548 put_unaligned(dst_char, &target[j]);
549 else if (charlen == 4) {
550 /* 4bytes UTF-8(surrogate pair) to 4bytes UTF-16
551 * 7-8bytes UTF-8(IVS) divided to 2 UTF-16
552 * (charlen=3+4 or 4+4) */
553 put_unaligned(dst_char, &target[j]);
554 dst_char = cpu_to_le16(*(wchar_to + 1));
556 put_unaligned(dst_char, &target[j]);
557 } else if (charlen >= 5) {
558 /* 5-6bytes UTF-8 to 6bytes UTF-16 */
559 put_unaligned(dst_char, &target[j]);
560 dst_char = cpu_to_le16(*(wchar_to + 1));
562 put_unaligned(dst_char, &target[j]);
563 dst_char = cpu_to_le16(*(wchar_to + 2));
565 put_unaligned(dst_char, &target[j]);
570 dst_char = cpu_to_le16(0x003f);
576 * character may take more than one byte in the source string,
577 * but will take exactly two bytes in the target string
580 put_unaligned(dst_char, &target[j]);
584 put_unaligned(0, &target[j]); /* Null terminate target unicode string */
590 * cifs_local_to_utf16_bytes - how long will a string be after conversion?
591 * @from - pointer to input string
592 * @maxbytes - don't go past this many bytes of input string
593 * @codepage - source codepage
595 * Walk a string and return the number of bytes that the string will
596 * be after being converted to the given charset, not including any null
597 * termination required. Don't walk past maxbytes in the source buffer.
601 cifs_local_to_utf16_bytes(const char *from, int len,
602 const struct nls_table *codepage)
608 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
609 charlen = codepage->char2uni(from, len, &wchar_to);
610 /* Failed conversion defaults to a question mark */
614 return 2 * i; /* UTF16 characters are two bytes */
618 * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
619 * @src - source string
620 * @maxlen - don't walk past this many bytes in the source string
621 * @utf16_len - the length of the allocated string in bytes (including null)
622 * @cp - source codepage
623 * @remap - map special chars
625 * Take a string convert it from the local codepage to UTF16 and
626 * put it in a new buffer. Returns a pointer to the new string or NULL on
630 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
631 const struct nls_table *cp, int remap)
636 len = cifs_local_to_utf16_bytes(src, maxlen, cp);
638 dst = kmalloc(len, GFP_KERNEL);
643 cifsConvertToUTF16(dst, src, strlen(src), cp, remap);