blob: fbb9da95184379bcec53499a7fcac55ff6078f17 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/*
2 * fs/cifs/cifs_unicode.c
3 *
Steve Frenchd185cda2009-04-30 17:45:10 +00004 * Copyright (c) International Business Machines Corp., 2000,2009
Linus Torvalds1da177e2005-04-16 15:20:36 -07005 * Modified by Steve French (sfrench@us.ibm.com)
6 *
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
Steve French221601c2007-06-05 20:35:06 +00009 * the Free Software Foundation; either version 2 of the License, or
Linus Torvalds1da177e2005-04-16 15:20:36 -070010 * (at your option) any later version.
Steve French221601c2007-06-05 20:35:06 +000011 *
Linus Torvalds1da177e2005-04-16 15:20:36 -070012 * 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.
16 *
17 * You should have received a copy of the GNU General Public License
Steve French221601c2007-06-05 20:35:06 +000018 * along with this program; if not, write to the Free Software
Linus Torvalds1da177e2005-04-16 15:20:36 -070019 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20 */
21#include <linux/fs.h>
Tejun Heo5a0e3ad2010-03-24 17:04:11 +090022#include <linux/slab.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070023#include "cifs_unicode.h"
24#include "cifs_uniupr.h"
25#include "cifspdu.h"
Steve French39798772006-05-31 22:40:51 +000026#include "cifsglob.h"
Linus Torvalds1da177e2005-04-16 15:20:36 -070027#include "cifs_debug.h"
28
29/*
Steve Frenchacbbb762012-01-18 22:32:33 -060030 * cifs_utf16_bytes - how long will a string be after conversion?
31 * @utf16 - pointer to input string
Jeff Layton69f801f2009-04-30 06:46:32 -040032 * @maxbytes - don't go past this many bytes of input string
33 * @codepage - destination codepage
34 *
Steve Frenchacbbb762012-01-18 22:32:33 -060035 * Walk a utf16le string and return the number of bytes that the string will
Jeff Layton69f801f2009-04-30 06:46:32 -040036 * be after being converted to the given charset, not including any null
37 * termination required. Don't walk past maxbytes in the source buffer.
38 */
39int
Steve Frenchacbbb762012-01-18 22:32:33 -060040cifs_utf16_bytes(const __le16 *from, int maxbytes,
Jeff Layton69f801f2009-04-30 06:46:32 -040041 const struct nls_table *codepage)
42{
43 int i;
44 int charlen, outlen = 0;
45 int maxwords = maxbytes / 2;
46 char tmp[NLS_MAX_CHARSET_SIZE];
Jeff Laytonba2dbf32011-01-20 13:36:51 -050047 __u16 ftmp;
Jeff Layton69f801f2009-04-30 06:46:32 -040048
Jeff Laytonba2dbf32011-01-20 13:36:51 -050049 for (i = 0; i < maxwords; i++) {
50 ftmp = get_unaligned_le16(&from[i]);
51 if (ftmp == 0)
52 break;
53
54 charlen = codepage->uni2char(ftmp, tmp, NLS_MAX_CHARSET_SIZE);
Jeff Layton69f801f2009-04-30 06:46:32 -040055 if (charlen > 0)
56 outlen += charlen;
57 else
58 outlen++;
59 }
60
61 return outlen;
62}
63
64/*
Jeff Laytonba2dbf32011-01-20 13:36:51 -050065 * cifs_mapchar - convert a host-endian char to proper char in codepage
Jeff Layton7fabf0c2009-04-30 06:46:15 -040066 * @target - where converted character should be copied
Jeff Laytonba2dbf32011-01-20 13:36:51 -050067 * @src_char - 2 byte host-endian source character
Jeff Layton7fabf0c2009-04-30 06:46:15 -040068 * @cp - codepage to which character should be converted
69 * @mapchar - should character be mapped according to mapchars mount option?
70 *
71 * This function handles the conversion of a single character. It is the
72 * responsibility of the caller to ensure that the target buffer is large
73 * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
74 */
75static int
Jeff Laytonba2dbf32011-01-20 13:36:51 -050076cifs_mapchar(char *target, const __u16 src_char, const struct nls_table *cp,
Jeff Layton7fabf0c2009-04-30 06:46:15 -040077 bool mapchar)
78{
79 int len = 1;
80
81 if (!mapchar)
82 goto cp_convert;
83
84 /*
85 * BB: Cannot handle remapping UNI_SLASH until all the calls to
86 * build_path_from_dentry are modified, as they use slash as
87 * separator.
88 */
Jeff Laytonba2dbf32011-01-20 13:36:51 -050089 switch (src_char) {
Jeff Layton7fabf0c2009-04-30 06:46:15 -040090 case UNI_COLON:
91 *target = ':';
92 break;
Jeff Layton581ade42011-04-05 15:02:37 -040093 case UNI_ASTERISK:
Jeff Layton7fabf0c2009-04-30 06:46:15 -040094 *target = '*';
95 break;
96 case UNI_QUESTION:
97 *target = '?';
98 break;
99 case UNI_PIPE:
100 *target = '|';
101 break;
102 case UNI_GRTRTHAN:
103 *target = '>';
104 break;
105 case UNI_LESSTHAN:
106 *target = '<';
107 break;
108 default:
109 goto cp_convert;
110 }
111
112out:
113 return len;
114
115cp_convert:
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500116 len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400117 if (len <= 0) {
118 *target = '?';
119 len = 1;
120 }
121 goto out;
122}
123
124/*
Steve Frenchacbbb762012-01-18 22:32:33 -0600125 * cifs_from_utf16 - convert utf16le string to local charset
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400126 * @to - destination buffer
127 * @from - source buffer
128 * @tolen - destination buffer size (in bytes)
129 * @fromlen - source buffer size (in bytes)
130 * @codepage - codepage to which characters should be converted
131 * @mapchar - should characters be remapped according to the mapchars option?
132 *
Steve Frenchacbbb762012-01-18 22:32:33 -0600133 * Convert a little-endian utf16le string (as sent by the server) to a string
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400134 * in the provided codepage. The tolen and fromlen parameters are to ensure
135 * that the code doesn't walk off of the end of the buffer (which is always
136 * a danger if the alignment of the source buffer is off). The destination
137 * string is always properly null terminated and fits in the destination
138 * buffer. Returns the length of the destination string in bytes (including
139 * null terminator).
140 *
141 * Note that some windows versions actually send multiword UTF-16 characters
Steve Frenchacbbb762012-01-18 22:32:33 -0600142 * instead of straight UTF16-2. The linux nls routines however aren't able to
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400143 * deal with those characters properly. In the event that we get some of
144 * those characters, they won't be translated properly.
145 */
146int
Steve Frenchacbbb762012-01-18 22:32:33 -0600147cifs_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400148 const struct nls_table *codepage, bool mapchar)
149{
150 int i, charlen, safelen;
151 int outlen = 0;
152 int nullsize = nls_nullsize(codepage);
153 int fromwords = fromlen / 2;
154 char tmp[NLS_MAX_CHARSET_SIZE];
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500155 __u16 ftmp;
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400156
157 /*
158 * because the chars can be of varying widths, we need to take care
159 * not to overflow the destination buffer when we get close to the
160 * end of it. Until we get to this offset, we don't need to check
161 * for overflow however.
162 */
163 safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
164
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500165 for (i = 0; i < fromwords; i++) {
166 ftmp = get_unaligned_le16(&from[i]);
167 if (ftmp == 0)
168 break;
169
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400170 /*
171 * check to see if converting this character might make the
172 * conversion bleed into the null terminator
173 */
174 if (outlen >= safelen) {
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500175 charlen = cifs_mapchar(tmp, ftmp, codepage, mapchar);
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400176 if ((outlen + charlen) > (tolen - nullsize))
177 break;
178 }
179
180 /* put converted char into 'to' buffer */
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500181 charlen = cifs_mapchar(&to[outlen], ftmp, codepage, mapchar);
Jeff Layton7fabf0c2009-04-30 06:46:15 -0400182 outlen += charlen;
183 }
184
185 /* properly null-terminate string */
186 for (i = 0; i < nullsize; i++)
187 to[outlen++] = 0;
188
189 return outlen;
190}
191
192/*
Steve Frenchacbbb762012-01-18 22:32:33 -0600193 * NAME: cifs_strtoUTF16()
Linus Torvalds1da177e2005-04-16 15:20:36 -0700194 *
195 * FUNCTION: Convert character string to unicode string
196 *
197 */
198int
Steve Frenchacbbb762012-01-18 22:32:33 -0600199cifs_strtoUTF16(__le16 *to, const char *from, int len,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700200 const struct nls_table *codepage)
201{
202 int charlen;
203 int i;
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500204 wchar_t wchar_to; /* needed to quiet sparse */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700205
206 for (i = 0; len && *from; i++, from += charlen, len -= charlen) {
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500207 charlen = codepage->char2uni(from, len, &wchar_to);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700208 if (charlen < 1) {
Steve Frenchacbbb762012-01-18 22:32:33 -0600209 cERROR(1, "strtoUTF16: char2uni of 0x%x returned %d",
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500210 *from, charlen);
Steve French69114082005-11-10 19:28:44 -0800211 /* A question mark */
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500212 wchar_to = 0x003f;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700213 charlen = 1;
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500214 }
215 put_unaligned_le16(wchar_to, &to[i]);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700216 }
217
Jeff Laytonba2dbf32011-01-20 13:36:51 -0500218 put_unaligned_le16(0, &to[i]);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700219 return i;
220}
221
Jeff Layton066ce682009-04-30 07:16:14 -0400222/*
Steve Frenchacbbb762012-01-18 22:32:33 -0600223 * cifs_strndup_from_utf16 - copy a string from wire format to the local
224 * codepage
Jeff Layton066ce682009-04-30 07:16:14 -0400225 * @src - source string
226 * @maxlen - don't walk past this many bytes in the source string
227 * @is_unicode - is this a unicode string?
228 * @codepage - destination codepage
229 *
230 * Take a string given by the server, convert it to the local codepage and
231 * put it in a new buffer. Returns a pointer to the new string or NULL on
232 * error.
233 */
234char *
Steve Frenchacbbb762012-01-18 22:32:33 -0600235cifs_strndup_from_utf16(const char *src, const int maxlen,
236 const bool is_unicode, const struct nls_table *codepage)
Jeff Layton066ce682009-04-30 07:16:14 -0400237{
238 int len;
239 char *dst;
240
241 if (is_unicode) {
Steve Frenchacbbb762012-01-18 22:32:33 -0600242 len = cifs_utf16_bytes((__le16 *) src, maxlen, codepage);
Jeff Layton066ce682009-04-30 07:16:14 -0400243 len += nls_nullsize(codepage);
244 dst = kmalloc(len, GFP_KERNEL);
245 if (!dst)
246 return NULL;
Steve Frenchacbbb762012-01-18 22:32:33 -0600247 cifs_from_utf16(dst, (__le16 *) src, len, maxlen, codepage,
Jeff Layton066ce682009-04-30 07:16:14 -0400248 false);
249 } else {
250 len = strnlen(src, maxlen);
251 len++;
252 dst = kmalloc(len, GFP_KERNEL);
253 if (!dst)
254 return NULL;
255 strlcpy(dst, src, len);
256 }
257
258 return dst;
259}
260
Jeff Layton84cdf742011-01-20 13:36:51 -0500261/*
262 * Convert 16 bit Unicode pathname to wire format from string in current code
263 * page. Conversion may involve remapping up the six characters that are
264 * only legal in POSIX-like OS (if they are present in the string). Path
265 * names are little endian 16 bit Unicode on the wire
266 */
267int
Steve Frenchacbbb762012-01-18 22:32:33 -0600268cifsConvertToUTF16(__le16 *target, const char *source, int srclen,
Jeff Layton84cdf742011-01-20 13:36:51 -0500269 const struct nls_table *cp, int mapChars)
270{
271 int i, j, charlen;
Jeff Layton84cdf742011-01-20 13:36:51 -0500272 char src_char;
Jeff Layton581ade42011-04-05 15:02:37 -0400273 __le16 dst_char;
274 wchar_t tmp;
Jeff Layton84cdf742011-01-20 13:36:51 -0500275
276 if (!mapChars)
Steve Frenchacbbb762012-01-18 22:32:33 -0600277 return cifs_strtoUTF16(target, source, PATH_MAX, cp);
Jeff Layton84cdf742011-01-20 13:36:51 -0500278
Jeff Layton581ade42011-04-05 15:02:37 -0400279 for (i = 0, j = 0; i < srclen; j++) {
Jeff Layton84cdf742011-01-20 13:36:51 -0500280 src_char = source[i];
Jeff Layton11379b52011-05-17 15:28:21 -0400281 charlen = 1;
Jeff Layton84cdf742011-01-20 13:36:51 -0500282 switch (src_char) {
283 case 0:
Jeff Layton581ade42011-04-05 15:02:37 -0400284 put_unaligned(0, &target[j]);
Steve Frenchacbbb762012-01-18 22:32:33 -0600285 goto ctoUTF16_out;
Jeff Layton84cdf742011-01-20 13:36:51 -0500286 case ':':
Jeff Layton581ade42011-04-05 15:02:37 -0400287 dst_char = cpu_to_le16(UNI_COLON);
Jeff Layton84cdf742011-01-20 13:36:51 -0500288 break;
289 case '*':
Jeff Layton581ade42011-04-05 15:02:37 -0400290 dst_char = cpu_to_le16(UNI_ASTERISK);
Jeff Layton84cdf742011-01-20 13:36:51 -0500291 break;
292 case '?':
Jeff Layton581ade42011-04-05 15:02:37 -0400293 dst_char = cpu_to_le16(UNI_QUESTION);
Jeff Layton84cdf742011-01-20 13:36:51 -0500294 break;
295 case '<':
Jeff Layton581ade42011-04-05 15:02:37 -0400296 dst_char = cpu_to_le16(UNI_LESSTHAN);
Jeff Layton84cdf742011-01-20 13:36:51 -0500297 break;
298 case '>':
Jeff Layton581ade42011-04-05 15:02:37 -0400299 dst_char = cpu_to_le16(UNI_GRTRTHAN);
Jeff Layton84cdf742011-01-20 13:36:51 -0500300 break;
301 case '|':
Jeff Layton581ade42011-04-05 15:02:37 -0400302 dst_char = cpu_to_le16(UNI_PIPE);
Jeff Layton84cdf742011-01-20 13:36:51 -0500303 break;
304 /*
305 * FIXME: We can not handle remapping backslash (UNI_SLASH)
306 * until all the calls to build_path_from_dentry are modified,
307 * as they use backslash as separator.
308 */
309 default:
Jeff Layton581ade42011-04-05 15:02:37 -0400310 charlen = cp->char2uni(source + i, srclen - i, &tmp);
311 dst_char = cpu_to_le16(tmp);
312
Jeff Layton84cdf742011-01-20 13:36:51 -0500313 /*
314 * if no match, use question mark, which at least in
315 * some cases serves as wild card
316 */
317 if (charlen < 1) {
Jeff Layton581ade42011-04-05 15:02:37 -0400318 dst_char = cpu_to_le16(0x003f);
Jeff Layton84cdf742011-01-20 13:36:51 -0500319 charlen = 1;
320 }
Jeff Layton84cdf742011-01-20 13:36:51 -0500321 }
Jeff Layton11379b52011-05-17 15:28:21 -0400322 /*
323 * character may take more than one byte in the source string,
324 * but will take exactly two bytes in the target string
325 */
326 i += charlen;
Jeff Layton581ade42011-04-05 15:02:37 -0400327 put_unaligned(dst_char, &target[j]);
Jeff Layton84cdf742011-01-20 13:36:51 -0500328 }
329
Steve Frenchacbbb762012-01-18 22:32:33 -0600330ctoUTF16_out:
Jeff Layton84cdf742011-01-20 13:36:51 -0500331 return i;
332}
333