1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /*
3  *   Some of the source code in this file came from fs/cifs/cifs_unicode.c
4  *
5  *   Copyright (c) International Business Machines  Corp., 2000,2009
6  *   Modified by Steve French (sfrench@us.ibm.com)
7  *   Modified by Namjae Jeon (linkinjeon@kernel.org)
8  */
9 #include <linux/fs.h>
10 #include <linux/slab.h>
11 #include <asm/unaligned.h>
12 #include "glob.h"
13 #include "unicode.h"
14 #include "uniupr.h"
15 #include "smb_common.h"
16 
17 /*
18  * smb_utf16_bytes() - how long will a string be after conversion?
19  * @from:	pointer to input string
20  * @maxbytes:	don't go past this many bytes of input string
21  * @codepage:	destination codepage
22  *
23  * Walk a utf16le string and return the number of bytes that the string will
24  * be after being converted to the given charset, not including any null
25  * termination required. Don't walk past maxbytes in the source buffer.
26  *
27  * Return:	string length after conversion
28  */
smb_utf16_bytes(const __le16 * from,int maxbytes,const struct nls_table * codepage)29 static int smb_utf16_bytes(const __le16 *from, int maxbytes,
30 			   const struct nls_table *codepage)
31 {
32 	int i;
33 	int charlen, outlen = 0;
34 	int maxwords = maxbytes / 2;
35 	char tmp[NLS_MAX_CHARSET_SIZE];
36 	__u16 ftmp;
37 
38 	for (i = 0; i < maxwords; i++) {
39 		ftmp = get_unaligned_le16(&from[i]);
40 		if (ftmp == 0)
41 			break;
42 
43 		charlen = codepage->uni2char(ftmp, tmp, NLS_MAX_CHARSET_SIZE);
44 		if (charlen > 0)
45 			outlen += charlen;
46 		else
47 			outlen++;
48 	}
49 
50 	return outlen;
51 }
52 
53 /*
54  * cifs_mapchar() - convert a host-endian char to proper char in codepage
55  * @target:	where converted character should be copied
56  * @src_char:	2 byte host-endian source character
57  * @cp:		codepage to which character should be converted
58  * @mapchar:	should character be mapped according to mapchars mount option?
59  *
60  * This function handles the conversion of a single character. It is the
61  * responsibility of the caller to ensure that the target buffer is large
62  * enough to hold the result of the conversion (at least NLS_MAX_CHARSET_SIZE).
63  *
64  * Return:	string length after conversion
65  */
66 static int
cifs_mapchar(char * target,const __u16 src_char,const struct nls_table * cp,bool mapchar)67 cifs_mapchar(char *target, const __u16 src_char, const struct nls_table *cp,
68 	     bool mapchar)
69 {
70 	int len = 1;
71 
72 	if (!mapchar)
73 		goto cp_convert;
74 
75 	/*
76 	 * BB: Cannot handle remapping UNI_SLASH until all the calls to
77 	 *     build_path_from_dentry are modified, as they use slash as
78 	 *     separator.
79 	 */
80 	switch (src_char) {
81 	case UNI_COLON:
82 		*target = ':';
83 		break;
84 	case UNI_ASTERISK:
85 		*target = '*';
86 		break;
87 	case UNI_QUESTION:
88 		*target = '?';
89 		break;
90 	case UNI_PIPE:
91 		*target = '|';
92 		break;
93 	case UNI_GRTRTHAN:
94 		*target = '>';
95 		break;
96 	case UNI_LESSTHAN:
97 		*target = '<';
98 		break;
99 	default:
100 		goto cp_convert;
101 	}
102 
103 out:
104 	return len;
105 
106 cp_convert:
107 	len = cp->uni2char(src_char, target, NLS_MAX_CHARSET_SIZE);
108 	if (len <= 0) {
109 		*target = '?';
110 		len = 1;
111 	}
112 
113 	goto out;
114 }
115 
116 /*
117  * is_char_allowed() - check for valid character
118  * @ch:		input character to be checked
119  *
120  * Return:	1 if char is allowed, otherwise 0
121  */
is_char_allowed(char * ch)122 static inline int is_char_allowed(char *ch)
123 {
124 	/* check for control chars, wildcards etc. */
125 	if (!(*ch & 0x80) &&
126 	    (*ch <= 0x1f ||
127 	     *ch == '?' || *ch == '"' || *ch == '<' ||
128 	     *ch == '>' || *ch == '|'))
129 		return 0;
130 
131 	return 1;
132 }
133 
134 /*
135  * smb_from_utf16() - convert utf16le string to local charset
136  * @to:		destination buffer
137  * @from:	source buffer
138  * @tolen:	destination buffer size (in bytes)
139  * @fromlen:	source buffer size (in bytes)
140  * @codepage:	codepage to which characters should be converted
141  * @mapchar:	should characters be remapped according to the mapchars option?
142  *
143  * Convert a little-endian utf16le string (as sent by the server) to a string
144  * in the provided codepage. The tolen and fromlen parameters are to ensure
145  * that the code doesn't walk off of the end of the buffer (which is always
146  * a danger if the alignment of the source buffer is off). The destination
147  * string is always properly null terminated and fits in the destination
148  * buffer. Returns the length of the destination string in bytes (including
149  * null terminator).
150  *
151  * Note that some windows versions actually send multiword UTF-16 characters
152  * instead of straight UTF16-2. The linux nls routines however aren't able to
153  * deal with those characters properly. In the event that we get some of
154  * those characters, they won't be translated properly.
155  *
156  * Return:	string length after conversion
157  */
smb_from_utf16(char * to,const __le16 * from,int tolen,int fromlen,const struct nls_table * codepage,bool mapchar)158 static int smb_from_utf16(char *to, const __le16 *from, int tolen, int fromlen,
159 			  const struct nls_table *codepage, bool mapchar)
160 {
161 	int i, charlen, safelen;
162 	int outlen = 0;
163 	int nullsize = nls_nullsize(codepage);
164 	int fromwords = fromlen / 2;
165 	char tmp[NLS_MAX_CHARSET_SIZE];
166 	__u16 ftmp;
167 
168 	/*
169 	 * because the chars can be of varying widths, we need to take care
170 	 * not to overflow the destination buffer when we get close to the
171 	 * end of it. Until we get to this offset, we don't need to check
172 	 * for overflow however.
173 	 */
174 	safelen = tolen - (NLS_MAX_CHARSET_SIZE + nullsize);
175 
176 	for (i = 0; i < fromwords; i++) {
177 		ftmp = get_unaligned_le16(&from[i]);
178 		if (ftmp == 0)
179 			break;
180 
181 		/*
182 		 * check to see if converting this character might make the
183 		 * conversion bleed into the null terminator
184 		 */
185 		if (outlen >= safelen) {
186 			charlen = cifs_mapchar(tmp, ftmp, codepage, mapchar);
187 			if ((outlen + charlen) > (tolen - nullsize))
188 				break;
189 		}
190 
191 		/* put converted char into 'to' buffer */
192 		charlen = cifs_mapchar(&to[outlen], ftmp, codepage, mapchar);
193 		outlen += charlen;
194 	}
195 
196 	/* properly null-terminate string */
197 	for (i = 0; i < nullsize; i++)
198 		to[outlen++] = 0;
199 
200 	return outlen;
201 }
202 
203 /*
204  * smb_strtoUTF16() - Convert character string to unicode string
205  * @to:		destination buffer
206  * @from:	source buffer
207  * @len:	destination buffer size (in bytes)
208  * @codepage:	codepage to which characters should be converted
209  *
210  * Return:	string length after conversion
211  */
smb_strtoUTF16(__le16 * to,const char * from,int len,const struct nls_table * codepage)212 int smb_strtoUTF16(__le16 *to, const char *from, int len,
213 		   const struct nls_table *codepage)
214 {
215 	int charlen;
216 	int i;
217 	wchar_t wchar_to; /* needed to quiet sparse */
218 
219 	/* special case for utf8 to handle no plane0 chars */
220 	if (!strcmp(codepage->charset, "utf8")) {
221 		/*
222 		 * convert utf8 -> utf16, we assume we have enough space
223 		 * as caller should have assumed conversion does not overflow
224 		 * in destination len is length in wchar_t units (16bits)
225 		 */
226 		i  = utf8s_to_utf16s(from, len, UTF16_LITTLE_ENDIAN,
227 				     (wchar_t *)to, len);
228 
229 		/* if success terminate and exit */
230 		if (i >= 0)
231 			goto success;
232 		/*
233 		 * if fails fall back to UCS encoding as this
234 		 * function should not return negative values
235 		 * currently can fail only if source contains
236 		 * invalid encoded characters
237 		 */
238 	}
239 
240 	for (i = 0; len > 0 && *from; i++, from += charlen, len -= charlen) {
241 		charlen = codepage->char2uni(from, len, &wchar_to);
242 		if (charlen < 1) {
243 			/* A question mark */
244 			wchar_to = 0x003f;
245 			charlen = 1;
246 		}
247 		put_unaligned_le16(wchar_to, &to[i]);
248 	}
249 
250 success:
251 	put_unaligned_le16(0, &to[i]);
252 	return i;
253 }
254 
255 /*
256  * smb_strndup_from_utf16() - copy a string from wire format to the local
257  *		codepage
258  * @src:	source string
259  * @maxlen:	don't walk past this many bytes in the source string
260  * @is_unicode:	is this a unicode string?
261  * @codepage:	destination codepage
262  *
263  * Take a string given by the server, convert it to the local codepage and
264  * put it in a new buffer. Returns a pointer to the new string or NULL on
265  * error.
266  *
267  * Return:	destination string buffer or error ptr
268  */
smb_strndup_from_utf16(const char * src,const int maxlen,const bool is_unicode,const struct nls_table * codepage)269 char *smb_strndup_from_utf16(const char *src, const int maxlen,
270 			     const bool is_unicode,
271 			     const struct nls_table *codepage)
272 {
273 	int len, ret;
274 	char *dst;
275 
276 	if (is_unicode) {
277 		len = smb_utf16_bytes((__le16 *)src, maxlen, codepage);
278 		len += nls_nullsize(codepage);
279 		dst = kmalloc(len, GFP_KERNEL);
280 		if (!dst)
281 			return ERR_PTR(-ENOMEM);
282 		ret = smb_from_utf16(dst, (__le16 *)src, len, maxlen, codepage,
283 				     false);
284 		if (ret < 0) {
285 			kfree(dst);
286 			return ERR_PTR(-EINVAL);
287 		}
288 	} else {
289 		len = strnlen(src, maxlen);
290 		len++;
291 		dst = kmalloc(len, GFP_KERNEL);
292 		if (!dst)
293 			return ERR_PTR(-ENOMEM);
294 		strscpy(dst, src, len);
295 	}
296 
297 	return dst;
298 }
299 
300 /*
301  * Convert 16 bit Unicode pathname to wire format from string in current code
302  * page. Conversion may involve remapping up the six characters that are
303  * only legal in POSIX-like OS (if they are present in the string). Path
304  * names are little endian 16 bit Unicode on the wire
305  */
306 /*
307  * smbConvertToUTF16() - convert string from local charset to utf16
308  * @target:	destination buffer
309  * @source:	source buffer
310  * @srclen:	source buffer size (in bytes)
311  * @cp:		codepage to which characters should be converted
312  * @mapchar:	should characters be remapped according to the mapchars option?
313  *
314  * Convert 16 bit Unicode pathname to wire format from string in current code
315  * page. Conversion may involve remapping up the six characters that are
316  * only legal in POSIX-like OS (if they are present in the string). Path
317  * names are little endian 16 bit Unicode on the wire
318  *
319  * Return:	char length after conversion
320  */
smbConvertToUTF16(__le16 * target,const char * source,int srclen,const struct nls_table * cp,int mapchars)321 int smbConvertToUTF16(__le16 *target, const char *source, int srclen,
322 		      const struct nls_table *cp, int mapchars)
323 {
324 	int i, j, charlen;
325 	char src_char;
326 	__le16 dst_char;
327 	wchar_t tmp;
328 
329 	if (!mapchars)
330 		return smb_strtoUTF16(target, source, srclen, cp);
331 
332 	for (i = 0, j = 0; i < srclen; j++) {
333 		src_char = source[i];
334 		charlen = 1;
335 		switch (src_char) {
336 		case 0:
337 			put_unaligned(0, &target[j]);
338 			return j;
339 		case ':':
340 			dst_char = cpu_to_le16(UNI_COLON);
341 			break;
342 		case '*':
343 			dst_char = cpu_to_le16(UNI_ASTERISK);
344 			break;
345 		case '?':
346 			dst_char = cpu_to_le16(UNI_QUESTION);
347 			break;
348 		case '<':
349 			dst_char = cpu_to_le16(UNI_LESSTHAN);
350 			break;
351 		case '>':
352 			dst_char = cpu_to_le16(UNI_GRTRTHAN);
353 			break;
354 		case '|':
355 			dst_char = cpu_to_le16(UNI_PIPE);
356 			break;
357 		/*
358 		 * FIXME: We can not handle remapping backslash (UNI_SLASH)
359 		 * until all the calls to build_path_from_dentry are modified,
360 		 * as they use backslash as separator.
361 		 */
362 		default:
363 			charlen = cp->char2uni(source + i, srclen - i, &tmp);
364 			dst_char = cpu_to_le16(tmp);
365 
366 			/*
367 			 * if no match, use question mark, which at least in
368 			 * some cases serves as wild card
369 			 */
370 			if (charlen < 1) {
371 				dst_char = cpu_to_le16(0x003f);
372 				charlen = 1;
373 			}
374 		}
375 		/*
376 		 * character may take more than one byte in the source string,
377 		 * but will take exactly two bytes in the target string
378 		 */
379 		i += charlen;
380 		put_unaligned(dst_char, &target[j]);
381 	}
382 
383 	return j;
384 }
385