blob: 0ea887fc859cc358dca5276f65ea1bcc9bb04e2f [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/*
2 * unistr.c - NTFS Unicode string handling. Part of the Linux-NTFS project.
3 *
Anton Altaparmakovc394e452005-10-04 13:08:53 +01004 * Copyright (c) 2001-2005 Anton Altaparmakov
Linus Torvalds1da177e2005-04-16 15:20:36 -07005 *
6 * This program/include file is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU General Public License as published
8 * by the Free Software Foundation; either version 2 of the License, or
9 * (at your option) any later version.
10 *
11 * This program/include file is distributed in the hope that it will be
12 * useful, but WITHOUT ANY WARRANTY; without even the implied warranty
13 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
15 *
16 * You should have received a copy of the GNU General Public License
17 * along with this program (in the main directory of the Linux-NTFS
18 * distribution in the file COPYING); if not, write to the Free Software
19 * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20 */
21
22#include "types.h"
23#include "debug.h"
24#include "ntfs.h"
25
26/*
27 * IMPORTANT
28 * =========
29 *
30 * All these routines assume that the Unicode characters are in little endian
31 * encoding inside the strings!!!
32 */
33
34/*
35 * This is used by the name collation functions to quickly determine what
36 * characters are (in)valid.
37 */
38static const u8 legal_ansi_char_array[0x40] = {
39 0x00, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
40 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
41
42 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
43 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
44
45 0x17, 0x07, 0x18, 0x17, 0x17, 0x17, 0x17, 0x17,
46 0x17, 0x17, 0x18, 0x16, 0x16, 0x17, 0x07, 0x00,
47
48 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17,
49 0x17, 0x17, 0x04, 0x16, 0x18, 0x16, 0x18, 0x18,
50};
51
52/**
53 * ntfs_are_names_equal - compare two Unicode names for equality
54 * @s1: name to compare to @s2
55 * @s1_len: length in Unicode characters of @s1
56 * @s2: name to compare to @s1
57 * @s2_len: length in Unicode characters of @s2
58 * @ic: ignore case bool
59 * @upcase: upcase table (only if @ic == IGNORE_CASE)
60 * @upcase_size: length in Unicode characters of @upcase (if present)
61 *
62 * Compare the names @s1 and @s2 and return TRUE (1) if the names are
63 * identical, or FALSE (0) if they are not identical. If @ic is IGNORE_CASE,
64 * the @upcase table is used to performa a case insensitive comparison.
65 */
66BOOL ntfs_are_names_equal(const ntfschar *s1, size_t s1_len,
67 const ntfschar *s2, size_t s2_len, const IGNORE_CASE_BOOL ic,
68 const ntfschar *upcase, const u32 upcase_size)
69{
70 if (s1_len != s2_len)
71 return FALSE;
72 if (ic == CASE_SENSITIVE)
73 return !ntfs_ucsncmp(s1, s2, s1_len);
74 return !ntfs_ucsncasecmp(s1, s2, s1_len, upcase, upcase_size);
75}
76
77/**
78 * ntfs_collate_names - collate two Unicode names
79 * @name1: first Unicode name to compare
80 * @name2: second Unicode name to compare
81 * @err_val: if @name1 contains an invalid character return this value
82 * @ic: either CASE_SENSITIVE or IGNORE_CASE
83 * @upcase: upcase table (ignored if @ic is CASE_SENSITIVE)
84 * @upcase_len: upcase table size (ignored if @ic is CASE_SENSITIVE)
85 *
86 * ntfs_collate_names collates two Unicode names and returns:
87 *
88 * -1 if the first name collates before the second one,
89 * 0 if the names match,
90 * 1 if the second name collates before the first one, or
91 * @err_val if an invalid character is found in @name1 during the comparison.
92 *
93 * The following characters are considered invalid: '"', '*', '<', '>' and '?'.
94 */
95int ntfs_collate_names(const ntfschar *name1, const u32 name1_len,
96 const ntfschar *name2, const u32 name2_len,
97 const int err_val, const IGNORE_CASE_BOOL ic,
98 const ntfschar *upcase, const u32 upcase_len)
99{
100 u32 cnt, min_len;
101 u16 c1, c2;
102
103 min_len = name1_len;
104 if (name1_len > name2_len)
105 min_len = name2_len;
106 for (cnt = 0; cnt < min_len; ++cnt) {
107 c1 = le16_to_cpu(*name1++);
108 c2 = le16_to_cpu(*name2++);
109 if (ic) {
110 if (c1 < upcase_len)
111 c1 = le16_to_cpu(upcase[c1]);
112 if (c2 < upcase_len)
113 c2 = le16_to_cpu(upcase[c2]);
114 }
115 if (c1 < 64 && legal_ansi_char_array[c1] & 8)
116 return err_val;
117 if (c1 < c2)
118 return -1;
119 if (c1 > c2)
120 return 1;
121 }
122 if (name1_len < name2_len)
123 return -1;
124 if (name1_len == name2_len)
125 return 0;
126 /* name1_len > name2_len */
127 c1 = le16_to_cpu(*name1);
128 if (c1 < 64 && legal_ansi_char_array[c1] & 8)
129 return err_val;
130 return 1;
131}
132
133/**
134 * ntfs_ucsncmp - compare two little endian Unicode strings
135 * @s1: first string
136 * @s2: second string
137 * @n: maximum unicode characters to compare
138 *
139 * Compare the first @n characters of the Unicode strings @s1 and @s2,
140 * The strings in little endian format and appropriate le16_to_cpu()
141 * conversion is performed on non-little endian machines.
142 *
143 * The function returns an integer less than, equal to, or greater than zero
144 * if @s1 (or the first @n Unicode characters thereof) is found, respectively,
145 * to be less than, to match, or be greater than @s2.
146 */
147int ntfs_ucsncmp(const ntfschar *s1, const ntfschar *s2, size_t n)
148{
149 u16 c1, c2;
150 size_t i;
151
152 for (i = 0; i < n; ++i) {
153 c1 = le16_to_cpu(s1[i]);
154 c2 = le16_to_cpu(s2[i]);
155 if (c1 < c2)
156 return -1;
157 if (c1 > c2)
158 return 1;
159 if (!c1)
160 break;
161 }
162 return 0;
163}
164
165/**
166 * ntfs_ucsncasecmp - compare two little endian Unicode strings, ignoring case
167 * @s1: first string
168 * @s2: second string
169 * @n: maximum unicode characters to compare
170 * @upcase: upcase table
171 * @upcase_size: upcase table size in Unicode characters
172 *
173 * Compare the first @n characters of the Unicode strings @s1 and @s2,
174 * ignoring case. The strings in little endian format and appropriate
175 * le16_to_cpu() conversion is performed on non-little endian machines.
176 *
177 * Each character is uppercased using the @upcase table before the comparison.
178 *
179 * The function returns an integer less than, equal to, or greater than zero
180 * if @s1 (or the first @n Unicode characters thereof) is found, respectively,
181 * to be less than, to match, or be greater than @s2.
182 */
183int ntfs_ucsncasecmp(const ntfschar *s1, const ntfschar *s2, size_t n,
184 const ntfschar *upcase, const u32 upcase_size)
185{
186 size_t i;
187 u16 c1, c2;
188
189 for (i = 0; i < n; ++i) {
190 if ((c1 = le16_to_cpu(s1[i])) < upcase_size)
191 c1 = le16_to_cpu(upcase[c1]);
192 if ((c2 = le16_to_cpu(s2[i])) < upcase_size)
193 c2 = le16_to_cpu(upcase[c2]);
194 if (c1 < c2)
195 return -1;
196 if (c1 > c2)
197 return 1;
198 if (!c1)
199 break;
200 }
201 return 0;
202}
203
204void ntfs_upcase_name(ntfschar *name, u32 name_len, const ntfschar *upcase,
205 const u32 upcase_len)
206{
207 u32 i;
208 u16 u;
209
210 for (i = 0; i < name_len; i++)
211 if ((u = le16_to_cpu(name[i])) < upcase_len)
212 name[i] = upcase[u];
213}
214
215void ntfs_file_upcase_value(FILE_NAME_ATTR *file_name_attr,
216 const ntfschar *upcase, const u32 upcase_len)
217{
218 ntfs_upcase_name((ntfschar*)&file_name_attr->file_name,
219 file_name_attr->file_name_length, upcase, upcase_len);
220}
221
222int ntfs_file_compare_values(FILE_NAME_ATTR *file_name_attr1,
223 FILE_NAME_ATTR *file_name_attr2,
224 const int err_val, const IGNORE_CASE_BOOL ic,
225 const ntfschar *upcase, const u32 upcase_len)
226{
227 return ntfs_collate_names((ntfschar*)&file_name_attr1->file_name,
228 file_name_attr1->file_name_length,
229 (ntfschar*)&file_name_attr2->file_name,
230 file_name_attr2->file_name_length,
231 err_val, ic, upcase, upcase_len);
232}
233
234/**
235 * ntfs_nlstoucs - convert NLS string to little endian Unicode string
236 * @vol: ntfs volume which we are working with
237 * @ins: input NLS string buffer
238 * @ins_len: length of input string in bytes
239 * @outs: on return contains the allocated output Unicode string buffer
240 *
241 * Convert the input string @ins, which is in whatever format the loaded NLS
242 * map dictates, into a little endian, 2-byte Unicode string.
243 *
244 * This function allocates the string and the caller is responsible for
245 * calling kmem_cache_free(ntfs_name_cache, @outs); when finished with it.
246 *
247 * On success the function returns the number of Unicode characters written to
248 * the output string *@outs (>= 0), not counting the terminating Unicode NULL
249 * character. *@outs is set to the allocated output string buffer.
250 *
251 * On error, a negative number corresponding to the error code is returned. In
252 * that case the output string is not allocated. Both *@outs and *@outs_len
253 * are then undefined.
254 *
255 * This might look a bit odd due to fast path optimization...
256 */
257int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
258 const int ins_len, ntfschar **outs)
259{
260 struct nls_table *nls = vol->nls_map;
261 ntfschar *ucs;
262 wchar_t wc;
263 int i, o, wc_len;
264
265 /* We don't trust outside sources. */
266 if (ins) {
Pekka Enberg2fb21db2005-05-25 21:15:34 +0300267 ucs = kmem_cache_alloc(ntfs_name_cache, SLAB_NOFS);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700268 if (ucs) {
269 for (i = o = 0; i < ins_len; i += wc_len) {
270 wc_len = nls->char2uni(ins + i, ins_len - i,
271 &wc);
272 if (wc_len >= 0) {
273 if (wc) {
274 ucs[o++] = cpu_to_le16(wc);
275 continue;
276 } /* else (!wc) */
277 break;
278 } /* else (wc_len < 0) */
279 goto conversion_err;
280 }
281 ucs[o] = 0;
282 *outs = ucs;
283 return o;
284 } /* else (!ucs) */
285 ntfs_error(vol->sb, "Failed to allocate name from "
286 "ntfs_name_cache!");
287 return -ENOMEM;
288 } /* else (!ins) */
289 ntfs_error(NULL, "Received NULL pointer.");
290 return -EINVAL;
291conversion_err:
292 ntfs_error(vol->sb, "Name using character set %s contains characters "
293 "that cannot be converted to Unicode.", nls->charset);
294 kmem_cache_free(ntfs_name_cache, ucs);
295 return -EILSEQ;
296}
297
298/**
299 * ntfs_ucstonls - convert little endian Unicode string to NLS string
300 * @vol: ntfs volume which we are working with
301 * @ins: input Unicode string buffer
302 * @ins_len: length of input string in Unicode characters
303 * @outs: on return contains the (allocated) output NLS string buffer
304 * @outs_len: length of output string buffer in bytes
305 *
306 * Convert the input little endian, 2-byte Unicode string @ins, of length
307 * @ins_len into the string format dictated by the loaded NLS.
308 *
309 * If *@outs is NULL, this function allocates the string and the caller is
310 * responsible for calling kfree(*@outs); when finished with it. In this case
311 * @outs_len is ignored and can be 0.
312 *
313 * On success the function returns the number of bytes written to the output
314 * string *@outs (>= 0), not counting the terminating NULL byte. If the output
315 * string buffer was allocated, *@outs is set to it.
316 *
317 * On error, a negative number corresponding to the error code is returned. In
318 * that case the output string is not allocated. The contents of *@outs are
319 * then undefined.
320 *
321 * This might look a bit odd due to fast path optimization...
322 */
323int ntfs_ucstonls(const ntfs_volume *vol, const ntfschar *ins,
324 const int ins_len, unsigned char **outs, int outs_len)
325{
326 struct nls_table *nls = vol->nls_map;
327 unsigned char *ns;
328 int i, o, ns_len, wc;
329
330 /* We don't trust outside sources. */
331 if (ins) {
332 ns = *outs;
333 ns_len = outs_len;
334 if (ns && !ns_len) {
335 wc = -ENAMETOOLONG;
336 goto conversion_err;
337 }
338 if (!ns) {
339 ns_len = ins_len * NLS_MAX_CHARSET_SIZE;
340 ns = (unsigned char*)kmalloc(ns_len + 1, GFP_NOFS);
341 if (!ns)
342 goto mem_err_out;
343 }
344 for (i = o = 0; i < ins_len; i++) {
345retry: wc = nls->uni2char(le16_to_cpu(ins[i]), ns + o,
346 ns_len - o);
347 if (wc > 0) {
348 o += wc;
349 continue;
350 } else if (!wc)
351 break;
352 else if (wc == -ENAMETOOLONG && ns != *outs) {
353 unsigned char *tc;
354 /* Grow in multiples of 64 bytes. */
355 tc = (unsigned char*)kmalloc((ns_len + 64) &
356 ~63, GFP_NOFS);
357 if (tc) {
358 memcpy(tc, ns, ns_len);
359 ns_len = ((ns_len + 64) & ~63) - 1;
360 kfree(ns);
361 ns = tc;
362 goto retry;
363 } /* No memory so goto conversion_error; */
364 } /* wc < 0, real error. */
365 goto conversion_err;
366 }
367 ns[o] = 0;
368 *outs = ns;
369 return o;
370 } /* else (!ins) */
371 ntfs_error(vol->sb, "Received NULL pointer.");
372 return -EINVAL;
373conversion_err:
374 ntfs_error(vol->sb, "Unicode name contains characters that cannot be "
Anton Altaparmakovf94ad382005-09-08 17:04:11 +0100375 "converted to character set %s. You might want to "
376 "try to use the mount option nls=utf8.", nls->charset);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700377 if (ns != *outs)
378 kfree(ns);
379 if (wc != -ENAMETOOLONG)
380 wc = -EILSEQ;
381 return wc;
382mem_err_out:
383 ntfs_error(vol->sb, "Failed to allocate name!");
384 return -ENOMEM;
385}