1 /* $FreeBSD: releng/12.0/sys/fs/msdosfs/msdosfs_conv.c 336966 2018-07-31 12:44:28Z emaste $ */
2 /* $NetBSD: msdosfs_conv.c,v 1.25 1997/11/17 15:36:40 ws Exp $ */
3
4 /*-
5 * SPDX-License-Identifier: BSD-4-Clause
6 *
7 * Copyright (C) 1995, 1997 Wolfgang Solfrank.
8 * Copyright (C) 1995, 1997 TooLs GmbH.
9 * All rights reserved.
10 * Original code by Paul Popelka (paulp@uts.amdahl.com) (see below).
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 * notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in the
19 * documentation and/or other materials provided with the distribution.
20 * 3. All advertising materials mentioning features or use of this software
21 * must display the following acknowledgement:
22 * This product includes software developed by TooLs GmbH.
23 * 4. The name of TooLs GmbH may not be used to endorse or promote products
24 * derived from this software without specific prior written permission.
25 *
26 * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR
27 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
28 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
29 * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
30 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
31 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
32 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
33 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
34 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
35 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
36 */
37 /*-
38 * Written by Paul Popelka (paulp@uts.amdahl.com)
39 *
40 * You can do anything you want with this software, just don't say you wrote
41 * it, and don't remove this notice.
42 *
43 * This software is provided "as is".
44 *
45 * The author supplies this software to be publicly redistributed on the
46 * understanding that the author is not responsible for the correct
47 * functioning of this software in any circumstances and is not liable for
48 * any damages caused by this software.
49 *
50 * October 1992
51 */
52
53 #include <sys/param.h>
54 #include <sys/systm.h>
55 #include <sys/dirent.h>
56 #include <sys/iconv.h>
57 #include <sys/mount.h>
58
59 #include <fs/msdosfs/bpb.h>
60 #include <fs/msdosfs/direntry.h>
61 #include <fs/msdosfs/msdosfsmount.h>
62
63 extern struct iconv_functions *msdosfs_iconv;
64
65 static int mbsadjpos(const char **, size_t, size_t, int, int, void *handle);
66 static u_char * dos2unixchr(u_char *, const u_char **, size_t *, int, struct msdosfsmount *);
67 static uint16_t unix2doschr(const u_char **, size_t *, struct msdosfsmount *);
68 static u_char * win2unixchr(u_char *, uint16_t, struct msdosfsmount *);
69 static uint16_t unix2winchr(const u_char **, size_t *, int, struct msdosfsmount *);
70
71 /*
72 * 0 - character disallowed in long file name.
73 * 1 - character should be replaced by '_' in DOS file name,
74 * and generation number inserted.
75 * 2 - character ('.' and ' ') should be skipped in DOS file name,
76 * and generation number inserted.
77 */
78 static const u_char
79 unix2dos[256] = {
80 /* iso8859-1 -> cp850 */
81 0, 0, 0, 0, 0, 0, 0, 0, /* 00-07 */
82 0, 0, 0, 0, 0, 0, 0, 0, /* 08-0f */
83 0, 0, 0, 0, 0, 0, 0, 0, /* 10-17 */
84 0, 0, 0, 0, 0, 0, 0, 0, /* 18-1f */
85 2, 0x21, 0, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
86 0x28, 0x29, 0, 1, 1, 0x2d, 2, 0, /* 28-2f */
87 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
88 0x38, 0x39, 0, 1, 0, 1, 0, 0, /* 38-3f */
89 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */
90 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */
91 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */
92 0x58, 0x59, 0x5a, 1, 0, 1, 0x5e, 0x5f, /* 58-5f */
93 0x60, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 60-67 */
94 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 68-6f */
95 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 70-77 */
96 0x58, 0x59, 0x5a, 0x7b, 0, 0x7d, 0x7e, 0, /* 78-7f */
97 0, 0, 0, 0, 0, 0, 0, 0, /* 80-87 */
98 0, 0, 0, 0, 0, 0, 0, 0, /* 88-8f */
99 0, 0, 0, 0, 0, 0, 0, 0, /* 90-97 */
100 0, 0, 0, 0, 0, 0, 0, 0, /* 98-9f */
101 0, 0xad, 0xbd, 0x9c, 0xcf, 0xbe, 0xdd, 0xf5, /* a0-a7 */
102 0xf9, 0xb8, 0xa6, 0xae, 0xaa, 0xf0, 0xa9, 0xee, /* a8-af */
103 0xf8, 0xf1, 0xfd, 0xfc, 0xef, 0xe6, 0xf4, 0xfa, /* b0-b7 */
104 0xf7, 0xfb, 0xa7, 0xaf, 0xac, 0xab, 0xf3, 0xa8, /* b8-bf */
105 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* c0-c7 */
106 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* c8-cf */
107 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0x9e, /* d0-d7 */
108 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0xe1, /* d8-df */
109 0xb7, 0xb5, 0xb6, 0xc7, 0x8e, 0x8f, 0x92, 0x80, /* e0-e7 */
110 0xd4, 0x90, 0xd2, 0xd3, 0xde, 0xd6, 0xd7, 0xd8, /* e8-ef */
111 0xd1, 0xa5, 0xe3, 0xe0, 0xe2, 0xe5, 0x99, 0xf6, /* f0-f7 */
112 0x9d, 0xeb, 0xe9, 0xea, 0x9a, 0xed, 0xe8, 0x98, /* f8-ff */
113 };
114
115 static const u_char
116 dos2unix[256] = {
117 /* cp850 -> iso8859-1 */
118 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 00-07 */
119 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 08-0f */
120 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 10-17 */
121 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, /* 18-1f */
122 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
123 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */
124 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
125 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */
126 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, /* 40-47 */
127 0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f, /* 48-4f */
128 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, /* 50-57 */
129 0x58, 0x59, 0x5a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */
130 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */
131 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */
132 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */
133 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */
134 0xc7, 0xfc, 0xe9, 0xe2, 0xe4, 0xe0, 0xe5, 0xe7, /* 80-87 */
135 0xea, 0xeb, 0xe8, 0xef, 0xee, 0xec, 0xc4, 0xc5, /* 88-8f */
136 0xc9, 0xe6, 0xc6, 0xf4, 0xf6, 0xf2, 0xfb, 0xf9, /* 90-97 */
137 0xff, 0xd6, 0xdc, 0xf8, 0xa3, 0xd8, 0xd7, 0x3f, /* 98-9f */
138 0xe1, 0xed, 0xf3, 0xfa, 0xf1, 0xd1, 0xaa, 0xba, /* a0-a7 */
139 0xbf, 0xae, 0xac, 0xbd, 0xbc, 0xa1, 0xab, 0xbb, /* a8-af */
140 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xc1, 0xc2, 0xc0, /* b0-b7 */
141 0xa9, 0x3f, 0x3f, 0x3f, 0x3f, 0xa2, 0xa5, 0x3f, /* b8-bf */
142 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xe3, 0xc3, /* c0-c7 */
143 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0x3f, 0xa4, /* c8-cf */
144 0xf0, 0xd0, 0xca, 0xcb, 0xc8, 0x3f, 0xcd, 0xce, /* d0-d7 */
145 0xcf, 0x3f, 0x3f, 0x3f, 0x3f, 0xa6, 0xcc, 0x3f, /* d8-df */
146 0xd3, 0xdf, 0xd4, 0xd2, 0xf5, 0xd5, 0xb5, 0xfe, /* e0-e7 */
147 0xde, 0xda, 0xdb, 0xd9, 0xfd, 0xdd, 0xaf, 0x3f, /* e8-ef */
148 0xad, 0xb1, 0x3f, 0xbe, 0xb6, 0xa7, 0xf7, 0xb8, /* f0-f7 */
149 0xb0, 0xa8, 0xb7, 0xb9, 0xb3, 0xb2, 0x3f, 0x3f, /* f8-ff */
150 };
151
152 static const u_char
153 u2l[256] = {
154 /* tolower */
155 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */
156 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */
157 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */
158 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */
159 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
160 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */
161 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
162 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */
163 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */
164 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */
165 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */
166 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */
167 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */
168 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */
169 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */
170 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */
171 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */
172 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */
173 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */
174 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */
175 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */
176 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */
177 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */
178 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */
179 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */
180 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */
181 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */
182 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */
183 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */
184 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */
185 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */
186 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */
187 };
188
189 static const u_char
190 l2u[256] = {
191 /* toupper */
192 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, /* 00-07 */
193 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, /* 08-0f */
194 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, /* 10-17 */
195 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, /* 18-1f */
196 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, /* 20-27 */
197 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, /* 28-2f */
198 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, /* 30-37 */
199 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0x3e, 0x3f, /* 38-3f */
200 0x40, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 40-47 */
201 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 48-4f */
202 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 50-57 */
203 0x78, 0x79, 0x7a, 0x5b, 0x5c, 0x5d, 0x5e, 0x5f, /* 58-5f */
204 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, /* 60-67 */
205 0x68, 0x69, 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, /* 68-6f */
206 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, /* 70-77 */
207 0x78, 0x79, 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, /* 78-7f */
208 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 80-87 */
209 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, /* 88-8f */
210 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 90-97 */
211 0x98, 0x99, 0x9a, 0x9b, 0x9c, 0x9d, 0x9e, 0x9f, /* 98-9f */
212 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, /* a0-a7 */
213 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, /* a8-af */
214 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7, /* b0-b7 */
215 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf, /* b8-bf */
216 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* c0-c7 */
217 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* c8-cf */
218 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xd7, /* d0-d7 */
219 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xdf, /* d8-df */
220 0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* e0-e7 */
221 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* e8-ef */
222 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* f0-f7 */
223 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* f8-ff */
224 };
225
226 /*
227 * DOS filenames are made of 2 parts, the name part and the extension part.
228 * The name part is 8 characters long and the extension part is 3
229 * characters long. They may contain trailing blanks if the name or
230 * extension are not long enough to fill their respective fields.
231 */
232
233 /*
234 * Convert a DOS filename to a unix filename. And, return the number of
235 * characters in the resulting unix filename excluding the terminating
236 * null.
237 */
238 int
239 dos2unixfn(u_char dn[11], u_char *un, int lower, struct msdosfsmount *pmp)
240 {
241 size_t i;
242 int thislong = 0;
243 u_char *c, tmpbuf[5];
244
245 /*
246 * If first char of the filename is SLOT_E5 (0x05), then the real
247 * first char of the filename should be 0xe5. But, they couldn't
248 * just have a 0xe5 mean 0xe5 because that is used to mean a freed
249 * directory slot. Another dos quirk.
250 */
251 if (*dn == SLOT_E5)
252 *dn = 0xe5;
253
254 /*
255 * Copy the name portion into the unix filename string.
256 */
257 for (i = 8; i > 0 && *dn != ' ';) {
258 c = dos2unixchr(tmpbuf, __DECONST(const u_char **, &dn), &i,
259 lower & LCASE_BASE, pmp);
260 while (*c != '\0') {
261 *un++ = *c++;
262 thislong++;
263 }
264 }
265 dn += i;
266
267 /*
268 * Now, if there is an extension then put in a period and copy in
269 * the extension.
270 */
271 if (*dn != ' ') {
272 *un++ = '.';
273 thislong++;
274 for (i = 3; i > 0 && *dn != ' ';) {
275 c = dos2unixchr(tmpbuf, __DECONST(const u_char **, &dn),
276 &i, lower & LCASE_EXT, pmp);
277 while (*c != '\0') {
278 *un++ = *c++;
279 thislong++;
280 }
281 }
282 }
283 *un++ = 0;
284
285 return (thislong);
286 }
287
288 /*
289 * Convert a unix filename to a DOS filename according to Win95 rules.
290 * If applicable and gen is not 0, it is inserted into the converted
291 * filename as a generation number.
292 * Returns
293 * 0 if name couldn't be converted
294 * 1 if the converted name is the same as the original
295 * (no long filename entry necessary for Win95)
296 * 2 if conversion was successful
297 * 3 if conversion was successful and generation number was inserted
298 */
299 int
300 unix2dosfn(const u_char *un, u_char dn[12], size_t unlen, u_int gen,
301 struct msdosfsmount *pmp)
302 {
303 ssize_t i, j;
304 int l;
305 int conv = 1;
306 const u_char *cp, *dp, *dp1;
307 u_char gentext[6], *wcp;
308 uint16_t c;
309
310 /*
311 * Fill the dos filename string with blanks. These are DOS's pad
312 * characters.
313 */
314 for (i = 0; i < 11; i++)
315 dn[i] = ' ';
316 dn[11] = 0;
317
318 /*
319 * The filenames "." and ".." are handled specially, since they
320 * don't follow dos filename rules.
321 */
322 if (un[0] == '.' && unlen == 1) {
323 dn[0] = '.';
324 return gen <= 1;
325 }
326 if (un[0] == '.' && un[1] == '.' && unlen == 2) {
327 dn[0] = '.';
328 dn[1] = '.';
329 return gen <= 1;
330 }
331
332 /*
333 * Filenames with only blanks and dots are not allowed!
334 */
335 for (cp = un, i = unlen; --i >= 0; cp++)
336 if (*cp != ' ' && *cp != '.')
337 break;
338 if (i < 0)
339 return 0;
340
341
342 /*
343 * Filenames with some characters are not allowed!
344 */
345 for (cp = un, i = unlen; i > 0;)
346 if (unix2doschr(&cp, (size_t *)&i, pmp) == 0)
347 return 0;
348
349 /*
350 * Now find the extension
351 * Note: dot as first char doesn't start extension
352 * and trailing dots and blanks are ignored
353 * Note(2003/7): It seems recent Windows has
354 * defferent rule than this code, that Windows
355 * ignores all dots before extension, and use all
356 * chars as filename except for dots.
357 */
358 dp = dp1 = NULL;
359 for (cp = un + 1, i = unlen - 1; --i >= 0;) {
360 switch (*cp++) {
361 case '.':
362 if (!dp1)
363 dp1 = cp;
364 break;
365 case ' ':
366 break;
367 default:
368 if (dp1)
369 dp = dp1;
370 dp1 = NULL;
371 break;
372 }
373 }
374
375 /*
376 * Now convert it (this part is for extension).
377 * As Windows XP do, if it's not ascii char,
378 * this function should return 2 or 3, so that checkng out Unicode name.
379 */
380 if (dp) {
381 if (dp1)
382 l = dp1 - dp;
383 else
384 l = unlen - (dp - un);
385 for (cp = dp, i = l, j = 8; i > 0 && j < 11; j++) {
386 c = unix2doschr(&cp, (size_t *)&i, pmp);
387 if (c & 0xff00) {
388 dn[j] = c >> 8;
389 if (++j < 11) {
390 dn[j] = c;
391 if (conv != 3)
392 conv = 2;
393 continue;
394 } else {
395 conv = 3;
396 dn[j-1] = ' ';
397 break;
398 }
399 } else {
400 dn[j] = c;
401 }
402 if (((dn[j] & 0x80) || *(cp - 1) != dn[j]) && conv != 3)
403 conv = 2;
404 if (dn[j] == 1) {
405 conv = 3;
406 dn[j] = '_';
407 }
408 if (dn[j] == 2) {
409 conv = 3;
410 dn[j--] = ' ';
411 }
412 }
413 if (i > 0)
414 conv = 3;
415 dp--;
416 } else {
417 for (dp = cp; *--dp == ' ' || *dp == '.';);
418 dp++;
419 }
420
421 /*
422 * Now convert the rest of the name
423 */
424 for (i = dp - un, j = 0; un < dp && j < 8; j++) {
425 c = unix2doschr(&un, &i, pmp);
426 if (c & 0xff00) {
427 dn[j] = c >> 8;
428 if (++j < 8) {
429 dn[j] = c;
430 if (conv != 3)
431 conv = 2;
432 continue;
433 } else {
434 conv = 3;
435 dn[j-1] = ' ';
436 break;
437 }
438 } else {
439 dn[j] = c;
440 }
441 if (((dn[j] & 0x80) || *(un - 1) != dn[j]) && conv != 3)
442 conv = 2;
443 if (dn[j] == 1) {
444 conv = 3;
445 dn[j] = '_';
446 }
447 if (dn[j] == 2) {
448 conv = 3;
449 dn[j--] = ' ';
450 }
451 }
452 if (un < dp)
453 conv = 3;
454 /*
455 * If we didn't have any chars in filename,
456 * generate a default
457 */
458 if (!j)
459 dn[0] = '_';
460
461 /*
462 * If there wasn't any char dropped,
463 * there is no place for generation numbers
464 */
465 if (conv != 3) {
466 if (gen > 1)
467 conv = 0;
468 goto done;
469 }
470
471 /*
472 * Now insert the generation number into the filename part
473 */
474 if (gen == 0)
475 goto done;
476 for (wcp = gentext + sizeof(gentext); wcp > gentext && gen; gen /= 10)
477 *--wcp = gen % 10 + '';
478 if (gen) {
479 conv = 0;
480 goto done;
481 }
482 for (i = 8; dn[--i] == ' ';);
483 i++;
484 if (gentext + sizeof(gentext) - wcp + 1 > 8 - i)
485 i = 8 - (gentext + sizeof(gentext) - wcp + 1);
486 /*
487 * Correct posision to where insert the generation number
488 */
489 cp = dn;
490 i -= mbsadjpos((const char**)&cp, i, unlen, 1, pmp->pm_flags, pmp->pm_d2u);
491
492 dn[i++] = '~';
493 while (wcp < gentext + sizeof(gentext))
494 dn[i++] = *wcp++;
495
496 /*
497 * Tail of the filename should be space
498 */
499 while (i < 8)
500 dn[i++] = ' ';
501 conv = 3;
502
503 done:
504 /*
505 * The first character cannot be E5,
506 * because that means a deleted entry
507 */
508 if (dn[0] == 0xe5)
509 dn[0] = SLOT_E5;
510
511 return conv;
512 }
513
514 /*
515 * Create a Win95 long name directory entry
516 * Note: assumes that the filename is valid,
517 * i.e. doesn't consist solely of blanks and dots
518 */
519 int
520 unix2winfn(const u_char *un, size_t unlen, struct winentry *wep, int cnt,
521 int chksum, struct msdosfsmount *pmp)
522 {
523 uint8_t *wcp;
524 int i, end;
525 uint16_t code;
526
527 /*
528 * Drop trailing blanks and dots
529 */
530 unlen = winLenFixup(un, unlen);
531
532 /*
533 * Cut *un for this slot
534 */
535 unlen = mbsadjpos((const char **)&un, unlen, (cnt - 1) * WIN_CHARS, 2,
536 pmp->pm_flags, pmp->pm_u2w);
537
538 /*
539 * Initialize winentry to some useful default
540 */
541 memset(wep, 0xff, sizeof(*wep));
542 wep->weCnt = cnt;
543 wep->weAttributes = ATTR_WIN95;
544 wep->weReserved1 = 0;
545 wep->weChksum = chksum;
546 wep->weReserved2 = 0;
547
548 /*
549 * Now convert the filename parts
550 */
551 end = 0;
552 for (wcp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0 && !end;) {
553 code = unix2winchr(&un, &unlen, 0, pmp);
554 *wcp++ = code;
555 *wcp++ = code >> 8;
556 if (!code)
557 end = WIN_LAST;
558 }
559 for (wcp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0 && !end;) {
560 code = unix2winchr(&un, &unlen, 0, pmp);
561 *wcp++ = code;
562 *wcp++ = code >> 8;
563 if (!code)
564 end = WIN_LAST;
565 }
566 for (wcp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0 && !end;) {
567 code = unix2winchr(&un, &unlen, 0, pmp);
568 *wcp++ = code;
569 *wcp++ = code >> 8;
570 if (!code)
571 end = WIN_LAST;
572 }
573 if (!unlen)
574 end = WIN_LAST;
575 wep->weCnt |= end;
576 return !end;
577 }
578
579 /*
580 * Compare our filename to the one in the Win95 entry
581 * Returns the checksum or -1 if no match
582 */
583 int
584 winChkName(struct mbnambuf *nbp, const u_char *un, size_t unlen, int chksum,
585 struct msdosfsmount *pmp)
586 {
587 size_t len;
588 uint16_t c1, c2;
589 u_char *np;
590 struct dirent dirbuf;
591
592 /*
593 * We already have winentry in *nbp.
594 */
595 if (!mbnambuf_flush(nbp, &dirbuf) || dirbuf.d_namlen == 0)
596 return -1;
597
598 #ifdef MSDOSFS_DEBUG
599 printf("winChkName(): un=%s:%zu,d_name=%s:%d\n", un, unlen,
600 dirbuf.d_name,
601 dirbuf.d_namlen);
602 #endif
603
604 /*
605 * Compare the name parts
606 */
607 len = dirbuf.d_namlen;
608 if (unlen != len)
609 return -2;
610
611 for (np = dirbuf.d_name; unlen > 0 && len > 0;) {
612 /*
613 * Comparison must be case insensitive, because FAT disallows
614 * to look up or create files in case sensitive even when
615 * it's a long file name.
616 */
617 c1 = unix2winchr(__DECONST(const u_char **, &np), &len,
618 LCASE_BASE, pmp);
619 c2 = unix2winchr(&un, &unlen, LCASE_BASE, pmp);
620 if (c1 != c2)
621 return -2;
622 }
623 return chksum;
624 }
625
626 /*
627 * Convert Win95 filename to dirbuf.
628 * Returns the checksum or -1 if impossible
629 */
630 int
631 win2unixfn(struct mbnambuf *nbp, struct winentry *wep, int chksum,
632 struct msdosfsmount *pmp)
633 {
634 u_char *c, tmpbuf[5];
635 uint8_t *cp;
636 uint8_t *np, name[WIN_CHARS * 3 + 1];
637 uint16_t code;
638 int i;
639
640 if ((wep->weCnt&WIN_CNT) > howmany(WIN_MAXLEN, WIN_CHARS)
641 || !(wep->weCnt&WIN_CNT))
642 return -1;
643
644 /*
645 * First compare checksums
646 */
647 if (wep->weCnt&WIN_LAST) {
648 chksum = wep->weChksum;
649 } else if (chksum != wep->weChksum)
650 chksum = -1;
651 if (chksum == -1)
652 return -1;
653
654 /*
655 * Convert the name parts
656 */
657 np = name;
658 for (cp = wep->wePart1, i = sizeof(wep->wePart1)/2; --i >= 0;) {
659 code = (cp[1] << 8) | cp[0];
660 switch (code) {
661 case 0:
662 *np = '\0';
663 if (mbnambuf_write(nbp, name,
664 (wep->weCnt & WIN_CNT) - 1) != 0)
665 return -1;
666 return chksum;
667 case '/':
668 *np = '\0';
669 return -1;
670 default:
671 c = win2unixchr(tmpbuf, code, pmp);
672 while (*c != '\0')
673 *np++ = *c++;
674 break;
675 }
676 cp += 2;
677 }
678 for (cp = wep->wePart2, i = sizeof(wep->wePart2)/2; --i >= 0;) {
679 code = (cp[1] << 8) | cp[0];
680 switch (code) {
681 case 0:
682 *np = '\0';
683 if (mbnambuf_write(nbp, name,
684 (wep->weCnt & WIN_CNT) - 1) != 0)
685 return -1;
686 return chksum;
687 case '/':
688 *np = '\0';
689 return -1;
690 default:
691 c = win2unixchr(tmpbuf, code, pmp);
692 while (*c != '\0')
693 *np++ = *c++;
694 break;
695 }
696 cp += 2;
697 }
698 for (cp = wep->wePart3, i = sizeof(wep->wePart3)/2; --i >= 0;) {
699 code = (cp[1] << 8) | cp[0];
700 switch (code) {
701 case 0:
702 *np = '\0';
703 if (mbnambuf_write(nbp, name,
704 (wep->weCnt & WIN_CNT) - 1) != 0)
705 return -1;
706 return chksum;
707 case '/':
708 *np = '\0';
709 return -1;
710 default:
711 c = win2unixchr(tmpbuf, code, pmp);
712 while (*c != '\0')
713 *np++ = *c++;
714 break;
715 }
716 cp += 2;
717 }
718 *np = '\0';
719 if (mbnambuf_write(nbp, name, (wep->weCnt & WIN_CNT) - 1) != 0)
720 return -1;
721 return chksum;
722 }
723
724 /*
725 * Compute the unrolled checksum of a DOS filename for Win95 LFN use.
726 */
727 uint8_t
728 winChksum(uint8_t *name)
729 {
730 int i;
731 uint8_t s;
732
733 for (s = 0, i = 11; --i >= 0; s += *name++)
734 s = (s << 7)|(s >> 1);
735 return (s);
736 }
737
738 /*
739 * Determine the number of slots necessary for Win95 names
740 */
741 int
742 winSlotCnt(const u_char *un, size_t unlen, struct msdosfsmount *pmp)
743 {
744 size_t wlen;
745 char wn[WIN_MAXLEN * 2 + 1], *wnp;
746
747 unlen = winLenFixup(un, unlen);
748
749 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
750 wlen = WIN_MAXLEN * 2;
751 wnp = wn;
752 msdosfs_iconv->conv(pmp->pm_u2w, (const char **)&un, &unlen, &wnp, &wlen);
753 if (unlen > 0)
754 return 0;
755 return howmany(WIN_MAXLEN - wlen/2, WIN_CHARS);
756 }
757
758 if (unlen > WIN_MAXLEN)
759 return 0;
760 return howmany(unlen, WIN_CHARS);
761 }
762
763 /*
764 * Determine the number of bytes necessary for Win95 names
765 */
766 size_t
767 winLenFixup(const u_char *un, size_t unlen)
768 {
769 for (un += unlen; unlen > 0; unlen--)
770 if (*--un != ' ' && *un != '.')
771 break;
772 return unlen;
773 }
774
775 /*
776 * Store an area with multi byte string instr, and returns left
777 * byte of instr and moves pointer forward. The area's size is
778 * inlen or outlen.
779 */
780 static int
781 mbsadjpos(const char **instr, size_t inlen, size_t outlen, int weight, int flag, void *handle)
782 {
783 char *outp, outstr[outlen * weight + 1];
784
785 if (flag & MSDOSFSMNT_KICONV && msdosfs_iconv) {
786 outp = outstr;
787 outlen *= weight;
788 msdosfs_iconv->conv(handle, instr, &inlen, &outp, &outlen);
789 return (inlen);
790 }
791
792 (*instr) += min(inlen, outlen);
793 return (inlen - min(inlen, outlen));
794 }
795
796 /*
797 * Convert DOS char to Local char
798 */
799 static u_char *
800 dos2unixchr(u_char *outbuf, const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp)
801 {
802 u_char c, *outp;
803 size_t len, olen;
804
805 outp = outbuf;
806 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
807 olen = len = 4;
808
809 if (lower & (LCASE_BASE | LCASE_EXT))
810 msdosfs_iconv->convchr_case(pmp->pm_d2u, (const char **)instr,
811 ilen, (char **)&outp, &olen, KICONV_LOWER);
812 else
813 msdosfs_iconv->convchr(pmp->pm_d2u, (const char **)instr,
814 ilen, (char **)&outp, &olen);
815 len -= olen;
816
817 /*
818 * return '?' if failed to convert
819 */
820 if (len == 0) {
821 (*ilen)--;
822 (*instr)++;
823 *outp++ = '?';
824 }
825 } else {
826 (*ilen)--;
827 c = *(*instr)++;
828 c = dos2unix[c];
829 if (lower & (LCASE_BASE | LCASE_EXT))
830 c = u2l[c];
831 *outp++ = c;
832 outbuf[1] = '\0';
833 }
834
835 *outp = '\0';
836 outp = outbuf;
837 return (outp);
838 }
839
840 /*
841 * Convert Local char to DOS char
842 */
843 static uint16_t
844 unix2doschr(const u_char **instr, size_t *ilen, struct msdosfsmount *pmp)
845 {
846 u_char c;
847 char *up, *outp, unicode[3], outbuf[3];
848 uint16_t wc;
849 size_t len, ucslen, unixlen, olen;
850
851 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
852 /*
853 * to hide an invisible character, using a unicode filter
854 */
855 ucslen = 2;
856 len = *ilen;
857 up = unicode;
858 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr,
859 ilen, &up, &ucslen);
860 unixlen = len - *ilen;
861
862 /*
863 * cannot be converted
864 */
865 if (unixlen == 0) {
866 (*ilen)--;
867 (*instr)++;
868 return (0);
869 }
870
871 /*
872 * return magic number for ascii char
873 */
874 if (unixlen == 1) {
875 c = *(*instr -1);
876 if (! (c & 0x80)) {
877 c = unix2dos[c];
878 if (c <= 2)
879 return (c);
880 }
881 }
882
883 /*
884 * now convert using libiconv
885 */
886 *instr -= unixlen;
887 *ilen = len;
888
889 olen = len = 2;
890 outp = outbuf;
891 msdosfs_iconv->convchr_case(pmp->pm_u2d, (const char **)instr,
892 ilen, &outp, &olen, KICONV_FROM_UPPER);
893 len -= olen;
894
895 /*
896 * cannot be converted, but has unicode char, should return magic number
897 */
898 if (len == 0) {
899 (*ilen) -= unixlen;
900 (*instr) += unixlen;
901 return (1);
902 }
903
904 wc = 0;
905 while(len--)
906 wc |= (*(outp - len - 1) & 0xff) << (len << 3);
907 return (wc);
908 }
909
910 (*ilen)--;
911 c = *(*instr)++;
912 c = l2u[c];
913 c = unix2dos[c];
914 return ((uint16_t)c);
915 }
916
917 /*
918 * Convert Windows char to Local char
919 */
920 static u_char *
921 win2unixchr(u_char *outbuf, uint16_t wc, struct msdosfsmount *pmp)
922 {
923 u_char *inp, *outp, inbuf[3];
924 size_t ilen, olen, len;
925
926 outp = outbuf;
927 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
928 inbuf[0] = (u_char)(wc>>8);
929 inbuf[1] = (u_char)wc;
930 inbuf[2] = '\0';
931
932 ilen = 2;
933 olen = len = 4;
934 inp = inbuf;
935 msdosfs_iconv->convchr(pmp->pm_w2u, __DECONST(const char **,
936 &inp), &ilen, (char **)&outp, &olen);
937 len -= olen;
938
939 /*
940 * return '?' if failed to convert
941 */
942 if (len == 0)
943 *outp++ = '?';
944 } else {
945 *outp++ = (wc & 0xff00) ? '?' : (u_char)(wc & 0xff);
946 }
947
948 *outp = '\0';
949 outp = outbuf;
950 return (outp);
951 }
952
953 /*
954 * Convert Local char to Windows char
955 */
956 static uint16_t
957 unix2winchr(const u_char **instr, size_t *ilen, int lower, struct msdosfsmount *pmp)
958 {
959 u_char *outp, outbuf[3];
960 uint16_t wc;
961 size_t olen;
962
963 if (*ilen == 0)
964 return (0);
965
966 if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) {
967 outp = outbuf;
968 olen = 2;
969 if (lower & (LCASE_BASE | LCASE_EXT))
970 msdosfs_iconv->convchr_case(pmp->pm_u2w, (const char **)instr,
971 ilen, (char **)&outp, &olen,
972 KICONV_FROM_LOWER);
973 else
974 msdosfs_iconv->convchr(pmp->pm_u2w, (const char **)instr,
975 ilen, (char **)&outp, &olen);
976
977 /*
978 * return '' if end of filename
979 */
980 if (olen == 2)
981 return (0);
982
983 wc = (outbuf[0]<<8) | outbuf[1];
984
985 return (wc);
986 }
987
988 (*ilen)--;
989 wc = (*instr)[0];
990 if (lower & (LCASE_BASE | LCASE_EXT))
991 wc = u2l[wc];
992 (*instr)++;
993 return (wc);
994 }
995
996 /*
997 * Initialize the temporary concatenation buffer.
998 */
999 void
1000 mbnambuf_init(struct mbnambuf *nbp)
1001 {
1002
1003 nbp->nb_len = 0;
1004 nbp->nb_last_id = -1;
1005 nbp->nb_buf[sizeof(nbp->nb_buf) - 1] = '\0';
1006 }
1007
1008 /*
1009 * Fill out our concatenation buffer with the given substring, at the offset
1010 * specified by its id. Since this function must be called with ids in
1011 * descending order, we take advantage of the fact that ASCII substrings are
1012 * exactly WIN_CHARS in length. For non-ASCII substrings, we shift all
1013 * previous (i.e. higher id) substrings upwards to make room for this one.
1014 * This only penalizes portions of substrings that contain more than
1015 * WIN_CHARS bytes when they are first encountered.
1016 */
1017 int
1018 mbnambuf_write(struct mbnambuf *nbp, char *name, int id)
1019 {
1020 char *slot;
1021 size_t count, newlen;
1022
1023 if (nbp->nb_len != 0 && id != nbp->nb_last_id - 1) {
1024 #ifdef MSDOSFS_DEBUG
1025 printf("msdosfs: non-decreasing id: id %d, last id %d\n",
1026 id, nbp->nb_last_id);
1027 #endif
1028 return (EINVAL);
1029 }
1030
1031 /* Will store this substring in a WIN_CHARS-aligned slot. */
1032 slot = &nbp->nb_buf[id * WIN_CHARS];
1033 count = strlen(name);
1034 newlen = nbp->nb_len + count;
1035 if (newlen > WIN_MAXLEN || newlen > MAXNAMLEN) {
1036 #ifdef MSDOSFS_DEBUG
1037 printf("msdosfs: file name length %zu too large\n", newlen);
1038 #endif
1039 return (ENAMETOOLONG);
1040 }
1041
1042 /* Shift suffix upwards by the amount length exceeds WIN_CHARS. */
1043 if (count > WIN_CHARS && nbp->nb_len != 0) {
1044 if ((id * WIN_CHARS + count + nbp->nb_len) >
1045 sizeof(nbp->nb_buf))
1046 return (ENAMETOOLONG);
1047
1048 memmove(slot + count, slot + WIN_CHARS, nbp->nb_len);
1049 }
1050
1051 /* Copy in the substring to its slot and update length so far. */
1052 memcpy(slot, name, count);
1053 nbp->nb_len = newlen;
1054 nbp->nb_last_id = id;
1055
1056 return (0);
1057 }
1058
1059 /*
1060 * Take the completed string and use it to setup the struct dirent.
1061 * Be sure to always nul-terminate the d_name and then copy the string
1062 * from our buffer. Note that this function assumes the full string has
1063 * been reassembled in the buffer. If it's called before all substrings
1064 * have been written via mbnambuf_write(), the result will be incorrect.
1065 */
1066 char *
1067 mbnambuf_flush(struct mbnambuf *nbp, struct dirent *dp)
1068 {
1069
1070 if (nbp->nb_len > sizeof(dp->d_name) - 1) {
1071 mbnambuf_init(nbp);
1072 return (NULL);
1073 }
1074 memcpy(dp->d_name, &nbp->nb_buf[0], nbp->nb_len);
1075 dp->d_name[nbp->nb_len] = '\0';
1076 dp->d_namlen = nbp->nb_len;
1077
1078 mbnambuf_init(nbp);
1079 return (dp->d_name);
1080 }
Cache object: b7652e5d598fab6adf8832b83992e16a
|