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)
118 * cifs_mapchar - convert a host-endian char to proper char in codepage
119 * @target - where converted character should be copied
120 * @src_char - 2 byte host-endian source character
121 * @cp - codepage to which character should be converted
122 * @map_type - How should the 7 NTFS/SMB reserved characters be mapped to UCS2?
124 * This function handles the conversion of a single character. It is the
125 * responsibility of the caller to ensure that the target buffer is large
126 * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
129 cifs_mapchar(char *target, const __u16 *from, const struct nls_table *cp,
137 if ((maptype == SFM_MAP_UNI_RSVD) && convert_sfm_char(src_char, target))
139 else if ((maptype == SFU_MAP_UNI_RSVD) &&
140 convert_sfu_char(src_char, target))
143 /* if character not one of seven in special remap set */
144 len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
151 /* convert SURROGATE_PAIR and IVS */
152 if (strcmp(cp->charset, "utf8"))
154 len = utf16s_to_utf8s(from, 3, UTF16_LITTLE_ENDIAN, target, 6);
166 * cifs_from_utf16 - convert utf16le string to local charset
167 * @to - destination buffer
168 * @from - source buffer
169 * @tolen - destination buffer size (in bytes)
170 * @fromlen - source buffer size (in bytes)
171 * @codepage - codepage to which characters should be converted
172 * @mapchar - should characters be remapped according to the mapchars option?
174 * Convert a little-endian utf16le string (as sent by the server) to a string
175 * in the provided codepage. The tolen and fromlen parameters are to ensure
176 * that the code doesn't walk off of the end of the buffer (which is always
177 * a danger if the alignment of the source buffer is off). The destination
178 * string is always properly null terminated and fits in the destination
179 * buffer. Returns the length of the destination string in bytes (including
182 * Note that some windows versions actually send multiword UTF-16 characters
183 * instead of straight UTF16-2. The linux nls routines however aren't able to
184 * deal with those characters properly. In the event that we get some of
185 * those characters, they won't be translated properly.
188 cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
189 const struct nls_table *codepage, int map_type)
191 int i, charlen, safelen;
193 int nullsize = nls_nullsize(codepage);
194 int fromwords = fromlen / 2;
195 char tmp[NLS_MAX_CHARSET_SIZE];
196 __u16 ftmp[3]; /* ftmp[3] = 3array x 2bytes = 6bytes UTF-16 */
199 * because the chars can be of varying widths, we need to take care
200 * not to overflow the destination buffer when we get close to the
201 * end of it. Until we get to this offset, we don't need to check
202 * for overflow however.
204 safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
206 for (i = 0; i < fromwords; i++) {
207 ftmp[0] = get_unaligned_le16(&from[i]);
210 if (i + 1 < fromwords)
211 ftmp[1] = get_unaligned_le16(&from[i + 1]);
214 if (i + 2 < fromwords)
215 ftmp[2] = get_unaligned_le16(&from[i + 2]);
220 * check to see if converting this character might make the
221 * conversion bleed into the null terminator
223 if (outlen >= safelen) {
224 charlen = cifs_mapchar(tmp, ftmp, codepage, map_type);
225 if ((outlen + charlen) > (tolen - nullsize))
229 /* put converted char into 'to' buffer */
230 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, map_type);
233 /* charlen (=bytes of UTF-8 for 1 character)
234 * 4bytes UTF-8(surrogate pair) is charlen=4
235 * (4bytes UTF-16 code)
236 * 7-8bytes UTF-8(IVS) is charlen=3+4 or 4+4
237 * (2 UTF-8 pairs divided to 2 UTF-16 pairs) */
240 else if (charlen >= 5)
245 /* properly null-terminate string */
246 for (i = 0; i < nullsize; i++)
253 * NAME: cifs_strtoUTF16()
255 * FUNCTION: Convert character string to unicode string
259 cifs_strtoUTF16(__le16 *to, const char *from, int len,
260 const struct nls_table *codepage)
264 wchar_t wchar_to; /* needed to quiet sparse */
266 /* special case for utf8 to handle no plane0 chars */
267 if (!strcmp(codepage->charset, "utf8")) {
269 * convert utf8 -> utf16, we assume we have enough space
270 * as caller should have assumed conversion does not overflow
271 * in destination len is length in wchar_t units (16bits)
273 i = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
274 (wchar_t *) to, len);
276 /* if success terminate and exit */
280 * if fails fall back to UCS encoding as this
281 * function should not return negative values
282 * currently can fail only if source contains
283 * invalid encoded characters
287 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
288 charlen = codepage->char2uni(from, len, &wchar_to);
290 cifs_dbg(VFS, "strtoUTF16: char2uni of 0x%x returned %d\n",
292 /* A question mark */
296 put_unaligned_le16(wchar_to, &to[i]);
300 put_unaligned_le16(0, &to[i]);
305 * cifs_utf16_bytes - how long will a string be after conversion?
306 * @utf16 - pointer to input string
307 * @maxbytes - don't go past this many bytes of input string
308 * @codepage - destination codepage
310 * Walk a utf16le string and return the number of bytes that the string will
311 * be after being converted to the given charset, not including any null
312 * termination required. Don't walk past maxbytes in the source buffer.
315 cifs_utf16_bytes(const __le16 *from, int maxbytes,
316 const struct nls_table *codepage)
319 int charlen, outlen = 0;
320 int maxwords = maxbytes / 2;
321 char tmp[NLS_MAX_CHARSET_SIZE];
324 for (i = 0; i < maxwords; i++) {
325 ftmp[0] = get_unaligned_le16(&from[i]);
328 if (i + 1 < maxwords)
329 ftmp[1] = get_unaligned_le16(&from[i + 1]);
332 if (i + 2 < maxwords)
333 ftmp[2] = get_unaligned_le16(&from[i + 2]);
337 charlen = cifs_mapchar(tmp, ftmp, codepage, NO_MAP_UNI_RSVD);
345 * cifs_strndup_from_utf16 - copy a string from wire format to the local
347 * @src - source string
348 * @maxlen - don't walk past this many bytes in the source string
349 * @is_unicode - is this a unicode string?
350 * @codepage - destination codepage
352 * Take a string given by the server, convert it to the local codepage and
353 * put it in a new buffer. Returns a pointer to the new string or NULL on
357 cifs_strndup_from_utf16(const char *src, const int maxlen,
358 const bool is_unicode, const struct nls_table *codepage)
364 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
365 len += nls_nullsize(codepage);
366 dst = kmalloc(len, GFP_KERNEL);
369 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
372 dst = kstrndup(src, maxlen, GFP_KERNEL);
378 static __le16 convert_to_sfu_char(char src_char)
384 dest_char = cpu_to_le16(UNI_COLON);
387 dest_char = cpu_to_le16(UNI_ASTERISK);
390 dest_char = cpu_to_le16(UNI_QUESTION);
393 dest_char = cpu_to_le16(UNI_LESSTHAN);
396 dest_char = cpu_to_le16(UNI_GRTRTHAN);
399 dest_char = cpu_to_le16(UNI_PIPE);
408 static __le16 convert_to_sfm_char(char src_char, bool end_of_string)
414 dest_char = cpu_to_le16(SFM_COLON);
417 dest_char = cpu_to_le16(SFM_DOUBLEQUOTE);
420 dest_char = cpu_to_le16(SFM_ASTERISK);
423 dest_char = cpu_to_le16(SFM_QUESTION);
426 dest_char = cpu_to_le16(SFM_LESSTHAN);
429 dest_char = cpu_to_le16(SFM_GRTRTHAN);
432 dest_char = cpu_to_le16(SFM_PIPE);
436 dest_char = cpu_to_le16(SFM_PERIOD);
442 dest_char = cpu_to_le16(SFM_SPACE);
454 * Convert 16 bit Unicode pathname to wire format from string in current code
455 * page. Conversion may involve remapping up the six characters that are
456 * only legal in POSIX-like OS (if they are present in the string). Path
457 * names are little endian 16 bit Unicode on the wire
460 cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
461 const struct nls_table *cp, int map_chars)
468 wchar_t *wchar_to; /* UTF-16 */
472 if (map_chars == NO_MAP_UNI_RSVD)
473 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
475 wchar_to = kzalloc(6, GFP_KERNEL);
477 for (i = 0; i < srclen; j++) {
478 src_char = source[i];
481 /* check if end of string */
485 /* see if we must remap this char */
486 if (map_chars == SFU_MAP_UNI_RSVD)
487 dst_char = convert_to_sfu_char(src_char);
488 else if (map_chars == SFM_MAP_UNI_RSVD) {
492 * Remap spaces and periods found at the end of every
493 * component of the path. The special cases of '.' and
494 * '..' do not need to be dealt with explicitly because
495 * they are addressed in namei.c:link_path_walk().
497 if ((i == srclen - 1) || (source[i+1] == '\\'))
498 end_of_string = true;
500 end_of_string = false;
502 dst_char = convert_to_sfm_char(src_char, end_of_string);
506 * FIXME: We can not handle remapping backslash (UNI_SLASH)
507 * until all the calls to build_path_from_dentry are modified,
508 * as they use backslash as separator.
511 charlen = cp->char2uni(source + i, srclen - i, &tmp);
512 dst_char = cpu_to_le16(tmp);
515 * if no match, use question mark, which at least in
516 * some cases serves as wild card
521 /* convert SURROGATE_PAIR */
522 if (strcmp(cp->charset, "utf8") || !wchar_to)
524 if (*(source + i) & 0x80) {
525 charlen = utf8_to_utf32(source + i, 6, &u);
530 ret = utf8s_to_utf16s(source + i, charlen,
537 dst_char = cpu_to_le16(*wchar_to);
539 /* 1-3bytes UTF-8 to 2bytes UTF-16 */
540 put_unaligned(dst_char, &target[j]);
541 else if (charlen == 4) {
542 /* 4bytes UTF-8(surrogate pair) to 4bytes UTF-16
543 * 7-8bytes UTF-8(IVS) divided to 2 UTF-16
544 * (charlen=3+4 or 4+4) */
545 put_unaligned(dst_char, &target[j]);
546 dst_char = cpu_to_le16(*(wchar_to + 1));
548 put_unaligned(dst_char, &target[j]);
549 } else if (charlen >= 5) {
550 /* 5-6bytes UTF-8 to 6bytes UTF-16 */
551 put_unaligned(dst_char, &target[j]);
552 dst_char = cpu_to_le16(*(wchar_to + 1));
554 put_unaligned(dst_char, &target[j]);
555 dst_char = cpu_to_le16(*(wchar_to + 2));
557 put_unaligned(dst_char, &target[j]);
562 dst_char = cpu_to_le16(0x003f);
568 * character may take more than one byte in the source string,
569 * but will take exactly two bytes in the target string
572 put_unaligned(dst_char, &target[j]);
576 put_unaligned(0, &target[j]); /* Null terminate target unicode string */
581 #ifdef CONFIG_CIFS_SMB2
583 * cifs_local_to_utf16_bytes - how long will a string be after conversion?
584 * @from - pointer to input string
585 * @maxbytes - don't go past this many bytes of input string
586 * @codepage - source codepage
588 * Walk a string and return the number of bytes that the string will
589 * be after being converted to the given charset, not including any null
590 * termination required. Don't walk past maxbytes in the source buffer.
594 cifs_local_to_utf16_bytes(const char *from, int len,
595 const struct nls_table *codepage)
601 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
602 charlen = codepage->char2uni(from, len, &wchar_to);
603 /* Failed conversion defaults to a question mark */
607 return 2 * i; /* UTF16 characters are two bytes */
611 * cifs_strndup_to_utf16 - copy a string to wire format from the local codepage
612 * @src - source string
613 * @maxlen - don't walk past this many bytes in the source string
614 * @utf16_len - the length of the allocated string in bytes (including null)
615 * @cp - source codepage
616 * @remap - map special chars
618 * Take a string convert it from the local codepage to UTF16 and
619 * put it in a new buffer. Returns a pointer to the new string or NULL on
623 cifs_strndup_to_utf16(const char *src, const int maxlen, int *utf16_len,
624 const struct nls_table *cp, int remap)
629 len = cifs_local_to_utf16_bytes(src, maxlen, cp);
631 dst = kmalloc(len, GFP_KERNEL);
636 cifsConvertToUTF16(dst, src, strlen(src), cp, remap);
640 #endif /* CONFIG_CIFS_SMB2 */