1/*
2 * unistr.c - NTFS Unicode string handling. Part of the Linux-NTFS project.
3 *
4 * Copyright (c) 2001-2006 Anton Altaparmakov
5 *
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 <linux/slab.h>
23
24#include "types.h"
25#include "debug.h"
26#include "ntfs.h"
27
28/*
29 * IMPORTANT
30 * =========
31 *
32 * All these routines assume that the Unicode characters are in little endian
33 * encoding inside the strings!!!
34 */
35
36/*
37 * This is used by the name collation functions to quickly determine what
38 * characters are (in)valid.
39 */
40static const u8 legal_ansi_char_array[0x40] = {
41	0x00, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
42	0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
43
44	0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
45	0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10, 0x10,
46
47	0x17, 0x07, 0x18, 0x17, 0x17, 0x17, 0x17, 0x17,
48	0x17, 0x17, 0x18, 0x16, 0x16, 0x17, 0x07, 0x00,
49
50	0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17, 0x17,
51	0x17, 0x17, 0x04, 0x16, 0x18, 0x16, 0x18, 0x18,
52};
53
54/**
55 * ntfs_are_names_equal - compare two Unicode names for equality
56 * @s1:			name to compare to @s2
57 * @s1_len:		length in Unicode characters of @s1
58 * @s2:			name to compare to @s1
59 * @s2_len:		length in Unicode characters of @s2
60 * @ic:			ignore case bool
61 * @upcase:		upcase table (only if @ic == IGNORE_CASE)
62 * @upcase_size:	length in Unicode characters of @upcase (if present)
63 *
64 * Compare the names @s1 and @s2 and return 'true' (1) if the names are
65 * identical, or 'false' (0) if they are not identical. If @ic is IGNORE_CASE,
66 * the @upcase table is used to performa a case insensitive comparison.
67 */
68bool ntfs_are_names_equal(const ntfschar *s1, size_t s1_len,
69		const ntfschar *s2, size_t s2_len, const IGNORE_CASE_BOOL ic,
70		const ntfschar *upcase, const u32 upcase_size)
71{
72	if (s1_len != s2_len)
73		return false;
74	if (ic == CASE_SENSITIVE)
75		return !ntfs_ucsncmp(s1, s2, s1_len);
76	return !ntfs_ucsncasecmp(s1, s2, s1_len, upcase, upcase_size);
77}
78
79/**
80 * ntfs_collate_names - collate two Unicode names
81 * @name1:	first Unicode name to compare
82 * @name2:	second Unicode name to compare
83 * @err_val:	if @name1 contains an invalid character return this value
84 * @ic:		either CASE_SENSITIVE or IGNORE_CASE
85 * @upcase:	upcase table (ignored if @ic is CASE_SENSITIVE)
86 * @upcase_len:	upcase table size (ignored if @ic is CASE_SENSITIVE)
87 *
88 * ntfs_collate_names collates two Unicode names and returns:
89 *
90 *  -1 if the first name collates before the second one,
91 *   0 if the names match,
92 *   1 if the second name collates before the first one, or
93 * @err_val if an invalid character is found in @name1 during the comparison.
94 *
95 * The following characters are considered invalid: '"', '*', '<', '>' and '?'.
96 */
97int ntfs_collate_names(const ntfschar *name1, const u32 name1_len,
98		const ntfschar *name2, const u32 name2_len,
99		const int err_val, const IGNORE_CASE_BOOL ic,
100		const ntfschar *upcase, const u32 upcase_len)
101{
102	u32 cnt, min_len;
103	u16 c1, c2;
104
105	min_len = name1_len;
106	if (name1_len > name2_len)
107		min_len = name2_len;
108	for (cnt = 0; cnt < min_len; ++cnt) {
109		c1 = le16_to_cpu(*name1++);
110		c2 = le16_to_cpu(*name2++);
111		if (ic) {
112			if (c1 < upcase_len)
113				c1 = le16_to_cpu(upcase[c1]);
114			if (c2 < upcase_len)
115				c2 = le16_to_cpu(upcase[c2]);
116		}
117		if (c1 < 64 && legal_ansi_char_array[c1] & 8)
118			return err_val;
119		if (c1 < c2)
120			return -1;
121		if (c1 > c2)
122			return 1;
123	}
124	if (name1_len < name2_len)
125		return -1;
126	if (name1_len == name2_len)
127		return 0;
128	/* name1_len > name2_len */
129	c1 = le16_to_cpu(*name1);
130	if (c1 < 64 && legal_ansi_char_array[c1] & 8)
131		return err_val;
132	return 1;
133}
134
135/**
136 * ntfs_ucsncmp - compare two little endian Unicode strings
137 * @s1:		first string
138 * @s2:		second string
139 * @n:		maximum unicode characters to compare
140 *
141 * Compare the first @n characters of the Unicode strings @s1 and @s2,
142 * The strings in little endian format and appropriate le16_to_cpu()
143 * conversion is performed on non-little endian machines.
144 *
145 * The function returns an integer less than, equal to, or greater than zero
146 * if @s1 (or the first @n Unicode characters thereof) is found, respectively,
147 * to be less than, to match, or be greater than @s2.
148 */
149int ntfs_ucsncmp(const ntfschar *s1, const ntfschar *s2, size_t n)
150{
151	u16 c1, c2;
152	size_t i;
153
154	for (i = 0; i < n; ++i) {
155		c1 = le16_to_cpu(s1[i]);
156		c2 = le16_to_cpu(s2[i]);
157		if (c1 < c2)
158			return -1;
159		if (c1 > c2)
160			return 1;
161		if (!c1)
162			break;
163	}
164	return 0;
165}
166
167/**
168 * ntfs_ucsncasecmp - compare two little endian Unicode strings, ignoring case
169 * @s1:			first string
170 * @s2:			second string
171 * @n:			maximum unicode characters to compare
172 * @upcase:		upcase table
173 * @upcase_size:	upcase table size in Unicode characters
174 *
175 * Compare the first @n characters of the Unicode strings @s1 and @s2,
176 * ignoring case. The strings in little endian format and appropriate
177 * le16_to_cpu() conversion is performed on non-little endian machines.
178 *
179 * Each character is uppercased using the @upcase table before the comparison.
180 *
181 * The function returns an integer less than, equal to, or greater than zero
182 * if @s1 (or the first @n Unicode characters thereof) is found, respectively,
183 * to be less than, to match, or be greater than @s2.
184 */
185int ntfs_ucsncasecmp(const ntfschar *s1, const ntfschar *s2, size_t n,
186		const ntfschar *upcase, const u32 upcase_size)
187{
188	size_t i;
189	u16 c1, c2;
190
191	for (i = 0; i < n; ++i) {
192		if ((c1 = le16_to_cpu(s1[i])) < upcase_size)
193			c1 = le16_to_cpu(upcase[c1]);
194		if ((c2 = le16_to_cpu(s2[i])) < upcase_size)
195			c2 = le16_to_cpu(upcase[c2]);
196		if (c1 < c2)
197			return -1;
198		if (c1 > c2)
199			return 1;
200		if (!c1)
201			break;
202	}
203	return 0;
204}
205
206void ntfs_upcase_name(ntfschar *name, u32 name_len, const ntfschar *upcase,
207		const u32 upcase_len)
208{
209	u32 i;
210	u16 u;
211
212	for (i = 0; i < name_len; i++)
213		if ((u = le16_to_cpu(name[i])) < upcase_len)
214			name[i] = upcase[u];
215}
216
217void ntfs_file_upcase_value(FILE_NAME_ATTR *file_name_attr,
218		const ntfschar *upcase, const u32 upcase_len)
219{
220	ntfs_upcase_name((ntfschar*)&file_name_attr->file_name,
221			file_name_attr->file_name_length, upcase, upcase_len);
222}
223
224int ntfs_file_compare_values(FILE_NAME_ATTR *file_name_attr1,
225		FILE_NAME_ATTR *file_name_attr2,
226		const int err_val, const IGNORE_CASE_BOOL ic,
227		const ntfschar *upcase, const u32 upcase_len)
228{
229	return ntfs_collate_names((ntfschar*)&file_name_attr1->file_name,
230			file_name_attr1->file_name_length,
231			(ntfschar*)&file_name_attr2->file_name,
232			file_name_attr2->file_name_length,
233			err_val, ic, upcase, upcase_len);
234}
235
236/**
237 * ntfs_nlstoucs - convert NLS string to little endian Unicode string
238 * @vol:	ntfs volume which we are working with
239 * @ins:	input NLS string buffer
240 * @ins_len:	length of input string in bytes
241 * @outs:	on return contains the allocated output Unicode string buffer
242 *
243 * Convert the input string @ins, which is in whatever format the loaded NLS
244 * map dictates, into a little endian, 2-byte Unicode string.
245 *
246 * This function allocates the string and the caller is responsible for
247 * calling kmem_cache_free(ntfs_name_cache, *@outs); when finished with it.
248 *
249 * On success the function returns the number of Unicode characters written to
250 * the output string *@outs (>= 0), not counting the terminating Unicode NULL
251 * character. *@outs is set to the allocated output string buffer.
252 *
253 * On error, a negative number corresponding to the error code is returned. In
254 * that case the output string is not allocated. Both *@outs and *@outs_len
255 * are then undefined.
256 *
257 * This might look a bit odd due to fast path optimization...
258 */
259int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
260		const int ins_len, ntfschar **outs)
261{
262	struct nls_table *nls = vol->nls_map;
263	ntfschar *ucs;
264	wchar_t wc;
265	int i, o, wc_len;
266
267	/* We do not trust outside sources. */
268	if (likely(ins)) {
269		ucs = kmem_cache_alloc(ntfs_name_cache, GFP_NOFS);
270		if (likely(ucs)) {
271			for (i = o = 0; i < ins_len; i += wc_len) {
272				wc_len = nls->char2uni(ins + i, ins_len - i,
273						&wc);
274				if (likely(wc_len >= 0 &&
275						o < NTFS_MAX_NAME_LEN)) {
276					if (likely(wc)) {
277						ucs[o++] = cpu_to_le16(wc);
278						continue;
279					} /* else if (!wc) */
280					break;
281				} /* else if (wc_len < 0 ||
282						o >= NTFS_MAX_NAME_LEN) */
283				goto name_err;
284			}
285			ucs[o] = 0;
286			*outs = ucs;
287			return o;
288		} /* else if (!ucs) */
289		ntfs_error(vol->sb, "Failed to allocate buffer for converted "
290				"name from ntfs_name_cache.");
291		return -ENOMEM;
292	} /* else if (!ins) */
293	ntfs_error(vol->sb, "Received NULL pointer.");
294	return -EINVAL;
295name_err:
296	kmem_cache_free(ntfs_name_cache, ucs);
297	if (wc_len < 0) {
298		ntfs_error(vol->sb, "Name using character set %s contains "
299				"characters that cannot be converted to "
300				"Unicode.", nls->charset);
301		i = -EILSEQ;
302	} else /* if (o >= NTFS_MAX_NAME_LEN) */ {
303		ntfs_error(vol->sb, "Name is too long (maximum length for a "
304				"name on NTFS is %d Unicode characters.",
305				NTFS_MAX_NAME_LEN);
306		i = -ENAMETOOLONG;
307	}
308	return i;
309}
310
311/**
312 * ntfs_ucstonls - convert little endian Unicode string to NLS string
313 * @vol:	ntfs volume which we are working with
314 * @ins:	input Unicode string buffer
315 * @ins_len:	length of input string in Unicode characters
316 * @outs:	on return contains the (allocated) output NLS string buffer
317 * @outs_len:	length of output string buffer in bytes
318 *
319 * Convert the input little endian, 2-byte Unicode string @ins, of length
320 * @ins_len into the string format dictated by the loaded NLS.
321 *
322 * If *@outs is NULL, this function allocates the string and the caller is
323 * responsible for calling kfree(*@outs); when finished with it. In this case
324 * @outs_len is ignored and can be 0.
325 *
326 * On success the function returns the number of bytes written to the output
327 * string *@outs (>= 0), not counting the terminating NULL byte. If the output
328 * string buffer was allocated, *@outs is set to it.
329 *
330 * On error, a negative number corresponding to the error code is returned. In
331 * that case the output string is not allocated. The contents of *@outs are
332 * then undefined.
333 *
334 * This might look a bit odd due to fast path optimization...
335 */
336int ntfs_ucstonls(const ntfs_volume *vol, const ntfschar *ins,
337		const int ins_len, unsigned char **outs, int outs_len)
338{
339	struct nls_table *nls = vol->nls_map;
340	unsigned char *ns;
341	int i, o, ns_len, wc;
342
343	/* We don't trust outside sources. */
344	if (ins) {
345		ns = *outs;
346		ns_len = outs_len;
347		if (ns && !ns_len) {
348			wc = -ENAMETOOLONG;
349			goto conversion_err;
350		}
351		if (!ns) {
352			ns_len = ins_len * NLS_MAX_CHARSET_SIZE;
353			ns = kmalloc(ns_len + 1, GFP_NOFS);
354			if (!ns)
355				goto mem_err_out;
356		}
357		for (i = o = 0; i < ins_len; i++) {
358retry:			wc = nls->uni2char(le16_to_cpu(ins[i]), ns + o,
359					ns_len - o);
360			if (wc > 0) {
361				o += wc;
362				continue;
363			} else if (!wc)
364				break;
365			else if (wc == -ENAMETOOLONG && ns != *outs) {
366				unsigned char *tc;
367				/* Grow in multiples of 64 bytes. */
368				tc = kmalloc((ns_len + 64) &
369						~63, GFP_NOFS);
370				if (tc) {
371					memcpy(tc, ns, ns_len);
372					ns_len = ((ns_len + 64) & ~63) - 1;
373					kfree(ns);
374					ns = tc;
375					goto retry;
376				} /* No memory so goto conversion_error; */
377			} /* wc < 0, real error. */
378			goto conversion_err;
379		}
380		ns[o] = 0;
381		*outs = ns;
382		return o;
383	} /* else (!ins) */
384	ntfs_error(vol->sb, "Received NULL pointer.");
385	return -EINVAL;
386conversion_err:
387	ntfs_error(vol->sb, "Unicode name contains characters that cannot be "
388			"converted to character set %s.  You might want to "
389			"try to use the mount option nls=utf8.", nls->charset);
390	if (ns != *outs)
391		kfree(ns);
392	if (wc != -ENAMETOOLONG)
393		wc = -EILSEQ;
394	return wc;
395mem_err_out:
396	ntfs_error(vol->sb, "Failed to allocate name!");
397	return -ENOMEM;
398}
399